High level of antibiotic production in a double polyphosphate kinase and phosphate-binding protein mutant of Streptomyces lividans


  • Margarita Díaz,

    1. Instituto de Biología Funcional y Genómica/Departamento de Microbiología y Genética, Consejo Superior de Investigaciones Científicas (CSIC)/Universidad de Salamanca, Salamanca, Spain
    Search for more papers by this author
  • Laura Sevillano,

    1. Instituto de Biología Funcional y Genómica/Departamento de Microbiología y Genética, Consejo Superior de Investigaciones Científicas (CSIC)/Universidad de Salamanca, Salamanca, Spain
    Search for more papers by this author
  • Sergio Rico,

    1. Instituto de Biología Funcional y Genómica/Departamento de Microbiología y Genética, Consejo Superior de Investigaciones Científicas (CSIC)/Universidad de Salamanca, Salamanca, Spain
    Search for more papers by this author
  • Felipe Lombo,

    1. Area de Microbiología, Departamento de Biología Funcional, Facultad de Medicina, Instituto Universitario de Oncología del Principado de Asturias (IUOPA), Universidad de Oviedo, Oviedo, Spain
    Search for more papers by this author
  • Alfredo F. Braña,

    1. Area de Microbiología, Departamento de Biología Funcional, Facultad de Medicina, Instituto Universitario de Oncología del Principado de Asturias (IUOPA), Universidad de Oviedo, Oviedo, Spain
    Search for more papers by this author
  • Jose A. Salas,

    1. Area de Microbiología, Departamento de Biología Funcional, Facultad de Medicina, Instituto Universitario de Oncología del Principado de Asturias (IUOPA), Universidad de Oviedo, Oviedo, Spain
    Search for more papers by this author
  • Carmen Mendez,

    1. Area de Microbiología, Departamento de Biología Funcional, Facultad de Medicina, Instituto Universitario de Oncología del Principado de Asturias (IUOPA), Universidad de Oviedo, Oviedo, Spain
    Search for more papers by this author
  • Ramón I. Santamaría

    Corresponding author
    • Instituto de Biología Funcional y Genómica/Departamento de Microbiología y Genética, Consejo Superior de Investigaciones Científicas (CSIC)/Universidad de Salamanca, Salamanca, Spain
    Search for more papers by this author

Correspondence: Ramón I. Santamaría, Instituto de Biología Funcional y Genómica, Zacarías González sn, E-37007 Salamanca, Spain. Tel.: +34 923 294899; fax: +34 923 224876; e-mail: santa@usal.es


Phosphate metabolism regulates most of the life processes of microorganisms. In the present work we obtained and studied a Streptomyces lividans ppk/pstS double mutant, which lacks polyphosphate kinase (PPK) and the high-affinity phosphate-binding protein (PstS), impairing at the same time the intracellular storage of polyphosphate and the intake of new inorganic phosphate from a phosphate-limited medium, respectively. In some of the aspects analyzed, the ppk/pstS double mutant was more similar to the wt strain than was the single pstS mutant. The double mutant was thus able to grow in phosphate-limited media, whereas the pstS mutant required the addition of 1 mM phosphate under the assay conditions used. The double mutant was able to incorporate more than one fourth of the inorganic phosphate incorporated by the wt strain, whereas phosphate incorporation was almost completely impaired in the pstS mutant. Noteworthy, under phosphate limitation conditions, the double ppk/pstS mutant showed a higher production of the endogenous antibiotic actinorhodin and the heterologous antitumor 8-demethyl-tetracenomycin (up to 10-fold with respect to the wt strain), opening new possibilities for the use of this strain in the heterologous expression of antibiotic pathways.


