Article Structural and Functional Basis of SARS-CoV-2 Entry by Using Human ACE2 Graphical Abstract Authors QihuiWang,YanfangZhang,LiliWu,..., HuanZhou,JinghuaYan,JianxunQi Correspondence [email protected](J.Y.), [email protected](J.Q.) In Brief ThecrystalstructureoftheC-terminal domainoftheSARS-CoV-2spikeprotein incomplexwithhumanACE2reveals insightsintothemechanismsofbinding ofthisvirusanditsdifferences fromSARS. Highlights d SARS-CoV-2interactswithhACE2viaSproteinCTD d A2.5-A˚ structureofSARS-CoV-2-CTDincomplexwith hACE2isresolved d TheSARS-CoV-2-CTDdisplaysstrongeraffinityforhACE2 comparedwithSARS-RBD d SARS-CoV-2-CTDisantigenicallydifferentfromSARS-RBD Wangetal.,2020,Cell181,894–904 May14,2020ª2020ElsevierInc. ll https://doi.org/10.1016/j.cell.2020.03.045 ll Article Structural and Functional Basis of SARS-CoV-2 Entry by Using Human ACE2 QihuiWang,1,2,3,15YanfangZhang,3,4,5,15LiliWu,1,4,15ShengNiu,3,6,15ChunliSong,1,7,15ZengyuanZhang,3,4 GuangwenLu,8ChengpengQiao,1YuHu,3,9Kwok-YungYuen,10,11QishengWang,12HuanZhou,12JinghuaYan,1,2,3,7,13,* andJianxunQi3,14,16,* 1CASKeyLaboratoryofMicrobialPhysiologicalandMetabolicEngineering,InstituteofMicrobiology,ChineseAcademyofSciences,Beijing 100101,China 2ShenzhenKeyLaboratoryofPathogenandImmunity,ShenzhenThirdPeople’sHospital,Shenzhen518112,China 3CASKeyLaboratoryofPathogenicMicrobiologyandImmunology,InstituteofMicrobiology,ChineseAcademyofSciences,Beijing 100101,China 4UniversityoftheChineseAcademyofSciences,Beijing100049,China 5LaboratoryofProteinEngineeringandVaccines,TianjinInstituteofBiotechnology,Tianjin300308,China 6CollegeofAnimalScienceandVeterinaryMedicine,ShanxiAgriculturalUniversity,Taigu030801,China 7InstituteofPhysicalScienceandInformation,AnhuiUniversity,Hefei230039,China 8WestChinaHospitalEmergencyDepartment(WCHED),StateKeyLaboratoryofBiotherapyandCancerCenter,WestChinaHospital, SichuanUniversity,andCollaborativeInnovationCenterofBiotherapy,Chengdu,Sichuan610041,China 9SchoolofLifeSciences,UniversityofScienceandTechnologyofChina,Hefei,Anhui230026,China 10StateKeyLaboratoryforEmergingInfectiousDiseases,TheUniversityofHongKong,Pokfulam,HongKongSpecialAdministrativeRegion 999077,China 11DepartmentofMicrobiology,TheUniversityofHongKong,Pokfulam,HongKongSpecialAdministrativeRegion999077,China 12ShanghaiSynchrotronRadiationFacility,ShanghaiAdvancedResearchInstitute,ChineseAcademyofSciences,Shanghai201204,China 13CollegeofLifeScience,UniversityoftheChineseAcademyofSciences,Beijing100049,China 14SavaidMedicalSchool,UniversityoftheChineseAcademyofSciences,Beijing100049,China 15Theseauthorscontributedequally 16LeadContact *Correspondence:[email protected](J.Y.),[email protected](J.Q.) https://doi.org/10.1016/j.cell.2020.03.045 SUMMARY The recent emergence of a novel coronavirus (SARS-CoV-2) in China has caused significant public health concerns.Recently,ACE2wasreportedasanentryreceptorforSARS-CoV-2.Inthisstudy,wepresentthe crystalstructureoftheC-terminaldomain ofSARS-CoV-2(SARS-CoV-2-CTD)spike(S)protein incomplex withhumanACE2(hACE2),whichrevealsahACE2-bindingmodesimilaroveralltothatobservedforSARS- CoV.However,atomicdetailsatthebindinginterfacedemonstratethatkeyresiduesubstitutionsinSARS- CoV-2-CTD slightly strengthen the interaction and lead to higher affinity for receptor binding than SARS- RBD. Additionally, a panel of murine monoclonal antibodies (mAbs) and polyclonal antibodies (pAbs) againstSARS-CoV-S1/receptor-bindingdomain(RBD)wereunabletointeractwiththeSARS-CoV-2Spro- tein, indicating notable differences in antigenicity between SARS-CoV and SARS-CoV-2. These findings shed light on the viral pathogenesis and provide important structural information regarding development of therapeutic countermeasures against the emerging virus. INTRODUCTION (Wu et al., 2020; Zhou et al., 2020; Zhu et al., 2020). Later, the International Committee on Taxonomy of Viruses (ICTV) Emerging and re-emerging viruses are a significant threat to officially designated the virus SARS-CoV-2 (Coronaviridae global public health (Gao, 2018). Since the end of 2019, Chi- Study Group of the International Committee on Taxonomy of nese authorities have reported a cluster of human pneumonia Viruses, 2020), although many virologists argue that HCoV-19 casesinWuhanCity,China(Wangetal.,2020),andthedisease is more appropriate (Jiang et al., 2020). As of February 24, was designated coronavirus disease 2019 (COVID-19). These 2020, 79,331 laboratory-confirmed cases have been reported cases showed symptoms such as fever and dyspnea and to the WHO globally, with 77,262 cases in China, including were diagnosed as viral pneumonia (Tan et al., 2020; Zhu 2,595deaths(https://www.who.int/).Inaddition,29othercoun- et al., 2020). Whole-genome sequencing results showed that tries have confirmed imported cases of SARS-CoV-2 infection the causative agent was a novel coronavirus that was initially (https://www.who.int/), raising great public health concerns named 2019-nCoV by the World Health Organization (WHO) worldwide. 