INFECTION AND IMMUNITY. Spt. 1983. p. 1234-1244 0019-9567/83/091234-11$02.00/0 Copyright 0 1983, American Society for Microbiology Vol. 41. No. 3 Precipitating Cross-Reactions Among Pneumococcal Types MICHAEL HEIDELBERGER Department of Pathology, New York University Medical Center. New York, New York 10016 Received 20 April 1983/Accepted 7 June 1983 Data accumulated over many years are brought together on cross-reactions of 46 among the more than 80 pneumococcal serological types, with the idea of correlating cross-reactions with the structures of the relevant type-specific capsular polysaccharides, insofar as these have been determined. The precipitin reaction was carried out with the polysaccharides and antibodies raised in horses, rabbits, and a mule. Quantitative values (micrograms of antibody nitrogen per milliliter of antiserum at 0 to 1°C) are given in many instances and discussed, together with arbitrary qualitative data, in terms of the known structures of the polysaccharides. Some precise relationships are uncovered, and an attempt is made to determine why some of the cross-reactions are reciprocal and why others are only unilateral. Even before the structures of any of the capsular type-specific polysaccharides of Strep- tococcus pneumoniae (pneumococci) were known, unilateral and reciprocal cross-reactions between the many serological types were re- corded (46, 58) as a result of attempts to raise type-specific diagnostic antisera in rabbits. The development and use of purified microbi- al polysaccharides as vaccines are presently expanding. There is also anticipation that dem- onstrated cross-reactivities might afford some measure of cross-protection. These consider- ations and the rapidly increasing knowledge of the structures of the relevant polysaccharides prompted this collection of published and hither- to unpublished data accumulated over the years. Cross-reactions among the pneumococcal (Pn) types, as demonstrated by the precipitin reac- tion, are shown in the tables. Data, mainly quantitative, from earlier papers have been add- ed to make the tables more complete and be- cause recently identified chemical structures of the Pn capsular polysaccharides have made some of the previously described cross-reac- tions more explicable. Quantitative values of antibody nitrogen precipitated are given for most of the more marked reactions. Data are given for cross-reactions of 46 serological types of S. pneumoniae with anti-Pn sera of 36 types. Cross-reactivity was measured by precipitation of antisera by type-specific capsular polysaccha- rides. Although the precipitin reaction is rela- tively insensitive, this was considered an advan- tage since positive results might be all the more significant in relating chemical structure to im- munological specificity, the main objective of the study. Predictions of structural features of Pn polysaccharides of unknown structure could often be made from cross-reactivity in antisera to types for which the structure of the capsular polysaccharides was known. In most instances these predictions were verified by subsequent studies. The cross-precipitations are assembled in Ta- ble l, and the reciprocal ones are brought to- gether in Table 2 and treated separately. Insofar as possible, the results are discussed in relation to the chemical structures of the type-specific capsular polysaccharides involved as dominant immunogens in the raising of antisera and as antigens in the cross-reactions. MATERIALS AND METHODS All tests were carried out at or near 0°C to obtain maximal values. My co-workers and I have repeatedly called attention to the usually larger differential shown by cross-reactions between 0 and 37°C than is ob- served with homologous precipitation. This is because cross-reacting antibodies