Targeting the Osmotic Stress Response for Strain Improvement of an Industrial Producer of Secondary Metabolites

Abstract

The transition from primary to secondary metabolism in antibiotic-producing Streptomyces correlates with expression of genes involved in stress responses. Consequently, regulatory pathways that regulate specific stress responses are potential targets to manipulate to increase antibiotic titers. In this study, genes encoding key proteins involved in regulation of the osmotic stress response in Streptomyces avermitilis, the industrial producer of avermectins, are investigated as targets. Disruption of either osaB_Sa, encoding a response regulator protein, or osaC_Sa, encoding a multidomain regulator of the alternative sigma factor SigB, led to increased production of both oligomycin, by up to 200%, and avermectin, by up to 37%. The mutations also conditionally affected morphological development; under osmotic stress, the mutants were unable to erect an aerial mycelium. In addition, we demonstrate the delivery of DNA into a streptomycete using biolistics. The data reveal that information on stress regulatory responses can be integrated in rational strain improvement to improve yields of bioactive secondary metabolites.

Introduction

Bacteria belonging to the genus Streptomyces are responsible for producing the majority of known antibiotics, together with some immunosuppressant compounds, antitumor agents, and anthelmintics such as avermectin [8]. These non-motile actinobacteria most commonly inhabit the soil, and in this habitat they frequently encounter and adapt to a variety of environmental stresses. Their secondary metabolism, and hence antibiotic production, is normally linked to morphological development and on solid media coincides with the development of an aerial mycelium that subsequently produces spores. Both processes can be triggered in response to environmental factors, such as osmotic stress [1,18]. Moreover, stress responses are integral to reprogramming the physiology of Streptomyces as they transition from primary to secondary metabolism. For example, proteomic analyses of Streptomyces coelicolor cultures sampled at successive time points during fermentations have identified elevated expression of stress proteins, including regulatory proteins such as the response regulator OsaB controlling the osmotic stress response, in the so-called transition phase immediately preceding the upregulation of secondary metabolism [14,19]. Similarly, a proteomic study of S. avermitilis identified several stress proteins that were highly expressed specifically at the onset of avermectin production [22].

Key components of regulation of the osmotic stress response that impact on both development and antibiotic production in the model streptomycete S. coelicolor have been described [2,6]. Loss of function of either OsaB or OsaC, respectively a response regulator and a multidomain regulator of the alternative sigma factor SigB, has contrasting effects on differentiation and secondary metabolite production. Under osmotic stress conditions, mutants in S. coelicolor cannot erect aerial hyphae but produce up to 5-fold greater antibiotic yields (actinorhodin and undecylprodigiosin) compared with the wild-type strain [2]. To examine if this regulatory paradigm is generally applicable and can be translated to improve production of commercially relevant secondary metabolites, we have investigated disruption of orthologous regulatory genes in S. avermitilis, a commercial producer of avermectins and oligomycin. The approach we adopted to construct one of the mutants was to introduce DNA by biolistics. There have been few reports of bacterial transformation by biolistics and none to date with Streptomyces; indeed, this method has been used primarily for DNA delivery into eukaryotic cells, notably into plant or animal tissue [3,13].

 

Materials and Methods

Microorganisms and Culture Conditions

Microorganisms and plasmids used are indicated in Table S1. E. coli strains were cultivated on LB agar plates [17]. Antibiotics used were kanamycin (25 μg/ml), ampicillin (100 μg/ml), chloramphenicol (25 μg/ml), hygromycin (100 μg/ml), spectinomycin (10 μg/ml) and apramycin (100 μg/ml). The sporulation medium was MS agar [9]. For S. avermitilis genomic DNA isolation, a Wizard Genomic DNA Purification kit (Promega, USA) was used. For that purpose, cultures were grown on LB agar for 36 h, and 100 mg of mycelium was scraped off and then processed as per the manufacturer’s instructions.

