Crosstalk between ROS Homeostasis and Secondary Metabolism in S. natalensis ATCC 27448: Modulation of Pimaricin Production by Intracellular ROS

Abstract

Streptomyces secondary metabolism is strongly affected by oxygen availability. The increased culture aeration enhances pimaricin production in S. natalensis, however the excess of O₂ consumption can lead to an intracellular ROS imbalance that is harmful to the cell. The adaptive physiological response of S. natalensis upon the addition of exogenous H₂O₂ suggested that the modulation of the intracellular ROS levels, through the activation of the H₂O₂ inducible catalase during the late exponential growth phase, can alter the production of pimaricin. With the construction of defective mutants on the H₂O₂ related enzymes SodF, AhpCD and KatA1, an effective and enduring modulation of intracellular ROS was achieved. Characterization of the knock-out strains revealed different behaviours regarding pimaricin production: whilst the superoxide dismutase defective mutant presented low levels of pimaricin production compared to the wild-type, the mutants defective on the H₂O₂-detoxifying enzymes displayed a pimaricin overproducer phenotype. Using physiological and molecular approaches we report a crosstalk between oxidative stress and secondary metabolism regulatory networks. Our results reveal that the redox-based regulation network triggered by an imbalance of the intracellular ROS homeostasis is also able to modulate the biosynthesis of pimaricin in S. natalensis.

Introduction

Streptomyces are Gram-positive, filamentous, soil-dwelling bacteria well known for their ability to produce a wide variety of secondary metabolites [1]. The biosynthesis of secondary metabolites occurs in a growth-phase dependent manner and is controlled by environmental and physiological factors [2]. Streptomyces secondary metabolism is regulated by a complex network that integrates multiple factors and takes place at different levels: from the so-called pathway-specific regulatory genes to pleiotropic regulators which control both secondary metabolism and morphological differentiation.

Streptomycetes secondary metabolism is an aerobic process and thus affected by oxygen availability. However, high levels of molecular oxygen consumption can lead to the formation of reactive oxygen species - ROS (hydrogen peroxide, H₂O₂; superoxide radicals, O₂ ^•− and hydroxyl radicals, HO^•) that can damage cell components such as proteins, nucleic acids and lipids [3]. To counteract the toxic effects of ROS, microorganisms have developed an adaptive response that extends from the modulation of gene expression to changes in enzymatic and non-enzymatic activities. The molecular machinery activated by this adaptive response is able to sense, scavenge ROS and repair the molecular damage. Concomitantly, it has been suggested that ROS can play an important role as secondary messengers on cell signalling, based on reductive-oxidative mechanisms [4]–[6]. Among ROS, H₂O₂ is the best studied as signalling molecule.

The ability to maintain cellular redox balance is essential to all organisms and is mainly achieved by the conversion of the redox signals into regulatory outputs, usually at the transcription level, which allows adaptation to the altered environment. Several studies suggest that the consequences of the adaptive response to oxidative stress extend beyond the primary effect of defence into alterations in the secondary metabolism profile. Although stress-induced regulatory mechanisms have been globally studied in Streptomyces, at the present there is a lack of knowledge on the influence, at the molecular level, of oxidative stress on the production of secondary metabolites. The S. coelicolor JH11 (ΔcatR) mutant strain that overproduces catalase (CatA), shows a reduced expression of the alkyl hydroperoxidase system (AhpCD) and produces lower amounts of actinorhodin [7]. Addition of a redox-cycling agent (phenazine methosulfate) to S. clavuligerus increases superoxide dismutase activity and also enhances clavulanic acid production by inducing the transcription of the pathway-specific regulator CcaR [8], [9]. The authors also report the same effect on the actinorhodin biosynthesis in S. coelicolor [9].

Streptomyces natalensis produces pimaricin, a 26-member tetraene macrolide antifungal antibiotic [10], widely used for the treatment of fungal keratitis and in the food industry to prevent mould contamination of non-sterile foods such as cheese, sausages, cured meat, among others. As a polyene, its antifungal activity lies in its interaction with membrane sterols, not causing membrane permeabilization as initially thought but inhibiting the sterol-dependent processes of membrane fusion and fission [11]. Pimaricin is synthesized by the action of a type I modular polyketide synthase (PKS) and its biosynthetic gene cluster has been previously sequenced and characterized [12]. The gene cluster contains 19 open reading frames including 5 multifunctional enzymes (PimS0-PimS4) that harbor 13 PKS modules [10], and 14 additional proteins involved in post-PKS modification of the polyketide skeleton (tailoring enzymes), export and regulation of gene expression [13]–[18]. Among these are two pathway-specific regulators, PimR and PimM. PimR is the archetype of a new class of regulators that combines an N-terminal domain corresponding to the SARP (Streptomyes antibiotic regulatory protein) family of transcriptional activators with a C-terminal homologous to guanylate cyclases and large ATP-binding regulators of the LuxR family (LAL)[13] . PimM combines an N-terminal PAS domain with a C-terminal helix–turn–helix (HTH) motif of the LuxR type [14], [19]. Both proteins seem to play an independent positive regulatory role in pimaricin biosynthesis by the transcriptional activation of different target genes. Additionally, a cholesterol oxidase encoding gene (pimE) is located in the centre of the cluster and although not classified as a pathway-specific regulator, it presents an important role, not yet fully understood, in the signaling transduction cascade leading to pimaricin biosynthesis [15].

Like in other streptomycetes, secondary metabolism in S. natalensis seems to be regulated in response to a variety of nutritional and environment signals in a growth-phase dependent manner [20]. In this study we present evidence for a functional molecular crosstalk between ROS homeostasis and secondary metabolism in S. natalensis. Using a combined approach of physiological and molecular characterization we present evidence that intracellular H₂O₂ levels are important to elicit pimaricin biosynthesis, particularly during late exponential phase. Modulation of intracellular H₂O₂ levels, either by prompting an adaptive response of S. natalensis to H₂O₂-induced oxidative stress or by the construction of knock-out mutants on the main H₂O₂-related enzymes, altered the pimaricin production profile.

Results

S. natalensis presents a catalase activity profile dependent on the growth-phase

In YEME liquid medium S. natalensis ATCC 27448 presents a typical growth curve, pimaricin is first detected during the late exponential phase and its production occurs until mid-stationary phase (Fig. 1A). For experimental purposes and in agreement to what was previously described for S. coelicolor [21], [22], the S. natalensis growth curve was divided into four growth stages: an early exponential phase characterized by a rapid growth (RG1); after a brief transition phase linked with the “metabolic switch” [23], there is a second rapid growth phase (RG2) with a lower growing rate that overlaps with the late exponential phase. Afterwards the cultures enter into the stationary phase (S). We have divided the stationary phase into an early- to mid-stationary phase when pimaricin biosynthesis occurs (S/P), and a late stationary phase, when pimaricin is no longer being synthesized by S. natalensis (S/NP) (Fig. 1A).

Figure 1. Pimaricin production and antioxidant growth-dependent profile of S. natalensis.

