ST1710–DNA complex crystal structure reveals the DNA binding mechanism of the MarR family of ...

13 Pages·2009·7.17 MB·English

Published online 9 June 2009 Nucleic Acids Research, 2009, Vol. 37, No. 14 4723–4735
doi:10.1093/nar/gkp496
ST1710–DNA complex crystal structure reveals
the DNA binding mechanism of the MarR family
of regulators
Thirumananseri Kumarevel1,*, Tomoyuki Tanaka1, Takashi Umehara2 and
Shigeyuki Yokoyama2,3
1RIKEN SPring-8 Center, Harima Institute, 1-1-1 Kouto, Sayo, Hyogo 679-5148, 2Systems and Structural Biology
Center, Yokohama Institute, RIKEN, 1-7-22 Suehiro-cho, Tsurumi, Yokohama 230-0045 and 3Department of D
o
Biophysics and Biochemistry, Graduate School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, w
n
Tokyo 113-0033, Japan loa
d
e
d
Received April 24, 2009; Revised May 21, 2009; Accepted May 21, 2009 fro
m
h
ttp
s
ABSTRACT marRAB operon, and repression of this operon is alle- ://a
viated by exposure to a variety of phenolic compounds, ca
ST1710, a member of the multiple antibiotic resis- d
most notably sodium salicylate (1). Similarly, MexR neg- e
tance regulator (MarR) family of regulatory proteins m
atively regulates an operon in Pseudomonas aeruginosa ic
in bacteria and archaea, plays important roles that, when expressed, encodes a tri-partite multi-drug .ou
in development of antibiotic resistance, a global efflux system that results in an increased resistance to p.c
o
health problem. Here, we present the crystal struc- multiple antibiotics, including tetracycline, b-lactams, m
ture of ST1710 from Sulfolobus tokodaii strain 7 chloramphenicol,novobiocin,trimethoprim,sulfonamides /na
complexed with salicylate, a well-known inhibitor and fluoroquinolones (4,5). Some members of the MarR r/a
of MarR proteins and the ST1710 complex with its family of DNA-binding proteins, for example hypotheti- rticle
promoterDNA,refinedto1.8and2.10A˚ resolutions, cal uricase regulator (HucR) and organic hydroperoxide -a
b
resistanceregulator(OhrR),mediateacellularresponseto s
rtoesppoelocgtiyveolyf.aTphoe-SSTT11771100–aDnNdAMcoarmRplperxotsehianrse,swtihthe rreaadciotidvueraonxsidHauticvRe wstaressssho(RwOnSt)o(r6e,p7r)e.ssThites oDweninoecxopcrceuss- tract/3
each subunit containing a winged helix-turn-helix sion as well as that of an uricase. This repression is alle- 7/1
(wHtH) DNA binding motif. Significantly large con- viatedbothinvivoandinvitrouponbindinguricacid,the 4/4
7
formational changes occurred upon DNA binding substrateforuricase.Asuricacidisapotentscavengerof 23
and in each of the dimeric monomers in the asym- reactive oxygen species, and D. radiodurans is known for /10
metric unit of the ST1710–DNA complex. Conserved its remarkable resistance to DNA-damaging agents, these 91
4
wHtH loop residues interacting with the bound observations indicate a novel oxidative stress response 5
9
DNA and mutagenic analysis indicated that R89, mechanism (8–10). Similar to HucR, the OhrR protein by
R90 and K91 were important for DNA recognition. of Bacillus subtilis also mediates a response to oxidative gu
e
Siniggnmifieccahnatnlyi,smth.e bound DNA exhibited a new bind- srterseidssu;ehobwyeovregra,nfoicrOhyhdrRro,pietriosxoixdiedsattihoantoafbarolognaetecsysDteNinAe st on
binding (11,12). 04
WehavereportedtwodifferentcrystalformsofST1710 A
p
INTRODUCTION (13) and others (14). The structure showed the winged ril 2
helix-turn-helix (wHtH) motif at the DNA binding site 0
1
Microbial antibiotic resistance is a result of either inacti- that obviously belonged to the MarR family of proteins. 9
vation or reduced accumulation of antibiotics within an ThecrystalstructuresofproteinsintheMarRfamilyhave
organism. Proteins belonging to the multiple antibiotic alsobeendeterminedfromanumberoforganismsinclud-
resistance regulators (MarR) family reportedly regulate ing MarR from E. coli (15), MexR from P. aeruginosa
theexpressionofproteinsconferringresistancetomultiple (16),SarRfromStaphylococcusaureus(17),Slya-likepro-
antibiotics, organic solvents, household disinfectants, teinfromEnterococcusfaecalis(18),OhrRfromB.subtilis
oxidative stress agents and pathogenic factors (1–3). For (19), HucR from D. radiodurans (20) and MTH313 from
example, in the absence of the appropriate stimulus, Methanobacterium thermoautotrophicum (21). Sequence
Escherichia coli MarR proteins negatively regulate the comparisons of these proteins with ST1710 showed less
*To whom correspondence should be addressed. Tel: [phone] (ext 7894); Fax: [phone]; Email: [email protected]
(cid:2)2009TheAuthor(s)
ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommonsAttributionNon-CommercialLicense(http://creativecommons.org/licenses/
by-nc/2.0/uk/)whichpermitsunrestrictednon-commercialuse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
4724 NucleicAcidsResearch,2009,Vol.37,No.14
than 25% identity. A homology search in the non-redun- distinct mode of DNA binding was observed with
dant protein database using Blastp revealed that ST1710 the bound DNA passing over the protein molecule
has about 51% identity to the Sulfolobus acidocaldarius rather than passing through the 2-fold related axis of the
(22) and Sulfolobus solfataricus (23) sequences and about molecule as previously observed (19).