Cloning of SAV2511 and Derivation of a Tn5062 Insertion

A 5.1 kb DNA fragment containing SAV2511 was released from cosmid CL-236-G10 by digestion with EcoRI and BglII enzymes; pME6 was simultaneously digested with the same enzymes. The purified restricted DNA fragments were ligated together and the recombinant plasmid pMBSa1 was obtained after electroporation of E. coli JM109. To reduce the size of the inserted fragment to contain just the SAV2511 gene, pMBSa1 was digested with KpnI to release a 2.1 kb fragment, which was then ligated with pME6 digested with the same enzyme. The recombinant plasmid pMBSa2L was recovered. To disrupt SAV2511, pMBSa2L was cut with AatII that recognizes a single site within the open reading frame. The restricted plasmid was subsequently treated with T4 DNA polymerase enzyme to generate blunt ends and ligated with a 3.4 kb PvuII fragment containing the transposon Tn5062. The plasmid pMBTn1 with a Tn5062 insertion in SAV2511 was recovered. To construct the corresponding S. avermitilis SAV2511 mutant, intergeneric conjugation of pMBTn1 from E. coli ET12567/pUZ8002 was used [7]. Apramycin-resistant, kanamycin-sensitive exconjugants were recovered. The identity of S. avermitilis mutants was confirmed by Southern hybridization [17] using a 3.4 kb PvuII fragment derived from Tn5062 as a probe.

Cloning of SAV2513 and Derivation of a Tn5062 Insertion

SAV2513 was amplified from S. avermitilis genomic DNA using oligonucleotides with the following sequences:

(Forward) 5’-TGA ATT CTT CCA CGA ACC GGA CAT ACC-3’

(Reverse) 5’-TTC TAG ATA CCT CCA GCT CCG TCT CGT-3’

DNA amplification was performed in a PTC-200 thermocycler (Bio-Rad, UK), with an initial 95℃ denaturation temperature for 3 min, followed by 30 amplification cycles (95℃, 1 min; 60℃, 1 min; 72℃, 2 min), using high-fidelity Pfu polymerase (Agilent Technologies Inc./Stratagene, USA). Subsequently, a 2.1 kb amplicon corresponding to SAV2513 was purified. For cloning, the SAV2513 amplicon and pME6 were digested with EcoRI and Xbal enzymes. They were then ligated together and the recombinant plasmid pMCSa1 was subsequently recovered after electroporation of E. coli JM109. To disrupt SAV2513, pMCSa1 was digested with EcoRV that has a single recognition site within the open reading frame. The digested plasmid was combined with the 3.4 kb PvuII fragment containing Tn5062 and the recombinant plasmid pMCTn2 containing the insertion subsequently recovered in E. coli JM109.

S. avermitilis Transformation Employing Biolistics

Instead of employing intergeneric conjugation, we investigated using biolistics as an alternative method to introduce pMCTn2 into S. avermitilis. MS-sorbitol (0.75 M) plates were spread with 1 × 106 spores and incubated for 24 h at 30℃ until a bacterial background lawn was obtained. Vector introduction was performed using a Biolistic PDS-1000/He Particle Delivery System (Bio-Rad, Mexico), selecting a 9 cm bombardment distance and a 1,200 psi pressure. The microcarriers of tungsten were prepared in 50% glycerol (30 mg/ml) mixed for 5 min on a platform vortexer to resuspend and disrupt agglomerated particles. Then 50 μl (3 mg) of microcarriers was transferred to a 1.5 ml microcentrifuge tube. Continuous agitation of the microcarriers was needed for uniform DNA coverage. While vortexing vigorously, the following were added in order: 5 μl of DNA (1 μg/μl), 50 μl of 2.5 M CaCl2, and 20 μl of 0.1 M spermidine (free base, tissue culture grade). Vortexing was continued for 2–3 min. The microcarriers were then allowed to settle for 1 min and pelleted by spinning for 2 sec in a microfuge. The liquid was removed and discarded and 140 μl of 70% ethanol (HPLC or spectrophotometric grade) was added. This washing was repeated twice and then 48 μl of 100% ethanol was added. The final pellet was gently resuspended by tapping the side of the tube several times, and then by vortexing at low speed for 2–3 sec. The DNA-coated microcarriers were fired into the bacterial lawn and the plates then incubated at 30℃ for 16 h prior to overlaying lawns with a solution of apramycin (100 μg/ml). They were incubated under the same conditions for 7 days. Single colonies were subsequently isolated and tested for sensitivity to kanamycin. The identity of the mutants was confirmed by Southern hybridization [17] using a 3.4 kb PvuII fragment derived from Tn5062 as a probe.

