Daunorubicin efflux in Streptomyces peucetius modulates biosynthesis by feedback regulation


  • Editor: Jose Gil

Correspondence: Ranjan Prasad, Department of Genetic Engineering, School of Biotechnology, Madurai Kamaraj University, Madurai 625 021, India. Tel.: +91 452 245 9115; fax: +91 452 245 9105; e-mail: ranjan_mku@yahoo.com


Streptomyces peucetius self-resistance genes drrA and drrB encode membrane-associated proteins that function like an ABC transporter for the efflux of daunorubicin and to maintain a constant subinhibitory physiological concentration of the drug within the cell. In this study, the drrA and drrB operons were disrupted for investigating drug production, self-resistance and regulation. The drrAdrrB null mutant was highly sensitive to daunorubicin. A 10-fold decrease in drug production was observed in the null mutant compared with the wild-type strain. We propose that the absence of a drug-specific efflux pump increases the intracellular concentration of daunorubicin, which is sensed by the organism to turn down drug production. Quantitative real-time PCR analysis of the mutant showed a drastic reduction in the expression of the key regulator dnrI and polyketide synthase gene dpsA. However, the expression of regulatory genes dnrO and dnrN was increased. Feedback regulation based on the intracellular daunorubicin concentration is discussed.


Streptomyces are soil bacteria that undergo morphological and physiological differentiation and produce many secondary metabolites in parallel (Bibb, 2005). Streptomyces peucetius produces daunorubicin (DNR) and its hydroxy derivative doxorubicin (DXR), which are anthracycline antibiotics used for cancer chemotherapy (Arcamone, 1981). DNR/DXR biosynthetic genes are located as a cluster in the bacterial chromosome. Additionally, genes that activate antibiotic synthesis as well as those that confer resistance against DNR are also found within the cluster (Piepersberg, 1997). Self-resistance is an important requirement for antibiotic-producing microorganisms and is mediated by drug inactivation, target site modification, reduction of the intracellular concentration via efflux and sequestration of drug by the formation of a protein–drug complex (Hopwood, 2007). A feed forward mechanism has been proposed in Streptomyces coelicolor, where a prodrug signals the cell to prepare for an efflux of drug that would accumulate at a later stage of growth (Tahlan et al., 2007). Multiple modes of self-protection against a single antibiotic are well documented (Cundliffe, 1992), often with one or more resistance determinants located adjacent to antibiotic biosynthetic genes.

Earlier studies have shown that self-resistance to DNR in S. peucetius occurs by the action of drrA, drrB (Guilfoile & Hutchinson, 1991) and drrC (Lomovskaya et al., 1996) genes. Adjacently placed drrA and drrB genes encode DrrAB proteins that belong to the ABC family of membrane transporters. DrrA is a peripheral membrane protein and DrrB is an integral membrane protein of 36 and 30 kDa, respectively (Kaur, 1997). They function together as a complex that may consist of two subunits of DrrA and two subunits of DrrB to efflux DNR (Kaur & Russell, 1998). The drrA and drrB genes have overlapping stop and start codons that are translationally coupled. Furthermore, it was observed that a functional complex could be achieved only when the genes were maintained in cis and in a translationally coupled manner (Pradhan et al., 2009). The drrC gene encodes a 764 amino acid protein that possibly inhibits or destabilizes the binding of DNR to genomic DNA (Lomovskaya et al., 1996).

DNR biosynthesis in S. peucetius is regulated by three sequentially activated transcriptional regulators dnrN, dnrO and dnrI. The dnrO gene is located adjacent to dnrN and is divergently transcribed. The DnrN protein binds specifically to the dnrI promoter region (Furuya & Hutchinson, 1996) and activates the transcription of the dnrI gene (Otten et al., 1995). DnrI activator protein binds to promoter elements of biosynthetic and resistance genes to turn them on. (Madduri & Hutchinson, 1995). DNR inhibits binding of DnrN to the dnrI promoter region. The dnrO gene encodes a DNA-binding protein that binds specifically to the dnrN/dnrO promoter region and activates dnrN (Otten et al., 2000). DnrO is an activator/repressor that activates dnrN and represses its own transcription. Repression is relieved in the presence of drug intermediate rhodomycin (Jiang & Hutchinson, 2006). Disruption of any regulatory gene leads to complete cessation of DNR production.