894 Cell181,894–904,May14,2020ª2020ElsevierInc. ll Article SARS-CoV-2 is the seventh coronavirus that is known to (Lai et al., 2007). S1 can be further divided into an N-terminal causehumandisease.Coronaviruses(CoVs)areagroupoflarge domain (NTD) and a C-terminal domain (CTD), both of which andenvelopedviruseswithpositive-sense,single-strandedRNA canfunctionasareceptor-bindingentity(e.g.,SARS-CoVand genomes(Laietal.,2007;LuandLiu,2012).Thevirusescanbe MERS-CoV utilize the S1 CTD to recognize the receptor (also classifiedintofourgenera:alpha,beta,gamma,anddeltaCoVs calledreceptorbindingdomain[RBD])(Lietal.,2005;Luetal., (Woo et al., 2009; https://talk.ictvonline.org/). Previously 2013), whereas mouse hepatitis CoV engages the receptor identified human CoVs that cause human disease include the with its S1 NTD (Taguchi and Hirai-Yuki, 2012)). The region in alphaCoVs hCoV-NL63 and hCoV-229E and the betaCoVs SARS-CoV-2SproteinthatisresponsibleforhACE2interaction HCoV-OC43, HKU1, severe acute respiratory syndrome CoV remainsunknown. (SARS-CoV), and Middle East respiratory syndrome CoV Inthisstudy,byutilizingimmunostainingandflowcytometryas- (MERS-CoV)(Luetal.,2015;WeversandvanderHoek,2009). says,wefirstidentifytheS1CTD(SARS-CoV-2-CTD)asthekey Both alphaCoVs and the betaCoVs HCoV-OC43 and HKU1 region in SARS-CoV-2 that interacts with the hACE2 receptor. cause self-limiting common cold-like illnesses (Chiu et al., Wesubsequentlysolveda2.5-A˚ crystalstructureofSARS-CoV- 2005; Gorse et al., 2009; Jean et al., 2013; Jev(cid:1)snik et al., 2-CTDincomplexwithhACE2,whichrevealsareceptor-binding 2012).However,SARS-CoVandMERS-CoVinfectioncanresult mode similar overall to that observed for the SARS-CoV RBD inlife-threateningdiseaseandhavepandemicpotential.During (SARS-RBD).However,SARS-CoV-2-CTDformsmoreatomicin- 2002–2003, SARS-CoV initially emerged in China and swiftly teractions with hACE2 than SARS-RBD, which correlates with spread to other parts of the world, causing more than 8,000 datashowinghigheraffinityforreceptorbinding.Notably,apanel infections and approximately 800 related deaths worldwide of monoclonal antibodies (mAbs) as well as murine polyclonal (WHO, 2004). In 2012, MERS-CoV was first identified in the antisera against SARS-S1/RBD were unable to bind to the MiddleEastandthenspreadtoothercountries(Ksiazeketal., SARS-CoV-2Sprotein,indicatingnotabledifferencesinantige- 2003;Zakietal.,2012).AsofNovember2019,atotalof2,494 nicity between SARS-CoV and SARS-CoV-2, suggesting that MERS cases with 858 related deaths have been recorded in thepreviouslydevelopedSARS-RBD-basedvaccinecandidates 27 countries globally (https://www.who.int/emergencies/ areunlikelytobeofanyclinicalbenefitforSARS-CoV-2prophy- mers-cov/en/). Notably, new cases of MERS-CoV infecting laxis.Takentogether,thesedatashedlightonviralentryandpath- humans are still being reported (https://www.who.int/csr/don/ ogenesisandwillhopefullyinspirenewtargetedtreatmentsfor archive/disease/coronavirus_infections/en/). SARS-CoV and thisemergingpathogen. MERS-CoV are zoonotic pathogens originating from animals. Detailed investigations indicate that SARS-CoV is transmitted RESULTS from civet cats to humans and MERS-CoV from dromedary camels to humans (Azhar et al., 2014; Ge et al., 2013; Guan SARS-CoV-2AppliestheCTDtoInteractwithhACE2 etal.,2003).ThesourceofSARS-CoV-2,however,isstillunder Throughbioinformaticsanalysis,theSARS-CoV-2Sproteinhas investigationbutlinkedtoawetanimalmarket(Zhuetal.,2020; beenshowntodisplaycharacteristicCoVSfeatures,includinga The2019-nCoVOutbreakJointFieldEpidemiologyInvestigation S1regioncontainingtheNTDandCTD,S2,atransmembranere- TeamandLi,2020). gion,andashortcytoplasmicdomain(FigureS1A).Phylogenetic Virus infections initiate with binding of viral particles to host studiesrevealthatSARS-CoV-2belongstoagroupcontaining surfacecellularreceptors.Receptorrecognitionisthereforean