Oligomycin Production

For pre-cultures, 25 ml of ϕ medium containing MgSO4·7H2O (0.5 g/l, J.T. Baker, Mexico), CaCl2·2H2O (0.7 g/l, J.T. Baker), glucose (10 g/l, J.T. Baker), tryptone (5 g/l, Bioxon), yeast extract (5 g/l, Bioxon), and Lab Lemco Powder (5 g/l, Oxoid), adjusted to pH 7 in a 250 ml Erlenmeyer flask was inoculated with 1 × 106 spores. Cultures were incubated at 30℃ with 225 rpm shaking for 17 h or before production of melanin was evident. Subsequently, 5 ml of the pre-culture was inoculated into a 250 ml Erlenmeyer flask with 20 ml of the production medium (AP-5), the composition of which was L-glutamic acid (0.6 g/l, J.T. Baker), FeSO4·7H2O (0.01 g/l, J.T. Baker), corn starch (80 g/l, Sigma), yeast extract (5 g/l, Bioxon), CaCO3 (7 g/l, J.T. Baker), MgSO4·7H2O (1 g/l, J.T. Baker), and K2HPO4 (1 g/l, J.T. Baker), adjusted to pH 6.9 and supplemented with 1 ml of trace metal solution containing ZnCl2 (20 mg/l, J.T. Baker), FeSO4·7H2O (20 mg/l, J.T. Baker), MnCl2·4H2O (20 mg/l, J.T. Baker), and distilled H2O (20 ml). Cultures were incubated at 28℃ with 225 rpm shaking for 10 days. For the experiments under osmotic stress conditions, KCl (J. T. Baker) was added to the production medium (AP-5) at a final concentration of 250 mM.

Oligomycin Bioassay with Aspergillus niger

Samples (400 μl) were obtained at different time points during the fermentation of strains in the (AP-5) production medium. The biomass and supernatants were separated by centrifugation, with the former subsequently dried and weighed. Then, 200 ml of AP-5 agar medium poured in 230 mm × 15 mm Petri dishes was inoculated with 1 × 109 spores of A. niger (A10). Eighteen wells (10 mm diameter) per plate were subsequently made with a punch, and 200 μl per well of fermentation supernatant or a range of concentrations of commercial oligomycin (Sigma) was added. Plates were placed at 4℃ for 30 min and subsequently incubated at 30℃ for 24 h or until vegetative growth was observed. The diameters of inhibition halos were measured.

Avermectin Production

Seed cultures were grown in 250 ml flasks containing 30 ml of seed medium (30 g/l soluble corn starch, 15 g/l yeast extract, (Bioxon), 5 g/l corn syrup solids, 0.4 g/l (Bioxon) KH2PO4 (J. T. Baker), and 2.5 g/l polyethyleneglycol (PEG 2,000, Sigma) at 28℃ and 200 rpm for 24 h. Then, 4 ml of seed culture was transferred to a 500 ml Erlenmeyer flask containing 40 ml of production medium (30 g/l soluble corn starch (Sigma), 14.4 g/l soybean flour (FloryVida), 21.6 g/l yeast extract, (Bioxon) 7.2 g/l corn syrup solids (Bioxon), 0.03 g/l CoCl2 (J. T. Baker), and 0.6 g/l KH2PO4 (J. T. Baker), pH 7.2). Fermentations were for up to 5 days at 28℃ and 200 rpm. To measure production under osmotic stress conditions, KCl (J. T. Baker) was added to the production medium at a final concentration of 250 mM.

