A multiply substituted G-H loop from foot-and-mouth disease virus in complex with a neutralizing antibody: A role for water molecules

Jou nal o Gene al Vi ology (2000), 81, 1495–1505. P in ed in G ea B i ain ................................................................................................................................................................................................................................................................................... A mul iply subs i u ed G–H loop om oo -and-mou h disease i us in complex wi h a neu alizing an ibody: a ole o wa e molecules Wendy F. Ochoa,1Susana G. Kalko,1Mau icio G. Ma eu,2Paula Gomes,3Da id And eu,3 Es eban Domingo,2Ignasi Fi a1and Nu ia Ve dague 1 1Ins i u o Biologı!a Molecula de Ba celona, Consejo Supe io de In es igaciones Cien ı! icas, Jo di-Gi ona 18-26, 08034 Ba celona, Spain 2Cen o de Biologı!a Molecula ‘Se e o Ochoah, Consejo Supe io de In es igaciones Cien ı! icas, Uni e sidad Au o !noma de Mad id, 28049 Can oblanco, Mad id, Spain 3Depa amen de Quı!mica O ga !nica, Uni e si a de Ba celona, 08028 Ba celona, Spain The c ys al s uc u e o a 15 amino acid syn he ic pep ide, co esponding o he sequence o he majo an igenic si e A (G–H loop o VP1) om a mul iple a ian o oo -and-mou h disease i us (FMDV), has been de e mined a 2n3A / esolu ion. The a ian pep ide includes ou amino acid subs i u ions in he loop ela i e o he p e iously s udied pep ide ep esen ing FMDV C-S8c1 and co esponds o he loop o a na u al FMDV isola e o sub ype C1. The pep ide was complexed wi h he Fab agmen o he neu alizing monoclonal an ibody 4C4. The pep ide adop s a compac old wi h a nea ly cyclic con o ma ion and a disposi ion o he ecep o - ecogni ion mo i A g–Gly–Asp ha is closely ela ed o he p e iously de e mined s uc u e o he i al loop, as pa o he i ion, and o unsubs i u ed syn he ic pep ide an igen bound o neu alizing an ibodies. New s uc u al indings include he obse a ion ha well-de ined sol en molecules appea o play a majo ole in s abilizing he con o ma ion o he pep ide and i s in e ac ions wi h he an ibody. S uc u al esul s a e suppo ed by molecula -dynamic simula ions. The mul iply subs i u ed pep ide de eloped compensa o y mechanisms o bind he an ibody wi h a con o ma ion e y simila o ha o i s unsubs i u ed coun e pa . One wa e molecule, which o s e ic easons could no occupy he same posi ion in he unsubs i u ed an igen, es ablishes hyd ogen bonds wi h h ee pep ide amino acids. The cons ancy o he s uc u e o an an igenic domain despi e mul iple amino acid subs i u ions has implica ions o accine design. In oduc ion Foo -and-mou h disease i us (FMDV) is he causa i e agen o he mos economically impo an disease o clo en- hoo ed animals wo ldwide (Pe ei a, 1981; B own, 1994). The i us belongs o he genus Aph ho i us o he amily Pico na- i idae (Ruecke , 1996). Con ol o he disease has been based on la ge-scale accina ions wi h whole- i us inac i a ed accines, limi a ion o animal mo emen s and des uc ion o he ds exposed o he i us ( he ‘s amping-ou ’ p ocedu e) ( e iews in Bach ach, 1968; B own, 1994; Domingo e al., 1990; Timoney e al., 1992). The a ailable accines show Au ho o co espondence: Nu ia Ve dague . Fax j34 93 2045904. e-mail n mc i!cid.csic.es gene ally good p o ec ion agains in ec ion wi h homologous i us and wi h an igenically ela ed isola es. Di icul ies acing he e adica ion o FMD include he an igenic di e si y o FMDV in na u e, which has been e lec ed in he iden i ica ion o se en se o ypes (A, O, C, SAT1, SAT2, SAT3 and Asia1), 65 sub ypes, un il sub yping was in e up ed (Pe ei a, 1977), and mul i udes o an igenic a ian s ha o en co-ci cula e in a gi en geog aphical a ea (Ma eu e al., 1988). An igenic a ia ion imposes a pe iodic upda ing o accine s ains, an igenic p ope ies o which mus ma ch hose o he ci cula ing i uses. Fu he mo e, some FMD ou b eaks ha e been aced o accine s ains o he i us (Beck & S ohmaie , 1987). Fo hese easons, i would be highly desi able o de elop e ec i e, syn he ic FMD accines. Neu alizing an ibodies a e impo an de e minan s o 0001-6861 #2000 SGM BEJF W. F. Ochoa and o he sW. F. Ochoa and o he s Table 1. Amino acid sequences o si e A pep ide a ian s Pep ides A15-C"-B escia, A15-C-S8c1 and A15-C-S30 ep esen sequences o FMDV isola es belonging o sub ype