In nature, microorganisms of the genus Streptomyces, a filamentous bacterium, grow in soil by hydrolyzing different complex carbon sources. The changing conditions of this way of life have forced these microorganisms, and others in similar habitats, to develop adaptive responses to different types of stress and nutritional deficiencies. One of these adaptive responses to the nutritional environment is mediated by the level of polyphosphate [poly(P)] (Rao & Kornberg, 1999; Manganelli, 2007). The poly(P) chain is a linear polymer of orthophosphate residues linked by high-energy bonds which is ubiquitous in all living organisms (Kulaev & Kulakovskaya, 2000). It constitutes a phosphate reservoir that is mobilized under Pi starvation conditions (Rao & Kornberg, 1996; Van Dien & Keasling, 1999). The enzyme polyphosphate kinase (PPK) synthesizes this polymer mainly from ATP and is a homotetrameric protein that is associated with the outer membrane in Escherichia coli (Ahn & Kornberg, 1990). A second PPK (PPK2), described as widely conserved in bacteria, can synthesize poly(P) from GTP or ATP (Zhang et al., 2002). Polyphosphate also functions as a source of the phosphate group for the phosphorylation of sugars, nucleotide diphosphate, and proteins, and its degradation is mainly carried out by phosphatases, although some kinases may use it as an ATP substitute, and even PPK and PPK2 can use it to generate ATP or GTP from the corresponding nucleotide diphosphate (Tzeng & Kornberg, 2000; Ishige et al., 2002).

To date, only one ppk gene has been studied in detail in S. lividans and it exerts a negative role in antibiotic production (Chouayekh & Virolle, 2002). Transcriptional studies of ppk have demonstrated that this gene is mainly expressed under conditions of Pi limitation, although a weak expression is also detectable with phosphate-rich medium. This expression is controlled by the two-component PhoR/PhoP system and by an unknown repressor that uses ATP as a corepressor (Ghorbel et al., 2006ab).

In previous work with S. lividans we described the increased accumulation of the PstS protein in a polyphosphate kinase-null mutant (Δppk; Diaz et al., 2005). The PstS protein is a high-affinity phosphate-binding protein that forms part of the high-affinity phosphate transport system encoded by the pst operon. This operon, which is expressed under the control of PhoR/PhoP, is induced under phosphate limitation and is also induced in the presence of an excess of certain carbon sources such as fructose (Diaz et al., 2005; Sola-Landa et al., 2005), suggesting a dual carbon-phosphate regulation (Esteban et al., 2008). Recently, and in relation to this complex regulation, it has been reported that the sugar phosphates affect Streptomyces development through genes that are under the positive control of the two-component system PhoR/PhoP (Tenconi et al., 2012).

In the present work we studied an S. lividans double mutant -Δppk/ΔpstS- to check the viability of this mutant under phosphate-limited conditions. Differences were detected in comparison with the wild type or the single ΔpstS or Δppk mutants upon incubation on asparagine-minimal solid medium (AMM) or in liquid R2YE under phosphate-limited conditions that suggested a cumulative effect of double mutation that partially suppresses the effects of separate single mutations. The most interesting feature of the double mutant was the overproduction of the pigmented antibiotic actinorhodin, when cultured in liquid R2YE under limited phosphate conditions. Additionally, when the double mutant was used as host to express the heterologous biosynthetic pathway for the antitumor compound 8-demethyl-tetramycin, a strong increase in production was obtained.

Materials and methods

Bacterial strains, plasmids and media

All strains and plasmids used are listed in Table 1. Streptomyces strains were grown at 30 °C on solid mannitol soy flour agar medium (MSA), or R2YE (Kieser et al., 2000) for normal cultures and sporulation. Asparagine-minimal medium (Martin & McDaniel, 1975; Sola-Landa et al., 2003) solidified with 3% agarose and supplemented with different amounts of phosphate (from 0 to 5 mM sodium phosphate, pH 7), was used to study the growth of the different mutants. Cultures in liquid AMM media with different amounts of phosphate were grown but only very limited growth was obtained even for the wt strain in all the concentrations assayed (data not shown). Submerged cultures were therefore normally carried out in YE medium (0.5% yeast extract) supplemented with different amounts of the carbon source studied, normally fructose or glucose plus 2 mM MgCl2. Other liquid media used were R2YE (the same as solid media without agar) supplemented with different amounts of sodium phosphate, pH 7. The Streptomyces culture conditions have been described previously (Fernández-Abalos et al., 2003).

