Changhua Hu and Guojian Liao, Institute of Modern Biopharmaceuticals, School of Pharmaceutical Sciences, Southwest University, Chongqing 400715, China. E-mails: email@example.com; firstname.lastname@example.org
Aims: Daptomycin, one of the A21978C factors produced by Streptomyces roseosporus, is an acidic cyclic lipopeptide antibiotic with potent activity against a variety of Gram-positive pathogens. To increase the titre of this extensively used and clinically important antibiotic, we applied a reported-guided rpsL mutation selection system to generate strains producing high levels of A21978C.
Methods and Results: In the reporter design, dptE was chosen as the overexpressing target, and neo-encoding neomycin phosphotransferase as the reporter. Using this reporter-guided selection system, 20% of the selected, streptomycin-resistant mutants produced greater amounts of A21978C than the starting strain. The selection system increased the screening efficiency about 10-fold with a frequency of 1·7% A21978C overproducing strains among strr mutants. A21978C production was increased approximately 2·2-fold in the rpsL K43N mutant.
Conclusions: The combination of ribosome engineering and reporter-guided mutant selection generated an A21978C overproducing strain that produced about twice as much A21978C as the parental strain.
Significance and Impact of the Study: The strategies presented here, which integrated the advantages of both ribosome engineering and reporter-guided mutation selection, could be applied to other bacteria to improve their yield of secondary metabolites.
An increasing need for the microbial production of antibiotics demands high-yielding strains suitable for industrial applications. Traditionally, strain improvement is achieved by multiple rounds of random mutagenesis and screening. Although the outcome of this strategy is fruitful, it is time intensive and laborious. A novel approach, called ‘ribosome engineering’, has been applied to obtain antibiotic-overproducing strains by simply screening for drug resistance mutants via simple selection on drug-containing plates (Hosoya et al. 1998). It has proven to be a powerful strategy for strain improvement, resulting in the generation of a number of high-producing strains (Hu and Ochi 2001; Hu et al. 2002; Tamehiro et al. 2003; Wang et al. 2008, 2009; Tanaka et al. 2009; Fukuda et al. 2010). The percentage of overproducing mutants among drug resistance mutants varied substantially depending on the species used, ranging from 2 to 20%.
Recently, a reporter-guided mutation selection method has been developed to facilitate the selection of antibiotic-overproducing strains. This strategy is based on the observation that the expression level of some genes in the antibiotic biosynthetic gene cluster is positively correlated with the titre of antibiotic production (Rokem et al. 2007; van Wezel and McDowall 2011). Its application in improving clavulanic acid production in Streptomyces clauligerus and lovastatin production in Aspergillus terreus resulted in the isolation of mutants with enhanced clavulanic acid or lovastatin production at a higher incidence (50–90%) (Askenazi et al. 2003; Xiang et al. 2009).
Daptomycin is a lipopeptide made by Streptomyces roseosporus, with bactericidal activity against a wide range of Gram-positive pathogens, including vancomycin-resistant Staphylococcus aureus, methicillin-resistant Staph. aureus (MRSA), penicillin-resistant Streptococcus pneumoniae, vancomycin-resistant enterococci and other antibiotic-resistant strains (Robbel and Marahiel 2010). Daptomycin is a member of the A21978C factors, consisting of 13 amino acids and a fatty acid which ranges from 10 to 13 carbon atoms (Liao et al. 2012). The biosynthesis of daptomycin has been widely studied, and most of its biosynthetic steps have been characterized. The putative daptomycin biosynthetic gene cluster consists of nine genes ranging from dptE to dptJ, and the contiguous genes from dptE to dptH may be transcribed on a giant polycitronic transcript (Miao et al. 2005). In this article, we describe the application of a reporter-guided rpsL mutation selection system to generate A21978C high-producing strains. The gene dptE, the first within the daptomycin cluster, was chosen as the target, and the kanamycin resistance gene neo as the reporter in reporter-guided mutation selection design (Fig. 1). Transformants harbouring the reporter system were subjected to ribosome engineering to obtain spontaneous streptomycin-resistant (strr) strains. Of the selected kanamycin-resistant mutants, 20% showed an improvement in A21978C production. Given that the frequency of A21978C overproducing strains among the strr mutants was 1·7%, the selection system had increased the screening efficiency about 10-fold.