C". MAb 4C4, which was aised agains C"-B escia, eac s wi h all h ee pep ides as shown by immunochemical and s uc u al analyses (Ma eu e al., 1990, 1992; Ve dague e al., 1995, 1997). Syn he ic pep ide A15-C- S8c1 LV p esen s a dec ease in binding a ini y o mos si e A-speci ic MAbs, including 4C4 (Ma eu e al., 1992). k, Less han 20% o he binding obse ed wi h he o he pep ides (j) in a quan i a i e enzyme immunodo assay desc ibed by Ma eu e al. (1992). No i us has been epo ed wi h a si e A sequence co esponding o pep ide A15-C-S8c1 LV. The one-le e amino acid code is used. Pep ide Sequence Binding o MAb 4C4 A15-C"-B escia YTASTRGDLAHLTATj A15-C-S8c1 YTASARGDLAHLTTTj A15-C-S8c1 LV YTASARGDLAHVTTTk A15-C-S30 YTTSTRGDLAHVTATj p o ec ion agains FMD and o he pico na i us diseases (Ma eu, 1995; McCullough e al., 1992; Misbah e al., 1992). An unde s anding o he ypes o in e ac ions be ween an ibodies and i uses ha lead o i us neu aliza ion is essen ial o accine design. One o he majo an igenic si es o FMDV is loca ed in he G–H loop o capsid p o ein VP1 (Bi le e al., 1982; P a e al., 1982; S ohmaie e al., 1982). This loop is diso de ed on he su ace o FMDV pa icles (Acha ya e al., 1989; Cu y e al., 1996; Lea e al., 1994, 1995; Logan e al., 1993). Fo FMDV o se o ype C, his an igenic si e has been e med si e A, and i beha es as an independen uni , wi h e y limi ed in luence o o he capsid esidues ega ding he in e ac ion o si e A wi h an ibodies (Hewa e al., 1997; Lea e al., 1994; Ma eu, 1995; Ve dague e al., 1999). The beha iou o his an igenic loop in FMDV pa icles in i s in e ac ion wi h an ibodies can be mimicked ai h ully wi h syn he ic pep ides ha ep esen he ele an amino acid sequences ound in au hen ic i us (Cla ke e al., 1983; Ma eu e al., 1989, 1990; Ma eu, 1995; Rowlands e al., 1983). In e es ingly, an igenic si e A includes a highly conse ed A g–Gly–Asp (RGD) iple ha se es as he ecogni ion si e o an in eg in ecep o (Be ins ein e al., 1995; Fox e al., 1989; He na !ndez e al., 1996; Jackson e al., 1997; Mason e al., 1994; Ne e al., 1998). The s uc u e o a syn he ic pep ide ep esen ing an igenic si e A o FMDV C-S8c1 [a biological clone o na u al isola e C- S a Pau Sp70 (Sob ino e al., 1983), a i us ep esen a i e o he Eu opean sub ype C"FMDVs] in a complex wi h he Fab o neu alizing monoclonal an ibodies (MAbs) aised agains he i us has been s udied by X- ay c ys allog aphy (Ve dague e al., 1995, 1996, 1998). The syn he ic pep ide spanned posi ions 136–150 o capsid p o ein VP1 (pep ide A15, Table 1). Two complexes we e s udied, in ol ing MAbs SD6 (Ve dague e al., 1995, 1996) and 4C4 (Ve dague e al., 1998). The wo neu alizing MAbs we e aised agains wo di e en FMDV ype C isola es (Ma eu e al., 1990): SD6 agains C-S8c1 and 4C4 agains C"-B escia, a i us ha di e s om C-S8c1 a wo posi ions (140 and 149) wi hin si e A (Table 1). In he wo complexes, he pep ide an igen acqui ed a e y simila quasi- ci cula con o ma ion and he RGD mo i pa icipa ed di ec ly in he in e ac ion wi h esidues o he complemen a i y- de e mining egions (CDRs) o he an ibodies (Ve dague e al., 1998). The RGD iple appea ed in an open u n con o ma ion, e y simila o ha o educed FMDV pa icles o se o ype O (Logan e al., 1993), and also simila o he con o ma ion in RGD-con aining in eg in ligands (P a , 1997). Rema kably, he Gly-142 and Asp-143 esidues o he RGD iple appea o be c i ical o in e ac ion bo h wi h an in eg in ecep o and wi h neu alizing an ibodies di ec ed o si e A (Ma eu e al., 1996; Ve dague e al., 1995). Bo h s uc u al and biochemical e idence sugges ha MAbs SD6 and 4C4 neu alize by mono alen binding o an igenic si e A (Ve dague e al., 1997, 1998; e iew in Domingo e al., 1999). The quasi-ci cula shape o he pep ide an igen in he complexes was s abilized by pep ide–an ibody in e ac ions and also h ough in apep ide hyd ogen bonds and an de Waalshin e ac ions (Ve dague e al., 1995, 1996, 1998). I a s uc u e simila o he one ound in hese complexes was also main ained wi h a ian an igens (including one o se e al amino acid subs i u ions), he s uc u e would gain g ound as he basis o he design o new an i-FMD immunogens. Se e al na u al isola es o FMDV o se o ype C ha e been cha ac- e ized p e iously wi h ega d o he amino acid sequence a an igenic si e A and eac i i y wi h MAbs (Ma ı!nez e al., 