A Thaarg of the Stnrcture and Process of Formation of antu.?odiss* Dwlq this past far yaare I ham been makinir: an effort to mderstz.nd md Interpret rerolo@al *enomsna in terns of raolecmlar rtn~0tu.m and molecular inter- act ions. The field of immunology is so axtensfva and the eqeriglanta2. obeomt~ona are eo oomplax (and acoasionally contradictory) that no one has fonnd it ~oasfb3.o to indwm a theory of the structure of antibadiee from the obsarrvatioml mtarinl. -Aa an alternative method of attaak we may yonomd and attmy& to answ?r the followin& qmst%ons: What ijr the ariqlest structure which can be mqypstad, on tke bassfar of the extensl-e infamatian now available about intrmolectiar `and. inter- muleau~r forc02, for a nolmmLe with the! -pro:)erties observed for antibodies, and 1 so0, fo zy of thi6 inforztion, L. Paull~, Bl%e Xatuxa of the Chemlaal Don 1e Stmcttxre of Noleeules and Crystals,* Cornell University Press, Ithaca, IJew York, Second Edltic.., @&. what la the &plast reasonable p~ocssa of fornation of such a mlecula? Pmcead& in this way, I hape dsaeIqed a detailed theory of the atmcture and pmcass of formation of antlbodfee and the nature of serologic& reactions which is nore deffnite and more widely a-+icable than kWW1i8r theorirs, and which is con~tlblrs w5th oux present knowledm of the strnctura and pxuyerties of uf~@a, moleculeer as well 8~ with most of the direct mnpirlcal inform&ion abotnt anti- bodies. This theory is detmribed and discusrsed below. 2 when an antigen ia fn$xWd into an ani~~al f3ostle of its; aMlectxlar am caytumd and held in the rs@~n of anttbody produotion2. Ax antibody to th%e Bordst , foraaa of qqmmiabl6 rrtrerugth betman the antibody ~~~lwul+ aRd tho anti&en surfwe configuration oanp3tmQatsry to that of the antigen. The rnle of ptarrinuray (the use of ths mfninm effort to achieve the xwmlt) -eat@ that them are only two such reg%onog that Ire that the antibudy realeaulsrr SF* at the mott blnerlent. The pzwpossit thbary la tsww% in this rwwmbls armqtlon. It wMd, of uourae, be posadbls to expand the theory in mwh a way aa to proride a rauchkirln for the formation of antibow latblsauler .tith ralsaue h%@er than two2 but thir woxilh antibodies of other olmaes 8~61 obvicms. The affect of (811 antigen fn dstomining tha rtruatlrre, af MI antibody that in the normal @o?mlia, ar amgmetsd by Breinl amuse, howtsver, that this ir not 110, but that all 4 antibodies with consfderkt4u tq;ecfficity for an apparently ualimitsd nurdwr of dfffemnt antigt2ne8, withcut the necessity of invoking 81~0 a variation in the T It hws been pointed ox& b A. Rothtm md I. Landsteirmr, Suieace x, 65 (1939)+ th8t th e posaibilitr of dlffment; wsys of foldinp: the uunc~ palgrt;rqptf&s abin to obtaia diffarsat antibodiew lo worth conaideri~. 1 *ss3 % I tio 0 t hii38 .