ebook img

Probing the structure of RNAIII, the Staphylococcus aureus agr regulatory RNA, and identification PDF

12 Pages·2000·0.73 MB·English
by  
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Probing the structure of RNAIII, the Staphylococcus aureus agr regulatory RNA, and identification

RNA(2000),6:668–679+CambridgeUniversityPress+PrintedintheUSA+ Copyright©2000RNASociety+ Probing the structure of RNAIII, the Staphylococcusaureusagrregulatory RNA, and identification of the RNA domain involved in repression of protein A expression YVONNE BENITO,1 FABRICEA. KOLB,2 PASCALE ROMBY,2 GERARD LINA,1 JEROME ETIENNE,1 and FRANÇOIS VANDENESCH1 1EA1655,FacultédeMédecineLaennec,69372LyonCedex08,France 2UPR9002duCentreNationaldelaRechercheScientifique,InstitutdeBiologieMoléculaireetCellulaire, 67084StrasbourgCedex,France ABSTRACT RNAIII,a514-ntRNAmolecule,regulatestheexpressionofmanyStaphylococcusaureusgenesencodingexoproteins andcell-wall-associatedproteins.WehavestudiedthestructureofRNAIIIinsolution,usingacombinationofchem- icalandenzymaticprobes.Amodelofthesecondarystructurewasderivedfromexperimentaldatawiththehelpof computer simulation of RNAfolding. The model contains 14 hairpin structures connected by unpaired nucleotides. Thedataalsopointtothreehelicesformedbydistantnucleotidesthatcloseoffstructuraldomains.Thismodelwas generally compatible with the results of in vivo probing experiments with dimethylsulfate in late exponential-phase cultures. Toe-printing experiments revealed that the ribosome binding site ofhld, which is encoded by RNAIII, was accessibletotheEscherichiacoli30Sribosomalsubunit,suggestingthattheinvitrostructurerepresentedatrans- latableformofRNAIII.Wealsofoundthat,withinthe39endofRNAIII,theconservedhairpin13andtheterminatorform anintrinsicstructuraldomainthatexertsspecificregulatoryactivityonproteinAgeneexpression. Keywords: agrlocus;chemicalandenzymaticprobing;invivoprobing;regulation;RNAIII INTRODUCTION toxin and enterotoxins), and represses the expression of exponential-phase surface proteins such as protein InStaphylococcusaureus, expression of several viru- A and coagulase+ RNAIII appears to be the specific lence factors is regulated by the global regulon agr+ regulator,asclonedRNAIII,defectiveindelta-hemolysin Theagrsystemiscomposedoftwodivergentoperons production, and expressed in an RNAIII-deletion mu- designated P2 and P3+ The P2 operon combines a tant,restoredboththepositiveandnegativeeffectson density-sensing cassette (agrD and B) and a two- theexpressionofexoproteinsandsurfaceproteins(Jan- componentsensorytransductionsystem(agrAandC), zon &Arvidson, 1990) and in anagr-null strain (Korn- both of which are required for autocatalytic activation blum et al+, 1990; Novick et al+, 1993)+ RNAIII acts ofthepromoterP2andfortranscriptionfromthediver- primarily on target gene transcription initiation and, in gentpromoterP3(Jietal+,1995;Linaetal+,1998)+The thecaseofalpha-hemolysin,independentlystimulates effector of the S.aureusagrsystem is a 514-nt tran- translation by a mechanism that appears to involve scriptoftheP3operon,designatedRNAIII,whichcon- direct interaction between the 59 end of RNAIII and a tains the 26-amino-acid delta-hemolysin gene (hld)+ stretchofapproximately80bpwithintheribosomebind- RNAIII stimulates the expression of postexponentially ingsiteofthehlatranscript(Morfeldtetal+,1995)+The expressed extracellular toxins and enzymes such as mechanism by which RNAIII controls the transcription alpha and beta-hemolysins, enzymes (lipases, prote- of its target genes is unknown+ One possibility is that asesandnucleases)andtoxins(toxicshocksyndrome RNAIII acts via trans-acting factors, as RNAIII is po- tentially capable of forming stable secondary struc- turesthatmaycreateproteinbindingsites(Novicketal+, Reprint requests to: Francois Vandenesch, EA1655, Faculté de 1993)+Theimportanceofthe59and39regionsofRNAIII MédecineLaennec,RueGuillaumeParadin,69372LyonCedex08, France;e-mail:denesch@univ-lyon1+fr+ for its regulatory function has been tested by deletion 668 StructureofS.aureusagr-RNAIII 669 analysisandbystudyingtheabilityofRNAIIIhomologs ionspriortomodification+WealsoverifiedthattheRNAs from non-aureus staphylococcal species to comple- foldedintoahomogenouspopulationbyusinggelelec- ment agr-deficient strains of S.aureus(Benito et al+, trophoresis in nondenaturing conditions+ A secondary 1998;Tegmarketal+,1998)+The39endofRNAIIImay structuremodelforRNAIIIwasderivedfromtheexper- beimportanttorepressproteinAexpression,andmay imental data