In this work, the tagged cytoplasmic domain of SenS SenSc, as well as the full-length differently tagged SenR, and corresponding mutant proteins carrying speci-fic amino acid exchanges we
Trang 1response to phosphorylation by the sensor histidine
autokinase SenS from Streptomyces reticuli
Gabriele Bogel, Hildgund Schrempf and Darı´o Ortiz de Orue´ Lucana
FB Biologie ⁄ Chemie, Universita¨t Osnabru¨ck, Germany
One of the major signal transduction systems
govern-ing bacterial responses and adaptation to
environmen-tal 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 mod-ule, 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 reg-ulatory region of target gene(s) or operons [3,4] The
Keywords
DNA binding; phosphorylation;
Streptomyces; two-component system
SenS–SenR
Correspondence
D Ortiz de Orue´ Lucana, Universita¨t
Osnabru¨ck, FB Biologie ⁄ Chemie,
Angewandte Genetik der Mikroorganismen,
Barbarastr 13, 49069 Osnabru¨ck, Germany
Fax: +49 541 9692804
Tel: +49 541 9693439
E-mail: ortiz@biologie.uni-osnabrueck.de
(Received 13 March 2007, revised 7 June
2007, accepted 7 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 mycel-ium-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
speci-fic 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 competi-tion and mutacompeti-tional 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 Com-parative analyses reveal that only the genomes of a few actinobacteria encode two-component systems that are closely related to SenS–SenR
Abbreviations
EMSA, electrophoretic mobility shift assay; LC, liquid chromatography; RR, response regulator; SenRP, phosphorylated SenR; SenS c , cytoplasmic domain of SenS; SenScP, phosphorylated SenS c ; SK, sensor histidine kinase; TCS, two-component system.
Trang 2well-studied receiver domain within the nitrogen
regu-latory protein C) controlling the transcription of
genes involved in nitrogen metabolism) has been
shown to change its topology upon activation by
phos-phorylation [5] Generally, the signaling pathway
includes a phosphatase that returns the RR to the
non-phosphorylated 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
combina-tion 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
Ger-mination of spores and subsequent elongation of germ
tubes lead to a network of vegetative hyphae In
response to nutritional stress and extracellular
signa-ling, 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
ecophysio-logic properties of the different strains [8] The
com-plete 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
actino-rhodin 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
lead-ing to altered DNA-bindlead-ing 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-span-ning portions SenR (23.2 kDa) has a C-terminal region
with a predicted helix–turn–helix motif, which is
char-acteristic 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
chromoso-mal disruption mutant showed that the presence of
SenS–SenR influences the transcription of the furS–
cpeB operon encoding the redox regulator FurS and
the catalase-peroxidase CpeB, and the hbpS gene for the secreted HbpS, representing a novel type of heme-binding 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 phosphoryla-tion cascade between the purified cytoplasmic domain
of SenS (SenSc) and SenR Using designed mutant pro-teins, the phosphorylation sites within SenScand SenR have been investigated Bandshift and footprinting analyses have allowed the characterization of the DNA-binding properties in response to phosphoryla-tion by the sensorkinase SenS
Results
Cloning of wild-type and mutant senScand senR genes and purification of fusion proteins
