• Keine Ergebnisse gefunden

Structural investigation of multivalent carbohydrate–protein interactions using synthetic biomolecules

N/A
N/A
Protected

Academic year: 2022

Aktie "Structural investigation of multivalent carbohydrate–protein interactions using synthetic biomolecules"

Copied!
8
0
0

Wird geladen.... (Jetzt Volltext ansehen)

Volltext

(1)

Structural investigation of multivalent carbohydrate–protein interactions using synthetic biomolecules

Valentin Wittmann

Understandingmultivalentcarbohydrate proteininteractions atthemolecularlevelrequiresaccesstostructuraldetailsof theseimportantbiologicalrecognitionprocesses.Recent developmentstowardthisgoalcomprisetheuseof

conformationallydefinedmolecularrulersincombinationwith bindingassays,crystallographicinvestigationofcomplexesof multivalentligandsandtheirtargetproteins,anddistance measurementsinthenanometerrangebyEPRspectroscopy.

Address

UniversityofKonstanz,DepartmentofChemistryandKonstanz ResearchSchoolChemicalBiology(KoRS CB),Universita¨tsstr.10, 78457Konstanz,Germany

Correspondingauthor:Wittmann,Valentin(mail@valentin wittmann.de)

Introduction

Molecularrecognitionofcarbohydratestructuresgoverns amultitudeofbiologicalprocessessuchas celladhesion andsignaltransduction.Examplesincludetheinflamma- toryresponse, viraland bacterialpathogenesis, fertiliza- tion,immune system modulation, and cellproliferation [1 3].Theproteins thatinteractwiththecarbohydrates are mostly lectins and antibodies. Generation of high- affinity ligands for these proteins is of great medicinal interestforseveralapplications.Blockingcarbohydrate proteininteractionsmaybebeneficialforthetreatmentof certaindiseasessuchasintoxicationsbybacterialtoxinsor acute or chronic inflammatory diseases. Lectin ligands carryingsuitableprobescanalsoserveasdiagnosticsfor certainconditions.

Ahallmarkofmanycarbohydrate proteininteractionsis multivalency[4,5].Indeed,manylectinscontainmultiple sugarbindingsitesmostlybecausetheyexistasoligomers ortheyareclusteredoncellsurfaces[6].Bindingpartners canbemultiplecarbohydrateswithinaglycanoramulti- valentdisplayofglycansona(cell)surface.Multivalency cannotonlyresultinincreasedbindingaffinity(referred to as avidity) but also in increased binding specificity

[7,8]. Several types of interactions can occur between multivalent ligands and receptors. The simultaneous bindingof amultivalent ligand to several binding sites of areceptor (also referredto as chelating binding)can leadto strongly increased affinity if the spacing of the individualcarbohydrateligandsmatchesthedistancesof thereceptorbindingsites[9].Besideschelatingligands, ligandswithasugarspacingtooshorttobridgeadjacent binding sites can also be effective. The mechanisms discussed to explain increasedbinding affinity in these cases include statisticalrebinding or the bindand slide mechanism[10].

Sincechelatingbindingisespeciallyeffectivetoarriveat high-affinityligands,manygroupsfocusedonthedesign ofsuitableligands[11 14].Accordingly,avastvarietyof scaffolds have been developed for the presentation of carbohydratesincluding dendrimers, peptides, proteins, nucleicacids,(cyclic)oligosaccharides,smallorganicscaf- folds, quantum dots, nanoparticles, surfaces, and poly- mers[15].Sucheffortsfollowdifferentstrategies.Large flexible ligands,such as glycopolymers, cancover wide areas of cell surfaces and bridge multiple membrane- located lectinsin a statistical manner (statistical multi- valency).Ontheotherhandsmalloligovalentligandsmay bindtoseveralbindingsitesofanindividual(oligomeric) lectin. In principle, with rigid well-designed ligands, selectivity for lectins with matched inter-binding-site distances can be achieved over other lectins with the samecarbohydrateselectivitybutunsuitablebindingsite spacing (targeted multivalency). Required for such a designisknowledgeaboutthe3Dstructureoftheprotein andthe locationof the bindingsites. Thisinformation, however,isoftenmissing.Butevenifthestructureofa lectinisknown,differentbindingmodesofamultivalent ligand can occur and it is a challenge to unravel the structuraldetailsof suchinteractions.