1991; Ma eu e al., 1988, 1989, 1990; Ve dague e al., 1995, 1996, 1998; Villa e de e al., 1991). One o he isola es analysed, e med C-S30 (o C"-Ba celona Sp 81), included ou eplacemen s wi hin an igenic si e A. Two o he changes we e coinciden wi h hose ound in he C"-B escia i us (Ala-140 Th , Th -149 Ala) and wo we e new (Ala-138 Th , Leu-147 Val) (Table 1). In spi e o he p esence o ou amino acid subs i u ions in he G–H loop o VP1, he eac i i y o FMDV C-S30 wi h MAb 4C4 was indis inguishable om ha o FMDV C-S8c1 (Ma eu e al., 1990). Single subs i u ions a posi ions 138 and, especially, 140 and 149 did no ha e a signi ican e ec on he in e ac ion o syn he ic pep ide an igens wi h mos MAbs, including 4C4, in ELISA (Ve dague e al., 1998). Howe e , he single eplacemen o esidue Leu- 147, highly conse ed among FMDVs o se o ype C, by aline diminished g ea ly he eac i i y o syn he ic pep ides wi h mos MAbs, including 4C4, in enzyme immunodo assays and compe i i e ELISA (Ma eu e al., 1992; Ve dague e al., 1998). Tha is, he ou amino acid eplacemen s ound a he an igenic si e A o C-S30 ela i e o C-S8c1 (Ma eu e al., 1989, 1990) appea o exe a compensa o y e ec ha es o es he a ini y o C-S30 (Ma eu e al., 1990) and o a syn he ic an igen BEJG S uc u e o a G–H loop a ian om FMDVS uc u e o a G–H loop a ian om FMDV ep esen ing i s an igenic si e A (Ma eu e al., 1992) o MAb 4C4. The e o e, he subs i u ed an igen o e ed an oppo uni y o in es iga e he s uc u e ha a a ian o m o an igenic si e A could acqui e as a complex wi h an ibody. The esul s desc ibed in his epo ea i m he quasi-ci cula con o ma ion o his an igenic loop and e eal new ea u es abou he o ces in ol ed in he s abili y o si e A and abou an igen–an ibody in e ac ions, in pa icula , key in e ac ions media ed by wa e molecules. Me hods MAb and syn he ic pep ides. The o igin and cha ac e iza ion o he an i-FMDV MAb 4C4 (Ma eu e al., 1987, 1989, 1990) and he p epa a ion and pu i ica ion o he Fab agmen o MAb 4C4 (Ve dague e al., 1998) ha e been epo ed p e iously. Pep ides we e syn hesized by solid-phase p ocedu es, pu i ied and analysed by using p ocedu es ha ha e been desc ibed p e iously (Ca en 4oe al., 1992; Ma eu e al., 1996). C ys alliza ion and da a collec ion. C ys als o he complex be ween he Fab agmen o he MAb 4C4 and he a ian pep ide A15- CS30 (Table 1) we e ob ained by he hanging d op- apou di usion Table 2. Fea u es o c ys als o he an igen–4C4 Fab complex analysed in he p esen s udy Fea u e Value Re e ence alue Band wid h C ys alliza ion and da a collec ion Space g oup P2"2"2" Cell pa ame e s (A /)48n18, 69n32, 146n54 Resolu ion (A /) 20–2n3 O e all comple eness (%) 93n0 A e age I σ7n5 Rsymm (%) 8n5 To al no. o esidues: Fab 429 Pep ide 13 To al no. o sol en molecules 166 Volume o sol en (%) 49 Di ac ion ag eemen Resolu ion (A /) 15–2n3 Numbe o e lec ions 17943 R ee 0n262 R- ac o 0n248 RMSD om ideal dis ance: Bond leng h (A /)0n009 Bond angle (m)1n48 A e age he mal ac o (A /#): Fab 23n6 Pep ide 23n2 S e eochemis y o main chain: Omega angle SD (m)1n26n03n0 Bad con ac s pe 100 esidues 0n87n610n0 Ze a angle SD1n33n11n6 S e eochemis y o side chain: Chi-1 pooled SD 13n320n35n0 echnique and successi e mic o- and mac oseedings (S u a & Wilson, 1992). Typically small, wined needles, ob ained wi h 18% PEG 4K, we e used o mic oseeding, which p oduced la ge needles a 16% PEG 4K. Finally, hese la ge needles we e used o mac oseedings in 2 µl d ople s con aining 7 mg ml Fab, 1n8mg ml pep ide, 6n5% PEG 4K, 0n2 M LiCl wi h 50 mM T is–HCl (pH 9), equilib a ed agains a ese oi con aining 13% PEG 4K equally bu e ed a oom empe a u e. C ys als we e o ho hombic, space g oup P2"2"2"wi h uni cell pa ame e s a l48n2A /, bl69n3A /and c l146n5A /, con aining one molecule o he complex pe asymme ic uni , which co esponds o a sol en con en olume o 49%. A da a se was collec ed, a 100 K wi h 20% glyce ol as a c yop o ec an , by using a Ma Reseach imaging pla e on a Rigaku o a ing anode. In ensi ies we e e alua ed and scaled in e nally wi h p og ams Denzo and Scalepack, espec i ely (O winowski & Mino , 1996). Da a we e 93% comple e a 2n3A / esolu ion, wi h an in e nal ag eemen ac o (Rsymm)o 8n5% (Table 2). S uc u e solu ion and e inemen . C ys als o he complex be ween pep ide A15-C-S8c1 LV and he Fab o MAb 4C4 seemed ela ed o c ys