(A) Growth (··•··), catalase activity (–Δ–) and pimaricin production (−▪−) of S. natalensis ATCC 27448 in YEME medium. Growth phases are indicated by solid lines at the top of the graph. Vertical bars indicate standard deviation of the mean values. Data are the average of triplicates from three independent experiments. (B)Native PAGE of cell extracts (30 µg protein per lane) from S. natalensis stained for catalase activity. S. natalensis was grown in YEME medium and samples were collected during the defined four growth phases (see Material and Methods): RG1, RG2, S/P and S/NP. Arrows show the two detectable catalase activity bands. (C) Native PAGE stained for catalase activity of cell extracts from S. natalensis ATCC 27448 cultures collected during RG1 phase after 1 mM H₂O₂ insult (+).

Regarding ROS homeostasis, catalases are key enzymes that have the ability to scavenge H₂O₂. The catalase activity profile of S. natalensis was monitored along the growth curve in YEME liquid medium (Fig. 1A). Total catalase activity gradually decreased during the exponential phase and at early stationary phase presented low levels (<2 U mg⁻¹). These low levels of catalase activity were maintained until mid-stationary phase and then a continuous increase was observed until the late stationary phase. Like S. coelicolor [24], [25] and S. avermitilis (Fig. S1), S. natalensis presents a growth-dependent behaviour of catalase activity. However, whereas S. coelicolor and S. avermitilis present a constant increase in catalase activity from the early exponential to the late stationary phases, S. natalensis profile showed low levels of catalase activity during RG2 and S/P phases, the pimaricin production phases. Furthermore, when compared to other Streptomyces strains, in particular S. coelicolor and S. avermitilis, the overall levels of total catalase activity in S. natalensis are considerably lower (Fig. S1).

Cell free protein extracts were also analysed by catalase activity staining after native PAGE (Fig. 1B). Two activity bands were detectable suggesting the expression of at least two catalase enzymes in S. natalensis under the conditions tested (Fig. 1B). A lower molecular weight band whose intensity increased during exponential phase up to RG2 phase and disappeared as the culture entered into the stationary phase. During stationary phase, a new activity band was induced contributing to the increase in total catalase activity quantified spectrophotometrically. When gels were stained for peroxidase activity, no activity was detected under the conditions tested (data not shown). To further confirm the absence of a bifunctional catalase peroxidase encoding gene, a common H₂O₂-dismutating enzyme in this bacterial group, a Southern and an immunoblot analysis were performed on genomic DNA (an internal probe from S. coelicolor catalase-peroxidase encoding gene was used) and on cytoplasmatic and extracytoplasmatic protein extracts from S. natalensis (using a primary antibody raised against a catalase-peroxidase purified from S. reticuli [26]) respectively. In both cases no positive signal was detected (data not shown).

Concerted regulation of katA1 and ahpC in response to H₂O₂

To study the response of S. natalensis ATCC 27448 to exogenous H₂O₂, total specific catalase activity was measured in cultures challenged with H₂O₂. Three different H₂O₂ concentrations were tested, 0.1 mM, 1 mM and 10 mM. Addition of 10 mM H₂O₂ at RG1 phase severely impaired growth of S. natalensis, and catalase activity induction with 1 mM H₂O₂ was 1.7-fold higher than with 0.1 mM (data not shown). Therefore, studies of catalase activity inducibility were carried out adding 1 mM H₂O₂ to the culture broth at RG1, RG2, S/P and S/NP phases. Specific catalase activity was measured in protein extracts collected 1 h after H₂O₂ insult (Fig. 2A). Addition of H₂O₂ prompted an increase in total catalase activity except in the RG2 phase where no induction of catalase activity was observed. Additionally, analysis of the protein extracts from RG1 phase by native PAGE revealed that the addition of H₂O₂ induces the catalase that is active during the stationary phase (Fig. 1C).

Figure 2. S. natalensis response to H₂O₂ induced stress.

(A) Induction of catalase activity by addition of H₂O₂ at different growth stages in control and iron supplemented cultures. Protein extracts were prepared from S. natalensis cultures collected 1 hour after 1 mM H₂O₂ (final concentration) treatment or an equal volume of water as the untreated control. Induction of catalase activity by H₂O₂ treatment was assessed at the four defined growth stages (see Material and Methods): RG1, RG2, S/P and S/NP. Results (average of triplicates and standard deviation) are representative of three independent experiments. The differences between treated and untreated samples at each time point were assessed by independent t-test. *, statistically significant (P<0.01). (B) Induction of AhpC expression by H₂O₂ induced stress. Protein extracts were analysed by comparative 2D gel electrophoresis to assess the expression profile of the S. natalensis proteome under induced stress conditions. The protein profile of the identified AhpC spot (indicated by an arrow) is enlarged. Fold variations of the AhpC spot between H₂O₂-treated and untreated samples are indicated in the top left of the H₂O₂-treated panels. Fold variations reflect the 2D-gel analysis of three independent experiments.

To further analyze the changes triggered by H₂O₂ at the proteome level two dimensional (2D) gel electrophoresis was carried out to compare the proteome of H₂O₂-treated and non-treated cultures. Protein expression differences of the 2D gels were analyzed using appropriate software tools (PDQuest, Biorad). The analysis revealed 43, 81, 121 and 39 differences in the RG1, RG2, S/P and S/NP phases, respectively. Among those changes it was detected, in H₂O₂-treated cultures, an 17.7- (during RG1 phase) and an 3.7-fold (in RG2 phase) induction of the alkyl hydroperoxide reductase (AhpC) protein, a component of the alkyl hydroperoxidase reductase system which exhibits activity against H₂O₂, organic peroxides and peroxynitrite (Fig. 2B). These results suggest that the H₂O₂ added at RG2 phase was detoxified through the action of the alkyl hydroperoxidase reductase system.

In S. coelicolor, the induction of the catalase by H₂O₂ is regulated in an iron-dependent manner by the fur-like protein CatR [7]. To assess the sensitivity of catalase induction to H₂O₂ in the presence of an excess of iron, total catalase activity was measured in iron-supplemented (20 µM FeSO₄) cultures with and without the addition of H₂O₂ (Fig. 2A). Under these conditions, during RG1 and S/P phases the same behaviour was observed regarding catalase response to H₂O₂. However, induction of catalase activity was also observed at the RG2 phase whereas the protein levels of AhpC did not change significantly (Fig. 2B). These results suggest an iron dependent crosstalk between the regulation of these two main H₂O₂ scavenging enzymes. RT-qPCR analysis was performed to confirm this concerted regulation between the catalase and the alkyl hydroperoxidase reductase system (Fig. 3). The results showed that the increase in total catalase activity during the exponential phase was due to an increase of the transcript levels of katA1. The transcripts of the H₂O₂-sensing regulator catR were also increased, suggesting the presence of a similar mechanism to that of S. coelicolor [7]. The ahpC and oxyR (the H₂O₂-sensing transcriptional regulator of the AhpCD system; see below) transcriptional profiles in iron-supplemented cultures (Fig. 3), are reflected in the AhpC protein expression profile (Fig. 2B) i.e. in the presence of iron, during RG2 phase, H₂O₂ was detoxified mainly by the action of the catalase.