usually combine with only a portion of the repeating unit of a polysaccharide, sometimes a single sugar. It therefore is not surprising that cross-precipitation at low temperatures does not always parallel immunogenicity in humans or other animals at 37°C or higher (cf. reference 41). An ex- treme example of the difference in precipitation be- tween 0 and 23°C is given in reference 19. Methods used and sources of most of the polysac- charides and antisera have been described in the papers cited. Antisera containing appreciable amounts of antibodies to the groupspecific C-polysaccharide were absorbed with this polysaccharide before use; this is denoted by the letter C after the number of the serum. Qualitative tests were allowed to stand in a cold room at 3 to 5°C for 6 days or more and were read on a scale of - to + + + + ; quantitative determinations were placed in a cold bath at 0 to 1°C for 2 days to 2 weeks or longer, depending on the rapidity of precipi- tation. Qualitative data are given within parentheses 1234 VOL. 41,1983 CROSS-REACTIONS OF PNEUMOCOCCAL TYPES 1235 when canied out in antisera which were negative or weakly positive in other tests. Quantitative figures within parentheses refer to single analyses instead of the usual mean of duplicates. The following abbrevia- tions are used: S. specific capsular polysaccharide; Gal. galactose; Glc, glucose; Rham, rhamnose; Man, mannose; GalA, galacturonic acid; GlcA, glucuronic acid; GlcNAc, N-acetylglucodne; dp. depyruvylat- ed. Except as indicated, the antisera used had been raised in horses. All sugars are in the pyranose form unless indicated as furanoses by f, as in Galf. RESULTS AND DISCUSSION Cross-precipitation of other Pn type-specific polysaccharides in anti-Pnl. S3 gave small pre- cipitates in three of the four equine anti-Pnl sera tested: 27 pg of antibody nitrogen per ml at 0°C was the largest amount, Nith 884C. S33 and SlOA (34 US.) precipitated 10 and 27 pg of N, respectively, from 1057C. Although the signifi- cance of these small reactions is not clear, a D- Galf residue of SlOA (11, 50) might fit into antibody spaces designed for S1, the structure of which is given (44) as: f*3)-2-acetamino-4-aino-2,4,6-trideoxy-~-Gal -a-(l -P 3)-D-GalA-a-(l -+ 4)-D-GalA-a-(1-&, S2, S5, S6, S6B, and anti-Pn sera (references 17, 18, and 28). The structure of S2 is (37): appreciable (18), SS failed to react in the anti- Pn2 sera at hand. S6B differs from S6 only in its 1,4-linkage between ~-Rham and ribitol (36); in S6 the con- nection is 1,3- (52). If the specificities are deter- mined almost entirely by the sugars, this would explain why S6 and S6B precipitate practically the same amount of antibody from anti-Pn6. Otherwise S6B cross-reacts very weakly (+) only in anti-Pn28 and therefore is omitted from Table 1. Reciprocal cross-reactions of type 2 in anti-Pnl9, anti-Pn20, and anti-Pn23 are given in Table 2 and discussed in a separate section, as are the other reciprocal reactions encountered. S4 and dp S4. Removal of pyruvic acid from its acetal linkage on positions 2 and 3 of D-Gal (43) in S4 occurs without marked degradation, but dpS4 precipitates far less antibody (21) from anti-Pn4, showing that the 2,3-pymvyl-D-Gd is immunodominant. However, dpS4 is more cross-reactive than S4 (Table l), owing to the liberation of two hydroxyl groups of the Gal. Moreover, dpS4 precipitates C-reactive protein and antibodies to Pn C-polysaccharide almost quantitatively (33). N- Acetyl-D-galactosamine is the sugar common to PnC and dpS4, but the extent of the reaction is unusual if only a single sugar in the repeating units is actually the cause. The 1 ,Clinked D-Gd (32) in dpS4 presumably is 3)-~-Rham-a-(l + 3)-~-Rham-a-(l ---* 3)-~-Rham+-(l ---* 4)-~-Glc-a-(l ~-GkA-a-(l + 6)-~-Glc-a-(1?