As shown previously, overexpression of the full-length senS gene resulted in the synthesis of an insoluble pro-tein 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 site-directed 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 resi-due in position 199 After induction with anhydrotetra-cycline, each of the corresponding E coli XL1-Blue transformants produced a SenSc fusion type in a sol-uble form within the cytoplasm Using streptactin affin-ity chromatography, the SenSc and the SenScH199A fusion protein, both with a predicted molecular mass of 27.1 kDa, were obtained (96 nmol per 1 L of culture)
in high purity (Fig 1) After proteolytic treatment with trypsin, each protein was analyzed by liquid chroma-tography⁄ 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 (car-rying designed codon exchanges) were cloned into the plasmid pET21a The resulting wild-type protein carry-ing a C-terminal His-tag (SenR) with a predicted molecular mass of 24.3 kDa was purified to homogen-eity from an E coli BL21(DE3)pLys transformant after induction with isopropyl thio-b-d-galactoside by Ni2+– nitrilotriacetic acid affinity chromatography (Fig 1)
Trang 3Correspondingly, the mutant SenRD60A and
Sen-RD65A 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
transform-ants by Ni2+–nitrilotriacetic acid affinity
chromatogra-phy (Fig 1) Surprisingly, SenRD60A seemed to be
partially degraded and aggregated From 1 L of E coli
culture, about 144 nmol of each SenR type was purified
SenScacts as a histidine autokinase in vitro
SenSc exhibited time-dependent autophosphorylation
during incubation with [32P]ATP[cP] The highest
sig-nal intensity was already achieved after 5 min of
incu-bation (Fig 2A) The subsequent addition of an excess
of unlabeled ATP resulted in a constant level of
phos-phorylated SenSc(SenScP) over a relatively long
per-iod (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
proce-dures) 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
phos-pholigand Thus, after treatment of SenScP with 1 m
HCl, the labeled phosphate group was lost from the protein, but it was retained in the presence of 1 m NaOH (Fig 2D) This is the characteristic feature of a phosphoamidate, which is stable under alkaline condi-tions but is sensitive to acidic condicondi-tions, under which rapid aminolysis at pH < 5.5 is induced [18] Taken together, the presented data show clearly that SenS is
a histidine autokinase
SenScphosphorylates and dephosphorylates SenR
As SenR was predicted to be the cognate RR of the
SK SenS, the transfer of radiolabeled phosphate from
Fig 1 Expression and purification of SenS c and SenR proteins
Sol-uble protein extracts containing SenS c obtained from E coli
XL1-Blue pASK2 (lane 1) after induction with anhydrotetracycline
(lane 2) were loaded onto a streptactin column After washing (see
Experimental procedures), SenS c was eluted with buffer W
contain-ing 2.5 m M 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 Ni 2+
–nitrilo-triacetic acid-containing agarose column Bound SenR was eluted
with solution A containing 250 m M 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.
Fig 2 Phosphorylation analysis of SenSc (A) To test its autokinase activity, the purified SenS c protein (74 pmol) was incubated in kin-ase buffer containing 0.05 lCi of [32P]ATP[cP] at 30 C for the indi-cated period Each sample was then separated by SDS ⁄ PAGE; subsequently, the gel was dried and exposed on an X-ray-sensitive film (B) After 4 min of self-phosphorylation of SenS c , an excess of unlabeled ATP was added to the samples Each reaction was ter-minated by adding an equal amount of 4 · sample buffer After electrophoresis, the gel was dried and exposed on an X-ray-sensi-tive film (C) SenSc (148 pmol) or SenScH199A (148 pmol) was incubated in the kinase buffer with 0.05 lCi of [ 32 P]ATP[cP] for
5 min at 30 C After the addition of 4 · 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 [ 32 P]ATP[cP] in kinase buffer for 5 min at 30 C, the reaction was terminated by adding 4 · sample buffer and subjected
to SDS ⁄ PAGE Each gel was treated with the indicated solutions, dried, and exposed on an X-ray-sensitive film.
