Positive
cooperativity
between
acceptor
and
donor
sites
of
the
peptidoglycan
glycosyltransferase
Daniel
Bury
a,1,2,*
,
Ismahene
Dahmane
b,2,
Adeline
Derouaux
b,
Shrinivas
Dumbre
c,
Piet
Herdewijn
c,
Andre´ Matagne
b,
Eefjan
Breukink
d,
Erika
Mueller-Seitz
a,
Michael
Petz
a,
Mohammed
Terrak
b,*
a
DepartmentofFoodChemistry,FacultyofMathematicsandNaturalSciences,UniversityofWuppertal,Gaussstr.20,42119Wuppertal,Germany
b
Centred’Inge´nieriedesProte´ines,Universite´ deLie`ge,Alle´edelaChimie,B6a,B-4000,SartTilman,Lie`ge,Belgium
c
LaboratoryofMedicinalChemistry,RegaInstituteforMedicalResearch,UniversityofLeuven,Leuven,Belgium
d
MembraneBiochemistryandBiophysics,DepartmentofChemistry,FacultyofScience,UtrechtUniversity,Padualaan8,3584CHUtrecht,TheNetherlands
1. Introduction
Theglycosyltransferasesoffamily51(GT51)areessentialenzymes foundinbacteriawithpeptidoglycancellwall[1].Theyexistintwo forms:asamonofunctionaldomainorlinkedtotheN-terminalendof penicillin-binding(PB) domain inbifunctional PB proteins(PBPs)
[1].BothformscatalyzethepolymerizationoflipidII(undecaprenyl pyrophosphate-MurNAc(pentapeptide)-GlcNAc) precursor to form
linearglycanchainswhich,aftercrosslinkingbythetranspeptidases, formthenet-likepeptidoglycanmacromoleculethatencasesbacteria and protects them from rupture under their high cytoplasmic pressure. Inhibitionofthe GTblocks peptidoglycansynthesisand leadstobacteriallysisanddeath,thereforetheGTisapromising targetinthesearchfornewantibacterialagentstocounterthearising antibioticresistantstrains.
Two main strategies are being followed to identify novel pharmacophoreleadmoleculesforthedesignandsynthesisofnew GT inhibitors that could be developed as antibiotics: high throughput screening of chemical libraries [2–4] and rational synthesisofsubstrateandmoenomycin(GTinhibitor)analogs[5– 8].ThesecondstrategyrequiresadeepunderstandingoftheGT mechanism and structural data on protein–ligand complexes. SubstantialprogresshasbeenmadeinthebiochemistryoftheGT
[9–14]andseveralcrystalstructureshavebeendetermined,both
ARTICLE INFO Articlehistory: Received24September2014 Accepted7November2014 Availableonlinexxx Keywords: Glycosyltransferase Peptidoglycan Moenomycin SPR LipidII ABSTRACT
Theglycosyltransferasesoffamily51(GT51)catalyzethepolymerizationoflipidIItoformlinearglycan chains,which,aftercrosslinkingbythetranspeptidases,formthenet-likepeptidoglycan macromole-cule.TheessentialfunctionoftheGTmakesitanattractiveantimicrobialtarget;thereforeabetter understandingofitsfunctionanditsmechanismofinteractionwithsubstratescouldhelpinthedesign andthedevelopmentofnewantibiotics.
Inthiswork,wehaveusedasurfaceplasmonresonanceBiacore1
biosensor,basedonanamine derivativeofmoenomycinAimmobilizedonasensorchipsurface,toinvestigatethemechanismof binding of substrate analogous inhibitors to the GT. Addition of increasing concentrations of moenomycinAtotheStaphylococcusaureusMtgAledtoreducedbindingoftheproteintothesensor chipasexpected.Remarkably,inthepresenceoflowconcentrationsofthemostactivedisaccharide inhibitors, binding of MtgA to immobilized moenomycin A was found to increase; in contrast competitionwithmoenomycinAoccurredonlyathighconcentrations.Thisfindingsuggeststhatatlow concentrations, thelipid IIanalogs bind to theacceptor site andinduce acooperative bindingof moenomycinAtothedonorsite.Ourresultsconstitutethefirstindicationoftheexistenceofapositive cooperativitybetweentheacceptorandthedonorsitesofpeptidoglycanGTs.
Inaddition,ourstudyindicatesthatamodificationoftworesidues(L119NandF120S)withinthe hydrophobicregionofMtgAcanyieldmonodisperseformsoftheproteinwithapparentlynochangein itssecondarystructurecontent,butthisisattheexpenseoftheenzymefunction.
ß2014ElsevierInc.Allrightsreserved.
* Correspondingauthors.
E-mailaddresses:bury@ipa-dguv.de(D.Bury),mterrak@ulg.ac.be(M.Terrak).
1
Presentaddress:InstituteforPreventionandOccupationalMedicine ofthe GermanSocialAccidentInsurance,InstituteoftheRuhr-Universita¨tBochum(IPA), Bu¨rkle-de-la-Camp-Platz1,44789Bochum,Germany.Tel.:+492343024796, fax:+492343024605.
2
D.BandI.Dcontributedequallytothiswork.
ContentslistsavailableatScienceDirect
Biochemical
Pharmacology
j our na l ho me p a ge : w ww . e l se v i e r . com / l oc a te / b i och e mph a rm
http://dx.doi.org/10.1016/j.bcp.2014.11.003
intheapoformandincomplexwithmoenomycinAandalipidII analog[15–20].ThesestructuresshowthattheGTiscomposedofa lysozyme-likeglobulardomainandasmallhydrophobicdomain (calledjawdomain)withadeepcleftbetweenthemcontainingthe activesite(Fig.1). Theextendedactive siteis dividedintotwo subsites,separatedbyamobileregion(composedofanalphahelix and a beta-hairpin), a donor site for the growing glycan chain bindingandanacceptorsiteforlipidIIbinding.Elongationofthe growing chainis achieved by theaddition ofdisaccharide unit MurNAc-GlcNAcoflipidIIinaprocessiveway.Comparisonofthe apoand moenomycin Abound structuresshows that thehead domainremainsunchangedaftermoenomycinAbindingwhereas amajorconformationalchangeoccursinthejawsubdomainwith partialrestructurationofthemobileregion[15].
