Introduction

Organic or inorganic phosphate is a very important phosphorus source for Streptomyces species and other bacteria.1 Phosphate forms part of DNA, RNA, phospholipids, phosphoproteins, nucleotides and other high-energy molecules and plays key roles in the respiratory chain2 and in regulation of enzyme activity by protein phosphorylation/dephosphorylation.3

In Streptomyces species, both primary and secondary metabolism are controlled by the two-component system PhoR–PhoP (orthologous of PhoR–PhoB in Escherichia coli). PhoR is a membrane-bound sensor histidine kinase that is able to autophosphorylate at His215.4, 5 The phosphorylated PhoR interacts with the cognate response regulator PhoP (a transcriptional factor member of the OmpR family) and transfers its phosphate group from PhoR to the receiver domain (Asp53) of PhoP (designated PhoP-P) that changes its configuration and binds to specific sequences (PHO boxes) in the promoter regions or, in a few cases, inside the open reading frames (ORFs).6, 7, 8, 9, 10 PhoR also has phosphatase activity and may dephosphorylate PhoP-P.11 The phosphorylated/dephosphorylated PhoP ratio is a key parameter in control of the pho regulon.

The structure and topology of the direct repeat units that conform each PHO box in Streptomyces coelicolor and their relevance in the control of expression of PhoP-regulated genes has been studied in detail.6, 7, 10, 12, 13, 14, 15, 16, 17 In most cases, the phosphorylated PhoP acts as a positive regulator of the genes containing PHO boxes in their promoter regions but, in some cases, PhoP-P binds PHO boxes in the −10 promoter region or inside the ORFs and exerts negative regulation,14 apparently by a ‘road block’ mechanism that prevents movement of the RNA polymerase on the DNA.8

In Streptomyces species, a PhoU-like protein of 229 amino acids (hereafter designated as PhoU) is encoded by a gene located adjacent to the PhoR–PhoP operon that is expressed divergently from the phoRP transcript.6 Surprisingly, the Streptomyces PhoU-like protein has little conservation (28.1% identical residues) with that of E. coli18 and the phoU gene is located in a different cluster. In E. coli, phoU is integrated in the pstSCAB-phoU cluster, whereas in Streptomyces it is linked to the PhoR−PhoP genes. Although the role of PhoU in E. coli is well studied,19, 20 the precise role of the PhoU-like protein in Streptomyces is still obscure.18, 21 In Streptomyces lividans and S. coelicolor, downstream of the phoU-like gene is mtpA; a gene encoding a metalloprotein of the metalothioneine family, related to the spx gene of Bacillus subtilis that appears to have a role in homeostasis of disulfide bonds protecting the cell against oxidative stress.18

The phoU-mtpA genes seem to be transcribed as a single transcript from the PhoP-regulated promoter of phoU but there is also evidence for a possible promoter located between these two genes based on conventional promoter-probe studies.18, 21 To clarify the control of expression of these two genes in this study, we have applied transcriptomic approaches.

In E. coli, PhoU serves as a modulator of the phosphate starvation stress response19 and it appears to play a role in the coupling of the PstSCAB complex to the PhoR sensor kinase.20, 22, 23 PhoU does not exist in B. subtilis and therefore is not an essential protein for bacterial phosphate metabolism; its ‘coupling’ effect may vary in different bacteria.24 Therefore, it is of great interest to clarify the role of the PhoU-like protein in S. coelicolor. In this bacterium the PhoR–PhoP system has been already shown to differ from that of E. coli in relation to ‘partner fidelity’.25 Thus, the main objective of this study was to elucidate the role of the PhoU protein in the control of the pho regulon. We show, using both promoter probe and quantitative reverse transcription-PCR (RT-PCR) analyses, that PhoU acts as a negative regulator of PhoP upregulated genes. The negative effect of PhoU may serve to compensate for the strong positive effect exerted by PhoP-P on expression of phoU and other pho regulon genes, thus establishing a balance in the control of the pho regulon.