41% identity to the Metallosphaera sedula sequence (24).
However, none of the proteins closely related to ST1710
have been biochemically or structurally characterized. MATERIALS AND METHODS
Sodium salicylate is well known to inhibit MarR activ-
Cloning, expression and purification of ST1710
ity both in vitro and in vivo, at millimolar concentration
levels (25). Sodium salicylate is routinely used as a model The gene encoding the MarR family regulator protein
inhibitor of MarR to induce MarA expression in E. coli (ST1710) from Sulfolobus tokodaii (S. tokodaii) strain 7
and Salmonella typhimurium, thereby conferring a Mar was amplified from genomic DNA by PCR, using the
phenotype (25–28). The structure of E. coli MarR was primers 50-ggaattCATATGTTAGAAAGTAATGAAAA Do
w
solved with two salicylate molecules per monomer and CAGAATAC-30 and 50-ggaattGGATCCTTATTACTGA n
both of them are highly exposed to the solvent. It seems CTAATTTCCTCAATTCTTTTC-30. The PCR fragment loa
d
that salicylate may have stabilized the crystal packing was digested with NdeI and BamHI and cloned into the ed
since in the absence of salicylate, the crystals could not pET21a(+) expression vector. The plasmid was trans- fro
m
be used for structure determination in the case of E. coli formed into the E. coli BL21-CodonPlus (DE3)—RIL-X
h
MarR (15). Recently, the structure of MTH313, a MarR strain (Stratagene), and the selenomethionine-containing ttp
homologfromM.thermoautotrophicum,wassolvedinthe ST1710 proteins were over-expressed and purified as s://a
free form and complexed with salicylate; these analyses described in our earlier studies (13). c
a
revealed a large asymmetrical conformational change d
e
m
that is mediated by the binding of sodium salicylate to Crystallization and data collection ic
two distinct locations in the dimer (21). .o
u
recTohgenimzeemdboeurbsleofstthraenMdeadrRDfNamAilybyoftrheegiurlawtoHrtyHprmotoetiinfss N(0a.1ti(cid:4)ve0c.1ry(cid:4)st0a.l1smofmS)Tw1i7t1h0inwtehreepperroiodducoedf atowmeeekdiautm20si8zCe p.com
(15–19). Footprinting experiments revealed that MarR bythesittingdropvapordiffusionmethod(32),byadding /n
a
binds as a dimer at two different, but similar, sites in 1mlofproteinsolutionto1mlofwellsolution,containing r/a
marO, protecting 21bp of DNA on both strands at a 18% PEG 8K, 0.2M calcium acetate, and 0.1M sodium rtic
single site without bending its target (29,30). One of the cacodylate, pH 6.5. For the ST1710–salicylate complex, le-a
MarR families of proteins, the OhrR protein complexed crystals were soaked in the mother liquor containing b
s
with the ohrA operator with a 29-bp duplex was solved, 0.2Mofsodiumsalicylatefor3min.Twentypercentethy- tra
which revealed the interactions between them. The leneglycolwasusedasacryo-protectantandthecomplete ct/3
protein–DNA contact region included the major groove datasetwascollectedforthenativeandsalicylatecomplex 7/1
of the (cid:2)10 element, and indicated that OhrR, and proba- with the in-house R-axis VII system (RIGAGU MSC). 4/4
bly MarR and MexR as well, repress transcription by These crystals belonged to the tetragonal space group, 72
3
blocking the access of RNA polymerase to this promoter P41212, and the processed data statistics are given in /1
0
element (19). In addition, the mutational analysis of the Table 1. The 30bp synthetic oligonucleotide containing 9
1
RNA polymerase binding site, the (cid:2)10 element of the the putative promoter sequence (50-AATAATGTCATT 45
OhrR-, MarR- and MexR-regulated promoters, revealed GTTAACAATAGCAAAAAT-30)anditscomplementary 9 b
the loss of DNA binding ability by (cid:3)10-fold when this oligonucleotide (50-ATTTTTGCTATTGTTAACAATG y g
region was altered (11,16,30,31). ACATTATT-30) were annealed completely to form the ue
s
On the basis of the sequence of the ohrA promoter, we DNA-duplex and complexed with the ST1710 protein t o
n
previously identified a putative promoter for ST1710 and at an equimolar ratio. Initial crystals of ST1710–DNA 0
4
showed binding ability by gel-mobility shift assays (13). complex were produced at 208C by the sitting drop A
TAhTiGsporfomthoetSerTi1s7l1o0cagteendeiamnmdeddoiwatneslytrueapmstroefamtheoSfTthSe1fi7r0s9t vDaNpoArcdoimffupslieoxnsomluettihoondto(312m),lboyfwaedldlisnoglu1tiomnl,ocfopnrtaoitneiinng– pril 20
gene. To understand the importance of MarR family 30% (v/v) 2-Methyl-2,4-pentanediol (MPD), 0.02M cal- 19
membersinantibioticresistanceandotherbiologicalpro- cium chloride dehydrate and 0.1M sodium acetate trihy-
cesses,here,wesolvedtheST1710inthreedifferentforms: drate,pH3.8.Preliminarytinycrystalsthatwereseededin
(i) apo-form (native), (ii) complexed with its inhibitor, theequilibrated dropsofthesameconditiongrewup toa
sodium salicylate (salicylate complex) and (iii) complexed maximum dimension of 0.3mm within a 2-week period.
with its promoter DNA. A slight conformational change Complete Multiple Anomalous Dispersion (MAD) data
on theside chains ofprotein residues’ was observed when setswereobtainedat100KusingaJupiter210CCDdetec-
boundtothesalicylateligand,comparedtotheapo-form. tor (RIGAGU MSC) on the RIKEN structural genomics
The DNA bound to the wHtH motif of one monomer on beamline I (BL26B1) at SPring-8, Hyogo, Japan. These
the dimeric ST1710, and specifically interacted with the crystals belonged to the centered orthorhombic space
residues R84, R89, R90 and K91. A significantly large group C222 , with cell dimensions a=94.44, b=106.73
1
conformational change was observed between the mono- and c=82.26A˚ . The native and Se-edge MAD data sets
mers of the dimeric protein bound to the DNA, and also wereprocessedupto2.10A˚ usingtheHKL2000suite(33)
withtheapo/salicylatecomplexstructures.Significantly,a (Table 1).