Trang 4SenSc to SenR was investigated For this purpose, the
purified SenR was added to the32
P-autophosphorylat-ed 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
acetyl-phosphate 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
phosphoryla-tion process [16] Site-directed mutagenesis showed
that each of the two residues was replaced by an
alan-ine SenRD60A and SenRD65A were subsequently
purified from corresponding E coli transformants (see
above) Further transphosphorylation analysis revealed
that the presence of SenScP provoked
phospholabe-ling of wild-type SenR and SenRD60A In contrast,
the mutant protein SenRD65A was not found to be
phosphorylated by SenScP (Fig 3C) D65 is therefore the phosphorylation site within SenR
As demonstrated by quantitative analysis (using a PhosphorImager system), during the transphosphoryla-tion reactransphosphoryla-tion dephosphorylatransphosphoryla-tion of phosphorylated SenR (SenRP) occurred after aproximately 3 min of incubation (Fig 3B); during this period, no rephospho-rylation of SenScwas recorded To investigate this pro-cess in more detail, phospholabeled SenR (carrying a His-tag) was separated immediately after phosphoryla-tion from SenSc(carrying a Strep-tag) by Ni2+ –nitrilo-triacetic acid affinity chromatography The addition of dephosphorylated SenScto a reaction mixture contain-ing phospholabeled SenR provoked a rapid (within
60 s) loss of the phosphoryl group from SenR (Fig 4A,B) In the absence of SenSc, autodephospho-rylation of SenRP occurred only after a longer (> 120 s) period of incubation (data not shown) These data show that SenScalso 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 4 Dephosphorylation rate of SenRP (A) SenR was first phos-phorylated by SenScP in a transphosphorylation reaction, and sub-sequently separated from it by Ni 2+ –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 4 · sample buffer, and the products were analyzed by SDS ⁄ PAGE Gels were dried and exposed on an X-ray-sensitive film (B) Dried gels were further analyzed using a Phos-phorImager 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 SenS c
Fig 3 Phosphotransfer from SenSc to SenR, SenRD60A or
Sen-RD65A (A, B) Purified SenSc (184 pmol) was incubated with
0.05 lCi of [32P]ATP[cP] for self-phosphorylation After 4 min, equal
amounts of purified SenR were added and incubated for the
indica-ted period at 30 C The reactions were terminated by adding
4 · 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 SenS c P (r) using a
PhosphorImager (B) (C) The wild-type SenR or SenR mutant
pro-teins (SenRD60A or SenRD65A), in each case 330 pmol of protein,
were mixed with 260 pmol of SenScP in transphosphorylation
buffer for 1 min at 30 C Reactions were terminated with 4 ·
sam-ple 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).
Trang 5transcripts (furS–cpeB and hbpS) and the
correspond-ing 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
ampli-fied 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
8 pmol) of SenR led to one type of retarded DNA spe-cies (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 5 Gene organization and EMSAs with isolated SenR proteins (A) The gene organization of furS–cpeB, hbpS, senS and senR is indica-ted 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 up-furS1 (5200 pmol) was incubated with constant amounts (16 pmol) of SenR and increasing amounts of unlabeled up-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% polyacryla-mide 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 0 (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.
Trang 6from left) Furthermore, SenR was not found to
inter-act with the upstream region (up-cpeB) of the cpeB
gene (Fig 5B, fourth box from left)
EMSAs (also known as bandshift assays) with
up-hbpS1 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
up-hbpS1 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)
Identification of the SenR-binding sites
To identify the exact DNA-binding site(s) within
up-furS1 and up-hbpS1, DNaseI footprinting
experi-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 pro-tected a region spanning 9 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 interac-tion 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 furS1 fragment ( 100 bp, named up-furS2) 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 com-petition using constant amounts of SenR and increas-ing amounts of unlabeled up-furS2 DI (Fig 7D)
Table 1 Relative binding affinity of wild-type and mutated SenR
proteins for 32 P-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
radio-activity level of the sample without protein was set at 100% The
experiments were repeated four times; the obtained data were
reproducible.
Dephosphorylated protein
Phosphorylated protein SenR SenRD60A SenRD65A SenR SenRD60A
up-furS1 6.2 > 41 11.1 34.6 > 41
Fig 6 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 m M Tris ⁄ HCl (pH 7.9), 5 lgÆmL)1sonicated salmon sperm DNA, 5% glycerol, 40 m M KCl, 2 m M MgCl2 and 2 m M
dithiothreitol After treatment with DNaseI, analyses were per-formed 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.