Materials and methods

Bacterial strains and plasmids

The bacterial strains, plasmids and oligonucleotides used in this work are listed in Table 1. S. coelicolor strains M14526 and its derivatives S. coelicolor ΔphoP15 and S. coelicolor ΔphoU were manipulated according to standard procedures.26, 27 E. coli ET12567 [pUZ8002] was used for intergeneric conjugative transfer of plasmid DNA into Streptomyces strains.28

Table 1 Bacterial strains, plasmids and primers used in this study
Full size table

The S. coelicolor ΔphoU mutant was obtained for this work. The mutant was constructed by interruption of phoU gene with kanamycin resistance cassette using pHZ1351 system.4 The phoU gene was cloned into pBluescript KS(+)29 from a 3 kb EcoRI–Ecl136II fragment of D46 cosmid30 and a 1.7 kb fragment EcoRI–NruI of pBSphoRP,4 obtaining pBSphoU. The disruption of phoU gene (kanamycin resistance gene in the opposite orientation) was derivated from Eco72I-fragment pBSphoU and 1.3 kb BamHI fragment pTC192-km, thus obtaining pBSphoUintKn. Finally, plasmid pHZphoU, which contains DNA inserts from SpeI–XhoI pBSphoUintKn and Ecl136II pHZ1351 fragments, was transferred into S. coelicolor M145. Candidate mutants were validated by Southern blot methodology (data not show).

The complementation of the mutant strain was performed with integrative conjugative plasmids. The parental plasmid pAV11b-N is a derivative of pAV11b,31 in which 24 bp were deleted by AvrII, Acc65I digestion, end-filling and ligation. The phoU gene, with its promoter, was amplified by PCR. The primers SM27 and SM28 amplified a 976 bp fragment corresponding to the phoU gene from −219 position (regarding ATG translation start triplet). The PstI (SM27) and XbaI (SM28) cloning sites were introduced via the primer sequences. First, the XbaI/PstI fragments were subcloned into pUC19 for validation of sequence and, then BamHI/AseI fragments were cloned into pAV11B-N for conjugation in Streptomyces. The final plasmid was pCOMphoU-b.

Culture conditions

S. coelicolor cultures were grown in defined liquid MG-3.2 medium (containing, as carbon and nitrogen sources, 50 g l−1 starch and 60 mM glutamate and a growth-limiting amount of 3.2 mM phosphate),14 or in complex liquid R5 medium26 with either 370 or 40 μM phosphate. Baffled flasks (0.5 l, 100 ml of medium) were inoculated with spores (106 ml−1) and incubated at 30 °C, 300 r.p.m., for reproducible and dispersed growth.14 Plates of solid media TBO, TSA and R5 were inoculated with a suspension of 108 spores and incubated at 30 °C. TBO medium was used to obtain spore preparations.32 Supplements of kanamycin (50 μg ml−1), hygromycin (100 μg ml−1) or apramycin (50 μg ml−1) were added when needed.

Luciferase assay

The reporter luciferase activity was measured in a Luminoskan luminometer (Labsystems, Helsinki, Finland) as follows. Culture samples were immediately cooled in ice until luminiscence readings were made. For measurement, 250 μl of 0.1% n-decanal was added to 500 μl of sample and the light emission was read after 20 s of integration time.

Antibiotic, growth and phosphate assays

The phosphate concentration of culture supernatants was measured using the malachite green assay.33 Antibiotic assays were performed as described by Kieser et al.26 For growth measurements, culture samples (2 ml) were centrifuged, washed twice with MilliQ (Merk, Kenilworth, NJ, USA) water and the dry weight was determined after desiccation at 80 °C for 3 days.

Transcriptomic data

A differential RNA-sequencing (dRNA-seq) analysis34 was carried out by Vertis Biotechnologie AG (Freising, Germany) using an Illumina HiSeq 2000 platform (San Diego, CA, USA). This approach discriminates primary and processed 5′ ends (identification of transcriptional start sites (TSSs)). Studies with two complementary DNA libraries were performed, one library from untreated total bacterial RNA and the other enriched for primary transcripts by terminator exonuclease treatment that degrades 5′P but not 5′PPP RNA. Total RNA was isolated using NucleoSpin miRNA kit (Macherey-Nagel, Dürem, Germany) ref. 740971.50. The sequenced RNA sample was a pool of RNA samples extracted from S. coelicolor M145 and INB201 (ΔphoP) cultures in MG-3.2 medium. The sampling comprised the late exponential phase (36–40 h, 6 samples), the response to phosphate depletion from the medium (40.5–45.5 h, 9 samples) and the response to 10 mM phosphate addition to the cultures after phosphate depletion; 40.5–45.5 h (13 samples). Reads were quality trimmed and mapped with BBTools (http://jgi.doe.gov/data-and-tools/bbtools/). Coverage values were calculated with BEDtools35 and visualized in the IGB browser.36

