|Home | About | Journals | Submit | Contact Us | Français|
Chloramphenicol, florfenicol, and thiamphenicol are used as antibacterial drugs in clinical and veterinary medicine. Two efflux pumps of the major facilitator superfamily encoded by the cmlR1 and cmlR2 genes mediate resistance to these antibiotics in Streptomyces coelicolor, a close relative of Mycobacterium tuberculosis. The transcription of both genes was observed by reverse transcription-PCR. Disruption of cmlR1 decreased the chloramphenicol MIC 1.6-fold, while disruption of cmlR2 lowered the MIC 16-fold. The chloramphenicol MIC of wild-type S. coelicolor decreased fourfold and eightfold in the presence of reserpine and Phe-Arg-β-naphthylamide, respectively. These compounds are known to potentiate the activity of some antibacterial drugs via efflux pump inhibition. While reserpine is known to potentiate drug activity against gram-positive bacteria, this is the first time that Phe-Arg-β-naphthylamide has been shown to potentiate drug activity against a gram-positive bacterium.
Membrane-bound efflux pumps commonly underlie drug resistance in pathogenic bacteria (21, 26, 33). These pumps function by actively expelling drugs from the cytosol of bacterial cells, reducing the effective intracellular drug concentration. There are five major families of efflux pumps in bacteria: the ATP binding cassette family, the major facilitator superfamily (MFS), the multidrug and toxic compound extrusion family, the resistance nodulation division family, and the small multidrug resistance family (33). Some efflux pumps exhibit high specificity for certain antimicrobial agents, while others act upon drugs from unrelated structural classes and thereby confer multidrug resistance (33). In both gram-positive and gram-negative bacteria, genes encoding efflux pumps are found on the chromosome and on plasmids (33). Multidrug resistance in human pathogens is often correlated with the overexpression of efflux pump genes (18, 34).
Efflux is a nondestructive mechanism of antibiotic resistance commonly observed in antibiotic-producing bacteria that provides self-resistance without compromising the biological activity of the biosynthesized antibiotics (9). Given that antibiotic-producing organisms are a reservoir of resistance genes (23), they are a likely source of the drug efflux pump genes found in pathogenic bacteria. Streptomyces bacteria produce two-thirds of the clinically used antibiotics and often have an efflux pump gene associated with self-resistance (1, 17, 31). In some cases, these gram-positive, soil-dwelling bacteria have efflux pumps that confer resistance to multiple antibiotics. For instance, the ptr gene in Streptomyces pristinaespiralis confers resistance to the pristinamycins and rifampin (rifampicin) (3). Likewise, an ATP binding cassette family transporter gene in Streptomyces rochei F20 confers resistance to macrolides (i.e., oleandomycin, erythromycin, and spiramycin), tetracycline, and doxorubicin (11). Although most Streptomyces bacteria are not pathogenic, they are of clinical interest because they are close relatives of pathogenic mycobacteria and are considered to be a reservoir of resistance genes (1, 9).
In this study, we examined efflux pump-mediated drug resistance in Streptomyces coelicolor, the model organism of the Streptomyces genus (1). S. coelicolor has an extensive array of efflux pump genes proposed to confer high-level resistance to a range of drug-related compounds that it does not produce, including macrolides, streptogramins, fosmidomycin, and chloramphenicol (1; http://streptomyces.org.uk; http://www.membranetransport.org/transporter2.php?oOID=scoe1). Presumably, S. coelicolor acquired these genes by horizontal transfer. We have established that two of the MFS efflux pump genes in S. coelicolor confer resistance to chloramphenicol, thiamphenicol, and florfenicol. These genes, cmlR1 and cmlR2, have low levels of homology to one another, yet each gene independently contributes to the chloramphenicol resistance of S. coelicolor. We also found that two known efflux pump inhibitors, reserpine and Phe-Arg-β-naphthylamide, significantly increased the susceptibility of S. coelicolor to chloramphenicol. While reserpine is known to potentiate drug activity against gram-positive bacteria, this is the first time that Phe-Arg-β-naphthylamide has been shown to potentiate the activity of a drug against a gram-positive bacterium.
The S. coelicolor strains used in this work are provided in Table Table1.1. The S. coelicolor strains were grown at 30°C on mannitol soy flour medium (SFM), Difco nutrient agar medium, yeast extract-malt extract medium, or minimal liquid medium (NMMP) (16). SFM was used for conjugations between S. coelicolor and Escherichia coli and for generating spore stocks. Escherichia coli strains DH5α and ET12567/pUZ8002 were grown on Luria-Bertani medium at 37°C for routine subcloning (29), and E. coli strain BW25113/pIJ790 was grown on Luria-Bertani medium at 30°C when selection for pIJ790 was maintained. For transcriptional analysis, S. coelicolor was grown in NMMP. For the selection of E. coli, ampicillin, apramycin, chloramphenicol, hygromycin, and kanamycin were employed at 100, 50, 25, 80, and 50 μg/ml, respectively. Nalidixic acid was used at 20 μg/ml to counterselect E. coli in conjugations with S. coelicolor. Apramycin and hygromycin were used at 50 μg/ml for the selection of S. coelicolor exconjugants.