Trang 7DNaseI protection assays using the amplified
inter-genic region between hbpS and senS (up-hbpS1)
revealed two SenR-binding sites (II and III) spanning
21 bp (II: ACCTCCAGTAGAGCCTGGGCT) and
19 bp (III: GGACCGGGCCGCGTCCCGT) (Fig 6B)
Site II is located near to the start codon of hbpS, and site III is relatively distant from it After incubation with SenRP, the ends of both sites became hyper-sensitive to DNaseI treatment This was accompanied
by an apparent extension of region II as well as of
A
B
C
D
Fig 7 EMSAs with mutated DNA regions up-furS and up-hbpS (A) Portions of the DNA fragments (up-furS2 or up-hbpS2, see below) con-taining 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 >>> <<<) or deleted (DPIII, dotted lines) perfect inverted repeat overlapping the binding site III are shown (B) The 100 bp upstream region of the furS–cpeB operon (up-furS2) (14 200 pmol) and the corresponding mutated region (up-furS2 DI) (14 200 pmol) were incubated without or with increasing amounts (8, 16 and 24 pmol; black triangle) of SenR (left) or SenRP (right) in incubation buffer at
30 C for 15 min (C) The intergenic region ( 100 bp) between hbpS and senS hbpS2) (9400 pmol) and the mutated counterparts (up-hbpS2 DII; DIII; DPIII; DII + DIII) (9400 pmol) were incubated without or with increasing amounts (8, 16 and 24 pmol; black triangle) of SenR (top) or SenRP (bottom), in incubation buffer (D) For competition experiments, labeled up-furS2 (14 200 pmol) was incubated with con-stant amounts (24 pmol) of SenR and increasing amounts of unlabeled up-furS2 DI (0, 14200, 21300 or 28400 pmol; open triangle) (left) In the same manner, unlabeled up-hbpS2 DII + DIII (0, 9400, 14 100 or 18 800 pmol; open triangle) was added to the mixture comprising labe-led up-hbpS2 (9400 pmol) and constant amounts (24 pmol) of SenR (right) The control DNA in the mixture without SenR is (B, C, D) marked
as lane 0 The analyses were performed with 5% polyacrylamide gels, and autoradiography.
Trang 8region III (Fig 6B) Further bandshift assays showed
that the shortened up-hbpS1 fragment ( 100 bp,
named up-hbpS2) containing both motifs (II + III,
Fig 7A) was still targeted by SenR and SenRP
(Fig 7C) The up-hbpS2 fragment lacking the perfect
inverted repeat (DPIII; Fig 7A) interacts only slightly
with SenRP (Fig 7C) Analyses with fragments with
either site II (DIII or site III (DIII) or both (DII +
DIII) deleted revealed that both binding sites are
required for interaction with SenR, independent of its
phosphorylation status (Fig 7C) The specificity of the
SenR–up-hbpS2 interaction was further corroborated
by competition using constant amounts of SenR and
increasing amounts of unlabeled up-hbpS2 DII + DIII
(Fig 7D)
Taken together, these data confirm the specificity of
the SenR-binding sites and corroborate the assumption
that phosphorylation of SenR by SenSP alters its
DNA-binding characteristics
Discussion
The designed cloning procedures allowed us to obtain
the cytoplasmic domain of SenS carrying a Strep-tag
(SenSc) and the full-length SenR protein with a
His-tag (SenR) at high purity as a basis for in vitro studies
SenScwas found to function as an efficient autokinase
Chemical stability assays revealed that the ligand
within the phosphorylated SenSc(SenScP) must be a
phosphoamidate, which is extremely acid-labile but
rel-atively base-stable This feature discriminates all
phos-phorylated amino acid residues (except arginine) from
phosphoamidates [18] Additional mutational
investi-gations demonstrated that SenSc requires the histidine
residue at position 199 for autokinase activity The
in vitrotransfer of the phosphate group from SenScto
SenR (dephosphorylated SenR) occurred very rapidly,
but did not occur in a designed mutant SenR protein
carrying a substitution of the aspartic acid residue at
position 65 (D65) The kinase SenSc was found to act
additionally as a phosphatase for SenRP
(phosphor-ylated SenR)
Bandshifts revealed that SenR, but not SenRP,
binds specifically to a region (site I) upstream of the
furS–cpeB operon encoding the redox regulator FurS
and the catalase-peroxidase CpeB The deletion of
site I abolishes the interaction with SenR Interestingly,
this site is located within the previously determined
FurS operator [19] and overlaps with its central region
The data imply that, in addition to FurS, SenR
