The RNA Polymerase Omega Factor RpoZ Is Regulated by PhoP and Has an Important Role in Antibiotic Biosynthesis and Morphological Differentiation in Streptomyces coelicolor^,

Abstract

The RNA polymerase (RNAP) omega factor (ω) forms a complex with the α₂ββ′ core of this enzyme in bacteria. We have characterized the rpoZ gene of Streptomyces coelicolor, which encodes a small protein (90 amino acids) identified as the omega factor. Deletion of the rpoZ gene resulted in strains with a slightly reduced growth rate, although they were still able to sporulate. The biosynthesis of actinorhodin and, particularly, that of undecylprodigiosin were drastically reduced in the ΔrpoZ strain, suggesting that expression of these secondary metabolite biosynthetic genes is dependent upon the presence of RpoZ in the RNAP complex. Complementation of the ΔrpoZ mutant with the wild-type rpoZ allele restored both phenotype and antibiotic production. Interestingly, the rpoZ gene contains a PHO box in its promoter region. DNA binding assays showed that the phosphate response regulator PhoP binds to such a region. Since luciferase reporter studies showed that rpoZ promoter activity was increased in a ΔphoP background, it can be concluded that rpoZ is controlled negatively by PhoP, thus connecting phosphate depletion regulation with antibiotic production and morphological differentiation in Streptomyces.

INTRODUCTION

In bacteria, the RNA polymerase (RNAP) complex plays a central role in transcription and is a target for regulation of primary metabolism (6, 7, 44). The rpoZ gene encodes the RNAP omega (ω) subunit, which forms a complex with the α₂ββ′ core of this enzyme. The ω subunit has been identified in the RNAPs of most free-living bacteria. This protein is functionally homologous to the RpoK subunit of the archaeal RNA polymerase complex and the RPB6 subunit of the eukaryotic RNA polymerases I, II, and III (32). In Escherichia coli the ω subunit interacts with the β′ subunit and promotes assembly of the RNA polymerase complex (14, 32), although it is not essential for survival in this bacterium (13).

Streptomyces spp. are soil-dwelling bacteria that are notorious for their ability to produce thousands of antibiotics, pigments, antitumor agents, immunomodulators, and a variety of other bioactive secondary metabolites (1, 2, 8). Differential expression of secondary metabolism genes occurs following nutrient depletion (34), but the transcriptional control mechanisms that govern the onset of secondary metabolites are still obscure (29). The rpoZ gene of Streptomyces kasugaensis has been shown to be required for antibiotic production and morphological differentiation but is not essential for growth (21). A DNA fragment containing the rpoZ gene was shown to complement an S. kasugaensis pleiotropic mutant deficient in aerial mycelium formation and kasugamycin biosynthesis. Although sigma factors in Streptomyces have received considerable attention in relation to the expression of antibiotic biosynthetic genes (9, 19, 20), the role of the RNAP ω subunit is still obscure.

The expression of many genes involved in antibiotic biosynthesis is negatively controlled by the phosphate concentration in the medium (reviewed in references 28 and 30). Limitation of inorganic phosphate produces drastic changes in primary metabolism and triggers the onset of secondary metabolism (34).

Phosphate control over the expression of several genes involved in primary or secondary metabolism (37) is mediated by the two-component system PhoR-PhoP (45, 46). Binding of the response regulator PhoP to specific sequences (named PHO boxes) in the promoter regions of phosphate-controlled genes allowed us to identify the consensus binding sequence of the PhoP operator in Streptomyces coelicolor (47).

Bioinformatic studies revealed that the rpoZ gene has a highly conserved PHO box in its promoter region. Taking into account the putative role of RpoZ in the transcription of genes involved in differentiation and secondary metabolism biosynthesis and its possible regulation by PhoP, it was of utmost interest to study the role of rpoZ in antibiotic biosynthesis and morphological differentiation in S. coelicolor and its possible regulation by PhoP. In this work, we report the deletion of rpoZ via the REDIRECT technique (16), phenotype restoration when a copy of the rpoZ wild-type allele is introduced, and in vivo and in vitro studies on the effect of PhoP on the expression of the rpoZ gene. The results indicate an important role of the RpoZ protein in sporulation and antibiotic biosynthesis. The work also establishes rpoZ as a new member of the pho regulon in S. coelicolor.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

Streptomyces coelicolor strains M145 (18) and INB201 (ΔphoP) (40) and gmk::Tn5062 and ΔrpoZ mutants (this study) were manipulated according to standard procedures (18). TBO medium (17) was used to obtain spores. S. coelicolor liquid cultures were grown in defined MG-3.2 medium containing starch (Scharlau) (50 g liter⁻¹), glutamate (8.83 g liter⁻¹), and phosphate (3.2 mM) (39). One hundred milliliters of MG-3.2 medium in 500-ml baffled flasks was inoculated with 10⁶ spores ml⁻¹ and incubated at 30°C and 300 rpm for dispersed and reproducible growth. Samples for antibiotic production and growth were taken after 35, 47, 52, 57, 70, 80, and 100 h of growth, and samples for promoter activity determination were taken after 41, 44, 47, 49, and 65 h of growth.