als o med wi h he same Fab and pep ide A15-C-S8c1 ( he names and amino acid sequence o pep ide an igens a e lis ed in Table 1), whose s uc u e had been sol ed p e iously (Ve dague e al., 1998). Howe e , he uni cell pa ame e c di e ed by abou 10 A /and he s uc u e was newly de e mined by molecula eplacemen by using he BEJH W. F. Ochoa and o he sW. F. Ochoa and o he s Fig. 1. (A) S e eo iew o he (2Fo-Fc) elec on densi y map co esponding o he an igenic pep ide A15-C-S30. The pep ide model and nea by sol en molecules (isola ed do s) a e also shown. (B) S e eo iew o he Fab–pep ide in e ac ions. Only he Fab esidues in di ec con ac wi h he pep ide a e depic ed (open ods). Wa e molecules (W) in di ec con ac wi h he pep ide o media ing he an ibody ecogni ion a e also shown. Hyd ogen bonds a e indica ed by b oken lines. (C) Supe - posi ion o he A15 pep ide con o ma ions ound in he 4C4 Fab complexes wi h he A15-C-S30 ( illed) and he A15-C- S8c1 (open) pep ides. AMoRe package wi h he 4C4 Fab coo dina es as sea ching model (Na aza, 1994). The ini ial solu ion was hen op imized by allowing o mo e as ou sepa a ed igid bodies he a iable hea y, a iable ligh , cons an hea y and cons an ligh domains. The esul ing R- ac o was 0n36% in he esolu ion ange o 15 o 4 A /. Examina ion o he elec on densi y maps, calcula ed a his s age, clea ly showed ex a densi y co esponding o he oligopep ide occupying he an igen-binding si e. The inal model o he s uc u e o he complex was ob ained by i e a i e cycles o model ebuilding by using he p og am O (Jones e al., 1991) and posi ional e inemen wi h XPLOR (B u $nge , 1992), including bulk BEJI S uc u e o a G–H loop a ian om FMDVS uc u e o a G–H loop a ian om FMDV Table 3. Hyd ogen bonds be ween he A15-CS30 and A15-CS8 pep ides and he 4C4 Fab The pa icipan s in he a ious hyd ogen bonds a e shown. C ys allog aphic da a a e desc ibed in Table 2 and he h ee-dimensional s uc u es o he wo complexes a e shown in Fig. 2. The CDR loops a e gi en o which he 4C4 Fab a oms lis ed belong. Pep ide a om Wa e molecule 4C4 Fab a om CDR loop Dis ance (A / ) A15-CS30 Th -137 O – Asp-104 N H3 3n2 Th -137 Oγ1 – Se -103 OγH3 3n1 Th -138 Oγ1w8 – –2n7 Se -139 N w3 Se -95 O L3 3n0 2n7 Se -139 N w3 Asp-104 OδH3 3n0 2n8 Se -139 Oγ– Asn-96 Oδ1L3 2n7 Se -139 O w2 – – 2n7 Th -140 N w9 Se -31 OγL1 2n7 2n6 Th -140 Oγ1w9 – –3n2 Th -140 Oγ1w4 – –3n3 A g-141 O – Asp-98 N L3 2n9 A g-141 O w6 – – 3n0 A g-141 NH1 – Glu-97 Oε1L3 2n7 Gly-142 N w1 w5 Asp-98 Oδ1L32n9 3n1 2n8 Gly-142 N w2 – – 3n4 Gly-142 O w1 – – 2n9 Asp-143 N w2 – – 3n3 Asp-143 Oδ2 – A g-99 Nη2H3 3n0 Asp-143 Oδ1 – A g-99 NεH3 2n8 Asp-143 Oδ1 – Th -50 Oγ1H2 2n9 Leu-144 N w2 – – 2n8 Leu-144 O w8 – – 2n9 His-146 N – Ty -59 OηH2 3n4 His-146 Nδ1 – Ty -59 OηH2 3n1 His-146 Nε2 – Th -33 Oγ1H1 2n8 Val-147 O w8 – – 2n5 –w7 w1 – – 3n0 A15-CS8c1 Th -137 N – A g-54 Nη2L2 2n6 Th -137 O – Asp-104 N H3 3n2 Th -137 Oδ1 – Asp-34 Nδ2L1 2n9 Se -139 Oδ– Asn-96 Oδ1L3 3n4 Asp-143 O – Ty -59 OηH2 3n4 Asp-143 Oδ1 – A g-9 Nη2H3 2n7 Asp-143 Oδ2 – A g-99 Nη1H3 2n8 Asp-143 Oδ2 – Th -50 Oγ1H2 3n0 His-146 Nδ1 – Ty -59 OηH2 2n6 His-146 Nδ2 – Th -33 OδH1 3n2 His-146 Nδ2 – Se -52 OδH2 2n8 sol en co ec ion. The e ined model con e ged o c ys allog aphic ag eemen ac o s R and R ee o 24% and 26%, espec i ely, o 17943 e lec ions in he esolu ion shell 15n0–2n3A /(Table 2). Molecula dynamics (MD) simula ion. Th ee se s o MD simula ions, co esponding o he a iable module (F ) o he 4C4 Fab in e ac ing wi h pep ides A15-C-S8c1, A15-C-S8c1 LV and A15-C-S30 (Table 1), we e pe o med by using he GROMOS96 package wi h i s s anda d p o ein and wa e o ce ields ( an Guns e en e al., 1996). C ys allog aphic coo dina es o pep ides A15-C-S8c1 and A15-C-S30 ( esidues 137–148) we e used as s a ing models in he co esponding simula ions. Th -148, no isible in he elec on densi y maps o he A15- C-S8c1 complex, was added o a oid end e ec s in he c i ical Leu-147 esidue, by using he in o ma ion a ailable in he A15-C-S30 s uc u e. S a ing coo dina es o he A15-C-S8c1 LV pep ide we e hen de i ed, wi h he g aphic p og am TURBO (Rousel & Cambillau, 1989), by BEJJ W. F. Ochoa and o he sW. F. Ochoa and o he s eplacing Leu-147 wi h aline. C ys allog aphic wa e molecules de- e mined in he p esen wo k o he complex o 4C4 Fab wi h he A15- C-S30 pep ide we e no included in he