~ Accordingly, there is no 1,4,&linked Glc, as previously thought, nor is there GlcA in the main chain, so the discussion of the Pn2-Pn20 cross-reaction in reference 30 requires revision. The immunodominant group of S2 is the isomal- touronic side chain (29), making it possible that the phospho-a-D-glucosyi residue of S20 (52a), responsible for the acquisition of cross-precipi- tation in anti-Pn8: intact S4 does not precipitate. Reactivity of both forms in anti-Pnl4 probably occurs because the 1,3-linked N-acetyl-D-galac- tosamine of S4 (34) fits partially into antibody sites designed for 1,3-linked D-G~. S7 and S7B. Although the structure of S7 is L (-OAc, partial) OH An with or without the adjacent a-1.6-linked D-Gk, might fit partially into combining sites on anti-S2 molecules designed for insomaltouronic acid. As S18. like S2, has Glc in 1.4- and 1 ,dlinkages and 1,flinked L-Rham in addition (161, it is not surprising that S2 precipitates anti-Pnl8. The reverse reaction does not occur with the anti- Pn2 sera available. A tentative structure for S5 retains the 1.2- linked D-~~cA (B. Lindberg et al., unpublished data). Although precipitation of antiPn5 by S2 is not yet fully known, nonreducing lateral end groups of D-G~ in the repeating unit have been identified (lo), as surmised in reference 30 and predicted in reference 56 because of the cross- reaction of S7 in anti-Pnl4. Precipitation in anti- Pn19 is probably due to the 1.4-linked D-Gk, as well as to 1,Zlinked L-Rham in S19 (47) and S7. The long-known reaction of S7B in rabbit anti- Pn24 (7) was confirmed. Several other weak cross-reactions were observed (Table 1). The structure is not known. 1236 HEIDELBERGER INFECT. IMMUN. TABLE 1. Cross-precipitation of Pn capsular polysaccharides in anti-Pn sera Cross-precipitationb in anti-Pn: 1 2 3 4 5 6 7 8 9 10 11 Polysaccharide" Homologous s1 2 3 4 dp4 5 6 7 7B (48 U.S.) 8 9 9A (33 U.S.) 10 10A (34 US.) 11 11A (43 U.S.) 12 12A (83 U.S.) 13 14 15 15A (30 U.S.) 15B (54 US.) 15C (77 US.) 16 18 18A (44 US) 18C (56 U.S.) 19 19A (57 US.) 20 22 23 24 24A (65 U.S.) 25 27 28 29 32 32A (67 U.S.) 37 12 1,288 + 205 70 58 (-) 62 + - - - 9 (-) to (2) + + - - I All - or + through 15 S9 (Danish 9N). A tentative structure, possibly overly complex, has been proposed for S9 (14). Simpler repeating units were shown to occur in other members of the group Le., in 9A (type 33 U.S.) (6) and 9V (type 68) (49). Cross-reactions within the group are discussed in reference 55: D-G~cA was immunodominant in 9A, 9V, and 9L since reduction of -COOH to -CH*OH greatly diminished serological activity. In the present study, only small cross-reactions of 9N were found with anti-Pn's 3, 8, 12, 14, 18,19, 22, and 25. S9 and the capsular polysaccharides of all of these types except possibly S25, the structure of which is not known, contain 1 ,Clinked D-GIc in their repeating units. This may be the cause of these minor precipitations. S10 and SlOA (34 U.S.) The structure assigned for S10 (M. B. Perry, V. Daoust, and R. Lowe, Abstr. W.H.O. 3rd Int. Conf. on Immunity, Immunization, and Cerebrospinal Meningitis, VOL. 41,1983 CROSS-REACTIONS OF PNEUMOCOCCAL TYPES 1237 TABLE 1-Continued 16 18 19 20 22 23 25 27 28 29 12 14 15 1 2 3 4 4 5 6 7 78 8 9 9A 10 1OA 11 1IA 12 12A 13 14 15 1 SA 15B 15C 16 18 18A 18C 19 19A 20 22 23 24 24A 25 27 28 29 32 32A 37 72 1,010 + + f 65 57 15' 36 (-1 11' 19 (-1 u 45 f +f 47 +++ ++++ +f + + ++f (-1 (fl - - f f - f ++ ++ - (2) (2) (-) (++I ++ ++ 2,200 2,250 878 620 3c ++ f 16' 8' (2) (++f) f ++ + ++ 49c 12' f ++f (21 - +f +f ++ +f - ++f - + f-) - - ++ (- or f) +f + ++ - (++) (f) (-f f 12Y - +f ++ - (-) (++f) +++ 38 + 389 ++ +f f ++ +f - ++ ff) (-) - 57 20 - +f - - - ++ ++ - - ++ f - - - f - (-1 ++ L (-1 f ++f 0) ++f ' S17 was + only in anti-Pn7 and anti-Pn9; S17A was ++ in anti-Pnll and + in anti-Pn's 12. 