Figure 3. Transcriptional analysis of ahpC, oxyR, katA1 and catR.

Transcription profiles of the monofuncional catalase (katA1) and alkyl hydroperoxide reductase (ahpC) encoding genes and their transcriptional regulators, catR and oxyR, in S. natalensis ATCC 27448 upon a H₂O₂ insult. The transcription of ahpC, oxyR, katA1 and catR was evaluated by RT-qPCR from S. natalensis grown in YEME medium and samples collected 30 min after 1 mM H₂O₂ treatment (+) or an equal volume of water as the untreated control (-).H₂O₂ treatments were applied independently during the early exponential phase (RG1) and late exponential phase (RG2). The Mean Normalized Fold Expression (±standard errors) of the target genes was calculated relative to the transcription of the reference genes (16 S rDNA and lysA) and the reaction internal normalization was performed using the sample from cells collected immediately before H₂O₂ addition to the culture broth (not shown). Results (average of triplicates and standard deviation) are representative of three independent experiments.

The adaptive response triggered by H₂O₂ modulates pimaricin biosynthesis

The adaptive response prompted by the addition of H₂O₂ in iron-supplemented cultures reflected in an increase in the total catalase activity, particularly during exponential phase. In E. coli, catalase is described as a more efficient enzyme at high levels of H₂O₂, whilst AhpC being a more efficient scavenger of trace H₂O₂ generated endogenously [27]. Induction of total catalase activity in iron-supplemented cultures reflected in a transient and reversible decrease in H₂O₂ intracellular levels, particularly during RG2 phase when a decrease of 25.6% in H₂O₂ intracellular concentration, 1 h after the H₂O₂ insult, was observed (data not shown). To determine whether the pimaricin production in S. natalensis is affected by the temporary redox imbalance created by the adaptive response to H₂O₂, pimaricin yields at 72 h were measured in the culture broths of H₂O₂-induced iron supplemented cultures (Fig. 4). When the H₂O₂ stimulus was introduced during RG1 phase no significant alteration in pimaricin yield was observed. However, when H₂O₂ stimulus was introduced during late exponential phase (RG2) the pimaricin yield was reduced to 62.5% of the production in the control culture (without the H₂O₂ stimulus). Although less pronounced, a decrease in pimaricin production (ca 10%) was also observed in cultures with H₂O₂-stress induced at S/P phase.

Figure 4. Pimaricin specific production in H₂O₂ growth dependent induced cultures.

S. natalensis ATCC 27448 was grown in iron-supplemented YEME medium, and pimaricin specific yield (per mg of total protein) measured at 72 h. 1 mM (final concentration) H₂O₂ was added to the culture broth either at the RG1, RG2 or S/P phase. Data are the means from three independent experiments. To assess the presence of significant differences between the tested growth phases, a one-way ANOVA was performed followed by post-hoc test (Tukey test; GraphPad Prism) in which each condition was compared to the control situation (pimaricin production without H₂O₂ addition; 100%). *, statistically significant (P<0.01); ns, not statistically significant.

Cloning of genes related with H₂O₂ from S. natalensis

The results from the native PAGE and 2D analyses showed the existence in the S. natalensis ATCC 27448 genome of genes coding for at least one H₂O₂ inducible monofunctional catalase (Fig. 1B), a Fe and Ni-containing SOD (Fig. 5A) and an alkyl hydroperoxide reductase (Fig. 2B). To clone those encoding genes we screened a S. natalensis ATCC 27448 cosmid library [10] with probes KA1, internal to katA1, AHP internal to ahpC and SF internal to sodF (see Materials and Methods).

Figure 5. Lack of SOD activity in strain CAM.02 decreases intracellular H₂O₂ levels.

(A) Native PAGE of cell extracts (30 µg protein per lane) from S. natalensis ATCC 27448 (upper panel) and S. natalensis CAM.02 (lower panel) stained for SOD activity. S. natalensis strains were grown in YEME medium and cells collected on the four previously defined growth stages. NiSO₄ 20 µM (final concentration) was added to the YEME medium for the induction of sodN expression. (B) Intracellular H₂O₂ levels in S natalensis ATCC 27448 (WT) and SodF defective mutant (CAM.02) at RG2 growth phase. Values are means from two independent experiments. To assess the presence of significant differences between the tested condition, a one-way ANOVA was performed followed by post-hoc test (Tukey test; GraphPad Prism) in which each condition was compared to the control situation (wild-type strain intracellular H₂O₂ levels in Ni non-supplemented cultures; 100%). *, statistically significant (P<0.01).

A 2.4 kb Nco I fragment from cosmid A1, was found to hybridize with probe KA1. The fragment was cloned into Nco I-digested pGEM^®-T Easy vector and sequenced. The fragment was 2430 bp and in silico analysis of the fragment showed the presence of two complete open reading frames (ORFs), katA1 (1464 bp) and catR (423 bp), divergently transcribed and separated by 161 bp (Fig. S2A). Both genes presented an overall codon usage pattern in good agreement with that of typical Streptomyces genes. The product of katA1 (487 amino acids with a deduced molecular mass of 54 kDa) showed high sequence identities with other Streptomyces monofunctional catalases particularly with the whole protein of SAV_3052 (87% identity), a putative catalase from S. avermitilis MA-4680. Like most monofunctional catalases, two conserved domains are identified in KatA1: in the N-terminal region of the protein it is found the catalase active site motif (amino acids 47 to 63) and in the C-terminal region the heme-ligand motif (amino acids 337 to 345). Upstream and divergently transcribed from katA1, catR was identified. CatR protein is 140 amino acids long and showed high sequence identity with other Streptomyces PerR like members of the Fur family of proteins (78% to 83% identity), in particular with hydrogen peroxide sensitive repressors. It contains the four cysteine residues conserved in PerR like proteins (C⁹²-X₂-C⁹⁵ and C¹³²-X₂-C¹³⁵) involved in structural Zn²⁺ binding as well as the two histidine residues (H³² and H⁸⁷) involved in Fe²⁺ (or Mn²⁺) coordination [28], [29].

A 1.8 kb Apa I fragment from cosmid SF6 was found to hybridize with probe SF internal to sodF. The fragment was cloned into Apa I-digested pGEM^®-T Easy vector and sequenced. The fragment was 1775 bp and in silico analysis of the fragment showed the presence of one complete ORF, sodF (642 bp) (Fig. S3A). Comparison of the protein encoded by sodF (213 aa) with the non-redundant protein sequence database (NCBInr) revealed high sequence identity (91 to 95% identity over the whole protein sequence) with FeZn superoxide dismutase proteins from Streptomyces. Analysis of the deduced amino acid sequence of SodF revealed the presence of the four conserved residues involved in binding to the metal cofactor (H²⁸, H⁷⁶, D¹⁶⁵ and H¹⁶⁹) [30], two of them included in the SOD conserved motif [D¹⁶⁵-A-W-E-H-A-F-Y¹⁷²] [31]. Furthermore, the upstream genomic region of the sodF gene revealed the presence of the Ni-responsive regulatory motif (TTGCAN₇TGCAA) suggested to be involved in the nickel dependent transcriptional repression of sodF [32].