SARS-CoVaswellastwobat-derivedSARS-likeviruses,ZC45 importantdeterminantofthecellandtissuetropismofavirus. andZCX21(FiguresS1B–S1D).Recently,hACE2wasreported Inaddition,thegainoffunctionofavirustobindtothereceptor tobethereceptorofSARS-CoV-2(Zhouetal.,2020).Because counterpartsinotherspeciesisalsoaprerequisiteforinter-spe- SARS-CoV utilizes its S1 CTD, otherwise known as the RBD, cies transmission (Lu et al., 2015). Interestingly, with the to recognize the same receptor, we decided to test whether exceptionofHCoV-OC43andHKU1,bothofwhichhavebeen the CTD in SARS-CoV-2 is also the key region for interaction shown to engage sugars for cell attachment (Li et al., 2005), withitsreceptorhACE2. theotherhumanCoVsrecognizeproteinaceouspeptidasesas WepreparedaseriesofmouseFc(mFc)-fusedSARS-CoV-2S receptors. HCoV-229E binds to human aminopeptidase N protein preparations, including S1 (SARS-CoV-2-S1), the NTD (hAPN) (Li et al., 2019), and MERS-CoV interacts with human (SARS-CoV-2-NTD),andtheCTD,andsubsequentlyvisualized dipeptidyl peptidase 4 (hDPP4 or hCD26) (Lu et al., 2013; Raj theirbindingtoGFP-taggedhACE2expressedonthecellsur- et al., 2013). Although they belong to different genera, face via confocal fluorescence microscopy. As a control, we SARS-CoV and hCoV-NL63 interact with human angiotensin- also prepared the Fc fusion proteins for SARS-RBD and converting enzyme 2 (hACE2) for virus entry (Hofmann et al., MERS-RBD and tested these in parallel with the SARS-CoV-2 2005; Li et al., 2003; Wu et al., 2009). After the outbreak of proteins.Asexpected,SARS-RBDshowedco-localizationwith COVID-19,ChinesescientistspromptlydeterminedthatSARS- hACE2andMERS-RBDwithhCD26.ForthenovelCoVproteins, CoV-2alsoutilizeshACE2forcellentry(Zhouetal.,2020). SARS-CoV-2-S1 and SARS-CoV-2-CTD co-localized with In CoVs, the entry process is mediated by the envelope- hACE2onthecellsurface.TheSARS-CoV-2-NTDprotein,how- embedded surface-located spike (S) glycoprotein (Lu et al., ever,wasincapableofbindingtohACE2.Inaddition,noneofthe 2015).ThisSproteinwould,inmostcases,becleavedbyhost SARS-CoV-2proteinsinteractedwithhCD26(Figure1). proteases into the S1 and S2 subunits, which are responsible Wefurther tested binding oftheviralproteinsto cell-surface for receptor recognition and membrane fusion, respectively hACE2 via flow cytometry. Consistently, SARS-CoV-2-S1 and Cell181,894–904,May14,2020 895 ll Article Figure1. SARS-CoV-2-S1andSARS-CoV-2-CTDCo-localizewithhACE2 HEK293TcellsweretransfectedwithpEGFP-N1-hACE2(leftpanels,hACE2-GFP)orpEGFP-C1-hCD26(rightpanels,hCD26-GFP).Twenty-fourhourslater,the cellswereincubatedwithsupernatantcontainingmFc-taggedSARS-CoV-2-S1(SARS-CoV-2-S1-mFc),SARS-CoV-2-NTD(SARS-CoV-2-NTD-mFc),SARS- CoV-2-CTD(SARS-CoV-2-CTD-mFc),MERS-RBD(MERS-RBD-mFc),orSARS-RBD(SARS-RBD-mFc)proteinsandsubsequentlyincubatedwithanti-mouse IgG(mIgG)antibodyconjugatedwithA594(anti-mIgG/A594).NucleiwerestainedwithDAPI.AllimageswereobtainedbyconfocalmicroscopyusingaLeicaSP8 (3100oilimmersionobjectivelens).Thescalebarineachpanelindicates8mm.Thedatashownarerepresentativeoftwoindependentexperiments. SeealsoFigureS2. SARS-CoV-2-CTD,butnotSARS-CoV-2-NTD,showedstrongaf- structural domains (Han et al., 2017). One is the conserved finityforhACE2(FigureS2A).NoneofthenovelCoVproteinsinter- coresubdomainwithfiveantiparallelbstrandsandaconserved actedwithhCD26orhAPN(FiguresS2BandS2C).Inaddition,sol- disulfide bond between bc2 and bc4 (Figures 2B and S1D). ublehACE2,butnothCD26orhAPN,wasshowntoinhibitthe The other isthe external subdomain, which is dominated bya interactionbetweenviralproteins,withcellsexpressinghACE2 disulfide bond-stabilized flexible loop that connects two small inadose-dependentmanner(FiguresS2D–S2I).Takentogether, bstrands.ThecomplexstructuredatashowthatSARS-CoV-2- theseresultsclearlydemonstratethatSARS-CoV-2iscapableof CTD utilizes its external subdomain to recognize subdomain I binding,viatheviralCTD,tohACE2. inthehACE2NTD(Figure2A;Towleretal.,2004). Further analysis was performed to identify key residues ComplexStructurebetweenSARS-CoV-2andhACE2 involved incomplex formation. Aminoacidslocated withinthe Wethenattemptedtostudythestructuralbasisofthevirus-re- van der Waals (vdw) contact distance (4.5-A˚-resolution cutoff) ceptor interaction. Weprepared theSARS-CoV-2-CTD/hACE2 between the viral ligand and receptor were selected (Table 2), complex by in vitro mixture of the two proteins and isolated