Molecular and cytological analysis of the expression of Streptomyces sporulation regulatory gene whiH


Correspondence: Klas Flärdh, Department of Biology, Lund University, Sölvegatan 35, 22362 Lund, Sweden.Tel.: +46-46-2228584; fax: +46-46-2224113; e-mail:


The whiH gene is required for the orderly sporulation septation that divides aerial hyphae into spores in Streptomyces coelicolor. Here, we use a whiHp–mCherry transcriptional reporter construct to show that whiHp is active specifically in aerial hyphae, fluorescence being dependent on sporulation sigma factor WhiG. The results show that the promoter is active before the septation event that separates the subapical compartment from the tip compartment destined to become a spore chain. We conclude that WhiG-directed RNA polymerase activity, which is required for whiH transcription, must precede this septation event and is not restricted to apical sporogenic compartment of the aerial hyphae. Further, it is demonstrated that WhiH, a predicted member of the GntR family of transcription factors, is able to bind specifically to a sequence in its own promoter, strongly suggesting that it acts as an autoregulatory transcription factor. Finally, we show by site-directed mutagenesis and a genetic complementation test that whiH is translated from a start codon overlapping with the previously identified transcription start point, implying leaderless transcription.


Streptomycetes are spore-forming members of the phylum Actinobacteria. Streptomyces spores are formed from reproductive aerial hyphae that emerge from the surface of the vegetative mycelium. Long sporogenic apical cells of the aerial hyphae are divided into prespore compartments by a specialized type of cell division that gives rise to multiple and synchronously forming sporulation septa (for recent reviews, see Jakimowicz & van Wezel, 2012; McCormick & Flärdh, 2012). Each prespore receives one copy of the genome and then matures to a spore in a process involving rounding up of the cylindrical compartments, thickening of the cell wall, condensation of DNA, and synthesis of a spore pigment (Elliot & Flärdh, 2012).

Central regulators in the developmental pathways that convert aerial hyphae to spores have been identified through analysis of Streptomyces coelicolor mutants that form aerial mycelium, but not the gray spore pigment, and therefore show a white colony appearance. The whiH locus is one of the classical whi loci that act early in the sporulation process and are required for sporulation septation (Chater, 1972; Ryding et al., 1998; Flärdh et al., 1999). The whiH gene is of special interest with respect to developmental control of cell division because many whiH mutants are phenotypically similar to two ftsZ mutants: one with a deletion of the sporulation-specific promoter upstream of ftsZ and the other having a missense mutation in ftsZ that prevents correct assembly of FtsZ into cytokinetic Z rings during sporulation (Flärdh et al., 1999, 2000; Grantcharova et al., 2003). The common phenotype includes reduced formation of spore pigment, the inability to correctly segregate nucleoids, the formation of only a few sporulation septa, and an absence of normal spores. Because the pleiotropic whiH mutant phenotype is largely reproduced in mutants specifically defective in septation, it is possible that WhiH controls genes directly involved in cell division.

The WhiH protein shows homology to the GntR family of transcription factors (Rigali et al., 2002; Hoskisson & Rigali, 2009). GntR regulators are well represented in Streptomyces genomes, and some of them have previously been implicated in developmental regulation, including DevA and DasR (Hoskisson et al., 2006; Rigali et al., 2008). WhiH belongs to the FadR subgroup of the GntR family, which carry a winged helix-turn-helix motif in the N-terminal DNA-binding domain (van Aalten et al., 2000; Rigali et al., 2002). However, no DNA-binding activity or interactions of WhiH with promoter sequences have hitherto been demonstrated, and no target genes are known. Streptomyces coelicolor whiH mutants show a strong increase in whiH transcription compared with their wild-type parents, suggesting that WhiH, like many other GntR family members, may negatively regulate its own transcription (Ryding et al., 1998; Aínsa et al., 1999).