A 3.75 kb Bam HI fragment from cosmid X10 was found to hybridize with probe AHP internal to ahpC. The fragment was cloned into Bam HI-digested pUC19 vector and sequenced. The nucleotide sequence analysis of the fragment revealed three complete ORFs: ORF1 (627 bp), oxyR (960 bp), ahpC (555 bp) and ahpD (534 bp) (Fig. S4A). The deduced protein sequences of oxyR, ahpC and ahpD showed high sequence identities with counterparts from Streptomyces, in particular a hydrogen peroxide sensing regulator of the LysR-family, an alkyl hydroperoxide reductase and an alkylhydroperoxidase respectively. The three conserved catalytic cysteine residues described to be involved in the peroxidatic activity of Mycobacterium tuberculosis AhpC [33] were identified in S. natalensis AhpC (C⁵¹, C¹⁶⁴ and C¹⁶⁶). Additionally, the conserved motif CXXC of electron transport proteins was identified in S. natalensis AhpD (C¹³¹ and C¹³⁴). ORF1 codes for a putative protein of unknown function that shows high identity with protein SGR3203 from S. griseus.

Knock-out mutants on the identified anti-oxidant encoding genes were constructed by replacing the target gene with an apramycin resistance cassette using the PCR targeting method (see Material and Methods) [34]. Three mutant strains were isolated: CAM.02 (ΔsodF::aac(3)IV-oriT), CAM.04 (ΔahpCD::aac(3)IV-oriT) and CAM.05 (ΔkatA1::aac(3)IV-oriT) (Table 1) and their identity was confirmed by Southern blot and PCR analyses (Supplementary material; Fig. S2, S3, S4). In the case of CAM.05 (ΔkatA1) mutant, to further confirm the identity of KatA1 as being inducible by H₂O₂, CAM.05 strain was challenged with a H₂O₂ insult during RG1 phase. The mutant strain was unable to induce catalase activity confirming that the catalase encoded by katA1 is inducible by H₂O₂ (data not shown). Antifungal activity and the identity of pimaricin produced by the knock-out mutants were confirmed by bioassay and ultra performance liquid chromatography (UPLC), respectively (data not shown).

Table 1. Strains and plasmids used in this study.

Strain or plasmid	Description	Reference
Strain
S. natalensis
ATCC 27448	Wild-type strain	[62]
CAM.02	sodF loci replaced by the oriT-aac(3)IV cassette; ΔsodF::aac(3)IV-oriT	This study
CAM.04	ahpCD loci replaced by the oriT-aac(3)IV cassette; ΔahpCD::aac(3)IV-oriT	This study
CAM.05	katA1 loci replaced by the oriT-aac(3)IV cassette; ΔkatA1::aac(3)IV-oriT	This study
E. coli
DH5α	General cloning strain	[63]
ET12567	Non-methylating strain used for conjugation with Streptomyces	[64]
BW25113	Strain used for PCR-targeted mutagenesis	[65]
BT340	DH5a [pCP20]	[66]
Cosmids & Plasmids
pGEM-T Easy	General cloning vector	Promega
pSET152neo	pSET152 derivative, neo	[17]
pIJ773	oriT aac(3)IV	[34]
pUZ8002	Mobilization plasmid; neo	[67]
A1	Cosmid from genomic library of S. natalensis ATCC 27448; katA1 catR	This study
SF6	Cosmid from genomic library of S. natalensis ATCC 27448; sodF	This study
X10	Cosmid from genomic library of S. natalensis ATCC 27448; oxyR ahpCD	This study

All three mutants grew normally in liquid and solid media. However, the CAM.05 mutant was unable to form aerial mycelia and to sporulate in solid media. This phenotypic trait hampered strain manipulation, in particular mutant complementation and pimaricin production assays. DNA delivery methodologies such as transformation or intergeneric conjugation using mycelium have been unfruitful for S. natalensis [35] and additionally, pimaricin production yields are significant lower when vegetative mycelium is used as inoculum [36]. This morphological phenotype is different from that reported for the S. coelicolor mutant strain YD9147, a catA defective mutant [37], that sporulated normally on R2YE medium.

Loss of SOD activity decreases pimaricin production

Superoxide dismutase (SOD) enzymes catalyze the dismutation of the superoxide anion radical into H₂O₂ and O₂. Two types of SODs have been characterized in Streptomyces, a nickel containing SOD (NiSOD) and both iron and zinc containing SOD (FeZnSOD) [38], [39]. Native PAGE analysis of S. natalensis crude extracts (Fig. 5A) showed a band with SOD activity corresponding to a FeZnSOD expressed constitutively throughout the growth curve. It has been reported a nickel dependent regulatory system that controls the expression of either the NiSOD or the FeZnSOD in S. coelicolor and S. griseus, i.e., in the presence of nickel ions NiSOD is preferentially expressed whereas FeZnSOD expression is down-regulated [39], [40]. The presence of the Ni-responsive regulatory motif upstream from sodF suggests a similar regulation mechanism in S. natalensis. Indeed, in NiSO₄ (20 µM) supplemented cultures, a second band displaying SOD activity can be observed in S. natalensis (Fig. 5A) suggesting the presence of a NiSOD. Moreover, the transcription of sodF was repressed in Ni-supplemented cultures as assessed by RT-qPCR (data not shown). When protein extracts from the sodF defective mutant (CAM.02) were analyzed by native PAGE, no FeZnSOD band was detected but it was still observed the band corresponding to the NiSOD in Ni-supplemented cultures (Fig. 5A).

SOD enzymes are key enzymes involved in intracellular ROS homeostasis. Deleting the FeZnSOD encoding gene created an imbalance in the intracellular ROS concentration of CAM.02 strain, throughout the growth curve. In particular, intracellular H₂O₂ levels at RG2 phase (when pimaricin biosynthesis begins) decreased 33% in CAM.02 mutant when compared to the wild-type. Ni-supplementation of the CAM.02 culture broth restored H₂O₂ to wild-type levels (Fig. 5B).

Quantification of pimaricin at 72 h in culture broths of CAM.02 strain grown in YEME medium, showed that the pimaricin specific production of the mutant strain was 8.2% of the wild-type. Interestingly, the pimaricin specific production was increased in CAM.02 mutant grown in Ni-supplemented medium (34.3% of the wild-type). Complementation of the CAM.02 with sodF under its own promoter restored pimaricin production to wild-type levels (Table 2).

Table 2. Pimaricin production of S. natalensis mutant strains.