--xl& us a.t30m tlz;nt Cha ~;lobuliri molemale cunsfsts of 8 rrlngle poQp3ptfde ain, contaLfnizg severs1 handma amirm-safrd realdues, and thst the ard#r of m&to-acirl x.6 i&ma is Budh twt far ths ountar sf the oM.n one of the msersibl6 amf@m~tianr ilar xwh more e&b%% t&-m w other+ w&rem the two end -rtrr of the chafn me of sash a aaprtmrs that theme oxlet for thea nasty cuafiguratitmer with aearly the emte eaerm. (TUB Faint irr d%rm.aeed fa dstafl 3s Section XV.) pour steps in ~IW poetuJ.atad p344@4 ai ftb=tw 0f 8 ~4-2 ~i0hiiin ~i44d4 434 fnu8t3ui4a 4n th bit da0 0f ~1~0 1. At atage I the polypeptida ahz&n hm hem qmtheaited, the amiao-ncid awiQuoa luwiw baea mm&alLed into the proyor ordez~, pembly tith the aid af ;xQymytldasee and protein ttm&&aa, SacT the two ends af the ah&x, h ma C, each WXW&XL~~ gezharp~~ tw kmzred r@AduBs, hme bow lfbsnt& with tha -z&able extended cmnfi@uatlon, (The horixont8l lin4 in 48431 dr4wi~ eqarater the rugion, h~low the line, in which the pGypeptlde -ia it not atizs to ahaage its aonf~tioa from the region, above the line, where this ir possible.) %ch of thaw &sin snds then coilrs q into the moat stable or one of the pgoat artable af the accerrnible coniigamtinnr (straQ;lt, 11) tend is, tied into thirr coafigxration by the fonarrtion af Q-drogen bmds ,znd other mak bondo between parts of the ohain. %`he central pati 1 of the ohain is then liberated (atap,e XXI> aad aammes' it8 atable folded wnfigkmstion (rtage IV) to give the cosrpleted There em rrlro indiaatsd 3.n lQu.re 1 rix stager in ths proasr# of form&ion . of lm aatiboag molo~r, In rtage I there sre shown an antigen molemle held at a platm of globulla pxwdtrcrtion and a @kabulin swlecmle with itr two end8 A aad 0 liberarted with the extreded w~igu.ration. At rtap If each of the ends har srmnned a stable ooiled wnfiguratian. These stable aoafigur&ions AD and Cl are nat, howawr, idoatiml with those A and C asemed in the absenas of the aatigm. The ataar and mupl which fonts the surfam of the antimw wlll attrast certain .oomplemntory part a of the globulfn chain (a ne~tivelpkarged group, far sxzusple, sttnrating a petitlmly-oharged group) and repel other ye,rts; as a result of these intemationcl the configwatianr At and C* of the chain endr which are stable in the pretmm of the euatigen a2td uhioh are aaaordi@y aemmed in the preseaao of the antigen will be euch that there is attmmtion between the ooiled globulin chain enda and the antigen, duis to their ooinplementarit~ Sm dww.re. Y2m BOPI figuration atsum by the ohain end nmq be any one of a large zmber, depending: upon whiah pax% of the aurfaco of the anhigm happenr to exert itr ifluence on ths chain sad tmd hew lwge a regiea of the starf%a happsnq to bs aorered bfr it. Vhen the aontml pm% B ef the globulin ChclSn is liberated from the plaae of Its ttynthealr (atage IIf), one of two procosser mafr ocuur. If the forcea of attraation batwean tha antigen aud the portions A1 and Ct are extmmely atmong, the7 till rmaia bonded to the aatigan for SD indefinite time, 9nd nothiw further of interert will happan. ff the foroer are somewhat maker, however, on4 will ih time break &wky-=-dir6ooiQtte frrna the axtti8en (eta@ IV). Then the portion B of the chain wiL1 fold up to aahiew itr noma stable configuration (stage V), asking a aompleted antibody swlwule. fn t ims thir will di5sociato f mm the aritigoa and float away (rtw vX>. It is rorsrible thart an auxili.wy mechanic 6 The middle part of the antib&y molemU thus gro8wwl umlb be WCS that af a normal glo'ltnrlin mohmite, and tha tmr bnda w&id have aonff@ratiaar n~ora cw lm3r camplQmQntary to part" of the BurfaCit of the antigen. 