Summarizedinthisarticlearerecentexamplesaimingat the determination of structural details of multivalent carbohydrate proteininteractionsusingsyntheticbiomo- lecules.

Bindingmodesofmultivalentinteractions

Carbohydrate-bindingproteinscaninteractwithglycans throughavarietyof mechanisms.Bothbindingpartners (glycanandprotein)canbeimmobilizedonacellsurface oroneorbothbindingpartnerscanexistinsolubleform leadingtodifferentscenarios.Evenwhenlookingatthe simplestsystemofamultivalentinteraction,thebinding Zuerst ersch. in : Current Opinion in Chemical Biology ; 17 (2013), 6. -

S. 982-989

http://dx.doi.org/10.1016/j.cbpa.2013.10.010

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-259437

(2)
(3)
(4)
(5)
(6)

tremendous effectoninhibitorypotency.Whereasdiva- lent6 hadanevenworsepotencycomparedtoamono- valent reference compound, the somewhat elongated 7 wasa545-fold(272-foldpersugar)andsecond-generation ligand 8 (IC50=2.7nM) was a7555-fold(3780-fold per sugar) better inhibitor. Molecular modeling confirmed thatboth7 and8 canbindin achelatingbindingmode to LecA whereas 6 is slightly too short to bridge both bindingsites.

In the examples discussed above systematic spatial screeningofcarbohydrate-modifiedmolecularrulerscon- firmedthedistancesbetweenbindingsitesofstructurally characterizedproteins.Thefindingsindicatethattheuse of suchmolecularrulersisavalidapproachto scan also undefinedproteinsforinter-binding-sitedistances.High- estaccuracyisobtainedwithconformationallyrestricted, rigidlinkers.Inthiscase,however,evensmalldeviations fromthematchingcasemayleadtoasignificantoreven completelossofbindingenhancements.Therefore,sev- eral groups developed combinatorial approaches to identify multivalent ligands from large ensembles of spatiallydiverseglycoclusters[37 41].

Crystallographicinvestigationofmultivalentinteractions X-raycrystallographyofcomplexesofmultivalentligands and theirtargetproteins providesthemostcomprehen- sive picture of the mode of interaction between the binding partners. Acommon problem in generating co- crystals of multivalent ligands and receptors isthe for- mation of random aggregates. Consequently, structural information onsuchcomplexes israre.

Lameignere et al. investigated BC2L-A,a Burkholderia cenocepaciasolublelectinrelatedtothelectinLecBfrom P.aeruginosa[42].BC2L-Aisamonovalentlectinthat formsdimers.IthasaspecificityformannosideswithaKd

inthemicromolarrangeandishypothesizedtoplayarole in biofilm formation. Examined ligands included three differentlylinkedmannosedisaccharides,thetrisacchar- ideMan(a1-3)[Man(a1-6)]Man9andtwosyntheticdiva- lentligandswitheitherarigid(10)oraflexiblelinker(11) (Figure 4a).ITC measurements indicatedthatonly the rigid divalent ligand 10 had a significantly increased binding affinity over a-methyl mannoside and induced clusteringofthelectin(proteinmonomer/ligandratio2:1) whereasthetrimannoside9wasatentimesweakerbinder andboundina1:1stoichiometrywithoutinducingclus- tering.Co-crystallizationoftheproteinwithtrisaccharide 9 required several weeks. Surprisingly, the high-resol- utionstructurerevealedthatthetrisaccharideexistsinan stretchedconformation inwhich itcanbridgetwoadja-

cent protein dimers leading to infinite chains (protein monomer/ligand ratio 2:1) (Figure 4a). Obviously, the different time scales of the ITC experiment and the crystallization process are the reason for the different bindingmode.