andaseriesofhydrophilicresidueslocatedalongtheinterface complexes via size exclusion chromatography. The complex werefoundtoformasolidnetworkofH-bondandsaltbridgein- structure was solved to 2.5-A˚ resolution (Table 1), with one teractions (Figure 2). These strong polar contacts include the SARS-CoV-2-CTD binding to a single hACE2 molecule in the SARS-CoV-2-CTDresidueA475interactingwithhACE2residue asymmetric unit. For hACE2, clear electron densities could be S19,N487withQ24(Figures2CandS3A),E484withK31,and tracedfor596residuesfromS19toA614oftheN-terminalpepti- Y453withH34(Figures2DandS3B).ResidueK417,locatedin dasedomainaswellasglycansN-linkedtoresidues53,90,and helixa3oftheCTDcoresubdomain,wasshowntocontribute 322(Figure2A). ionicinteractionswithhACE2D30(Figures2DandS3B).Notably, Inthecomplexstructure,theSARS-CoV-2-CTDcontains195 thebulgedloopsinSARS-CoV-2-CTD,thea1’/b1’loopandb20/ consecutivedensity-traceableresiduesspanningT333toP527 h1’loop,properlypositionseveralresidues(G446,Y449,G496, togetherwithN-linkedglycosylationatN343.Similartootherre- Q498, T500, and G502) in close proximity with hACE2 amino ported betaCoV CTD structures, this protein also exhibits two acidsD38,Y41,Q42,K353,andD355,formingaconcentration 896 Cell181,894–904,May14,2020 ll Article Ca atoms (Figure 3A). In comparison with the SARS-RBD, the Table1. DataCollectionandRefinementStatistics majorityofthesecondarystructureelementsarewellsuperim- SARS-CoV-2-CTDandhACE2 posed in SARS-CoV-2-CTD, with the exception of the b10/b20 DataCollection loop, whichshowedthemostsequencevariation betweenthe Spacegroup P41212 twoligands(Figures3AandS1D). CellDimensions TheoverallstructureoftheSARS-CoV-2-CTD/hACE2complex a,b,c(A˚) 104.45,104.45,229.79 isverysimilartothepreviouslyreportedstructureofSARS-RBD a,b,c((cid:2)) 90.00,90.00,90.00 bound to the same receptor with anRMSDof 0.431 A˚ for 669 Resolution(A˚) 50.00–2.50(2.59–2.50) equivalentCaatoms(Lietal.,2005;Figures3A–3C).Consistent withthishighdegreeofsimilarity,thesolubleSARS-RBDblocks Uniquereflections 44,981(43,84) theinteractionbetweentheSARS-CoV-2ligandwithhACE2ina Completeness(%) 100.0(100.0) concentration-dependentmanner(FiguresS2JandS2K).Further Rmerge 0.129(1.147) detailedcomparisonofthereceptorbindinginterfacebetweenthe I/sI 26.7(3.3) twovirusesreveals that,amongthe24residuesinhACE2that CC (%) 0.999(0.867) make vdw contacts with either CTD, 15 amino acids display 1/2 Redundancy 21.6(22.3) morecontactswiththeSARS-CoV-2-CTD(Table2).TheSARS- CoV-2-CTD binding interface also has more residues than Refinement SARS-RBD(21versus17)thatdirectlyinteractwithhACE2,form- Resolution(A˚) 34.50-2.50 ingmorevdwcontacts(288versus213)aswellasH-bonds(16 Numberofreflections 44861 versus11)(Tables2andS1).Consistently,SARS-CoV-2-CTDin Rwork/Rfree 0.1846/0.2142 complex with hACE2 buries larger surface areas than SARS- NumberofAtoms RBD(1773A˚2versus1686A˚2). Protein 6,461 Notably,themostvariableloop(b1’/b20loop)contributessub- Ligand/ion 1 stantiallymorevdwcontactsinSARS-CoV-2-CTDthanforthe SARS-RBD(115versus53)(Figure3D;TableS1).Specifically, Water 322 F486 in SARS-CoV-2, instead of I472 in SARS-RBD, forms B-factors strong aromatic-aromatic interactions with hACE2 Y83, and Protein 44.1 E484 in the SARS-CoV-2-CTD, instead of P470 in the Ligand/ion 38.3 SARS-RBD,formsionicinteractionswithK31(Figure3D). Water 40.4 RMSDs TheInteractionbetweenSARS-CoV-2-CTDandhACE2 Bondlengths(A˚) 0.005 IsSpecificandDisplays4-FoldStrongerAffinity Bondangles((cid:2)) 0.799 ComparedwiththeSARS-RBD In light of the increased atomic interactions between hACE2 RamachandranStatistics(%) with the SARS-CoV-2-CTD compared with the SARS-RBD, Favored 98.60 we speculated that the former should bind to the receptor Allowed 1.02 with stronger affinity than the latter. To test this hypothesis, Disallowed 0.38 we performed real-time surface plasmon resonance (SPR) as- Valuesinparenthesesareforthehighest-resolutionshell. says. The mFc-tagged S-domain proteins were captured by anti-mouse IgG (mIgG) antibodies that were immobilized on the chip and tested for binding with gradient concentrations ofH-bonds(Figures2EandS3C).Furthervirus-receptorcontacts of the soluble ectodomain proteins of hACE2 and hCD26. As include SARS-CoV-2-CTD Y489 and F486 packing against assay controls, SARS-RBD and MERS-RBD were found to hACE2residuesF28,L79,M82,andY83,formingasmallpatch readily interact with their respective canonical receptors (Fig- of