15, and 20. They are omitted from the table. Data are expressed either qualitatively (see text) or quantitatively (as micrograms of antibody N per milliliter). From reference 1. From reference 18. From reference 28. ' From reference 30. From reference 21. From reference 16. ' From reference 49a. ' In anti-Pnl 1057C; - in anti-Pnl 884. ' t + + in rabbit anti-Pnll 59 (New York State). ' From reference 22. Mean of 43 and 50. dpS27 gave 176 pg of N per ml. dpS27 precipitated 45. Rectpitation by dpS27. 1238 HEIDELBERGER Marburg, 1980) is: INFECT. IMMUN. 9 2)-ribitolG-P-O + 6)-P-~-Galf-(l + 3)-a-~-Gal-(l + 4)-p-~-GalNAc-(l + 3)-a-D-Gal-(l OH P-D-Gal-(1f6 and that for SlOA (11, 50) is: ~2)-nbitol-(5-O-P-O-3)-o-Galf-(l + 3)-D-Gal-(1 + 4)-~-GalNAc-(l --., 3)-~-GaI-(l 9 OH (-OAc, partial) L In spite of structural similarities, cross-precip- itation of S10 in rabbit anti-PnlOA (R13 N.Y. State) was only about 80 of 1,965 pg of N per ml and 6% of the total in the reverse direction with H627. There were marked differences in the cross-reactivities of S10 and SlOA in other anti- sera (Table 1). Rabbit anti-10A was tested only with S1, S8, S14, S25, S27, and S29, giving 2, +, It, + 5, -, and + +, respectively. S11 and SllA (43 U.S.). The structure for S11 (M. B. Perry, personal communication) is: respectively, but SllA gave + + + + and 278 bg of N per ml (Table 1). As pointed out in refer- ence 38, SllA and S18 have the probable se- quence +3)-D-Gal-(1-4)-a-D-Gk-(l+6)-~-Glc- (1- and also glycerophosphate and 0-acetyl groups (16) in common, so that massive cross- reactivity is to be expected. The structure of S16 is not known. S12, S12A (83 U.S.), and S13. The only marked cross-reaction of S12 was in anti-Pnl4 635C and was predictable since both S12 (15,42) 6)-a-~-GlcNAc-(l + +a-D-Gal-(l + 3)-p-~-Gal-(l + 4)-P-D-Gk-(' -- Ribitol-7-0 -0Ac i OH and that for SllA (38) is: 21 46)-a-D-Gk-(l + 4)-cr-~-Gal-(l + 3)-P-~-Gal-(l + 4)-P-D-Gk-(l+ OAc 1 4T 9 0-P-0- CH*-CH( OH)-CHzOH OH -0Ac Removal of the acetyl groups from SllA abol- ished the precipitin reaction in a rabbit anti- PnllA (38). The smaller-than-expected precipi- tation of SllA in anti-Pnll 613 (16%) is perhaps due to blocking of antibodies by the as yet unlocated -0Ac groups in S11 and SllA. Both polysaccharides react heavily in anti-Pnl5 528C, probably because of the -0Ac sugar in their repeating units, since 0-deacetylation of S15, as with S11 A, abolishes its precipitation in homolo- gous rabbit serum. Precipitation of S4, S25, and S29 in a potent rabbit anti-11A was + 2, +, and +*, respectively. Precipitation by Sll in anti- Pn16 594C and antiPnl8 495 was only + and -, and S14 (4, 45) have nonreducing lateral end groups of D-Gal in their repeating units. S12A precipitated anti-Pnl2 very heavily and anti-Pn6 and anti-Pn9 only moderately. S12A differs from S12 only in the substitution of N-acetyl-D-gluco- samine for N-acetyl-D-galactosamine in the main chain. (K. Leontein, B. Lindberg, J. Lonngren, and D. J. Carlo, personal communication). S13 (57) gave only small cross-precipitations (Table 1). In anti-Pn6, reactivity might be due to a small amount of anti-ribitolphosphate in the antiserum; in anti-Pn14 and anti-Pn20, it might be due to 1,3-linked gal. S14. New methods have established a simpler VOL. 41, 1983 structure for S14 (45): CROSS-REACTIONS OF PNEUMOCOCCAL TYPES 1239 anti-Pnl2, and anti-Pn27, but the structure of ~~ ~ Anti-Pnl4 635C has been a useful reagent for detecting and predicting the presence of im- munodominant lateral D-G~ in many other mi- crobial and plant polysaccharides. Cross-precip- itation of S14 in anti-Pn22 is apparently caused by the 1.4-linked D-GIc residues in the repeating units of S14 and S22 (9). S15, S15A, SSB, and S15C. The structure of S1S (48) is: S16 is not known. S17 and S17A (78 U.S.). S17 and S17A are omitted from Table 1 because the strongest reaction was of S17A, + * in anti-Pnll 613. However, S17 (for structure, see M. B. Perry, D. R. Bundle, V. Daost, and D. J. Carlo, Can. J. Biochem. Cell Biol., in press) precipitated mas- sively in a rabbit antistreptococcal group F, type 4 serum (31), as did the group F type 4 polysac- +~)-P-D-GICNAC-(~ + 3)-P-~-Gal-(1 + 4)-@-D-Glc-(1 + 3)-a-D-Gd-(l - 2)-D-Gal-l+ -0Ac -0Ac 1 S15 and Sl5A precipitated anti-Pn4 609C to the same extent. Possibly the 1.3-linked D-Gal resi- dues of S1S fitted partially into combining sites of anti-Pn4 designed for the 1.3-linked gal- NAc residues of S4. The structures of SlSA. SlSB, and SlSC have not yet been determined, but SlSC has the same main chain as Sl5B but lacks the -0-acetyl group of the latter (P. E. Jansson, unpublished data). In this group, S15 reacted most heavily in anti-Pnl0 627C, S15A and S15C precipitated less, and SlSB did not react at all. The 1,3-linked residues of D-Gal in S1S and S29 are apparently the mediators of the reaction of S1S in anti-Pn29, whereas the 1,4- charide in rabbit anti-Pn17 (N.Y. State no. 29). S17A has several structural differences from that of S17 (P. E. Jansson et al., in press). S18, S18A (44 U.S.) (22), and S18C (56 U.S.). Two alternative probable structures have been proposed for S18 (16). 1,4-Linked D-glUCOSyl residues in S18 and S23 are undoubtedly respon- sible for the cross-reaction of S18 in anti-Pn23. As the structures of S16, S18A, and S18C are not known, reasons for the heavy cross-precipi- tation of the S18 group in anti-Pn16 cannot be given. S19 and S19A (57 U.S.). The structure of S19 (47) is: 4)-p-~-ManNAc-(l + 4)-a-D-Glc-(l --* 2)-a-~-Rham-(l + O-bt -f 01 bH 1'' ~ ~~ ~ ~~ linked residues of D-GIc are undoubtedly the reason for precipitation of S15 in anti-Pn20 616. S1SA and SlSB precipitate antiPnl5 to the same extent but differ sharply in anti-Pnl4, anti-Pnl8, and anti-Pn27. Because of the strong reaction of SlSB in anti-Pnl4, SlSB may be expected to have a lateral nonreducing end group of D-G~ in its repeating unit; SlSC gave only + . Although SlSA precipitated anti-Pn4 609 quite strongly (Table 11, the reciprocal reaction in a potent rabbit anti-PnlSA was only + 2. Precipitation of S4 in rabbit anti-PnlSB was only (2). S20, however, gave ++, but the reciprocal reaction was only (+I. S16. Small cross-reactions occur in anti-Pn7, In antiPn2 and anti-Pn7, S19 gave definite pre- cipitates, but S19A (40, 41) did not (Table 1). Precipitation of anti-Pn8 1008 by S19 was much stronger than that by S19A. probably because of inhibition by the side chains of S19A on the 1.4- linked Glc. S19A reacted weakly in anti-Pn4. anti-PnS, anti-Pn6, and anti-Pnl0, in which S19 was negative. Except for 1,3-linked D-Gal in S10 and S19A, the reasons for the other reactions are obscure. S20 (p. 1237) and 521. The strong precipitation of S20 in anti-Pnl8 appears to be due to the 1,6- linked D-~~UCOSY~ and 1.3-linked D-galactosyl residues in the repeating units of both S18 and S20. The latter residues and those of 1.4-linked 1240 HEIDELBERGER INFECT. IMMUN. D-GIc probably account for the weaker reaction in anti-Pn27. S21, - to f in all antisera tested, is omitted from Table 1. S22. As in other instances, the stronger reac- tions (Table 1) of S22 (9) could be due to the 1,4- linked D-linked D-Gk of the polysaccharides of the types involved. S23. The weak reaction in anti-Pn2 might be caused by 1,Clinked D-Glc even though (or because) it is substituted by a phosphoryl group on S23 (Perry et al., Abstr. W.H.O. 3rd Int. Conf. in Immunity, Immunization, and Cerebro- spinal Meningitis, 1980.) D-GIcPO~ might even fit loosely into the combining sites of anti-Pn2 meant for D-GkA. S24 