DNA-binding characteristics of the regulator SenR in response to phosphorylation by the sensor histidine autokinase SenS from Streptomyces reticuli ´ ´ Gabriele Bogel, Hildgund Schrempf and Darıo Ortiz de Orue Lucana FB Biologie Chemie, Universitat Osnabruck, Germany ă ă Keywords DNA binding; phosphorylation; Streptomyces; two-component system SenS–SenR Correspondence ´ D Ortiz de Orue Lucana, Universitat ă Osnabruck, FB Biologie Chemie, ă Angewandte Genetik der Mikroorganismen, Barbarastr 13, 49069 Osnabruck, Germany ă Fax: +49 541 9692804 Tel: +49 541 9693439 E-mail: ortiz@biologie.uni-osnabrueck.de (Received 13 March 2007, revised June 2007, accepted June 2007) doi:10.1111/j.1742-4658.2007.05923.x The two-component system SenS–SenR from Streptomyces reticuli has been shown to influence the production of the redox regulator FurS, the mycelium-associated enzyme CpeB, which displays heme-dependent catalase and peroxidase activity as well as heme-independent manganese peroxidase activity, and the extracellular heme-binding protein HbpS In addition, it was suggested to participate in the sensing of redox changes In this work, the tagged cytoplasmic domain of SenS (SenSc), as well as the full-length differently tagged SenR, and corresponding mutant proteins carrying specific amino acid exchanges were purified after heterologous expression in Escherichia coli In vitro, SenSc is autophosphorylated to SenScP at the histidine residue at position 199, transfers the phosphate group to the aspartic acid residue at position 65 in SenR, and acts as a phosphatase for SenRP Bandshift and footprinting assays in combination with competition and mutational analyses revealed that only unphosphorylated SenR binds to specific sites upstream of the furS–cpeB operon Further specific sites within the regulatory region, common to the oppositely orientated senS and hbpS genes, were recognized by SenR Upon its phosphorylation, the DNA-binding affinity of this area was enhanced These data, together with previous in vivo studies using mutants lacking functional senS and senR, indicate that the two-component SenS–SenR system governs the transcription of the furS–cpeB operon, senS–senR and the hbpS gene Comparative analyses reveal that only the genomes of a few actinobacteria encode two-component systems that are closely related to SenS–SenR One of the major signal transduction systems governing bacterial responses and adaptation to environmental changes is the two-component system (TCS) A typical TCS consists of an autophosphorylating sensor histidine kinase (SK) and a cognate response regulator (RR) [1] SKs detect stimuli via an extracellular input domain or intracellular signals via cytoplasmic regions, or use transmembrane regions and sometimes additional short extracellular loops for sensing [2] In addition to the N-terminal input domain, SKs contain a C-terminal portion representing the transmitter module, with several blocks of amino acid residues being conserved among these kinase types Phosphorylation within a typical SK usually takes place at a conserved histidine residue; the phosphoryl group of the SK is subsequently transferred to a conserved aspartic acid residue within the receiver domain of the RR As a result, its C-terminally located output domain has an altered DNA-binding capacity for the regulatory region of target gene(s) or operons [3,4] The Abbreviations EMSA, electrophoretic mobility shift assay; LC, liquid chromatography; RR, response regulator; SenRP, phosphorylated SenR; SenSc, cytoplasmic domain of SenS; SenScP, phosphorylated SenSc; SK, sensor histidine kinase; TCS, two-component system 3900 FEBS Journal 274 (2007) 3900–3913 ª 2007 The Authors Journal compilation ª 2007 FEBS G Bogel et al well-studied receiver domain within the nitrogen regulatory protein C ) controlling the transcription of genes involved in nitrogen metabolism ) has been shown to change its topology upon activation by phosphorylation [5] Generally, the signaling pathway includes a phosphatase that returns the RR to the nonphosphorylated state The phosphatase can exist as an