CHEMO-IMMUNOLOGTCA.L STUDIES ON CONJUGATED CARBOHYDRATE-PROTEINS ~11. WOLOGICAL SPECIFICITY OF ANTIGENS hEPARED BY CO&t- B-G a- AND ~GLUCOSIDES OP GLUCOSE WITH PROTEINS BY OSJVALD T. AVERY, M.D., y%THEF. GOEBEL, PH.D., AND FRANK H. (From tke Eospitd of The Rockefh In&h& fw Medical Rmarck) (Received for publication, February 24,1932) The study of antigens prepared by chemically combining sugar derivatives with protein has revealed the fact that carbohydrates exert a determinin g inlluence on the immunological specificity of the compounds of which they form a part. In preceding papers evidence has been presented showing that mere differences in the structural configuration of a single carbon atom may determine the serological specificity of two sugar derivatives identical save in this one respect. As previously pointed out, the p-aminophenol &glucosides of glu- cose and galactose, differing from each other only in the interchange h of the H and OH groups on a single carbon atom, exhibit an individual specificity irrespective of the protein to which they are attached (1,2). Moreover, the specific r8Ie of the sugar radical in the reactive part of the antigen is revealed by the fact that the glucosides alone, unattached to protein, specifically inhibit the precipitin reaction between the corresponding antiserum and homologous antigen. The inhibition of the antigen-antibody reaction by the intervention of the specifk sugar hapten was found to occur both in vitro and in Go. In guinea pigs passively sensitized with the precipitating serum of immune rabbits, the fatal anaphylactic shock which invariably f&lows the intravenous administration of the homologous sugar-protein was completely prevented by a single prophylactic injection of the speci& glucoside alone. The uncenjugated glucosides, although &msdvea not capable of inducing shock, specificaJly inhibited the anaphyhxtic reaction in sensitive animals when given immediately before the introduction of the shock-producing antigen (3). The specifxity of the &lucosides of glucose and galactose was demonstrated by the fact that the inhibition of the anaphylactic reaction, like the inhibition of the precipitin test in vitro, occurred only when the glumside used was the same as that present in the specifically reacting antigen. 769 770 CONJTJGiWZD CARBOHYDRATE-PROTEINS. VII These earlier studies concerned themselves with the speciiicity of the fi-glucosides of two different hexoses, glucose and galactose. The present paper deals with the specificity of two different glucosides of the same sugar, namely; the paminofihenol .a- and /3-glucosides of glucose., :: :,I ; .:; .!.`.' 1, j. * The "CY and fl antigens" formed by the union of the diazonium de+ rivatives of these glucosides with protein d$fer from each other only in the spatial relations of the terminal carbon atom in each sugar component. Since I the ultimate composition of both antigens is chemically the same, any observable differences in their immunological specificity can be referable only to known differences in the molecular structure of each. The follo+g experiments were carried out to determine to what extent' two isomeric glucosides of the same sugar, in combination with a single protein, would influence the specificity of two antigens otherwise identical. The present paper deals with the question of the specificity of the precipitins present in the serum of rabbits immunized with CC- and p-gluco-globulin. EXPEEIMENTAL. Methods The synthesis of the a- and @+-aminophenol glucosides of glucose and the method of coupling the diazonium derivatives to protein are given in detail in the preceding paper (4). The method of the intravenous immunization of rabbits and the technique of the precipitin and iuhibitioi tests differ in no essential respect from those described in the earlier studies (2). The inwnuniaing a?zfigens were prepared by combining each of the glucose deriva- tives to globulin derived f&m normal horse serum; on the other hand the tGst a&gm used in the precipitin reactions, contained the respective glucosides bound to the proteins of chicken serum. The use of proteins of remote biological origin in the two sets of antigens emludes the possibiity of amunon antiprotein precipi- tins masking the specificity of the carbohydrate reactions. I. Anticarbohydrate Antibodies (1) Specific Precipitin React&c-The sera prepared by immuniza- tion of rabbits with a- and B-gluco-globulin were first tested for the presence of precipitins specifically reactive with an antigen containing the same carbohydrate radical. The test antigens employed in the precipitin reactions were in each instance prepared by combining the 0. T. AVERY, W. F. GOEBEL, AND F. H. BABERS 771 rqxctive glucoside with a protein unrelated