Attheinitiationphaseofpolymerization,thelipidIIsubstrate molecules bind to the donor and acceptor sites which makes bindingstudiestodeterminetheaffinityofthesubstrateforeach siteparticularlycomplex.Differencesinaffinitywoulddetermine theorderofbindingintheinitiationstageandmightconsequently affecttheaffinityforthesecondsitebyanallostericeffect.Wehave used lipid II mimic inhibitors and moenomycin A, which is proposed to mimic the growing chain, to investigate these questions.AnassaywasdevelopedbyamodificationoftheSPR methoddescribedbyWelzelandcolleagues[21].Thistechniqueis basedontheinteractionofanimmobilizedmoenomycinAamino derivativewithpeptidoglycanglycosyltransferaseandwas origi-nally used to test moenomycin derivatives for their inhibitory activity.Weimprovedthesensitivityofthisassayanduseditto characterizeseverallipidIIanaloginhibitorsaswellas enzymati-cally inactive GT mutants as to their donor site functionality. MoenomycinAnecessaryforthesynthesisoftheaminoderivative isnotcommerciallyavailable.Hencewealsodevelopedamethod to isolate moenomycin A in good purity and yields from a Flavomycin1
standard containing a mixture of moenomycin congeners.
WeshowthatinthepresenceoflowconcentrationsoflipidII analogs,supposedlyboundto theacceptorsite, moenomycinA binding to the donor site of the GT increased, indicating an allostericactivationofthedonorsite.Athigherconcentrationsof theseanalogs,lowerGTbindinglevelswereobserved,indicating competitionwithmoenomycinAatthedonorsite.
AllGTscontainaconservedhydrophobicsurface(Fig.1)that mediatestheir interactionwiththecytoplasmicmembrane and rendersthepurifiedproteinspolydisperse[18,22].Thisproperty
makesathoroughquantitativebindingstudyoftheMtgAbySPR complex.Thepossibilitytoimprovethesolubility (monodisper-sity)oftheproteinbymodificationofresiduesinthehydrophobic surfacewasinvestigated.Amonodisperseformoftheproteinwas obtainedwithoutdetergentandcharacterized.
2. Materialsandmethods
2.1. Materials
Surface plasmon resonance (SPR) biosensor analyses were performedonaBiacoreQ1
(BiacoreAB,Uppsala,Sweden). Semi-preparativeHPLC(highperformanceliquidchromatography)was performedusinganHPLCconsistingofaMerck-HitachiLaChrom1
PumpL7100(HitachiLtd.,Tokyo,Japan),amanualsampleinjector valvemodel7125(Rheodyne Inc.,Cotati,USA),aNUCLEODUR1
C18HTeccolumn(dimensions:25010mm,particlesize:5
m
m) (Macherey-Nagel GmbH & Co. KG, Du¨ren, Germany) in a thermostatedcolumncompartmentmodelERC125(ERCGmbH, Riemerling,Germany),aMerck-HitachiLaChrom1UV–vis Detec-torL-7420withsemi-microflowcell(HitachiLtd.,Tokyo,Japan), and a 162 chromatography signal interface (Autochrom Inc., Milford, USA). Analytical HPLC was performed using an HPLC consistingofaMerck-HitachiLaChrom1
PumpL7100(HitachiLtd., Tokyo,Japan), a Merck-Hitachi LaChrom1
Programmable Auto-samplerL-7250(HitachiLtd.,Tokyo,Japan),aNUCLEODUR1
C18 HTec column (dimensions: 2504mm, particle size: 5
m
m) (Macherey-NagelGmbH&Co.KG,Du¨ren,Germany) ina Merck LaChrom1Column Oven L-7350 (Merck KGaA, Darmstadt, Germany), a Merck-Hitachi LaChrom1
Diode ArrayDetector L-7450(HitachiLtd.Tokyo,Japan),andaMerck-HitachiInterface D-7000(HitachiLtd.,Tokyo,Japan).HPLCeluentsandSPRrunning bufferswerefilteredthroughaWhatman1
OE67celluloseacetate membrane filter (poresize 0.45
m
m; Whatman GmbH, Dassel, Germany)anddegassedpriortouse.High-puritywaterwasproducedfromdeionizedwaterusinga Milli-Q1
GradientA10(Millipore,Molsheim,France).Acetonitrile (FisherScientificUKLimited,Loughborough,UK),methanol(VWR International,Fontenais-Sous-Bois,France),acetone(VWR Inter-national, Fontenais-Sous-Bois, France), and chloroform (Fisher Scientific UK Limited, Loughborough, UK) were HPLC grade. Acetonitrile(puriss.)forsemi-preparativeHPLC,dimethyl sulfox-ide (p.a.), 25% ammonia, sodium hydroxide (99%), sodium chloride(p.a.),HEPES(p.a.),TRIS(99.9%),andEDTA(disodium
Fig.1.(a)RibbonrepresentationofthestructureofMtgA-E100QincomplexwithmoenomycinAboundtothedonorsite(PDB3HZS).ThemodifiedresiduesL119N,F120Sare showninstick.(b)ElectrostaticsurfacerepresentationofMtgAwithpositivelychargedresiduesinblue,negativelychargedresiduesinredandneutralresiduesingrey.The residuesofthehydrophobicsurfaceareindicated.FiguresweregeneratedbyThePyMOLMolecularGraphicsSystem,Version1.5.0.4Schro¨dinger,LLC).
salt,dihydrate;p.a.)werepurchasedfromCarlRothGmbH+Co (Karlsruhe,Germany).Ethanolamine(98%),potassium dihydro-gen phosphate (puriss, p.a.), anhydrous dipotassium hydrogen phosphate(purum,p.a.),TWEEN1
20,SDS(99.0%), 5-amino-2-nitrobenzoic acid(97%), 1,10-carbonyldiimidazole (>90%),
N-(3-dimethylaminopropyl)-N0-ethylcarbodiimide hydrochloride
(purum)andN-hydroxysuccinimide(98%) werepurchased from Sigma-AldrichChemieGmbH (Steinheim,Germany).85% ortho-phosphoricacid(p.a.),ammoniumhydrogen carbonate(puriss.), sodium acetate trihydrate (p.a.) and sodium nitrite (p.a.) were purchasedfromMerckKGaA (Darmstadt, Germany).Anhydrous pyridine (p.a.) was obtained from Fluka (Seelze, Germany), hydrochloric acid (36%) from Fisher Scientific UK Limited (Loughborough, UK), and CHAPS (98.0%) from A.G. Scientific, Inc. (San Diego, USA). Flavomycin1
reference standard was purchasedfrombiovet1
(Peshtera,Bulgaria).
UnlabledlipidIIwaspreparedasdescribed[23].Thesyntheses oflipidIIanalogs21,43,44,56,5761,62havebeendescribedin
[5].