partici-pates in regulating the transcription of the furS–cpeB
operon This conclusion is supported by the earlier
finding that the absence of a functional furS or senR
gene provokes enhanced transcription of the furS–cpeB operon [16,20] Overlapping DNA-binding sites have also been described for other known regulators Depending on the physiologic condition, either the activator NhaR or the RR RcsB from E coli interacts with overlapping motifs within the upstream region of osmC This gene encodes a predicted envelope protein that is required for resistance to organic peroxides and also for long-term survival in the stationary phase [21,22] The regulator PutR and the activator CRP from Vibrio vulnificus bind simultaneously to overlap-ping sites but probably to opposite faces This process leads to activation of the transcription of the operon encoding a proline dehydrogenase and a proline perm-ease [23] PutR has been suggested to facilitate the DNA binding of CRP by direct protein–protein inter-action or to induce a change in DNA topology that allows more efficient recruitment of CRP
Comparative transcriptional and biochemical studies have revealed that SenS–SenR modulates the transcrip-tion of the furS–cpeB operon as well as the hbpS gene encoding a novel heme-binding protein Interestingly, SenR has a high affinity for the intergenic region between hbpS and senS, spanning 21 bp (site II) and
19 bp (site III) Both became hypersensitive to DNaseI treatment at their ends after incubation with SenRP, indicating altered DNA topology The phosphorylated form of RRs has been shown to provoke oligomerization and to bind cooperatively to target DNA sequences [24,25] Altered DNA binding upon phosphorylation was observed, for example, for the
RR RegR of the RegS–RegR system from Bradyrhizo-bium japonicum, controlling the expression of numerous genes, the products of which are either directly involved
in nitrogen fixation or in functions associated with the microaerobic lifestyle of this symbiont [26] A corres-ponding observation was also made for MisR of the TCS MisR–MisS from Neisseria meningitides, which is required for its pathogenicity [24], and for NtrC of the NtrB–NtrC system, which controls the expression of genes involved in nitrogen metabolism in Rhodobacter capsulatus[27]
The two SenR-binding sites II and III share a com-mon motif CNTCCNGT in the same orientation Additionally, binding site III is localized within a region (CGGCCCGGACCGGGCCG) representing a perfect inverted repeat (Fig 7) The use of DNA frag-ments lacking either binding site II, binding site III or both showed that each of them is necessary for
speci-fic targeting by SenR Further single replacements within each binding site (I, II or III) may reveal the essential role of single nucleotides in the specific inter-action with SenR The position of the SenR operator
Trang 9(sites II and III) indicates that the transcription of
senS–senR is autoregulated As reported previously
[16], SenS–SenR shows similarity to the ChrS–ChrA
system from Corynebacterium diphtheriae ChrA has
so far not been purified, but it has been predicted to
modulate the transcription of the heme oxygenase
gene (hmuO) [28,29] On the basis of our data, we
identified a DNA region upstream of hmuO with high
similarity to SenR-binding site II and an additional
shared sequence (GGGCGTCGG) near to its 3¢-end
(data not shown) This is in accordance with the fact
that the helix–turn–helix DNA-binding domains of
SenR and ChrA share 61% amino acid identity (data
not shown)
The designed SenRD60 and SenRD65A proteins
showed reduced DNA-binding affinity for up-furS as
well as for up-hbpS D60 and D65, together with other
aspartic acid residues (in positions 19 and 20), in SenR
correspond to those that have been predicted to form
an acidic pocket within RRs containing a CheY-like
receiver domain [30] Mutations at any of the acidic
pocket aspartates result in loss of functionality [31]
SDS⁄ PAGE analysis of purified SenR proteins
revealed that SenRD60A appeared to be partially
degraded and aggregated (Fig 1) Both SenRD60A
and SenRD65A seem to be perturbed in their
confor-mation and hence show altered DNA-binding abilities
Our previous data revealed that the presence of
SenS–SenR considerably enhances the resistance of
S reticuli to hemin or the redox cycling compound
plumbagin, suggesting its relevance in sensing of redox