with the help of a computer simulation of therefore be an intrinsic domain (Novick et al+, 1993; RNA folding (Gultyaev et al+, 1995)+ This procedure Benito et al+, 1998)+ As mentioned above, besides its folds the molecule by generating consecutive transi- regulatory function, RNAIII is also an mRNA that en- tions between favorable intermediate structures, that codes the delta-hemolysin gene+Translation of RNAIII is,bysimulatingfoldingpathwaysratherthanbysearch- intodelta-hemolysinwasdelayedby1hfollowingtran- ingfortheglobalminimumstate(Gultyaevetal+,1995; scription and this delay was abolished by deletion of Gerdesetal+,1997)+Thesecondarystructuremodelis the 39 half of RNAIII, suggesting that a conformational characterized by 14 hairpin structures and three heli- changeoftheRNAisrequiredfortranslation(Balaban ces (A–C) which involve distant nucleotides (Fig+ 2)+ & Novick, 1995)+ Secondary structure predictions indi- The model can be divided into three mains domains: catethatthe59and39endsofRNAIIImaypair,thereby the 59 end domain (nt 1–31), a central domain delimi- sequestering the ribosomal binding site ofhld(Novick tated by helix B (nt 32–382) and the 39 end domain et al+, 1993)+ Taken together, these data indicate that (nt 383–514)+ thestructureofRNAIIIplaysanessentialrolenotonly The59enddomaincontainsawell-characterizedstem- inhldtranslation but also in regulatory functions+ loopmotif+Thishairpinstructure1bearsaGNRAloop Here, we probed the secondary structure of S.au- thatisknowntoadoptaparticularstructureclosedbya reus RNAIII in vitro, using chemical and enzymatic shearedG-Abasepair(Heus&Pardi,1991;Pleyetal+, probes+ The results were compared with those of in 1994)+ The poor accessibility of G9 and A12 towards vivo probing of RNAIII produced during the late expo- chemicalsandenzymessupportstheexistenceofsuch nential phase+The proposed model of RNAIII second- a G-A base pair (Fig+ 1A,E)+ Two long-range inter- ary structure is characterized by 14 hairpin structures actionsinvolvingresiduesA24-G30/U452-458(helixC) andthreelong-rangeinteractionsthatbringthe39and and U32-A42/U371-G382 (helix B) can be proposed+ 59 ends of RNAIII into close proximity+ A number of Both helices are susceptible to RNase V1 hydrolysis hairpin structures appear to be conserved within the and several nucleotides are weakly or not reactive at Staphylococcusspecies+ One, located in the 39 end of one of their Watson–Crick positions+ However helix C S.aureusRNAIII, efficiently repressed protein A ex- presentsaweakstabilityasevidencedbytheconcom- pression in vivo+ The proposed model for the second- itantpresenceofseveralsingle-strand-specificRNase ary structure of RNAIII is discussed in the light of its cleavages (Fig+ 2)+ Thus, helix C brings in close prox- regulatory functions+ imitythe39and59domainswhereashelixBdelimitates the central domain+ Inthislargecentraldomain,thereactivitypatternsof RESULTS nucleotides in hairpin 2–11 were characteristic of hair- pin structures (Fig+ 2), as most of the nucleotides lo- Structural probing of catedintheloopswerereactiveattheirWatson–Crick StaphylococcusaureusRNAIII positions+ This was accompanied by single-strand- WeinvestigatedtheconformationofRNAIII(nt1–514) specific RNase cleavages (Figs+ 1 and 2)+ Further- with various chemical and enzymatic probes (Ehres- more, the stems were all cleaved by RNase V1, and mann et al+, 1987) (Fig+ 1)+ We chose an identical pH, most of the nucleotides were weakly or not reactive temperature, ionic strength, and divalent ion concen- towards chemicals+ This is particularly well-illustrated trationforthedifferentprobestominimizetheeffectsof for the stem-loop motifs 3, 4, 5, and 7–11 (Fig+ 2)+ He- theseparametersonconformationalchanges+Thefour lix2,whichisA-Urich,formsametastablehelicalstruc- basesweretestedatoneoftheirWatson–Crickpairing ture,whichmayaccountfortheconcomitantpresence positions by using DMS (N1A . N3C) and CMCT of RNase V1 and RNase T2 cleavages and for the (N3U . N1G)+ Single-stranded regions were probed reactivityofseveralnucleotidesatWatson–Crickposi- with RNases T2 and T1, and helical regions were de- tions (Fig+ 2)+ The long hairpin structure 6, character- tectedwiththedouble-strandedspecificRNaseV1+In- izedbyastemrichinA-Ubasepairsandinterruptedby formationonthe39and59endsofRNAIIIwasobtained an internal loop, required magnesium for stabilization: using enzymes, DMS (for N3C), and nickel complex most of the residues within the helix became strongly (which modifies N7G)+ Nickel complex is sensitive to