Table 1. Bacterial strains
Streptomyces lividans TK24str-6 SLP2 SLP3Parental strainKieser et al. (2000)
Streptomyces lividans Δppk str-6 SLP2 SLP3 ΔppkPolyphosphate kinase-defective mutantChouayekh & Virolle (2002)
Streptomyces lividans ΔpstS str-6 SLP2 SLP3 ΔpstSMutant defective in the high-affinity phosphate protein PstSPresent study
Streptomyces lividans ΔpstS/Δppk str-6 SLP2 SLP3 ΔpstS/ΔppkMutant defective in the high-affinity phosphate protein PstS and in the polyphosphate kinase, PpkPresent study
Escherichia coli DH5αF, ϕ80dlacZΔM15, Δ(lacZYA-argF)U169, recA1, endA1, hsdR17(rk, mk+), supE44, λ-, thi-1, gyrA, relA1Cloning, plasmid isolationHanahan & Meselson (1983)
E. coli BW25113/pIJ790 E. coli K12 derivative ΔaraBAD, ΔrhaBADGene replacementDatsenko & Wanner (2000)
E. coli ET12567/pUZ8002 dam, dcm, hsdS, cat, tet E. coli/S. lividans conjugationMacNeil et al. (1992)

Escherichia coli was grown in Luria broth (LB) at 37 °C, supplemented with kanamycin (25–50 μg mL−1) when needed.

DNA manipulations and transformations of S. lividans and E. coli

Total DNA isolation (genomic + plasmid), transformation, and protoplast manipulation were done as indicated previously (Diaz et al., 2005). Intergeneric conjugation was used to transfer cosmids from E. coli to S. lividans as described in Gust et al. (2003).

Phosphate uptake

Phosphate incorporation in S. lividans cultures was studied in cells grown in liquid YE + 5% fructose for 60 h (30 °C, 200 rpm). Cells were washed with 0.9% NaCl and 32P-labeled Na2HPO4 was added (2 × 105 cpm mL−1). Phosphate uptake was measured after 2 min at 30 °C with a liquid scintillation counter (Wallac 1409-001). The phosphate uptake results were normalized to dry weight of the corresponding cells used in the assay.

Construction of S. lividans ∆ppk/∆pstS mutant

Deletion of the pstS gene was accomplished using the REDIRECT technology (Gust et al., 2003). The ppk::Ωhygro mutant strain (TK24 derivative; Chouayekh & Virolle, 2002) was used as a host to obtain the double mutant. A pstS-deletion cassette generated previously to delete the pstS gene, in S. lividans 1326 and in Streptomyces coelicolor M145, was used (Diaz et al., 2005). The recombinant cosmid [SCD84 pstS::acc(3)IV-oriT] was introduced into S. lividans TK24 and the S. lividans ppk::Ωhygro mutant to obtain the pstS and the pstS/ppk null mutants, respectively, by intergeneric conjugation (E. coli/Streptomyces). Correct replacement was checked in Southern blot experiments.

Protein analysis

Total cell protein was obtained by breaking the cells in a fast prep (MP-Biomedicals) and boiling the extract in SDS-polyacrylamide loading buffer for 10 min. Protein electrophoresis was accomplished in denaturing polyacrylamide gels (SDS-PAGE), as described elsewhere (Ruiz-Arribas et al., 1995). Coomassie blue staining was done to visualize proteins. Western blot analyses of the proteins separated in SDS-PAGE were done as in Esteban et al. (2008). Anti-PstS antibodies were used as primary antibodies and horseradish peroxidase-conjugated secondary donkey-anti-rabbit antibody was used. The blot was developed with ECL reagents obtained from General Electric, used according to the manufacturers' instructions.

Alkaline phosphatase assay

Alkaline phosphatase activity was measured following the method described by Moura et al. (2001). In summary, 50 μL of sample was added to 50 μL of 25 mM Tris-HCl buffer pH 8 containing 10 mM p-nitrophenyl phosphate (PNPP) and 0.4 mM CaCl2 and incubated at 37 °C for 10 min. The reaction was stopped by adding 1 mL of 0.5 M Na2CO3 and absorbance was measured at 410 nm. The growth rates of all the strains tested were similar (data not shown).