S. coelicolor microarrays were obtained from Agilent (Santa Clara, CA, USA) as a custom design. Two time series were done with S. coelicolor M145 and INB201 (ΔphoP) strains. Briefly, both strains were cultured in flasks containing defined MG-3.2 medium and samples were taken from 36 to 41 h of culture (13 samples) and stabilized with RNAprotect Bacteria reagent (Qiagen Venlo, Netherlands). The time span included the point when phosphate becomes depleted from the medium and the pho regulon is activated. Total RNA was purified from the culture samples by acid phenol–chloroform extraction and ethanol precipitation. RNA was labeled with Cy3 using the Label IT reagent (Mirus Bio, Madison, WI, USA). Genomic DNA (strain M145) was labeled with Cy5 by Exo-Klenow polymerase (BioPrime kit, Invitrogen (Carlsbad, CA, USA)). Transcriptional values (Mg) were calculated as the normalized log ratio Cy3/Cy5, as indicated previously.37

Reverse transcription quantitative PCR (RT-qPCR)

RT-qPCR was performed in 96-well format with a Stratagene MX3005P (Agilent) and the Brilliant II SYBR Green qPCR Master Mix (Agilent). Complementary DNA was obtained from 0.5 μg of RNA with AffinityScript reverse transcriptase and random primers (Agilent). PCR cycling consisted of a single incubation at 95 °C for 5 min, followed by 40 cycles of 95 °C for 30 s and 65 °C for 1 min. A final melting curve analysis (95 °C for 1 min, 55 °C for 30 s and 95 °C for 30 s) was always included.

The quantification of gene expression was performed by MxPro QPCR Software (Agilent Technologies) according to the ΔΔCt method.38 The S. coelicolor gene SCO1638 was chosen for the reference (housekeeping) assay, as analyses of our microarray data with NormFinder39 revealed this gene as having one of the most stable transcription profiles in our culture conditions (not shown). This was confirmed with a subsequent geNorm40 analysis of RT-qPCR data.

Electron microscopy

Scanning electron microscopy of sporulated mycelium was performed with a JSM-6480 LV scanning electron microscope (JEOL, Tokyo, Japan). S. coelicolor strains were cultured in TBO agar for 7 days at 30 °C. Samples were fixed with glutaraldehyde and osmium tetroxide and dehydrated with increasing concentrations of ethanol. Subsequently, samples were critical-point dried in a CPD 030 critical-point dryer (Bal-Tec, Liechtenstein, Balzers, Liechtenstein) using liquid CO2. Finally, the samples were coated with a gold layer in a Balzers Union (Liechtenstein) SCD 004 sputter coater.

Results

Phenotypic studies in solid medium and sporulation: the phoU mutant is defective in spore pigment

A phoU mutant of S. coelicolor was isolated by insertion of a kanamycin resistance cassette, as indicated in the Materials and Methods. To determine the phenotype of the phoU mutant, it was grown on R5, TBO and TSA media (Figure 1a). The S. coelicolor phoU mutant showed delayed growth in all media tested. Moreover, sporulation of this mutant was delayed in comparison with that of the parental strain in the three media. Noteworthily, the phoU mutant sporulated weakly and its spores remained white after 2 weeks, although the spores were viable (Figure 1b). Therefore, PhoU is required for the formation of the polyketide pigment of the spores (see Discussion). Complementation of the phoU mutant with the wild-type phoU allele restored the sporulation of the mutant and, partially, the spore pigmentation

Figure 1
figure 1

(a) Phenotypic effect of the mutants on growth, sporulation and pigmentation on different solid media: R5, TSA and TBO (left to right). Plates are shown after 3 days of growth at 30 °C. (b) Scanning electron microscopy of spores of the parental (M145) and phoU mutant strains after growth on TBO for 7 days at 30 °C. A full colour version of this figure is available at the Journal of Antibiotics journal online.

Full size image

Transcriptomic analysis of the phoU-mptA region

The S. coelicolor phoU gene (SCO4228) has its own promoter activated by PhoP when the phosphate concentration in the medium is low.6, 12 These previous results are confirmed and extended in this article by the transcriptomics results of the genes in this region (SCO4228–SCO4225). The microarray profile in MG-3.2 showed that the transcription of phoU is totally dependent on PhoP (Figure 2a). In contrast, the downstream genes SCO4227 (mptA) and SCO4226 (hypothetical protein) share a profile that indicates only a partial transcriptional activation by PhoP. A clear promoter at the 5′ end of the SCO4227 coding sequence was deduced from the results of the dRNA-seq assay (see below) (Figure 2b). This promoter appears to be independent of PhoP, as no conserved PHO boxes are shown upstream and it would account for the transcription of the SCO4227–SCO4226 genes in the strain INB201 (ΔphoP) (Figure 2a). Both these genes would form a transcriptional unit as the downstream gene SCO4225, encoding a membrane integral protein, behaved differently and appears to contain its own promoter (Figure 2a).