and S24A (65 U.S.) (composition and struc- tures not known). S24 precipitated in anti-Pn4 and anti-Pn 8 (Table 11, but S24A did not; both reacted equally weakly in anti-Pn14 and nearly equally weakly in mule anti-Pn25. S24A was the only Pn S tested which precipitated strongly in anti-Pn28 510C, giving 189 of pg N per ml. It also gave 109 pg of N in anti-Pn23 912C; recipro- cal reactions were negligible in the rabbit anti- Pn24A available. S25. Precipitation was ++ in antiPn5 and + + + in anti-Pn8. S25 contains D-G~A (13), but the structure has not been determined. Cross- reactions are shown in Tables 1 and 2. S27. S27 (5) was the first Pn S in which pyruvic acid was found (21); because of the cross-reactivity of the extracellular polysaccha- ride of pyruvyl-containing Klebsiella type K32 in both anti-Pn27 and anti-Pn4, pyruvic acid was suspected in S4 as well and was identified (21). S28. The structure of S28 is not known. Cross- reactions are given in Table 1. S29. S29 has two residues of 1,6-linked D-Galf in its repeating unit (Sl), and these might func- tion as somewhat hindered nonreducing end groups. This would explain tfie reactivity of S29 in anti-Pn6 and anti-Pnl4. Both S29 and S14 have 1,3-linked D-galactopyranose in their re- peating units. This could reinforce the precipita- tion in anti-Pn14 and account for that in anti-Pn9 as well, since S9 contains 1,3-linked D-G~~NAc which would be partly equivalent immunologi- cally to similarly bound D-Gal. S32 and S32A (67 US.) (structures unknown). S32 contains Gal, Glc, Rham, and uronic acid (1) and does not precipitate as strongly in anti-Pn23 (a long-known cross-reaction [46, 581) as does S32A (Table 1). The latter, at least, may there- fore be assumed to have nonreducing lateral end groups of L-Rham in its repeating units (P. Allen, B. Prescott, and M. Heidelberger, unpublished data); phosphorylcholine is also a constituent. S32A reacts in anti-Pn's4, 14, 16, 18, and 25, whereas S32 does not, but gives 28 pg of N in anti-Pn19 631C. Both cross-reacted roughly equally in anti-Pn27. S37. The only noteworthy cross-reaction of the unusual glucan S37 (39) (not tested in all antisera) was in anti-Pn12 2%C, giving 34 pg of N per ml. S37 is said to have the nonreducing lateral end groups D-Glc-P-(1~2)-~-Glc-P-(l-, in contrast to the corresponding a-linked kojibiosyl of s12 (15). S72. All tests of S72 were - to f, except for heavy precipitation in anti-Pn16 594C (290 pg of N per ml). As structural formulas have not been proposed for S16 or S72 (12) and there are several sugars in common, an explanation is presently lacking. RECIPROCAL CROSS-REACTIONS Reciprocal, as well as unilateral, cross-reac- tions were often encountered in early efforts to obtain a reliable supply of strictly type-specific, diagnostic rabbit anti-Pn sera (46, 58). In this part of the present paper, an attempt is made to find out why some cross-precipitations between serological types are unilateral and others are reciprocal. One factor that might be involved is that different animals of a single species produce antisera mainly reactive toward different por- tions of the repeating units of bacterial polysac- charides (see, for example, reference 28). There- fore, if one could have a sufficiently large collection of potent antisera, possibly all cross- precipitations would be reciprocal. It is also likely that a target sugar in the main chain of the more highly branched or more highly O-acetylat- ed of two polysaccharides would be relatively inaccessible to cross-reactive antibodies. This would tend to make a reaction unilateral. These considerations apply also to bacterial agglutina- tion, especially as this is a precipitation reaction at the cellular surface (23). Data on reciprocal cross-precipitations among the Pn types