9h06tJ two aatirs Qmll am affwtiro in tiffsrant Birthions, w that, after the euatibbdr is compXstely Samad, oal~r ane of them at a time tmn &tnasp a partlaal~s antigen aobmula,. The mtig!m molecals, after fts itrssertfaa by the newly-foxwd antibody mt3lecule. may tmm as the pattern for another, and. uontlm6 to rem6 until it6 tmrfacs ia sarsmd by rary etrongly held ant2bodtse ot portiona of arntlbodfar or until tht~ conaentmtion of antibodies bsqmer 60 Cjred that emn with weak forear aptmating; the antigen is carmbfnad tith antibodier most of the tim (as illurtmted in l%um 11, or until the ant ie;sa ~~~leuul~ lr dscet zvyed or ermpes f mm the m&on of &.olmZin fomation. of the tturfacft of an antigen mlilbaale. Great Paristy in this consplamntary conflgnmtSon WI&% be expected to rmi%t from the acoidmtal appmxi;rrstfon to one or another mzfrma m&on, an& further varisty fmn variaticrn in poaftion a2 the antit?;en salmuls mlativa, to ?&a point of lfberatlan of tha &hulJs chain end md from acrrld6mtal eol;lfng and lhk%ng of the ah&% ~lrd before It QOBMHI m&w the inilnaaO6 of the i%digUa. %UQ of the mtibadzy m~]Leatxrau w&d be UXliYd6at. one of the ehain end6 having, bwauw of its too gr6at &i&awe fmtt the antigen, fold4 fnto a normal pJ&mXin eonfigtmstion. 'Bhl3"4t~rb~iCSti0~6 aT@ WXifi&i v 6%J?tWiJtl~tti XV3Sdt.tlt. It i6 W&3, bOWl3 that an i.ms a4arw~ to on6 suntigan wlI1, a6 a ti6, rmct with a mlatsb hetsmlogma antlgm, ouad that after ektmustion with the latter thare ramfar a frwtfon which till at113 react with the original antlfleza, S.4wdstafner iund van dtbr sa&a61r" , neixg as mitfgenr amprotein car3yixg marfour ha>tmnr mmtaini3qg 10 K. Lan&stsiner and J. van der S&mm, . a, 325 (1936). the $&se sctim gmwp, have &WWB tbt tha 24ntirsrt.m fOP on6 antig6n contab% varfouft fmctl6n6 diffe 7 l.`in& in tb sb%~@h of *hair atti~atibn fbr the hapten6. ;Sg tbe cpmtitativ8 study of prcbuipitin mactionr H~i~slberger and Km&&lrz M. Rsfdelbergw and F. E. &m&U, J. &QJ. Med. j&, 559 (1935); &, 467, 697 ?o??? o mached tha uwtm co~cltaaion, and thawad in addition that event after `prolm#I i-~p;sltion the antiaarw6 cstutied (anti-egg albumin) aontainecl lnuah luw-mds antibody, inctqmble br Itself 41 fomtin$ a precipitate with the antigen, but tith tht, pruprty of bdag card& Qua in the precripitata flamed vith a mom mactiw f ractid2. ____,_. 73. Th Biva of ,-ha theury 8truatUral Christ it is O&~MF that thi6 pi&IX0 of the prtmfpitate imat be tm?XWt. a T~Q gmst ~eaificsit~ of aurhlbody+mtig6n, fnt6matimr mqulraa th2Lt di dsfinite band be formed b&men an antibody m3bi?uls and an aatigm aolsc~le. If anti- badlea of mti@M6 v6re animblmt, this would lead to CObT$116%66 of one &nti@n ntalecmle mad one or nare mtibody molebats,a (or of arm antibody molsdrs md ona OT wre a&i@m mleculer), and m kmw f rem tPxperiaurm with protsina that them aggregates would in general IpmmiB in mlution, If both antibody and anti- am mltivalenCt howemrr the cmplex til3, gzcav to = aggragats of indeftite 8i66, aich is the pr6c;ipitate. This pXYOC666 i6 obwrved disaotly in tha %@.UtiRation of cell+. 