Standard cloning procedures (29) were employed to generate the plasmids listed in Table Table1.1. The site-specific integrating vector pMS81 (12) was used for genetic complementation (32). pIJ10257, derived from pMS81, was the vector used for the overexpression of the efflux pump genes in S. coelicolor (14). DNA sequencing was performed by Davis Sequencing (Davis, CA). All primers used in the work were synthesized by Invitrogen. PCR was performed with Taq (Invitrogen) and Pfu (Stratagene, Agilent Technologies). All PCRs were performed with 5% (vol/vol) dimethyl sulfoxide (16).
PCR-targeted mutagenesis was used to replace the S. coelicolor SCO7526 (cmlR1) and SCO7662 (cmlR2) genes with an apramycin resistance marker, apr (13). The requisite PCR products were amplified from the apramycin resistance cassette of pIJ773 with primers 5 to 8 (Table (Table2).2). The PCR products were introduced into E. coli BW25113/pIJ790 harboring either cosmid St8G12 (cmlR1) or cosmid St10F4 (cmlR2) and expressing bacteriophage λ RED recombinase. The resultant recombinant cosmids, St8G12 ΔcmlR1::apr and St10F4 ΔcmlR2::apr, were introduced into E. coli strain ET12567/pUZ8002 and then into wild-type strain S. coelicolor M600 (16) by way of conjugation, as described previously (15). Exconjugants lacking the cmlR1 and cmlR2 genes were identified by selection for apramycin resistance and kanamycin sensitivity. Gene replacements in the cmlR1-null strain (strain B754) and in the cmlR2-null strain (strain B756) were confirmed by PCR analysis of genomic DNA isolated from the null mutants with primers 9 to 12 (Table (Table22).
A 1,724-bp DNA fragment containing cmlR1 and its promoter region was excised from S. coelicolor cosmid St8G12 by restriction digestion with SfcI and SmlI. This fragment contained the cmlR1 open reading frame (ORF) with 430 upstream base pairs. The excised fragment was treated with DNA polymerase I large (Klenow) fragment (New England Biolabs), according to the manufacturer's protocol, and ligated into the SmaI site of pBluescript II KS+. The fragment was then excised from pBluescript II KS+ with SpeI and EcoRV and ligated into pMS81 pretreated with the same enzymes to yield pJS331. pJS331 was introduced by transformation into nonmethylating E. coli strain ET12567/pUZ8002 and then into the S. coelicolor cmlR1-null mutant via conjugation.
A 1,882-bp DNA fragment containing cmlR2 and its promoter region was excised from S. coelicolor cosmid St10F4 by restriction digestion with SphI. This fragment contained the cmlR2 ORF with 579 upstream base pairs. The excised SphI fragment was treated with DNA polymerase I large (Klenow) fragment (New England Biolabs), according to the manufacturer's protocol, and ligated into the SmaI site of pBluescript II KS+. The fragment was then excised from pBluescript II KS+ and ligated into pMS81 to yield pJS332, which was introduced into the S. coelicolor cmlR2-null mutant as described above.
The cmlR1 and cmlR2 ORFs were amplified from wild-type S. coelicolor genomic DNA with primers 1 to 4 (Table (Table2).2). The PCR products were cloned into pBluescript II KS+ and sequenced to confirm their identities. The genes were transferred into the integrative, constitutive expression vector pIJ10257 (14), yielding pJS333 (cmlR1) and pJS334 (cmlR2). The resultant plasmids were introduced into E. coli strain ET12567/pUZ8002 and then into the chloramphenicol-sensitive S. coelicolor cmlR2-null strain (strain B756) by way of conjugation, as described previously (15). Exconjugants with either pJS333 (strain B758) or pJS334 (strain B759) were identified by selection for hygromycin resistance.
Chloramphenicol, florfenicol, thiamphenicol, reserpine, and Phe-Arg-β-naphthylamide were all purchased from Sigma-Aldrich.
All MIC assays were performed on Difco nutrient agar medium supplemented with the indicated concentrations of each antimicrobial agent. Growth was assessed after incubation at 30°C for 48 h.