2.2. PreparationofmoenomycinA
MoenomycinAwasisolatedfromtheFlavomycin1
reference standard aspreviously reported [24]by semi-preparativeHPLC followedbydesaltingviaSPE.TheHPLCseparationwasperformed with6mL/min50mMKH2PO4pH2.3/acetonitrile58:42(v/v)at
458C and a detection wavelength of 280nm. With each chromatographic run approximately 35mg Flavomycin1
in 100
m
L water/acetonitrile 8:2 (v/v) were injected. Fractions containing moenomycin A were analyzed via analytical HPLC (1mL/min100mM NH4HCO3 pH8.0/acetonitrile 55:45 (v/v)at458C) and chromatographically pure moenomycin A fractions werecombined,diluted1:2with50mMphosphatebufferpH7.0 (toimproveretentionontheSPEphase)anddesaltedviaSPEon CHROMABOND1
HR-X SPE cartridges (1000mg sorbent) (Macherey-Nagel GmbH & Co. KG, Du¨ren, Germany). After conditioning with methanol, followed by water and 50mM phosphatebufferpH7.0(15mLeach),oneachSPEcartridgethe combinedmoenomycinAfractionsfromthreechromatographic separations (corresponding to approximately 100mg Flavomy-cin1
) were loaded with 10mL/min. After washing with water (3times5mL),moenomycinAwaselutedwithacetone/water6:4 (v/v)with1mL/minatmaximum(2times10mLhaveproventobe sufficient).Acetonewasremovedviarotary evaporatorand the remainingsolutionswerelyophilized.
2.3. SynthesisofmoenomycinAaminoderivative
5-Amino-N-(2-amino-ethyl)-2-nitrobenzamide was synthe-sizedaccordingtoStembera etal. [25]exceptfortheactivated 5-amino-2-nitrobenzoic acid beingslowly added to the cooled solutionofethylenediamineandtheproductbeingpurifiedby columnchromatographyon silicagel(washedwithHClandthe water content set to 1.5% (w/w) according to [26]) with 25% ammonia/methanol1:25(v/v).
Compound3bin [25]wassynthesizedfrom100mg moeno-mycinAaccordingtoStemberaetal.[25]exceptthatnoshieldgas wasusedandexceptforcleanup,whichwasperformedbySPEon CHROMABOND1
HR-XSPEcartridges(3cartridgeswith1000mg sorbenteach)(Macherey-NagelGmbH&Co.KG,Du¨ren,Germany) afteradjustingthepHto7.5(theaminoderivativeispoorlysoluble in acidic solution). After washing with50mM K2HPO4 pH7.5/
acetonitrile8:2(v/v)(2times5mLpercartridge)andsubsequent washingwithwater(5mLpercartridge),theproductwaseluted with acetone/water 6:4 (v/v) (4 times 10mL per cartridge). Acetonewas removedviarotary evaporator and theremaining
solutionwaslyophilizedtoyield 87mg of thedesiredproduct. Identitywasconfirmedbyitsmeasuredaccuratemass(negative mode ESI–ToF–MS). The residual moenomycin A content was determined via HPLC with 1mL/min 50mM KH2PO4 pH 2.3/
acetonitrile47:53(v/v)afterdissolvinganaliquotoftheproductin theeluentmixtureand membranefiltration(CHROMAFIL1
Xtra PET-45/25(Macherey-NagelGmbH&Co.KG,Du¨ren,Germany)to removeundissolvedproduct.Residualcontentof 5-amino-N-(2-amino-ethyl)-2-nitrobenzamide was determined by HPLC with 1mL/min 100mM NH4HCO3/acetonitrile 95:5 (v/v) (detection
wavelength: 380nm).The productcontained4%(w/w) moeno-mycin A and <0.1% (w/w) 5-amino-N-(2-amino-ethyl)-2-nitro-benzamide(correspondingto<1%molefraction).
2.4. PreparationofSPRbiosensorchipsurface
The moenomycin A amino derivative was tethered to the surface of a Sensor Chip CM5 (GE Healthcare Bio-Sciences AB, Uppsala,Sweden)asdescribedbyStemberaetal.[25].
2.5. DirectSPRbiosensorassayfordeterminingglycosyltransferase bindingactivity
The glycosyltransferases (GT) were diluted in SPR running buffer(25mMTrisbufferpH7.5containing300mMNaCland0.3% CHAPS) and injected over the sensor chip surface for 120s. Regenerationwasperformed by injectionof 1%SDS inrunning bufferfor30sfollowedbya60sinjectionofpurerunningbuffer. Allstepswereperformedataflowrateof40
m
L/min.ForeachGTa dilutionserieswasanalyzed.SinceaCHAPSmoleculewasfoundinthecrystalstructureof PBP1a[20],weverifieditseffectontheGTactivityandfoundno effectforupto2timestheconcentrationusedintheassay(data notshown),showingthatCHAPSdoesnotinterferewithourassay. TheSPRbiosensorusedforourexperimentscannot simulta-neously inject a solution over different chip surfaces, thus subtractionofdatafromareferencecellcouldnotbeperformed inreal-time.Tocompensatepossiblebaselineshiftduringanalysis, sensorgrams are expressed as relative response values (actual response minus response 10s before main injection of the respectivecycle)versustimeinsteadofabsoluteresponsevalues versustime(thesameappliestoSection2.6).
2.6. CompetitiveSPRbiosensorassayoninhibitoractivity
For each inhibitora dilution serieswaspreparedinrunning buffer((25mMTrisbufferpH7.5;300mMNaCl)/DMSO9:1(v/v) containing0.3%CHAPS).Thesolutionsweremixedwithastock solution of Staphylococcus aureus MtgA wild type (in running buffer)resultinginafinalMtgAconcentrationof400nM.Eachof thesemixtureswasinjectedoverthesensorchipsurfacefor150s withaflowrateof10
m
L/min.GTbindinglevelswerecalculated fromtheresponse30saftertheendoftheinjection(toeliminate thebulkeffectofthesolution)minustheresponse10sbeforestart oftheinjection(toaccountforbaselineshiftinthecourseofthe analysis).Regenerationwasperformedbyinjectionof1%SDSin runningbufferfor30sfollowedbya60sinjectionofpurerunning buffer(bothstepsat40m
L/min).2.7. Site-directedmutagenesisandpurificationofproteins
Single and double mutations were generated using Quick-Changesitedirectedmutagenesismethod(AgilentTechnologies) usingtheoligonucleotidesinTable1andtheplasmidpDML2004 astemplate[27].Allconstructswerecheckedbysequencing.For theexchangeof6aminoacids(MtgA-6M:L112N,L119N,F120S,
F150S,V154S, F158S),a modified mtgA genewasorderedfrom GeneArt(Lifetechnologies).
Escherichia coliPBP1b was expressedand purified using the pDML924plasmidaspreviouslydescribed[11].
