BldD from Streptomyces coelicolor is a non-essential global regulator that binds its own promoter as a dimer

Authors

  • Marie A Elliot,

    1. Department of Biological Sciences, CW405 Biological Sciences Building, University of Alberta, Edmonton, AB, Canada T6G 2E9
    Search for more papers by this author
    • 1

      Department of Molecular Biology, John Innes Centre, Colney, Norwich NR4 7UH, UK.

  • Troy R Locke,

    1. Department of Biological Sciences, CW405 Biological Sciences Building, University of Alberta, Edmonton, AB, Canada T6G 2E9
    Search for more papers by this author
    • 2

      Institute of Biomolecular Design, University of Alberta, Edmonton, AB, Canada.

  • Claire M Galibois,

    1. Department of Biological Sciences, CW405 Biological Sciences Building, University of Alberta, Edmonton, AB, Canada T6G 2E9
    Search for more papers by this author
  • Brenda K Leskiw

    Corresponding author
    1. Department of Biological Sciences, CW405 Biological Sciences Building, University of Alberta, Edmonton, AB, Canada T6G 2E9
    Search for more papers by this author

*Corresponding author. Tel.: +1 (780) 492-1868; Fax: +1 (780) 492-9234, E-mail address: brenda.leskiw@ualberta.ca

Abstract

We have shown that the bldD gene of Streptomyces coelicolor, while required for antibiotic production and morphological differentiation, is not essential for viability. We have also demonstrated that BldD forms a higher order complex both in solution and when bound to target DNA. Purified BldD exists in three forms in solution, as a tetramer, dimer and monomer, but only in the dimeric form when bound to its own promoter/operator.

1Introduction

The life cycle of the filamentous bacterium Streptomyces coelicolor involves a series of metabolically and structurally distinct states. The life cycle begins with germination of a spore and the growth of a mat of vegetative, substrate mycelium that later undergoes both physiological differentiation to an antibiotic-producing state, and morphological differentiation to erect spore-forming, aerial hyphae. Antibiotic production in the substrate mycelium and the formation of aerial hyphae coincide during colony development and this, together with the existence of mutations that block both of these developmental processes, suggests that they share regulatory elements. Among the so-called bld (for ‘bald’, lacking aerial hyphae) mutants with pleiotropic defects in both antibiotic production and morphological differentiation is the single, known bldD mutant [1]. On minimal medium containing glucose as carbon source, the bldD53 mutant produces none of the four well-characterized antibiotics known to be produced by S. coelicolor, and the colonies have a fragmented surface lacking aerial structures. As is the case for many of the bld mutants, the morphological defect, but not the loss of antibiotic production, is overcome by growth on minimal medium containing mannitol as carbon source.

The bldD gene encodes a 167 amino acid protein with a deduced molecular mass of 18 167 Da. The only known mutant allele has an A→G point mutation changing a tyrosine residue at amino acid 62 to a cysteine [2]. The biochemical role for BldD has been addressed using electrophoretic mobility shift assays in which purified BldD was shown to be a DNA binding protein capable of binding its own promoter/operator, as well as the promoter/operators of several key developmental genes [3], including the sporulation-specific sigma factor, σWhiG[4], and the extracytoplasmic function sigma factor, σBldN[5]. Independent studies showed that a promoter/operator fragment from the S. coelicolor sigH gene, encoding the stress response sigma factor, σH[6], could also be retarded in electrophoretic mobility shift assays using extracts from a bldD+ but not a bldD strain, and BldD was later shown to be the sigH promoter binding factor. The use of DNase I footprinting analysis to delimit the DNA binding site for some of the BldD target genes suggested the imperfect inverted repeat, AGTgA (n)m TCACc, as the consensus sequence for BldD binding [3], and the location of the binding site, together with a comparison of the putative bldD target gene transcript levels in the wild-type and bldD mutant strains, were consistent with BldD functioning as a repressor during vegetative growth. How BldD repression of its various target genes might be relieved during development is currently unknown. To gain a further understanding of how BldD exerts its regulatory effects, we have constructed a chromosomal deletion of the bldD open reading frame (ORF) in S. coelicolor M600, and we have begun to address how BldD binds to its target DNAs.