Transcription of whiH depends on the alternative RNA polymerase σ factor encoded by whiG, which is required for commitment of aerial hyphae to sporulation (Chater et al., 1989; Ryding et al., 1998; Aínsa et al., 1999; Kormanec et al., 1999). Transcription studies have indicated a delay between the appearance of whiG transcripts and the onset of whiH expression, suggesting that additional levels of control are involved (Ryding et al., 1998). In addition to the autoregulation mentioned above, whiH transcription is also negatively influenced by whiI, encoding an atypical response regulator, but it is not known whether this is due to any direct effect of WhiI on the whiH promoter (Aínsa et al., 1999). σWhiG RNA polymerase initiates transcription at a site that overlaps with the first putative translation start codon of whiH (Ryding et al., 1998), suggesting that whiH may be expressed from a leaderless transcript (Janssen, 1993).

In this study, we wanted to clarify the expression and function of WhiH by, first, using a reporter fusion of the whiH promoter to the gene for a fluorescent protein to analyze when and where whiH is transcribed in developing aerial hyphae and, second, experimentally testing whether it binds to the whiH promoter region and to identify a binding motif.

Materials and methods

Bacterial strains and growth conditions

Bacterial strains and plasmids used in this work are listed in Table 1. The construction of plasmids is described in the Supporting Information, Data S1. Escherichia coli strain ET12567/pUZ8002 was used to produce nonmethylated plasmid DNA for conjugation with strains of S. coelicolor A3(2), which has a methyl-specific restriction system. Growth conditions and media followed general protocols for E. coli and S. coelicolor (Kieser et al., 2000; Sambrook & Russel, 2001).

Table 1. Bacterial strains and plasmids used in this study
StrainsDescription or genotypeReference
E. coli
Tuner (DE3)F ompT hsdSB (rB mB) gal dcm lacY1 (DE3)Novagen
DH5αsupE44 ΔlacU169 (Φ80lacZΔM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1Hanahan (1983)
ET12567dam dcm hsdS (used to obtain nonmethylated plasmid DNA)MacNeil et al. (1992)
S. coelicolor A3(2)
M145Prototrophic, SCP1 SCP2Kieser et al. (2000)
J2408M145 ΔwhiH::ermEFlärdh et al. (1999)
J2400M145 whiG::hygFlärdh et al. (1999)
J2401M145 whiA::hygFlärdh et al. (1999)
J2402M145 ΔwhiB::hygFlärdh et al. (1999)
J2450M145 whiI::hygAínsa et al. (1999)
K202M145 attBΦC31::pKF41[Φ(ftsZ-egfp)Hyb]Grantcharova et al. (2005)
K308M145 hupS::pKF293[Φ(hupS-egfp)Hyb]Salerno et al. (2009)
K310M145 ∆whiH::ermE hupS::pKF292[hupS-egfp]Salerno et al. (2009)
pET15bExpression vector for His-tagged proteinsNovagen
pIJ2925pUC-derived E. coli vector with a modified polylinker; blaJanssen & Bibb (1993)
pIJ8600Integrative vector for Streptomyces; oriT(RK2) int attPΦC31 aac(3)IVSun et al. (1999)
pKF18PinAI-SalI-fragment with whiHp cloned in pIJ2925K. Flärdh, unpublished data
pKF51Codon 1 to 295 of whiH in pET15bThis work
pKF53Codon 47 to 295 of whiH in pET15bThis work
pKF200pRT801 with whiHp-whiHThis work
pKF201pRT801 with whiHp-whiH, first start codon changed to alanine codonThis work
pKF202pRT801 with whiHp-whiH, start codon 47 changed to alanine codonThis work
pKF203pRT801 with whiHp-whiH, codon 28 changed to a stop codonThis work
pKF204Transcriptional fusion between whiHp and mCherry in pIJ8600This work
pKF205Promoterless mCherry in pIJ8600This work
pKF208Translational fusion between whiHp and mCherry in pIJ8600This work
pKS-mCh-ET3mCherry cloned in the EcoRV site of pBluescript II KS +N. Ausmees
pRT801Integrative vector for Streptomyces; oriT(RK2) int attPΦBT1 aac(3)IVGregory et al. (2003)
pTYB12Expression vector, IMPACT systemNew England BioLabs
pUZ8002tra, neo, nontransferable derivative of RP4Kieser et al. (2000)