S. aureusMtgA andits mutantswereprepared, purified and theiractivitiestestedaspreviouslydescribed[27].MtgAmutants werepurifiedunderthesameconditionsasthewildtypeunless indicated.AlltheMtgAvariantsusedarewithouttransmembrane (TM)segment,this deletiondoesnotaffectthestructureof the enzyme[18].DeletionoftheTMregionwasfoundtoreducethe activityoftheenzyme,mostprobablybecauseTMcontributesto thehydrophobicinteractionwiththesubstrate[19].Thisconstruct ismoresuitedforourstudiesbecauseitwillbetterhighlightthe specificitiesoftheenzymeforthepolarregions(sugars,peptide andphosphates)ofsubstrateanalogs.
2.8. AnalysisofMtgAandmutantsbygelfiltration
GelfiltrationofMtgAandmutantsMtgA-NSandMtgA-F120S (2mg/mL)wasperformed ona 24mLSuperdex-75column (GE healthcare)in buffer25mM Tris–HCl pH8, 500mM NaCl,10% glycerol using Akta Explorer (GE healthcare). Protein fractions werecollectedandanalyzedbySDS-PAGE.
2.9. Circulardichroism(CD)
Far-UVCDspectra(200–260nm)wererecordedwithaJasco J-810spectropolarimeterat208Cin25mMTris–HClpH8,500mM NaCl,usinga1mmpathlengthquartzSuprasilcell(Hellma),with proteinconcentrations of ca.0.1mg/mL (i.e.4
m
M).Four scans (10nm/min,1nmbandwidth,0.1nmdatapitchand1sDIT)were averaged, base lines were substracted and no smoothing was applied.Data are presented as the residue ellipticity ([u
]MRW),calculatedusingthemolarconcentrationofproteinandnumberof residues.
2.10. Largeunilamellarvesicle(LUV)preparationandinteraction withMtgA-WTandvariants
E. colilipidsextract (20mg)solutions in chloroform(Avanti PolarlipidsInc.,Alabama,USA)weremixedandevaporatedunder agentlestreamofnitrogen.Thelipidfilmwassubsequentlydried for20minundervacuum.Thefilmwashydratedbytheadditionof thebuffer(25mMHepes pH7.5,150mM NaCl,10mM MgCl2)
under mechanical agitation and submitted to 10 freeze–thaw cyclesusingliquidnitrogenandawaterbath.Thelipidsuspension wasthenextruded10timesthroughapolycarbonatemembrane filterwithaporesizeof200nm(WhatmanInternational).
MtgAandvariantsat80
m
Morlysozymeusedascontrolwere preincubatedwithorwithoutLUVfor30minat48Cundergentle mixing.Themixturewasultracentrifugedat122,000g,148Cfor 30min.Thesupernatantandpelletfractionswereseparatedand the pellet was washed 3 times in 25mM Tris buffer pH 8, containing300mMNaClfollowingthesameprocedure.Fractionsrecoveredaftereachcentrifugationstep wereanalyzedby SDS-PAGEandtheamountofbound(pellet)and unbound (superna-tant)proteinswereestimated.
3. Results
3.1. IsolationofmoenomycinAfromthemixtureofcongeners Moenomycin A was obtained as a colorless powder with a purity>99%(HPLC-UV,258nm,Fig.2).Duetotheuseof basic bufferaseluentfortheanalyticalHPLCinsteadofacidicbuffer(as used for semi-preparative HPLC), the elution order is changed makingiteasytoidentifyanyremainingmoenomycinA12inthe
isolated moenomycin A. Overall 225mg moenomycin A were isolatedfrom1050mgFlavomycin1
.ByHPLC-UVwedetermineda moenomycinAcontentof30%inthestartingmaterial,accordingly theyieldwas>70%.IdentityoftheisolatedmoenomycinAwas confirmedby1HNMR (conditionsasdescribedbyHennigetal.
[28])andESI–MSinnegativemode(datanotshown).
3.2. OptimizationoftheSPRassayandanalysisofGTbindingto moenomycinA
MoenomycinAinhibitsthetransglycosylationstepbybinding tothedonorsiteoftheglycosyltransferase.Thus,theabilityofaGT tobindmoenomycinAcorrespondstoitsabilitytobinditsnatural donorsubstrate(i.e.thegrowingglycanchain)andindicatesthat thedonorsiteiscorrectlyfoldedandfunctional.
Stembera et al. [21,25] described an SPR biosensor assay utilizingamoenomycinAaminoderivativetetheredtothechip surfaceandsuccessfullyappliedittotestdifferentmoenomycinA analogsfortheirrelativeaffinitytowardsE.coliPBP1b[21](the analogsinsolutioncompetewithimmobilizedmoenomycinAfor theGTdonorsites).Wehaveoptimizedthisassaybycomparing PBP1bandS.aureusMtgAandfoundconditions,whereMtgAwas more sensitive and data more reproducible than with PBP1b (probablyduetothepresenceofthetransmembranesegmentin PBP1b and requirementof higher detergent concentrations for solubility).UndertheseconditionsmoenomycinAat concentra-tionsaslowas0.5
m
McausedaclearinhibitionofMtgAbinding (Fig.3a).InStemberaetal. [21]100m
MmoenomycinAclearly inhibited PBP1b binding, whereas 10m
M did not cause any inhibition.Wereproducedtheirexperimentsandconfirmedthis lowersensitivityforPBP1b.ThereforeMtgAwaschosentoanalyze itsinteractionwithsubstrateanalogs.First,2-folddilutionseriesofMtgAfromtwodifferentbatches rangingfrom125nMto2
m
Mwereusedtoanalyzeitsbindingto immobilized moenomycin A. Clear concentration dependent binding responses were observed (Fig. 3b). However, the sensorgramsobtainedcannotbedescribedbyasingleexponential function(yetthisshouldbepossibleforasimple1:1interaction withoutmasstransportlimitation[29]).Theydonotreachasteady stateforanyconcentration(thusequilibriumanalysiscannotbe performedtocalculatedissociationconstants).Instead,afterafirst exponential phase, the binding curves reach a phase with practicallyconstant(yetincomparisonflat)slope,whichcannot beexplainedbya1:1interaction.Thiscomplexityinthebinding responseswasalsoobservedonaflowcellwithlowliganddensity (310RUmoenomycinAaminoderivativecomparedto720RUfor the other experiments) and even at low MtgA concentrations (Fig.3c). Even here and despite the prolonged injection (300s insteadof120s)asteadystatewasnotreached.Sincenon-specific bindingofGTswasnotobservedatconcentrationsbelow5m
M, we presume this complexity of binding responses is caused by aggregation of GTs to one another after reaching a critical concentrationonthechipsurfacemostlikelyviathehydrophobicTable1
Sequencesofoligonucleotidesusedformutagenesis. mtgAmutants Oligonucleotides
L119N-F120S 50-GGTACAACTAGAGTCAACTCTTCAACGATTAGCGAC-30 50-GTCGCTAATCGTTGAAGAGTTGACTCTAGTTGTACC-30 F120S 50-CTGTCGCTAATCGTTGAAGATAAAGCTCTAGTTGTAC-30 50-GTACAACTAGAGCTTTATCTTCAACGATTAGCGACAG-30 F150S 50-GATAATGATCGCAGCTTCACCCGCAAAAGCAAAG-30 50-CTTTGCTTTTGCGGGTGAAGCTGCGATCATTATC-30 F104A 50-CAATGGAAGATGAACGAGCCTACAATCATCATGGAT-30 50-GTTACCTTCTACTTGCTCGGATGTTAGTAGTACCTA-30
jawregion (attempts toimprove solubility of MtgA have been investigated;seebelow).Sincerateconstantsvaluesderivedfrom such complex binding data are known to provide ambiguous results [30], we do not present any value for association and dissociation rate constants characteristic of the interaction betweentheGTsandmoenomycinA,butinsteadlimitourselves toqualitativeoratmostsemi-quantitativeconclusions.