S. natalensis strain (genotype)	Conditions	Pimaricin production(% of WT)a
CAM.02 (ΔsodF:: oriT-aac(3)IV)	-	8.2±1.45
	20 µM NiSO4	34.3±0.99
	pSET-sodF	86.0±0.22b
CAM.04 (ΔahpCD:: oriT-aac(3)IV)	-	129.9±4.66
	20 µM Fe2+	156.3±5.60
	10 mM H2O2 @ RG2	169.4±7.82
	20 µM Fe2+ + 10 mM H2O2 @ RG2	174.1±0.23
	pSET-ahpCD	101.2±0.76b
CAM.05 (ΔkatA1:: oriT-aac(3)IV)	-	156±4.05c
	1 mM H2O2 @ RG2	203.8±9.41c

^aValues are means of triplicates from three independent experiments; 100% represents the pimaricin production of S. natalensis ATCC 27448 at 72 h in YEME medium.

^b100% represents pimaricin production of S. natalensis harbouring pSET152neo at 72 h in YEME medium.

^cPimaricin production of S. natalensis CAM.05 strain was assessed at 72 h in YEME medium inoculated with vegetative mycelium. 100% represents the pimaricin production of S. natalensis ATCC 27448 at 72 h in YEME medium inoculated with vegetative mycelium.

Mutants defective in either H₂O₂ detoxification enzymes show a pimaricin overproducer profile

To counteract the increase in intracellular H₂O₂ levels, microorganisms have enzymes that present H₂O₂-scavenging activity such as catalases and the alkyl hydroperoxide reductase system. The S. natalensis mutants defective on such H₂O₂-detoxifying enzymes, CAM.04 (ΔahpCD) and CAM.05 (ΔkatA1) showed different phenotypic traits but both behaved as pimaricin overproducers, reaching 130% and 156% of the pimaricin produced by the wild-type strain, respectively (Table 2). Also, it is worth noting that the pimaricin overproducer behaviour of the mutant strains, CAM.04 and CAM.05, was enhanced by the addition of H₂O₂ to the culture broths during RG2 phase (Table 2). Finally, complementation of the CAM.04 with ahpCD under its own promoter restored pimaricin production to the wild-type levels (Table 2).

The intracellular H₂O₂ levels of the S. natalensis CAM.04 (ΔahpCD) strain were higher than those of the wild-type strain, particularly during RG1 phase with a 3-fold increase (Fig 6C). During RG2 and S/P growth phases the H₂O₂ levels were 30% to 40% above the wild-type strain levels (Fig. 6C). Concomitantly with the high H₂O₂ intracellular levels, the total catalase specific activity, namely during the exponential growth phases (RG1 and RG2), had a significant boost, reaching a 30-fold increase in RG1 phase when compared to the wild-type (Fig. 6A). Native PAGE analysis showed that the increase in catalase activity is due to the increase of the levels of the catalase inducible by H₂O₂ (data not shown). Regarding total specific SOD activity, CAM.04 showed a similar profile to the wild-type, although with lower activities (Fig. 6B).

Figure 6. Characterization of S. natalensis CAM.04 strain.

(A) Total catalase specific activity, (B) SOD specific activity and (C) intracellular H₂O₂ levels of S. natalensis wild-type (white bars) and S. natalensis CAM.04 (black bars) at the four defined growth stages: RG1, RG2, S/P and S/NP. Results (average of triplicates and standard deviation) are representative of three independent experiments. The presence of statistically significant differences between wild-type (WT) and CAM.04 samples was determined by t-test. *, statistically significant (P<0.01).

The high levels of intracellular H₂O₂ present a vulnerability to the CAM.04 strain because of the H₂O₂ reduction to the highly toxic HO^• by the metal-catalysed Fenton reaction. To determine the outreach of such toxicity regarding pimaricin biosynthesis, CAM.04 strain was grown in iron-supplemented medium and pimaricin specific production was assessed with and without the addition of H₂O₂ to the culture broth (Table 2). In both situations the overproducer phenotype of the CAM.04 strain persisted (174% and 156% of the pimaricin produced by the wild-type, respectively).

Discussion

Optimization of the strain ideal aeration conditions in submerged cultures is an essential step of industrial fermentation processes as microbial secondary metabolism is strongly affected by oxygen availability. Previous studies have shown that pimaricin production is highly dependent on dissolved oxygen levels, namely an increase in culture aeration lead to higher yields of pimaricin [36]. However, an increased consumption of O₂ can generate ROS that would trigger an adaptive response in order to sense and scavenge ROS and repair the damage in cell components. In this work we addressed the role of ROS homeostasis, in particular intracellular H₂O₂ levels, and of the adaptive response triggered by H₂O₂, on the production of pimaricin by S. natalensis ATCC 27448.

Catalases play a key role in ROS homeostasis due to their ability to perform the dismutation of H₂O₂ into water and molecular oxygen. Analysis of catalase activity in crude extracts of S. natalensis showed that, under the conditions tested, two different catalases contribute to the specific total catalase activity profile presented by S. natalensis. The expression of both catalases is growth-phase dependent (Fig. 1B): while one catalase was mainly active during the exponential growth phases (RG1 and RG2) the second catalase was present during the stationary phase. Characterization of the response of S. natalensis wild-type and CAM.05 (ΔkatA1) strains to exogenous H₂O₂ allowed us to confirm that the stationary phase specific catalase from S. natalensis (KatA1) is inducible by H₂O₂. However, its behaviour regarding the response to H₂O₂ was considerably different from that of S. coelicolor [24], [41]. During RG2 phase no induction of catalase activity was observed upon the addition of H₂O₂. Instead, an increase in the expression of the alkyl hydroperoxide reductase protein (AhpC) suggested a detoxification of H₂O₂ by the alkyl hydroperoxide system (AhpCD). Conversely, in iron supplemented cultures, an induction of total catalase activity is achieved during RG2 phase, whereas no induction of the AhpC expression is observed. Thus, in the presence of an excess of iron, H₂O₂ was being detoxified mainly by the action of the monofunctional catalase, KatA1. In S. coelicolor, the two main H₂O₂ scavenging enzymes (CatA and AhpCD) are regulated independently by two proteins able to sense H₂O₂, CatR and OxyR [42], respectively. During the early exponential phase (before the metabolic switch) and stationary phase, S. natalensis counteracts the addition of H₂O₂ by inducing catalase activity. However, during the late exponential phase (RG2), in order to maintain H₂O₂ homeostasis, S. natalensis activates the AhpCD system upon a H₂O₂ challenge instead of catalase, which has been described as a more effective enzyme at high concentrations of H₂O₂ [27].

Assuming similar tertiary structures between the H₂O₂-sensing regulators from S. natalensis (CatR_SNA) and S. coelicolor (CatR_SCO) and PerR from Bacillus subtilis , and taking into consideration the postulated models for CatR_SCO and PerR DNA binding mechanisms [7], [28], the behaviour of CatR_SNA in iron-supplemented experiments led us to hypothesise about its mechanism. We suggest that the mechanism of sensing H₂O₂ and DNA binding activity of CatR_SNA is not through disulfide bond formation between the C-terminal cysteine residues, as proposed for S. coelicolor [7], but in a metal-dependent manner through the metal-catalyzed oxidation (MCO) of the histidine residues (H³² or H⁸⁷), responsible for iron coordination as postulated for PerR in B. subtilis [28]. The higher availability of Fe²⁺ in iron-supplemented cultures increases the interaction of CatR_SNA with Fe²⁺, enhancing CatR_SNA susceptibility to MCO by H₂O₂ thus increasing katA1 transcription. However, this needs to be further investigated experimentally to fully elucidate the CatR_SNA H₂O₂ sensing and DNA binding mechanism.