Generation of an integrative version of 8-demethyl-tetracenomycin C cosmid clone cos16F4

Cosmid clone cos16F4 is a pKC505 derivative (Kieser et al., 2000) that contains most of the genes from the elloramycin gene cluster from Streptomyces olivaceus Tü2353 and is responsible for the biosynthesis of 8-demethyl-tetracenomycin C aglycon (Ramos et al., 2008).

To convert this replicative (low copy number) and apramycin-resistant cosmid into an integrative one, a 6.9-kb SpeI DNA fragment from pFL1139 was cloned into the unique XbaI restriction site of cos16F4, which is located at its multiple cloning site. This 6.9-kb DNA fragment contained the conjugative oriT (for conjugation from E. coli), the tetracycline resistance cassette, the site-specific recombination attP site, the int integrase gene from ΦC31, and the ermE erythromycin resistance cassette (for selection in Streptomyces). pFL1139 is a pBluescriptSK derivative that contains the ermE cassette cloned as an EcoRV-StuI 1.7-kb DNA fragment into the unique EcoRI site (blunt-ended) of pFL1138. pFL1138 is a pBluescriptSK derivative that contains a 5.2-kb DraI-BsaI DNA fragment from pIJ787 (kindly provided by Dr. Bertold Gust, Universität Tübingen, Germany) cloned at the pBluescriptSK SmaI site.

HPLC analysis and quantification of 8-demethyl-tetracenomycin C

Liquid cultures (10 mL in R2YE with different phosphate concentrations, see below) were incubated at 28 °C for 5 days and then extracted with 1 vol. ethyl acetate and the organic layer was dried in vacuo. The dry extracts were finally resuspended in methanol. These extracts were analyzed by reversed phase chromatography in an Acquity UPLC device with a BEH C18 column (1.7 mm, 2.1 × 100 mm, Waters) and equipped with a DAD (Waters 2996). The two mobile phase solvents were acetonitrile and 0.1% trifluoroacetic acid (in water). Samples were chromatographed using this elution program: 10% acetonitrile for 1 min, followed by a linear gradient from 10% to 80% acetonitrile over 7 min at a flow rate of 0.5 mL min−1 and a column temperature of 30 °C. Detection and spectral characterization of the peaks were performed by photodiode array detection and empower software (Waters), extracting two-dimensional chromatograms at 280 nm. The peaks corresponding to 8-demethyl-tetracenomycin C eluted at 4.04 min and were quantified by area integration compared with pure 8-demethyl-tetracenomycin C.

Actinorhodin was quantified using the standard spectrophotometric method (Kieser et al., 2000).

Enzymes and reagents

The products used were purchased from Bio-Rad, Boehringer Mannheim, Invitrogen, Merck, Panreac, Promega, Qiagen or Sigma, and were used following the manufacturers' guidelines.

Results and discussion

The double mutant ppk/pstS restores deficiencies of the pstS mutant growth under phosphate-limited conditions

We have previously reported the over-accumulation of the PstS protein in the S. lividans Δppk mutant (Diaz et al., 2005). To check the effect that the join deletion ppk-pstS has on the growth of S. lividans under limited phosphate conditions, a double mutant, ppk/pstS, was generated in S. lividans TK24. The S. lividans ppk mutant (Chouayekh & Virolle, 2002) was used as a host to delete the pstS gene with the Redirect technology (Gust et al., 2003). The apramycin cassette was used to replace the pstS gene, as described in Diaz et al. (Diaz et al., 2005). To obtain isogenic strains, a single pstS mutant was also generated in the S. lividans TK24 strain that was the parental strain of the ppk mutant. DNA-DNA hybridization and PCR analyses were used to corroborate pstS gene replacement in both mutants (data not shown). The absence of PstS protein in the cells (Fig. 1a) and in the culture supernatant (not shown) of these ΔpstS and ∆ppk/pstS mutants was also corroborated by SDS-PAGE and Western blot with anti-PstS antibodies.

Figure 1.