For the solid cultures TBO, MS, ISP4, TSA, R5, and YPD media (18) were inoculated with a suspension of 10⁸ spores ml⁻¹ and grown at 30°C (normal conditions) or 40°C (for temperature resistance assay).

rpoZ disruption by gene replacement.

The rpoZ deletion mutant was obtained by REDIRECT technology (16). Two primers (rpoZ-F [5′-CCGACCGAATCTTTCCCATCCATCGGAAGGTAGAGCGTGATTCCGGGATCCGTCGACC] and rpoZ-R [5′-GCAAGGCTGAAGATATTACCGCCGATGCTGCTGGTACTATGTAGGCTGGAGCTGCTTC]) were designed for the amplification of the apramycin resistance cassette of pIJ774. The PCR product was electroporated in E. coli BW25113/pIJ790 carrying the cosmid St9C5, an S. coelicolor cosmid containing target SCO1478. Since E. coli BW25113/pIJ790 has a high recombination activity, replacement of the rpoZ coding region by the apramycin resistance cassette takes place with high efficiency. The recombinant cosmid was transformed in E. coli ET12567/pUZ8002 and then transferred to Streptomyces by conjugation. Deletion mutants were selected by their Apr^r Kn^s phenotype and confirmed by PCR and Southern hybridization.

Complementation of ΔrpoZ strains.

A 550-bp fragment of S. coelicolor M145 that contains the rpoZ gene with its promoter and terminator sequences was amplified with primers RPOZ-Cu1 (5′-GGGCCTCTAGATAAGTCAGCGCA) and RPOZ-PD (5′-ACGAGCTGGATCCGCAGGCTCA), which had been modified with XbaI and BamHI restriction sites, respectively. The PCR product was cloned into pRA (36), an integrative conjugative plasmid. In order to introduce the complementation plasmid in the apramycin-resistant ΔrpoZ mutant strain, a 1.4-kb fragment containing the Tn5/Neo resistance gene was cloned in the EcoRV site of pRA. The complementation was confirmed by Southern hybridization using as probes the complementing DNA fragment (550 bp) and an internal fragment of the Apr^r gene (0.9 kb).

Disruption of the gmk gene.

SCO1479 (gmk)-disrupted mutants were constructed using the apramycin-resistant Tn5062 insertion 7D08.2.A11 as described by Fernández-Martínez et al. (10). The exconjugants were checked by Southern blot hybridization using the Tn5062 itself (3,442 bp) as a probe.

Growth, phosphate, and antibiotic determinations.

Dry weight determinations were made by drying pellets of 2 ml of cultures after washing twice with MilliQ water. The phosphate concentration of the culture supernatants was measured with the malachite green assay (24). Antibiotic assays were performed as described by Kieser et al. (18). Calcium-dependent antibiotic (CDA) was determined using bioassays with Bacillus mycoides CECT 4128.

Cloning of the rpoZ promoter region for luciferase assays.

The rpoZ promoter region was amplified by PCR using total DNA as the template and the primers RPOZ-PD (5′-ACGAGCTGGATCCGCAGGCTCA) and RPOZ-PU (5′-GAGGACATATGCTACCTTCCGAT). The 246-bp amplified fragment encompasses the intergenic region between SCO1479 and SCO1478 (from position −62 to −4 with respect to the translation start triplet of SCO1478) and the last 176 nucleotides (nt) of the SCO1479 coding region. The BamHI and NdeI cloning sites (underlined) were introduced in the primer sequences of RPOZ-PD and RPOZ-PU, respectively, and cloned into the BamHI-NdeI pLUXAR-neo digested vector (40) to obtain (pLUX-rpoZ), which was confirmed by sequencing.

The luxAB activity driven from the promoter was determined in a Sirius V3.2 luminometer (Berthold Technologies). The samples for the luciferase assay were kept on ice until all were collected, and then 500 μl of each sample was mixed with 250 μl of 0.1% n-decanal by injection and measured in a raw data program with an integration time of 20 s after a 5-s delay time step. Measurement units are given as relative light units (RLU) per second.