reactive in the absence of magnesium (Fig+ 1B)+ The stacking of base rings; for example, the N7 position of central domain also contained a branched structure aguaninewithinahelixisnotreactiveunlessthedeep constitutedbyhairpins7–9andclosedbyhelixA(Fig+2)+ groove is widened (Chen et al+, 1993)+ The RNAmol- ThishelixinvolvingthetwodistantregionsG207–U216 ecules were renatured in the presence of magnesium andA310–U319 is supported by the presence of sev- 670 Y.Benitoetal. eral RNase V1 cleavages in both strands, and by the the helix became reactive under semidenaturing con- nonreactivity of several residues towards chemicals ditions (Fig+ 2D)+ The stem, rich in A-U base pairs, is (Figs+ 1 and 2)+ The single-stranded conformation of interruptedbytwobulgednucleotidesandclosedbya the interhelical regions is supported mainly by the re- 13-nt long loop+ Finally, hairpin 14, corresponding to activity of nucleotides at their Watson–Crick position the terminator, is rich in G-C base pairs and is one of (Figs+ 1 and 2)+ This was notably the case for nucleo- the most stable stem-loops of RNAIII+ Its existence is tides A42–U47, U65–A68, A105–A109, C138–C140, supported by the presence of RNAse V1 cuts on both A157–U163,A218–A220,U258–A260,U321–U327,and sides of the helix, and by the fact that none of the U350–C353+ cytosines at position N3 and the guanines at position The 39 end domain is constituted by three hairpins, N7 (located in the helix) were reactive towards chem- 12–14, and by a long single-stranded region involving icals(Fig+1C)+Inaddition,cytosines495–496and498– residues459–483(Fig+2)+Thestem-loopstructure13 500attheirN3positionswerehighlyreactivewithDMS, is clearly identified, but its folding required the pres- and several RNaseT2 cleavages occurred in the loop enceofmagnesium,becausemanynucleotideswithin (Fig+ 1C)+ FIGURE1. Probing the structure of RNAIII in vitro and in vivo+ Selected autoradiograms showing enzymatic (A) and chemical(B,C)probingperformedat378ConinvitrotranscribedRNAIII+A:RNaseT1(T1),RNaseT2(T2)andRNaseV1 (V1)hydrolysison59-labeledRNAIII,shortmigration+LanesI,I9:incubationcontrolsinbufferusedforRNaseT1,andRNase V1hydrolysis,respectively;lanesT1:0+05and0+15UofRNaseT1;lanesT2:0+02and0+04UofRNaseT2;lanesV1:0+1 and0+2UofRNaseV1;lanesT1andU2:RNasesT1andU2indenaturingconditions,respectively;laneL:alkalineladder+ B: DMS and CMCT modifications, long migration+ Lane I: incubation control in buffer used for DMS; lane DMS: native conditions,1mLofDMSdiluted1:16;lanesCMCT:nativeconditions(N)andsemidenaturingconditions(SD),2mLofCMCT (100mg/mL);lanesG,A,T,andCcorrespondtosequencingladders+C:DMSmodificationofcytosinesatN3on39-end- labeledRNAIII,shortmigration+LaneI:incubationcontrol;lanesDMS:nativeconditions,1,2,and4mLofDMSdiluted1:8; lanesU2andLasinA+DandE:selectedautoradiogramsshowingDMSandCMCTmodificationsoninvitrotranscribed RNAIII(asinB)andacomparisonwithDMSmodificationonRNAIIIinvivo,shortmigration+Thedifferencesinreactivities aredenotedbyarrows+LanesI:incubationcontrolsinbufferusedforDMS;lanesDMSinvivo:25mLofDMSin20mLof staphylococcalcultureatlateexponentialgrowthphase;lanesDMSinvitro:nativeconditions(N),semi-denaturingcondi- tions(SD),1mLofDMSdiluted1:16;lanesCMCT:nativeconditions(N),semi-denaturingconditions(SD),1mLofCMCT (100mg/mL);lanesG,A,T,andCcorrespondtosequencingladders+(Figurecontinuesonfacingpage.) StructureofS.aureusagr-RNAIII 671 Probing the structure of the 39 end of RNAIII main alone, all these nucleotides were reactive at one of their Watson–Crick positions and were cleaved by Becauseithadbeensuggestedthatthe59and39ends RNaseT2(Fig+3A),indicatingthatthisregionissingle of RNAIII would pair, and because specific regulatory stranded+ activitywasattributedtothe39end(Novicketal+,1993; Benito et al+, 1998), the conformation of this region (nt394–490)wasprobedseparatelywithchemicaland In vivo DMS probing of RNAIII enzymatic probes (Fig+ 3)+ Interestingly, the enzymatic cleavageandchemicalreactivitypatternsofnt409–483 The structure of RNAIII was studied in vivo by using in the 39 domain were identical to those in the entire DMS, which alkylates cytosine at N3 and adenine at RNAIIImoleculeandarethusconsistentwiththeexis- N1 (Mayford & Weisblum, 1989)+ In a typical experi- tenceofstem-loopstructure13(Figs+2and3)+Indeed, ment,S.aureusstrain RN6390 was grown to the late RNase V1 cuts occurred at positions 410–418, 421– exponentialphaseandtreatedwithDMS+SitesofRNAIII 422, 439, 443–444, and 447–450, whereas the main modificationweredetectedbytheprimerextensionre- RNase