In this study, simultaneous targeted disruption of drrA and drrB was performed to obtain a null mutant strain with a low self-resistance and drug production. Quantitative real-time (qRT)-PCR was carried out to understand the negative feedback regulation activated by the disruption of a specific antibiotic efflux pump. Feedback inhibition of antibiotic biosynthesis by DNR discussed in earlier studies is revisited and supported by our new findings.

Materials and methods

Biochemicals and reagents

Taq DNA polymerase, DNR, fine chemicals and oligonucleotide primers were purchased from Sigma Aldrich Chemicals Pvt Ltd, India. Antarctic alkaline phosphatase was purchased from New England Biolabs Inc. Culture media components were obtained from HiMedia Laboratories Pvt Ltd, India. Restriction enzymes, T4 DNA ligase and polynucleotide kinase were purchased from Promega. Other analytical-grade laboratory reagents were procured from standard commercial sources.

Bacterial strains and plasmids

Strain plasmids and genes with accession numbers used in the study are provided as Supporting Information, Table S1.

Construction of the drrAdrrB disruption cassette

The drrAdrrB disruption cassette consists of a left homologous box and a right homologous box and the apramycin gene acc(3)IV in between. The three elements were obtained by amplification from appropriate templates and assembled into a cassette by ligation. The construction of the drrAdrrB suicide plasmid is schematically represented in Fig. 1. The left homologous box was amplified as 555 bp DNA using the KpnI forward adapter primer (RASKF 5′-TAATGGTACCGTGAACACGCAGCCGAC) and the EcoRI reverse adapter primer (RADER 5′-GACAGAATTCCAGAGCCCGCACGATG). This DNA includes sequences downstream of the drrA start codon. The left homologous box was cloned in pBluescript SK (Stratagene) and named as pSKA2. Apramycin resistance gene acc(3)IV was amplified from the pSET152 template with the XbaI forward adapter primer (accF 5′-GCCGTCTAGAGTTTATCACCACCGACTATTTGC) and the EcoRI reverse adapter primer (accR 5′-ATACGAATTCAGCGTCTGCTCCGCCATTC). This was ligated to pSKA2 restricted with XbaI–EcoRI to place it next to the left homologous box and named pSKA2Apr. The right homologous box was amplified as 456 bp DNA using the XbaI forward adapter primer (RBDXF 5′-CGCTCTAGAGGCAGTCTCCTCGGTG) and the SacII reverse adapter primer (RBESR 5′-ATTATTCCGCGGTCAGTGGGCGTTCTTG). This DNA includes sequences upstream of the drrB stop codon. The right homologous box was cloned in pSKA2Apr to place it next to the acc(3)IV gene and the clone named as pABDD. The gene disruption cassette comprising the left box, the apramycin marker and the right box was subcloned in pSET151 (Bierman et al., 1992) as the HindIII–BamHI fragment. For this, compatible ends were generated by PCR amplification utilizing a pABDD template and adapter primers. The resultant disruption construct pSETDD was transformed into Escherichia coli ET12567 (MacNeil et al., 1992).

Figure 1.

 Construction of the pSETDD–drrAB suicide plasmid.

Gene disruption

Disruption plasmid pSETDD carrying the RK2 oriT was mobilized from E. coli ET12567 (carries pUZ8002 with transfer functions) to S. peucetius using a protocol described by Kieser et al. (1998). A conjugation agar plate was incubated at 30 °C for 18 h and overlaid with a 1 mL solution of 0.1% apramycin and 0.05% nalidixic acid. Exconjugant colonies appeared after further incubation for 5 days at 30 °C. The colonies were replica patched on SMA (2% soyabean meal and 2% mannitol) agar to check for apramycin resistance and thiostrepton sensitivity. Double homologous recombination would result in the loss of the plasmid marker (thiostrepton) and the cell gains apramycin resistance by site-specific chromosomal integration. The transfer plasmid lacks ori for replication in Streptomyces and therefore it cannot survive as a free plasmid. Apramycin-resistant and thiostrepton-sensitive colonies were propagated further.