EMSA.

For the electrophoretic mobility shift assay (EMSA) analyses, the 246-bp fragment containing the rpoZ promoter was cloned into the vector pGEM-T to obtain pGEM-rpoZ. The fluorescent DNA probe was obtained by PCR using pGEM-rpoZ as the template and 6-carboxyfluorescein (6-FAM)-labeled primers (6FAM-T7 and 6FAM-SP6). The labeled probe (425 bp) was purified from an agarose gel using GFX columns. The conditions for DNA-protein binding were described previously by Sola-Landa et al. (46). For the control competition reactions, the unlabeled 246-bp fragment containing the rpoZ promoter (specific) and an unlabeled 1,190-bp fragment internal to the SCO1661 coding region (nonspecific) were used.

Samples were run in 0.5× Tris-borate-EDTA (TBE) buffer on a 5% polyacrylamide native gel for 2 h at 80 V using a Bio-Rad Mini Protean III instrument. After the run, the gel was scanned in an Ettan DIGE Imager charge-coupled device (CCD) camera (GE Healthcare) and analyzed using the software ImageQuant TL.

Primer extension analysis.

RNA samples were taken at 48 h from S. coelicolor M145 (wild type) and INB201 (ΔphoP) MG-3.2 cultures. The isolation of RNA was performed with the RNeasy Midikit (Qiagen). RNA concentration and quality were checked using a NanoDrop ND-1000 (Thermo Fisher Scientific) and a 2100 Bioanalyzer (Agilent). The transcription start site was determined as described by Santos-Beneit et al. (39) using Superscript III reverse transcriptase (Invitrogen) and the rpoZ-FAM + 99 primer (5′-CGCGTAGATCACGAGGCTGTACTTC), which is complementary to the rpoZ coding region from nucleotide +75 to +99.

Electron microscopy.

Microscopic morphological observations were performed by scanning electron microscopy (JSM-6480 LV; JEOL, Japan), from 6-day-old MS plates (30°C). Glutaraldehyde-fixed samples were dehydrated through a graded ethanol series and critical-point dried in a Bal-Tec (Liechtenstein) CPD 030 critical-point dryer using liquid CO₂. Samples were then coated with a 2-nm gold layer in a Balzers Union (Liechtenstein) SCD 004 sputter coater.

Information theory analysis of binding sites.

To evaluate the potential binding of PhoP, we calculated the individual information content (43) of each 11-nt stretch using the weight matrix of model I of Sola-Landa et al. (47). This weight matrix gathers the information of 37 direct-repeat units (DRu) that form the core of experimentally demonstrated PhoP operators.

RESULTS

The SCO1478 open reading frame (ORF) encodes a small protein homologous to RpoZ.

The SCO1478 gene of S. coelicolor encodes a small protein (90 amino acids) with a predicted mass of 9.7 kDa that is homologous to the RpoZ of S. kasugaensis (98% identity) (21) and to a hypothetical protein identified in the genomes of Streptomyces griseus and other Streptomyces species, but it is not similar to the WhiG-like sigma factor designated RpoZ in Streptomyces aureofaciens (22, 23) (see Discussion). The RpoZ protein contains a PRK02950 motif, which is characteristic of the DNA binding domain of the RNA polymerase ω subunit.

In E. coli the rpoZ gene is linked to the spoT gene (12), encoding a pyrophosphatase that controls the ppGpp levels. As shown in Fig. 1, the S. coelicolor rpoZ gene is located downstream of a guanylate kinase gene (gmk) that is involved in the biosynthesis of guanine nucleotides (35). Since the gmk-rpoZ arrangement is kept in E. coli and in S. coelicolor (Fig. 1), the possible role of gmk in differentiation and secondary metabolite production was also investigated by gene disruption (see below).

Fig. 1.

Physical map of the S. coelicolor DNA region containing the rpoZ gene (SCO1478). Note the linkage to the gmk gene (SCO1479), encoding a putative guanylate kinase. For comparison, the E. coli DNA region containing the rpoZ, gmk, and spoT genes is shown at the bottom.

Deletion of the rpoZ gene alters the phenotype of the mutant.