Materials and methods
Bacterial strains, plasmids and growth conditions
Bacterial strains and plasmids used in this study are listed in Table 1. Streptomyces roseosporus NRRL11379, the producer of daptomycin, was obtained from the ATCC. Micrococcus luteus was used as indicator strain for A21978C bioassay. Escherichia coli JM109 was used for the cloning and subcloning. Escherichia coliET12567 harbouring pUZ8002 was used for conjugal transfer according to the established protocol (McHenney et al. 1998).
Table 1. Bacterial strains and plasmids used in this study
Strains or plasmids
Escherichia coli–Streptomyces shuttle vector capable of integration into PhiC31 attB site in Streptomyces
Containing pSRdptE and with mutation in rpsL (K43N)
S. roseosporus DE64
Containing pSRdptE and with mutation in rpsL (K43N)
S. roseosporus SR90
Containing pSRdptE and with mutation in rpsL (K88E)
S. roseosporus SR114
Containing pSRdptE and with mutation in rpsL (K88R)
Streptomyces roseosporus and its derivative strains were grown at 28°C in various media. Solid medium AS-1 and liquid medium TSB were prepared as described elsewhere (Nguyen et al. 2006). F10A medium (CaCO3 0·3%, distillers soluble 0·5%, soluble starch 2·5%, yeast extract 0·5%, glucose 0·5% and bactopeptone 0·5%) was used for daptomycin production. Escherichia coli strains were incubated in LB medium at 37°C. When necessary, antibiotics were used at the following concentrations: apramycin, 10 μg ml−1 in AS-1 for S. roseosporus and 100 μg ml−1 in LB for E. coli; kanamycin, 20–400 μg ml−1 in AS-1 for S. roseosporus and 100 μg ml−1 in LB for E. coli; and choramphenicol 34 μg ml−1 in LB for E. coli.
Primers and polymerase chain reactions
The oligonucleotide primers used to amplify the promoter sequence of dptE were P1 (TCCCCACCACCTGCCCAGTGT) and P2 (GACTCCGGTCGAACGACGCATC). rpsL-F (CAGCAGCTGGTCCGGAAG) and rpsL-R (CCTTCTTGGCGCCGTAGC) were used to amplify the rpsL DNA fragment from genomic DNA of S. roseosporus. Polymerase chain reactions (PCRs) were performed using FastPfu DNA polymerase (TransGen Biotech, Beijing, China): an initial denaturation at 95°C for 4 min followed by 30 cycles of amplification (95°C for 1 min, 63°C for 30 s and 72°C for 1 min) and an additional 10 min at 72°C.
Construction of the reporter plasmid
The putative promoter sequences of dptE were obtained by PCR with S. roseosporus genomic DNA as the template and primers P1 and P2. The resulting 366-bp fragment was inserted into the EcoRV site of pSR2 to generate pdptE. The integrity of the plasmid was confirmed by sequencing. The resulting plasmids pdptE, pSR2 and pSRE were passed through E. coliET12567/pUZ8002 and then introduced into S. roseosporus by conjugal transfer. The resulting transformant was inoculated on AS-1 plates to form spores.
Generation and selection of streptomycin-resistant strains
A total of 108 spores of transformants harbouring pSRE were spread out on an AS-1 plate containing various concentrations of streptomycin. The resulting spontaneous streptomycin-resistant mutants were dotted on plates containing various amounts of kanamycin and 10 μg ml−1 apramycin and then incubated at 28°C for 3 days. The resulting kanamycin-resistant mutants were used for further fermentation study, and those genomic DNAs were used as the template to amplify the rpsL DNA fragment by PCR with primers rpsL-F and rpsL-R.