Figure 2
figure 2

(a) Microarray transcription profiles of SCO4225–SCO4228 genes from cultures of M145 (black line) and ΔphoP (gray line) strains in MG-3.2 medium. Horizontal axis corresponds to the time (h) after inoculation in the cultures. The vertical axis shows the normalized log transcription values (i.e., log2 abundance). Note the strong dependence of phoU on a functional PhoP regulator. (b) S. coelicolor transcription start sites (TSSs) revealed by differential RNA-sequencing (dRNA-seq) in SCO4228–SCO4225 region. Coverage graphs (i.e., number of reads mapped at each nucleotide position) are shown for the control library (normal RNA-seq, upper graph) and for the library generated from the same RNA treated with Terminator exonuclease (bottom graph) that is enriched in primary 5′ ends. Determined three TSSs are shown as arrows on annotated genes SCO4228, SCO4227 and SCO4225.

Full size image

Analysis of the TSSs was based on RNA-seq approach (see Materials and Methods). The TSS of the phoU PhoP-dependent promoter was previously identified by primer extension.6 The dRNA-seq approach revealed the same TSS with a minor difference of 2 nt with respect to the data of Sola-Landa et al.6 (2005) (genomic coordinates 4633052) (Figure 2b). The transcription start point of mptA was identified at genome position 4632207 (Figure 2b). A less clear TSS is suggested for the SCO4225 gene at coordinates 4631506.

phoU disruption delays growth and phosphate uptake

The growth of S. coelicolor phoU mutant was compared with that of the parental S. coelicolor M145 strain in both phosphate-replete (MG-18.5) and phosphate-limited (MG-3.2) media. Results of several experiments (see also below studies in R5 medium) showed that the phoU mutant grows more slowly than the parental strain M145 in MG-18.5; although it reaches finally a similar biomass (dry weight) to that of the parental strain (Figure 3). Similarly, growth of the phoU mutant in phosphate-limited (MG-3.2) medium was also delayed with respect to the parental strain and the rate of phosphate utilization by this mutant was slightly lower than that of other strains (Figure 3b). The effect of the phoU mutation slightly reducing the uptake of phosphate is consistent with the modulating effect of the PhoU protein on other genes of the pho regulon, particularly with the pstS transport system. In conclusion, PhoU is required for optimal phosphate uptake in S. coelicolor and this role is independent of the amount of extracellular phosphate.

Figure 3
figure 3

Growth and phosphate uptake rate by the parental strain M145 (white circles) and phoU mutant (gray diamonds) strains in MG-18.5 medium (a) and MG-3.2 medium (b). Error bars correspond to the s.e.m. for three biological replicates.

Full size image

Antibiotic production in complex R5-370 medium and defined MG medium by the phoU mutant

In order to quantify the effect of the mutation on antibiotic production in submerged cultures, the parental strain and the mutant phoU complemented strain were grown in phosphate-limited MG-3.2 medium and medium R5-370 (R5 medium with 370 μM phosphate concentration), a complex medium that has been optimized by us for antibiotic production.

Growth of the phoU mutant was again slower than that of the parental during the first 60 h of culture. However, the final biomass was similar in this mutant and in the other strains. Antibiotic actinorhodin (ACT) and undecylprodigiosin (RED) production in liquid cultures was delayed and lower in the phoU mutant than that of the M145 strain in both defined and complex media (Figures 4a–c). The respective phoU complemented strain was grown in MG-3.2 medium and the growth and RED and ACT production were partially restored (Figure 5). Therefore, PhoU modulates ACT and RED biosynthesis through its effect on phosphate regulation. As shown below, PhoU modulates negatively the pho regulon and this includes genes involved in antibiotic biosynthesis.

Figure 4
figure 4

Production of actinorhodin (ACT) and undecylprodigiosin (RED) by the parental strain M145 (white circles) and the phoU mutant (gray diamonds) strain in MG-18.5 medium (a), MG-3.2 medium (b) and R5 medium (370 μM phosphate) (c). Error bars correspond to the s.e.m. for three biological replicates.