to that used in the im- m&&g antigen. Under these conditions, the results of the tests are not confused by the presence of a common protein and the specifkity of the reactions are directly interpretable in terms of known differences in the chemical structure of sugar radical in the reactive part of each antigen. TABLE I @&f.c Pmci@itim in Sera of Rabbits Immunisd with u-lZluco-GloWi~* J Tat antigen: usltlachurur -* ~UUC r&it - 1:5,000 1: 10JJoo 1:2o,lm 1:40,000 1 ++++ +++ ++ +j= 2 f 3 zi++ E -ii + ++A-+ - Complete precipitation with compact d&like precipitate. * III order to avoid the reactions of a common protein, the irnmvnising antigen was prepared by combining a-glucoside with the globulin from horse serum, while the test antigen was similarly prepared by coupling the same glucoside to the pro- teins of cllickeIl serum. TABLE II S#e+ Precipitins in the Sera of Rabbits lmmuniscd with P%4co-Globulin Tat u~tigen: ~&uoJ~~ amum* Immuuc rabbit aera l:S,OOO 1: 10,ooo 1:2o,ooo ,1:4O,OW 1 ' +++ . ++ +h 2 ++++ ++++ +++ +:+ 3 ++++ ++++ +++ +++ + + + + y Complete precipitation with compact disk formation. * Unrelated proteins in immunizing and test antigens (see footnote, Table I). The precipitin reactions between the homologous test antigens and the antisera of rabbiti immunized with 4~- and &gluco-globulin are ' given in Tables I and II. . . `. The results of the prekipitin tests illustrate the capacity of the immune sera to react with an antigen containing the homologous glu- coside irrespe&ve of the protein to which it is attached. They con- firm the earlier observations concerning the orienting influence of the 772 CONJUGATED CAXBOHYDRATE-PROTEINS. VII sugar radical on the specificity of the protein with which it is com- bined. . (2) Specijit Inhibition Tests.-The selective specificity of the anti- bodies reactive with the two isomeric derivatives of glucose is clearly demonstrated by the results of the inhibition tests given in Tables III and IV. Analysis of the data presented in Table III shows that the ar-gluco- side alone, when added in excess to the homologous antiserum specifi- cally binds the precipitins and thus renders them unable to react subse- TABLE m SpGcifiG Inhibition of PrGGipitins in dXm-G~obu&a Ad..wum by Homologous Glucoside cc. 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 = I Glddrs n/U I' B .- cc. salt solution to volume cc. ' 0.3 0.3 2 0.3 ti 6 0.3 - 0.3 - ; 0.3 Pp - . - + + + - Disk-like precipitate 0 - No precipitation, showing complete inhibition. * See footnote, Table I. = I . _ - Redt +++ f++ f++ 0 0 0 ++* -I-++ +++ quently with the test antigen containing the same sugar derivative. The specificity of this inhibition is shown by the fact that the addition of an equal concentration of the @glucoside to CT immune serum has little or no appreciable effect on the precipitin reaction with the Q antigen. Similar relationships are shown to exist in the spe- cific inhibition of the precipitins in fi antiserum by the B-glucoside (Table IV). The results of the inhibition tests reveal the specikity of the anti- carbohydrate reactions, since in the case of both (Y and B antisera the precipitins for the homologous antigen are inhibited only by the corre- 0. T. AVERY, W. F. GOEiEL, AND F. H. BABERS 773 spending glucoside. Although the glucosides by themselves, un- attached to protein, fail to produce precipitation in immune sera, each by binding the specific antibodies in its homologous antiserum prevents them from again reacting with a test antigen containing the same glucoside. In this sense the glucosides unattached to protein function as haptens, the chemo-specific groups of which unite with the corresponding antibodies without causing visible change in the reaction mixture. Marrack and Smith (5) using a direct method have TABLE IV ,Qecifi Inhibitian b Homologw Gk~osidk of PrecilUins in &Gk&obtdin = Ghoslda ~113 cc. 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2. 0.2 +++T- = Complete precipitation with disk-like precipitate. 0 = No precipitation, showing complete inhibition. * See footnote, Table I. f `, (z cc. 0.3 0.3 0.3 = I B cc. 0.3 0.3 0.3 salt aokloa 0 volrunc cc. 0.3 0.3 0.3 A = , I, 1 -- ~ - recently brought experimental proof of Landsteiner's original view that the union between hapten andantibody is a specific combination between the chemo-spe& groups of the rkcting~agents. As Land- Steiner has pointed out in the case of non-protein radicals which func- tion as haptens, the union of these glucosides with antibody does not lead to precipitation unless the reacting groups are attached to protein. The preceding experiments have dealt with the homologous immune reactions of two antigens of. knoti chemical constitution which differ one from the other only in the CY