Fermentation of S. roseosporus and A21978C bioassay
Spores of S. roseosporus and its derivatives were inoculated in TSB. The cultures were grown at 28°C on a rotary shaker (220 rev min−1) for 48 h and used as a seed culture. One millilitre (2% v/v) of seed culture was inoculated into flasks containing 50 ml of F10A medium and then fermented at 28°C on a rotary shaker (220 rev min−1) for 7 days. The culture filtrates harvested by centrifugation were used for the daptomycin bioassay.
The microbiological assay was adopted for A21978C bioassay. Micrococcus luteus was used as the assay organism. One millilitre of tested bacteria at a concentration of 105 CFU ml−1 was add to the LB agar medium that was supplemented with 5 mmol CaCl2 at 50°C and then pour plate. Ten-millimetre wells were made in the plate. The sample or standards were pipetted into the wells in aseptic condition. Plates were incubated for 18–24 h at 37°C after which time zone sizes were measured. The experiment was performed in triplicate, and the concentration of A21978C in the samples was calculated using the data from the curves derived from the drug standards. The standard curves of the zone size versus the natural logarithm of the drug concentration were linear between 0·5 and 8 μg ml−1 when the standards were prepared in methanol (r = 0·99).
Construction and verification of reporter system
The first gene in the daptomycin biosynthetic gene cluster dptE was chosen as the target, and its putative promoter (PdptE) was used to construct the reporter plasmid. The resistant gene neo, encoding neomycin phosphotransferase, was chosen as the reporter in this study (Fig. 2). pSR2 containing a promoterless neo and its transformants showed a resistance level of less than 20 μg ml−1 of kanamycin, while transformants of the control plasmid pSRE bearing the strong promoter PermE* were resistant to more than 100 μg ml−1. Transformants harbouring pSRdptE (pdptE in front of neo) were resistant up to 40 μg ml−1, demonstrating that dptE was moderately expressed in S. roseosporus (Fig. 3).
Generation and reporter-guided selection of str mutants
Streptomyces roseosporus pSRdptE as the starting strain was subjected to ribosome engineering. To obtain spontaneous streptomycin-resistant (strr) strains, spores of S. roseosporus pSRdptE were spread and incubated on AS-1 agar containing streptomycin at 5–20 times the MIC (MIC = 10 μg ml−1), and strr mutants developed after 4–10 days at a frequency of 2–4 × 10−8. A total of 117 strr mutants were obtained, 44, 52 and 21 from AS-1 containing streptomycin at 50, 100 and 200 μg ml−1, respectively.
All 117 strr mutants were spotted and incubated on AS-1 agar containing kanamycin (400 μg ml−1) at 28°C. A total of 10 strains developed colonies after 3 days’ incubation. Subsequently, the resulting 10 strains were fermented and their A21978C titres measured. Two (20%) mutants showed higher A21978C titres (2·2-fold) than the starting strain. To determine the frequency of A21978C overproducing strains among the strr mutant pool, we randomly selected 20 streptomycin-resistant strains that did not confer a high-level of kanamycin resistance (kans) for fermentation and measured their A21978C titres. Most of them produced the same amounts of A21978C and some even decreased their A21978C production compared with the starting strain (data not shown). Taken together, ribosome engineering using streptomycin as selective antibiotic could lead to the generation of A21978C high-producing mutants at a relatively low frequency (2/117, 1·7%), while the reporter plasmid was applied to mutation selection, the positive rate increased significantly (2/10, 20%).