2Materials and methods

2.1Bacterial strains, plasmids and culture conditions

The strains used in this work are listed in Table 1. The plasmids used in this work are listed in Table 2. For routine purposes, Escherichia coli strains were grown in Luria–Bertani (LB) medium at 37°C and S. coelicolor was grown at 30°C on R2YE agar [7]. Ampicillin (Sigma), apralan (50% apramycin (wt/wt), Provel), and spectinomycin (Sigma) were all used at 100 μg ml−1, and kanamycin (Sigma) and chloramphenicol (Sigma) were used at 50 and 25 μg ml−1, respectively.

Table 1.  Strains used in this study
StrainGenotypeReference and/or source
S. coelicolor  
M600prototroph, SCP1, SCP2John Innes Centre
916hisA1 mthB2 pheA1 strA1 SCP1NF (SCP2*)[1]
1169hisA1 mthB2 pheA1 strA1 bldD53 SCP1NF (SCP2*)[1]
E. coli  
BW25113(Δ(araD-araB)567, ΔlacZ4787(::rrnB-4), lacIp-4000(lacIQ), λ, rpoS3699(Am), rph-1, Δ(rhaD-rhaB)568, jsdR514Bertolt Gust; Plant Bioscience Ltd., Norwich, UK, see [8]
ET12567F, dam-13::Tn9, dcm-6, hsdM, hsdR, recF143, zjj-202::Tn10, galK2, galT22, ara14, lacY1, xyl-5, leuB6, thi-1, tonA31, rspL136, hisG4, tsx78, mtl-1, glnV44[9]
BL21(DE3)Fdcm ompT hsdS(rB mB) gal metλ(DE3)Stratagene
JM109recA1, supE44, endA1, hsdR17, gyrA96, relA1, thiΔ(lac-proAB)[17]
Table 2.  Plasmids used in this study
PlasmidFeaturesPhenotypeReference or source
E. coli/Streptomyces shuttle plasmids
pSETΩderivative of φC31 att site-integrating pSET152 [18]: aadA replaces aac(3)IVspecRgift from J. Nodwell; [12]
pSETΩbldDpSETΩ with a 1.3 kb EcoRI/XbaI fragment from pAU181 inserted into the MCSspecRthis study
pAU181pSET152 containing bldD on a 1.3 kb SphI/XmnI fragmentapraR[2]
E. coli plasmids   
cosmid SC9C5Supercos 1 derivativekanR carbRH. Kieser, John Innes Centre
pIJ773aac(3)IV+oriTampR apraRBertolt Gust; Plant Bioscience Ltd., Norwich, UK, see [8]
pUZ8002RK2 derivative with defective oriT (aph), driver plasmid for conjugationkanRM. Buttner, John Innes Centre
pIJ790λ-RED (gam, bet, exo), cat araC, repA101tscamRBertolt Gust; Plant Bioscience Ltd., Norwich, UK, see [8]
pGEX2T(bldD)high copy number GST fusion protein overexpression vector containing the bldD coding sequenceampR[3]
pQE9(BldD+)high copy number 6×His tag fusion protein overexpression vector containing the bldD coding sequenceampR[10]