PstS expression and strains growth in AMM with different amount of phosphate. (a) Western blot to detect cell-bound PstS in the indicated strains (3 μg of total protein were loaded per lane) using anti-PstS. (b) Growth of the different strains in minimal medium (AMM) supplemented with the indicated amount of sodium phosphate buffer, pH 7. The growth of these strains in R2YE and MSA is also included as a control.

The effects of the different concentrations of phosphate (from 0 μM to 5 mM) on the different mutants were studied on AMM solid medium. A total of 100 viable spores from each strain were deposited in 5 μL water onto the surface of the medium and incubated at 30 °C for several days and the growth of all the strains monitored. The growth of the single pstS mutant was the most affected. After 3 days of incubation, this mutant was unable to grow on any of the phosphate concentrations used, in contrast to all the other strains which grew well (data not shown). Longer incubations (4–6 days) permitted the growth of the single pstS mutant in media containing 1 mM phosphate and more, whereas all the other strains (including the double ppk/pstS mutant) were able to grow even in the absence of added phosphate (Fig. 1b). All the strains tested grew perfectly well in other complex media, such as R2YE and MSA (Fig. 1b).

These results suggested that another way to obtain and capture phosphate might be activated in the double mutant. At least two possibilities may explain these results: first, an increase the extracellular phosphatase level and/or secondly, an increase in phosphate incorporation. To study this we used the media YE containing 5% fructose and 2 mM MgCl2 that was used in our previous work on pstS gene (Diaz et al., 2005). Phosphatase activity was similar in the wt strain and in the pstS mutant at all the times assayed. However, phosphatase activity increased up to 2.5 times in the ppk and in the ppk/pstS double mutant at 60-h cultures (Fig. 2a). Inorganic phosphate incorporation was studied in the different strains with 32P-labeled phosphate uptake. As described previously for S. lividans 1326 ΔpstS (Diaz et al., 2005) a striking reduction in phosphate uptake was observed for S. lividans TK24 ΔpstS, which was only able to incorporate 5.2% of the amount taken up by the wt strain. The ppk mutant was able to incorporate 82% of the amount incorporated by the wt strain, whereas the pstS/ppk double mutant was able to take up 28% of the phosphate incorporated by the wt strain (5.3-fold the amount from the pstS mutant; Fig. 2b). Both higher phosphatase activity and an increase in phosphate uptake may explain the previous observation that this Δppk/pstS double mutant grew better than the single ΔpstS mutant under limited phosphate conditions.

Figure 2.

Extracellular phosphatase and phosphate transport. (a) Extracellular phosphatase activity (μmol PNP mL−1) of the different strains: wt (♦), ΔpstS (■), Δppk (▲) Δppk/pstS (X). (b) Uptake of 32P-labeled phosphate after 2 min at 30 °C of the indicated strains. The results were normalized to dry weight of the corresponding cells used in the assay. The results presented are the means of three independent experiments.

This opens the possibility that another high affinity phosphate transport system could be activated in the double mutant. Although a putative orthologous pst operon is present in S. lividans 1326 and S. coelicolor (ORFs: SCO6814, SCO6815 and SCO6816), that operon is missing in S. lividans TK24 genome (Lewis et al., 2010). Therefore, we do not have a clear candidate responsible for the increase of phosphate transport in the double mutant under low phosphate concentrations.

Because the ppk mutant of S. lividans displays a higher expression of the PhoP regulator (Ghorbel et al., 2006a) and a higher expression of the complete pst operon (data not shown) higher phosphate transport would be expected in this mutant. However, under our experimental conditions, the incorporation of radioactive phosphate was slightly lower in this strain than in the wt strain, indicating the existence of another level of regulation, perhaps triggered by a saturation of the concentrations of intracellular phosphate that was not processed into polyphosphate in this mutant. Streptomyces lividans has another putative functional polyphosphate kinase encoded by the SSPG_07441.1 ORF, which is identical to SCO0166 from S. coelicolor. That protein, classified as a putative regulator, in both databanks shares 64% identity and 77% similarity with the PPK2A (NCgl0880) and 53% identity and 71% similarity with the PPK2B (NCgl2620) from Corynebacterium glutamicum (Lindner et al., 2007). The protein encoded by SCO0166 also shares high similarity (60% identity and 72% similarity) with Pseudomonas aeruginosa PPK2, whose activity has been demonstrated experimentally (Zhang et al., 2002; Rao et al., 2009). Future studies addressing the activity of the putative Streptomyces PPK2 may clarify the role of this enzyme in phosphate storage and uptake.