and @ type of glucosidic union. The results show. that mere dEerences in the spatial arrangement of the 774 CONJUGATED CARBOHYDRATE-PROTEINS. M groups on the terminal carbon atom in these two derivatives of. glucose sufKce to confer on each antigen a marked degree of difbzrential spec&ity. For, as shown by the specific inhibition reactions, it is possible to differentiate selectively between the two isomeric glucosides of the same sugar by the use of immune sera. Despite the reflection in serological specificity of differences in the stereochemical relationships of the first carbon atom bearing the non- sugar constituent of each glucoside, it must be borne in mind that in both the a! and ,9 derivatives the spatial arrangement of the polar groups on the r emaining five carbon atoms is identical. Quite dif- ferent, however, are the structural relations existing in the case of the @-glucosides of glucose and galactose, the immunological specificity of which has previously been shown to be absolute (2). In these latter glucosides the stereochemical arrangement of the groupings on the terminal carbon atom are both in the @ position, but the symmetry of the polar groups on the remaining five carbon atoms is not identical, for on the fourth carbon atom of galactose the H and OH groups are rotated through an angle of 180o. This difIerence in molecular con- figuration, with its consequent change in the spatial relationship of the polar groups, is the important if not the sole factor determining the individual specificity of the @-glucosides of glucose and galactose. On the basis of these facts, one might anticipate an absolute specificity in the case of the similar glucosides of two different hexoses, but an overlapping specificity in isomeric glucosides of the same hexose. (3) Cross-Precipitin Reactions.-It is therefore not surprising to find that this partial similarity in the chemical structure of the (r and @ derivatives of glucose is reflected in the serological relationships of both substances. The results of the cross-precipitin tests presented in Table V show that antibodies present in a and B antisera cross-react in each instance with the heterologous test antigen. The cross-precipitin reactions are sharply defined and are quantita- tively only slightly less than the reactions between each antiserum and its homologous antigen. Interesting relationships in the overlapping speciflcity of both sugar derivatives are brought out in the reciprocal inhibition tests, the results of which are given in Table VI. Analysis of the data presented in Table VI shows that the addition of cr-glucoside to Q! antiserum ,inhibits the precipitins from reacting 0. T. AVERY, W. F. GOEBEk, AND F. H. BABERS 775 witi both a and @ tat antigens. However, the addition of an equal concentration of &glucoside to a antiserum inhibits &e precipitin TABLR V CroJs-Preci#t*tin RediofzS of a- and &%U&~obdin Antisera with Hete&og~ Antigens Test atuigam* Immuoo- a-glucoshIcLn - b-g- serum 1:5,000 1: 10,Ow 1:2o,om l:S.am 1: 10,ocnJ l:2O,Ow ------. a-gluco-globulin. . . , . . . . . . . . . ++++ +++ ++ ++* ++ &glum-globulin. . . . . , . . . , . . . . ++ ++ + ++++++++ +:+ `++++ = Complete precipitation with compact disk-like precipitate. * See footnote, Table I. TABLE VI Inhibitiun OJ ' Cross-Precipitin Reucticma I I Homologous and Huterotogm us = I GZucosi& Glued s. o! 0.2 P 0.2 a0.2 uo.2 a0.2 a0.2 @ 0.2 80.2 8 0.2 6 0.2. 8 0.2 80.2 GlUdo x/l.5 ' a 8 -- cc. cc. 0.3 - 0.3 0.3 - 0.3 -- 0.3 0.3 - 0.3 0.3 - = I 2A to volume 04; 0.3 0.3 0.3 ,- I Tat mti,gcn' 1: 10,000 8 cc. 0.5 0.5 0.5 0.5 0.5 0.5 A Basulta +++ +L ++ 0 0 ++++ 0 +++* ++ 0 0 Plus signs indicate positive reaction with gradation in amount of precipitate. 0 = no precipitation, showing complete inhibition. - = reagent not used. o See footnote, Table L ,. reaction with the /3 test antigen without appreciably diminishing its capacity to precipitate the a antigen. Similar relationships hold in 776 CONJUGATED CARBOHYDRATE-PROTEINS. VII the case of the @ antiserum with respect to the inhibition of precipitins by the homologous and heterologous glucosides. It is evident from these results, that the addition of the homologous glucoside to its antiserum completely inhibits the precipitins for both the homologous and heterologous test antigens. On the other hand, the addition of the heterologous glucoside to immune serum inhibits only the precipitins for the heterologous test antigen and has but slight effect upon the antibodies reactive with the homologous antigen. The results of the cross-inhibition tests with the a- and @-glucosides show that the reaction of an immune serum with its homologous anti- gen is specifically inhibited only by the homologous glucoside; while the cross-reaction between this serum and the heterologous antigen is TABJaE VU Precipilins for Globulin in Sera oj Rabbits Immun&ed with cy- and &GkuoGlobdin* Immune ma I Globulin from horse - 1:s.ooo I 1:10.ooo I 1:20*wm I 1:4o.o00 cc-&o-globulin. . .,. . . . . . . . . . . . . . . . . . . . . . . 