T2 cleavage sites were located in the hairpin action on RNAextracts and were compared to modifi- loop at positions 423–430 and 434–437 (Fig+ 3)+ Fur- cations observed in vitro+ As shown in Figure 1D,E, thermore, the nucleotides in the external loop were all reactivitywithRNAIIIinvivoandinvitrowasessentially highly reactive with chemical probes at their Watson– identical, as only nucleotides in the hairpin loop and Crickposition(Fig+3)+ThepresenceofseveralRNase interhelical regions were highly sensitive to DMS+ Dif- V1 cuts in the upper part of the stem (A419–U424/ ferencesbetweentheinvitroandinvivomappingdata A438–U443) and the reactivity of several uridines mainly concerned the reactivity of the adenine resi- pointed to the formation of a metastable helix+ duesinvolvedintheproposedlong-rangeinteractions, The main reactivity changes observed between the particularlyhelixB(U371–G382/U32–A42)+Thesead- entire RNAIII molecule and the 39 end domain alone enines were more sensitive to DMS alkylation at their were located at nt U452–U466+ Indeed, in the 39 do- N1 position in vivo than in vitro (Fig+ 1D)+Another sig- FIGURE1. (continued.) 672 Y.Benitoetal. FIGURE2. Secondary structure model of RNAIII showing the enzymatic cleavages (A) and reactivity of nucleotides to chemicalsprobes(B)+Thehairpinstructuresarenumbered1to14;longdistanceinteractionsaredesignatedA,BandC; theShine–Dalgarnosequenceandthestartandstopcodonsofhldareinboldface+A:RNaseT1cleavagesaredenoted by( ),RNaseT2by( ),thick,medium,orthinarrowsforstrong,moderateorlowreactivity,respectively;RNaseV1 cleavagesaredenotedbyblack,gray,oremptytrianglesforstrong,moderate,orlowreactivity,respectively+B:Reactivity of Watson–Crick positions to DMS (N1A, N3C) and CMCT (N3U, N1G) are represented as follows: reactive in native conditions(darkenedblackorgreycirclesforstrong,moderateorlowreactivity,respectively);reactiveonlyinsemidena- turing conditions (squares); more reactive in semidenaturing conditions (cross); nonreactive in both native and semi- denaturing conditions (no symbol)+ Dots indicate not determined due to unspecific cleavages or pauses of RT in the incubation control+ Guanines reactive at position N7 to NiCr are denoted by black squares, and empty squares denote unreactive guanines+ Information about guanine accessibility at N7 position is only provided for region 410–512+ The differencesbetweenthereactivityofadenineatN1andcytosinesatN3toDMSinvitroandinvivoareindicatedasfollows: increasedreactivityinvivobyanasteriskandprotectioninvivobyblacktriangles+Thesedatawereobtainedfromfour independentexperiments+ nificantdifferencebetweeninvivoandinvitromapping only 1 h later (Balaban & Novick, 1995)+ This delay is resultsconcernedseveraladenineresidueslocatedat not dependent on a specific timing signal during the the ribosome initiation site+ Indeed, adenines 58, 71, growthcycle,andisabolishedbydeletionofthe39end and 72 were clearly nonreactive in vivo, whereas ad- of RNAIII (Balaban & Novick, 1995)+ We therefore enines 58 and 57 became more reactive at their N1 probed the accessibility of the hldribosome binding positions in vivo (Figs+ 1E, 2)+ site in vitro by using a toe-printing assay (Hartz et al+, 1988)+Inthisprimerextensioninhibitionassay,forma- tionoftheternarycomplexcomposedofthe30Ssub- unit, RNAIII, and the initiator tRNAMet blocks the Ribosomes bind toS.aureusRNAIII elongation of a cDNA primer by reverse transcriptase Although RNAIII is transcribed in the mid-exponential (Hartzetal+,1988)+Indeed,the30S/tRNAMet complex f phase of growth, it is translated into delta-hemolysin incubated with the refolded RNAIII produced a strong StructureofS.aureusagr-RNAIII 673 FIGURE3. Probing the secondary structure of in vitro-transcribed RNAIII 39 domain+ A: autoradiogram showing enzymatic cleavage sitesonthe39domain+RNaseV1(V1)andRNaseT1(T1)hydro- lysis,shortmigration+LanesI,I9:incubationcontrols;lanesT1:0+05 and0+15UofRNaseT1;lanesV1:0+1and0+2UofRNaseV1;lanes G,A,T,Careforsequencingladders+B:secondarystructuremodel ofthe39enddomainwiththeenzymaticcleavagesandthechemical reactivities to DMS and CMCT+ Symbols as in Figure 2+ The data correspondtothreeindependentexperiments+ toe-print at A100, 15 nt downstream of the initiation ribosomes bind efficiently to hldRBS and that the in codon(Fig+4),aspreviouslyshown(Hartzetal+,1988)+ vitrostructurerepresentsatranslatableformofRNAIII+ The yield of the toe-print was identical when RNAIII was denatured prior to the experiment (Fig+ 4)+ Given the possibility that the primer