individual protein, or reside on a module, which is linked either to the RR or to the kinase A combination of kinase and phosphatase activity ensures rapid coordination of the cell response [6] Streptomycetes are Gram-positive and G + C-rich bacteria with a complex developmental life cycle Germination of spores and subsequent elongation of germ tubes lead to a network of vegetative hyphae In response to nutritional stress and extracellular signaling, aerial hyphae develop, in which spores mature [7] As soil-dwelling organisms, streptomycetes need to respond to highly variable conditions The range of environmental stimuli to which a bacterium can respond is expected to correlate with the number of functional SKs and RRs These are assumed to have evolved by selection pressure for different ecophysiologic properties of the different strains [8] The complete genome sequence of Streptomyces coelicolor A3(2) comprises 84 SK genes and 80 RR genes [9] The physiologic roles of only a few of them have been investigated experimentally For instance, the AbsA1– AbsA2 system negatively regulates the production of several antibiotics [10,11], and the VanR–VanS system activates the expression of vancomycin resistance [12,13] Phosphate control of the production of actinorhodin and undecylprodigiosin in S lividans and S coelicolor A3(2) is mediated by the two-component PhoR–PhoP system, which also controls the alkaline phosphatase gene (phoA) and other phoA-related genes [14,15] To date, however, the phosphorylation cascade between a Streptomyces SK and its cognate RR leading to altered DNA-binding affinity of the RR has not been analyzed in detail The cellulose degrader S reticuli has been reported to contain the neighboring genes senS and senR, which encode an SK and an RR, respectively SenS (42.2 kDa) comprises five predicted membrane-spanning portions SenR (23.2 kDa) has a C-terminal region with a predicted helix–turn–helix motif, which is characteristic for different DNA-binding proteins [16] It was concluded that SenR is the cognate RR for the SK SenS Comparative transcriptional and biochemical studies with a designed S reticuli senS–senR chromosomal disruption mutant showed that the presence of SenS–SenR influences the transcription of the furS– cpeB operon encoding the redox regulator FurS and Response regulator SenR the catalase-peroxidase CpeB, and the hbpS gene for the secreted HbpS, representing a novel type of hemebinding protein [16] Physiologic studies showed that the production of HbpS is positively influenced by hemin in S reticuli; this correlated with increased hemin resistance Interestingly, the presence of HbpS leads to enhanced synthesis of the heme-containing CpeB [17] In this study, we describe the in vitro phosphorylation cascade between the purified cytoplasmic domain of SenS (SenSc) and SenR Using designed mutant proteins, the phosphorylation sites within SenSc and SenR have been investigated Bandshift and footprinting analyses have allowed the characterization of the DNA-binding properties in response to phosphorylation by the sensorkinase SenS Results Cloning of wild-type and mutant senSc and senR genes and purification of fusion proteins As shown previously, overexpression of the full-length senS gene resulted in the synthesis of an insoluble protein in Escherichia coli [16] To obtain a truncated SenS (comprising its predicted cytoplasmic portion; see Experimental procedures) with an N-terminal Strep-tag (SenSc), the corresponding portion of senS was cloned into the plasmid pASK-IBA7 Furthermore, using sitedirected mutagenesis, a mutant gene was designed and cloned into plasmid pASK-IBA7 (see Experimental procedures), leading to the mutant SenScH199A, which carried an alanine residue in place of the histidine residue in position 199 After induction with anhydrotetracycline, each of the corresponding E coli XL1-Blue transformants produced a SenSc fusion type in a soluble form within the cytoplasm Using streptactin affinity chromatography, the SenSc and the SenScH199A fusion protein, both with a predicted molecular mass of 27.1 