MissionBring together scientists from the School of Medicine, the College of Arts & Science, and theSchool of Engineering who create a complex pipeline that integrates structural biology with bigdata for the discovery of novel antibodies that will spur the development of next-generationvaccines.

Aims• FormpipelinebetweentheVanderbiltVaccineCenter(VVC) andtheCenterforStructuralBiology(CSB) thatcharacterizesthemostpromisingnovelantibodiesintermsofthestructuraldeterminantsthatdriveneutralizationofpathogenicviruses

• IntegratebigdataandcomputationalengineeringwiththeVVCandCSBtoengineernovelantibodiesandvaccines

•Graduateandundergraduatestudenttraininginintegratingstructuralbiologywithbigdataforthedesignofnextgenerationvaccines

VanderbiltProgramforNextGenerationVaccines

JensMeiler (PI)Professor,ChancellorFacultyFellowDepartmentsofChemistry,Pharmacology,andBiomedicalInformaticsCenterforStructuralBiology,InstituteforChemicalBiologyjens.meiler@vanderbilt.edu

WalterChazin,Ph.D.Chancellor'sProfessorandIngramProfessorofCancerResearchDepartmentofBiochemistryandChemistryDirectorofCenterforStructuralBiology

MelanieOhi,Ph.D.AssociateProfessorDepartmentofCellandDevelopmentalBiologyandCenterforStructuralBiology

BenjaminSpiller,Ph.D.AssociateProfessorDepartmentsofPharmacologyandPathology,MicrobiologyandImmunology

KevinLSchey,Ph.D.ProfessorDepartmentofBiochemistryDirectorofMassSpectrometryCore

JamesECroweJr,M.D.,(co-PI)ProfessorandAnnScottCarellChairDepartmentsofPediatricsandPathology,MicrobiologyandImmunologyDirectorofVanderbiltVaccineCenterjames.crowe@vanderbilt.edu

Ivelin Georgiev,Ph.D.AssistantProfessorDepartmentofPathology,MicrobiologyandImmunology

Yevgeniy Vorobeychik,Ph.D.AssistantProfessorDepartmentofElectricalEngineeringandComputerScience

JamesC.Slaughter,Ph.D.AssociateProfessorDepartmentofBiostatistics

DanielFabbri,Ph.D.AssistantProfessorDepartmentofBiomedicalInformatics

FacultyParticipants

- Overview-IntegratingStructuralBiologywithBigData

Opportunitiesexistfor:•Undergraduatesummerresearchinternships•Graduatestudents•Post-Doctoralfellows

Inquiries:Dr.JensMeiler – jens.meiler@vanderbilt.edu

Dr.JamesE.Crowe– james.crowe@vanderbilt.edu

StipendsAvailable!

Structural Biology of Poxvirus Protein and Human Antibody Interaction

q The initial focus – determining structures of neutralizingantibodies to orthopoxviruses bound to antigen by using X-raycrystallography and single particle electron microscopy.

q Computational approaches will be used to dock highresolution crystal structures of the complexed antibody andantigen into the lower resolution electron density map.

q The long term goal is to be able to computationally predict thestructure of the antibody-antigen complex once the sequenceof the antibody is known.

Crystal of poxvirus protein D8 diffracted to 2.4 Å(below)

Amritraj PatraMentors:MelanieD.Ohi,JensMeiler,andJamesE.CroweJr.

Human Neutralizing Antibodies for NorovirusGoals:

– ProvideadetailedunderstandingofthehumanhumoralresponsetoNorovirusinfection– DefinethemolecularandstructuralbasisforinhibitionofNorovirusbyhumanantibodies.

Norovirus(NoV)– Leadingcauseofsporadicandepidemicgastroenteritisinhumans.– Novaccines,therapeutics,orprophylacticsavailabletopreventortreatinfection– Vaccinedesigndifficultduetoantigenicvariationwithin&betweenNoV genogroups

Gabriela AlvaradoMentors: James E. Crowe Jr., Ivelin Georgiev

Weintendtocharacterizethree-dimensionallywherebroadlybinding&blockinghumanmonoclonalantibodiesbindtocurrentstrainsofNoV.

Delineationofpolyclonalantibodyspecificitiesinserafromdonorsinfectedbyorvaccinatedagainstvirusesofbiomedicalsignificance

Nagarajan RajuMentors:Ivelin Georgiev,JamesE.CroweJr.

Wearedevelopinganovelalgorithmforsimulatingantibodyneutralizationfingerprints.PreliminaryanalysisshowsthatthealgorithmcansuccessfullyrecapitulateknownHIV-1antibodyfingerprints.Ultimately,thealgorithmwillbeusedtodesignandidentifynovelfingerprintsandepitopes.

Ø Computationalneutralizationfingerprintinguseslarge-scaleantibody-antigenfunctionaldatatodelineatethecomplexepitopespecificitiesfoundinpolyclonalresponsestoinfectionandvaccination.

Ø Thegoalistodevelopnext-generationneutralizationfingerprintingtechnologyforidentifyingnovel,previouslyuncharacterized,antibodyspecificitiesinpolyclonalresponsesfrominfectionorvaccination

Ø ApplicationtoHIV-1anddengue

Designingantibodyfingerprintsfortargetingnovelantigenepitopes

ConservationofdengueantibodyepitopesonproteinE

Nagarajan RajuMentors:Ivelin Georgiev,JamesE.CroweJr.