The induction of KatA1 activity during RG2 phase by addition of H₂O₂ in iron-supplemented cultures lowered the intracellular H₂O₂ levels. However the effects of catalase induction went beyond the adaptive response towards H₂O₂ and reflected in a decrease in the specific production of pimaricin by S. natalensis. Although it was observed a growth phase dependent response, regarding pimaricin biosynthesis, upon the addition of H₂O₂ in iron-supplemented cultures, the harmful effect of hydroxyl radicals generated by the Fenton reaction may also contribute to the decrease in the pimaricin specific production during RG2 phase. Nevertheless, these results suggest that the intracellular levels of ROS are able to modulate the biosynthesis of pimaricin. Previous studies had suggested that ROS, in particular H₂O₂, play an important role as secondary messengers on cell signalling based on reductive-oxidative mechanisms [4], [43]. We hypothesize that in iron-supplemented cultures, the “forced” activation of a monofunctional catalase during RG2 phase, detoxified H₂O₂ to levels below those of the wild-type strain, altering the secondary metabolism regulatory network at the pleiotropic or at the pathway-specific level. Interestingly, the pimaricin biosynthetic gene cluster in S. natalensis comprises the positive regulator PimM which has an N-terminal PAS domain [14]. PAS sensor domains are able to sense a wide range of signals including cellular redox status [44]. This feature turns PimM into a possible (in)direct target of a redox-dependent regulation.

To test the possibility of a regulation of secondary metabolism by intracellular H₂O₂, a functional approach was pursued with the construction of knock-out mutants in genes coding for hydrogen peroxide related enzymes. The construction of the mutant strains defective in anti-oxidative enzymes (SodF, KatA1 or AhpCD), allowed an effective modulation in H₂O₂ intracellular concentration in contrast to the punctual imbalance created by the addition of H₂O₂ to the culture broth that elicits an adaptive response to restore normal cell redox status.

An important source of intracellular H₂O₂ is the dismutation of superoxide anion by SOD. In fact, the SodF defective mutant, S. natalensis CAM.02, showed lower levels of intracellular H₂O₂ along the growth phase, in particular during RG2 phase with a 30% decrease (Fig. 5B). Simultaneously, a drastic decrease in the production of pimaricin was also observed when compared to the wild-type. In contrast, supplementation of the culture broth with Ni²⁺ ions not only increased the expression of the Ni-containing SOD but also, presumably as a result of the NiSOD action, increased the intracellular H₂O₂ levels. These conditions increased the pimaricin specific production showing a possible correlation between H₂O₂ intracellular levels and pimaricin production in S. natalensis. However it is expected an increase in the intracellular levels of the superoxide anion in the CAM.02 mutant. As a consequence, part of the negative effect observed in the pimaricin production could be related to the toxic effect of the superoxide anion particularly in the proteins containing iron-sulphur clusters. Recently, heterologous expression of a FeZnSOD and NiSOD from S. peucetius in S. clavuligerous and S. lividans increased the production of clavulanic acid and actinorhodin respectively [45]. The authors ascribe a positive effect of SODs in secondary metabolite production. The results obtained in S. natalensis lead us to go further on and suggest that the positive effect observed was due to an increase in the intracellular H₂O₂ levels.

Increased levels of intracellular H₂O₂ were observed, particularly during the exponential phase, in the H₂O₂ scavenging defective mutant, CAM.04 (ΔahpCD) (Fig. 6C). Probably due to high H₂O₂ levels, catalase activity in CAM.04 strain is induced as a compensatory effect. This concerted regulation between the two H₂O₂ scavenger enzymes has been already reported for Xanthomonas campestris pv phaseoli and Staphylococcus aureus [46], [47]. Although the activation of the H₂O₂-inducible catalase is clearly observed, the intracellular H₂O₂ levels in the mutant strain were still higher than in the wild-type. Simultaneously, a decrease in SOD activity is observed when compared to the wild-type. This can either be due to a compensatory effect towards re-establishing the cell redox status or to a SOD inactivation by H₂O₂ [39]. Regardless of the oxidative stress response triggered by the absence of AhpCD it should be noticed that CAM.04 presented a pimaricin overproducer phenotype that was reinforced by the addition of H₂O₂ to the culture broth during RG2 phase. A similar behaviour was observed in the CAM.05 (ΔkatA1) strain. On the other hand, the high levels of intracellular H₂O₂ will trigger the production of the highly toxic hydroxyl radical as a consequence of the oxidation of unincorporated intracellular ferrous iron (Fenton reaction). The effect that the hydroxyl radicals may have on the pimaricin biosynthesis was assessed in CAM.04 iron-supplemented cultures. Although a negative effect could be expected due to the toxic nature of the hydroxyl radicals, no decrease in pimaricin specific production was observed. Instead there was an increase in iron-supplemented cultures that could be explained by the optimization of iron conditions in the growth medium.

Finally, the bald phenotype exhibited by the CAM.05 strain unveils a correlation between the H₂O₂-inducible catalase, KatA1, and morphological differentiation of S. natalensis on solid culture. Unlike liquid cultures where total catalase activity increases during the mid-late stationary growth phase, S. natalensis presents high catalase activity levels (from KatA1) during and after the emergence of aerial mycelia on solid culture (Beites T.B., unpublished results). It was suggested that during the development of aerial mycelia cells undergo a transient state of oxidative stress [48], [49] that could result from the nutrition limitation and other signals that trigger morphological differentiation or from a higher exposure to oxygen by aerial mycelia. The absence of the H₂O₂ inducible catalase in CAM.05 severely hampers the ability of the strain to cope with induced oxidative stress which may prevent a normal morphological differentiation of S. natalensis on solid cultures.

From an industrial point of view, oxygen supply is a key parameter on the production of secondary metabolites. While low O₂ concentrations limit growth and product formation, increased O₂ concentration usually can improve secondary metabolite production [50], [51]. However when above a certain threshold, O₂ may have regulating and toxic effects on microbial cultures mainly due to induction of oxidative stress. Like other environment induced responses, the oxidative stress related mechanisms are embedded in global networks that extend the effect of the oxidative stress response into secondary metabolism. Thus, it is reasonable to assume that oxygen, or the by-products of its reduction, may have a regulatory effect on the Streptomyces secondary metabolism. The results obtained in this work point out to an important role of H₂O₂ intracellular levels towards the regulation of pimaricin biosynthesis, presumably through redox-based mechanisms. Maintaining H₂O₂ intracellular levels, particularly during late exponential – early stationary phases, proved to be important to elicit pimaricin biosynthesis. Although the punctual H₂O₂ modulation achieved by the addition of exogenous H₂O₂ suggested such a crosstalk, the correlation between H₂O₂ and secondary metabolism has been clearly shown by the characterization of the H₂O₂-related enzymes defective mutants, in which a long-term modulation of H₂O₂ levels is achieved.