used for reverse tran- The 39 domain of RNAIII represses scriptionperturbedthestructureofRNAIII,formationof proteinAgene expression theternary30S/RNAIII/tRNAMetcomplexwasalsostud- iedinfilterbindingassays(notshown)+Againthedata We have shown here that the 39 end of RNAIII is an showed no major differences between native and de- independent structural domain+ Furthermore, we and natured RNAIII+ Altogether, these results indicate that others have previously shown that the 39 end of the 674 Y.Benitoetal. structure of RNAIII has been proposed (Novick et al+, 1993)+This model was based on minimum energy so- lutionsandwascharacterizedbyextensivelong-range interactions between the 59 and 39 domains (Novick et al+, 1993)+ This prompted us to analyze the confor- mationofRNAIIIinvitrobyusingseveralchemicaland enzymatic probes, and in vivo by using DMS+ These experimental data formed the basis for a secondary structuremodelderivedwiththehelpofcomputersim- ulation of RNAIII folding (Gultyaev et al+, 1995)+ The presentsecondarystructureofRNAIIIischaracterized byanumberofhairpinstructures(Fig+2),mostofwhich hadbeenpredicted(Novicketal+,1993;Tegmarketal+, 1998)+Wealsobringevidencethatmostofthesestem- loopstructuresoccurredinvivo+ThealignmentofRNAIII fromS.aureuswith that of non-aureusstaphylococcal FIGURE4. Accessibility of the ribosome binding site in RNAIII in species (S.epidermidis, S.warneri, S.simulans, and vitro+Theexperimentswereconductedinparallelonnativeandde- S.lugdunensis)revealsthatthestem-loopmotifs1,7, natured (unfolded) RNAIII molecules+ The position of the toeprint (A100) is indicated+ Lane c: incubation control of free RNAin the 13, and 14 are predicted to be highly conserved by presenceof30Ssubunit;lanes30S:experimentinthepresenceof computer folding, and the hairpins 6, 8, 9, 11, and 12 increasingconcentrationsof30Ssubunits(125,250and625nM); must be present in all RNAIII but their sequences di- lanesG,A,TandCareforsequencingladders+ verged(Vandeneschetal+,1993;Tegmarketal+,1998)+ Interestingly,severalbasepaircompensatorychanges are found in helices 1, 7, 13, and 14 amongStaphylo- coccusspecies(Fig+6)+Thusthehighconservationof RNAIIIisimportantforrepressingproteinAgene(spa) hairpins1,7,13,and14suggestsacommonfunction+ expression,andmaythusbeafunctionaldomain(No- Forinstance,withtheexceptionofS.epidermidisRNAIII, vick et al+, 1993; Benito et al+, 1998)+ To further inves- which carries a 5-nt-long loop (Fig+ 6), hairpin 1 con- tigate this question, we subcloned selected fragments tains a conserved GAGA loop motif+ It is well known of RNAIII under the control of the RNAIII promoter on thatstablestem-loopstructureslocatedatthe59endof the staphylococcal plasmid pE194 and introduced the resulting constructs in S.aureusstrain WA400 (agr2, mRNAmayincreaseitsstability(forareview,seeBech- hofer, 1993)+ It is therefore tempting to propose a 59 lackingRNAIII)+Downregulationofspaexpressionwas stabilizer function for this conserved hairpin motif in achieved normally in WA400/pE194 expressing either the RNAIII fragment comprising residues 391–514 (Fig+5,lane4)orresidues403–455and484–514(both RNAfragments contain hairpin 13 and the terminator) (Fig+ 5, lane 5)+ Interestingly, WA400/pE194 express- ing a fragment of RNAIII that contains only the tran- scriptionterminator(nt484–514)wasstillabletorepress spaexpression,albeittoalesserextent(Fig+5,lane3)+ DISCUSSION The secondary structure model ofS.aureusRNAIII S.aureusRNAIII regulates the expression of several genes encoding exoproteins and cell-wall-associated FIGURE5. EffectoftruncatedRNAIIIonproteinAgenetranscrip- proteins(Novicketal+,1993)+RNAIIIalsoencodesdelta- tion+NorthernblotanalysisofRNAIIIandproteinAmRNAexpres- hemolysin, but inactivation of the open reading frame sion+ RNAs from postexponential phase cultures were hybridized does not affect the regulatory function of RNAIII (Jan- withprobescorrespondingtoRNAIIIandproteinAgene(spa)+Lane1: RN6390(agr1);lane2:WA400(agrmutant);lane3:WA400/pLUG310 zon & Arvidson, 1990; Novick et al+, 1993)+ There is (terminatortranscriptionalone,nt484–514);lane4:WA400/pLUG300 alsoevidencethatRNAIIImayundergoconformational (39endRNAIII,nt391–514);lane5:WA400/pLUG315(hairpinstruc- changes that activate either its translation or its regu- ture13transcriptionterminator,nt403–455::484–514)+RNAIIItran- scripts are not seen in lanes 4 and 5 because of their very small latoryfunctions(Novicketal+,1993;Balaban&Novick, sizes+Positionsofmigrationof16SrRNA(1541nt)and23SrRNA 1995)+ A computer-simulated model