NucleicAcidsResearch,2009,Vol.37,No.14 4725
Structure determination and refinement were expressed and purified in a manner similar to the
native protein (13).
The native and salicylate complex of ST1710 structures
Toevaluatetheprotein–DNAinteractionsinsolution,a
were determined by the molecular replacement method,
gel-mobilityshiftassaywasused.ThepurifiedST1710was
using our previous ST1710 structure (PDB code, 2eb7),
incubated in binding buffer (10mM Tris–HCl, pH 8.0,
as a search model. The solution was found by auto-
200mM NaCl, 20mM MgCl and 5mM b-ME) at pre-
mated-MOLREP, within the CCP4 program suite, and 2
determined concentrations, and the 30-mer DNA
the refinement was carried out using CNS (34). The pro-
(100nM) was added. The reaction mixture was incubated
tein model was built using the programs Quanta (35) and
for 20min at room temperature and mixed with 1ml of
Coot (36). The native and salicylate complexes were
50% glycerol before loading onto the gel. The free DNA
refined to resolutions 1.80 and 2.0A˚ , respectively
and ST1710–DNA complexes were resolved on a 10%
(Table 1). Since the molecular replacement method was
polyacrylamidegel(runningbuffer,1XTBE,constantvol-
not successful for phasing the ST1710–DNA complex, D
tage, 200V; temperature, 48C). Inhibition of ST1710 was o
we collected and processed the Se-MAD data sets. The w
analyzed in the presence of increasing concentrations of n
ST1710–DNA complex structure was successfully phased lo
sodium salicylate by gel-mobility shift analysis. The free a
by the MAD method with the three different wavelength d
and complexed nucleic acids were stained by fluorescent ed
dSaOtLaVsEets(37c)o.llSeocltveedntafltatthteeniSneg-eadngde,inuitsiianlgmothdeelpbruoigldrianmg SYBR Green (EMSA Kit, Invitrogen) and the bands from
visualized with a UV transilluminator (LAS-3000,
were performed by RESOLVE (37). Improvement of the h
partial model derived from RESOLVE was performed FUJIFILM, Japan). ttps
with the ARP/wARP program (38). We observed unam- ://a
c
biguous density for the DNA bases and built the DNA- RESULTS AND DISCUSSION ad
e
model using the program Quanta (35). The final model m
was refined and manually fitted using CNS (33), Coot Binihoipbhityiosincaalssaanyaslysis of salicylate binding to ST1710 and ic.o
(36) and Quanta (34). The final model with 285 protein up
residues and 23 nucleic acid bases, except for 4 and 3 Previous in vitro and in vivo analyses of MarR fam- .co
m
residuesintheN-terminalofAandBchains,respectively, ily of proteins suggested that salicylate is a broad inhibi- /n
w(Rafsreer=efi0n.e2d87)toata2.c1r0yAs˚talrleosgorlauptihoinc, Ru-sfiancgtosrynocfhro0t.2ro3n7 tleovrelfsor(1M).aTrRo ainctvievsittiygaatet twhehemthielrlimsoallaicrylcaotencebnintrdastiotno ar/artic
radiation X-ray data collected at cryo temperature (see ST1710, we used a differential scanning calorimetric le
Table 1). Figures were prepared with the program method for the binding analysis. Various concentrations -ab
s
Pymol (39). The coordinates and structure factors for of sodium salicylate (0–300mM) were mixed with the tra
the native, salicylate, and ST1710–DNA complex have protein and the heat capacities (Cp) measured with the ct/3
beendepositedintheProteinDataBank,undertheacces- scan rate of 908C/hr (Figure 1A). The binding constant 7
/1
sion codes 3GEZ, 3GF2 and 3GFI, respectively. (K ) was 20(cid:5)4.9mM for sodium salicylate as calculated 4
d /4
by the GraphPad Prism software, while the Origin 7
2
Differential scanning calorimetric (DSC) analysis of program produced similar values. This analysis clearly 3/1
salicylate binding to ST1710 suggests that salicylate binds to ST1710, and the binding 09
1
constants were comparable with other members of MarR 4
DSC experiments were carried out using a VP-capillary 5
family (21). 9
DSC platform (Microcal, USA). For the DSC measure- b
Next, to visualize the protein–DNA inhibition by y
ments,theproteinconcentrationwasfixedat0.5mg/mlin sodium salicylate, we used the gel-mobility shift assays. gu
e
20mM Tris–HCl buffer (pH 8.0) containing 150mM s
NaCl. Dialyzed protein sample and the sodium salicylate WthahnilefrteheeDSNT1A71a0lo–nDeN, tAhecoadmdpitleioxnfoorfmsaedlicmylaotreeicnohmibpitaecdt t on
were filtered through a 0.22mM pore size membrane and the DNA–protein complex. At concentrations above 04
complexedwithdifferentconcentrations ofsodiumsalicy- A
100mM, the most of the DNA in the salicylate treated p
late (0–300mM) and were loaded on to the capillary complex released from ST1710. This analysis clearly ril 2
system. The scan rate was 908C/hr for all experiments. 0
demonstrates that the ST1710–DNA complex formation 1
9
Wecalculatedthebindingconstant(K )ofsodiumsalicy-
d was inhibited with increasing concentrations of salicylate
late using the DSC curves analysis using the Origin soft-
(Figure 1B). Taken together, the data indicates that
ware (Microcal, USA) and the Graphpad Prism 2.0, a
salicylate bound to the ST1710 in solution and inhibited
non-linear curve-fitting algorithm (GraphPad software).
the protein–DNA complex when it exceeded the intracel-
lular concentration levels.