Schwefeletal.studiedtheinteractionofseveraldi-,tri-, and tetravalent ligands with wheat germ agglutinin (WGA) by an ELLA and by X-ray crystallography [43]. WGA exists as a dimer with eight binding sites (fourprimarywithhighaffinityandfoursecondary)forN- acetylglucosamine (GlcNAc). Divalent 12 (Figure 4b) showed a remarkable multivalency effect (1170-fold binding enhancement per sugar). X-ray crystallography of the complex of WGA and the second best, closely relatedligand13explainedthisaffinity.Fourmolecules of13simultaneouslybindtotheproteinwitheachligand bridging pairs of adjacent binding sites. This structure showedforthefirsttimethatalleightbindingsitesofthe WGA dimer are simultaneously functional. In contrast, thecrystalstructureofWGAincomplexwithmonovalent GlcNAc derivative 14 revealed that only the primary (high-affinity) binding sites were occupied. Thus, the increasedbindingaffinityofadivalent,chelating ligand providesameansfortheidentificationofsecondary(low- affinity) binding sites. Reported was also tetravalent neoglycopeptide15withabindingpotency25,500times higherthanthatofGlcNAc(6400timespersugar).X-ray crystallographyrevealedthat15bindstoWGAwithtwo sugarsinachelatingbindingmode.ComparisonoftheX- raystructurewiththeNMR-derivedsolutionstructureof 15suggeststhatthepeptideispreorganizedinsolutionin awaysupportingchelatingbinding.

ApplicationofEPRspectroscopy

Sincebindingmechanismsofmultivalentinteractionsina denselypackedcrystalmaydifferfromthoseinsolution, Braun et al. employed distance measurements in the nanometer range by using EPR spectroscopy of spin- labeled WGA ligands to study multivalent binding in (frozenglassy)solution[28].Doubleelectron electron resonance techniques (DEER or PELDOR) delivered distance distributions between two nitroxide labels at oppositeendsofdivalentligands16(Figure4c).Analyses ofthedistancedistributionsshowedadetailedpictureof the binding mechanisms of the divalent ligands. For example, a ligand with a flexible linker gave rise to a broaddistance distributionwhenfreeinsolution. Upon chelatingbindingtoaprotein,itsflexibilitywasreduced leadingto anarrowerdistancedistributionwith amaxi- mum indicative for the distance between the binding sites (cf. Figure 2c). In this way, chelating binding is

987

(Figure4Legend)MultivalentlectinligandsusedincrystallographicandEPRspectroscopicinvestigations.(a)SyntheticligandsforBC2L Aand crystalstructureof9incomplexwiththeprotein(PDBID:2WRA).(b)Mono ,di ,andtetravalentWGAligandsandcrystalstructureof13incomplex withtheprotein(PDBID:2X52).(c)Spin labeledWGAligandsforEPRspectroscopicdistancemeasurements.

(7)

directly detectedandcanbedifferentiated frommono- valentbindingofmultiplemolecules.Inaddition,itwas possibletouseincreasingconcentrationsofspin-labeled monovalentligand17toeffectsuccessivebindingtothe multivalent protein. This approach allowed to measure distancesbetweenbindingsitesandtodetectdifferences inthebindingaffinitiesofindividualbindingsites.

Conclusions

Thegenerationofhigh-affinitymultivalentlectinligands isincreasinglyrecognizedaspromisingapproachtoobtain drugs for therapeutic and diagnostic applications.

Although a wealth of scaffolds has been elegantly designedandmanyoftheobtainedligandsshowtremen- dousbindingenhancements overtheirmonovalentcon- stituents,theunderlyingmechanismsoftheaffinitygain areoftennotwellunderstood.Itisgenerallyacceptedthat chelatingbindingleadstostrongmultivalencyeffectsand potentially also to enhanced selectivity, especially with rigid, perfectly fitting linkers between the sugar residues. The rational design of such linkers,however, requiresknowledgeof theproteinstructure.Systematic scanningofdistancesbetweenbindingsitesispossibleto someextentwithmolecularrulersthatallowforprecise control of spatial presentation of carbohydrates. Much deeper insight in the structural details of multivalent interactionsisgainedbycrystalstructureanalysis.How- ever,co-crystallizationmaybe hamperedbyprecipitate formation. Therefore,innovativemethodsfor structural investigationofmultivalentinteractionssuchasdistance measurementsbyEPRspectroscopyareurgentlyneeded.

Theywillbeafurthersteptowardadeeperunderstand- ingofmultivalentinteractions.

Acknowledgements

TheauthorthankstheDeutscheForschungsgemeinschaft,theUniversity ofKonstanz,andtheKonstanzResearchSchoolChemicalBiologyfor continuoussupport.

