Streptomyces strains used in this study are: S. albus J1074, S. argillaceus ATCC 12596, S. coelicolor M145, S. glaucescens Tü49, S. griseus ATCC13273, S. lividans 1326, S. parvulus JI2283, S. peucetius ATCC 27952, S. rochei CECT 3329, S. steffisburgensis NRRL3193, S. vinaceus JI2838, and 8 Streptomyces sp. strains isolated from different soil samples (Supplementary Table S1). These strains were grown on R2YE, MS, PGA, and NA solid media for transformation, sporulation, conjugation, and phenotypic assays, respectively (Coco et al., 1991; Kieser et al., 2000). YES xylose (Sevillano et al., 2016) or NMMP (Kieser et al., 2000) containing 1% of xylose were used in the overexpression assays. Routine plasmid construction and plasmid isolation was done in Escherichia coli DH5α, and E. coli ET12567, a non-methylating strain, was used to obtain the plasmids to be transformed into S. coelicolor. E. coli strain BW25113 (pIJ790) (containing the λ Red system) (Datsenko and Wanner, 2000) and ET12567 (pUZ8002) (harboring the tra genes in the non-transmissible RP4-derivative plasmid pUZ8002) (MacNeil et al., 1992) were used for PCR-targeted mutagenesis of S. coelicolor M145 and conjugation plasmid transfer to the different Streptomyces species. Staphylococcus aureus MB5393, E. coli ATCC25922 and Candida albicans ATCC64124 were used in the antibiogram analysis. Antibiotics were used when needed for plasmid selection (E. coli: 100 μg mL^-1 for ampicillin; 50 μg mL^-1 for apramycin; 50 μg mL^-1 for kanamycin; 34 μg mL^-1 for chloramphenicol, and 25 μg mL^-1 for nalidixic acid. S. coelicolor: 20 μg mL^-1 for neomycin, 10 μg mL^-1 for apramycin and 10 μg mL^-1 for thiostrepton).

DNA transformation and manipulation of E. coli and S. coelicolor were carried out using the methods by Green and Sambrook (2012) and Kieser et al. (2000), respectively. The plasmids used in this work are listed in Table 1.

The coding regions of SCO4441, SCO4442 or both genes were replaced by an apramycin resistance cassette (aac(3)IV gene) by using REDIRECT PCR-targeting technology (Gust et al., 2003). The primers LS-090 and LS-091 (Supplementary Table S2) were used to generate the mutation cassette from the plasmid pIJ773 (Gust et al., 2003), which was used as the template. The mutated genes were obtained using cosmid SCD6 (Redenbach et al., 1996) and transferred by conjugation from ET12567 (pUZ8002) to S. coelicolor M145. The desired mutants were selected by apramycin resistance and sensitivity to kanamycin. PCR assays confirmed the deletion of the SCO4441, SCO4442 or both genes in S. coelicolor M145.

Qualitative actinorhodin (ACT) production of the S. coelicolor strains was observed on different solid media. Approximately 10⁵ spores were deposited in 5 μL drops onto plates that were incubated at 30°C. ACT production was detected after three to 10 days of growth as a blue halo around the colonies.

Colorimetric quantification of prodiginines (RED) and ACT production was determined by the spectrophotometric method described in Yepes et al. (2011). All experiments were performed in triplicate.

Liquid cultures (10 mL) of S. coelicolor or S. lividans in YES+Xylose, or NMMP+Xylose, containing 20 μg mL^-1 for neomycin or 10 μg mL^-1 for apramycin, depending on the plasmid used, were incubated at 28°C for 8 days at 200 rpm. Then, 1 mL of the culture was extracted using 0.7 volume of 1% formic acid acidified ethyl acetate and the organic layer was dried in vacuo. The dry extracts were finally resuspended in methanol (100 μL). LC-HRMS-analyses were performed as previously described using a Bruker maXis QTOF mass spectrometer coupled to an Agilent 1200 LC (Martin et al., 2014; Perez-Victoria et al., 2016). Differential peaks where selected by direct comparison of the DAD signal base peak chromatograms.