Site-directed mutagenesis of ST1710 and gel-mobility
shift assays Structure of ST1710 complexed with salicylate
Initially, the ST1710 plasmid was prepared with a Qiagen Our DSC and gel-mobility shift analyses confirmed that
miniprep kit. The Quickchange site-directed mutagenesis salicylate bound totheST1710 and inhibited theprotein–
kit (Stratagene) was used to create the DNA-binding site DNA complex formation. To see how salicylate binds to
mutants(R89A,R90AandK91A),andtheresultantplas- ST1710, we determined the ST1710–salicylate complex at
mids were transformed into JM109 cells. N-terminal a resolution of 1.80A˚ , and refined to a final R value of
sequencing was carried out for all the mutants, which 23.3% and an R value of 26.5%. The overall structure
free
4726 NucleicAcidsResearch,2009,Vol.37,No.14
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
s
://a
c
a
d
e
m
ic
Figure1. SalicylatebindingandinhibitionofST1710–DNA complexassays.(A)DSCanalysisofsalicylatebindingtoST1710.Typicalexcessheat .ou
capacity curves of ST1710 in the absence/presence of sodium salicylate ligand, at a scan rate of 908C/hr. Salicylate concentration and peak tem- p.c
perature are noted on each curve. The binding constant was calculated using the non-linear regression fit, using the GraphPad Prism software. o
m
(B)Gel-mobilityshiftassayshowingST1710–DNAcomplexinhibition.Allreactionswerecarriedoutinbindingbuffercontaining150mMofprotein /n
andvariousamountsofsodiumsalicylate(toafinalconcentrationof0–250mM;lanes1–16with0,1,5,10,20,30,40,50,60,70,80,90,100,150, a
200and250mM,respectively)wereaddedandincubatedatroomtemperaturefor20min.Tothisreactionmixture,100nMof30-merDNAadded, r/a
and after for 20min., the reactions were fractionated by 10% native PAGE and the DNA stained by SYBR Green (EMSA Kit, Invitrogen). rtic
The positions of the free and complex DNA are indicated by an arrow and an arrowhead, respectively. le
-a
b
s
tra
c
of ST1710 within the complex was similar to our recently domains formed a salicylate-binding pocket. The salicy- t/3
7
reported native structure, showing that it belongs to the lateligandhasmanyinteractionswiththeproteinresidues /1
a/b family of proteins and resembles those of the MarR (Figure 2B). The O20 of salicylate is bonded to the side 4/4
family of proteins. It consists of six a-helices and two chain oxygens of Y37 and Y111; in addition, side chain 72
3
b-strands, arranged in the order of a1-a2-a3-a4-b1-b2- oxygen of Y37 is also bonded to the O10 of the ligand /10
a5-a6intheprimarystructure(Figure2A).Theasymmet- molecule. The ligand oxygen O 0 is hydrogen bonded to 9
1 1
roifc6s0ub(cid:4)u3n9it(cid:4)co2n6tAa˚i.nTswonoemmoonloecmuelers,wariethreolvaeteradllbdyimacernyssiotanls- tthhee Osid2eofchliagianndamminoolecgurloeuips h(NydHro2)geonfbreosnidduede tRo20th,ewshidilee 459 by
lographic 2-fold axis to form the dimer (Supplementary chainnitrogenofK17.Thelattertwointeractionsarefrom gu
Figure S1), and this is consistent with our gel-filtration the symmetrically related molecule. Thus, the bound sali- e
s
analysis (13) as well as with studies of other MarR cylate has many interactions. All of the residues which t o
n
familyproteins(15–19).TheN-andC-terminalequivalent interact with the ligand are highly conserved among the 0
4
residues of each monomer, located at the a1, a5 and a6 closely related species (Figure 1E). A
p
hdeolmicaesin,,awrehicclhosiselsytaibnitleizretwdibnyedhyadnrdopfhoormbicaanddimheyrdizraotgioenn inNtheexts,atloicyolbasteervliegaifnadneydccoonmfoprlemxawtiohnenalccohmanpgaereodcctuorrtehde ril 20
1
bonding interactions between the residues located within nativestructure,wecollectedthenewnativedatasetfrom 9
these regions. Apart from the dimerization domain, as the crystal grown under the same conditions, solved
observed in many DNA binding transcriptional regula- at 2.0A˚ and refined to a final R value of 21.1% and
tors, the residues located at the a2-a3-a4-b1-b2 formed a an R value of 25.2%. The overall conformation of
free
wHtH DNA binding motif. In the dimer, the distances the complex is very similar to the native structure, with
between the recognition helix (a4) to the recognition an rmsd of 0.11A˚ for superposition of 141 C atoms
a
helix and the loop to loop (connecting the b-strands) of (Figure 2C). However, a minor conformational change
the wHtH domains are (cid:3)30 and (cid:3)70A˚ , respectively. was observed in the side chain orientations of the ligand
The fine quality of the electron density map allowed us interacting residues and the DNA-binding wHtH motifs.