8-gluco-globulin . . . . . . . . . . . . . . . . . . . . . . . . . . . + + + - flocculent precipitate, not compact. * Both immunizing antigens were prepared by combining each glucoside with globulin from horse serum. completely inhibited by either glucoside. This lack of reciprocal inhibition of the precipitins in LY and @ antisera may be interpreted as further evidence of the lack of the immunological identity of the two isomeric glucosides. II. Antijwotein Antibodies As previously described the immunizing antigens were prepared by combining the cx and 0 derivatives of glucose with globulin obtained from normal horse serum. The sugar-protein antigens were standardized on the basis of their nitrogen con- tent so as to contain 5 mg. of protein per cc. Each rabbit received during the course of immunization a total of approximately 160 mg, of protein antigen. In addition to the antibodies reactive with the specific glucosides, there are also present in both immune sera precipitins for the globulin used in preparing the immunizing antigens (Table VII). 0. T. AVERY, W. P. GOEBEL, AND I?. H. BABERS 777 The precipitating antibodies for horse serum globulin may arise in response to the presence in both antigens of free protein unbound by the &a~~t.id glucosides. However, it is also possible that even in the absence of free protein, the sugar-protein antigen as a whole may stimulate the formation of two qualitatively different antibodies, each specifically related to the corresponding component of the antigenic complex. This concept of the dual antigen+ of a single complex antigen involves the assumption that the coupling of the glucoside to the protein has not masked the groups essential to the specificity of the protein, and that while the sugar radical through conjugation acquires specific antigenicity, the protein molecule retains, in part at least, its original antigenic properties. DISCUSSION Landsteiner and Lamp1 (6), working with ortho, meta, and para substituted aromatic compounds, have pointed out the significance of the spatial arrangement of the groupings upon which the speticity of these compounds depends. In more resent studies on the serologi- cal differentiation of steric isomers of p-aminobenzoylamino acetic acid and tartaric acid, Landsteiner and van der Scheer (7) have further emphasized the dependence of immunological specificity upon chemical structure. The results of the present study add further evidence in support . of the view that the immunological specificity of carbohydrates is determined by their chemical constitution. Di&rences in the specific behavior of the a- and /?-glucosides of glucose may be accounted for by known differences in the spatial position of the groups on the first carbon atom of each glucoside. These structural changes are so sharply reflected in serological specificity that it is possible by means of immune sera to diiIerentiate selectively between the two isomeric glucosides of the same sugar. However, granting the difference in the chemical structure of the two glucosides it becomes necessary to account for their overlapping specificity. The chemical basis for this immunological crossing may lie in the fact that the spatial arrange ment of the polar groups on the mmaining five carbon atoms is identi- cal in both glucosides. This partial similarity of molecular grouping might then account for the degree of immunological likeness exhibited 778 CONJUGATED CARBOHYDRATE-PROTEINS. VII by the two substances. Although in a portion of the molecule of both glucosides the structural relationship is identical, nevertheless they `behave chemically as separate entities and possess serologically a separate and distinct specificity. If Complete reciprocal inhibition of precipitins is accepted as the criterion of s&ological identity then the failure of both glucosides to exhibit ,this capacity may be taken as flirther evidence of differences in the immunological specificity of each. The lack of reciprocal absorption of aggltitinins by two organisms mutually agglutinable in the immune serum of each is generally con- ceded to indicate a lack of immunological identity. `Relationships of this order are known to exist between Pneumococcus Type II and Friedlgnder bacillus Type 13. In both instances the reactive sub- stance has been identified chemicallly as the specific poIysaccharide peculiar to the capsule of each organism (8). While the structural constitution of these complex sugars is not as yet fully