of the secondary (2904nt)indicated+ StructureofS.aureusagr-RNAIII 675 FIGURE6. SequenceofseveralconservedhairpinstructuresofRNAIIIfromS.aureus(typesI–IV),andcomparisonwith four different species,S.epidermidis(e),S.warneri(w),S.simulans(s), andS.lugdunensis(l)+ Nucleotide insertion is shown by an open arrow, and D is for deletion+The hairpin structures have been assigned according to the secondary structuremodelpresentedinFigure2+ RNAIII+ The central domain (nt 32–382) is closed by proposed model of RNAIII, the two long-range inter- helix B (Fig+ 2)+Analogous long-range interaction may actions B and C hide part of the hla-mRNA binding also arise in S.epidermidis, S.simulans, and S.war- sequence, indicating that conformational changes of neriassuggestedbycomputerfolding,despitethefact RNAIII are required for efficient binding+ Interestingly, that the sequence is not conserved (data not shown)+ the greater accessibility of adenines 372–378 in vivo The central domain encloses the open reading frame suggeststhathelixBisdestabilized,likelyastheresult encodingfordelta-hemolysinthatispresentinallRNAIII, ofhlamRNAbindingtothe59regionofRNAIII,explain- exceptS.lugdunensis,withahighdegreeofsequence ing the apparent discrepancy with the in vitro data+ identity(Vandeneschetal+,1993;Tegmarketal+,1998)+ Different pairing pathways that result in the formation Thecentraldomainalsocontainsabranchedstructure of stable antisense RNA and mRNA complexes have constituted by three hairpins, 7–9, closed by a long- been described (for a review, see Zeiler & Simons, range interaction (helix A)+ The existence of helix A, 1998)+ In most cases, binding initiates between two which was clearly demonstrated in vitro, was also loops, or between a loop and a single-stranded RNA supported by the in vivo probing+ In this branched segment involving a restricted number of nucleotides, structure, hairpin 7 is of particular interest+ Several and more stable complexes are subsequently formed base pair compensatory changes are found among by pairing of distal RNAsegments+ In region 42–72 of Staphylococcusspecies, and the loop contains a con- RNAIII, several nucleotides show high accessibility to served sequence (UCCCAA) that resembles that of single-strand-specific probes, indicating that they are theloopsofhairpin13andtheterminator(Fig+6)+This potential sites for initial pairings+Afragment of RNAIII highconservationindicatesthathairpin7isapotential containingnt1–380,whichpotentiallyformshelixB,is recognition site fortrans-acting factor(s)+ indeedcompetentforbindingtohlamRNAinvitro(Mor- feldt et al+, 1995)+ The central domain also contains the ribosomal ini- The central domain may undergo tiationsiterequiredforthetranslationofdelta-hemolysin, specific conformational changes which is the only known translation product of RNAIII+ Thus,thetwofunctionsofRNAIII(translationofhldand The 59 side of the central domain of RNAIII contains translational regulation of hla) are mutually exclusive+ theregulatoryregion(nt19–87)involvedintheactiva- Structural probing of RNAIII in vitro revealed that the tionofalpha-toxintranslation(Novicketal+,1993;Mor- RBSsequenceandtheAUGinitiationcodonofhldare feldtetal+,1995)+BindingofRNAIIItotheleaderregion both accessible to single-strand-specific probes+ The ofalpha-toxinmRNAinducedaconformationalchange toe-printing and filter binding assays showed that the ofthehlamRNA,renderingtheribosomalinitiationsite ribosome efficiently binds both to native RNAIII and to available for translation (Morfeldt et al+, 1995)+ In the thedenaturedmoleculeatthehldinitiationsite(Fig+4)+ 676 Y.Benitoetal. Footprinting experiments indicated that the initiating showed an almost identical nucleotide reactivity pat- ribosome protects ;40–50 nt (Murakawa & Nierlich, tern, a finding consistent with the existence of a stem- 1989; Huttenhofer & Noller, 1994)+ As the 39 side oc- loop motif, as further supported by the in vivo probing curredatposition115fromtheadenineoftheinitiation (Fig+ 3)+ The stem of hairpin 13 is interrupted by two codon (as determined by the toe-printing approach), bulged residues (U417 andA418)+ Despite the appar- weassumethattheribosomemaycoverpartofstem- entlylowerstabilityoftheupperstem,theprobingdata loop structure 2 from the 59 side+ Many groups have pointtotheexistenceofa13-nt-longloop(Fig+2)+More- investigatedtheeffectsofRNAstructureontranslation over,this39domainexertedspecificregulatoryactivity, initiation (e+g+, for a review, see de Smit & van Duin, as the last 124 nt cloned downstream