to identify unambiguously the specific salicylate binding We believe that allosteric changes might occur at the
site in the complex structure. The bound salicylate is molecular level in solutions in the presence of inhibiting
located at the interface between the helical dimerization ligand,salicylate,whichisnotseeninthecrystalstructure.
and wHtH DNA-binding domains (Figures 2A and B). A divalent metal ion, Ca2+, was observed in the native
The large portion of the DNA-binding and dimerization structure, and it interacted with the C-terminal Q and
146
NucleicAcidsResearch,2009,Vol.37,No.14 4727
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
s
://a
c
a
d
e
m
ic
.o
u
p
.c
o
m
/n
a
r/a
rtic
le
-a
b
s
tra
c
t/3
7
/1
4
/4
7
2
3
/1
0
9
1
4
5
9
b
y
g
u
e
s
t o
n
0
4
A
p
SFTig1u7r1e02–.saSlitcryulcattuerecoomfpSleTx1.7T1h0ensaetciovendaanrdysstarluiccytluarteeacsosmignpmlexenatnsdansdeqtuheenNce-/cCo-mtepramriisnoinarweiltahbcelloedseolynrtehleatsetdrupcrtuorteei.nTs.he(Ab)ouAndribsabloicnyldaitaegirsasmhoowfnthine ril 2
0
the stick model. (B) An electrostatic representation of the ST1710 monomer. The basic regions are shown in blue, and the acidic regions are red. 1
9
Close-up view of salicylate binding site interactions with protein residues. The hydrogen bonds are indicated by broken lines. (C) Structural
comparison of the ST1710–salicylate complex with the native structure. The native structure is shown in orange color. The salicylate and key
residuesinvolvedininteractionsareshowninstickmodelsandtheboundCa2+ioninthenativestructureisrepresentedbyanorangesphere.The
Ca2+ionbindingsiteisenlarged.(D)SuperimpositionoftheST1710–salicylatecomplexalongwithanotherknownMarRfamilyproteincrystallized
with salicylate. The E. coli MarR is represented by blue color and M. thermoautotrophicum MTH313 is shown in red. (E) Sequence analysis of
ST1710(S.tok)anditscloselyrelatedproteinsfromdifferentspecies:S.acidocaldarius(S.aci),S.solfataricus(S.sol),andM.sedula(M.sed)along
with the OhrR protein. Conserved residues are indicated by red letters. The secondary structural elements in the primary sequences of ST1710 are
indicatedasa-helices (bars),b-strands(arrows)andloops(lines).ThesalicylateandDNAcontacting residuesinST1710areshowninblue-shaded
and yellow-shaded boxes, respectively. DNA contacting residues ofthe wing in OhrR are indicated by red asterisks. Selected homologous/identical
residues that interact with DNA in OhrR are indicated by blue asterisks.
E residues, which is closer to the DNA-binding loop complex structures available; MarR from E. coli and
98
of the symmetrically related molecule (Figure 2C). the more recently solved MTH313 from M. thermoauto-
It is interesting to compare the salicylate complexes trophicum. Although these proteins displayed sequence
within the MarR family of regulators. There are two similarity (cid:3)26% to ST1710, the superposition of the
4728 NucleicAcidsResearch,2009,Vol.37,No.14
ST1710–salicylate complex with the two others revealed moleculeofthefull-lengthpromoterDNA.Thefull-length
that the overall topology is similar, with an rmsd of (30-mer)oftheDNA-duplexinthecrystalswerealsocon-
1.90A˚ for 119 C atoms and 2.1A˚ for 131 C atoms for firmed by dissolving the complex crystals (ST1710–DNA)
a a
E. coli MarR and M. thermoautotrophicum MTH313, inwater,afterwashingthemafewtimesinreservoirsolu-
respectively(Figure2D).However,E.coliMarRwascrys- tion, and then analyzed DNA-duplex, using agarose gel
tallizedwithtwomoleculesofsalicylateperdimer,bothof electrophoresis. These analyses indicated that the DNA-
which were highly exposed to the solvent (15), and these duplex present in the crystal was, indeed, the full-length
salicylate binding sites are not comparable to that with DNA (Supplementary Figure S3). To clarify further the
ST1710. The salicylate ligands in MarR hydrogen DNAbindingtoST1710,wepreparedanadditionalDNA
bonded to some of the amino acid residues (A70, T72, duplex with exactly 2-fold related sequence based on the
R77 and R86); however, the physiological relevance of observedfragmentsinthepresentcrystal(50-TTGCTATT
either salicylate binding site could not be determined GTTAACAATAGCAAAAAT-30) (Figure 3A, bottom
D
becausetheligandswereinvolvedininteractionswithpro- sequence),crystallized,andsolvedthestructure.Theover- o
w
tein molecules within the crystal which may stabilize the all structure of this new complex resembles the present n
lo
crystal lattice. On the other hand, the mode of salicylate one, and it recognizes the DNA fragment irrespective of a
d
bindingbetweenST1710andMTH313iscomparable,and the sequence heterogeneity (our unpublished data). As ed
the ligand adjusted up (cid:3)2A˚ and (cid:3)3A˚ towards the a5 shown in Figure 3E, one full-length duplex-DNA was fro
m
helix upon binding. In contrast to the salicylate binding interacted with four molecules of the ST1710, which was h
in ST1710, two direct and one water mediated protein coming from four different asu of the dimeric molecules. ttp
s
residues were in contact with the ligand in MTH313. A Thus, the analysis of the ST1710–DNA complex in the ://a
comparison between the apo and salicylate complex of present study suggests that the protein:DNA ratio is 2:1. c
a