Quantification of Avermectin by HPLC

Samples (1 ml) were obtained at different time points during the fermentation of strains in the avermectin production medium. One half of each sample was used to determine the biomass dry weight. The remainder was centrifuged to separate the mycelia and supernatant. Avermectins were extracted from the separated biomass by mixing with an equal volume of methanol for 30 min. A 10 μl sample of methanol extract supernatant was analyzed by HPLC. A Nova-pak C18 column (Waters Ltd, USA: 3.9 mm in inner diameter, 150 mm in length) was developed with methanol water (85:15 (v/v)) at a flow rate of 0.8 ml/min in ambient conditions. The quantities of total avermectins (AVMs) were calculated from the integration value at 246 nm using an authentic sample of avermectin B1a (SigmaAldrich, Mexico) as a standard. The B1a proportion among avermectins was calculated as a ratio of B1a peak area and the total area at 246 nm.

Reproducibility of Results

Experiments were performed in duplicate and repeated at least twice. The variation coefficient was always lower than 8%.

 

Results

Comparison of Osmotic Stress Regulation in S. avermitilis and S. coelicolor

The S. avermitilis genome contains an orthologous gene cluster of osmoadaptation genes [20] with a similar architecture to that found in the S. coelicolor genome. Genes encoding an atypical two-component signal transduction system are located in one arm of the chromosome, with a hybrid histidine kinase encoded by osaASa (SAV2512) that shares 92.3% amino acid identity with the corresponding S. coelicolor protein. Downstream of osaASa is the response regulator gene osaBSa (SAV2511), whose protein product shares 93.8% amino acid identity with its S. coelicolor ortholog. OsaA and OsaB proteins from both organisms have very similar domain architectures. Upstream and divergently transcribed from osaASa is a multidomain regulator encoded by osaCSa (SAV2513) sharing 87.5% amino acid identity with OsaCSc. Comparison of the domain architectures revealed that the RsbW-like N-terminal kinase domain (HATPase_c domain) of S. coelicolor OsaC is less well conserved in the S. avermitilis protein; indeed, it is not predicted by SMART (http://smart.embl-heidelberg.de/) [12] (Fig. 1A). More specific alignment of the corresponding regions of the respective OsaC proteins with other better characterized RsbW-like kinase domains found in Bacillus subtilis proteins revealed the absence of conserved amino acid residues adjacent to and within the putative ATP-binding Bergerat fold [5] of the S. avermitilis protein (Fig. 1B). Of particular note is the substitution of a conserved acidic aspartate residue by glycine in the Bergerat fold G3 box. The RsbW-like kinase domain of OsaCSc has a function in regulating the activity of the alternative sigma factor SigB [6], which can in turn regulate other alternative sigma factors involved in the osmotic stress response in S. coelicolor [11]. Comparison of SigBSa and SigBSc revealed that they share 77.8% amino acid identity.

Fig. 1.Comparison of OsaCSc and OsaCSa. (A) The Simple Modular Architecture Tool (SMART) was used to predict the domain structures of the respective proteins, indicating the absence of the N-terminal HATPase_c domain in OsaCSa (HATPase_c, histidine kinase-like ATPase domain; PAS, Per-Arnt-Sim signal sensor domain; GAF, cAMP/cGMP binding regulatory domain typical of cGMP-specific phosphodiesterases, adenylyl and guanylyl cyclases, and phytochromes; PP2CSIG, Sigma factor PP2C-like phosphatase domain). (B) An alignment of residues 41-197 of OsaCSa with the HATPase-c/RsbW-like kinase domain of OsaCSc and related kinase proteins from B. subtilis, indicating the conserved N-box and G1, G2, and G3 boxes of the Bergerat ATP-binding fold. Residues highlighted in black are conserved, those in dark grey share >75% identity, and those in light grey share >50% identity.