0x1 th6 addition of su1 &&glutinin to a osll nzapm~ion the cell6 arta reen to clung together. ft la obvious that thn aggJutiti& aplecul6s which am hoMing tha cells tomther am biwLIent" -eat& bar two aotiw ends, with tonfigtm&ion complarwntary 40 that 23 Z%llowi~ H6idelber&w and Kendall, I u&e th6 terminology of chemical wlenrire theory in dirc!nsrifk& the ageaific nmtua~ attrmtton of antigen WI& antibody. The antibodpanttgen %alentse Imd# am not, of course, to b6 uwfu66d with ordinary comalent &umioal bando: they am due inatslrd to the integmtad weak forces distmwed in hs. Iv, of a partion of the &m6 of the cell6; the ag#Jutinin mol66ular hold th6 oellr topcethsr at their re&iom of a(lntabt, ~6 shtavn in Blgur6 2. &s-j. ft 66dm6 gmb~~ble that sell antilmifea ham thi@ 6truatum=-t&i+ they 6x6 bivalsnt, with their two aatiw ~&&WAS oppositely directed. H~bdslbar@s and Pmfeswr Eeid.el?mr@w baa fnibnted n&e that in their tp&ntitatioe treatsmt of d&a ML the prscipitin reaation he and Dr. &uui.aIl have found no incanpati- bllity vith this metricrtiont In their paper% they Bircutmed the fienmrral CXWJ of mltiv&mos of antibo0y am well AS of mitfgeY%. (ignoring the sxmgMm%l cam of the attackxmnt of two or nom anti@tns OT h,qtana to the mame wd r&on of ;urr mtibody) hzss not pretiously been made. 9 aa in the cuprwua o;rida arranhFaraQat (copyer and oxygen &vms beI= rqlaaed by 10 11 have the valsnca Jj = 12, If a31 regions of the euLtigen aurfave were active m;d if tha antibody mleemlee wem vphvrlva3.t t!m aseunption of elongated antibv& mvleculaa would pm&t the mlence to be vowwhat larger. The ValuB 12 of g oorwsyond~s to the W&XB 6 fVF the ratlo ,R. lor l8lpptr antigem ZaPger oaluer of ,B wotAld be axpmtedt ad for tm8llar one8 ller Y8luQs, hen for ant Lgem dth molscular weight as mall at 111300 the pradictad ~TB&RW mlns of 2 is 4 tfar syheriaal mtiWUar) or larger. In f0 Qnd 7001W0 lie tmtwesn &boat 2.5 for the smaller antiF;sna and 1% for the l.ar@m 5~~458 19 . It is man t-hat the walnsa of: are T9 It IS our sestr~crtian of the val%nue of the antikrdy$kthe Rlgdmsa v&m two whleh 1%&s to oar axplanatfon of the antlbdly-antl~n ratla$, `Fha g%taral observation of mkluer of g aem5lderabl~ greatar than 1 18 not aoatmktsd for lay a fmmmrk th%ary in whfch anti'body and anti- rnelsaalsa am both sultlvabnt, Ud488 8onw SW%i3ialFJr ~08tUbkf@ i6 inv&ad to &%&I the effeativa Vabxm~ of antl~n cansldmably greatar than that af aatlbodg, soaawhat fsas thaa th6 corr%8po~%ina; w~lamr from Table 1, which indi5ates that n5t all of the %WfaC% r%e;ians of th% antigsns am effsotiY%. '!h% data gi~@n la Table 2 ark thorre af Hsl%.elberger an% his craU.aboratore$ the ~1~68 reported by other iavestfgstora are ritmllsrr in rm&tra&e. Table 2 for PracipitatQ8 froa B&&bit kti8W@ IA 8 precipitate ftmB8d fram a Ml~tlon C@At&itiA#-$ AA mm611 of antibdly Aot an of the aastibo%y WA&JXl5*8 will be 8&ttWW?d. bt the limft cf antibody exam6 the grecsipitclts till be a network af linear ag®aterr 4th a rtrwsture mch a8 t&t rUpm8fHtted in ?&gim 4. &we &a& aAtig6A Btd.BtiS (tith am aooarioti axaagticm) f8 eurrrwnded w E antitmby iIl5lWU&38, OAly tw5 of which band it t5 nsi&borfn@ ant i@m IB516CtilQI. 