The multicopy overexpression vectors pNX4441, pNX4442, and pNX4441-42 were derivatives of the pN703GEM3 plasmid (Fernández-Ábalos et al., 2003) and carry the neomycin resistance gene as selective marker. They were obtained by cloning the corresponding coding sequences SCO4441 (scr1), SCO4442 (scr2) or both, previously amplified by PCR using the corresponding oligonucleotides (see Supplementary Table S1), into the NdeI and XhoI sites of the intermediate plasmid pXHis1 (Adham et al., 2001b), yielding pX4441His, pX4442His, and pX4441/42His, respectively. In a second step, the corresponding BgIII/HindIII fragments of these three plasmids were cloned into the same sites of pN702GEM3, obtaining pNX4441, pNX4442, and pNX4441-42. In these constructs, the strong xylanase promoter xysAp (Rodríguez et al., 2005) controls the ORFs of scr1, scr2 and scr1 (scr2 was under the control of its own promoter in pNX4441-42), respectively. In these plasmids the corresponding Scr1, Scr2, and Scr2 were tagged, respectively, with a six His tag at the carboxy terminus.

To obtain conjugative plasmids, a BamHI fragment of 1380 bp containing the oriT and the apramycin resistance, from pIJ773 (Gust et al., 2003), was cloned in the BlgII site of the plasmids pNX4441/42 and pN702Gem3, respectively, obtaining the plasmids pNX4441/42c and pN702Gem3c. Later a 3900-bp band was obtained from the plasmid pNX4441/42c and a 1600-bp band was obtained from the plasmid pN702Gem3c using the enzymes HindIII and NheI. These fragments were individually cloned into the low copy plasmid pHJL401 (Larson and Hershberger, 1986) digested with HindIII and XbaI to obtain the final plasmids pHAX4441/42c and its control pHJL401c, respectively. These plasmids carry resistance genes to thiostrepton and apramycin.

The SCO4441 gene (scr1) encodes a 295 aa protein with a 7.18 isoelectric point, has a molecular weight of 33.59 kDa and has been predicted to be a putative antitoxin. This protein contains two regions, the HTH-XRE domain (aas 23 to 80) and the HipB domain (aas 24 to 62) (NCBI database⁴). The gene SCO4442 (scr2) encodes a 63 aa protein with a 6.49 isoelectric point, has a molecular weight of 6.48 kDa and has been predicted to be a putative toxin. This protein contains a DUF397 region (aas 11 to 62).

As Scr1 and Scr2 are proposed to form part of a putative toxin–antitoxin system (TAS) (Shao et al., 2011; Xie et al., 2018), here we carried out a study to analyze the effect of its putative toxicity using the over-expression of Scr1, and/or Scr2 in S. coelicolor M145 and in S. lividans 1326. Three multicopy plasmids, pNX4441, pNX4442, and pNX4441/42 (see section “Materials and Methods”) were generated, where the expression of these genes was controlled by the xylanase strong promoter xysAp (Rodríguez et al., 2005). The plasmids and the control, the empty plasmid pN703Gem3, were introduced into both Streptomyces species to assess the possibility of Scr2 acting as a toxin.

The S. coelicolor cells overexpressing the putative toxin in plasmid pNX4442 were viable, generating colonies which were able to differentiate like those obtained with both the control plasmid and the putative antitoxin (pNX4441). A slight delay in differentiation was observed when both genes of the putative TAS were overexpressed (pNX4441/42) (Figure 1A). The same results were obtained when the plasmids were transformed into S. lividans where no toxic effect was observed when the putative toxin encoded in the plasmid pNX4442 was expressed (data not shown).

These results indicated that the protein encoded by scr2 was not acting as a conventional toxin, inducing lethality when overexpressed in S. coelicolor and S. lividans, at least under the conditions used (Figure 1A), as was the case of YefM/YoeBsl, isolated from S. lividans, which was previously characterized in our laboratory (Sevillano et al., 2012).