Production and purification of His-tagged variants of WhiH

Two hexahistidine-tagged versions of WhiH, one starting at codon 1 and one at codon 47, were heterologously produced and purified. Escherichia coli strain Tuner (DE3) containing pKF51 or pKF53 was grown to an OD600 nm of 0.6–0.7 at 37 °C. Expression was induced with 0.04 mM isopropyl-β-D-thiogalactopyranoside (IPTG) at 20 °C for 16 h. After harvesting by centrifugation, the cells were lysed in phosphate buffer [50 mM sodium phosphate, 300 mM NaCl, 10 mM imidazole, 10% glycerol, EDTA-free protease inhibitor cocktail (Roche), and DNaseI, pH 8.0] by French press. The lysate was centrifuged at 4 °C, 48 400 g, 45 min, passed through a 0.2-μm filter, and mixed with Ni-NTA resin (Qiagen) for 1 h to permit binding. The mixture was loaded into a column and washed with phosphate buffer (50 mM sodium phosphate, 300 mM NaCl, 20 mM imidazole, 10% glycerol, pH 8.0), before WhiH was eluted by increasing the imidazole concentration to 250 mM. The eluted fractions were examined using SDS-PAGE, performed in a Mini-Protean 3 system (Bio-Rad). The samples containing WhiH were pooled, and buffer exchange was carried out using PD-10 columns (GE Healthcare). The WhiH protein samples were stored at −20 °C in storage buffer (20 mM Tris–HCl pH 7.5, 60 mM KCl, 0.5 mM EDTA, 0.5 mM DTT, 50% glycerol).

Electrophoretic mobility shift assays

A 187-bp DNA fragment containing the whiH promoter region was produced by PCR using pKF18 as template and primers KF82 and KF83 (Table S1). 10 pmol of the PCR product was 5′-end-labeled with 20 pmol [γ-32P] ATP using 10 units of T4 polynucleotide kinase (Fermentas) at 37 °C for 1 h. The enzyme was inactivated at 70 °C for 10 min. Unincorporated ATP was removed using ProbeQuant G-50 Micro columns (GE Healthcare). 20 fmol of labeled DNA was incubated with purified His-tagged WhiH at 25 °C for 15 min in 15 μL binding buffer (20 mM Tris–HCl pH 7.9, 100 mM KCl, 10 mM MgCl2, 1 mM EDTA, 1 mM DTT, 12.5% glycerol) containing 1 μg poly(dI-dC) and 4.5 μg bovine serum albumin. The DNA–protein complexes were separated on nondenaturing 4% polyacrylamide–Tris–borate–EDTA gels, and the results were visualized by phospho-imaging.

DNaseI footprinting

For the footprinting reactions, a whiH promoter probe labeled at the 5′-end of either the sense or the antisense strand was generated by PCR using pKF18 as template and primers KF82 and KF83 (one of the primers was radiolabeled at the 5′-end using [γ-32P] ATP and T4 polynucleotide kinase). In 50 μL binding reaction mixtures (20 mM Tris–HCl pH 7.9, 100 mM KCl, 10 mM MgCl2, 1 mM DTT, 12.5% glycerol, 4.5 μg bovine serum albumin), 20 fmol of labeled whiH promoter probe was mixed with purified His-tagged WhiH and incubated at 25 °C for 15 min. 50 μL of 10 mM MgCl2 and 5 mM CaCl2 were added and then the reaction mixtures were incubated at room temperature for 1 min. Digestion was started by the addition of 5 U DNase I (New England Biolab). The reactions were stopped after 1 min of incubation at room temperature using 200 mM NaCl, 30 mM EDTA, 1% SDS, and 9 μg yeast tRNA. From each reaction, DNA was extracted with 200 μL phenol/chloroform/isoamyl alcohol (25 : 24 : 1) and ethanol-precipitated. The precipitated DNA was dissolved in loading buffer (33 mM NaOH, 67% formamide, 1.2 M urea, 0.1% bromophenol blue) and loaded onto a denaturing 6% polyacrylamide–Tris–borate–EDTA gel. Dideoxy sequencing ladders were generated using Sequenase version 2.0 DNA sequencing kit (USB) and the same labeled primers as those used to prepare the probes.