hydrophobic interactions at the interface (Figures 2C and ures 4A and 4D). SARS-CoV-2-S1 and SARS-CoV-2-CTD S3A). Overall, the virus-receptor engagement is dominated by bound to hACE2 but not to hCD26 (Figures 4E, 4F, 4I, and polarcontactsmediatedbythehydrophilicresidues.Insupport 4J). The recorded binding profiles revealed typical slow-on/ of this hypothesis, a single K353A mutation was sufficient to slow-off kinetics, as observed with the SARS-CoV and abolishtheseinteractions(FigureS2L). MERS-CoV proteins. The equilibrium dissociation constants (K ) of SARS-CoV-2-S1 and SARS-CoV-2-CTD binding to D ComparisonoftheBindingInterfacesbetweenhACE2/ hACE2 were calculated to be 94.6 ± 6.5 nM and 133.3 ± SARS-CoV-2-CTDandhACE2/SARS-RBD 5.6 nM, respectively. These values represent ~4-fold higher SARS-CoV-2-CTDexhibitssignificantstructuralhomologytoits binding affinities than that observed for the SARS-RBD SARS-CoVhomolog,inagreementwithhighsequenceidentity engaging the same receptor, which was determined to be betweenthetwomolecules(~73.9%)(FigureS1C).Superimpo- 408.7 ± 11.1 nM (Figure 4). Taken together, the increased sition of the SARS-CoV-2-CTD structure onto a previously atomic interactions between the hACE2 and SARS-CoV-2- reported SARS-RBD structure (PDB: 2GHV) revealed a root- CTD binding region leads to ~4-fold higher binding affinity mean-square deviation (RMSD) of 0.475 A˚ for 128 equivalent comparedwiththe SARS-RBD. Cell181,894–904,May14,2020 897 ll Article Figure2. TheComplexStructureofSARS-CoV-2-CTDBoundtohACE2 (A)Acartoonrepresentationofthecomplexstructure.ThecoresubdomainandexternalsubdomaininSARS-CoV-2-CTDarecoloredcyanandorange, respectively.hACE2subdomainIandIIarecoloredvioletandgreen,respectively.Therightpanelwasobtainedbyanticlockwiserotationoftheleftpanelalonga longitudinalaxis.Thecontactingsitesarefurtherdelineatedin(C)–(E)fortheaminoacidinteractiondetails. (B)AcartonrepresentationoftheSARS-CoV-2-CTDstructure.Thesecondarystructuralelementsarelabeledaccordingtotheiroccurrenceinsequenceand locationinthesubdomains.Specifically,thebstrandsconstitutingthecoresubdomainarelabeledwithanextrac,whereastheelementsintheexternalsub- domainarelabeledwithanextraprimesymbol.ThedisulfidebondsandN-glycanlinkedtoN343areshownassticksandspheres,respectively. (C–E)Keycontactsitesaremarkedwiththeleft,middleandrightboxin(A)andfurtherdelineatedforinteractiondetails,respectively.Theresiduesinvolvedare shownandlabeled. SeealsoFiguresS1,S2,andS3andTableS1. SARS-CoV-2ExhibitsDistinctEpitopeFeaturesinthe distinctstructuralcharacteristicsintheRBDexternalsubdomain RBDfromSARS-CoV thatmediatesreceptorbinding(Lietal.,2005;Luetal.,2013).In To conclude, we set out to investigate the epitope features of the positive control, anti-SARS-RBD antibodies, but not anti- SARS-CoV-2SbyusingapanelofmurinemAbsdirectedagainst MERS-RBD antibodies, potently bound to cells expressing SARS-CoV S, including the B30A38, A50A1A1, and C31A12 SARS-CoV S, as expected (Figures 5C and 5D). Nonetheless, antibodies, which recognize SARS-CoV S1, and mAbs 1–3, neitheroftheantibodypreparationsboundtoSARS-CoV-2S(Fig- whichrecognizetheSARS-RBD(FigureS1D;Wenetal.,2004; ures 5E and 5F). In agreement with this observation, although Zhang et al., 2009). Using flow cytometry, all six mAbs were SARS-CoV-2-CTDisstructurallysimilartotheSARS-RBDstruc- observed to effectively bind to cells expressing SARS-CoV S. tures(Figure3),theelectrostaticsurfacepotentialmapsofthese NoneofthemAbs,however,interactedwithSARS-CoV-2S(Fig- proteinsweredifferent(Figure5Gand5H),whichmightexplain ures5Aand5B). thedifferingimmunogenicityofthetwoligands.Therefore,there- IncomparisonwithalimitednumberofmAbs,polyclonalanti- sultshighlightdistinctepitopefeaturesbetweenSARS-RBDand bodiesprovideamorecomprehensiveviewonpotentialepitope SARS-CoV-2-CTD,althoughbothcanengagehACE2. differences. In light of the determinant role of SARS-RBD and MERS-RBD in receptor recognition (Li et al., 2003, 2005; Lu DISCUSSION et al., 2013; Raj et al., 2013), the majority of neutralizing antibodieshavebeenshowntotargettheRBD,exertingneutral- TherecentemergenceofSARS-CoV-2infectioninChinahasled izationactivitybydisruptingvirus/receptorengagement(Duetal., tomajorpublichealthconcerns.ACE2hasbeenreportedtobe 2009;Wangetal.,2016).Wethereforefurtherpreparedmurine thereceptorforthisnovelCoV(Hoffmannetal.,2020;Zhouetal., polyclonalantibodiesagainstSARS-RBDandMERS-RBD.These 