Mutagenesis of SAV2511 (osaBSa)

To disrupt osaBSa, the gene was first subcloned from cosmid CL-236-G10. A copy of Tn5062 was then introduced at the single AatII site within the osaBSa coding sequence (at position 167 within the 687 bp coding DNA sequence) to create pMBTn1. This plasmid, which cannot replicate in Streptomyces, was transferred into S. avermitilis from E. coli by intergeneric conjugation, exploiting the oriT carried by the transposon. Six apramycin-resistant clones were selected for further analysis. Southern hybridization of genomic DNA isolated from each clone with a Tn5062 probe indicated replacement of the wild-type gene by the disrupted allele (Fig. 3).

The six mutants, together with the parental strain, were grown on MS sporulation medium with and without 250 mM KCl (Fig. 2). Whereas all strains could erect an aerial mycelium on non-supplemented MS, all six mutants, in contrast to the wild-type, were unable to do so when challenged by osmotic stress (Fig. 2B), paralleling the phenotype of an S. coelicolor osaB mutant [2].

Fig. 2.S. avermitilis osaB and osaC mutants are conditionally bald. (A and B) Six independent osaB mutants (1–6) and the parental wild-type strain (P) were plated on MS medium (A) and MS medium supplemented with 250 mM KCl (B) and grown for 5 days. The mutants failed to erect an aerial mycelium on the supplemented medium. (C) The parental strain (P), an osaB mutant (1), and the mutant with pPM04 expressing osaBSc (2) were grown on MS medium supplemented with 250 mM KCl for 7 days. (D and E) Six independent osaC mutants (1–6) and the parental wild-type strain (P) were plated on MS medium (D) and MS medium supplemented with 250 mM KCl (E) and grown for 7 days. The mutants failed to erect an aerial mycelium on the supplemented medium. (F) The parental strain (P), an osaC mutant (1), and the mutant with pSHOsaC1 expressing osaCSc (2) were grown on MS medium supplemented with 250 mM KCl for 7 days.

Fig. 3.Verification of osaBSa mutants. Southern hybridization of genomic DNA isolated from each clone with a Tn5062 probe indicated replacement of the wild-type gene by the disrupted allele. Genomic DNA from the six apramycin-resistant mutants (lanes 1-6) was digested with NotI, as was pMBTn1 (lane C). A hybridizing 3.1 kb fragment common to each lane indicated the presence of the disrupted allele in each mutant. A lambda DNA HindIII digest was included as a molecular size marker.

Further evidence for a conservation of function between the OsaB proteins from both species was obtained by genetic complementation of the S. avermitilis osaB mutant by osaBSc. An integrating plasmid carrying osaBSc, pPM04 [2], was introduced into a representative S. avermitilis osaB mutant and could restore development of an aerial mycelium when the strains were grown under osmotic stress (Fig. 2C).

Mutagenesis of SAV2513 (osaCSa)

The osaCSa gene sequence was amplified from cosmid CL-236-G10 and cloned as an EcoRI-XbaI fragment. It was disrupted by insertion of Tn5062 at the single internal EcoRV site (at position 358 within the 2,748 bp coding DNA sequence), generating plasmid pMCTn2. This plasmid, which is unable to replicate in Streptomyces, was introduced by biolistics into young hyphae of S. avermitilis growing on solid agar. This, we believe, is the first report of delivery of DNA into a streptomycete by this method. From 110 colonies, six apramycin-resistant mutants were selected for further analysis. PCR analysis of genomic DNA confirmed the replacement of osaCSa by the mutated allele in each mutant. Phenotypic analysis also indicated that the mutants were unable to erect an aerial mycelium when grown under osmotic stress (Fig. 2E), reproducing the phenotype of the corresponding S. coelicolor osaC mutant [6]. However, unlike the example of interspecific complementation observed for osaB, when the S. coelicolor osaC gene on plasmid pSHOsaC1 [6] was introduced into a representative mutant, there was no evidence for complementation of the developmental phenotype of the osaCSa mutant (Fig. 2F).