Fluorescence microscopy

Fluorescence microscopy was carried out as described previously (Salerno et al., 2009).

Results and discussion

Monitoring whiH promoter activity in live cells

Transcription of whiH is temporally controlled during colony development and is dependent on the sporulation sigma factor specified by whiG (Ryding et al., 1998; Aínsa et al., 1999). To monitor in more detail when and in which cell type whiHp is active, the promoter was transcriptionally fused to the gene encoding the red fluorescent protein mCherry. In aerial hyphae of the wild-type strain M145 carrying this fusion on plasmid pKF204, fluorescence signal was observed both in early-sporulating hyphae, where spore chains had not yet developed, and in mature spore chains (Fig. 1), while vegetative hyphae showed no fluorescence signal above the very low hyphal autofluorescence (data not shown, similar to the whiG panel in Fig. 1). Thus, the whiH promoter activity is confined to aerial hyphae. In the whiG mutant J2400, introduction of pKF204 gave no clear mCherry signal in aerial hyphae, confirming that the reporter construct reflects the expected whiG dependence of whiHp (Fig. 1). In mutants lacking whiA, whiB, or whiI, pKF204 gave signals in aerial hyphae, which is the expected behavior of a whiHp reporter construct because whiH transcription does not depend on any of these genes (Aínsa et al., 1999; Jakimowicz et al., 2006). It has previously been reported that whiH transcripts are overproduced in whiI mutants, suggesting negative regulation by WhiI (Ryding et al., 1998; Aínsa et al., 1999), but no striking upregulation of whiHp-mCherry could here be observed in the whiI mutant. Some putative targets of WhiI have been identified by microarray analysis (Tian et al., 2007; Zhang et al., 2012), but they did not include whiH, and it remains unclear whether WhiI binds directly to whiHp.

Figure 1.

Expression of a whiHp-mCherry reporter in aerial hyphae. A transcriptional fusion of whiHp-mCherry in plasmid pKF204 was integrated at the attBϕC31 site of S. coelicolor strain M145 and its mutant derivatives J2400 (whiG::hyg), J2402 (ΔwhiB::hyg), J2408 (ΔwhiH::ermE), and J2450 (whiI::hyg). Strain J2401 (whiA::hyg) gave results similar to those with J2402 and is not shown. Strains were grown on MS agar and aerial hyphae sampled and investigated by fluorescence microscopy to detect mCherry (shown in inverted gray scale in the right-hand panel of each pair). Representative images are shown. Exposure and adjustment of gray scale were identical for all images. Phase-contrast images of the cells are shown to the left. Bar, 8 μm.

To gain more precise information on the developmental expression of whiH, we studied when and where the whiH promoter is active in relation to two molecular markers of sporulation: the formation of multiple FtsZ rings that precedes sporulation septation and the appearance of the spore-specific nucleoid-associated protein HupS (Schwedock et al., 1997; Grantcharova et al., 2005; Salerno et al., 2009). In a strain having FtsZ fused to EGFP, all observed aerial hyphae with clear FtsZ-EGFP signal showed also signal from whiHp-mCherry, while many aerial hyphae with whiHp-mCherry activity did not contain FtsZ-EGFP structures or visible septa (Fig. 2a). These results show that the whiH promoter is upregulated at an early stage of sporulation in aerial hyphae, before the assembly of multiple FtsZ rings. This conclusion was further confirmed by studies of the HupS-EGFP-producing strain K308, showing whiHp-mCherry signal in preseptational aerial hyphae, as well as in both the apical and subapical compartments of the aerial hyphae that were septating into prespores and showed the spore-specific HupS signal associated with nucleoids (arrow in Fig. 2b).