Full size image
Figure 5
figure 5

Growth and production of actinorhodin (ACT) and undecylprodigiosin (RED) by the parental strain M145::pAV11b-N (white circles), the phoU::pAV11b-N mutant (gray diamonds) and the complemented phoU::pCOMphoU-b (black squares) strains in MG-3.2 medium. Error bars correspond to the s.e.m. for four biological replicates.

Full size image

Quantification of expression of regulatory genes actII-ORF4 and redD

In order to validate the effect of phoU mutation on ACT and RED production, we analyzed the expression of antibiotic biosynthesis regulatory genes, including SCO5085 (actII-ORF4) and SCO5877 (redD) by RT-qPCR. The expression profile of these genes in the wild-type and phoU mutant strains was studied in samples taken after 35, 44 and 60 h of cultivation in MG-3.2 medium. Results of this study showed that the actII-ORF4 regulatory gene of ACT had a lower expression level in the phoU mutant than in the parental strain at all times studied (Figure 6b). These results explain the modulation effect exerted by phoU on ACT production.

Figure 6
figure 6

Validation of genes involved in antibiotic production by reverse transcription quantitative PCR (RT-qPCR). Quantitative RT-PCR of genes is indicated below each panel. Error bars were calculated by measuring the s.d. among technical replicates of each sample. (a) SCO5877, encoding RedD, the undecylprodigiosin regulator and (b) SCO5085, encoding ActII-ORF4, the actinorhodin activator. Parental M145 (white circles) and phoU mutant (gray triangles) strains.

Full size image

In the mutant phoU the redD gene was activated later; it showed a lower level of expression than wild-type strain at 44 h. However, at later times redD was activated and showed a higher level of expression than the parental strain (Figure 6a).

Activation of the glpQ1 promoter as an ‘in vivo’ monitor of the PhoP-mediated response

Plasmid pLUX-glpQ1, containing the glpQ1 promoter coupled to luxAB genes, was used previously to determine the glpQ1 promoter activity in Pi shift-down experiments, because expression of this gene is totally and specifically dependent upon PhoP and serves to monitor the pho response in vivo.16 The strict dependence of the glpQ1 promoter in PhoP activation has been confirmed recently (M. Ordóñez-Robles, unpublished data). However, the effect of the PhoU in the expression of the pho regulon genes may be distinct for different PhoP-regulated promoters. In order to elucidate the molecular basis of the role of phoU gene on the pho regulon, plasmid pLUX-glpQ1 was introduced into the parental and in the phoU mutant strains. Cultures for the reporter luciferase analysis were made in media containing a limiting phosphate concentration, that is, MG-3.2 and R5-40 (R5 medium supplemented with 40 μM phosphate is growth limiting). The disruption of phoU caused consistently higher glpQ1 promoter activities throughout most of the time course of the cultures in both media. When promoter activities were compared, the glpQ1 promoter activity in the phoU mutant was more than 2.5 or 5 times (depending on the medium) higher than that of the parental strain transformed with the same construction (Figure 7). This higher expression of glpQ1 promoter in the mutant points to an important negative role of the PhoU protein in the control of the pho regulon. To compensate for this negative effect, the cell might reduce phoU transcription by lowering PhoP phosphorylation, establishing a pho regulon balance under Pi limitation.

Figure 7
figure 7

Expression of the glpQ1 promoter coupled to the luxAB reporter genes in S. coelicolor M145 and mutants derived from it. M145::pAV11b-N (circles), phoU::pAV11b-N mutant (triangles) and the complemented phoU::pCOMphoU-b strain (squares) in MG-3.2 (a), R5 40 μM (b) and MG-18.5 (c) medium.

Full size image

Validation studies by the quantitative RT-PCR chain reaction confirms the negative regulation exerted by PhoU on the pho regulon

In order to validate the luciferase reporter results and to extend the study to other PhoP-regulated genes, we analyzed by RT-qPCR the expression of three representative genes of the pho regulon; that is, phoR (SCO4229), glpQ1 (SCO1565) and pstS (SCO4142). The expression profile of these genes in the parental and mutant strains was determined in samples taken after 35, 40, 42, 44 and 60 h of cultivation in MG-3.2 medium. The gene SCO1638 of S. coelicolor (encoding a peptidyl-prolyl cis–trans isomerase) was chosen as the reference (housekeeping) gene as analyses of our microarray data with the geNorm program (not shown) highlighted this gene as having the more stable expression in these experimental conditions. Data on gene expression profile of glpQ1 by RT-qPCR are shown in Figure 8 and agree with the profile obtained by the luxAB reporter method; glpQ1 was activated at the same time in both strains but at later times (44 and 60 h) the expression in the phoU mutant strain of glpQ1 was higher than in the parental strain; that is, PhoU exerts a negative modulation on the pho regulon. Similarly, the pstS gene showed a higher expression in the phoU mutant than in the parental strain at later times (44 and 60 h), thus extending the luciferase results to other genes of the Pho regulon. Finally, phoR expression does not seem to change significantly in the phoU mutant, perhaps because in addition to its regulation by PhoU, the phoRP operon is regulated by other interacting proteins.23 These results indicated that PhoU has a distinct effect on phoP-regulated promoters and this may be because of a different degree of interaction of PhoP with those promoters, depending on their structure, not only to a simple effect of PhoU on the phosphorylation balance of PhoP.