known, con- siderable knowledge has been gained concerning their chemical prop- erties. For example, it is known that the Pneumococcus Type II polysaccharide is built up of glucose units and that chemically it bears a close resemblance to the polysaccharide recovered from the Type B FriedEnder bacillus. However; the two substances are not chemi- cally identical, although the similarity between them is sufficient to result in a certain likeness in immunological speciiicity (9). In the absence of precise knowledge of the structural relations of the two .polysaccharides, it seems reasonable to assume that both contain in a portion of the complex molecule the same or a closely similar configura- tion of atoms. This &nilarity of molecular grouping might then account for the immunological similarity of the two substances. Considerable evidence for this point of view is found`in the results of the present study concerning the specificity of the cy- and /3-gluco- sides of glucose. A, comparison of the serological relatiohships be- tween the two isomeric. derivatives of glucose and the capsular poly- saccharides of the two organisms in question is presented inTable VIII. Analogous relationships ex,tending even to the cross-immunity?eac- tions and the lack of reciprocal absoi-ption and inhibition of antibodies are evident-in both the groups of simple and complex carbohydrates. While the comparison is instructive it is not necessarily, valid in .the case of the capsular polysaccharides, for the final intekpretation of 0. T. AVERY, W. F. GOEBEL, AND F. H. BABERS `779 these relationships must await further knowledge of the structural relations of the more complex bacterial sugars. TAtiLE VIII Comparison of the Serological Re&ztimship between I$omeric Derivatives of Glucose and the' Capstitar Polysaccharides of Two Unrelated Species of Bacteria Antipneumococcus serum & II . a&co-globulin antiserum By addition of Precipitios are inhibited for I lot inhibited fool _ - a-glucoside a and fl test antigens &gIucoside @ test anti- a test anti- gen gen ,. @-gluco-globulin antiserum /3-glucoside Anti-jriedlaendcri serum Type B - Precipitins are Agglutinins `are : . Absorbed with inhibited for 1 not inhibited lo] removed for not removed for - . a andi test *. B.. frkdt&z- Pneumococ- antigens deri Type cus Type B." `.' II B. fried&n- deri Type B a-glucoside `a test anti- B test anti- Pneumococ- Pneumococ- B. friedlam- gen en cus Type cus' Type deri Type II II ,B, - Absorbed with Pneumococ- cus Type II i.. ffiedaen- deri' Type B Agglutinins are removed for I not removed for Pneumococ- cus Type II B. fried&z- deri Type B, B. f&d&m Pneumococ- deri Type cus Type B II The test antigens in all instances were prepared by cdmbinhg tie respective glucoside with a protein biologically unrelated to that in the imm&ing antigen; see footnote, Table I. Enders (10) has ,recently suggested that crdss-relationships between Type II Pneumococcus and' Type B Friedlgnder bacillus `may be. `180 CONJKfGATED C!ARBOE!i!D~TE-PXOTEINS. VII _ due to the presence in pneumococci of a type-specific agglutinogen unrelated to the specific carbohydrate, ;.., `.`i -;. . . In the case of the synthetic antigens containing the a and fl com- pounds of glucose alone, the evidence indicates that the immunological relationships of the reactive glucosides are determined by known variations in their chemical constitution and are independent of the protein to which they are attached. In view of these findings it seems not unlikely that in the case of the polysaccharides, because of their more complicated structure and the greater possibility for variation in molecular configuration, there may be found many examples of a sin&r overlapping specificity among carbohydrates of unrelated OIigiIL S-Y The chemical and immuno1ogica.l properties of the p-aminophenol a- and @-glucosides of glucose are described and correlated. The results are discussed with reference to their possible bearing on the chemo-immunologicad nature of the speciftc polysaccharides of bac- terial origin. BIBLIOGIUPHY 1. Goebel, W. F., and Avery, 0. T., J. Exp. dbed., 1929,50,521. 2. Avery, 0. T., and Goebel, W. F., J. Exfi. Med., 1929,50,533. 3. Tillett, W. S., Avery, 0. T., and Goebel, W. F., J. Exfi. Med., 1929,50,551. 4. Goebel, W. F., Babers, F. H., and Avery, 0. T., J. Ex$. Med., 1932,55,761. 5. Marrack, J. R., andsmith, F. C.,Nawe, 1931,128,1077. 6. Landsteiner, K., and Lamp& H., Biochem. Z., 1918,86,388. 7. Landsteiner, K., and van der Scheer, J., J. a?&#. Med., 1928,48,315; 1929, 50, 407. 8. Heidelberger, M., Goebel, W. F., and Avery, 0. T., J. l&j. Med., 1925,42, 701. 9. Avery, 0. T., Heidelberger, hf., and Goebel, W. F., J. E.@. Med., 1925,42, 709. 10. Enders, J. F., J. Ex#. Med., 1932,55, 191.