changes [16] Further preliminary comparative analysis
(data not shown) revealed that under different
redox-stress conditions, the presence of SenS–SenR is
required for the production of additional extracellular
proteins, whose characteristics remain to be clarified
One key part of the sensing processes is expected to be
orchestrated by the heme-binding protein HbpS [17]
How it participates in delivering signals to SenS will
be explored
Sequence comparisons showed that the relative
organization of senS and senR is identical to those of
homologous genes within the S coelicolor A3(2)
gen-ome [32]; these genes are also preceded by an
uncharacterized gene that is closely related to hbpS
[16] Further sequence alignments revealed the presence
of other predicted TCSs showing high amino acid
identity with SenS–SenR within Rhodococcussp
RHA1 [33] and Arthrobacter aurescens TC1 [34]
Inter-estingly, each of the corresponding homologous genes
is also preceded by a close homolog of hbpS, the
organization of which is identical to that within the
S reticuli genome The fact that each corresponding
intergenic region comprises motifs that are related to the SenR-binding sites (II and III) (data not shown) indicates that these homologous systems are also auto-regulated According to these findings, it could be assumed that HbpS and SenS–SenR, and probably also the corresponding homologs from the other men-tioned actinobacteria, interact together to mediate an appropriate response to environmental changes Such a mode of interaction has been recently postulated for the lipoprotein LpqB and the TCS MtrA–MtrB, which together might form an actinobacterial three-compo-nent system [35] The elucidation of the exact role of accessory proteins for the modulation of bacterial TCSs might give new insights into the complex net-work of signaling processes
Taking the presented and previous data into account, the TCS SenS–SenR from S reticuli is a model that is well suited to elucidate the role of other related TCSs from different actinobacteria
Experimental procedures
Bacterial strains, plasmids, media and culture conditions
The plasmid pUC18 [36] was a gift from J Messing (State University of New Jersey, Piscataway, NJ, USA) The
E coli strains DH5a [37], BL21(DE3)pLys (Novagen, Darmstadt, Germany) and XL1-Blue [38], and the plasmids pET21a (Novagen) and pASK-IBA7 (IBA, Go¨ttingen, Germany), were used for routine cloning purposes The constructs pUKS10 (pUC18 derivative) and pWKB1 (pWHM3 derivative), which contain the furS–cpeB operon, have been described previously [16,20] E coli strains were grown in LB medium at 37C [36]
Chemicals and enzymes Chemicals for SDS⁄ PAGE were obtained from Serva (Hei-delberg, Germany) Molecular weight markers were sup-plied by Sigma (Steinheim, Germany) Restriction enzymes, T4 ligase, T4 polynucleotide kinase, DNaseI and Pfu DNA polymerase for PCR were obtained from New England Bio-labs (Frankfurt am Main, Germany), Roche (Mannheim, Germany), or Promega (Mannheim, Germany)
Isolation of DNA and transformations Plasmids were isolated from E coli with the aid of a mini plasmid kit (Qiagen, Hilden, Germany) E coli DH5a and XL1-Blue were transformed with plasmid DNA by electro-poration [39], whereas BL21(DE3)pLys was transformed with the CaCl2method as previously described [36]
Trang 10PCR, DNA sequencing and computer analysis
PCR was performed with Pfu DNA polymerase To test
the correctness of cloned genes, sequencing was done using
the Ready Reaction mix and ABI PRISM equipment (PE
Biosystems, Foster City, CA, USA) by the departmental
sequence service (U Coja, FB Biologie, University of
Osnabrueck) For DNaseI footprinting studies, sequencing
was done using the AutoRead Sequencing Kit (Amersham
Biosciences, Freiburg, Germany); however, radioactively
labeled primers corresponded to those utilized for the
EMSAs (see below) Sequence entry, primary analysis and
ORF searches were performed using clone manager 5.0
Database searches using the PAM120 scoring matrix were
carried out with blast algorithms (blastx, blastp and
tblastn) on the NCBI file server [40] Multiple sequence
alignments were generated by means of the clustalw
(1.74) program [41]
Site-directed mutagenesis within senS and senR
A point mutation in the senS gene on plasmid pQS1 [16]
was introduced using the QuikChange Site-Directed