kDa, were obtained (96 nmol per L of culture) in high purity (Fig 1) After proteolytic treatment with trypsin, each protein was analyzed by liquid chromatography ⁄ mass spectrometry (LC-MS), and was found to comprise the correct N-terminal and internal peptides (data not shown) The full-length senR gene and mutant senR genes (carrying designed codon exchanges) were cloned into the plasmid pET21a The resulting wild-type protein carrying a C-terminal His-tag (SenR) with a predicted molecular mass of 24.3 kDa was purified to homogeneity from an E coli BL21(DE3)pLys transformant after induction with isopropyl thio-b-d-galactoside by Ni2+– nitrilotriacetic acid affinity chromatography (Fig 1) FEBS Journal 274 (2007) 3900–3913 ª 2007 The Authors Journal compilation ª 2007 FEBS 3901 Response regulator SenR G Bogel et al Fig Expression and purification of SenSc and SenR proteins Soluble protein extracts containing SenSc obtained from E coli XL1Blue pASK2 (lane 1) after induction with anhydrotetracycline (lane 2) were loaded onto a streptactin column After washing (see Experimental procedures), SenSc was eluted with buffer W containing 2.5 mM desthiobiotin (lane 3) SenSCH199A was purified in the same manner (lane 4) To obtain SenR, a cytoplasmic protein extract (lane 5) containing SenR obtained from E coli BL21(DE3)pLys pETR1 after induction (lane 6) was loaded onto an Ni2+–nitrilotriacetic acid-containing agarose column Bound SenR was eluted with solution A containing 250 mM imidazole (lane 7) as described under Experimental procedures SenRD60A (lane 8) and SenRD65A (lane 9) were purified in the same manner The molecular masses of the protein markers (S) are indicated Correspondingly, the mutant SenRD60A and SenRD65A fusion proteins (24.3 kDa), which carried an alanine instead of the original aspartic acid residue at position 60 or 65, were purified to homogeneity from the corresponding E coli BL21(DE3)pLys transformants by Ni2+–nitrilotriacetic acid affinity chromatography (Fig 1) Surprisingly, SenRD60A seemed to be partially degraded and aggregated From L of E coli culture, about 144 nmol of each SenR type was purified SenSc acts as a histidine autokinase in vitro SenSc exhibited time-dependent autophosphorylation during incubation with [32P]ATP[cP] The highest signal intensity was already achieved after of incubation (Fig 2A) The subsequent addition of an excess of unlabeled ATP resulted in a constant level of phosphorylated SenSc (SenScP) over a relatively long period (at least 20 min; Fig 2B) Sequence alignments showed that the histidine residue at position 199 within SenS is predicted to be the phosphorylation site [16] To corroborate this assumption, the corresponding H199 codon was replaced by one for alanine using site-directed mutagenesis (see Experimental procedures) The purified SenScH199A (Fig 2C, left) failed to undergo autophosphorylation after incubation with [32P]ATP[cP] (Fig 2C, right) Chemical stability tests were applied to characterize the nature of the phospholigand Thus, after treatment of SenScP with m 3902 Fig Phosphorylation analysis of SenSc (A) To test its autokinase activity, the purified SenSc protein (74 pmol) was incubated in kinase buffer containing 0.05 lCi of [32P]ATP[cP] at 30 °C for the indicated period Each sample was then separated by SDS ⁄ PAGE; subsequently, the gel was dried and exposed on an X-ray-sensitive film (B) After of self-phosphorylation of SenSc, an excess of unlabeled ATP was added to the samples Each reaction was terminated by adding an equal amount of · sample buffer After electrophoresis, the gel was dried and exposed on an X-ray-sensitive film (C) SenSc (148 pmol) or SenScH199A (148 pmol) was incubated in the kinase buffer with 0.05 lCi of [32P]ATP[cP] for at 30 °C After the addition of · sample buffer, the reaction was stopped, and the mixture was subsequently subjected to SDS ⁄ PAGE The gel was stained with Coomassie Brilliant Blue (left), or alternatively dried and exposed on an X-ray-sensitive film (right) (D) After autophosphorylation of 74 pmol of SenSc with 0.05 lCi of [32P]ATP[cP] in kinase