Oligomycin Production Is Increased in Osmoadaptation Mutants of S. avermitilis

One representative osmoadaptation mutant of each type was chosen for analysis of antibiotic production. To monitor oligomycin production, S. avermitilis strains were grown in submerged culture in AP-5 medium with and without 250 mM KCl to induce osmotic stress. Culture supernatants were sampled at different time-points and their oligomycin contents measured using a bioassay against Aspergillus niger, [16] comparing with known concentrations of commercial oligomycin. For the wild type, production remained at between 0.1 to 0.15 μg/mg dry weights of biomass from 26 h until 160 h fermentation, irrespective of osmotic stress (Figs. 4A and 4B). In contrast, when grown in non-supplemented production medium, both mutant strains displayed a peak in oligomycin production at 26 h, with titers 3- to 3.5-times greater than produced by the wild type at the corresponding time-point. For longer fermentations, the antibiotic yields from the mutants subsequently declined but remained consistently higher than those of the wild type. These production profiles were reproduced but tended to be amplified when the osaCSa was grown under osmotic stress. At 26 h, the oligomycin titer of this mutant reached 0.52 μg/mg dry weight of biomass, approximately 5 times greater than that of the wild type. However, after 52 h, production declined to levels similar to the wild type. An increase in production after 26 h fermentation in osmotic stress conditions was also noted for the osaBSa mutant, and this also declined after longer fermentations (Fig. 4B).

Fig. 4.Comparison of oligomycin production kinetics of the wild type and osmoadaptation mutants. Quantification of oligomycin titers (μg/mg dry weight of biomass) produced by the parental strain (open circles), an osaB mutant (solid squares), and an osaC mutant (solid triangles) fermented in AP-5 medium (A) and in AP-5 medium supplemented with 250 mM KCl (B).

Avermectin Production Is Increased in Osmoadaptation Mutants of S. avermitilis

Avermectin production was analyzed using HPLC separations of culture extracts grown under non-stressed conditions. Under these conditions, avermectin production was increased 37% with respect to the parental strain in both osmoadaptation mutants, peaking at 285 ng/mg dry weight of biomass in the mutants compared with 180 μg/mg dry weight of biomass in the wild type after 120 h fermentation (Fig. 5). No significant differences in avermectin production were observed between the strains fermented in osmotic stress conditions, with yields peaking at 150 ng/mg dry weight of biomass after 120 h fermentation.

Fig. 5.Comparison of avermectin production kinetics of the wild type and osmoadaptation mutants. Quantification of avermectin B1a titers (ng/mg dry weight of biomass) produced by the parental strain (open circles), an osaB mutant (solid squares), and an osaC mutant (solid triangles) fermented in production medium.

 

Discussion

Empirical strain improvement programs applied to antibiotic-producing streptomycetes have exploited random mutagenesis and selection of overproducing variants [21]. More recently, this approach has been combined with protoplast fusion to shuffle combinations of useful traits between genomes [23,24]. A “semi-rational” approach has been to isolate mutations resulting in antibiotic resistance due to changes in ribosome or RNA polymerase function that, for reasons that are not entirely clear, result in improved antibiotic production [20]. Rational engineering of strains has targeted optimizing primary metabolism, and hence precursor supply, or rate-limiting steps in secondary metabolic pathways to improve overall flux and corresponding yields of specific antibiotics [8,15]. In addition, pathway-specific regulator genes, encoding either transcriptional activators or repressors, have been targeted to increase expression of genes within specific secondary metabolite biosynthetic gene clusters [4]. Lastly, pathways, often activated or de-repressed, have been expressed in heterologous hosts that have been modified to delete competing secondary metabolite pathways. For example, using S. avermitilis for heterologous expression of the streptomycin biosynthetic gene cluster, a 6-fold increase in streptomycin production, from 3 to 18 μg/ml, was obtained in a genome-minimized strain compared with the wild type [10]. Some of these approaches have been applied to S. avermitilis to increase production of its own antibiotics. For example, ribosomal mutations and genome shuffling improved production of an avermectin analog, doramectin, 11.2-fold [23]. Increased expression of aveR, encoding the transcriptional activator of the avermectin biosynthetic genes, improved avermectin yield by 50% [25].