simula ions. In e e y simula ion, he s uc u e o he co esponding complex was placed ini ially a he cen e o a unca ed oc ahed on, he dimensions o which we e chosen such ha he minimum dis ance o any p o ein a om om he closes wall was 7 A /. The edge leng hs o he co esponding cubic boxes we e abou 73 A /. Sys ems we e ea ed as imme sed in o an equilib ium con igu a ion o bulk simple poin cha ge (SPC) wa e (Be endsen e al., 1986). Wa e molecules ou side he box o wi h a dis ance o a solu e a om o less han 2n3A /we e emo ed. The numbe s o wa e molecules conside ed in he h ee simula ions we e 5271, 5272 and 5264, espec i ely. To elax s ong wa e –wa e and wa e –p o ein non-bonded in e ac ions, s eepes -descen ene gy minimiza ion was pe o med un il s abiliza ion was eached. A e ha , coun e -ions we e added o neu alize cha ged p o ein si es, wi h a subsequen ene gy minimiza ion. Simula ions we e pe o med a cons an olume and empe a u e (300 K) wi h pe iodic bounda y condi ions and an in- eg a ion s ep o 2 s. Tempe a u e was kep cons an by weak coupling o an ex e nal ba h (Be endsen e al., 1984, 1986). Bond leng hs we e cons ained o equilib ium alues by using he SHAKE algo i hm (Rickae e al., 1977). Equilib a ions we e achie ed wi hin 50 ps and, a e wa ds, he h ee sys ems we e simula ed o a o al ime o 300 ps wi h he o al po en ial ene gy emaining essen ially cons an . Only amino acids om he pep ides and all wa e molecules we e allowed o mo e du ing calcula ions. Analysis we e pe o med, mainly wi h p og ams con ained in he GROMOS96 package, using he s uc u es gene a ed e e y 0n1 ps du ing he in e al spanning om equilib a ion un il he end ( om abou 50 o 350 ps). Coo dina es. Coo dina es ha e been deposi ed in he B ookha en p o ein da abase unde accession numbe 1EJO. Resul s S uc u e o he 4C4 Fab in complex wi h he A15-CS30 pep ide The inal elec on densi y maps o he complex be ween he 4C4 Fab and he A15-C-S30 pep ide, a 2n3A / esolu ion, allowed he posi ioning o he main and side chains o 431 o he o al o 436 4C4 Fab esidues (216 om he ligh chain and 220 om he hea y chain) and o 12 o he 15 pep ide esidues (Fig. 1A). Hea y-chain Fab esidues 136–143, in he cons an domain, and e minal pep ide esidues Ala-149, Th -150 and he side chain o Ty -136 we e diso de ed, and hei co- o dina es ha e no been in oduced in he p esen model. The quali y o he inal maps is also e lec ed in he 166 well-de ined sol en molecules ound, nine o which pa icipa e in he in e ac ions be ween he pep ide and he Fab CDRs (Table 3), and h ee wa e molecules es ablish b idges among pep ide esidues (Fig. 1B). The A15-C-S30 pep ide shows an o e all con o ma ion closely ela ed o he one ound in o he si e A complexes de e mined p e iously (Ve dague e al., 1995, 1998) (Fig. 1 C). The RGD mo i , esidues 141–143, is loca ed in an open u n con o ma ion p eceded by an ex ended egion, esidues 136–140, and ollowed by a sho helix, posi ions 144–148 (Fig. 1). All o he pep ide esidues inco po a ed in o he model 100 75 50 25 0 B Residue numbe Con ac a ea (% o esidue su ace) 100 75 50 25 0 A 136 Residue numbe Con ac a ea (Å2) 137 *138 139 *140 141 142 143 144 145 146 *147 148 136 137 *138 139 *140 141 142 143 144 145 146 *147 148 Fig. 2. Con ac a eas o he A15 pep ides in he 4C4 Fab complexes wi h he A15-C-S30 ( illed ba s) and he A15-C-S8c1 (open ba s) pep ides exp essed as absolu e alues (A) o as pe cen ages o he esidue su ace (B). As e isks indica e A15-C-S30 esidues ha di e in A15-C-S8c1. Pep ide sequences a e shown in Table 1. in e ac di ec ly wi h he hype a iable egions o he Fab (Table 3 and Fig. 2). In pa icula , he en i e molecula su aces o Asp-143, om he RGD mo i , and Leu-144 a e in con ac wi h he an ibody, simila o wha had been obse ed in complexes in ol ing Fab A15-C-S8c1 (Fig. 2) (Ve dague e al., 1995, 1998). Howe e , he subs i u ed esidues Th -138 and, pa icula ly, Th -140 and Val-147 do no p esen ex ensi e in e ac ions wi h he an ibody (Fig. 2). The ou h subs i u ion, Ala-149, no isible in he elec on densi y and no included in he p esen model, mus also emain exposed o he sol en and a om he an ibody. Compa isons o he si e A a ian pep ides A15-CS30 and A15-CS8c1 The oo -mean-squa e de ia ion (RMSD) be ween he main-chain pep ide a oms in he A15-C-S30 and A15-C-S8c1 Fab 4C4 complexes is only 0n4A / o he