of the P3 pro- 1994)+However,chemicalprobingindicatedthatstem- moter in an RNAIII-deficient strain were sufficient to loopstructure2ismetastable,andtheribosomeisthus repress spaexpression (Fig+ 6)+ It is remarkable that able to bind to the unfolded form+ It was also shown onlyhairpin13(nt403–455)andtheterminatorhairpin that a 5-bp hairpin located 3 nt upstream of the SD (nt 484–514) were required for efficientspatranscrip- sequenceoftheT4rIIBgenecanbeincreasedto9bp tioncontrol,asdeletionoftheconservedresiduesU456 without having an inhibitory effect (Shinedling et al+, to U489 did not affect regulation (Fig+ 2)+ It has previ- 1987)+Therefore,ourdataindicatethattheinvitrostruc- ouslybeenshownthatRNAIIIfromS.lugdunensisdoes tureofRNAIIIcorrespondstoitstranslatableform+One not repress spa transcription, while a chimeric con- intriguing finding is that translation ofhldin vivo is de- structconsistingofthe59endofS.lugdunensisRNAIII layedby1hfollowingtranscription(Balaban&Novick, and the 39 end ofS.aureusRNAIII repressedspaex- 1995)+ It was postulated that a conformational change pression more efficiently in a S.aureusbackground ofRNAIIItoatranslatableformprecededitsregulation (Benitoetal+,1998)+Severalsequencedifferenceswere of target genes (Balaban & Novick, 1995)+This appar- observedbetweenS.lugdunensisandS.aureusRNAIII ently conflicts with the fact that the in vitro structure is in the hairpin loop of the terminator and stem-loop 13 efficiently recognized by the ribosome+ It is possible (Fig+ 6)+ Conversely, other coagulase-negative staph- that, owing to transcription–translation coupling, some ylococci in which the last 150 nt of RNAIII (including transientinhibitorystructuresoccurringinvivoseques- stem-loop13,Fig+6)aremorestronglyconservedwere ter the ribosome binding site or influence the kinetic able to repressspaexpression (Tegmark et al+, 1998)+ foldingpathwayofRNAIIIinvivo+Interestingly,theover- These data point to a shared function of the 39 end all pattern of nucleotide reactivity to chemical probes domainofseveralRNAIIIspeciesinsparegulation+We was very similar in vivo and in vitro+ One major differ- also found that the last 31 nt of RNAIII (containing the encewaslocatedwithintheRBS:inparticular,thetwo transcriptionterminator)partiallyrepressedspaexpres- adenines(71and72)oftheShine–Dalgarnosequence sion,suggestingthattheterminatoralsoparticipatesin becameprotectedinvivo(Fig+2),aphenomenonwhich theregulatorymechanism+Whetherthe39domainreg- may correlate with occupancy by the ribosome of its ulates transcription directly or indirectly is unknown+ initiationhldbinding site+ Indeed, in the initiation com- One plausible explanation is that hairpin 13 and the plex,base-specificprotectionmainlyoccursintheRBS terminator of RNAIII act via a ribonucleoprotein com- and at the initiation codon, reflecting the formation of plex+Apossible role of RNAin modulating the activity stable base pairings with the 39 end of 16S rRNAand of DNAbinding proteins has been reported (Retallack the anticodon sequence of the tRNAMet, respectively & Friedman, 1995; Sledjeski & Gottesman, 1995)+ It is f (Huttenhofer & Noller, 1994)+ However, in vivo probing also noteworthy that both hairpin loops contain a con- of RNAIII took place in the late exponential phase of served CCCAAsequence that may provide a bipartite growth when RNAIII is also involved in transcriptional protein binding site (Fig+ 6)+ regulation of target genes+ It is likely that RNAIII un- In summary, RNAIII has the unique property of act- dergoesspecificconformationalchangesthatarenec- ing as a messenger RNAand having pleiotropic regu- essaryfordefinedfunctions+Thus,severalpopulations latory functions+ The secondary structure model of of RNAIII may coexist in vivo but are difficult to distin- RNAIII proposed here should help to clarify functional guish by chemical mapping+ domains+Wedemonstratedherethatthe39domainof RNAIII contains an independent structure and has a regulatory function in proteinAsynthesis+ The 39 end domain is sufficient for the control of proteinAexpression MATERIALSAND METHODS We also identified an intrinsic structural domain in the 39 end of RNAIII, with hairpin 13 at its center+ Stem- Bacterial strains and growth conditions loop13hasbeenpredictedtobepartiallybasepaired withtheregionencompassingnt130–98(Novicketal+, Escherichiacoli TG1 and S.aureus RN4220, a nitroso- 1993)+ However, structural mapping of the full-length guanidine-inducedmutantcapableofacceptingE.coliDNA RNAIII and of the RNAIII domain