PCR-based confirmation of gene disruption

PCR analysis was performed with a genomic template of the drrAdrrB null mutant. The forward primer anneals 282-bp upstream of the drrA start codon and the reverse primer to the internal region of the apramycin gene. Streptomyces peucetius wild type (WT) served as a negative control. The PCR product was sequenced and the chromosome–apramycin junction was confirmed on the drrA side of integration.

Cloning of junction fragments

Genomic DNA of the drrAdrrB null mutant was cut with BamHI and ligated to the dephosphorylated BamHI ends of pBluescript SK. Escherichia coli cells transformed with ligated DNA were selected on ampicillin and apramycin (positive selection) plates. This recombinant clone pRESAB was sequenced with appropriate primers (Genetic Analyzer ABI310) to confirm the presence of the chromosome–marker junction sequence on the drrB side of integration. Digestion of pRESAB with XbaI–BamHI released a 2.1-kb fragment comprising a drrB carboxy end and the adjacent drrD/dnrW gene. This was cloned in pOK12 (Vieira & Messing, 1991) and sequenced with M13 forward and reverse primers.

HPLC analysis of DNR

The medium used for the study was prepared as described previously (Dekleva et al., 1985). Single-colony S. peucetius was inoculated into 25 mL nitrate defined medium with 0.5% yeast extract and grown for 36 h at 30 °C, 180 r.p.m. Mycelia were collected by centrifugation at 2000 g for 20 min at 4 °C. One gram wet weight of mycelia was inoculated in 100 mL of nitrate defined medium (NDM) with 5% maltose as the carbon source and grown for 120 h at 30 °C. Anthracylines were extracted and analyzed by HPLC (Shimadzu, Japan) as described earlier (Bartel et al., 1990). A C18 reverse-phase octadecyl column (Shimadzu) was used. The mobile phase was 65% methanol and 35% phosphorylated water, pH 2.0. DNR (Sigma Aldrich, Bangalore, India) was used as the standard. HPLC was set at a flow rate of 1 mL min−1 and A254 nm was measured. A series of dilutions were analyzed by HPLC to construct the standard graph. DNR levels were estimated based on the peak area of the DNR standard.

Antibiotic resistance assay

The drrAdrrB null mutant and WT cells were tested for levels of resistance to DNR in the culture medium. R2YE plates with 0, 1, 2, 4, 6, 8 and 10 μg mL−1 DNR were prepared. The cells were grown in NDM liquid for 120 h and 10 μL of the culture was placed on an agar surface. Plates were incubated for 90 h and photographed to record growth inhibition.


Total RNA was prepared using the RNeasy Plant Mini Kit (Qiagen) according to the instructions of the manufacturer. The RNA was treated with Turbo DNAse (Ambion) according to the manufacturer's instructions. RNA was quantified using a Nanodrop ND-1000 spectrophotometer, and the quality of RNA was analyzed on an agarose gel as described by Kieser et al. (1998).

In a 10-μL reaction, 1 μg RNA, 1 mM dNTP mix and 250 ng of random hexamer (Promega) were heated to 80 °C for 5 min and rapidly chilled on ice. Two hundred units of M-MLV reverse transcriptase and 20 U Rnasin (Sigma Aldrich) were added and the volume was made up to 20 μL. The mixture was incubated at 37 °C for 60 min; the reaction was stopped by heating at 90 °C for 5 min. Control reactions were carried out without reverse transcriptase.