3.3. ModificationofthehydrophobicregioninMtgAand characterizationofthemutants
The GTs are characterized by a hydrophobic surface that mediates the interaction of the proteins with the cytoplasmic membrane (Fig. 1b). MtgA devoid of its transmembrane (TM) segmenthastendencytoaggregateduetoitshydrophobicregion. Thisbehaviorand theneed ofhighconcentrationsofdetergent preventfromdetailedkineticandbindingcharacterizationinvitro. Wehavetriedtoreducethepolydispersityoftheproteinthrough mutagenesisoftheresiduesinthehydrophobicsurfacetofacilitate invitrostudybySPR.InMtgA,thissurfaceincludesF110,L112, L119,F120,F150,V154,L157,F158.Thefollowingsinglemutations
F143T,V154T,L157TorF158T havenoeffecton theactivityof MtgAbutdidnotimprovethesolubilityoftheproteinsignificantly
[18]. F110 may be essential for binding of lipid II and its modification would affect activity [18]. We have prepared the followingsinglemutationsF120SandF150S,thedoublemutation L119N-F120S (MtgA-NS) and the multiple mutations simulta-neouslyMtgA-6M(L112N,L119N,F120S, F150S,V154S, F158S). ThemutantF150SandMtgA-6Mcouldnotbeexpressedinsoluble form(e.g.presenceinthesupernatantfractionwithoutdetergent after cell lysis and centrifugation) using different expression conditions (18–378C). Since a Pheat position150seemstobe importantforproperfoldingoftheprotein,anothermutantofall 6residues,butF150waspreparedandalsofoundinsoluble.
ThesinglemutantF120SandthedoublemutantL119N-F120S couldbeexpressedinsolubleform.Analysisofthedoublemutant (2mg/mL)bygelfiltrationwithoutdetergentshowedtheprotein eluted as a monomer in a single peak without any aggregate (Fig.4a).UndertheseconditionsthewildtypeMtgAwasentirely found in the void volume with highmolecular mass (Fig. 4a). Without detergent, monodispersity of the singlemutant F120S was in between that of wild type MtgA and MtgA-NS. In the
Fig.2.(a)StructuresofthemoenomycinmaincongenersinFlavomycin1
.(b)Chromatogramofthesemi-preparativeHPLC(pH2.3)separationof35mgFlavomycin1
. (c)ChromatogramoftheanalyticalHPLCseparation(pH8.0)ofFlavomycin1
incomparisontotheisolatedmoenomycinA.MA,MA12,MC1,MC3,MC4:MoenomycinA,A12,
C1,C3,C4.
Fig.3.SensorgramsofS.aureusMtgAbindingtomoenomycinchip.(a)SensorgramsofaconcentrationseriesofmoenomycinAinpresenceof400nMMtgA;injectionofthe solutionsstartsat0sandlastsfor150s.(b)MtgAconcentrationseries;injectionofthesolutionsstartsat0sandlastsfor120s.(c)complexbindingresponsesofMtgA.
presence of CHAPS about 80% of the wild type protein was monodisperseand 20% aggregate(datanot shown).Despite its monodispersityMtgA-NSwasstillabletobindlipidvesicles(LUV) (datanotshown).Far-UVcirculardichroism(CD)comparisonof MtgA-NSandMtgA-F120SwiththewildtypeMtgAshowedsimilar spectra (Fig. 4b), characteristic of
a
-helices profile with two characteristicminimaat208nmand222nm,inagreementwith theproteinstructure.Thisresultsuggeststhatthemutationdid notmodifythesecondarystructuresoftheproteinsignificantly. ThemutantsMtgA-NSandMtgA-F120Swereunabletoconvert radioactivelipidIIsubstratetopolymericpeptidoglycanevenat high protein concentration indicating that they were inactive. Previous results showed that the MtgA mutant E100Q which retains 0.2% of the wild type activity binds to moenomycin A[18,27].ComparisonofE100QwithanotherMtgAmutantF104A (1%ofwildtypeactivity)bySPR(at400nM),showedthatE100Q bindsmoenomycinAinthesameorderofmagnitudeasthewild type while the binding response for F104A was drastically decreased (Fig. 5a). Binding responses of both mutants were specificviatheirdonorsiteaswastestedbyanalyzingGTbinding tothechipsurfaceinpresenceofmoenomycinA(datanotshown). ThisdatashowsthatinactiveorbarelyactiveMtgAmutantscould stillbe abletobindmoenomycin Aindicating that atleast the donorsiteisfunctional.Totestthispossibilityinthecaseofthe mostsolublemutantL119N-F120S,westudieditsinteractionwith moenomycinAusingSPR.
When testedat125nMto2
m
M,MtgA-NSdidnotshowany bindingresponses,itwasthenanalyzedathigherconcentrations(Fig.5b and c; additional experiment withdouble referencing:
Fig. 5d–f). The binding responses were rather random than concentrationdependent,stronglysuggestingnon-specific bind-ingof MtgA-NSto moenomycinA (e.g.bindingof hydrophobic regionsoftheGTtothemoenocinolmoiety)(Fig.5).Hencethe aminoacidexchangesL119NandF120Sinthemembranebinding surfaceofMtgAresultedinmonodisperseproteinandacomplete lossof specificmoenomycinA bindingcapability. Theseresults suggest that the hydrophobicity of this region is important to maintainthecorrectstructureofthejawsubdomainand/orforthe interactionwiththesubstrateand moenomycin.Another possi-bilitycouldbethatthetwomutations(L119N,F120S)mayaffect theouterhelixtransition/conformationalchangeduringcatalysis.