common esidues 137–147. The A g-141 side chain has di e en disposi ions in he wo pep ides (Fig. 1C), al hough he elec on densi y co esponding o his esidue was weak in bo h complexes (Fig. 1A; see also Ve dague e al., 1995, 1998). The eplaced esidues, Th -138 and Val-147, show he la ges main-chain BFAA S uc u e o a G–H loop a ian om FMDVS uc u e o a G–H loop a ian om FMDV Fig. 3. Space- illing ep esen a ion o he h ee pep ide models A15-C- S8c1 (A), A15-C-S8/LV (B) and A15-C-S30 (C). Pep ide esidues ha con ibu e o in apep ide hyd ophobic in e ac ions a e highligh ed in da k- blue. The wa e molecule ound in con ac wi h he subs i u ed Th -138 in he A15-C-S30 pep ide is displayed in ed (C). de ia ions o 0n7 and 1n0A /, espec i ely. Despi e he high s uc u al simila i y be ween he wo pep ides A15-C-S30 and A15-C-S8c1, di e ences in he main-chain con o ma ional angles a e impo an , pa icula ly a ound he RGD mo i (Fig. 2). Thus, in he A15-C-S30 pep ide, esidues Th -140 and A g- 141 a e si ua ed in he Ramachand an egion co esponding o α-helices while, in he A15-C-S8c1 pep ide, Ala-140 and A g- 141 a e in he egions o β-s ands and le -handed helices, espec i ely. A hi d si ua ion is ound in he s uc u e epo ed o he G–H loop o he educed FMDV se o ype O"(Lea e al., 1993, 1995), whe e he co esponding esidues Leu-144 and A g-145 a e bo h ound in he Ramachand an egion co - esponding o β-s ands. The lexibili y o Gly-142 allows compensa o y main-chain o sional angles ha esul in he o e all s uc u al simila i y o he h ee pep ides. I is impo an o no ice he p esence, in he A15-CS30 s uc u e, o one wa e molecule (named w8 in Fig. 1) ha is hyd ogen-bonded wi h he side chain o Th -138 and he main-chain oxygen a oms o esidues Leu-144 and Leu-147. This wa e molecule canno be p esen , o s e ic easons, in he A15-C-S8c1 s uc u e, whe e he bulkie Leu-147 side chain ills he a ailable space (Fig. 3). MD analysis Th ee MD simula ions co esponding o complexes o he F 4C4 wi h pep ides A15-C-S8c1, A15-C-S8c1 LV and A15- C-S30 we e ca ied ou as desc ibed in Me hods. The main di e ence be ween he a e aged con o ma ions o he h ee pep ides is he opening o he C- e minal egion in he simula ion o he single-subs i u ed A15-C-S8c1 LV pep ide (Fig. 4A). O he signi ican di e ences can also be seen a he amino e mini and o he A g-141 side chain. The s abili y o he a e aged con o ma ions was analysed by e alua ing he RMSD alues o he pep ide luc ua ions in e e y simula ion. The lowes alues co espond o he A15-C-S30 pep ide, while he mo e cons an main-chain egion includes he RGD mo i and Leu-144 in he h ee pep ides, wi h Asp-143 ha ing he smalles a ia ion. The side chains o Asp-143 and Leu-144 also appea ex emely in a ian , p obably e lec ing he ex en and he s abili y o he in e ac ions o hese esidues wi h he an ibody. The s abili y o he e a-subs i u ed pep ide A15-C- S30 obse ed du ing he simula ion is con ibu ed by a hyd ogen-bonded wa e molecule loca ed in a si e e y close o he posi ion o he b idging c ys allog aphic wa e (w8) (Figs 1B and 4B). This wa e si e p esen s a high occupa ion and a low RMSD o abou 0n1A /. Hyd ogen-bond in e ac ions ake place be ween he wa e molecule ha occupies his si e and pola a oms om he subs i u ed esidues Th -138, Leu- 144 and Val-147 (Table 3). The inc eased hyd ophobic su ace exposed o he sol en appea s o explain he depa u e om he ‘co ec hsi e A-like con o ma ion obse ed by MD wi h he A15-CS8 LV pep ide (Fig. 3). BFAB W. F. Ochoa and o he sW. F. Ochoa and o he s Fig. 4. (A) S e eo iew o he supe posi ion o he A15-C-S8c1 ( illed), A15-C-S8/LV (shaded) and A15-C-S30 (open) pep ides co esponding o he a e aged con o ma ions calcula ed in 300 ps o MD simula ions o he co esponding F –pep ide complexes. (B) S e eo iew o he supe posi ion o he A15-C-S30 coo dina es om he c ys al ( illed) and he MD simula ion (open). The wa e molecule ound in he c ys al s uc u e (shaded) is si ua ed close o he wa e si e de e mined in he MD simula ion (open). Discussion The e ec s o single o mul iple amino acid subs i u ions a an igenic si e A o FMDV o se o ype C on he eac i i y o si e A-speci ic neu alizing MAbs ha e been s udied ex- ensi ely by immunochemical assays employing a ian i uses and subs i u ed syn he ic pep ides (Ma eu e al., 1989, 1990, 1992; Ma ı!nez e al., 1997; No ella e