Materials and Methods

Bacterial strains, plasmids and growth conditions

The strains, plasmids, and cosmids used in this study are listed in Table 1. S. natalensis was routinely grown by inoculating a spore suspension (3×10⁸ c.f.u.) into 100 mL of YEME medium [52] without sucrose and prepared with deionized water, in 1 L baffled flasks. Cultures were incubated in an orbital incubator shaker at 200 r.p.m. and 30 °C for 72 hours. Sporulation was achieved in TBO medium [10]. Induced oxidative stress studies were carried out by challenging cultures with 1 mM H₂O₂ over the different growth phases which were defined by the change in the rate of increase in cell density at 600 nm [21], [22]. Growth phases were defined as follows: early exponential (RG1) between OD_600nm 1.5 and 4; late exponential (RG2) between OD_600nm 5 and 8; stationary phase with pimaricin production (S/P) between 38 h and 48 h of culture and late exponential without pimaricin production (S/NP) at 72 h after inoculation. E. coli cells were routinely grown in LB medium.

Nucleic acids and protein procedures

Standard genetic techniques with Streptomyces, E. coli and in vitro DNA manipulations were carried out as previously described [52], [53]. Isolation of total DNA from Streptomyces was performed using the MasterPure^TM Gram Positive DNA Purification Kit (Epicentre). Southern hybridization was carried out with probes labelled with digoxigenin by using the DIG DNA labelling kit (Roche).

For S. natalensis total RNA isolation 1 ml-aliquots of cultures grown in YEME medium were mixed with two volumes of RNA Protect Bacteria Reagent (Qiagen). Mycelia were harvested by centrifugation and immediately frozen by immersion in liquid nitrogen. Buffer RLT (Qiagen) in the presence of 1.0% (v/v) β-mercaptoethanol was added to the frozen mycelium, mixed thoroughly and cells broken by sonication (Branson Sonifier, Model B-15). RNeasy Mini Spin columns were used for RNA isolation according to manufacturer's instructions. DNA was removed by two serial DNase treatments, an in-column DNase I RNase-free (Qiagen), followed by a batch treatment using the DNA-free Kit (Ambion). Total RNA concentration was determined with a NanoDrop ND-1000 spectrophotometer (Thermo Scientific), and RNA quality and integrity were checked in an Experion^TM Automated Electrophoresis System (Bio-Rad).

For S. natalensis crude protein extracts preparation, mycelia from 1 mL of culture were harvested by centrifugation. Cell pellets were washed once with 50 mM K-phosphate buffer, pH 6.8, and resuspended in 0.5 mL of the same buffer containing 25% (v/v) of a protease inhibitor (Roche). Cells were disrupted by sonication (Branson Sonifier, Model B-15), the lysate was centrifuged and the pellet discarded. Protein content of cellular extracts was determined by the BCA^TM Protein Assay Kit (Pierce) using bovine serum albumin as a standard.

Intergeneric conjugation between E. coli ET12567 [pUZ8002] and S. natalensis ATCC 27448 was performed as previously described [35].

Determination of catalase and superoxide dismutase activities

Enzymatic assays were performed to determine total catalase and superoxide dismutase (SOD) activity in S. natalensis protein extracts. Catalase activity was quantified spectrophotometrically by following the rate of decrease in absorbance at 240 nm caused by the disappearance of H₂O₂ [54]. The assay mix contained 30 µL of protein extract and 10 mM H₂O₂ in 50 mM phosphate buffer pH 6.8 in a final volume of 1 mL. Assays were carried out at 25 °C. One unit of enzyme activity is defined as the amount required for the conversion of 1 µmol substrate into product per min. SOD activity was assayed by the method previously described [39], [55]. One unit of enzyme activity was defined as the amount of SOD required to inhibit the reduction of nitroblue tetrazolium by 50% under the reaction conditions.

Catalase, peroxidase and superoxide dismutase activities of S. natalensis protein extracts were also analyzed on 7.5% –10% (w/v) non-denaturing polyacrylamide gels. Catalase activity was visualized on the gel using the method previously described [41], [56]. Peroxidase gel activity was detected using a mixture of chromogenic substrates [57]. SOD activity was detected on native gels by its ability to deplete superoxide, which can reduce nitroblue tetrazolium [39], [58].

Two-dimensional (2D) gel electrophoresis (PAGE)

2D electrophoresis was performed as previously described [59]. A total of 150 µg of crude protein extract was treated with 3% (v/v) of Benzonase Nuclease (Sigma) at 37 °C for 30 min, subjected to a clean-up process using the 2-D Clean up Kit (GE Healthcare), and used for IEF in 17-cm precast IPG strips (Bio-Rad) with linear pH gradient of 4.0–7.0 using a PROTEAN IEF Cell (Bio-Rad). Second dimension was run in 12.5% (w/v) SDS-PAGE gels using an Ettan Dalt apparatus (GE Healthcare) as recommended by the manufacturer. Gels were silver stained following an MS-compatible protocol [60]. LMW low molecular weight marker (GE Healthcare) was used as size markers. Image analysis was performed with biological triplicates by using the PDQuest 2-D analysis software (Bio-Rad).

Protein spots were excised from gels and digested with trypsin. Samples were analyzed using the 4700 Proteomics Analyzer MALDI-TOF/TOF (Applied Biosystems) as described previously (Pinho et al., 2009). Data were analyzed using GPS Explorer (Version 3.6; Applied Biosystems). The alkyl hydroperoxide reductase C, ahpC, protein spot was identified using combined data from PMF (Peptide Mass Fingerprint) and tandem mass (MS/MS) spectra. The Mascot (Matrix Science, UK) protein identification algorithm, which reflects these two levels of information, had a score associated with this identification of 155 with an expected value of 2.9e-009, and a confidence interval of 100%. Two tryptic peptide peaks have been selected for MS/MS peptide sequencing. The sequences of the fragmented peptides in the identified protein were LNDEFADR and ALGIEGEDGFAQR.

Quantification of intracellular reactive oxygen species (ROS)

Levels of intracellular H₂O₂ were detected with dihydrorhodamine 123 (DHR) (Invitrogen). Aliquots from the culture broth were taken at selected time points, pellets were resuspended in 50 mM K-phosphate buffer pH 6.8 and DHR was added to a final concentration of 15 µg ml⁻¹. Cells were incubated for 60 min at 30 °C in the dark. Mycelia were washed twice in 50 mM K-phosphate buffer pH 6.7 and cells were broken by sonication. The amount of ROS was quantified with a spectrofluorometer (Fluoromax-4, Horiba) emitting at 504 nm and measuring at 534 nm. Protein content of the crude extracts was used as normalization factor.