2020).Inthisstudy,wedeterminedthekeyregioninSARS-CoV- twoviralRBDsshareverylimitedsequenceidentityandexhibit 2 that is responsible for the interaction with the receptor and 898 Cell181,894–904,May14,2020 ll Article toutilizethehACE2receptorforcellentry.Thecomplexstruc- Table2. ComparisonofhACE2BindingtoSARS-CoV-2-CTDand SARS-RBD tures of hCoV-NL63 CTD and SARS-RBD bound to hACE2 havebeenreportedpreviously(Lietal.,2005;Wuetal.,2009). hACE2 SARS-CoV-2-CTD SARS-RBD Although hCoV-NL63 CTD and SARS-RBD are structurally S19(7/1) A475(3,1),G476(4) P462(1) distinct, the two viral ligands recognize and engage sterically Q24(24/6) A475(4),G476(5), N473(6,1) overlappingsitesinthereceptor(Li,2015).Thecomplexstruc- N487(15,1) tureofSARS-CoV-2-CTDtogetherwithhACE2revealsthatthe T27(15/8) F456(5),Y473(1), L443(3),Y475(5) majorityofbindingsitesofSARS-CoV-2inhACE2alsooverlap A475(2),Y489(7) theSARS-CoVbindingsite.Theobservationsfavorascenario F28(7/7) Y489(7) Y475(7) wheretheseCoVshaveevolvedtorecognizea‘‘hotspot’’region D30(10/2) K417(4,1),L455(2), Y442(2) inhACE2forreceptorbinding. F456(4) During the revision of our manuscript, the full-length hACE2 K31(19/12) L455(2),F456(5), Y442(6),Y475(6) structure was reported to form a dimer in the presence of E484(1),Y489(6), B0AT1(anaminoacidtransporter),asrevealedbycryoelectron F490(2),Q493(3) microscopy(cryo-EM)analysis(Yanetal.,2020).Theyalsore- H34(20/10) Y453(5,1),L455(9), Y440(5,1), ported the cryo-EM structure of dimeric hACE2-B0AT1 bound Q493(6) Y442(1),N479(4) to two SARS-CoV-2-CTDs, with each molecule bound to an E35(8/0) Q493(8) – hACE2monomer,withalocalresolutionof3.5A˚ attheinterface. E37(7/4) Y505(7) Y491(4) OurcrystalstructureofSARS-CoV-2-CTD/hACE2iswellsuper- D38(15/11) Y449(9,1),G496(5), Y436(9,2), imposedwiththecryo-EMstructure,withanRMSDof1.019A˚ Q498(1) G482(1),Y484(1) over 722 pairs of Ca atoms. Notably, two cryo-EM structures Y41(23/25) Q498(8),T500(7,1), Y484(9),T486(8,1), oftrimericSARS-CoV-2Sproteinswerealsopublishedrecently, N501(8,1) T487(8) withthereceptorbindingregionburiedorexposed(Wallsetal., Q42(16/9) G446(4,1),Y449(4,1), Y436(5,1),Y484(4) 2020;Wrappetal.,2020),whichisconsistentwiththestructural Q498(8,3) featuresofMERS-CoVandSARS-CoVSproteins(Yuanetal., L45(4/3) Q498(3),T500(1) Y484(2),T486(1) 2017).Furtherstructurealignmentsshowthatthecrystalstruc- ture of SARS-CoV-2-CTD in the complex also fits well with its L79(2/2) F486(2) L472(2) counterpartsinthecryo-EMstructures,withRMSDsof0.724A˚ M82(9/4) F486(9) L472(4) (exposed state) and 0.742 A˚ (buried states) related to PDB: Y83(20/10) F486(11),N487(8,1), N473(8,2),Y475(2) 6VSB and 0.632 A˚ (exposed state) and 0.622 A˚ (buried state) Y489(1) relatedtoPDB:6VYB,respectively.Theseresultsindicatethat Q325(0/4) – R426(2),I489(2) the crystal structure of the complex is consistent with the E329(0/6) – R426(6,1) respectivecryo-EMstructuresandprovidemoredetailedbind- N330(8/11) T500(8) T486(11,1) inginformation. K353(50/48) G496(7,1),N501(11), Y481(1),G482(3), ConsideringthehighsequenceidentitybetweenSARS-CoV- G502(4,1),Y505(28) Y484(2),T487(11), 2-CTD and SARS-RBD, atomic comparisons of the two viral G488(6,1),Y491(25) ligandsbindingthesamereceptorwereperformed.Atomicde- G354(11/10) Y502(7),Y505(4) G488(7),Y491(3) tailsrevealmoreinteractionsinSARS-CoV-2-CTD/hACE2than D355(9/15) T500(8,1),G502(1) T486(8),T487(3), inSARS-RBD/hACE2,includingmoreengagedresidues,more G488(4) vdw contacts, more H-bonds, as well as larger buried surface R357(3/4) T500(3) T486(4) areas.Interestingly,theb1’/b20loop,whichisthemostvariable R393(1/1) Y505(1) Y491(1) region between SARS-CoV-2-CTD and SARS-RBD, confers moreinteractionstoSARS-CoV-2-CTD/hACE2,includingstrong Total 288(16) 213(11) interactions, such as aromatic-aromatic interactions and ionic ThenumbersinparenthesesofhACE2residuesrepresentthenumberof interactions, in contrast to the SARS-RBD b1’/b20 loop. A vdwcontactsbetweentheindicatedresiduewithSARS-CoV-2-CTD(the recentlypublishedpaperalsoindicatesthattheSARS-CoV-2S former)andSARS-RBD(thelatter).Thenumbersinparenthesesofeither ligandresidues representthenumbersofwdwcontactstheindicated proteinbindshACE2withhigheraffinitythantheSARS-CoVS residuesconferred.Thenumberswithunderlinesuggestnumbersofpo- protein (Wrapp et al., 2020), which was shown in this report tential H-bonds between the pairs of residues. wdw contact was aswell. analyzedatacutoffof4.5A˚ andH-bondsatacutoffof3.5A˚.Seealso