Characterization of str mutants derived from S. roseosporus
Streptomycin inhibits protein synthesis by binding to the small ribosomal subunit. Streptomycin resistance frequently results from a mutation in the open reading frame of rpsL (Ochi et al. 2004). We amplified, sequenced and compared the rpsL gene of the wild-type and kanr strains. Three different types of missense mutations were identified in ten high-level kanr mutants at codons 129, 262 and 263 (Table 2). All missense mutations were caused by single-base substitutions, and the most frequent mutation was observed at nt 263 (Lys88-Arg, 70%), the second at nt 129 (Lys43-Asn, 20%) and the third at nt 262 (Lys88-Glu, 10%). The introduction of rpsL mutations at lys88 (K88E and K88R), frequently found in antibiotic-overproducing strains, failed to increase the production of A21978C. However, rpsL mutations at lys43 (K43N) resulted in a 2·2-fold increase in A21978C production (Fig. 4).
Table 2. Characterization of streptomycin resistance mutations of Streptomyces roseosporus
Position of mutation
Amino acid substitution*
Frequency of mutants with same mutation
*Numbered in accordance with the numbering system for Escherichia coli.
The application of ribosome engineering in various Streptomyces strains has resulted in the generation of a number of high-producing antibiotic mutants. When streptomycin was used as the selection drug, Strr mutants with enhanced antibiotic production usually contained a mutation within the rpsL gene leading to an alteration of Lys88 to Arg/Glu. For example, the salinomycin overproducing industrial strain of Streptomyces albus harboured the K88R mutation (Hu and Ochi 2001). Those mutant ribosomes sustained a high level of protein synthesis, especially in the late growth phase, resulting in a high expression of stationary phase-specific genes, including those for secondary metabolism (Okamoto-Hosoya et al. 2003) The mechanisms of those mutations’ beneficial effects are not as yet completely understood. In S. roseosporus, the K88R or K88E mutants in S. roseosporus failed to enhance A21978C production. In contrast, K43N mutants increased the yield significantly. Given that the efficacy of mutation on productivity often depends on the medium used, we employed several media to determine their effect on A21978C production. Unfortunately, K88R or K88E mutants could not produce higher levels of A21978C than the starting strain in all tested media (data not shown). Thus, the K43N alteration represents a new rpsL mutation that supports high-level A21978C production. Our results were consistent with similar results from other groups, showing that the K43N mutation was effective at increasing erythromycin production in Saccharopolyspora erythraea and oligomycin production in Streptomyces avermitilis (Tanaka et al. 2009). The K43N mutant of Saccharopolyspora erythraea exhibited a higher level of bldD expression at late growth phase than that of the wild-type strain. bldD was a key developmental regulatory gene that positively regulated the biosynthetic gene cluster of erythromycin (Chng et al. 2008).
In previous reports, reporter genes were used to select the pool of mutants generated by chemical (NTG) or physical (UV) mutagens to identify the overproducing strain. Although the positive rate was reasonably high, those strategies carried the inherent limitation of random mutagenesis, including generation of suboptimal mutants with the introduction of second-site detrimental mutations, and a difficulty in determining mutation loci. Thus, any improvement observed in one strain could be hard to transfer to another strain via inverse metabolic engineering. In this study, the overproducing phenotype can be ascribed to a point mutation in the rpsL gene and can be transferred to another high-producing strain by directed genetic manipulation. Moreover, the resulting high-producing strains could be used as a new starting strain to perform further ribosome engineering, as demonstrated whereby cumulative drug resistance mutations could significantly increase antibiotic production by up to two orders of magnitude (Wang et al. 2008).
In conclusion, we reported the combination of ribosome engineering and reporter-guided mutant selection to generate an A21978C overproducing strain. This resulted in desirable strains being produced at the positive rate of 20%, which produced 2·2-fold more A21978C than the starting strain. The strategy we used was simple, rapid and inexpensive and could be widely utilized for the improvement of other bacterial strains.
We would like to thank Prof. Huarong Tan for kindly providing pSET152 and E. coliET12567 and Keqian Yang for kindly providing pSER2 and pSRE. This work was supported by a grant (31100069) from the National Natural Science Foundation of China and a grant (XDJK2011C053) from the Fundamental Research Funds for the Central Universities. The authors declare that they have no competing interests.