PCR and RT-qPCR

DNA amplification by PCR was performed with GoTaq Flexi DNA Polymerase (Promega) or Pfu DNA polymerase (Fermentas) according to the manufacturer's instructions. For gene expression studies 1 µg of DNase I-treated (DNA-free Kit, Ambion) total RNA was transcribed with the iScript™ Select cDNA Synthesis Kit (Bio-Rad), using the random primers supplied, and following the manufacturer's instructions. qPCR amplifications were performed using the primer pairs listed in Table S1 and using 0.2 µM of each primer,10 µl of iQ™ SYBR® Green Supermix (Bio-Rad) and 2 µl of template cDNA. qPCRs were carried out in the iCycler iQ5 Real-Time PCR detection system (Bio-Rad) and conditions were as follows: 95°C for 5 min, 40 cycles of 95°C for 30 sec, 55, 60 or 65 °C (depending of the set of primers used) for 30 sec, and 72 °C for 30 sec. Standard dilutions of the cDNA were used to check the relative efficiency and quality of primers. Negative controls (no template cDNA) were included in all qPCR. A melting curve analysis was performed at the end of each qPCR assay to exclude the formation of nonspecific products. The data obtained were analyzed using the method described in Pfaffl [61]. For each analysis 16 S rRNA and lysA were used for normalization. The identity of each amplified product was corroborated by sequencing the PCR product.

Gene identification, cloning and sequencing

For the identification of the antioxidant enconding genes (monofunctional catalase, SOD and the alkyl hydroperoxidase system) of S. natalensis, internal hybridization probes were obtained by PCR amplification using genomic DNA of S. natalensis as template. Primer pairs were designed based on alignments of known nucleotide sequences of these proteins from S. avermitilis, S. coelicolor and S. griseus. Once confirmed the genetic identity of the PCR products by sequencing, they were labelled with digoxigenin and used as probes for screening a S. natalensis ATCC 27448 cosmid library [10].

Plasmid DNA was isolated from E. coli cultures using the GenElute™ Plasmid Miniprep Kit (Sigma-Aldrich, Saint Louis, MO), and sequenced at STAB Vida (Lisbon). Each nucleotide was sequenced a minimum of three times independently on both strands. Alignment of sequence contigs was performed using the Vector NTI ContigExpress program (Invitrogen). DNA and protein sequences were analyzed with the NCBI World Wide Web BLAST server.

Published sequences were retrieved from GenBank and computer-assisted sequence comparisons were performed using Vector NTI Advance 10 (Invitrogen).

Construction of mutant strains and complementation

The strategy used for gene disruption was based on the PCR targeting system developed in S. coelicolor [34]. The coding sequences of sodF, ahpCD and katA1 genes from S. natalensis ATCC 27448 were replaced by the aac(3)IV/oriT cassette from plasmid pIJ773. The primers used for amplifying the cassette are listed in Table 2. The PCR targeting strategy originated mutant stains CAM.02, CAM.04 and CAM.05 respectively. The identity of all mutants was confirmed by Southern hybridization and PCR. Complementation of defective mutants was carried out by inserting into the integrative vector pSET152neo a DNA fragment containing the wild-type gene (either ahpCD or sodF) plus their own promoters. The originating plasmids (pSETahpCD and pSETsodF) were transfered to S. natalensis CAM.04 and CAM.02 respectively, by conjugation. pSET152neo was also introduced into S. natalensis wild-type as control.

Pimaricin Production

The production of pimaricin in liquid cultures was routinely quantified by spectrophotometric determination at 304 nm. A 100 µL aliquot of culture was extracted with 8 vol of butanol, and the organic phase was diluted in water-saturated butanol to bring the absorbance at 304 nm in the range of 0.1–0.8 U. Pimaricin was quantified using a solution of pure pimaricin (Calbiochem) as standard. To confirm the identity of pimaricin, a UV–visible absorption spectrum (absorption peaks at 319, 304, 291 and 281 nm) was routinely determined. The fungicidal activity of pimaricin was tested by bioassay using C. utilis CECT 1061 as test organism. In addition to spectrophotometric determination, mutants broth extracts were also analysed by UPLC using a Waters ACQUITY System coupled to a PDA detector, fitted with a reverse-phase BEH C18 column (2.1×50 mm, particle size, 1.7 µm.). Elution was performed with a methanol/water gradient (0.4 ml/min) according to the following program (methanol concentration): 50% 0–0.73 min, up to 90% 0.73–2.90 min, 90% 2.90–4.83, down to 50% 4.83–6.02 min and 50% 6.02–7.21 min. Under these conditions pimaricin eluted at 2.60 min.

Nucleotide sequence accession numbers

The sequence data associated with this study have been submitted to the GenBank database under accession numbers JN005772, JN005773 and JN005774.

Supporting Information

Figure S1

Total catalase activity of S. natalensis ATCC 27448, S. coelicolor M145 and S. avermitilis in YEME medium. Samples were collected at the four defined growth phases (see Experimental Procedures): RG1, RG2, S/P and S/NP. Results (average of triplicates and standard deviation) are representative of three independent experiments.

(TIF)

Figure S2

Construction of strain CAM.05 by gene replacement of katA1 . A) Predicted restriction enzyme polymorphism caused by gene replacement. The NcoI-XbaI restriction pattern before and after replacement is shown. The probe used for southern hybridization is indicated by thick lines. N, NcoI; X, XbaI. B) Confirmation of gene disruption by PCR. A pair of primers, cKatA_F and ckatA_R, covering the deleted region in the chromosome were used for quick screening to identify double crossover mutants. C) Confirmation of gene disruption by Southern hybridization of the NcoI-XbaI digested chromosomal DNA of the wild type (lane 1), and CAM.05 (ΔkatA1; lane 2) strains. Lane M, DIG-labeled DNA Molecular Weight Marker II (Roche). Extra bands hybridizing on the Southern blot are cross-hybridization with other genomic fragments.

(TIF)

Figure S3

Construction of strain CAM.02 by gene replacement of sodF . A) Predicted restriction enzyme polymorphism caused by gene replacement. The NotI restriction pattern before and after replacement is shown. The probe used for southern hybridization is indicated by thick line. A, Apa I; N, NotI. B) Confirmation of gene disruption by PCR. A pair of primers, CsodF-F and CsodF-R, covering the deleted region in the chromosome were used for quick screening to identify double crossover mutants. C) Confirmation of gene disruption by Southern hybridization of the NotI digested chromosomal DNA of the wild type (lane 1), and CAM.02 (ΔsodF; lane 2) strains. Lane L, DIG-labeled DNA Molecular Weight Marker II (Roche).

(TIF)

Figure S4

Construction of strain CAM.04 by gene replacement of ahpCD . A) Predicted restriction enzyme polymorphism caused by gene replacement. The PvuII restriction pattern before and after replacement is shown. The probe used for southern hybridization is indicated by thick lines. B, BamHI; N, NcoI; P, PvuII. B) Confirmation of gene disruption by PCR. A pair of primers, Cahp-F and Cahp-R, covering the deleted region in the chromosome were used for quick screening to identify double crossover mutants. C) Confirmation of gene disruption by Southern hybridization of the PvuII digested chromosomal DNA of the wild type (lane 1), and CAM.04 (ΔahpCD; lane 2) strains. Lane L, DIG-labeled DNA Molecular Weight Marker II (Roche).

(TIF)

Table S1

Primers used in this work.

(PDF)