qRT-PCR of the selected genes was performed using the Power SYBR Green PCR Master Mix (Applied Biosystems) in an ABI PRISM 7000 Sequence Detection system (Applied Biosystems). All primers were designed using perlprimer (Marshall, 2004). The oligonucleotide sequences of the primers used in this study are listed in Table 1. 16S rRNA gene was used as an endogenous control. Fifty picograms of cDNA from both the WT and the mutant was used for analysis. Real-time PCR conditions were as follows: 94 °C for 10 min, 50 cycles of 94 °C for 30 s, 60 °C for 30 s and 72 °C for 30 s. The reactions were subjected to melting-curve analysis to confirm that a single DNA PCR product was prepared from the cDNA template. The amplification was performed in duplicate or in triplicate wells. For each sample analyzed, reverse transcriptase without controls and nontemplate controls were performed.

Table 1.   Primers used for real-time PCR
16S rRNA gene
 Forward primerReverse primer
 Forward primerReverse primer
 Forward primerReverse primer
 Forward primerReverse primer
 Forward primerReverse primer

After PCR amplifications, the threshold cycle (CT) was calculated using abi prism 7000 sds software (Applied Biosystems). The target gene mRNA levels were normalized internally to the level of 16S rRNA gene. ΔΔCT values and SD were calculated from experimental replicates (Table S2). The S. peucetius transcript was considered as 1.0 for comparison with the null mutant for each of the genes analyzed. Serial dilution of the cDNA was subjected to real-time PCR for all the genes tested. For each transcript, plots of the log dilution factor against the ΔCTCT target−ΔCT 16S rRNA gene) values provided an estimate of the efficiency of the amplification. The relative quantification of gene expression was performed as described in section VII of ‘Guide to performing relative quantification of gene expression using Real-Time quantitative PCR’ (Applied Biosystems).

Results and discussion

Inactivation of drrA and drrB genes

Targeted disruption was performed by the insertion of the apramycin resistance marker gene that replaced 830 bp out of 1841 bp of drrA and drrB coding sequences. Apramycin-based disruption plasmid pSETDD can be delivered to Streptomyces from E. coli. The plasmid's marker gene confers resistance for thiostrepton and lacks ori for replication in Streptomyces. The recipient cell can only survive when single crossover occurs, in which case the whole plasmid integrates along with the disruption cassette. In the event of recombination occurring on either side of the apramycin gene, the likely result is the disruption of drrAdrrB and the simultaneous loss of the transfer plasmid backbone. In the present study, two thiostrepton-sensitive apramycin-resistant colonies out of 24 thiostrepton- and apramycin-resistant colonies were obtained following the introduction of pSETDD into S. peucetius. Genuine double-crossover disruption was tested by amplification of the junction sequence using a primer that anneals to the apramycin resistance gene sequence and the other annealing to the chromosomal sequence. The amplified 1.1 kb DNA (Fig. 2b) was sequenced and the data confirm the appropriate left junction region. To confirm the right junction sequence, genomic DNA was cut with BamHI and ligated to pBluescript SK. Cells transformed with ligated DNA were selected on an apramycin plate (positive selection). There is no BamHI site in the apramycin resistance gene and the next site is in the chromosome at a considerable distance from the cassette sequence. In this way, the junction region along with the neighboring drrD/dnrW (Lomovskaya et al., 1998) could be cloned. The resulting plasmid pRESAB (Fig. 2c) was used as a template to sequence the right junction between chromosome and acc(3)IV utilizing appropriate primers. A 2.1-kb fragment from pRESAB was subcloned in pOK12 and the presence of the drrD gene was confirmed by sequencing. The above experiments confirmed the disruption of drrAdrrB in the S. peucetius chromosome.

Figure 2.

 Confirmation of drrA and drrB disruption. (a) The disrupted region in the Streptomyces peucetius chromosome with adjacent dnrF and drrD/dnrW on either side. The horizontal solid arrows indicate primers that anneal to dnrF on the left side and acc(3)IV on the right to amplify a 1.1 kb DNA. (b) PCR amplification, lane 1, S. peucetius; 2, amplified DNA from mutant 1.1 kb; 3, 1-kb ladder. (c) BamHI fragment ligated to pBluescript SK to obtain pRESAB, which was sequenced to confirm adjacent chromosomal regions. Sequenced regions are shown as dotted lines.