buffer for at 30 °C, the reaction was terminated by adding · sample buffer and subjected to SDS ⁄ PAGE Each gel was treated with the indicated solutions, dried, and exposed on an X-ray-sensitive film HCl, the labeled phosphate group was lost from the protein, but it was retained in the presence of m NaOH (Fig 2D) This is the characteristic feature of a phosphoamidate, which is stable under alkaline conditions but is sensitive to acidic conditions, under which rapid aminolysis at pH < 5.5 is induced [18] Taken together, the presented data show clearly that SenS is a histidine autokinase SenSc phosphorylates and dephosphorylates SenR As SenR was predicted to be the cognate RR of the SK SenS, the transfer of radiolabeled phosphate from FEBS Journal 274 (2007) 3900–3913 ª 2007 The Authors Journal compilation ª 2007 FEBS G Bogel et al SenSc to SenR was investigated For this purpose, the purified SenR was added to the 32P-autophosphorylated SenSc (see previous section) Very rapid (within 5– 10 s) labeling of SenR was observed, together with a concomitant reduction of the phospholabel within SenSc (Fig 3A,B) Autophosphorylation activity of SenR using [32P]ATP[cP] or the phosphodonor acetylphosphate could not be detected (data not shown) The deduced SenR comprises aspartic acid residues at position 60 (D60) and position 65 (D65), each of which is a candidate to participate in the phosphorylation process [16] Site-directed mutagenesis showed that each of the two residues was replaced by an alanine SenRD60A and SenRD65A were subsequently purified from corresponding E coli transformants (see above) Further transphosphorylation analysis revealed that the presence of SenScP provoked phospholabeling of wild-type SenR and SenRD60A In contrast, the mutant protein SenRD65A was not found to be Response regulator SenR phosphorylated by SenScP (Fig 3C) D65 is therefore the phosphorylation site within SenR As demonstrated by quantitative analysis (using a PhosphorImager system), during the transphosphorylation reaction dephosphorylation of phosphorylated SenR (SenRP) occurred after aproximately of incubation (Fig 3B); during this period, no rephosphorylation of SenSc was recorded To investigate this process in more detail, phospholabeled SenR (carrying a His-tag) was separated immediately after phosphorylation from SenSc (carrying a Strep-tag) by Ni2+–nitrilotriacetic acid affinity chromatography The addition of dephosphorylated SenSc to a reaction mixture containing phospholabeled SenR provoked a rapid (within 60 s) loss of the phosphoryl group from SenR (Fig 4A,B) In the absence of SenSc, autodephosphorylation of SenRP occurred only after a longer (> 120 s) period of incubation (data not shown) These data show that SenSc also acts as a phosphatase for SenRP DNA-binding properties of SenR depend on its phosphorylated state Comparative analysis of wild-type S reticuli and the senS–senR disruption mutant showed that the presence of SenS–SenR correlates with a significant reduction of Fig Phosphotransfer from SenSc to SenR, SenRD60A or SenRD65A (A, B) Purified SenSc (184 pmol) was incubated with 0.05 lCi of [32P]ATP[cP] for self-phosphorylation After min, equal amounts of purified SenR were added and incubated for the indicated period at 30 °C The reactions were terminated by adding · sample buffer After SDS ⁄ PAGE, the gel was dried and exposed on an X-ray-sensitive film (A) or quantified by detection of the radioactivity emitted by SenRP (j) or SenScP (r) using a PhosphorImager (B) (C) The wild-type SenR or SenR mutant proteins (SenRD60A or SenRD65A), in each case 330 pmol of protein, were mixed with 260 pmol of SenScP in transphosphorylation buffer for at 30 °C Reactions were terminated with · sample buffer, subjected to SDS ⁄ PAGE, and stained with Coomassie Brilliant Blue (left), or alternatively the gel was dried and exposed on an X-ray-sensitive film (right) Fig Dephosphorylation rate of SenRP (A) SenR was first phosphorylated by SenScP in a transphosphorylation reaction, and subsequently separated from it by Ni2+–nitrilotriacetic acid affinity chromatography