MTH313 revealed a significant asymmetrical conforma- Interestingly, the bound DNA passes over the wHtH de
m
tional change that is mediated by the binding of sodium motif, making contacts only at the loop regions. ic
salicylate to two distinct locations in the dimer (21), .o
u
p
whereas we did not observe such changes in the case of Interactions between the ST1710 and promoter DNA .c
o
ST1710.Theavailableinvivoandinvitroanalysessuggest m
that the MarR family of regulators inhibits the activity in As explained above, Figure 4A represents the full length /na
thepresenceofsalicylate.Sincewecouldseeonlythefixed duplex-DNA bound to four monomers of the symmetry- r/a
side-chain orientation of the protein residues contacting related dimeric molecules. Although we used only a rtic
30-mer duplexed DNA for our crystallization studies, we le
the bound salicylate ligand, we believe possible dynamic -a
couldseethe duplexed-DNA consistingof T5to A27and b
or allosteric changes occurring upon salicylate binding s
to ST1710 to inhibit its activity would not be captured T50 toA270 ofthebasesboundtotheprotein(Figures4A tra
by crystallization. and B). The 4 and 3 bases at the 50- and 30-end, respec- ct/3
tively,werehighlydisorderedinbothoftheDNA-strands 7
/1
and hence not modeled. The bound DNA adapted a 4
Overall structure of the ST1710–DNA operator complex /4
B-form right handed structure, passing over the protein 7
2
3
In our earlier studies, we identified the putative promoter molecule by only contacting at the wHtH loop regions. /1
0
for ST1710, which is located upstream of the st1710 The protein–DNA interactions seen in the asu of the 9
1
gene (13). The gel-mobility shift analysis suggested that complex (Figures 3B and C) were essentially same in all 4
5
9
protein–DNA interactions were competitive and concen- of the four symmetry-related molecules. Interestingly, as b
tration dependent (Supplementary Figure S2). Here, we observed in the OhrR-ohrA operator complex, the –10 y g
u
complexed the ST1710 and 30-mer duplexed DNA region (TAACAAT) of the promoter DNA (15–21) was e
s
(Figure 3A), crystallized, and collected the data set up to recognized by the wHtH domains (Figures 4B–F). Of the t o
a resolution of 2.1A˚ , under the space group C2221. bound 54 nucleotides, only 22 nucleotides make 36 con- n 0
4
Initially, we were unsuccessful in solving the structure by tacts with six protein residues (Figures 4B–F). The side A
mstroulecctuurlae.rLreaptlearc,ewmeesnotluvesidngthteheSTco1o7r1d0i–nDatNesAofcoomurpnleaxtibvye cInhtaeirnesotixnygglye,nthoefsSid65e cwhaasinb(oNnHded) otforetshieduOe5R0 offoTrhmye5d0. pril 2
1 84 0
the MAD method using three different data sets collected water-mediated hydrogen bonds to the N3 of bases G130 19
at the Se-edge, and the structure was refined to a final R andAde .Inaddition,sidechain(NH )ofR hydrogen
17 1 89
valueof23.7%andanR value of28.7%.Ahighqual- bonded to the backbone phosphate oxygen (O P) of
free 2
ity electron density map enabled us to build the nucleic Thy 0. The residue R hydrogen bonded to the O 0 and
14 90 4
acidsunambiguously(Figure3B).Theoverallstructureof O of base Cyt and the same residue made two salt-
2 18
ST1710–DNA complex is shown in Figures 3C–E. The bridge contacts with D . This salt bridge may assist in
88
protein crystallized as a dimer in the asymmetric unit fixing the conformation of residue R in order to make
90
(asu)withtheDNAfragmentboundtooneofthedimeric contactwiththenucleicacidbase,Cyt .Besides,theside
18
molecules (Figure 3C). Part of the packing diagram with chainatom(CD)bondedtothebasesofGua130 (N )and
2
two and four asymmetric unit of the molecules related by Thy140 (O ). The side chain of K interacted with back-
2 91
the crystallographic 2-fold axis is also given (Figures 3D bone phosphate of Ade and Ile to C of Ade . Thus,
19 91 5 20
andE).Theduplex-DNAfragmentsobservedintwoadja- thefollowing residuesS ,R ,D ,R ,R ,K andI
65 84 88 89 90 91 92
cent asu of the dimeric molecules shown in magenta and interacted with the bound promoter DNA. To evaluate
red(Figure3E),probablyformonedouble-strandedDNA the protein–DNA interactions at the loop region,
NucleicAcidsResearch,2009,Vol.37,No.14 4729
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
s
://a
c
a
d
e
m
ic
.o
u
p
.c
o
m
/n
a
r/a
rtic
le
-a
b
s
tra
c
t/3
7
/1
4
/4
7
2
3
/1
0
9
1
4
5
9
b
y
g
u
e
s
t o
n
0
4
A
p
ril 2
0
1
9
Figure3. StructureofST1710–DNAcomplex.(A)DNAsequenceusedforcrystallization(30-merand26-merduplex).Thebluebackgroundshows
the2-foldrelatedsequence.Theredboxedresidueswereobservedinthecrystalstructure.(B)Thefinal2Fo-Fcomitelectrondensitymapwiththe
nucleic acids contoured at 1s level is shown in blue mesh. The template strand and its complementary strands are shown in blue and yellow stick
models.(C)StereoviewoftheasymmetricunitofST1710–DNAcomplex.Thesecondarystructuralassignments,N-andC-terminalendsarelabeled
inoneofthedimeric monomers.Thebound nucleicacidsareshown asstick representations. (D)Stereoviewofthetwodimersofthe asymmetric
unitrelatedby2-foldaxistoformthetetramerforrecognitionofonepromoterDNA.(E)Partofthecrystalpackingisshown(stereoview).Each
asymmetric unit of the complex is shown in a unique color. The 50- and 30- ends of each nucleotide chain is labeled.