In order to elucidate the role of rpoZ, the gene was deleted using the REDIRECT technique. For this purpose, cosmid St9C5, containing the pIJ774 cassette replacing rpoZ, was transformed into E. coli ET12567/pUZ8002 (see Materials and Methods). After conjugation into S. coelicolor, four Apr^r Kn^s transformants (named T107, T150, T162, and T165) were selected. The deletion of the rpoZ gene in these four transformants was confirmed by PCR and Southern hybridization. The four ΔrpoZ transformants behaved similarly, and one of them, T165, was selected for further studies. Complementation of the ΔrpoZ mutant (T165) was performed by conjugation with E. coli ET12567/pUZ8002 carrying the construction pRA-rpoZ (see Materials and Methods), in which the rpoZ gene is expressed from its own promoter and carries its own terminator sequence. The complementation was also confirmed by PCR and Southern hybridization.

To determine the phenotype of the ΔrpoZ mutant, the parental M145, ΔrpoZ (T165), and complemented (C1) strains were grown on several media. After 2 days of growth, the onset of actinorhodin (ACT) production was clearly precocious in the ΔrpoZ strain on MS medium (Fig. 2A). On the other hand, S. coelicolor ΔrpoZ showed delayed growth on all media tested, especially TBO and ISP4 (Fig. 2B), at this time. After 4 days of growth, actinorhodin production was already observed in both the wild-type and complemented strains on R5, TSA, and ISP4 but not on YPD, MS, and TBO. On the other hand, the ΔrpoZ strain produced significant amounts of actinorhodin on R5, ISP4, MS, and TBO, although not on TSA (Fig. 2C). Sporulation of the ΔrpoZ mutant was delayed in comparison to those of the parental and complemented strains (Fig. 2B and C). After 6 days, all strains were able to sporulate on TBO, MS, and ISP4, although the spores of the ΔrpoZ mutant remained white (Fig. 2D). This white phenotype of the mutant remained after 2 weeks of growth. Another interesting observation was the sensitivity of S. coelicolor ΔrpoZ to high temperatures; indeed, when the temperature was increased to 40°C, the mutant was impaired in aerial mycelium formation, showing a bald phenotype, while the wild type was still able to sporulate (Fig. 2E). However, despite the lack of spore pigmentation and the lower resistance to heat stress, the mutant spores were viable and showed normal morphology, as confirmed by electron microscopy (Fig. 3).

Fig. 2.

Phenotypic effect of the ΔrpoZ mutation on growth, sporulation, and pigmentation on six different solid media: MS, TBO, ISP4, R5, TSA, and YPD. (A) MS plate after 2 days of growth. Left image, top view. Right image, bottom view. Note the precocious pigmentation of the ΔrpoZ strain in this medium. (B) Top view of TBO and ISP4 plates after 2 days of growth. Note the delay in growth of the ΔrpoZ mutant on both media. (C) Bottom view (except top view for TBO) of R5, MS, TSA, YPD, TBO, and ISP4 plates after 4 days of growth. Note the changes in pigmentation of the ΔrpoZ mutant on MS, TSA, and TBO plates and the delay in sporulation of the mutant on TBO medium. (D) Top view of TBO, MS, and ISP4 plates after 6 days of growth. Note the lack of spore gray pigmentation of the ΔrpoZ mutant in all plates. (E) Top view of an MS plate after 6 days of growth at 40°C. Note the lack of sporulation of the ΔrpoZ mutant.

Fig. 3.

Scanning electron microscopy of spores of the parental (M145) and ΔrpoZ mutant strains after growth on MS for 6 days at 30°C. Magnifications, ×13,000 (A and B) and ×30,000 (C and D).

To test if the putative guanylate kinase gene (gmk) had any role in differentiation or antibiotic production, this gene was disrupted as described in Materials and Methods. Disruption of gmk did not alter the phenotype of the mutant (data not shown), excluding a role of this gene in the control of growth, differentiation, or secondary metabolism.

In summary, on most of the solid media tested, deletion of rpoZ produced a retardation of growth, a lack of the gray pigmentation, and an early onset of actinorhodin biosynthesis.

ACT and RED pigment production is drastically altered in ΔrpoZ liquid cultures.

In order to quantify the effect of the rpoZ deletion on growth and antibiotic production in submerged cultures, the ΔrpoZ mutant, M145, and the complemented rpoZ (C1 and C2) strains were grown in liquid MG-3.2 medium (38, 39, 40, 41, 42). Phosphate is depleted in this medium (the residual level is below 0.1 mM) after 44 h of growth, time in which the pho regulon genes are induced and the onset of secondary metabolism takes place (39, 40).