Purified SenRP ( 82 pmol) was incubated at 30 °C alone (top) or with (bottom) 148 pmol of dephosphorylated SenS for the indicated times Each reaction was stopped by adding an equal amount of · sample buffer, and the products were analyzed by SDS ⁄ PAGE Gels were dried and exposed on an X-raysensitive film (B) Dried gels were further analyzed using a PhosphorImager The diagram shows the quantified results representing the measured radioactivity at the indicated times (j) with SenRP alone or for the mixture (r) of SenRP and SenSc FEBS Journal 274 (2007) 3900–3913 ª 2007 The Authors Journal compilation ª 2007 FEBS 3903 Response regulator SenR G Bogel et al transcripts (furS–cpeB and hbpS) and the corresponding proteins [16] For further analyses, different DNA fragments (Fig 5A) corresponding to the upstream region (310 bp, named up-furS1) of the furS–cpeB operon or the upstream region (548 bp, named up-hbpS1) located between hbpS and senS were amplified by PCR Electrophoretic mobility shift assays (EMSAs) were performed with labeled DNA (5200 pmol of up-furS1 or 2900 pmol of up-hbpS1) and increasing quantities (0–16 pmol) of the purified SenR or SenRP Interestingly, in contrast to SenRP, SenR interacted with up-furS1 (Fig 5B) The addition of 12 pmol of SenR to the reaction mixture led to an  84% decrease of free up-furS1, whereas the same amount of SenRP provoked only a  10% reduction (Fig 5D) The presence of small quantities (4 and pmol) of SenR led to one type of retarded DNA species (Fig 5B, arrow b); an additional one was formed if the protein concentration (12 and 16 pmol) was increased (Fig 5B, arrow a) These data suggested the presence of multiple SenR-binding sites The specificity of this interaction was verified by competition using constant amounts of SenR and additional increasing amounts of unlabeled up-furS1 (Fig 5B, third box Fig Gene organization and EMSAs with isolated SenR proteins (A) The gene organization of furS–cpeB, hbpS, senS and senR is indicated The labeled DNA regions are marked in gray (B, C) The upstream region of the furS–cpeB operon (5200 pmol of up-furS1) (B) or the intergenic region between hbpS and senS (2900 pmol of up-hbpS1) (C) was incubated without or with increasing amounts (0, 4, 8, 12 or 16 pmol; black triangle) of SenR or SenRP in incubation buffer (see Experimental procedures) For competition experiments, labeled upfurS1 (5200 pmol) was incubated with constant amounts (16 pmol) of SenR and increasing amounts of unlabeled up-furS1 (0, 5200, 7800, 10 400 or 13 000 pmol; open triangle) (B, third box from left) In the same manner, unlabeled up-hbpS1 (0, 2900, 4350, 5800 or 7250 pmol; open triangle) was added to the mixture comprising labeled up-hbpS1 (2900 pmol) and constant amounts (16 pmol) of SenR (C, third box from left) For further corroboration of the binding specificity, SenR (0–16 pmol; black triangle) was incubated with the upstream region of cpeB (up-cpeB, 5500 pmol) (B, fourth box from left) After incubation at 30 °C for 15 min, the mixtures were separated on 5% polyacrylamide gels, and then subjected to autoradiography The retarded DNA fragments are indicated (a, b, c and d) The control DNA in mixtures without SenR is everywhere marked as lane (D) In addition, gels were dried and analyzed by a PhosphorImager System The radioactivity level of the DNA probe alone was set at 100% The reaction products up-furS1 + SenR (j), up-furS1 + SenRP (h), up-hbpS1 + SenR (r), and up-hbpS1 + SenRP (e) as well as the quantities of SenR used are indicated 3904 FEBS Journal 274 (2007) 3900–3913 ª 2007 The Authors Journal compilation ª 2007 FEBS G Bogel et al Response regulator SenR from left) Furthermore, SenR was not found to interact with the upstream region (up-cpeB) of the cpeB gene (Fig 5B, fourth box from left) EMSAs (also known as bandshift assays) with uphbpS1 and varying amounts (0–16 pmol) of SenR or SenRP showed that the DNA-binding affinity was enhanced after phosphorylation (Fig 5C) This was indicated by the observation that SenR (4 pmol) led to a  63% decrease in free up-hbpS1, whereas the same