ProteolysisoftheSproteinintoS1andS2isanotherprereq- TableS1. uisite for CoVs infection. MERS-Uganda and the bat CoV HKU4canreadilyinteractwithhCD26,buttheybothrequirepro- teaseactivationforcellentry(Kametal.,2009;Matsuyamaand solved the crystal structure of SARS-CoV-2-CTD in complex Taguchi,2009;Menacheryetal.,2015,2020;Wangetal.,2014). withhACE2. Arecentstudyshowsthat,incontrastwithSARS-CoVS,which ConsideringthenewlyidentifiedSARS-CoV-2,atotalofseven does not contain furin recognition sites between S1 and S2, humanCoVshavebeenreportedsofar.Oftheseviruses,three SARS-CoV-2Scontainsonepotentialcleavagesiteandcould (hCoV-NL63, SARS-CoV, and SARS-CoV-2) have been shown beefficientlyprocessedintoS1andS2(Hoffmannetal.,2020). Cell181,894–904,May14,2020 899 ll Article Figure3. ComparisonoftheSARS-CoV-2-CTD/hACE2andSARS-RBD/hACE2BindingSites (A)OverallsimilarreceptorbindingmodeswereobservedforSARS-CoV-2-CTDandSARS-RBD.SuperimpositionofthestructureofSARS-CoV-2-CTD(external subdomaininorangeandcoresubdomainincyan)boundtohACE2(violet)andacomplexstructureofSARS-RBD(ingray)withhACE2(yellow)areshown.The loopexhibitingvariantconformationsishighlightedbyadashedoval. (B)hACE2displayedinsurfaceview.ResiduesthatinteractwiththeSARS-CoV-2-CTDaremarked. (C)hACE2displayedinsurfaceview.ResiduesthatinteractwiththeSARS-RBDaremarked. (D)ResiduessubstitutionsinSARS-CoV-2-CTDslightlystrengthentheinteractionwiththereceptorcomparedtotheSARS-RBD.Theaminoacidsequencesof theloopspecifiedin(A)werealignedbetweentheSARS-CoV-2-CTDandtheSARS-RBD.Thenumbersshowthevdwcontactsbetweenthereceptorwiththe indicatedSARS-CoV-2-CTDresidues(abovethesequence)orSARS-RBDresidues(belowthesequence).Numbersinparenthesesindicatethenumberof potentialH-bondsconferredbytheindicatedresidues.Theredandbluearrowsrepresenttheaminoacidsthatformionicandaromatic-aromaticinteractionswith thereceptor,respectively. SeealsoFigureS1andTableS1. TheserineproteaseTMPRSS2hasbeenreportedtocontribute teins have been shown to efficiently induce production of toprimingoftheSARS-CoV-2Sprotein,andaTMPRSS2inhib- neutralizing antibodies (Du et al., 2009; Wang et al., 2016). itorapprovedforclinicalusewasabletoblockentry.Theauthors However, because of the observed differences in antigenicity postulated that the TMPRSS2 inhibitor might be a treatment and electrostatic distribution between SARS-CoV and SARS- option(Hoffmannetal.,2020). CoV-2,itisunclearwhetherpreviouslydevelopedSARS-RBD- AlthoughSARS-CoVandSARS-CoV-2sharemorethan70% basedvaccinecandidates,suchassubunitvaccines,willconfer sequenceidentityintheSprotein,andbothengagehACE2via effective SARS-CoV-2 prophylaxis. During the revision of our the CTD, we find that the two viruses CTDs are antigenically manuscript,otherstudieshavereportedthatSARS-CoVS-eli- distinct.WhenusingapanelofmAbstargetingSARS-CoVS1/ citedpolyclonalantibodiesinmiceandpatientspotentlyneutral- CTD, none of the antibodies were able to recognize SARS- izedSARS-CoV-2S-mediatedentryintocells(Hoffmannetal., CoV-2 S. mAb1 and mAb2/mAb3 used in the above assay 2020; Walls et al., 2020). Notably, the S2 regions between havebeendeterminedtobindtoSARS-CoVSprotein330–350 SARS-CoV and SARS-CoV-2 exhibit higher sequence identity and 380–399, respectively (Zhang et al., 2009). However, the (~90%) and also contain neutralizing epitopes (Duan et al., bindingsitesfortheotherthreemAbs(B30A38,A50A1A1,and 2005;Wangetal.,2015).Thus,theefficacyofSARS-CoVvac- C31A12), which were generated using SARS-CoV S1 as the cinestargetingSproteinsforSARS-CoV-2prophylaxisrequires immunogen,remainelusive.Consistently,arecentlypublished furtherevaluationandstudy. paper also reported similar results showing that three SARS- Inconclusion,CoVsarezoonoticpathogensandinfecthumans RBD-directed mAbs, S230, m396, and 80R, were unable to viainter-speciestransmission.SARS-CoVandMERS-CoVaretwo bind to SARS-CoV-2 (Wrapp et al., 2020). Furthermore, we notoriousexamplesofCoVscrossingthespeciesbarrierandre- also demonstrate that polyclonal antisera directed against sultinginhumaninfection.Previousstudieshaveshownthatthe SARS-RBD do not recognize the S protein of SARS-CoV-2. A twovirusesfirstjumpedfromtheirnaturalhosts(bats)toaninter- comparison of the two viral ligands shows that they display mediate adaptive animal (e.g., dromedary camels for MERS- divergentelectrostaticpotential,whichlikelyresultsindiffering CoV) before infecting humans (Azhar et al., 2014; Wang et al., immunogenicity despite both ligands showing a similar