References andrecommendedreading

Papersofparticularinterest,publishedwithintheperiodofreview, havebeenhighlightedas:

ofspecialinterest ofoutstandinginterest

1. VarkiA,CummingsRD,EskoJD,FreezeHH,StanleyP,Bertozzi CR,HartGW,EtzlerME(Eds): EssentialsofGlycobiology.Cold SpringHarbor,NY:ColdSpringHarborLaboratoryPress;2009.

2. WangB,BoonsG J(Eds): CarbohydrateRecognition:Biological Problems,Methods,andApplications.Hoboken:JohnWiley&

Sons;2011.

3. GabiusH J(Ed): TheSugarCode.Weinheim:Wiley VCH;2009.

4. FastingC,SchalleyCA,WeberM,SeitzO,HechtS,KokschB, DerneddeJ,GrafC,KnappE W,HaagR:Multivalencyasa chemicalorganizationandactionprinciple.AngewChemIntEd 2012,51:10472 10498.

5. MammenM,ChoiS K,WhitesidesGM:Polyvalentinteractions inbiologicalsystems:implicationsfordesignanduseof multivalentligandsandinhibitors.AngewChemIntEd1998, 37:2755 2794.

6. LisH,SharonN:Lectins:carbohydrate specificproteinsthat mediatecellularrecognition.ChemRev1998,98:637 674.

7. LeeRT,LeeYC:Affinityenhancementbymultivalentlectin carbohydrateinteraction.GlycoconjJ2000,17:543 551.

8. KrishnamurthyVM,EstroffLA,WhitesidesGM:Multivalencyin liganddesign.In Fragment BasedApproachesinDrug Discovery.EditedbyJahnkeW,ErlansonDA.Weinheim:Wiley VCH;2006:11 53.

9. WittmannV,PietersRJ:Bridginglectinbindingsitesby multivalentcarbohydrates.ChemSocRev2013,42:4492 4500.

10. DamTK,BrewerCF:Effectsofclusteredepitopesinmultivalent ligand receptorinteractions.Biochemistry2008,47:8470 8480.

11. ChabreYM,RoyR:Designandcreativityinsynthesisof multivalentneoglycoconjugates.AdvCarbohydrChem Biochem2010,63:165 393.

12. DeniaudD,JulienneK,GouinSG:Insightsintherationaldesign ofsyntheticmultivalentglycoconjugatesaslectinligands.Org BiomolChem2011,9:966 979.

13. LundquistJJ,TooneEJ:Theclusterglycosideeffect.ChemRev 2002,102:555 578.

14. KiesslingLL,GrimJC:Glycopolymerprobesofsignal transduction.ChemSocRev2013,42:4476 4480.

15.

RenaudetO,RoyR:Multivalentscaffoldsinglycoscience:an overview.ChemSocRev2013,42:4515 4520.

Aneditorialoverviewofthethemedcollection‘‘Multivalentscaffoldsin glycosciences’’.Thearticlesofthecollectionreferredtointhisoverview provideacomprehensiveselectionofcurrentlyusedcarbohydratedis plays.

16. HunterCA,AndersonHL:Whatiscooperativity? AngewChem IntEd2009,48:7488 7490.

17. ErcolaniG,SchiaffinoL:Allosteric,chelate,andinterannular cooperativity:amiseaupoint.AngewChemIntEd2011, 50:1762 1770.

18. CecioniS,PralyJ P,MatthewsSE,Wimmerova´ M,ImbertyA, VidalS:Rationaldesignsynthesisofoptimizedglycoclusters formultivalentlectin carbohydrateinteractions:influenceof thelinkerarm.ChemEurJ2012,18:6250 6263.

19. PickensJC,MitchellDD,LiuJ,TanX,ZhangZ,VerlindeCLMJ, HolWGJ,FanE:Nonspanningbivalentligandsasimproved surfacereceptorbindinginhibitorsofthecholeratoxinB pentamer.ChemBiol2004,11:1205 1210.

20. DamTK,BrewerCF:Multivalentprotein carbohydrate interactions:isothermaltitrationmicrocalorimetrystudies.

MethodsEnzymol2004,379:107 128.

21. FreyerMW,LewisEA:Isothermaltitrationcalorimetry:

experimentaldesigndataanalysis,andprobing macromolecule/ligandbindingandkineticinteractions.