4730 NucleicAcidsResearch,2009,Vol.37,No.14
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
s
://a
c
a
d
e
m
ic
.o
u
p
.c
o
m
/n
a
r/a
rtic
le
-a
b
s
tra
c
t/3
7
/1
4
/4
7
2
3
/1
0
9
1
4
5
9
b
y
g
u
e
s
t o
n
0
4
A
p
ril 2
0
1
9
Figure4. ST1710–promoter DNAinteractions. (A)AribbondiagramofST1710–DNA complex coloredasinFigure 4E. Thefull-length promoter
DNA is shown in stick model. The T5-A27 and T50-A270 strands are in blue and yellow, respectively. (B) Schematic representation showing all
ST1710–DNAcontacts.Thebasesareshowninrectangles.Theprotein–DNAcontactsineachoftheST1710monomerarerepresentedbythesame
color in (A). (C–F) The close-up view of the critical protein–DNA interactions in the complex is shown in A–D. The residues of nucleic acids and
protein are shown in stick models. The hydrogen bonds are shown in broken lines.
NucleicAcidsResearch,2009,Vol.37,No.14 4731
we prepared three mutant proteins (R A, R90A and mode of recognition, the wHtH loop regions were stabi-
89
K A) and analyzed the binding ability by gel-mobility lized. The temperature factor for this loop region in the
91
shift assays. All three mutants failed to bind to DNA, B-chainincreasedsignificantly,althoughitformedadimer
suggestingthatthesethreeresiduesareimportantforpro- related by a non-crystallographic 2-fold axis, observed in
tein–DNA interactions (Supplementary Figure S4). We theasymmetricunitofthecell(SupplementaryFigureS7).
also crystallized all three mutant proteins under the It is noteworthy to mention that the temperature factor
native protein conditions and solved their structures by was higher for the DNA-complex B-chain when not only
molecular replacement method as explained previously compared to the DNA-complex A-chain, but also to the
(Table 1). All three mutant structures resembled the native, salicylate complex and all three of the mutant
native ST1710, except very little changes were observed structures. Apparently, when protein binds to the DNA,
in the wHtH loop regions (Supplementary Figure S5). the wHtH loop becomes stabilized and the temperature
Furthermore, DNA-binding residues in ST1710 were factorofthat region islowered;this observation wassim-
D
highly conserved among the closely related proteins and ilar to that of the salicylate-complex which was solved at o
w
OhrR (Figure 2E). The winged loop region connecting the highest resolution. It is also interesting to note that n
lo
the strands b1 and b2 apparently plays a major role in in the dimer, the distances between the loop to loop a
d
modulating their conformation for binding to the DNA (connecting the b-strands) of the wHtH domains were ed
molecule, and this mode of recognition is anticipated for reduced by (cid:3)10A˚ for the ST1710–DNA complex, com- fro
m
theproteinscloselyrelatedtoST1710aswellasthefamily pared to the native and salicylate complexes. h
of MarR regulators. We observed Ca2+ ions in all of the ttp
s
mutantandnativestructures,butnotinthesalicylateand Mode of nucleic-acid binding ://a
DNA-complex structures. Superimposition of the native, c
salicylateandDNA-complexstructuressuggestedthatthe There are several MarR families of regulators reported ade
to date from different organisms including E. coli m
Cfr-otmermthienamlhetealilxio(an6b)iinndtihnegDsiNteAancodminpltehxedseavliiacyteladtgerceoamtly- (15), P. aeruginosa (16), S. aureus (17), E. faecalis (18), ic.o
u
B.subtilis(19),D.radiodurans(20)andM.thermoautotro- p
plexslightchangesinsidechainorientationswereobserved .c
phicum (21). However, only the structure of OhrR from o
(SupplementaryFigureS6).However,thefunctionsofthis m
metal ion observed in the native and mutant structures of B.subtilisisavailablewithitspromoterandrevealedtheir /na
ST1710will need tobe furtherinvestigated. interactions.ItisintriguingtocompareourST1710–DNA r/a
complex with OhrR-ohrA operator binding to clarify the rtic
binding mechanism. The superimposition of the ST1710– le
Conformational changes -a
DNA complex on the OhrR-ohrA operator revealed large b
s
Inourrecentreport,wenoticedasmallchangeonlyatthe conformationalchanges,withanrmsdof4.4A˚ for215Ca tra
c
loop region connecting strands b1 and b2 in the protein atoms when compared to ST1710 before binding to its t/3
conformers crystallized in two different space groups and cognate promoter (rmsd 2.80 for 238 C atoms) 7
a /1
theoverallstructurewasidenticalwithanrmsdof0.519A˚ (Figures6AandB).ComparedtotheST1710–DNAcom- 4
/4
for 141 C atoms (13). In a similar way, when we com- plexwithitsnativestructureandsalicylatecomplexinthe 7
a 2
3
pared the present ST1710–salicylate complex and native previoussection,wefoundasimilarchangeintheprotein /1
0
structure crystallized under the same conditions, a similar and unexpectedly found conformational changes in the 9