Figure 2.

Cellular localization of whiHp-mCherry promoter probe activity in relation to molecular markers of sporogenic cells. (a) FtsZ-EGFP indicates the assembly of FtsZ rings in sporulating aerial hyphae. The whiHp-mCherry plasmid pKF204 (top panels) and the control plasmid pKF205 with promoterless mCherry (bottom panels) were introduced into the ftsZ-egfp-expressing strain K202. (b) HupS-EGFP highlights the condensing nucleoids in the prespores produced from a sporogenic cell. pKF204 was introduced into strain K308 expressing a hupS-EGFP fusion. Strains were grown on MS agar and aerial hyphae sampled and investigated by fluorescence microscopy to detect mCherry (shown in inverted gray scale in middle panels) and FtsZ-EGFP or HupS-EGFP (shown in inverted gray scale in right-hand panels). An arrowhead indicates a preseptational aerial hypha that shows expression of whiHp-mCherry, and an arrow indicates a septated apical sporogenic compartment that has been converted to prespores. Representative images are shown. Bar, 8 μm.

WhiH is hypothesized to be an autorepressor (Ryding et al., 1998), but the mCherry reporter gene failed to reveal any effects of the whiH autoregulation (Fig. 1). We could not detect signs of a premature onset or general increase in whiHp activity or any evidence that whiHp is hyperactive in the partially spore-like aerial hyphal fragments that are produced by whiH mutants. Such spore-like compartments were unequivocally identified in strain K310 because they show strong HupS-EGFP signal, but the whiHp-mCherry signal was not higher in these cells than in other parts of aerial hyphae (Fig. S1). Similar results were obtained with a whiH deletion strain (Fig. 1) and a strain carrying the whiH119 allele (data not shown), which has a missense mutation affecting the DNA-binding domain and was used in the previous investigation of whiH transcription (Ryding et al., 1998). Also a translational whiHp-mCherry fusion (pKF208) showed similar patterns and levels of expression in both wild-type and the whiH deletion mutant (data not shown). Possibly, monitoring this type of fluorescence-based reporter construct at the single-cell level may not be very sensitive to some types of developmental modulations in gene expression that are readily detected by monitoring specific transcript levels in bulk RNA preparations.

On the other hand, the reporter constructs provide useful information. Because the whiHp-mCherry constructs depend on whiG, the results show that σWhiG is active at an early stage of aerial hyphal development, and its activity is not restricted to the sporogenic apical cell. Thus, expression of σWhiG-controlled genes like whiH does not depend on specific activation of σWhiG within the sporogenic cell, unlike the situation with σF-dependent genes in Bacillus subtilis, which depend on cell-type-specific activation of σF in the prespore (Barak & Wilkinson, 2005). Instead, whiH is transcribed before delimitation of the sporogenic compartment and may contribute both to the formation of this compartment and to its subsequent conversion into spores.