Muta-genesis kit (Stratagene, Amsterdam, The Netherlands) with
the following specific primers: H199A1, 5¢-CCCGGGAGA
TCGCCGACACCCTCGC-3¢; and H199A2, 5¢-GCGAGG
GTGTCGGCGATCTCCCGGG-3¢ These oligonucleotides
were designed to replace the selected histidine codon at
posi-tion 199 (H199) with an alanine codon (underlined) The
constructs were transformed into E coli XL1-Blue
Subse-quently, each of the cloned inserts in the plasmid DNA,
iso-lated from several individual transformants, was analyzed
with restriction enzymes and by sequencing The resulting
correct plasmid was named pQS1H199A
For introducing specific mutations in the senR gene, a
dif-ferent strategy was used First, the subconstruct pUR1 was
created by ligation of the longer SphI–BamHI fragment of
pUC18 with the 1.8 kb SphI–BamHI fragment (containing
senR) of pWKB1 The plasmid pUR1 was then used for
PCRs and further ligation The desired PCR products were
obtained using the forward primer R1NSph (5¢-CAGCGC
ATGCTGCTCCAGGCAGCCGAC-3¢), and one of the
reverse primers R1D65Pst (5¢-CGAGCTGCAGGGCCAT
CAGGACGACGTC-3¢) or R1D60Pst (5¢-CGAGCTGCAG
GTCCATCAGGACGACGGCCGGGGCGGTCTTGC-3¢)
The PstI restriction site is in bold type within R1D60Pst
and R1D65Pst, and the codon for alanine replacing the
ori-ginal codon (GTC) for aspartic acid is underlined; the SphI
restriction site within R1NSph is in bold type The cleaved
PCR product resulted in a 0.76 kb DNA fragment
contain-ing a part of the mutagenized senR gene (D60A or D65A)
Each of these was then ligated with the large SphI–PstI
frag-ment (3.75 kb) of plasmid pUR1 The resulting constructs
(isolated from E coli DH5a transformants) were named
pUR1D60A or pUR1D65A, respectively The correctness of
the in-frame replacement was controlled by restriction and sequencing
Cloning of genes in E coli The DNA region (from the nucleotide at position 505 to that
at position 1197 of senS) encoding the cytoplasmic part (comprising the segment from the aspartic acid at position 169 to the arginine at position 398) of SenS was amplified by PCR with the following primers: SenS3, 5¢-CTAGAATTCGACGACCTGGTC-3¢, harboring an EcoRI restriction site (in bold type), and SenS4, 5¢-GATCTGCAG TCATCTCGGCTC-3¢, containing the PstI restriction site (in bold type) The plasmids pWKB1 [16] and pQS1H199A were used as DNA templates Each PCR product was diges-ted with EcoRI and PstI and ligadiges-ted with EcoRI–PstI-digested pASK-IBA7 The resulting construct pASKS2 or pASKS2H199A was transformed into E coli XL1-Blue, recovered, and sequenced Transformants containing the cor-rect constructs and producing the designed fusion protein with Strep-tag codons (SenScor SenScH199A) were selected The senR-coding region of plasmid pWKB1, pUR1D60A
or pUR1D65A was amplified by PCR with following prim-ers: SenR1, 5¢-CCCATATGACCCCCACCCCGCAGCCG CC-3¢, consisting of an NdeI restriction site (in bold type), followed by the sequence encoding the N-terminal amino acids of SenR, and SenR2, 5¢-CGCGCTCGAGAGACAGG AGGCGTTGTTC-3¢, determining the C-terminal amino acids of SenR, followed by the XhoI restriction site (in bold type) The PCR products were digested with NdeI and XhoI, ligated with the NdeI–XhoI-cleaved pET21a, and sub-sequently transformed into E coli XL1-Blue Having sequenced several of the resulting plasmids pETR1, pET-R1D60A or pETR1D65A, we verified the correctness of the senRgene and the in-frame fusion with the His-tag codons
Purification of the fusion proteins
An E coli XL1-Blue transformant containing the pASKS2
or pASKS2H199A plasmid was inoculated in LB medium with ampicillin (100 lgÆmL)1) and cultivated at 37C The synthesis of the SenSc or the SenScH199A fusion protein was induced (at a D600of 0.6) by adding anhydrotetracycline (200 ngÆmL)1) The cells were grown for 4 h, harvested, washed with buffer W (100 mm Tris⁄ HCl, pH 8.0, 150 mm NaCl, 1 mm EDTA), and disrupted in the same buffer W by ultrasonication (10· 10 s, with 10 s intervals) using a Bran-son Bran-sonifier (Danbury, CT, USA) The fusion protein was purified directly from a cleared cell lysate using streptactin affinity chromatography according to the instructions of the manufacturer (IBA)
Plasmid pETR1, pETR1D60A or pETR1D65A was transformed into E coli BL21(DE3)pLys A selected trans-formant containing the desired plasmid was inoculated
in LB medium containing ampicillin (100 lgÆmL)1) and