'he prrddsd 8triIk($s forned in thie way ar8 tiod tug&her by an ouuartonal cra~.t-link to form the preicri~itate. l%e antibody- antifln muleaular Yatio $8 80%~~ to be eloa8 to ; - 1, aiich i8 OAO let&s than twicQ th8 value y2 for the nrlena+-tatarst;sd preuigitate. ai prdicted relertisn betuwn thssr ZVAtiO8 &t ibOdy MCQo88 - ~qubalema XcmIb - l $8 man from the %&t& in Table 2 to be verified ayprox~t81~ by expsrfmunt for the ant%gem other than thyme;lobulia. The bframpmay &awn by i+?qcro~l~bUlin is, fndwd., ta be expected for aa sntl@m tith raalecular weight @mt:4r than that of the antibocty. The reqnizvmanta of getmtryare such that an amaqymemt in which sachaxtt2gen irbondsd 14 Haicra for large antigen moleaules tha ~~Gecular rastio oikn exceed 6 fa the vale~5e- saturated preoipitatsr only if two or more antibody aaolsctllsrr are shared between the %WX@ pidr of a.attfym B?d@ti%%, Wh%M8 in th% ~tibo~--~xaO$e re&On the entire swf%c~e of the large antigen ssq be crav%rs% by antibody moleonles. varies between 1 m.td the tinismg value k/Z. XC i e, obserwd that in general no precipitate fame in the ragian af great antigen %m388s, and Heidelbsrpat aad hfr csallztbaratosrr hsrve in faot aroigned mlues 1 an% I/:! to the molerzular ratiae for the coxqler-ae in eolution. pchaX%fks pre&$tatiOn i8 iahibitud b i%ZhtigQn u4081, it UsaeLlly OC@kWI %Ttm with @`az%t aiatibody ~~888, althou& dohbf8 WQ&83C86 with n&mulaa ratio z and th8 8tWturS OhOWn iR ?i@U'8 7 tW8 %q@Ct%d tC 0%%8t. It %%@m% prCbmb&B that ths biffarmwm ia bekvi5r of ripItem in %he excm~b aatiasf, re&w aad 6~~8s~ antibodg retion fa to be attrlbutad t5 the fast that the m?i48C~h~ ratio for yreci?itate ;wn8 eolttble cxNEp,lex %iffers by a f~mtor if8 great i?k8 tw5 for the fmmer cease, an% by 5nlyIJ!@ - 1) for the latter, #b&i ?l b e d T 8 a8 the 'f Molecule.- Thsra are two ~pgs in which a~ antibody moleaulrs with two opposed arative, reg:ions w?lementaFg to the ant&&en @ght be prodwed. One is the wsy %ercribe% in &GtiOA If. !I?he other would invalrs the EIMmfactwe of the antibow molecule in its final confi.guwtlon h&&en two antfgen #ioleSUfee, one of which would eem S8 the, pa+,tcpm for OA% &Atibd.y an8 aA& thta other far the ee5ond. &Zt attempt t5 de.cide botwwn thase alte~&tivee eaezw to Hera been IWL%~ beforej thesre, exist8 evidence, howmm, soA6 of which ia mentioned below, to indicolte that the first method of antibody productioA, 1Avolvlng only one antikym niol8~ule, oocurs prdt%oMi?laAt ly. It is for this reauo~ th\.st I hwe d%velo?%d the rsther complicatad theory described above, with the two end :~orfiana of the antibody fanning first, me (or both) thQa saparatlA& frost the antfgtm, and the 5entml ;art of the anti- bo&y then r8sn'&4g Its &ape mui hold&g- the active ends la :~osition for attaohmztt to two tJ&i&OA mokculer. Thf6 thtt0~ X'O~Ui~ti th%t the fOZ'!tBtioA of antibody b% 8 rtWCtiQA of th8 first order with r~:spect to the aAti@A, w!ww&e the other nlternatiPe would require it to be of the second arder. There 4%istt very little o+fRence ati to whether on immmiaatioA wfth i~~~btll. aAotmta of antigctn the antibody production is ~rwpartional to the anoant of antigen iajected or to ita BQS;"Q), S5me tbupport for the one-antigen-m5lemle theory is provided by the exp%tieAta dealiqg with the hjeotlon of a XbiX+tW@ of aAti&eAs. Xf two antigea m0hsl81) w%re requ$red for antibody fomw.tion, it wmiL% be @%pe~tebd that antibodiear At-Et, At-C*, EI*-Ca,~~* complementary to two different Antigms A md 23, A Eurd C, 3 and C,-ae a# well a8 those A'-A', B'-38, C'-et,... CoE@BB%At9~ to a 8iA&% iSAtf@3A wOti% be ionned. The CSV~ihh~8 eviclence speaka strcmgly qainst thirs, !