al., 1993; Ve dague e al., 1998). Replacemen s a posi ion 147 al e ed he binding o mos an ibodies subs an ially, including 4C4, e en hough esidue 147 had only ma ginal in e ac ions wi h ha an ibody. In pa icula , subs i u ion o Leu-147 by aline diminished he binding o mos si e A-speci ic MAbs analysed (Ma eu e al., 1992). Howe e , he e ec o his single subs i u ion Leu-147 Val was appa en ly compensa ed o , a leas in pa , by he eplacemen Ala-138 Th (Ma eu e al., 1992). These obse a ions and he ac ha na u al FMDV isola e C-S30 (Ma ı!nez e al., 1991; Villa e de e al., 1991) con ains hese c i ical eplacemen s in i s VP1 G–H loop ela i e o clone C-S8c1 (Table 1) encou aged he p esen s udy. In he s uc u e o he 4C4 Fab complexed wi h he a ian pep ide A15-C-S30 epo ed he e, he wo amino acids Th - 138 and Val-147 ha we e eplaced in he pep ide show only mino in e ac ions wi h he Fab (Table 2 and Fig. 2). Di e ences in binding a ini ies o pep ides wi h subs i u ions in hose wo esidues should he e o e be due mainly o ee-ene gy di e ences when adop ing a ‘co ec hsi e A-like loop s uc u e, he one ecognized by an ibodies o by he ecep o (Fig. 5). The educed s abili y o ha con o ma ion in he pep ide wi h he single subs i u ion Leu-147 Val would hen explain he educed a ini y obse ed o he an ibody. Ins ead, he double- subs i u ed pep ide could eco e binding a ini y by e- s abilizing he ‘co ec hcon o ma ion wi h he addi ional wa e molecule (w8) ha b idges he chain hyd oxyl g oup o he eplaced Th -138 wi h he main-chain oxygen a oms o esidues Leu-144 and Leu-147 (Table 3 and Fig. 3C). In he MD simula ion o he A15-C-S30 a ian pep ide, a wa e molecule wi h a low RMSD is placed in almos exac ly he same loca ion as he one ound in he c ys al s uc u e (Fig. 4B). These s uc u al esul s p o ide a a ionale o how a gene al change in speci ici y o a la ge di e si y o an ibodies can be achie ed by a des abiliza ion o he o iginal an igen s uc u e. In hose cases, u he subs i u ions can es o e he a ini y o he an ibodies by inc easing he s abili y o he o iginal an igen con o ma ion, e en when none o he esidues subs i u ed BFAC S uc u e o a G–H loop a ian om FMDVS uc u e o a G–H loop a ian om FMDV Fabc ys al + Pep idec ys al Fab – Pep idecomplex Fabsolu ion + Pep idesolu ion ∆Gbinding ∆G o al ∆G olding Fig. 5. A simpli ied he modynamic cycle o a Fab in e ac ing wi h an an igenic pep ide can be desc ibed as a wo-s ep p ocess. In he i s s ep, he pep ide adop s indi idually he s uc u e seen in he complex ha is o med in he second s ep, so ha ∆G o al l∆G oldingj∆Gbinding, whe e ∆G olding e e s o he ee-ene gy di e ence be ween he pep ide s uc u e in solu ion and ha ound in he complex ( he Fab eo ganiza ion, i equi ed, can also be included in his e m) and ∆Gbinding e e s o he ee- ene gy di e ence due o he o ma ion o he in e ac ions be ween he pep ide and he Fab. A ‘lock-and-key’ ecogni ion mechanism, in which s uc u es emain unchanged on complex o ma ion, co esponds o ∆G olding l0. Ins ead, ∆G olding 0 desc ibes an ‘induced- i ’ mechanism, whe e he complex would only be s abilized by a a ou able ∆Gbinding con ibu ion. Analysis o Fab 4C4 complexes wi h a numbe o A15 pep ides, bo h om he a ailable c ys al s uc u es and wi h MD simula ions, sugges ha ∆Gbinding(A15-C-S30) $∆Gbinding(A15-C-S8/LV), while he expe imen al da a on he ela i e s abili ies o hese complexes (Ma eu e al., 1992) imply ha ∆G o al(A15-C-S30) ∆G o al(A15-C- S8/LV), which would allow us o conclude ha ∆G olding(A15-C-S30)  ∆G olding(A15-C-S8/LV). The educed s abili y o he ‘co ec ’ si e A-like con o ma ion o he A15-C-S8/LV pep ide would explain he lack o a ini y o mos si e A MAbs o his a ian pep ide (see ex ). pa icipa es in di ec an igen–an ibody in e ac ions (F ei e, 1999). Recogni ion o con inuous epi opes, like he VP1 G–H loop o FMDV s udied in his wo k, depends no only on he amino acid sequence bu also on he spa ial con o ma ion o he epi ope, whe e sol en molecules o en play a c i ical ole. In e p e a ion o mechanisms o i us escape om an ibody ecogni ion, essen ial in RNA i us e olu ion and accine design, mus also ake in o conside a ion he e ec s o amino acid eplacemen s in connec ion wi h he sol en ea ange- men s imposed. In spi e o di icul ies in p edic ing he e ec s o amino acid eplacemen s in an igen–an ibody ecogni