Mutational analysis of putative translation start sites of whiH

To further clarify the expression of whiH, we determined which is the likely translational start point. The annotated start codon of whiH is located right at the experimentally determined mRNA 5′-end, implying that the transcript is leaderless. However, there is also an alternative start codon at valine 47, preceded by a potential ribosome-binding sequence, which would generate a product of similar size as many other related GntR-like proteins (Fig. 3a). To establish experimentally whether the start codon allocation was correct, point mutations were introduced into a plasmid-borne whiH construct, and the effects on functionality of whiH were assayed by a complementation test. The conversion of codon 47 from putative start codon GTG to alanine codon GCG in pKF202 did not influence the ability to complement the whiH disruption in strain J2408. Both pKF202 and the plasmid carrying native whiH (pKF200) restored sporulation and gray colony pigmentation to J2408 (Fig. 3b). In contrast, changing the first putative start codon (codon 1) to an alanine codon (pKF201) or introducing a stop codon at codon 28 (pKF203) abolished the complementation activity (Fig. 3b). The results show that the putative start codon at codon 47 is dispensable and that codon 1 is required for whiH function. This conclusion was supported by comparative genomics: blastp comparisons to Streptomyces genome sequences available on StrepDB ( indicated that the N-terminal region was conserved in all cases (Streptomyces avermitilis, Streptomyces clavuligerus, Streptomyces griseus, Streptomyces scabies and Streptomyces venezuelae), and the possible internal translation start was absent from S. avermitilis and S. griseus.

Figure 3.

(a) Nucleotide sequence of whiHp promoter region and first part of the coding sequence. The nucleotide sequence is numbered in relation to the two mapped transcription start points (Ryding et al., 1998). Those positions are numbered 0 and +1 and marked by asterisks. Possible translation initiation codons 1 and 47 are underlined. The expression plasmid pKF51 includes codon 1-295, while pKF53 contains codon 47-295. Two additional possible translation starts, codon 8 and 9, are present, but not preceded by putative ribosome-binding sites. The changes in codon 1 and 47 introduced by site-directed mutagenesis in pKF201 and pKF202, respectively, are indicated. Codon 28 is underlined, and its change to a stop codon in pKF203 is shown. The −10 and −35 motifs of the whiH promoter recognized by σWhiG are indicated. Two inverted repeats that were highlighted in a previous article are indicated by arrows (Ryding et al., 1998). The WhiH-binding site, identified in this work by DNaseI footprinting, is shown with black background for the main binding site and gray background for the extended footprint seen at higher WhiH concentration. (b) Mutational analyses of putative translation start sites. The strains were grown on MS agar at 30 °C. The wild-type strain M145 produces gray aerial mycelium. Strain J2408 is a whiH deletion mutant and has a white appearance due to lack of spores and spore pigment. Plasmid pRT801 is the empty vector, and pKF200 contains the whiH gene and complements the J2408 mutant phenotype. Specific point mutations were introduced into pKF200 to yield plasmids KF201 (putative translational start at codon 1 inactivated), pKF202 (putative translational start at codon 47 inactivated), and pKF203 (stop codon introduced at codon 28).

WhiH binds to the whiH promoter region

To test the DNA-binding activity of WhiH, the protein was produced in E. coli with a hexahistidine tag and partially purified by Ni-NTA affinity chromatography. Because the role of the N-terminal extension on WhiH was unclear, two versions were made; 34 kDa (codon 1-295) and 29 kDa (codon 47-295). We used the purified proteins in electrophoretic mobility shift assays to test whether WhiH binds to a DNA probe containing the whiH promoter region. Binding of the short version of His6-WhiH (codon 47-295) to the whiH promoter was clearly detected (Fig. 4). Binding was reduced when increasing molar excess of the unlabeled probe was added, confirming that the interaction was specific (Fig. 4). Similar molar excess of nonspecific DNA did not influence binding (data not shown). Surprisingly, the long version of WhiH (codon 1-295) did not bind to the whiH promoter in similar assays (data not shown). We find in a separate study that will be published elsewhere that the full-length WhiH from S. venezuelae is able to bind to the same DNA probes as an N-terminally truncated S. venezuelae WhiH (starting at codon 47), although the full-length protein binds with somewhat lower affinity (J. Persson, M.J. Bibb, E. Barane, M.J. Buttner & K. Flärdh, unpublished). It is not clear why the full-length S. coelicolor WhiH, when produced heterologously in E. coli, is not functional in DNA binding, but it could be due to artificial effects like poor folding. Further investigation will be required to clarify exactly how the N-terminal extension affects the function of WhiH.

Figure 4.