MethodsCellBiol2008,84:79 113.

22. Fernandez AlonsoMdC,DiazD,BerbisMA,MarceloF,CanadaJ, Jimenez BarberoJ:Protein carbohydrateinteractionsstudied byNMR:frommolecularrecognitiontodrugdesign.Curr ProteinPeptSci2012,13:816 830.

23. PalmerRA:X raycrystallographicstudiesofprotein ligand interactions.In Protein LigandInteractions:Structureand Spectroscopy.EditedbyHardingSE,ChowdryBZ.Oxford:Oxford UniversityPress;2001:1 98.

24. ShenoySR,BarrientosLG,RatnerDM,O’KeefeBR, SeebergerPH,GronenbornAM,BoydMR:Multisiteand multivalentbindingbetweencyanovirin nandbranched oligomannosides:calorimetricandNMRcharacterization.

ChemBiol2002,9:1109 1110.

25. KitovPI,SadowskaJM,MulveryG,ArmstrongGD,LingH, PannuNS,ReadRJ,BundleDR:Shiga liketoxinsare neutralizedbytailoredmultivalentcarbohydrateligands.

Nature2000,403:669 672.

26. MerrittEA,ZhangZ,PickensJC,AhnM,HolWGJ,FanE:

Characterizationandcrystalstructureofahigh affinity 988

(8)

pentavalentreceptor bindinginhibitorforcholeratoxinandE.

coliheat labileenterotoxin.JAmChemSoc2002,124:8818 8820.

27. ZhangZ,MerrittEA,AhnM,RoachC,HouZ,VerlindeCLMJ, HolWGJ,FanE:Solutionandcrystallographicstudiesof branchedmultivalentligandsthatinhibitthereceptor bindingofcholeratoxin.JAmChemSoc2002,124:

12991 13000.

28.

BraunP,Na¨geleB,WittmannV,DrescherM:Mechanism ofmultivalentcarbohydrate proteininteractionsstudied byEPRspectroscopy.AngewChemIntEd2011,50:

8428 8430.

Thefirst reportontheuseofspin labelEPRspectroscopy tostudy multivalentinteractionsin(frozen)solution.Employingdistancemeasure mentsbetweentwonitroxideswithinsingledivalentligands,chelating bindingcanbedistinguishedfrommonovalentbindingandthedistance betweenbindingsitescanbedetermined.

29. MatsuuraK,HibinoM,IkedaT,YamadaY,KobayashiK:Self organizedglycoclustersalongDNA:effectofthespatial arrangementofgalactosideresiduesoncooperativelectin recognition.ChemEurJ2004,10:352 359.

30. GorskaK,HuangK T,ChaloinO,WinssingerN:DNA templated homo andheterodimerizationofpeptidenucleicacid encodedoligosaccharidesthatmimickthecarbohydrate epitopeofHIV.AngewChemIntEd2009,48:7695 7700.

31. ScheibeC,BujotzekA,DerneddeJ,WeberM,SeitzO:DNA programmedspatialscreeningofcarbohydrate lectin interactions.ChemSci2011,2:770 775.

32. ScheibeC,SeitzO:PNA sugarconjugatesastoolsforthe spatialscreeningofcarbohydrate lectininteractions.Pure ApplChem2012,84:77 85.

33.

ScheibeC,WedepohlS,RieseSB,DerneddeJ,SeitzO:

Carbohydrate PNAandaptamer PNAconjugatesforthe spatialscreeningoflectinsandlectinassemblies.

ChemBioChem2013,14:236 250.

DemonstratesthatDNA PNAconjugatesaresuitedtoscanfordistances between binding sites in multivalent lectins. Identification of a new distance possiblytoasofarunknownsecondarybindingsite has beenachieved.DifferentdensitiesofLselectinassembliesonleukocytes andnanoparticlesweredetected.

34. SchlegelMK,Hu¨tterJ,ErikssonM,LepeniesB,SeebergerPH:

DefinedpresentationofcarbohydratesonaduplexDNA scaffold.ChemBioChem2011,12:2791 2800.

35. PerticiF,PietersRJ:Potentdivalentinhibitorswithrigid glucoseclickspacersforPseudomonasaeruginosalectin LecA.ChemCommun2012,48:4008 4010.