pro- 2014).Delineatingthiscross-speciestransmissionroutecouldbe teinfold. highly instructive for disease control. Nevertheless, the natural ConsideringthekeyroleoftheCTDinreceptorbinding,this hostandtheintermediateadaptiveanimal,ifany,forSARS-CoV- receptorengagemententityisanidealimmunogenforvaccine 2remainsunknown.ThestructuralinformationforSARS-CoV-2- development. For instance, SARS-RBD and MERS-RBD pro- CTDandhACE2showninthisstudyshouldshedlightontheviral 900 Cell181,894–904,May14,2020 ll Article Figure4. SpecificInteractionsbetweenSARS-CoV-2-S1andSARS-CoV-2-CTDwithhACE2,CharacterizedbySPR TheindicatedmFc-taggedproteinsinthesupernatantwerecapturedbyanti-mIgGantibodiesthatwereimmobilizedonthechipandsubsequentlytestedfor bindingwithgradientconcentrationsofhACE2orhCD26,withthefollowingbindingprofilesshown. (A)SARS-RBDbindingtohACE2. (B)SARS-RBDbindingtohCD26. (C)MERS-RBDbindingtohACE2. (D)MERS-RBDbindingtohCD26. (E)SARS-CoV-2-S1bindingtohACE2. (F)SARS-CoV-2-S1bindingtohCD26. (G)SARS-CoV-2-NTDbindingtohACE2. (H)SARS-CoV-2-NTDbindingtohCD26. (I)SARS-CoV-2-CTDbindingtohACE2. (J)SARS-CoV-2-CTDbindingtohCD26. (K)CulturesupernatantofHEK293Tcellswithouttransfection(NC)bindingtohACE2. (L)CulturesupernatantofHEK293Tcellswithouttransfection(NC)bindingtohCD26. Thevaluesshownarethemean±SDofthreeindependentexperiments. inter-speciestransmissionroutebycharacterizingtheinteractions B SPRanalysis betweenSandhACE2ofdifferentspeciesinthefuture. B Indirect immunofluorescence analysis and confocal microscopy STAR+METHODS B Immunizationofmice B Crystallization Detailedmethodsareprovidedintheonlineversionofthispaper B Datacollectionandstructuredetermination andincludethefollowing: B Sequencesusedinthealignments d QUANTIFICATIONANDSTATISTICALANALYSIS d KEYRESOURCESTABLE B Bindingstudies d LEADCONTACTANDMATERIALSAVAILABILITY B Flowcytometryanalysis d EXPERIMENTALMODELANDSUBJECTDETAILS d DATAANDCODEAVAILABILITY B Cells d METHODDETAILS SUPPLEMENTALINFORMATION B Genecloning B Proteinexpressionandpurification Supplemental Information can be found online at https://doi.org/10.1016/j. B Flowcytometry cell.2020.03.045. Cell181,894–904,May14,2020 901 ll Article Figure5. DifferentAntigenicitybetweentheSARS-CoV-2SandSARS-CoVSProteins (AandB)HEK293TcellsweretransfectedwithpCAGGSplasmidscontainingFlag-taggedSARS-CoVS(A)orSARS-CoV-2S(B).Theindicatedpurifiedmurine mAbsweresubsequentlyaddedtothetransfectedcellsbeforetheywerefixed,permeabilized,andstainedwithanti-Flag/fluoresceinisothiocyanate(FITC). (CandD)HEK293TcellsweretransfectedwithpCAGGSplasmidsexpressingFlag-taggedSARS-CoVS.ThemurinepolyclonalseraagainstSARS-RBD(C)or MERS-RBD(D)weresubsequentlyaddedtothetransfectedcellsbeforetheywerefixed,permeabilized,andstainedwithanti-Flag/FITC. (EandF)HEK293TcellsweretransfectedwithpCAGGSplasmidsexpressingFlag-taggedSARS-CoV-2S.ThemurinepolyclonalseraagainstSARS-RBD(E)or MERS-RBD(F)weresubsequentlyaddedtothetransfectedcellsbeforetheywerefixed,permeabilized,andstainedwithanti-Flag/FITC. (G)ElectrostaticsurfaceviewofSARS-CoV-2-CTD.Thefirstpanelrepresentsthetopview.Theothersareyieldedbyrotationoftheformerpanelalonga horizontalaxis. (H)ElectrostaticsurfaceviewofSARS-RBD.Thefirstpanelrepresentsthetopview.Theothersareyieldedbyrotationoftheformerpanelalongahorizontalaxis. ACKNOWLEDGMENTS XDB29040302), the National Key Research and Development Program of China(2016YFD0500305),theNationalNaturalScienceFoundationofChina WearegratefultoW.N.D.Gaoforrevisingthemanuscript.WethankY.Chen (81922044,81673358and81973228),andtheZhejiangUniversitySpecialSci- andZ.Yang(InstituteofBiophysics,ChineseAcademyofSciences[CAS])for entificResearchFundforCOVID-19PreventionandControl(2020XGZX031). technicalsupportfortheSPRanalysis.WethankX.Zhang(InstituteofMicro- Q.W. is supported by Youth Innovation Promotion Association CAS grant biology,CAS)forhelpwithperformingconfocalimaging.WearegratefultoL. 2018119.G.L.issupportedbythespecialresearchfundonCOVID-19ofSi- Dai(BeijingInstituteofLifeSciences,CAS)andK.Wen(SouthernMedicalUni- chuanProvince(2020YFS0010)andthespecialresearchfundonCOVID-19 versity) for kindly providing the mAbs/pAbs against SARS-CoV. We thank ofWestChinaHospitalSichuanUniversity(HX-2019-nCoV-004).J.Yissup- L.Wangforhelpwiththephylogeneticsanalysis.Thisworkwassupported ported by Foundation of the NSFC Innovative Research Group grant by the Strategic Priority Research Program of CAS (XDB29010202 and 81621091. 902 Cell181,894–904,May14,2020