The pstS/ppk double mutant expresses higher amounts of endogenous actinorhodin and of heterologous 8-demethyl-tetracenomycin C than the other strains under low phosphate concentrations

During the study of the growth of the different strains on solid AMM with different phosphate concentrations (see above), the production of the blue-colored antibiotic actinorhodin was detected on plates containing 250 and 500 μM phosphate for the ppk/pstS double mutant and on plates with phosphate concentrations of 500 μM and 1 mM for the ppk mutant, whereas higher concentrations impaired antibiotic production in both strains (Fig. 1b).

The effect of the different phosphate concentrations on actinorhodin production by all four strains was also studied and quantified in liquid R2YE medium with three different amounts of added phosphate: medium without phosphate (R2-P), medium supplemented with the normal amount of phosphate (0.37 mM), and medium with a higher amount of phosphate (1.85 mM). Actinorhodin production by the ppk and ppk/pstS mutants was clearly observed when grown in R2-P and in normal R2 after 4 days of culture. Higher production was obtained in the ppk/pstS mutant in both conditions (Fig. 3a). However, the addition of a high concentration of phosphate (R2 + 1.85 mM P) blocked antibiotic production in these strains.

Figure 3.

Antibiotic production by the different strains. (a) Histogram showing the production of actinorhodin in R2YE with different amounts of phosphate (P): without phosphate (■), supplemented with the normal amount of phosphate (0.37 mM) (■), and with a higher amount of phosphate (1.85 mM) (□). (b) Histogram showing the production of 8-demethyl-tetracenomycin C from an integrated plasmid in all the different strains. The cultures were carried out in R2YE without phosphate (■) and in the same medium supplemented with 1.83 mM phosphate (□). The results presented are the means of two independent experiments.

Overproduction of actinorhodin by the ppk mutant under phosphate-limiting conditions was described previously (Chouayekh & Virolle, 2002). These authors reported that the expression of actII-ORF4 increased drastically in the ppk mutant and originated an increase in actinorhodin production. Although this strain has a functional pst operon that permits a phosphate incorporation almost similar to the wt strain, the incapacity to accumulate polyphosphate may originate a phosphate starvation under low phosphate culture concentrations. This starvation is increased in the ppk/pstS double mutant on which a limitation on phosphate transport is observed when compared with the phosphate incorporation on the ppk mutant. This phosphate famine might explain the higher actinorhodin production of the double mutant ∆ppk∆pstS compared with the single one ∆ppk.

Also studied was the effect of phosphate on antibiotic production on the ability of these strains to produce heterologous compounds. The integrative cosmid cos16F4iE, which directs the biosynthesis of the polyketide antitumor 8-demethyl-tetracenomycin C, was introduced into all the strains and the production of this antitumor agent was carried out in R2-P or R2 + 1.85 mM P. Production of the antitumor agent was quantified by HPLC and was observed to be higher under phosphate limitation: threefold higher than in media with the phosphate supplement. The best producer under both conditions was the pstS/ppk double mutant, which attained a production of about 9.7 μg mL−1 under phosphate limitation and 3 μg mL−1 under an excess of phosphate. These yields represent over 10-fold more antibiotic than that obtained with the wt strain and about threefold more than that obtained with the ppk single mutant. These results open the future possibility of using the ∆ppk/∆pstS strain as a host for the industrial production of metabolites of interest.


This work has been supported by grants BFU2006-13668 and EUI2008-03631 to R. I. Santamaría from the Ministerio de Educación y Ciencia. We thank Dr. M. Virolle for the gift of the strains S. lividans TK24 and Δppk. We also thank M. J. Jiménez Rufo and A. Esteban for their excellent technical work. The IBFG acknowledges the institutional support of the Ramón Areces Foundation during 2011–2012. The authors have no conflict of interest to declare.