are summarized in Table 2. Except for certain intergroup reactions, those are omit- ted which are recorded as less than + + in either direction. Types 2,5 and 2,6 are included, even though S5 and S6 failed to precipitate several anti-Pn2 sera, since types 5 and 6, originally 2A and 2B, were "weakly and incompletely aggluti- nated" (3) in anti-Pn2. Types 2,5; 2,6; 2,19; 2,20; and 2,22. Reciprocal cross-precipitation of types 2.5 is undoubtedly due to the lateral nonreducing end groups of D- GlcA in the repeating unit of S2 and 1,Zlinked D-GkA in SS. Like 1,2-linked D-Gal in S6 (52). 1 ,2-linked D-GkA in S5 may react serologically as a somewhat hindered lateral end group (18). The cross-reaction may also be reinforced by the 1,4-linked D-Glc in the main chains of both S2 and S5, a feature presumably also involved in the two-way 2,19 reaction. The rather small precipitation in the 2,6 pair is referable to the 1,3-linked ~-Rham in the repeating units of both. It is difficult to account for the 2,20 pair unless VOL. 41, 1983 CROSS-REACTIONS OF PNEUMOCOCCAL TYPES 1241 TABLE 2. Reciprocal cross-precipitation of pneumococcal types" Polysaccharide Antisem 2b 3 4 5 6 7 8 10 1OA 11 11A 14 15 15A 18 V555 63, V606 79, VI681 21 VI771 155. XIX631 8, XX616 44, XXII566 + + VIIIlooS 205 RXXIV + +, MXXV 16, RXXIV +++ XIV635 15, XVIII495 +*, xXIx641++ XIV 34, XVIII 91 XVIII 14, XIX 127, RXXIV +++ RXA Q 79, XIV 15, XV628 ++, RXVA, +++ XX 35, MXXV 49, XXVII 86, XXIX 57 XXIX 20 RXIA +++ XV (682)', XVI594 ++++, XVIII 278 XV 21, RXVA +++ RXVA ++++, RXVB + + + + , XVIII 57 RXVB ++++, XXVII 95 RXVIIIA 321". RXXXIIA +++' XXVII668 18' Polysaccharide Antiserum 5, 6, 19, 20, 22 8 24, 25. 27 24 14, Id, 29 14,18 18, 19, 24 10A, 14, 15, 15A, 20, 25. 27, 29 29 11A 15, 16, 18 15, 15A lSA, lSB, 18 15B, 27 18A, 32A' I1464 +, 11513 0. 20, 56, ++f 111792 113 IV60950, +++, 14Od v555 ++ VI 681 17.10, ++ VI1937 35, 7 VI11 47, 105, 68 X627 51, 79, 62. 28 163,43, ll', 61 RXA ++ XI613 125 RXIA ++, +++f, XIV 47, ++f XV 301, 293, + + RXVA ?., +++ ++ XVIII 357h, + + + " Qualitative data are expressed on a scale of - to + + + + ; quantitative values are expressed as micrograms of antibody nitrogen per milliliter of antiserum. M, Mule; R, rabbit (antisera without these letters were raised in horses). Sera containing appreciable antibody to Pn groupspecilic C-polysaccharide were absorbed with C before use. In this table serum types are given in Roman numerals to avoid confusion with serum numbers and amounts of precipitated nitrogen. See text for reason why group pairs 2,6 and 6.2 are included. dpS4 gave 25 p,g of N. dpS27 gave 176 pg of N. o dpS27 gave 45 pg of N. Single determination only. E Test was not made. ' From reference 22. ' P. Z. Allen, B. Rescott, and M. Heidelberger, unpublished data. the glucose PO4 of S20 could fit loosely into combining sites of anti-Pn2 designed for GlcA. As for types 2 and 22, the polysaccharides of both contain sugars bound in the same ways, yet the cross-precipitations are small. Types 3,s. Since one-half of S8 consists of the repeating unit of S3 (35). heavy mutual cross- precipitation is to be expected. This long-known reciprocal cross-reaction has been extensively discussed (24, 25). as have also the inhibitory properties of oligosaccharides formed on partial hydrolysis of each S (8). The strong reciprocal cross-precipitation between types 8 and 19 may now be accounted for by the presence of a- and p-1 ,elinked D-Glc (35) in S8 and by a-l,4-linked D-Glc and p-1 &linked D-ManNAc (partially equivalent immunologically to D-Gk) in S19. Types 424; 4,s; and 4,27. S24 and S25, of unknown structure, appear not to have been tested for pyruvic acid. This acid, however, has little to do with the reciprocal cross-reactivity of 4,27, as precipitation is greater with the