2.2Construction of a bldD null mutant

To construct the bldD null mutant, the REDIRECT© technology was employed according to the directions of the supplier (Bertolt Gust; Plant Bioscience Ltd., Norwich, UK, see [8]) using the provided E. coli strains and plasmids. Briefly, a disruption cassette containing oriT and the aac(3)IV gene conferring apramycin resistance was generated by polymerase chain reaction (PCR) using a gel-purified, 1.4 kb EcoRI–HindIII, aac(3)IV+oriT fragment from pIJ773 as template and the oligonucleotide primers CGA17 (5′-CAGCCTGCCGCGTCGACACCTTGTCCGGGGAGCCATATGATTCCGGGGATCCGTCGACC) and CGA16 (5′-ACCCCGGCGGCACGTTTCTGCTGAGGTGGTGGGGGCTCATGTAGGCTGGAGCTGCTTC), both of which have 39 nucleotide extensions (underlined) specific for the sequences extending upstream and downstream of the ATG start and TGA stop codons of the bldD ORF, respectively. The linear, PCR-amplified disruption cassette was then introduced by electroporation into E. coli BW25113 containing the bldD cosmid, SC9C5 (provided by H. Kieser, John Innes Centre), and a second plasmid, pIJ790, with the λ RED functions that allowed high frequency recombination between the linear disruption fragment and the SC9C5 cosmid. The disrupted cosmid was passaged through E. coli ET12567 [9] to generate unmethylated cosmid DNA, and the presence of oriT in the disruption cassette then allowed the use of conjugation [7] to introduce the PCR-targeted SC9C5 cosmid into S. coelicolor M600 for the selection of disruption mutants. Despite the presence of at least 16 kb of flanking DNA on both sides of the disrupted bldD gene, none of the resulting exconjugants had the apramycinR, kanamycinS phenotype expected for double crossover mutants. Therefore, spores isolated after non-selective growth of two of the apramycinR (aprR), kanamycinR (kanR) colonies that would have arisen by single crossover into the chromosome were plated (∼2200 colony-forming units (CFU) from each, ∼100 per plate) on R2YE agar [7] in the absence of selection. After 5 days of incubation, 0.3% and 0.5% of the colonies resulting from non-selective growth of aprR/kanR colony #1 and colony #2, respectively, displayed a phenotype identical to that of the known bldD point mutant strain (S. coelicolor 1169) and were found to be apramycinR, kanamycinS. Chromosomal DNA from one candidate bald colony derived from each of the aprR/kanR colonies was digested with SphI and XmnI and subjected to Southern analysis with either a gel-purified, 32P random primer-labeled, 1.3 kb SphI–XmnI bldD fragment (isolated from pAU181 [2]), or a gel-purified, 32P random primer-labeled, 1.4 kb EcoRI–HindIII aac(3)IV fragment (isolated from pIJ773) as probe.

2.3BldD overexpression and purification

BldD was overexpressed as an amino-terminal His6 fusion protein as previously described [10]. The His6-tagged BldD was initially purified using nickel affinity chromatography [10] and then subjected to a further round of purification by ion exchange chromatography. The nickel affinity-purified His6-tagged BldD protein was applied to a Pharmacia Mono Q HR 5/5 ion exchange column equilibrated with 20 mM Tris–HCl (pH 8.5) and then eluted with a linear gradient of 0–0.5 M NaCl. The BldD-containing fractions were then analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and BldD was judged to be >99% pure (not shown). The BldD-containing fractions were pooled, dialyzed against 20 mM Tris–HCl (pH 8.5), and then applied to a Pharmacia Superose 12 HR 10.30 gel filtration column that had been equilibrated in the same buffer. The elution profile of BldD was then compared to the profile of protein standards (bovine serum albumin (66 kDa), β-lactamase (29.5 kDa) and cytochrome c (12 kDa)) run under the same conditions.

BldD with an amino-terminal glutathione S transferase (GST) tag was overexpressed as previously described [3]. The soluble protein fraction was purified by affinity chromatography using glutathione agarose (Sigma) [3]. The column was washed with 1×phosphate buffered saline (PBS) and the protein was eluted with reduced glutathione (10 mM, Sigma). The purity of the GST-BldD fusion protein was assessed by SDS–PAGE analysis of aliquots from each fraction. In order to concentrate the protein and to replace the column buffer with 20 mM Tris–HCl (pH 8.5) buffer containing 150 mM NaCl, those fractions judged to be >99% pure after Coomassie staining were subjected to centrifugation in an Ultrafree®-4 Centrifugal Filter (Millipore).

2.4Electrophoretic mobility shift assay

Electrophoretic mobility shift assays were performed using the 77 bp MAE16-4 probe fragment spanning nucleotides 31–+40 of the bldD promoter as described previously [10]. 0–40 pmol of His6-BldD and/or GST-BldD and 1–2 ng (20–40 fmol) of end-labeled MAE16-4 probe were added to each 20 μl reaction, and after incubation at 30°C the bound and free DNA products were separated on a 1.5% glycerol-containing, 8% polyacrylamide gel. Gels were dried and exposed to a phosphor screen before analysis using a phosphorimager (Molecular Dynamics model 445 SI).

3Results and discussion

3.1Constructed bldD null mutants have the same phenotype as the bldD point mutant

Several failed attempts to construct a bldD null mutant strain had suggested that BldD might, in addition to its known role in the regulation of antibiotic production and morphological differentiation, have an essential function. To address whether the previous failure to generate a null mutant was due to an essential function being disrupted, or was due to a very low frequency in which a double crossover into the chromosome might be expected to occur with relatively short flanking sequences in the disruption cassette, we decided to use a newly developed, PCR-targeted mutagenesis system (called REDIRECT©) [8] to generate a bldD null mutant. The REDIRECT© technology (adapted from [11]) allows disruption of the gene of interest on the cosmid that contains the gene. The disrupted gene is then subsequently transferred to the chromosome by a homologous recombination step in which the large (usually several kilobases) flanking sequences on the cosmid increase the likelihood of crossover into the chromosome.