To exclude the possibility that a single gene copy of scr1, present in the S. coelicolor genome (wild-type), could be sufficient to counteract the toxicity of Scr2 overexpression, a deletion mutant strain lacking both putative TAS genes, ΔSCO4441/42 (Δscr1/scr2), was generated (see section “Materials and Methods”) and used as a host for plasmids pNX4441, pNX4442, and pNX4441/42. The results obtained when overexpressing these plasmids in the new mutant strain were similar to those described for the wild-type (Figure 1B). Hence, the function of these two genes did not correspond to a TAS under these particular conditions. Interestingly, the mutant Δscr1/scr2 did not show any phenotype divergent from that of the wild-type in the conditions used (see section “Discussion”).

Furthermore, a clear phenotype of colored antibiotic induction was observed when the strains of S. coelicolor wild-type and Δscr1/scr2, carrying the overexpression plasmid pNX4441/42, were grown on several solid media such as NMMP+Xyl. A high overproduction of the blue-red antibiotic ACT was observed when scr1 was overexpressed either alone or with scr2 in the wild-type strain. Nevertheless, when the S. coelicolor Δscr1/scr2 strain was used as a host, this high overproduction of ACT (blue color) was only observed with the overexpression of both genes at the same time (pNX4441/42) (Figure 2). These results indicated that Scr1 was a positive regulator of ACT production and required the presence of Scr2 to function.

To further delve into the functionality of these genes, two new mutants lacking only one of the two genes, S. coelicolor Δscr1 and S. coelicolor Δscr2, were generated (see section “Materials and Methods”). These two new single mutant strains, the double Δscr1/scr2 mutant and the wild-type were transformed with four plasmids, pNX4441, pNX4442, pNX4441/42, and pN702Gem3, and the production in liquid cultures was analyzed. The increase of the production of colored antibiotics was observed in all of the liquid media assayed [R2YE, NMMP+Xyl (data not shown) and YES+Xyl] when pNX4441/42 was used. Since the highest production was obtained in YES+Xyl, this medium was used to carry out the rest of the experiments.

Once again, a high level of ACT was induced by Scr1 (pNX4441) in the wild-type strain. This induction was also observed in the S. coelicolor Δscr1 strain, but not in the two strains lacking the SCO4442 gene (Δscr2 and Δscr1/scr2) (Figure 3A). Overproduction of the colored antibiotics was also observed in strains Δscr2 and Δscr1/scr2, but only when the plasmid pNX4441/42 was used (Figures 3A,B). Antibiotic production in these cultures was quantified, and the yield of ACT was between 450 and 550 μM after expressing Scr1 alone or together with Scr2 in the wild-type M145 strain. The production of prodiginines (RED) was also induced under the same conditions reaching concentrations of 35–45 μM in this strain (Figure 3C). The quantification of these antibiotics in the Δscr2 and Δscr1/scr2 transformant strains reinforced the hypothesis that Scr1 acts as a positive regulator of antibiotic production and requires the presence of the protein encoded by scr2. Therefore, these two genes must act in conjunction (Figures 3B,C).

Additionally, the analysis of the S. coelicolor M145 cultures overexpressing Scr1 and Scr2 by LC-HRMS was performed for identifying new putative compounds, originating from silent pathways that may be promoted by Scr1/Scr2. Several compounds were produced de novo by the action of Scr1/2 and were putatively identified as SEK 4, SEK 4b, ε-actinorhodin and γ-actinorhodin, and one putative new compound with the molecular formula C₃₂H₂₄O₁₄ and an UV absorbance similar to actinorhodin (Figure 4). SEK 4 and SEK 4b have been previously described as shunt polyketide intermediates, non-enzymatically cyclized products from the biosynthesis of actinorhodin (Fu et al., 1994; Jetter et al., 2013).