3.4. Positivecooperativitybetweentheacceptoranddonorsitesofthe GT
SeveralanalogsofGT’slipidIIsubstrate(21,43,44,56,57,61, and62) (Fig.6)havebeensynthesizedpreviouslyandfoundto inhibittheGTactivityinvitroandcausebacterialgrowthdefect
[5]. These compoundsare valuable toolsfor understanding the complexcatalyticmechanismof theGT.However,their precise modeofaction remainselusive,suchastheir specificityforthe donorand/oracceptorsites.Inordertoanswerthesequestionswe haveanalyzedtheireffectonmoenomycinAbindingofMtgAusing SPR.Inthisassay,inhibitorscapableofbindingtothedonorsiteof GTsshouldcompeteforthisbindingsitewiththemoenomycinA aminoderivativetetheredtothesensorchipsurface,resultingina decreasedGTbindingresponse(Fig.3a).
Mixtures of MtgA with varying concentrations of the com-poundsweretestedfortheireffectonMtgAbindingtoa sensor chipsurfacemodifiedwiththemoenomycinAaminoderivative.In addition,thenaturalsubstratelipidII(Fig.6)wasanalyzedinthe presenceofEDTAintherunningbuffertoavoidpolymerization.All substanceswereanalyzedin3-folddilutionseriesandmeasuredin duplicate.Forinhibitors21,43,44,and62theconcentrationrange was 7nM to400
m
M, for inhibitors57 and 61 it was 3nM to 200m
M,forinhibitor56itwas3nMto148m
M,andforlipidII 2nMto120m
M.AllsubstancestestedshowedinhibitionofMtgA binding at high concentrations (Fig. 7a–d). Interestingly, at concentrations below approximately 50m
M, the disaccharide analogs 21, 44, and 62 clearly showed an increased binding responseincomparisontoMtgAinabsenceofinhibitor(Fig.7a). Inhibitor57(belowapproximately10m
M)andlipidIImighthave thesameenhancementeffect,buttoalesserextent(Fig.7candd). In case of lipid II, the binding responses had a high spread, preventing a thorough interpretation, most likely due to poor solubility.EnhancingsolubilityoflipidIIbyincreasingtheDMSO content inthe runningbufferdid resultin a drasticchangein refractiveindexleadingtoSPRanglesoutsidethedetectorrange. Also,EDTAisknowntoformcomplexeswithmoenomycins[31]whichmightaddtothepoorqualityofthedata.The monosaccha-ride analogs 43, 56, and 61 clearly showed a concentration dependentinhibitionbutincreasedbindingwasnotobservedwith thesecompounds(Fig.7b),thusindicatingasimplecompetition withmoenomycin A for thedonor site of the GT.Non-specific bindingoftheinhibitorstothesensorchipsurfacecanberuledout asacausefortheincreasedbindingbehavior,becauseitwouldnot explain the decrease of binding responses at higher inhibitor concentrations.Fromthestructuralsimilarityofthedisaccharide analogs withlipid IIsubstrate theyare expectedtobindtothe acceptor site of MtgA. Therefore, we suggest that at low concentrations the disaccharide compoundsbind selectively to the acceptorsite and increase theaffinity of the donor site to moenomycin A by heteroallosteric activation leading to an increased MtgA binding response. This effect is rivaled by the
Fig.4.ComparativeanalysisofMtgAandthemutantsMtgA-NSandF120Sbygel filtrationandcirculardichroism.(a)SizeexclusionprofileofMtgA,mutants MtgA-NSandF120SonSuperdex75column.(b)FarUVCDspectraofMtgA,mutants MtgA-NSandF120S.
abilityoftheinhibitorstobindtotheGT’sdonorsiteathigher concentration, which explains the decrease in the binding responses(competitionforthedonorsiteoutweighstheenhanced binding).Suchanallostericactivationwouldaffecttheassociation and/ordissociationratesforinteractionswiththedonorsite.Fig.8
shows an overlay of sensorgrams for inhibitor 62 in the concentrationrangearoundthemaximumofMtgAbinding.The highestbindingresponsewasobtainedataninhibitor concentra-tionof4.9
m
M.From thereon thebindingresponsesdecreased bothwithincreasinganddecreasinginhibitorconcentration.Since dissociation oftheGT-inhibitor complexes followsa firstorder kinetic (i.e. dependent on the complex’ concentration, d[LA]/ dt=kd[LA]), the sensorgram for the highest GT bindingresponse(4.9
m
M)shouldhavethesteepestslopeindissociation phase,unlesskdhadchanged.Thesensorgramsforthetwohighestinhibitorconcentrations(44and 133
m
M),however,featurethe steepest slopein thedissociation phase (most clearly observed after180s),suggestingthatkdhaschanged,hencesupportingourhypothesis.
4. Discussion
IntermsofGTactivityinhibition,thestudiedcompoundscanbe ranked as follows from the best active to the least active 62>44>43>21>61=57>56(Fig.6)[5].Compounds21and 57 differ in the dipeptide moiety LAla-DGluin 57. Compound 21showedbetterenzymaticinhibitionandhigherMtgAbinding enhancement tomoenomycin A (cooperativeeffect) than com-pound57.Thisshowsthatthepeptideaffectsatleastpartiallythe bindingtotheacceptorsite.Compounds62and57bothhavea dipeptide moiety but differ in the phosphate linker, 57 has a monophosphategroupand62hasaphosphoglycerate,amimicof thepyrophosphateofthesubstrate.62causeshigherincreasein GTbindingtomoenomycinAthan57.Thisdifferencesuggeststhat
thephosphoglycerategroupmaybeanimportantdeterminantfor bindingtotheacceptorsite.Compound43isa monosaccharide-phosphoglyceratewithGTinhibitionactivitycomparableto44and 62 but did not show any enhancing effect on moenomycin A binding,yetonlyinhibitsbindingathigherconcentration.These resultssuggestthat43mayhaveapooraffinityfortheacceptor siteandhigherbindingaffinitytothedonorsite.Itisnoticeable thatthefourlipidIIanalogswithdisaccharidestructure(inhibitors 21,44,57and62)weretheonesenhancingMtgAbindingatlow concentrations. Therefore it seems likely that the disaccharide moiety is moreimportantfor theaffinity towardstheacceptor site thanfor theaffinity towardsthedonorsite. Adisaccharide compoundsuchas44,withoutapeptideandwith phosphoglyc-erate-lipidgroup,seemstobethebestligandtotargetboththe donor and acceptor sites. Co-crystallization of a GT with this compoundwouldprovidevaluableinformationtobetter under-standthecatalyticmechanismandcouldhelpinthedesignofnew inhibitors.