As shown in Fig. 4, growth of the ΔrpoZ mutant was slower than that of the parental and complemented strains during the first 70 h of culture. In fact, the phosphate in the medium was depleted in the mutant cultures 10 h later than in the control cultures (indicated by vertical lines in Fig. 4A). However, the final biomass was higher in the mutant strain than in the parental and complemented strains (Fig. 4A). When the cultures were observed under the optical microscope, differences were observed between the wild-type and mutant strains in terms of the shape and complexity of the mycelial pellets. While the parental strain developed branches of mycelium around the initial pellets, which grow throughout the time culture, no branch formation was observed in the ΔrpoZ mutant cultures (see Fig. S1 in the supplemental material). This observation points toward a role of RpoZ in the control of mycelium development in S. coelicolor in both liquid and solid cultures (26).

Fig. 4.

Growth and production of actinorhodin (ACT) and undecylprodigiosin (RED) by the parental strain M145, the ΔrpoZ mutant, the complemented mutant strains (C1 and C2), and strain M145 plus the plasmid without the complementation cassette (control) in MG-3.2 medium. (A) Dry weight; (B) ACT production; (C) RED production. Black circles, strain M145; black squares, M145 control; white circles, ΔrpoZ mutant; black triangles, complemented strains. The vertical lines indicate when phosphate is depleted in the medium (<50 μM) in both parental (dashed lines) and ΔrpoZ (dotted lines) cultures. Error bars correspond to the standard errors of the means for three biological replicates in M145 and ΔrpoZ cultures and four replicates for the complemented strains (two cultures of each complemented strain).

In relation to the antibiotic synthesis in liquid cultures, there was a burst of actinorhodin (ACT) production in the ΔrpoZ mutant at the first stages of culture growth, in contrast to the case for the parental strain, in which antibiotic production took place after 47 h (Fig. 4B). The volumetric ACT values for the parental strain increased from 1.7 ± 0.6 μg ml⁻¹ at 47 h to 254 ± 27 μg ml⁻¹ at 70 h, in contrast to the case for the ΔrpoZ mutant, where the volumetric ACT values did not change from 47 h (28.7 ± 0.3 μg ml⁻¹) to 70 h (29.4 ± 1 μg ml⁻¹). This indicates that the synthesis of ACT is precocious in the mutant strain but that it fails to maintain the normal production rate after 47 h. This phenomenon agrees with the clear burst of ACT production observed in complex MS and TBO solid media, where the ΔrpoZ mutant strain started to produce ACT sooner than the parental strain, in spite of the significant delay in growth of the mutant (Fig. 2A and C). In conclusion, the onset of actinorhodin biosynthesis took place earlier in the ΔrpoZ mutant, although maximum ACT values were quite low in comparison to those in the parental M145 strain.

A more drastic effect of the rpoZ deletion on undecylprodigiosin (RED) production was observed. Almost no detectable synthesis of the RED pigment was observed in the ΔrpoZ strain, at least until very late time points, when a low production of this pigment took place (Fig. 4C). Strikingly, while in the parental strain ACT production and RED production took place almost simultaneously, with RED production closely followed by ACT production, in the mutant the synthesis of these antibiotics occurred in very separate stages of the culture (with the ACT pigment produced first and the RED one later).

Almost full restoration of ACT production and partial restoration of RED production were observed in the complemented strain (Fig. 4B and C). This lack of full restoration is due to the effect of integration of the plasmid on ACT and RED production, as was shown when the plasmid without the complementation cassette was introduced in the wild-type strain (Fig. 4B and C). This indicates that the drastic reduction in formation of these secondary metabolites is due to the absence of a functional RpoZ (see Discussion). Not all secondary metabolites are regulated in the same manner. The production of calcium-dependent antibiotic (CDA) in nutrient agar was slightly increased in the ΔrpoZ mutant strain compared to the parental strain, and its production was restored to the parental levels in the complemented strain (Fig. 5).

Fig. 5.

Effect of rpoZ deletion on CDA production. For the bioassay, strain M145, the ΔrpoZ mutant, and C1 (ΔrpoZ mutant complemented with the wild-type rpoZ) were grown on nutrient agar for 48 h. Agar plugs (5 mm in diameter) were then transferred to soft nutrient agar inoculated with Bacillus mycoides. The plate on the left was supplemented with 17 mM calcium nitrate to induce CDA activity. The plate on the right, without calcium nitrate, acts as a control.

The gmk-disrupted mutant was also tested for ACT, RED, and CDA production and did not show any significant difference with respect to the parental strain, excluding a polar effect of rpoZ deletion on the gmk mutant phenotype (data not shown).

Binding of PhoP to the promoter region of rpoZ.