• Maximumantibody-interactiveareaforeachantigenresidueisplottedasaheatmap,cappedat100Å2

• Darker color corresponds toantigenresidues that are asubstantial partof atleast one antibodyepitope

Computational Antibody Design• Computational design in the combinatorial antibody and virus sequence space• New paradigm for designing antibodies with increased breadth, high potency

and high similarity to antibodies in the human repertoire• Goal: to develop highly scalable methods for sequence-based design of

antibodies for HIV that exhibit very broad recognition of diverse isolates• Algorithms: combination of game theoretic modeling, computational

structure design, scalable sequence search, and scalable bi-level combinatorialoptimization

H

L L

HH

G G

The native VRC01-GP120 complex (left) andgame-theoretically designed antibody incomplex (right). GP120 needs at least 7mutations to escape binding to the designedantibody, as compared to just 1 mutation incase of VRC01.

Swetasudha PandaMentors: Yevgeniy Vorobeychik, Jens Meiler

Broadly Binding Antibody Design

Swetasudha PandaMentors: Yevgeniy Vorobeychik, Jens Meiler

Game-Theoretic Antibody Design

Swetasudha PandaMentors: Yevgeniy Vorobeychik, Jens Meiler

Structure Based Antibody and Antigen Designfor Sialic Acid Binding Domains

• Goal - to characterize sialic acid (SIA) binding in proteins by looking at the electrochemical environment as well as the stereochemical conformation bias inherent in the SIA carbohydrate

• Computational screening and model building done using EON and ROCS– Results show high levels of homology between hemaggluttinin and neuraminidase SIA binding pockets

• Experimental testing of predicted trends to be run using ELISA and biolayerinterferometry

Clayton AndersonMentors: James E. Crowe Jr., Jens Meiler

1.Detectionofsequenceliabilities 3.Guidedcomputationalantibodydesign 2.Detectionofstructureliabilities

Humanization,whichisalsoreferredtoasreshaping,complementaritydeterminingregion(CDR)-grafting,veneering,resurfacing,specificity-determiningresidue(SDR)-transfer,orDeImmunizationTM,comprisesstrategiesforreducingtheimmunogenicityofmonoclonalantibodies(mAbs)fromanimalsourcesandforimprovingtheiractivationofthehumanimmunesystem.

Awebserverwascreatedtodetectandscore13commonliabilities,whicharelikelytocausecomplicationsforelicitationasavaccine(Figureabove).Thecurrentexperiencebasedscoringmodelwillbereplacedbyastatisticalapproach,theBayesestimator,whichenablesharvestofthehugesequencepoolofNextGenerationSequencing(NGS).Forthispurpose,theCroweLab antibodysequencedatabasewillbeharvestedtocollectmutationinformationandfrequenciesofhumanantibodiesextractedfromrealpatients.Theknowledgeaboutinvivomutationswillallowbetterevaluationofpossiblesequenceliabilities.Theadditionofdatafromongoingstudies,maybeincorporatedintothemodelusingBayesianinference.

Residuepaircouplingsextractedfromnaturalhomologoussequencesgiveinsightintotheevolutionofproteinsandencodefunctionalandstructuralliabilitiesofagivenproteinfamily.

Incorporationofresiduepaircouplingshasproventobeasuccessfulstrategytoenhancecomputationalproteindesign.

Usingactualevolutionarydatamaycompensateforthelimitationsofcommondesigntechniques,attributabletoasimplifiedmolecularforcefieldorlackingknowledgeaboutproteinfunctionalityduringthedesignprocess.

Aphysicalmodel,theIsing modeliscommonlyusedtoestimateresiduepaircouplingsinsequencedatagremlin.bakerlab.org.Aspartofmywork,IwillupdatethecouplingpredictionbyreplacingtheIsing modelbythePottsmodel.

Forthispurpose,I’vebeguncollaboratingwithoneofthedevelopersofGremlin,SergeyOvchinnikov.

ROSETTA,asoftwaresuiteforcomputationalproteindesign,iscapabletodockanddesignantibodiestospecifictargets.Therearetwomajordrawbackswhichcomewiththecurrentversionofthisprocedure.First,itrequiresasignificantamountofcomputationalresourcesbynaivelysamplingrandomsequencemodificationsfromthehugevarietyofpossiblesequences.Second,ROSETTAisguidedbystructuralstability,butomitsfunctionalandstructuralconstraints.Asystemtoincorporateknowledgeontheseconstraintscouldcircumventantibodysequenceswhichwouldcauseimmunogenicityeffectsorexhibitlowexpressionratesinthemodelsystem.

Toovercometheseissues,sequenceandstructureliabilities(goal1and2)canbeincorporatedintotheROSETTADesignprocess.Thisadditionalinformationwillreducethesequencespacetobesampleddramaticallyandsavecomputationalresources.Furthermore,thedesignsareexpectedtobemorehuman-like,whichwillreducethepostprocessingeffortdramatically.

Withthesavedcomputationalresources,designonmultipleconformations(multistatedesign)willbecomeevenmorefeasibleandallowmultispecifity ofantibodiesforbroadlyneutralizingvaccines.

Samuel SchmitzMentors: Jens Meiler, James E. Crowe Jr.

Restricting the Sequence Space in Rosetta Antibody Design to Human-like Antibodies

Symposium on Modeling Immunity

April 27th1220 MRBIII

• 12:30pm lunch – Attendees must register by April 7th

• 1:00-4:30pm – 5 confirmed speakers

• Frank DiMaio – University of Washington• Sarel Fleishman – Weizmann Institute of Science• Steven Kleinstein – Yale Center for Medical Informatics• Jens Meiler – Vanderbilt University• Andrew Ward – The Scripps Institute

– Coffee/tea break

• 4:30-5:30pm poster session and reception

• https://my.vanderbilt.edu/modelingimmunity/eventpage/