36.

PerticiF,deMolN,KemminkJ,PietersRJ:Optimizingbivalent inhibitorsofPseudomonasaeruginosalectinLecAusinga rigidspacer.ChemEurJ2002,3DOI:10.1002/chem20130346.

SpatialscreeningoflectinligandswithrigidlinkersledtoaLecAinhibitor withunprecedentedbindingenhancement.Itisdemonstratedthatminor structuralvariationsofarigidlinkercanhaveatremendousimpacton inhibitorypotency.

37. WittmannV,SeebergerS:Combinatorialsolid phasesynthesis ofmultivalentcyclicneoglycopeptides.AngewChemIntEd 2000,39:4348 4350.

38. WittmannV,SeebergerS:Spatialscreeningofcyclic neoglycopeptides:identificationofpolyvalentwheat germ agglutininligands.AngewChemIntEd2004,43:900 903.

39. Andre´ S,MaljaarsCEP,HalkesKM,GabiusH J,KamerlingJP:

Discoveryofgalectinligandsinfullyrandomized combinatorialone bead one compound(glyco)peptide libraries.BioorgMedChemLett2007,17:793 798.

40. MacPhersonIS,TemmeJS,HabeshianS,FelczakK,

PankiewiczK,HedstromL,KraussIJ:Multivalentglycocluster designthroughdirectedevolution.AngewChemIntEd2011, 50:11238 11240.

41. ReymondJ L,BergmannM,DarbreT:Glycopeptidedendrimers asPseudomonasaeruginosabiofilminhibitors.ChemSocRev 2013,42:4814 4820.

42.

LameignereE,ShiaoTC,RoyR,WimmerovaM,DubreuilF, VarrotA,ImbertyA:Structuralbasisoftheaffinityfor oligomannosidesandanalogsdisplayedbyBC2L A,a Burkholderiacenocepaciasolublelectin.Glycobiology2010, 20:87 98.

X raycrystallographyunraveledthemechanismforbindingofanoligo mannosidetoBC2L A.Interestingly,crystallizationoccurredonlywithin weeksandresultedininfinitechainsoflectindimersthatareconnected by the mannoside acting as a bridge, whereas ITC measurements indicatedadifferentstoichiometryofthecarbohydrate proteincomplex.

Thisdemonstratedthatthetimescaleofamultivalentinteractionisan importantparameter.

43.

SchwefelD,MaierhoferC,BeckJG,SeebergerS,DiederichsK, Mo¨llerHM,WelteW,WittmannV:Structuralbasisofmultivalent bindingtowheatgermagglutinin.JAmChemSoc2010, 132:8704 8710.

Athoroughcharacterizationofthemultivalentinteractionofwheatgerm agglutinin withaseriesofsynthetic ligandscombining enzyme linked lectinassays,X raycrystallography,andNMRspectroscopy.Alleight bindingsitesoftheproteincouldbeshowntobesimultaneouslyfunc tional.Thehighaffinityofacyclicneoglycopeptidewasrationalizedbya conformationalpreorganizationofthecyclopeptidescaffoldsupporting multivalentbinding.

989

Referenzen

ÄHNLICHE DOKUMENTE

We have shown how both rigorous nonequilibrium free energy calculations based on MD simulations, as well as the latest Rosetta protocols, are able to quantitatively predict changes

To obtain information about the overall stability and the changes in amino- acid mobility of the protein between the ligand-bound and -unbound systems, we calculated the

• Different sensitivity to ions and regulation of nucleotide binding in hippo- campal and cortical membranes proposes that 5-HT 1A receptors in these tissues are coupled to

Assuming that no large-scaled structural changes occur upon ligand-binding that would lead to the desolvation or solvation of protein atoms far from the binding site,

Moderate to fast

If the structure of a target receptor is available, information about the binding-site and principles of protein-ligand interactions can be used to estimate the binding affinity of

Comparison of the X-ray structure and the solution NMR structure as well as comparison of the IC 50 values of dipeptide 28 and divalent ligand 23 (Table 1) suggests that the

After processing the reads as done for the LL36 iCLIP targets, we arrived at 469 transcripts with significant XL sites in at least two of the three AtGRP7::AtGRP7-GFP grp7-1