Using REDIRECT©, apramycinR, kanamycinS colonies with a phenotype identical to that of the known bldD point mutant strain were isolated after introduction of the disrupted bldD cosmid into S. coelicolor M600. Two of these apparent double crossover exconjugants were confirmed by Southern analysis to have the bldD coding region replaced by the disruption cassette (Fig. 1) and designated S. coelicolorΔbldD1 and S. coelicolorΔbldD2. Hybridization to the expected 1.3 kb SphI–XmnI bldD fragment in the parent M600 strain is seen in Fig. 1B. Deletion in S. coelicolor strains ΔbldD1 and ΔbldD2 of the bldD ORF and replacement with the aac(3)IV-containing disruption cassette resulted in the expected introduction of an XmnI site and the release of two fragments (951 bp and 1262 bp; see Fig. 1A) that would hybridize to the bldD flanking regions on the 1.3 kb bldD probe (Fig. 1B). Hybridization of the same two fragments to the aac(3)IV probe (Fig. 1C) and failure to detect any aac(3)IV-hybridizing fragments in the wild-type M600 strain confirmed the deletion of bldD and replacement with the aac(3)IV-containing disruption cassette. As is the case for the known point mutant strain (Fig. 2A), both null mutants were unable to sporulate and produced red pigment when grown on the rich R2YE medium (Fig. 2B) and sporulation was restored on minimal medium containing mannitol as carbon source (not shown). Since bldD has been shown to be monocistronic [2], deletion of the bldD ORF and replacement with an antibiotic resistance marker was not expected to have polar effects. Despite this, a bldD complementing plasmid was constructed to confirm that the phenotypic changes were due to deletion of bldD. The plasmid, designated pSETΩbldD, was a derivative of pSETΩ (gift from J. Nodwell; [12]) in which a 1.3 kb, EcoRI–XbaI bldD fragment from pAU181 [2] had been inserted into the EcoRI and XbaI sites in the multiple cloning site. While introduction of the parent vector, pSETΩ, had no effect on the phenotype of the bldD null mutant strains, introduction of pSETΩbldD caused the expected restoration of wild-type levels of antibiotic production and sporulation (Fig. 2C).

Figure 1.

Southern analysis confirming bldD deletion and replacement by aac(3)IV in S. coelicolorΔbldD1 and ΔbldD2. A is a schematic diagram of the chromosomal DNA of wild-type and disrupted bldD showing the location of the SphI and XmnI sites and the sizes of the resulting digestion products. FRT=flip recombinase target, aac(3)IV=apramycin resistance gene, oriT=origin of transfer. Thicker lines indicate flanking chromosomal DNA. 1–2 μg of total DNA were digested with SphI and XmnI, separated on a 1% agarose gel, transferred to nylon membrane (Amersham) and hybridized with either (B) a gel-purified, random primer-labeled, 1.3 kb SphI–XmnI bldD fragment (isolated from pAU181 [2]), or (C) a gel-purified, random primer-labeled, 1.4 kb EcoRI–HindIII aac(3)IV fragment (isolated from pIJ773). The sizes of PstI-digested λ molecular mass marker bands are shown on the left.

Figure 2.

Phenotype of the constructed bldD null mutant strains. A shows wild-type S. coelicolor strain 916 and its congenic bldD point mutant 1169 grown for 3 days at 30°C on R2YE agar. B shows S. coelicolor M600 and the bldD null mutant strains, ΔbldD1 and ΔbldD2, grown under the same conditions. C shows restoration in ΔbldD1 and ΔbldD2 of wild-type levels of antibiotic production and sporulation after introduction by conjugation of the φC31 att site-integrating plasmid, pSETΩbldD. Strains containing pSETΩbldD, or pSETΩ as a negative control, were grown on R2YE agar plates containing spectinomycin (100 μg ml−1) for 3 days at 30°C.