Transcriptional analyses of the S. coelicolor SCO7526 and SCO7662 genes were performed as described previously (32). Two NMMP cultures were inoculated with 2 × 109 S. coelicolor spores to test for transcription in the presence and the absence of chloramphenicol. One culture was grown for 21 h prior to treatment with 50 μg/ml chloramphenicol for 3 h. The cells were then washed once with 10.3% (wt/vol) aqueous sucrose; resuspended in lysis buffer consisting of 50 mM Tris-HCl, pH 8.0-1 mM EDTA supplemented with 10 mg/ml lysozyme (Sigma-Aldrich); and incubated for 15 min at 30°C. Total RNA was then isolated with an RNeasy minikit (Qiagen), according to the manufacturer's protocol. The concentration of each purified, DNA-free total RNA isolate was measured with a NanoDrop ND-1000 spectrophotometer. An equal quantity of total RNA (1.8 μg) was employed in all reverse transcription-PCRs (RT-PCRs). RT-PCR was performed with a OneStep RT-PCR kit (Qiagen), according to the manufacturer's protocol for transcripts with high GC contents (by the use of Q solution). Primers 13 to 16 (Table (Table2)2) were used for RT-PCR analyses. The PCR program used for the detection of both transcripts was 50°C for 30 min; 95°C for 15 min; 35 cycles of 94°C for 30 s, 58°C for 30 s, and 72°C for 40 s; and a final elongation at 72°C for 10 min. No signals were detected in control experiments with Pfu polymerase, indicating that the RT-PCR signals correspond to the amplification of transcripts.
S. coelicolor is highly resistant to chloramphenicol (MIC, 80 μg/ml) (Table (Table3).3). Bioinformatic analysis of the S. coelicolor genome sequence revealed two putative MFS chloramphenicol efflux pump genes, cmlR1 (SCO7526) and cmlR2 (SCO7662) (1). These genes are separated by more than 100 kb on the chromosome and encode proteins that are 40% identical in amino acid sequence. The products of cmlR1 and cmlR2 are related to the MFS chloramphenicol exporters CmlA in E. coli and FloR in Salmonella enterica (4, 5). Interestingly, Mycobacterium tuberculosis has an uncharacterized MFS protein that is also homologous to the products of cmlR1 and cmlR2 (7).
Initially, we used RT-PCR to assess the transcriptional activity of these genes. Both cmlR1 and cmlR2 were constitutively transcribed; the transcripts were detected in both the presence and the absence of chloramphenicol (Fig. (Fig.1).1). To validate their functions and determine their relative contributions to chloramphenicol resistance, we used a PCR-targeting procedure to replace them with an apramycin resistance marker in wild-type strain S. coelicolor M600 (13). The gene replacements were confirmed by PCR analysis of genomic DNA isolated from the null strains. Both null strains, S. coelicolor B754 cmlR1::apr and S. coelicolor B756 cmlR2::apr, were grown on Difco nutrient agar to assess their chloramphenicol susceptibilities. The chloramphenicol MIC of the strain lacking cmlR1 (50 μg/ml) was 1.6-fold lower than that of the wild-type strain, and the MIC of the strain lacking cmlR2 (5 μg/ml) was 16-fold lower (Table (Table3).3). These results indicate that cmlR2 encodes the major chloramphenicol resistance determinant in S. coelicolor. In both null strains, chloramphenicol resistance was restored by genetic complementation in trans (strains B755 and B757 in Table Table1).1). To further validate the conclusion that cmlR1 is a minor contributor to chloramphenicol resistance, we generated strains B758 and B759 in which either the cmlR1 gene or the cmlR2 gene, respectively, was overexpressed in the chloramphenicol-sensitive cmlR2-null strain (Table (Table1).1). The overexpression of the cmlR1 gene increased the chloramphenicol MIC from 5 μg/ml to 10 μg/ml (a 2-fold increase), while the overexpression of the cmlR2 gene increased the MIC from 5 μg/ml to 70 μg/ml (a 14-fold increase).
The null strains were also tested for their susceptibilities to florfenicol and thiamphenicol. The florfenicol MIC of the wild-type S. coelicolor strain was 30 μg/ml (Table (Table3).3). Strains lacking cmlR1 and cmlR2 exhibited 1.8-fold and 3-fold florfenicol MIC reductions, respectively (Table (Table3).3). In contrast, the thiamphenicol MICs of wild-type S. coelicolor and the strain lacking cmlR1 were both found to be greater than 500 μg/ml (Table (Table3).3). The thiamphenicol MIC of the strain lacking cmlR2 was reduced 10-fold compared to that of wild-type S. coelicolor (Table (Table3).3). These results indicate that cmlR2, and not cmlR1, is a thiamphenicol resistance determinant. Thiamphenicol and florfenicol resistance was restored to the null strains by genetic complementation in trans.