The ability of Scr1, Scr2, and Scr1/Scr2 from S. coelicolor to induce antibiotics was also tested in S. lividans 1326 in the same way as above. The plasmids pN702Gem3, pNX4441, pNX4442, and pNX4441/42 were introduced in this strain, cultured in liquid YES+Xyl and compared. As in S. coelicolor, production of ACT (59 μM) and RED (430 μM) was detected when this strain was transformed with the plasmids pNX4441 and pNX4441/42, but not observed when transformed with pNX4442. Interestingly the production of RED was almost ten times higher than the production reached for this antibiotic in S. coelicolor (Figures 5A,B). LC-HRMS analysis of the cultures of S. lividans 1326 overexpressing scr1 and scr2 detected the production of SEK 4, SEK 4b, fogacin, a putative new compound with the molecular formula C₃₂H₂₈O₁₄ and a prodigiosin of the molecular formula C₂₅H₃₃N₃O with six isomeric prodigiosins found in the Dictionary of Natural Products (DNP) (Figure 6). Fogacin has been previously described as a cyclic octaketide obtained from a Streptomyces sp. (strain Tü 6319) isolated from a contaminated soil in Romania (Radzom et al., 2006).

A potential strategy for improving or inducing cryptic metabolites of different species of Streptomyces producers is the cloning of a positive regulator of antibiotic production. Based on the results obtained from the overexpression of Scr1/2 in S. coelicolor and S. lividans, these genes appeared to be good candidates for obtaining new natural products using this type of strategy. Therefore, we generated a plasmid that could be used to transfer these regulators into other Streptomyces species. To do so, a conjugative plasmid derived from pHJL401 was generated (pHJL401c) and used to introduce scr1 and scr2 in the final plasmid pHAX4441/42c. These two new plasmids were transferred from E. coli to nine Streptomyces species known to produce different antibiotics: S. albus, S. argillaceus, S. glaucescens, S. griseus, S. parvulus, S. peucetius, S. rochei, S. steffisburgensis and S. vinaceus. The conjugants obtained were grown in liquid YES+Xyl containing 10 μg mL^-1 of apramycin for 8 days and the metabolites produced and extracted with acidified ethyl acetate were analyzed using LC-HRMS to look for the increased production of known or new compounds. This strategy was successful in S. peucetius and S. steffisburgensis. In S. peucetius, two putative new compounds were produced (peaks B and C) and an additional one (peak A) was greatly induced by Scr1/Scr2 (Figure 7). The analysis of the differentially detected peaks did not correspond with those of known compounds. This may have occurred in two of the cases (peaks A and B) due to the lack of ionization in both positive and negative ESI. In the case of peak C, the only component found in the DNP with the predicted molecular formula C₁₈H₁₄N₂O₄ (Supplementary Figure S1) corresponded to Flazine methyl ester, which had an UV spectrum that did not coincide with the spectrum determined experimentally (Buckingham, 2017). In S. steffisburgensis, an induced compound, N-[1-Hydroxy-2-(1H-indol-3-yl)-2-oxoethyl]acetamide, was highly produced and putatively identified (Figure 8). No changes were observed in the peaks obtained in the other studied strains (Supplementary Figures S2–S5).

Moreover, eight additional strains of Streptomyces sp., with no clear antibiotic activity against the microorganisms tested (S. aureus, E. coli, and C. albicans) when grown in seven different types of media, were also used in the experiment. Conjugants were only obtained for two of the strains: Streptomyces sp. CA-240608 and Streptomyces sp. CA-258987. LC-HRMS-analysis of liquid culture extracts of Streptomyces sp. CA-240608 detected the overproduction of one component with the molecular formula of C₂₅H₁₆N₄O₆ (Supplementary Figure S1) that could correspond to any of the isomers Izumiphenazine A, Izumiphenazine B, Phenazinoline D or Phenazinoline E (Figure 9). These were the only compounds described in the DNP with this predicted molecular formula. Interestingly, the antibiogram activity of the clones of this strain overexpressing Scr1/Scr2 exhibited slight antibiotic activity against E. coli. By contrast, the control strain did not show any antibiotic activity (Figure 9). No changes in the production pattern of the strain Streptomyces sp. CA-258987 were detected (Supplementary Figure S5).

Therefore, these results validate the use of Scr1/Scr2 as a biotechnological tool for increasing or promoting different metabolite pathways in some Streptomyces species. The overexpression of Scr1/Scr2 strongly activates antibiotic production in S. coelicolor and S. lividans, and may be useful as a means for searching for new natural products in some Streptomyces species.