3.2BldD exists mainly as a dimer in solution and as a dimer when bound to its own promoter

Since we had previously predicted that BldD binds to an inverted repeat sequence in its own promoter/operator region [10], it seemed reasonable to suggest that it might bind DNA in a dimeric form with one monomer interacting with each half site. To address this possibility, the oligomeric state of BldD was examined both in solution and when bound to a fragment of DNA encompassing the known BldD binding site [3] in its own promoter/operator region.

In order to determine the oligomeric state in solution, BldD was overexpressed as an amino-terminal His6 fusion protein [10] and purified as described in Section 2. BldD judged to be >99% pure was then subjected to size exclusion chromatography and its elution profile was compared to the profile of protein standards run under the same conditions (Fig. 3). BldD eluted in three peaks which, by comparison to the elution profiles of the standards, were believed to correspond to tetrameric, dimeric and monomeric forms of the protein. In two separate gel filtration experiments, most of the BldD protein was in the dimer form, as has been found for many transcription factors (reviewed in [13]), and estimation of the areas under the peaks allowed us to determine that the ratio of tetramer:dimer:monomer was approximately 2:7:1. Very recently, gel filtration chromatography also suggested that BldB, a small transcription factor that appears to regulate its own synthesis [14], dimerizes in solution [15].

Figure 3.

Gel filtration profile of BldD. Bold arrows indicate the fractions in which size standards eluted when run under identical conditions. mAU, milliabsorbance units; BSA, bovine serum albumin.

In order to determine the oligomeric state of BldD when bound to DNA, an electrophoretic mobility shift assay was conducted in which both the His6-BldD (described above), as well as GST-BldD (see Section 2), were employed. Electrophoretic mobility shift assays were performed using the 77 bp MAE16-4 probe [10]. As had been seen previously for the His6-BldD [10], two shifted bands were observed both in the presence of His6-BldD and GST-BldD (Fig. 4). Since the GST-BldD was significantly larger than the His6-BldD, the GST-BldD protein-bound DNA was less mobile in the gel. When equal amounts of the two fusion proteins were mixed together and incubated with the end-labeled probe, the two bands corresponding to the GST-BldD-bound DNA, and the two bands corresponding to the His6-BldD-bound DNA were seen. In addition, a protein-bound fragment of intermediate mobility was also observed. Since no intermediate fragment would be expected if BldD bound as a monomer, and more than one intermediate fragment would be expected if BldD bound as a tetramer (for example, intermediates would include [His6-BldD][GST-BldD]3; [His6-BldD]2[GST-BldD]2; [His6-BldD]3[GST-BldD]), the intermediate fragment most likely represents binding by a [His6-BldD][GST-BldD] heterodimer.

Figure 4.

Oligomeric state of BldD when bound to DNA. Electrophoretic mobility shift assays were conducted using 0–40 pmol of His6-BldD, GST-BldD, or an equal mixture of the two fusion proteins, and 1–2 ng (20–40 fmol) of the 77 bp, end-labeled MAE16-4 [10] DNA probe. The 20 μl reactions were performed as described previously [10] and the protein-bound and free DNA were separated by electrophoresis on a continuously running 8% non-denaturing polyacrylamide gel containing 1.5% glycerol and 1×TBE buffer. Free DNA and DNA complexed with His6-BldD, GST-BldD and the [His6-BldD][GST-BldD] heterodimer are indicated with arrows.

Although our findings suggest that BldD exists primarily as a dimer in solution and that it is also in the dimeric form once bound to DNA, we cannot conclude with certainty that it is the dimer form of BldD that makes initial contact with its target DNA. The presence of a small fraction of the monomer in solution leaves open the possibility that BldD binding in vivo occurs via the monomer pathway [16] in which two monomers could bind sequentially to each of the half sites and then assemble their dimerization interface when bound to the DNA. Although use of the monomer pathway has been shown to increase the specificity of protein:DNA complex formation [16], it seems unlikely that BldD follows the monomer pathway for binding to its target promoters since no binding was observed when the left half site fragment of the bldD promoter/operator was separated from the right half site and used as a probe in gel mobility shift experiments [10].

Acknowledgements

This work was supported by the Alberta Heritage Foundation for Medical Research and the Natural Sciences and Engineering Research Council of Canada. We are grateful to Mark Glover, University of Alberta, for helpful discussions about the electrophoretic mobility shift assays, to Plant Bioscience Ltd. for the gift of the REDIRECT constructs and strains, and to Justin Nodwell for the pSETΩ.

Ancillary