Figure 8
figure 8

Validation of the luciferase results by reverse transcription quantitative PCR (RT-qPCR). Quantitative RT-PCR of genes: (a) SCO1565 (glpQ1), (b) SCO4142 (pstS) and (c) SCO4229 (phoR). Parental M145 (white circles) and phoU mutant (gray diamonds) strains. Error bars were calculated by measuring the s.d. among technical replicates of each sample.

Full size image

Discussion

It is well established that the phosphate starvation response is controlled by the two-component system PhoR–PhoP in Streptomyces species.4, 5 PhoP-P is able to bind the pstS and phoU promoters in phosphate-limiting conditions.6

In this article, as shown in Figure 2, we proved that expression of phoU is entirely dependent on PhoP activation. Similarly, in S. lividans, phoU was totally PhoP-P dependent.20 On the other hand, we have confirmed using transcriptomic studies that expression of SCO4227–SCO4226 is partially PhoP independent.

Based on the transcriptomic and RNA-seq analysis performed in this work we conclude that promoters of the genes phoU and mtpA drive the overlapping transcription of genes SCO4228 to SCO4226 and it is likely that the promoters respond to different transcriptional factors; for example, the phoU promoter is clearly dependent on PhoP, whereas mtpA is not.

Two TSSs were clearly identified upstream of phoU and mtpA. This agrees with the recently published dRNA-seq data.41 The mtpA TSS lies three codons inside the annotated coding sequence, implying that the bioinformatically deduced annotation is wrong. The next possible translational start codon in the open reading frame is a ATG codon at position 4632132. In favor of the alternative translation start is the existence of an upstream, well-located, putative ribosome-binding sequence (AGGAGG). In fact, the S. lividans TK24 homolog region is annotated (accession AIJ14424.1) as coding for a shorter protein containing 31 identical amino acid residues.

In S. coelicolor, our results show that PhoU is a negative modulator of PhoP- mediated activation of the Pho regulon. The PhoU negative control of the PhoP-P-mediated activation prevents uneconomical pho regulon activation. However, contrary to E. coli, disruption of phoU in S. coelicolor has a moderate effect (as shown in RT-qPCR studies), but disruption of phoU does not produce a fully constitutive activation of the S. coelicolor pho regulon. The stronger PhoU effect under Pi-limitation conditions matches with the PhoP-P-dependent expression profile of phoU in S. coelicolor (Figure 2). In summary, phoU expression is activated by PhoP-P in Pi-limited conditions. The synthesized PhoU protein, in turn, represses PhoP-P modulation, thus preventing this system to be out of control.

Interestingly, in the phoU mutant in phosphate-repleted conditions (MG-18.5) transcription from the glpQ1 promoter is upregulated during the growth phase (up to 42 h) as compared with the parental strain (Figure 7). In the phosphate-limited MG-3.2 medium the phoU mutant also has an upregulation of the glpQ1 promoter throughout the entire culture time (Figure 7). As expression of glpQ1 is totally dependent on the phosphorylated form of PhoP, we conclude that the PhoU protein contributes to maintain the response regulator in its inactive unphosphorylated form, probably by favoring the phosphatase/kinase ratio of PhoR on the PhoP, as proposed in E. coli.22

It is well known that expression of antibiotic and pigment biosynthetic genes is regulated by PhoP.1, 10 Indeed, as observed in this study, PhoU modulates expression of actinorhodin and undecylprodigiosin and also spores pigment. It is known that PhoP regulates expression of secondary metabolite biosynthetic genes at transcriptional and post-transcriptional levels.5, 8, 9 Therefore, PhoU affects expression of antibiotic and pigment biosynthetic genes through different regulatory proteins (for example, AfsS-AfsR) by a cascade mechanism.42, 43

Dedication

Dedicated to Prof. H. Umezawa for his friendship.