ion, he p esen esul s o FMDV encou age he design o quasi- cyclic s uc u es ep oducing an igenic si e A. Da a o he G–H loop om a numbe o di e en an ibodies, including he mul iply subs i u ed pep ide epo ed he e, sugges a obus and s ongly immunogenic s uc u e despi e he mobili y o his an igenic si e on he i ion su ace (Pa y e al., 1990; Logan e al., 1993). Cock ails o quasi-ci cula pep ides ep esen ing he mos equen escape mu an s ound a si e A should be candida es o inco po a ion in o accine o - mula ions. Such o mula ions should include addi ional B-cell and T-cell epi opes o ensu e a b oad immune esponse o minimize he selec ion o escape mu an s and accine ailu es (Domingo & Holland, 1992; Taboga e al., 1997). We wish o hank E. B occhi o he gene ous supply o MAb 4C4 and he con inuous help wi h ou esea ch. Wo k in Ba celona was unded by DGICYT g an s PB95-0218 and PB97-0873. Wo k in Mad id was suppo ed by g an s om DGES PM97-006-C02-01 and Fundacio !n Ramo !n A eces. W.F.O. is a ecipien o a ellowship om he Minis e io de Educacio !n y Cul u a (Spain). P.G. hanks he Fundaçao C. Gulbenkian o a PhD ellowship and he Uni e si y o Po o o empo a y lea e. Re e ences Acha ya, R., F y, E., S ua , D., Fox, G., Rowlands, D. & B own, F. (1989). The h ee-dimensional s uc u e o oo -and-mou h disease i us a 2n9A / esolu ion. Na u e 337, 709–716. Bach ach, H. L. (1968). Foo and mou h disease i us. Annual Re iew o Mic obiology 22, 201–244. Beck, E. & S ohmaie , K. (1987). Sub yping o Eu opean oo -and- mou h disease i us s ains by nucleo ide sequence de e mina ion. Jou nal o Vi ology 61, 1621–1629. Be endsen, H. J. C., Pos ma, J. P. M., an Guns e en, W. F., DiNola, A. & Haak, J. R. (1984). Molecula dynamics wi h coupling o an ex e nal ba h. Jou nal o Chemical Physics 81, 3684–3690. Be endsen, H. J., Van Guns e en, W. F., Zwinde man, H. R. & Geu sen, R. G. (1986). Simula ions o p o eins in wa e . Annals o he New Yo k Academy o Sciences 482, 269–286. Be ins ein, A., Roi ainen, M., Ho i, T., Mason, P. W. & Bax , B. (1995). An ibodies o he i onec in ecep o (in eg in alpha V be a 3) inhibi binding and in ec ion o oo -and-mou h disease i us o cul u ed cells. Jou nal o Vi ology 69, 2664–2666. Bi le, J. L., Hough en, R. A., Alexande , H., Shinnick, T. M., Su cli e, J. G., Le ne , R. A., Rowlands, D. J. & B own, F. (1982). P o ec ion agains oo -and-mou h disease by immuniza ion wi h a chemically syn hesized pep ide p edic ed om he i al nucleo ide sequence. Na u e 298, 30–33. B own, F. (1994). Foo -and-mou h disease. In Syn he ic Vaccines, pp. 416–432. Edi ed by B. H. Nicholson. Ox o d: Blackwell Scien i ic. B u $nge , A. T. (1992). XPLOR manual, e sion 3.0. Yale Uni e si y, New Ha en, CT, USA. Ca en 4o, C., Roig, X., Cai o, J., Cama e o, J., Ma eu, M. G., Domingo, E., Gi al , E. & And eu, D. (1992). S udies on an igenic a iabili y o C s ains o oo -and-mou h disease i us by means o syn he ic pep ides and monoclonal an ibodies. In e na ional Jou nal o Pep ide and P o ein Resea ch 39, 41–47. Cla ke, B. E., Ca oll, A. R., Rowlands, D. J., Nicholson, B. H., Hough en, R. A., Le ne , R. A. & B own, F. (1983). Syn he ic pep ides mimic sub ype speci ici y o oo -and-mou h disease i us. FEBS Le e s 157, 261–264. Cu y, S., F y, E., Blakemo e, W., Abu-Ghazaleh, R., Jackson, T., King, A., Lea, S., Newman, J., Rowlands, D. & S ua , D. (1996). Pe u ba ions in he su ace s uc u e o A22 I aq oo -and-mou h disease i us accompanying coupled changes in hos cell speci ici y and an igenici y. S uc u e 4, 135–145. Domingo, E. & Holland, J. J. (1992). Complica ions o RNA he - e ogenei y o he enginee ing o i us accines and an i i al agen s. In Gene ic Enginee ing,P inciples and Me hods, ol. 14, pp. 13–32. Edi ed by J. K. Se low. New Yo k: Plenum. Domingo, E., Ma eu, M. G., Ma ı!nez, M. A., Dopazo, J., Moya, A. & Sob ino, F. (1990). Gene ic a iabili y and an igenic di e si y o oo - and-mou h disease i us. In Applied Vi ology Resea ch, pp. 233–266. Edi ed by E. Ku ks ak, R. G. Ma usyk, S. A. Mu phy & M. H. V. Van Regenmo el. New Yo k: Plenum. Domingo, E., Ve dague , N., Ochoa, W. F., Ruiz-Ja abo, C. M., Se illa, N., Ba anowski, E., Ma eu, M. G. & Fi a, I. (1999). Biochemical and s uc u al s udies wi h neu alizing an ibodies aised agains oo -and- mou h disease i us. Vi us Resea ch 62, 169–175. BFAD