Inorganic phosphate is known to control expression of genes involved in both primary and secondary metabolism. Many of the primary metabolism genes are controlled by direct binding of PhoP to their promoter regions (37, 47), whereas secondary metabolism genes are regulated through signal transduction cascades involving pathway-specific regulators (29, 40).

Bioinformatic studies indicated that the region upstream of rpoZ contains two direct repeat units, designated DRu-1 and DRu-2, i.e., a PHO box. They are adjacent and located at 33 nt from the rpoZ translation start triplet. The individual information content (the Ri value) (43) of each 11-nt DRu obtained using model I of the PhoP binding site (47) indicates that the two DRus have a high score (9.7 and 7.3 bits for DRu-1 and DRu-2, respectively), although these values were lower than in the consensus PHO box (Ri 14.6 bits) since the second nucleotides of both DRu-1 and DRu-2 did not match the consensus T nucleotide. This feature is also observed in the promoter region of one of the major sigma factor-encoding genes (hrdA) in S. coelicolor, which has been previously shown to be bound and repressed by PhoP (37, 47) (Fig. 6A). Therefore, binding of PhoP to the 223-nt DNA fragment carrying the rpoZ promoter region was tested. As shown in Fig. 6B, binding of PhoP to this region was clearly observed, giving rise to one shifted band (DNA-PhoP complex), as expected for a core-core (CC)-type operator (47). This result suggests that expression of rpoZ is modulated by PhoP.

Fig. 6.

(A) Individual information analysis of the PhoP binding sites in the rpoZ and hrdA promoter regions using the sequence walker method (43) and the model I weight matrix of Sola-Landa et al. (47). Boxes contain the individual information content (Ri, bits) of each 11-nt direct repeat unit (DRu). The height of the letters represents the Ri contribution of each position to the total information content. Letters extending downward represent unfavorable protein-DNA contacts. Note that in both cases the second base is unfavorable for each of the DRus. (B) Electrophoretic mobility shift assays of the rpoZ promoter with the glutathione S-transferase (GST)-PhoP^DBD protein and competition with unlabeled probe. Lane C, control (labeled probe without protein). Lanes 1 to 5, labeled probe with protein at 2 μM (lane 1) and 4 μM (lane 2), 4 μM protein and a 500× excess of unlabeled unspecific probe (lane 3), 4 μM protein and a 100× excess of unlabeled specific probe (lane 4), and 4 μM protein and a 500× excess of unlabeled specific probe (lane 5).

PhoP control of rpoZ expression.

To check the in vivo effect of the PhoP binding on rpoZ expression, the promoter of this gene was coupled to the luxAB genes and introduced into both the parental and ΔphoP strains as described in Materials and Methods. The cultures for the luciferase analysis contained a limiting phosphate concentration (MG-3.2) in order to achieve the PhoP response once this nutrient is depleted from the medium. The results showed that when the phosphate became depleted (vertical line in Fig. 7), the rpoZ promoter activity, which was identical in both strains until that time, was higher in the S. coelicolor ΔphoP mutant for as long as the cultures went on. A similar difference occurred in terms of growth; the ΔphoP strain was not able to continue growing after 44 h, while the parental strain continued growing at least till 65 h.

Fig. 7.

Expression of the rpoZ promoter coupled to the luxAB reporter genes in S. coelicolor M145 (black circles) and INB201 (ΔphoP; white squares) strains in MG-3.2 medium. Luciferase specific activity (black lines, left y axis) and growth (gray lines, right y axis) are shown. The time when the phosphate in the medium is depleted (<50 μM) in both M145 and INB201 cultures is represented with a black vertical line. Error bars correspond to the standard errors of the means for four biological replicates (two replicates of two different exconjugants per condition).

This higher expression of rpoZ in the ΔphoP strain points to a negative role of PhoP in the regulation of the rpoZ gene. This observation is also supported by transcriptomic studies using the M145 and ΔphoP mutant strains, in which the rpoZ expression in the ΔphoP mutant increased an average of 3.7-fold with respect to that in the wild type (see Fig. S2 in the supplemental material).

Characterization of the rpoZ promoter.

In order to elucidate the molecular basis of the negative role of PhoP in rpoZ expression, the transcription start point (TSP) of rpoZ was determined by primer extension analysis. The aim of the experiment was to determine the position of the PHO box with regard to the TSP and to check whether the TSP is the same in the wild-type and ΔphoP strains. For this purpose, RNA samples from 48-h MG-3.2 phosphate-limited cultures were used. As shown in Fig. 8, the same TSP was obtained in both the wild-type and ΔphoP strains, and the PHO box was located at positions −54 to −33.