MostofthecrystalstructuresofGTinapoformorincomplex withmoenomycinA,showanunstructuredormobileregioninthe jaw domain between donor and acceptor sites. This region (Ser121-Gly130)locatedbetweenconservedmotifs1and2was proposedtohaveanimportantroleinthebindingoftheacceptor substrate[16].Weproposethatthedisaccharideanalogs(orlipid II)bindpreferentiallytotheacceptorsiteandstabilizethemobile regionwhichinduceslocalstructurerearrangementinthedonor siteincreasingitsaffinityformoenomycinA(orthesecondlipidII). Thismechanismmaybeparticularlycriticalduringtheinitiation stage(lipidIIinthedonorsite).Theallostericactivationofbinding inthedonorsitecouldalsoapplytotheelongatingglycanchain and would play animportant role inthe processiveelongation mechanism.Consequently,aftereachcatalyticcyclethe translo-cationoftheproductfromtheacceptorsitetothedonorsite is concomitantwiththeunfoldingormovementofthemobileregion,
Fig.5.SensorgramsofS.aureusMtgAmutantsbindingtomoenomycinAchip.(a)WildtypeMtgAandmutantsE100QandF104Aanalyzedat400nM;injectionofMtgA solutionsstartsat0sandlastsfor150s.(b)ConcentrationseriesofMtgA-NS;injectionstartsat0sandlastsfor120s.(c)Dissociationphaseofthesensorgramsinb). (d–f)TestofS.aureusMtgA-NSforspecificityofbindingwithMtgAaspositivecontrol:BindingdatafromthereactioncellwithmoenomycinAaminoderivative tetheredtothechipsurface(d)wascorrectedforbulkshiftandbaselinedriftbysubtractionofdatafromanunmodifiedreferencecell(e)andbufferblanksubtractionto yieldthepurebindingresponses(f).ThespikesatthebeginningandtheendoftheGTinjectionsin(f)arecausedbyaslightmisalignmentbetweenthedatafromboth chipsurfaces.
whichcontributetothis processbeforea newlipidIImolecule bindsandstabilizesitagaintostartanewcatalyticreaction.
The bi-substrate reaction of the GT at initiation stage of polymerizationmakes distinctionbetween theaffinities for the substrateofthedonorsiteandacceptorsitecomplex.Moenomycin AbindstothedonorsiteoftheGTandisproposedtomimicthe growingchain.LipidIIbindstobothsitesaswoulddoitsanalogs. The concentration dependent effect of lipid II analogs on moenomycin A binding indicates the presence of allosteric activation of the donor site following ligand binding to the acceptorsite.DuetostructuralsimilaritiesbetweenlipidIIandthe investigatedinhibitorsandbasedontheobtaineddataweexpect thesamebehaviorforlipidIIitself.Solubilityproblemshowever, resultinginahighspreadofthebindingresponses,preventeda thoroughinvestigationforthiscompound.
Thiseffecthasgreatimpactonthecatalyticmechanismand couldexplaintheroleofthemobileregionbetweenthetwosites.
The binding of lipid II to the acceptor site could induce local structuralrearrangement,whichalsoaffectsthedonorsiteviathe mobileregion.Duringelongationphase,stabilizationofthemobile region would increase binding to the elongating chain and contributetotheprocessivepolymerization.Aftereachcatalytic cycle,producttranslocationfromacceptortodonorsiteisassisted by higher affinity of the pyrophosphate-lipid moiety (larger positivelychargedpatchesinsidethedonorsite)andthemobility oftheunstructuredregion,concomitantwiththevacancyofthe acceptorsite.
So farweshowedenhancementof MtgAbindingto moeno-mycinA,whichissupposedtomimicthegrowingglycanchain. Further studies using other approaches are needed to better characterizethepostulatedcooperativitymechanismby investi-gatingtheinteractionbetweenlipidIIandMtgAandeventually otherGTs.Tobebiologicallyrelevantforpeptidoglycansynthesis, suchanenhancementshouldalsoincreasetheaffinityofthedonor
siteofMtgAtothegrowingglycanchainitself.SPRexperiments withthetetrasaccharidelipidIVtetheredtothesensorchipsurface insteadofmoenomycinAcouldprovebeneficialhere.
Acknowledgments
This work was supported by the Fond de la Recherche FondamentaleCollectiveFRFCno.2.4543.12,theBelgianScience PolicyprogramsIAP-VIIiPROSproject(no.P7/44),FRIAfellowship toIDandtheBelgianSciencePolicyreturngranttoAD.Mohammed TerrakisresearchassociateoftheF.R.S-FNRS.WethankProfessor Heiko Hayen for valuable advice during development of the isolationprocedureformoenomycinAandEricSauvageforhelpin preparationofFig.1.
References
[1]SauvageE, KerffF,TerrakM,AyalaJA, CharlierP.Thepenicillin-binding proteins:structureandroleinpeptidoglycanbiosynthesis.FEMSMicrobiol Rev2008;32:234–58.
[2]Huang SH,Wu WS,Huang LY, HuangWF,Fu WC,Chen PT,et al.New
continuousfluorometric assayforbacterialtransglycosylase usingForster resonanceenergytransfer.JAmChemSoc2013;135:17078–89.
[3]GampeCM,TsukamotoH,DoudEH,WalkerS,KahneD.Tuningthe moeno-mycinpharmacophoretoenablediscoveryofbacterialcellwallsynthesis inhibitors.JAmChemSoc2013;135:3776–9.
[4]ShihHW,ChenKT,ChenSK,HuangCY,ChengTJ,MaC,etal.Combinatorial approachtowardsynthesisofsmallmoleculelibrariesasbacterial transgly-cosylaseinhibitors.OrgBiomolChem2010;8:2586–93.
[5]DumbreS,DerouauxA,LescrinierE,PietteA,JorisB,TerrakM,etal.Synthesis ofmodifiedpeptidoglycanprecursoranaloguesfortheinhibitionof glycosyl-transferase.JAmChemSoc2012;134:9343–51.
[6]GampeCM,TsukamotoH,WangTS,WalkerS,KahneD.Modularsynthesisof diphospholipidoligosaccharidefragmentsofthebacterialcellwallandtheir usetostudythemechanismofmoenomycinandotherantibiotics. Tetrahe-dron2011;67:9771–8.
[7]WelzelP. Synthesesaroundthetransglycosylation stepin peptidoglycan biosynthesis.ChemRev2005;105:4610–60.
[8]Ostash B, Walker S.Moenomycin family antibiotics: chemical synthesis, biosynthesis,andbiologicalactivity.NatProdRep2010;27:1594–617.
[9]TerrakM,SauvageE,DerouauxA,DeharengD,BouhssA,BreukinkE,etal. Importanceoftheconservedresiduesinthepeptidoglycanglycosyltransferase moduleoftheclassApenicillin-bindingprotein1bofEscherichiacoli.JBiol
Chem2008;283:28464–70.