&ua Deep, Taylor, Rnd hdafr 20 hstve reported that th% s%mm produaed by immmlzation with a alxtww H, 1. Deal, Q. L. Taylor, and M. P, Adair, k. *giens 3, 69 (1935). of egg albunin and s%ntm aDu~ia oantaina distinot antibodies h5PablogOus to the two aatigtm6, +%A& that proaipitation dth 5nu aatigsn lsanss the #mouAt of the hetsrologous antibody ansltersd, ah @WA SKWO ri@!WUS d~Af.!It~tiOA t#%I Yumished by Reidelbergm and &data, who, from the @%rum of a cow whioh had N, R%idelbarg%r and E. A. G&at, J. Xxp. we?. 6& U51 (19$5). been iA,jecteh with types I, 11, rarrd 1x1 pA%umococci, isolated in mooeselon, with the oorraspondiae; tpeaif tc -pol$laactiri%a, the three antiosrbahgarates, 8acpfr in a~ appawntly -pure state rsnd tith 110 a~pr%*oiabLe 5rore-mm5tivltg as to 16 pnmmocaacur tpa . gn another striking experiment Rektaen an4 Boor 22 fcmld `22 L, P-aktoen wnd A. K. Boar, J. XRfect, Dia. s, fjB8 (1931). that the smwm obt?+iasd oa irtjectiw a rabbit w%th ar. mixture of 35 antIgear reacted with 34 of the aatlgens, zsnd that absorption with any me had in the main little effsat on eubeeqymt mctfon with another. &we oa the tw-Lantigen- molecule theory the wwmnta of antibodies A&k', 33*-B', o . . cqmble of cmsixyl; precf?ltation with B sfnale ant&en mmld be ~:~11 co-ed with the total amount of antibody (of the order af I&, for2 mtigene-about J$ in thfs care), thersa qualitative observatianr pravida significmt e~5donce in favor sf the alt sraztlve theory. B `3 ,-There has lmn m&ensiva, d.lrsaussion of th6 qwtstion of what laake$ a nzbstawloe CUL antigen, but x~o gen@rnLly aoce:-Aad concltislana hsve `bean renuhed. Our throw -pmnit$ the farmlntfm of the folkw- lng maenmble ariteria for eratig~~%c ,zctitityt 1. The antigen molemle mot contain attim .qmups, cs%ble of mfflciently etrorf intar-.r)tion with ths &~bulin chsin to irrirttance it8 mniigwatian. 2. The ceafipratloa cd' the anttgen aalemile must be well-defined over snrfamt regions larget snowh to glw rfee to ,m integrated antiboP+antim force mfffaPent to hold the m&m&m together. 3. The zntigan molecule must be larce enou& to have two OP more such surf~,ce m&ons, and ia ceree th%t the antfgenie activity de:mnds upon a pwtfmlar aacaharida~s, invertebrate &m0&@+8k~~), and txme lipide and carbohydrate-lipid 17 prvttAns in the body. Man-aatfgenic mabstmam km been rqmrted to beaome aatZgmi(3 Vmen adsorbs& on pwtiolea (Farsmwn antigen on kavlin 24 )g in this caw the ymrtiale with adsorbed haptem is to be &mridemU the antigen *m~~eoule~ of our theory, I pmwllct that raletiwly simple mvletiee ovntaining two vr more -tens till ha fotmd to be antf,genic: e~xpmkaeats to test thla preifietion are now ttmlar my. ctut mae pmro a@# 26 that the well-UefinmI prvpvrti-ts of n&im protsiaa rmplm a A. 3. Nireky snd L. Pauline, Prvc. Ghiness 3. PhyriaL 3, 32l.