Reserpine is a plant alkaloid known to potentiate the activities of fluoroquinolones and tetracycline in gram-positive bacteria by inhibiting multidrug efflux pumps (6, 8, 24, 30). In light of the ability of reserpine to potentiate drug activity in gram-positive bacteria, we assessed the chloramphenicol susceptibilities of wild-type S. coelicolor and the cmlR1-null and cmlR2-null strains over a range of reserpine concentrations (Table (Table4).4). Wild-type S. coelicolor was rendered more susceptible to chloramphenicol at all reserpine concentrations tested, but the most significant potentiation of chloramphenicol activity (i.e., fourfold) was observed at reserpine concentrations at or above 50 μg/ml (Table (Table4).4). In contrast, the activity of chloramphenicol against the cmlR1- and cmlR2-null strains was potentiated less than fourfold at all reserpine concentrations tested (Table (Table4).4). Clearly, potentiation was dependent on both the dose of reserpine and the genetic background of the strain being tested. In control experiments, reserpine alone had no apparent toxicity to S. coelicolor. In fact, the reserpine MIC of S. coelicolor could not be determined at concentrations below 250 μg/ml.
Phe-Arg-β-naphthylamide is a synthetic C-capped dipeptide known to potentiate drug activity in gram-negative bacteria via inhibition of the resistance nodulation division family of drug efflux pumps (2, 19-21, 25, 27). It has been reported that E. coli strains harboring multidrug efflux pumps are more susceptible to chloramphenicol in the presence of Phe-Arg-β-naphthylamide (28). On the basis of those observations, we sought to determine if this compound would increase the chloramphenicol susceptibilities of wild-type S. coelicolor and the cmlR1- and cmlR2-null strains (Table (Table5).5). The chloramphenicol susceptibility of wild-type S. coelicolor increased in proportion to the concentration of Phe-Arg-β-naphthylamide, with an eightfold potentiation observed at 75 μg/ml (Table (Table5).5). In the presence of Phe-Arg-β-naphthylamide, chloramphenicol activity was potentiated to a lesser degree in the cmlR1- and cmlR2-null strains than in the wild-type strain (Table (Table5).5). As was the case for reserpine, the potentiation of chloramphenicol activity by Phe-Arg-β-naphthylamide was dependent on both the compound dose and the genetic background of the strain being tested. In control experiments, Phe-Arg-β-naphthylamide alone had no apparent toxicity to S. coelicolor. The Phe-Arg-β-naphthylamide MIC of S. coelicolor was 200 μg/ml.
Bacterial drug efflux pumps have the potential to critically compromise the efficacies of many antibacterial drugs in use today (33). In many cases, the genes encoding efflux pumps are found on plasmids that are easily transferred from one bacterium to another. Certain bacteria have multiple genes encoding efflux pumps that mediate multidrug resistance. S. coelicolor is a multidrug-resistant bacterium (9, 14, 16, 32) with a large number of putative drug efflux pumps (1). The mechanism of chloramphenicol resistance described here is particularly interesting because S. coelicolor has two distinct MFS efflux pumps that confer chloramphenicol resistance, yet it does not produce the antibiotic. Transcriptional analyses clearly indicate that both genes are constitutively transcribed. Using reverse genetic analysis, we demonstrated that the cmlR2 gene is the major contributor to chloramphenicol resistance. A close ortholog of cmlR2 has been observed in Streptomyces lividans (10). It is intriguing that S. coelicolor has retained and simultaneously transcribes two antibiotic resistance genes with redundant functions.
We also found that S. coelicolor was sensitized to chloramphenicol in the presence of reserpine and Phe-Arg-β-naphthylamide, two well-known efflux pump inhibitors from distinct structural classes (21, 22). Interestingly, Phe-Arg-β-naphthylamide has never been reported to enhance the antibacterial susceptibility of a gram-positive bacterium. It is also noteworthy that the potentiation of chloramphenicol activity by reserpine and Phe-Arg-β-naphthylamide was diminished in S. coelicolor strains lacking either the cmlR1 or the cmlR2 gene. This observation suggests that inhibition of CmlR1 and CmlR2 may be the underlying mechanism of potentiation. In any case, our findings provide further evidence that efflux pump inhibitors can be useful against multidrug-resistant, gram-positive bacteria (8).
Brown University is gratefully acknowledged for financial support. J.J.V. was supported by a National Science Foundation EPSCoR graduate fellowship. B.A. was supported by a summer research fellowship sponsored by the Leadership Alliance and by grant NHLBI, R25 HL08892.
Published ahead of print on 17 August 2009.