Fig. 8.

Primer extension analysis of the rpoZ promoter using automated fluorescent capillary electrophoresis. Primer extension reactions were done with the rpoZ-FAM + 99 primer using both wild-type and ΔphoP RNA samples. Results of a control reaction in which no RNA was added are shown in the upper fluorogram. The filled traces represent the LIZ-500 standard that was included in each sample to determine the size of the extension product (unfilled trace). Note that in both the wild-type and ΔphoP RNA samples, the same extension product was obtained. The box at the bottom summarizes the promoter elements of the gene. The translation start triplet of rpoZ and the translation end triplet of the upstream gene are in boldface. The transcription start point (+1) is indicated with an arrow above. The putative −10 element and the DRus that form the PhoP operator are boxed. The position with respect to the translation start triplet is indicated under the nucleotide sequence.

As shown in Fig. 8, the TSP of rpoZ coincides with the first nucleotide (G) of its translation start triplet. Although the majority of archaeal transcripts are leaderless, this is not the case with bacterial transcripts (3). In any case, at least 11 leaderless transcripts of different Streptomyces species are described in the review by Strohl (48). Of these 11 transcripts, only one has a GUG as the translation start triplet, so this feature does not seem to be usual in Streptomyces. In all compiled promoters, the −10 box ends with a T (48). Taking into account this feature and also the presence of an A (86%) in the second position of the hexamer and the presence of a purine nucleotide in the fourth and fifth positions (69% and 72%, respectively), we have identified the hypothetical −10 element of the rpoZ promoter (Fig. 8). However, no obvious −35 hexamer was found around the −35 region of the rpoZ promoter, and the PHO box was located upstream of this region. In other pho genes, the PHO box is located in the −35 region. PhoP usually acts as a transcriptional activator when it binds to the −35 region. In our case, PhoP acts a transcriptional repressor. This might occur by a change in the conformation of the DNA that affects the transcription process (see Discussion).

DISCUSSION

All bacterial RNA polymerases consist of five subunits, α₂ (two identical monomers), β, β′, and ω, which perform DNA-dependent transcription (14). In addition, a variety of sigma (σ) factors are used in the recognition of different promoters (9, 19, 20, 27). The role of the omega subunit of the bacterial RNAP complex has remained obscure for decades. In vitro studies with E. coli revealed that the ω subunit is an important component of RNAP complex, being required to restore functionality to denatured RNAP (33). In fact, the ω subunit has been shown to interact with the α₂ββ′ core through binding to the β′ subunit (15). This interaction was also defined when the crystal structure of the Thermus aquaticus RNAP was determined (32). The presence of the ω subunit is not strictly required for growth, although E. coli mutants lacking the ω subunit show a low growth rate and a lower final cell density (50). On the other hand, a truncated version of RpoZ containing the N-terminal region (51 out of 91 amino acids) is sufficient to interact with the α₂ββ′ core, leading to a full growth rate of the complemented strain (15).

The deletion of the rpoZ gene in Mycobacterium smegmatis (31) produced a fragmentation of the β′ subunit, although the fragments remained associated with the RNAP core. This mutant showed a reduced growth rate and abnormal colony morphology and pigmentation (31).

The S. coelicolor 90-amino-acid RpoZ protein is highly conserved in all sequenced Streptomyces genomes. No other rpoZ-like gene was found in the genome of S. coelicolor. Surprisingly, an unrelated gene encoding a 278-amino-acid sigma factor homologous to WhiG, which also influences sporulation, was named rpoZ in Streptomyces aureofaciens (22, 23). However, the homology of S. coelicolor RpoZ to the ω subunits of E. coli (12), M. smegmatis (31), S. kasugaensis (21), and other bacteria clearly indicates that the 90-amino-acid protein encoded by SCO1478 is the authentic ω subunit of this actinomycete.

As shown in this article, deletion of the rpoZ gene in S. coelicolor causes a reduction of the growth rate, as occurs in E. coli and M. smegmatis. Disruption of the guanylate kinase gene (gmk), located upstream of rpoZ, had no effect on growth or differentiation. Therefore, gmk and rpoZ appear to be transcribed independently from each other. The slow-growth phenotype of S. coelicolor ΔrpoZ in several solid and liquid media suggests that the RNAP complex lacking the ω subunit is inefficient in terms of expression of the genes required for a full growth rate. However, it is clear that RpoZ is not essential for growth, since the rpoZ-defective mutants are able to reach the same levels of growth in prolonged cultures as the parental strain.