Electrophoretic mobility shift assay to analyze the interaction of WhiH with the whiH promoter region. A 32P-labeled probe containing the whiH promoter region (position −173 to +14) was used in all reactions (20 fmol). Increasing amounts of the truncated version of His6-WhiH (codon 47-295) were used in lanes 2–6; a fixed concentration was used in lanes 8–11, while no WhiH protein was added to lanes 1 and 7. Competition assays were carried out using unlabeled probe containing the whiH promoter at 0.016, 0.16, 1.6, 16 pmol in lane 8–11, respectively, giving a 0.8-to 800-fold molar excess of unlabeled to labeled probe.

Identification of the WhiH-binding site in the whiH promoter region

To localize the binding site for the truncated WhiH protein in the whiH promoter, DNase I footprinting assays were performed with a DNA fragment containing position −173 to +14 with respect to the whiH transcription start. This identified a region of DNase I protection in the whiH promoter; from −62 to −34 on the upper strand and from −59 to −36 on the lower strand (Figs 5 and 3a). Increased concentration of WhiH in the assay led to an extended footprint to about position −110, suggesting the presence of an additional weaker binding to DNA further upstream in the promoter region (Fig. 5).

Figure 5.

DNase I footprinting analysis of WhiH binding to the whiH promoter region. Different amounts of His6-WhiH (codon 47-295) were added to 20 fmol of the whiH probe labeled on the coding (upper) or noncoding (lower) strand and digested with DNase I for a definite time. The panels show the following (left to right). Lanes 1–4 contain dideoxy sequencing reactions using the same labeled primers as the ones used to generate the corresponding probe. Lane 5 shows the probe without DNase I treatment (the faster migrating band is formed by some remaining undenatured double-stranded DNA and is not a contaminating DNA molecule). Lanes 6–11 contain DNase-treated probe that was incubated without (lane 6) or with increasing concentrations of WhiH (1.6–20 μM). The protected regions are shown by bars and nucleotide numbers to the right of the gel images. The sequences of both the coding and noncoding strands of the whiH promoter region are shown below the panels, numbered in relation to the transcription start point. The areas that are protected by WhiH against DNase I digestion are shown on the respective strand using black background (for the main site protected at lower concentration) and gray background (for the extended footprint seen at higher concentration of WhiH). The positions of the known promoter motifs of whiHp are shown in bold and underlined. A DNase I hypersensitive site between the two footprints on the upper strand is indicated by an arrowhead in the gel image and by a bold capital G in the sequence.

The results demonstrate that WhiH indeed is a DNA-binding protein and that the N-terminally truncated protein binds to a site in its own promoter, supporting previous indications that WhiH is an autoregulator (Ryding et al., 1998). WhiH does not bind directly to either of the two inverted repeats that were previously speculated to be possible binding sites (Ryding et al., 1998) and binds instead to a motif between these repeats (Fig. 3a). The binding site contains another inverted repeat (here indicated in capitals TTTcGTGtgcttttCACcAAA), and it conforms with the common 5′-(N)yGT(Nx)AC(Ny)-3′ sequence motif arrived at by alignment of multiple palindromic operator sequences for GntR family proteins (Rigali et al., 2002), although the spacing of the two half-sites is unusually long. The DNA sequence upstream of the whiH promoter region, including the main binding site and parts of the extended footprint, is highly conserved between different streptomycete genomes (Fig. S2). The binding motif is conserved also in Streptomyces aureofaciens, but does not coincide with any of the binding sites for two so far uncharacterized proteins from S. aureofaciens cell extracts that were reported to bind to the whiHp promoter region (Kormanec et al., 1999; Homerova & Kormanec, 2001).

The identified WhiH-binding motif should be helpful in the search for other WhiH target promoters, but we find no obvious candidate sites with similar sequences associated with predicted promoters in the S. coelicolor genome (data not shown). Importantly, the ftsZ promoter region does not contain any putative WhiH-binding motifs, suggesting that WhiH affects developmental control of cell division via some other route than direct control of ftsZ transcription.