[10]SchwartzB,MarkwalderJA,SeitzSP,WangY,SteinRL.Akinetic characteriza-tionoftheglycosyltransferaseactivityofEschericiacoliPBP1band develop-mentofacontinuousfluorescenceassay.Biochemistry2002;41:12552–61.
Fig.7.BindinglevelsofMtgAinpresenceoflipidIIanalogsorlipidII:MtgApreincubatedwithvaryingconcentrationsoflipidIIorthedifferentlipidIIanalogswasinjected overthesensorchipsurfacemodifiedwiththemoenomycinAaminoderivative.MtgAbindinglevelswerecalculatedfromtheresponse30saftertheendoftheinjection minustheresponse10sbeforestartoftheinjection.(a)Disaccharideinhibitors21,44,and62;(b)monosaccharideinhibitors43,56,and61;(c)disaccharideinhibitor57; (d)lipidII.
Fig.8.SensorgramsofMtgA(400nM)inpresenceof0.02mMto133mMinhibitor62.(a)Completesensorgrams.(b)Dissociationphaseofthesensorgramsin(a).Thebox marksthetimeframeforcalculationofthebindingresponses.Injectionofthesolutionsstartsat0sandlastsfor150s.Theoffsetinthesensorgramfor133mMduring injectionisduetoaslightmismatchinDMSOcontentbetweentheinjectedsolutionandtherunningbuffer.
[11]TerrakM,GhoshTK,vanHeijenoortJ,VanBeeumenJ,LampilasM,AszodiJ, etal.Thecatalytic,glycosyltransferaseandacyltransferasemodulesofthecell wallpeptidoglycan-polymerizingpenicillin-bindingprotein1bofEscherichia coli.MolMicrobiol1999;34:350–64.
[12]FraipontC,SapunaricF, ZervosenA,AugerG,DevreeseB,LiouxT,etal. Glycosyl transferase activity of the Escherichia coli penicillin-binding protein 1b: specificity profile for the substrate. Biochemistry 2006;45: 4007–13.
[13]PerlsteinDL,ZhangY,WangTS,KahneDE,WalkerS.Thedirectionofglycan chain elongation by peptidoglycan glycosyltransferases. J AmChem Soc 2007;129:12674–75.
[14]PerlsteinDL,WangTS,DoudEH,KahneD,WalkerS.Theroleofthesubstrate lipidinprocessiveglycanpolymerizationbythepeptidoglycan glycosyltrans-ferases.JAmChemSoc2010;132:48–9.
[15]LoveringAL,deCastroLH,LimD,StrynadkaNC.Structuralinsightintothe transglycosylationstepofbacterialcell-wallbiosynthesis.Science2007;315: 1402–5.
[16]LoveringAL,DeCastroL,StrynadkaNC.Identificationofdynamicstructural motifsinvolvedinpeptidoglycanglycosyltransfer.JMolBiol2008;383:167– 77.
[17]SungMT,LaiYT,HuangCY,ChouLY,ShihHW,ChengWC,etal.Crystal structureofthemembrane-boundbifunctionaltransglycosylasePBP1bfrom Escherichiacoli.ProcNatlAcadSciUSA.2009;106:8824–9.
[18]HeasletH,ShawB,MistryA,MillerAA.CharacterizationoftheactivesiteofS. aureusmonofunctionalglycosyltransferase(Mtg)bysite-directedmutation andstructuralanalysisoftheproteincomplexedwithmoenomycin.JStruct Biol2009;167:129–35.
[19]HuangCY,ShihHW,LinLY,TienYW,ChengTJ,ChengWC,etal.Crystal structureofStaphylococcusaureustransglycosylaseincomplexwithalipidII analogandelucidationofpeptidoglycansynthesismechanism.ProcNatlAcad SciUSA.2012;109:6496–501.
[20]YuanY,BarrettD,ZhangY,KahneD,SlizP,WalkerS.Crystalstructureofa peptidoglycanglycosyltransferase suggestsamodelforprocessiveglycan chainsynthesis.ProcNatlAcadSciUSA.2007;104:5348–53.
[21]StemberaK,VogelS,BuchynskyyA,AyalaJA,WelzelP.Asurfaceplasmon resonanceanalysisoftheinteractionbetweentheantibioticmoenomycinA andpenicillin-bindingprotein1b.ChemBioChem2002;3:559–65.
[22]Nicholas RA,Lamson DR,SchultzDE. Penicillin-bindingprotein 1Bfrom Escherichiacolicontainsamembraneassociationsiteinadditiontoits trans-membraneanchor.JBiolChem1993;268:5632–41.
[23]BreukinkE,vanHeusdenHE,VollmerhausPJ,SwiezewskaE,BrunnerL,Walker S,etal.LipidIIisanintrinsiccomponentoftheporeinducedbynisinin bacterialmembranes.JBiolChem2003;278:19898–903.
[24]BuryD,Mu¨ller-SeitzE,HayenH,PetzM.IsolierungvonMoenomycinAaus demMastfo¨rdererFlavomycinzurSynthesevonDerivatenfu¨reinen SPR-Assay. Weitere erweiterte Posterbeitra¨ge vom Lebensmittelchemikertag 2012.Lebensmittelchemie 2013;67:2–7. http://dx.doi.org/10.1002/lemi. 201390005.
[25]StemberaK,BuchynskyyA,VogelS, KnollD,OsmanAA, AyalaJA,etal. Moenomycin-mediatedaffinitypurificationofpenicillin-bindingprotein1b.
ChemBioChem2002;3:332–40.
[26]EsterbauerH.U¨ berdieAutoxydationvonLinolsa¨ure-methylesterinWasser III:ChromatographischeAuftrennungderwasserlo¨slichenReaktionsprodukte. FetteSeifenAnstrichm1968;70:1–4.
[27]TerrakM,Nguyen-Diste`cheM.Kineticcharacterizationofthemonofunctional glycosyltransferasefromStaphylococcusaureus.JBacteriol2006;188:2528– 32.
[28]HennigL,FindeisenM,WelzelP,HaessnerR.1
HNMRspectroscopicstudiesof
themoenomycins.MagnResonChem1998;36:615–20.
[29]MortonTA,MyszkaDG,ChaikenIM.Interpretingcomplexbindingkinetics fromopticalbiosensors:acomparisonofanalysisbylinearization,the inte-gratedrateequation,andnumericalintegration.AnalBiochem1995;227: 176–85.
[30]RichRL,MyszkaDG.Surveyoftheyear2007commercialopticalbiosensor literature.JMolRecognit:JMR2008;21:355–400.
[31]WallhausserKH, Nesemann G, Prave P, SteiglerA. Moenomycin, a new
antibiotic I. Fermentation and isolation. Antimicrob Agents Chemother 1965;5:734–6.