Streptomyces peucetius drrA and drrB genes encode an ABC transporter for efflux of DNR to maintain a constant subinhibitory physiological concentration of the drug within the cell. DrrA is a peripheral membrane protein that binds ATP in a DNR-dependent manner and DrrB is a membrane-localized transporter that effluxes DNR from the cell (Kaur & Russell, 1998). Disruption of drrAdrrB was not lethal to the cell unlike the disruption of drrC (Lomovskaya et al., 1996). Mutation of the mtrA gene in mitramycin-producing Streptomyces argillaceus was lethal, suggesting that the efflux pump was essential for survival in that case (Fernández et al., 1996). A lethal effect or a severe reduction in the viability of the drrAdrrB null mutant is expected in the absence of a specific DNR efflux system. In contrast, disruption of drrAdrrB genes did not affect the growth of the cells as evident by the fact that mutant cell density was greater by 1.5-fold compared with WT in a 100 mL NDM for 120 h (Table 2). Therefore, it is likely that S. peucetius senses intracellular drug levels and turns up/down biosynthesis accordingly. An alternative low-efficiency efflux system may operate to efflux DNR that is produced at a low level in the mutant.

Table 2.   Comparison of the drrAdrrB null mutant and Streptomyces peucetius WT for DNR production and mycelial wet weight
ExperimentdrrAdrrB mutantS. peucetius WT
(μg mL−1)
Mycelial wet weight
(g L−1)
(μg mL−1)
Mycelial wet
weight (g L−1)
Mean89.2 ± 15.640.06 ± 2.3804.0 ± 95.227.2 ± 1.7

Disruption of the drrAB gene affects DNR production and self-resistance

Although the drrAdrrB mutation was not lethal to the cell, it was considerably more sensitive to DNR added externally in the culture medium. A sensitive plate assay was performed to determine the maximum concentration of DNR tolerated by WT and the drrAdrrB null mutant. The maximum DNR concentration at which WT can grow is somewhere between 20 and 25 μg mL−1 (data not shown) and that for the mutant is between 4 and 6 μg mL−1 (Fig. 3). This implies that drrA- and drrB-mediated resistance is a major mechanism by which the producing organism survives the toxic effects of DNR. Estimation of DNR production by HPLC analysis showed that the mutant produced 10 times less DNR than WT per unit volume of liquid culture (Table 2). This observation suggests that inhibition of efflux limits drug production and a feedback inhibition operates in S. peucetius, which is governed by intracellular drug levels. The positive activator of DNR biosynthesis is the DnrI, which is activated by DnrN by binding to an upstream element near the promoter (Furuya & Hutchinson, 1996). This sequence is also a preferential DNR-intercalating site where a mutually exclusive competitive binding of DNR and DnrN occurs. This may be the mechanism that senses the intracellular DNR level to either turn on or turn off the expression of DnrI, which is the key activator for DNR biosynthesis. This study shows the circular nature of regulation, where three elements namely the DnrI activator, the DrrA–DrrB efflux pump and DNR are acting in sequence. At a steady-state level of antibiotic production, DnrI activates the drrAdrrB operon as well as major biosynthetic operons. The efflux system maintains the intracellular DNR at an optimum concentration, and a micro increase in the intracellular DNR level leads to preferential intercalation at the DnrN-binding site that shuts down dnrI transcription temporarily. The intercalated drug must leave the site before DnrN can bind and reactivate dnrI, which is possibly affected by DrrC (Lomovskaya et al., 1996). Yet another regulation is by the control of DnrN expression, which is dictated by its activator DnrO that binds at the upstream element near the dnrN promoter. This site is also a preferential intercalating site for DNR (Otten et al., 2000). These combined factors possibly fine tune the feedback regulation of drug biosynthesis.

Figure 3.