from nt 391–490 (Kreiswirthetal+,1983),wereusedforplasmidamplification StructureofS.aureusagr-RNAIII 677 and genetic manipulations+S.aureusRN6390 derives from Bam1081andagr-sa757);thesetworegionswereseparated 8325-4andisourstandardagr1strain+S.aureusWA400is by a short cloning sequence that contained the restriction aderivativeof8325-4inwhichtheP2operonisfunctionalbut sitesStuIandBamHI+Toexpresshairpinstructure13under theP3operonisdeletedandreplacedbythechloramphen- thecontroloftheP3promoter,thecorrespondingsequence icoltransacetylasegene(cat86)(Janzon&Arvidson,1990) (nt 1168–1116) was amplified using primers Stu1168 and (seeTable 1)+ Staphylococci were grown either on BM agar Bam1116, digested by StuI and BamHI, and ligated to plates(1%peptone,0+5%yeastextract,0+1%glucose,0+5% pLUG310digestedbythesameenzymes+ NaCl, 0+1% K2HPO4), or in brain-heart infusion (BHI), with Forinvitrotranscription,theRNAIIIcodingsequence(cor- erythromycin(5mgmL21)whenappropriate+ respondingtont1–514ofRNAIII)wasamplifiedusingprim- ersStu1572andBam1055+ThePCRproductwasdigested byStuIandBamHIandligatedtothepUT7vector(Serganov Molecular cloning and plasmid construction etal+,1997)cleavedwiththesamerestrictionenzymes+The Total DNA and plasmid DNA were prepared with standard resultingplasmidwasintroducedbytransformationinE.coli methods (Sambrook et al+, 1989)+ Transformation of E.coli TG1+ Similarly, the 39 end of RNAIII (nt 391–490) lacking TG1wasperformedwithcellsrenderedcompetentbytreat- the transcription terminator was amplified by using primers ment with CaCl2, andS.aureusRN4220 and WA400 were Stu1180andBam1079andwasthenclonedontopUT7vec- transformedbyelectroporation(Bio-Radgenepulser)+ torinE.coliTG1+ RNAIIIwasexpressedinStaphylococcusbyusingplasmid pE194(Horinouchi&Weisblum,1982),whichhadbeenmod- Northern blots ifiedbyaddinganEcoRVrestrictionsitenexttotheXbaIsite formingpLUG274+PCRwasperformedfromtheRN6390agr Electrophoresis of total RNA through 1% agarose gel con- sequence (GenBank accession number X52543)+ To clone taining2+2Mformaldehyde,vacuumtransfertonylonmem- the39domainofRNAIIIundercontroloftheP3promoter,the brane, hybridizations with specific digoxigenin-labeled RNA promotersequencewasamplifiedusingprimersagr-sa1819/ probes and luminescent detection were carried out as de- Stu1569 (Table 2), and the 39 end of RNAIII was amplified scribed(Benitoetal+,1998)+PCRprimersfortheproduction using primers agr-sa757/Stu1180+ The PCR products were of agrRNAIII and protein A-specific gene probes were se- digested byStuI and ligated together before being reampli- lectedfromGenBankaccessionnumberX52543andA04518 fied with the external primers agr-sa1819 and agr-sa757+ respectively+ The resulting PCR product was digested with HindII and XbaI (site present on each primer) and ligated to the modi- Preparation of RNA fiedpE194vectordigestedbyEcoRVandXbaI+Theligation product was electroporated to RN4220 and the resulting RNAIIIandthe39domainofRNAIIIweretranscribedinvitro plasmid (pLUG300) was introduced into strain WA400+ To with T7 RNA polymerase from BamHI linearized pLUG322 express intramolecular sequences of RNAIII under the con- andpLUG281,respectively+TranscriptionofRNAIIIyieldeda troloftheP3promoter,weconstructed,usingasimilarmethod, run-offproductof522ntincludingfiveadditionalnucleotides plasmidpLUG310thatcontainedtheP3promoterregionlinked atthe59endandthreeatthe39end,whichallcamefromthe to the transcriptional terminator (amplified using primers vector+The39domaincorrespondedtont391–490ofRNAIII+ TABLE1+ Maincharacteristicsofstrainsandplasmids+ Strainsorplasmids Relevantcharacteristics Referenceorsource S.aureusstrains 8325-4 NCTC8325curedofthreeprophages Novick,1963 RN4220 restriction-mutantof8325-4 Kreiswirthetal+,1983 RN6390 derivativeof8325-4,agrpositive Pengetal+,1988 WA400 8325-4DRNAIIIregion::cat86 Janzon&Arvidson,1990 LUG450 WA400/pLUG300 thiswork LUG460 WA400/pLUG310 thiswork LUG467 WA400/pLUG315 thiswork E.coliplasmids pUT7 pUC119::T7promoter Serganovetal+,1997 PLUG322 pUT7::RNAIIInt1–514 thiswork pLUG281 pUT7::RNAIIInt391–490 thiswork Staphylococcalplasmids pE194 3+728kbS.aureusplasmid,inducible Horinouchi&Weisblum,1982 MLSresistance(erm) pLUG300 pE194::P3promoterlinkto39-endRNAIII thiswork (nt391–514) PLUG310 pE194::P3promoter-TT(nt484–514) thiswork pLUG315 pLUG310::hairpin13(nt403–455) thiswork

Description:
ABSTRACT. RNAIII, a 514-nt RNA molecule, regulates the expression of many Staphylococcus aureus genes encoding exoproteins.
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.