The sporulation process was also affected in the ΔrpoZ mutant, although viable spores, with normal morphology, are formed. However, the spores of the mutant lack the typical gray pigment associated with S. coelicolor and other Streptomyces spores. In addition to the white phenotype, the ΔrpoZ mutant is less resistant to temperature stress than the parental strain. A bald phenotype (no aerial mycelium formation) is obtained when the mutant strain is grown at 40°C. A role of RpoZ in resistance to heat stress in E. coli was also reported. In this bacterium, overproduction of RpoZ suppressed the temperature-sensitive phenotype of rpoZ mutants (32).

In this work we show that the deletion of rpoZ in S. coelicolor has a very strong effect on antibiotic biosynthesis, particularly on undecylprodigiosin, almost preventing its expression. Kojima and coworkers (21) observed a similar effect on the biosynthesis of the aminoglycoside kasugamycin and proposed that this might be a general effect, also affecting secondary metabolites other than aminoglycosides. Indeed, in S. coelicolor, the deletion of rpoZ affected not only the biosynthesis of the proline-derived undecylprodigiosin but also that of the polyketide actinorhodin, the lipopeptide CDA, and the gray pigment associated with spores, suggesting that RpoZ modulates the expression of different secondary metabolites. These unrelated metabolites were affected differently by the deletion of rpoZ; undecylprodigiosin and the gray pigment were almost abolished in the ΔrpoZ mutant, whereas CDA synthesis increased. On the other hand, production of actinorhodin was impaired after an initial burst in the first stages of the culture. The precocious production of actinorhodin by the ΔrpoZ mutant may be explained as the result of defective RNAP complex formation which induces a low growth rate that triggers a burst of this antibiotic. These differences are likely to be due to the distinct organization and nucleotide sequences of promoters in the actinorhodin, undecylprodigiosin, or CDA clusters that affect their interaction with the RNAP complex and the ω subunit.

The molecular mechanism of interaction of the RNAP complex with promoters of genes controlling secondary metabolism is also modulated by the action of specific DNA binding proteins such as PhoP or AfsR (40). AfsR seems to act by recruiting the RNAP complex to interact with higher affinity with those AfsR-regulated promoters (49). AfsR is, indeed, a positive regulator of afsS (encoding a sigma factor-like small protein) that triggers expression of the pathway-specific regulators ActII-ORF4 and RedD (11).

PhoP recognizes a DNA binding sequence similar to that of AfsR (40) and appears to act by a similar mechanism, recruiting the RNAP complex for interaction with the afsS promoter. In E. coli and Bacillus subtilis, the PhoB protein (homologous to the Streptomyces PhoP) also interacts with the RNAP holoenzyme (5, 25). In summary, the RpoZ protein may affect expression of secondary metabolite genes by interaction with the core RNAP and with wide-domain regulators such as AfsR or PhoP.

As shown in this work, PhoP binds to the rpoZ promoter, in agreement with the presence in its promoter of a consensus PHO box. Two direct repeats, DRu-1 and DRu-2, are adjacent, forming a class I (CC) operator (47). Although both DRus have an unconserved nucleotide in the second position, PhoP was shown to bind the promoter with high affinity. The binding of PhoP to this region formed a single DNA-protein complex, in agreement with the behavior of class I operators. The effect of PhoP on this promoter is negative, since the promoter activity was higher in the ΔphoP mutant than in the parental strain.

According to Browning and Busby (4), there are three general mechanisms of transcriptional repression. In the first mechanism, the repressor binds in or close to the core of the promoter elements and produces a steric hindrance of RNA polymerase binding. In the second, the repressor binds to promoter-distal sites and does not prevent binding of RNA polymerase to the promoter but instead interferes with postrecruitment steps in transcription initiation. In the third mechanism, the repressor functions as an antiactivator.

We have previously described that PhoP acts as a repressor with the first mechanism mentioned above. Thus, PhoP represses pitH2 transcription when it binds overlapping the −10 element (39). PhoP also has been shown to function as an antiactivator (third mechanism). PhoP represses the glnA and afsS genes competing with the binding sites of the transcriptional activators GlnR and AfsR (38, 40, 42). In this study, we also describe PhoP acting as a transcriptional repressor according to the second mechanism. Thus, rpoZ expression is slightly decreased when PhoP binds upstream of the promoter elements. Only the steric hindrance mechanism (where the binding of the RNA polymerase is prevented when PhoP binds to the −10 hexamer) seems to produce a drastic repression effect on the transcription of the genes controlled by this protein.