1
structural conformation was revealed with an rmsd of mode of DNA recognition also. In the OhrR-ohrA com- 4
5
0.11A˚ for superposition of 141 Ca atoms. Additionally, plex, the distance between the subunits loop-to-loop was 9 b
the subunits in the dimer were identical. In contrast, the around 67A˚ and the recognition helix (a4)(cid:2)helix was y g
superimposition of the ST1710–DNA complex subunits about 20A˚ (Figure 6A), although the wings of the subu- ue
s
(A and B chains) on one another revealed a large local nits translocated about 16A˚ compared to the structure of t o
n
conformational change all along the structure, excluding reduced OhrR (19). The bound 2-fold related promoter 0
the helices a1 and a5 with an rmsd of 2.85A˚ for 142 Ca sequence recognized the protein wHtH loop-to-loop, 4 A
ailtaorm(Fs;ighuorwee5vAer),.tAhedoisvpelraaclelmsternutctoufra(cid:3)l3t.o5p–o5l.o5gAy˚ wwaass ssiemen- wgriothovseubthstaatntrieaslulwteiddenfrinogmainndserdteioenpenoifngtheofrethcoegnmitaijoonr pril 2
0
allalongthewingedHtHmotifregionandtheC-terminal helix (a4) of the HtH motif. In contrast to this mode of 19
helix showed the displacement of around 2–3A˚ . The binding, the bound DNA in ST1710 passed over the
winged HtH motif of A-chain where the DNA is recog- wHTH motif without deepening the structure through
nized was elevated up compared to the B-chain, while the the 2-fold axis, even though the protein contacting resi-
C-terminal helix a6 was shifted down. duesarehighlyconservedbetweenthesetwoproteinsand
As seen in Figure 5B, significant changes were also amongtheMarRfamilyofregulators(Figures2E,4Aand
observed between the subunits of the DNA-complex and 6B). This unexpected mode of DNA-binding originated
the native and salicylate complex. Superposition of the due to the translocation of one of the subunits around
native and salicylate complexes on to the A-chain 13A˚ towards the 2-fold axis, reducing the distance
showed greater differences than on the B-chain: an rmsd between the recognition helix of the subunits to 13A˚
of 2.6 and 2.9A˚ with the A-chain; and 0.9 and 1.0A˚ with (Figures 6A and B). Thus, the DNA passing through the
the B-chain, for the native structure and salicylate com- 2-fold axis deepening the recognition helices as observed
plex,respectively.OneofthewHtHmotifsmayhavebeen in OhrR-ohrA operator complex would be impossible for
relocated to establish contact with the DNA, and by this that of ST1710.
4732 NucleicAcidsResearch,2009,Vol.37,No.14
K91A 1.54178P42211ab==45.96,c=138.29150–2.10(2.18–2.10)916312.2(12.9)98.1(100)0.095(0.302) 20.0–2.108997 8258 0.218739 0.2701138 127 136.9 96.23.8 3GFM measurements.
i
0) all
9
1. or
R90A 1.54178P42211ab==45.90,c=137.42140–1.90(1.97– 1208513.0(12.6)97.7(95.1)0.065(0.326) 20–1.9011906 10943 0.205963 0.2391136 160 133.5 97.72.3 3GFL Ihvalueof()f
n D
a o
DatacollectionandrefinementstatisticsoftheST1710native,salicylate,DNAcomplexandtheirmutantsTable1. DatacollectionST1710-salicylateST1710-nativeST1710–DNAcomplexR89A PeakEdgeRemote(Lowenergy) Wavelength1.541781.541780.978960.979301.0001.5417822P422C222P422SpacegroupP41111111˚ababababCelldimensions(A)==46.14,==46.04,=94.44,=106.73,==45.86,cccc=137.64=138.45=82.26=138.17No.ofmolecules/asu1121˚a40–2.0(2.07–2.0)50–2.10(2.18–2.10)40–2.20(2.28–2.20)ResolutionRange(A)40–1.80(2.07–2.0) Uniquereflections13381103802434924204242318143Redundancy13.4(7.6)18.5(19.1)9.6(9.8)12.5(13.0)Completeness(%)91.3(99.8)96.0(94.4)99.1(98.7)99.1(98.8)99.0(98.6)100(100)bR0.100(0.264)0.104(0.321)0.080(0.294)0.072(0.284)0.067(0.281)0.091(0.389)merge Refinementstatistics˚ResolutionRange(A)20–1.8020–2.020.0–2.1020–2.20Reflectionsusedinthe1313810244242027877refinementTotalno.ofreflections122469423224767292usedforworkingsetcR0.2330.2110.2370.201work8928211726585Totalno.ofreflectionsusedforRfreedR0.2650.2520.2870.245freeNo.ofproteinatoms1142114223051136No.ofnucleicacidatoms––466No.ofwatermolecules152156154125No.ligandmolecules12+ions–1–1No.Ca˚2)29.631.851.228.1AverageBfactor(A RamachandranstatisticsMostfavoredregions(%)97.798.598.197.0Allowedregions(%)2.31.51.93.0 3GEZ3GF23GFI3GFJPDBcode ˚RMSDinbondanglesandbondlengthsarevariedfrom0.005to0.007(A)and1.0–1.1(8).Solventcontentisabout55%.aValuesinparenthesesareforthehighestresolutionshell.PPPPb(cid:2)RIhi<Ih>IhiIhiih<Ih>=|(,)()|/(,),where(,)istheintensityvalueofthethmeasurementofand()isthecorrespondingmehihimergePc(cid:2)RFFFFFfactor=||||||/||,where||and||aretheobservedandcalculatedstructurefactoramplitudes,respectively.obscalcobsobscalcdRisthesameasRfactor,butfora5–7%subsetofallreflections.free wnloaded from https://academic.oup.com/nar/article-abstract/37/14/4723/1091459 by guest on 04 April 2019