To assess whether modifying pleiotropic regulatory pathways could be another broadly applicable rational approach to improve antibiotic production, we have created specific mutants of S. avermitilis impaired in the regulatory pathways controlling osmoadaptation. A conserved signal transduction system, comprising a hybrid histidine kinase, OsaA, and an atypical response regulator, OsaB, was disrupted by knocking out the function of the latter. As is the case of the corresponding osaB null mutant of S. coelicolor, the S. avermitilis mutant was conditionally “bald,” being unable to erect an aerial mycelium when grown on medium supplemented with osmolyte. Moreover, the mutant overproduced two different antibiotics. In non-supplemented fermentations, a 3-fold increase in oligomycin production was obtained after 26 h growth; this was moderately enhanced further by inclusion of osmolyte in the fermentation. For avermectin, a 37% increase in yield was obtained in non-supplemented fermentations. The OsaB protein consists of a conserved N-terminal receiver domain, but, unlike most response regulators, it lacks a C-terminal DNA binding domain. Both S. avermitilis and S. coelicolor proteins contain a coiled-coil region that is required for dimerization (unpublished results), but the mechanism for relaying a signal to effect a response in conditions of high external osmolarity is unknown. The observation that expression of OsaBSc can restore aerial development in an osaB mutant of S. avermitilis implies that the mechanism is common to both species.

The respective OsaC proteins of both S. avermitilis and S. coelicolor share less identity than their OsaB proteins. However, the osaCSa null mutant was, like the corresponding S. coelicolor mutant, conditionally bald. Antibiotic production was also elevated: in non-supplemented fermentations, oligomycin production was increased 3.5-fold and avermectin production improved by 37%. An even greater increase in oligomycin production, up to 5-fold, was obtained in fermentations with added osmolyte. OsaC is the paradigm for a multidomain regulator protein family specific to Streptomyces. To date, the function of only one domain has been investigated in detail: the N-terminal RsbW-like kinase (HATPase_c) domain. This domain confers on OsaCSc an anti-sigma factor activity required to modulate the activity of the alternative sigma factor SigB subsequent to the osmotic stress response [6]. There is evidence that, under specific growth conditions that induce the osmotic stress response, this sigma factor sits at the top of a cascade of alternative sigma factors, including SigL and SigM [11]. The molecular targets for OsaC’s regulatory functions may be similar in both species. Even though SMART fails to predict a HATPase_c domain in OsaCSa, manual alignment with RsbW-like kinase domains indicates the presence of the majority of conserved residues. The inability of OsaCSc to restore aerial development in the osaCSa mutant may be a consequence of divergence of the molecular targets; for example, the respective SigB proteins share only 78% identity. The approach we adopted to construct the osaCSa mutant was to introduce the non-replicating plasmid carrying the Tn5062-disrupted allele into S. avermitilis by biolistics. To the best of our knowledge, this is the first report of DNA delivery to Streptomyces by biolistics, and this opens possibilities for manipulation of less genetically tractable actinobacteria.

In conclusion, we have demonstrated that targeted mutations to disrupt regulatory pathways involved in the osmotic stress response lead to antibiotic overproduction in a species of great importance to biotechnology. An inference of this study is that these mutations can be incorporated in rational strain improvement programs to improve antibiotic yields in other streptomycetes. In addition, as the effects are pleiotropic, affecting more than one pathway, they can be utilized in generating genetic backgrounds in which so-called cryptic pathways may be activated, expanding the range of bioactive secondary metabolites produced by the genus.