amount of SenRP (4 pmol) enhanced it to  94% (Fig 5D) Interestingly, SenR induced the formation of two shifted species (Fig 5C, arrows c and d), suggesting the existence of at least two binding sites within uphbpS1 One of them (marked as d) was only observed in the presence of small quantities (4 pmol) of SenR but not with SenRP Further EMSAs using different quantities of proteins showed that, to obtain a 50% decrease in free up-hbpS1, at least three times as much SenR as SenRP was required (Table 1) Competition studies using constant amounts of SenR and increasing amounts of unlabeled up-hbpS verified the specificity of the SenR–up-hbpS1 interaction (Fig 5C, third box from left) Taken together, these data revealed that SenR binds specifically to up-furS1 and up-hbpS1, and the phosphorylation of SenR by SenSP substantially alters its DNA-binding characteristics Further bandshift assays using different amounts of purified SenR proteins demonstrated that each of the SenR mutant proteins (SenRD60A and SenRD65A) has reduced binding affinity for up-furS1 and up-hbpS1 (Table 1) ments with the purified RRs SenR and SenRP, after their phosphorylation in the presence of ATP by SenSc, were performed Footprinting experiments with radioactively labeled up-furS1 showed that SenR protected a region spanning bp (I, AACTTGGGG) against DNaseI cleavage (Fig 6A, left) In addition, a short region (marked by a white block) upstream of region I was protected, implicating probable binding sites (as observed by bandshift experiments), or a change in DNA topology being induced after interaction with SenR Increasing amounts of SenR neither extended nor altered the extent of the protection SenRP had no affinity for this DNA region, even at high concentrations (up to 60 pmol) (Fig 6A, right) A truncated up-furS1 fragment ( 100 bp, named upfurS2) comprising site I (I, Fig 7A) still interacted with SenR, as shown by bandshift assays (Fig 7B) Studies with this fragment having a deleted site I (DI) (Fig 7A) showed that it was targeted neither by SenR nor by SenRP (Fig 7B) The specificity of the SenR– up-furS2 interaction was further corroborated by competition using constant amounts of SenR and increasing amounts of unlabeled up-furS2 DI (Fig 7D) A B Identification of the SenR-binding sites To identify the exact DNA-binding site(s) within up-furS1 and up-hbpS1, DNaseI footprinting experiTable Relative binding affinity of wild-type and mutated SenR proteins for 32P-labeled DNA-fragments EMSAs were done (as described in Experimental procedures) using increasing (0– 100 pmol) amounts of the mentioned proteins and analysis was done with a PhosphorImager The indicated amount (in pmol) of each protein is required to obtain a 50% decrease of the intensity of free DNA (up-furS1 or up-hbpS1) For this purpose, the radioactivity level of the sample without protein was set at 100% The experiments were repeated four times; the obtained data were reproducible Dephosphorylated protein SenR up-furS1 up-hbpS1 Phosphorylated protein SenRD60A SenRD65A SenR SenRD60A 6.2 3.7 > 41 10.7 11.1 7.0 34.6 1.2 > 41 2.5 Fig Footprinting studies (A) up-furS1 (6900 pmol) and (B) up-hbpS1 (5800 pmol) were incubated without SenR or SenRP, or with increasing amounts (20.5, 41 and 61.5 pmol) of SenR or SenRP in 10 mM Tris ⁄ HCl (pH 7.9), lgỈmL)1 sonicated salmon sperm DNA, 5% glycerol, 40 mM KCl, mM MgCl2 and mM dithiothreitol After treatment with DNaseI, analyses were performed with 6% polyacrylamide-urea gels, and autoradiography The protected DNA regions (I, II and III) are indicated by black blocks The additional protected region within up-furS1 is indicated by a small, open rectangle FEBS Journal 274 (2007) 3900–3913 ª 2007 The Authors Journal compilation ª 2007 FEBS 3905 Response regulator SenR G Bogel et al A B C D Fig EMSAs with mutated DNA regions up-furS and up-hbpS (A) Portions of the DNA fragments (up-furS2 or up-hbpS2, see below) containing the complete (I, II, III and II + III; underlined) or deleted (DI, DII, DIII and DII + DIII; dotted lines) binding motifs, or the complete (PIII, marked by >>>