 Analysis of resistance to externally added DNR in gradient plates. A, Streptomyces peucetius and B, the drrAdrrB mutant were plated in an R2YE agar medium containing a series of DNR concentrations. Plates 1–7 contain 0, 1, 2, 4, 6, 8 and 10 μg mL−1 DNR.

Mutation in the drrAB efflux pump affects the regulatory genes

We analyzed the effect of the drrAB mutation on the three regulatory genes dnrN, dnrO and dnrI along with the structural gene dpsA, which is essential for polyketide biosynthesis (Grimm et al., 1994). qRT-PCR results show that both dnrI and dpsA are downregulated to 1/8th and 1/16th, respectively, when compared with the WT (Fig. 4b). The melting-curve analysis shows a single peak for the respective amplicons and the amplification efficiency plot had a slope <0.1 (Fig. 4a). This finding confirms the hypothesis that an increase in the DNR level is sensed and the key activator of drug biosynthesis DnrI is downregulated. This results in a decline of dpsA expression, which is essential for polyketide biosynthesis. In the null mutant, DnrN has failed to activate dnrI transcription in spite of a 2.2-fold increase in the dnrN transcript relative to WT as seen in qRT-PCR results. The DnrN-binding site at the dnrI promoter region is a high-affinity site for DNR intercalation (Furuya & Hutchinson, 1996). Therefore, a small increase in the DNR level within the cell is sufficient to exclude DnrN from its activation site. It is intriguing that dnrN/O has an upstream element that is intercalated by DNR in competition with DnrO, which is an activator protein of dnrN transcription (Otten et al., 2000). The possible reason for the increase in the dnrN transcript is that DnrO possibly binds to a second activation site indicated in a previous report (Jiang & Hutchinson, 2006). Nevertheless, the slight increase in the dnrN transcript in the mutant remains unexplained. qRT-PCR shows that the DnrO transcript level increases by 3.4-fold in the mutant relative to WT. DnrO, in addition to being an activator, is also an autorepressor, and DNR intercalation is expected to alleviate DnrO repression. Therefore, an increase in dnrO transcription is in expected lines (Fig. 4b). Figure 5 illustrates the feedback regulation of DNR biosynthesis in S. peucetius.

Figure 4.

 qRT-PCR (a) Amplification efficiency plot for the target genes dnrN, dnrO, dnrI and dpsA and the slope is 0.075, 0.065, 0.002 and 0.037, respectively. (b) Log relative quantity of dnrI, dpsA, dnrN and dnrO in Streptomyces peucetius and the mutant. The relative quantity of S. peucetius WT was considered to be one. Error bars indicate the SD from the experimental replicates.

Figure 5.

 Regulation of DNR biosynthesis in Streptomyces peucetius. (a) Scheme of regulation in S. peucetius WT where the drug efflux is normal and (b) in the drrAdrrB null mutant, where the drug efflux is limited in the absence of DNR-specific efflux pump DrrAB.

Overexpression of drrAB genes under the control of a strong constitutive promoter has been shown to increase DNR production by 2.2-fold (Malla et al., 2009). It would be interesting to study the effect of dnrI overexpression along with drrAB genes.

For the first time, a feedback mechanism of drug production has been studied in a drug efflux without a mutant. The study highlights the use of the drug-producing organism itself rather than in a heterologous system for the analysis of a regulatory mechanism. We have shown that disruption of the DNR-specific efflux pump exerts a negative effect on drug production due to the innate ability of the cell to sense the drug levels within the cell and halt biosynthesis when it reaches a threshold level. For this to occur, the transcription of dnrI is downregulated by the intercalation of DNR at a specific DNA sequence that prevents activation by DnrN. We suggest that similar studies in other antibiotic-producing Streptomyces could shed more light into the regulatory mechanisms operating in them.


P.S. thanks CSIR for funding. The authors thank Dr K. Dharmalingam for his critical comments and technical support. Instrument support provided by DBT Centre for Genetic Engineering and Strain Manipulation and UGC SAP, at Madurai Kamaraj University, is acknowledged.