dp polysaccharides (Table 1). A possible reason for the reaction is the partial serological equivalence of 1,4-~-ManNAc in S4 and 1,4-linked D-GIC in S27. Types 5,24. There is no explanation for the reciprocal reactions of 5.24 at present. Types 6,14; 6,18; and 6,29. The rather weak reciprocal cross-reactions of 6.14 are accounted for by the evidence (28) that the 1 .2-linked D-Gal of S6 shows the serological behavior of a partial- ly hindered lateral end group of D-Gal such as exists unobstructed in S14 (see above). The still weaker 6,18 reactions reflect the presence of 1,3-linked L-Rham in both polysaccharides. In the 6,29 pair, the two 1,Glinked D-Galf residues of S29 might fit loosely into antibody sites designed for the gal of S6. Types 7,14 and 7,18. Reciprocal cross-reactiv- 1242 HEIDELBERGER INFECT. IWMUN. ity of types 7 and 14 is due to the nonreducing lateral end groups of D-Gal in the repeating units of S7 and S14, with probable reinforcement by the p-1,4-linked D-glucosyl .residues present in both, The 7,18 pair also has 1,Clinked D-GIc and, in addition, 1,3-linked L-Rham. Types 8,18; 8,19; 8,24. Each repeating unit of S8 and S18 has two 1.4-linked D-Gk residues; these appear to suffice for a limited degree of reciprocal cross-reactivity. In the much heavier cross-reactions of the 8,19 pair, it is possible that the phosphoryl-(l~4)-p-D-ManNAc residues of S19 might fit to some extent into anti-Pn8 com- bining sites intended for the cellobiouronic acid groups of S8, aided and abetted by the 1,4-~- glucosyl residues of S19. For the opposite direc- tion, the cellobiouronic acid residues of S8 might fit into the anti-Pnl9 combining sites for PO4- ManNAc residues. The composition and struc- ture of S24 are not known. 9A, 9L, 9N (9 U.S.), and 9V. Cross-precipita- tion among the members of this group has been studied quantitatively (55). Reciprocal crossing between 9A and 9N is massive, as would be expected from their similar main chains. Mutual cross-reactivity of the 9A,9L pair is almost complete, but the structure of 9L is not known. 9V has 0-acetyl groups on GlcA and Glc; these probably give it an individual specificity since 9N does not precipitate the rabbit anti-9V used (55). Types lO,lOA, 10,14; 10,15; 10,15A; 10,W 10,27; 10A,25; and lOA,29. S10, SIOA, and S15 (see above for structures) have two 1,3-linked D- galactosyl residues, and 514 and 527 have one in their repeating units; these sugars appear to be principally responsible for reciprocal reactions 10,lOA; 10,14; 10,15; and 10,27. In the 10,20 pair the nonreducing lateral end groups of p-D- Galf are probably the main cause, whereas SlOA and S29 have the same furanose in the chain of sugars and 1,3-linked gal in addition. The structures of SlSA and 525 are not yet available. Types 11,llA and llA,15. The structures of S11, SllA, and S15 are given above. As all three polysaccharides contain the sequence 1,3-p-~- Gal-l,4-~-Glc in their repeating units, this ac- counts for the reciprocal precipitation in antisera to both pairs. Why the reaction of Sll and SllA in anti-Pnl5 is so massive is not clear. Types 14,15 and 14,lSA. The reason for these reciprocal reactions appears to be the same Gal- Glc sequence as described above, since this also occurs in S14, Types 15,15A; 15,15B; 15A,15B; 15,18; and 15A,27. 518 also has the 1,3-~-Gd-1,4-~-Glc sequence, but the anomeric forms of the sugars were not determined (16). A rabbit anti-Pnl5A gave +++ with S27; S15A precipitated anti- Pn2? 668C similarly (95 pg of N per mi). Types 18,18A and 18,32A. S18A contains all components of S18 except the 0-acetyl group and has GlcNAc in addition (22), but its struc- ture is not known, nor is that of S32A. Types 19,19A. This reciprocal cross-precipita- tion was demonstrated in reference 40. S19A differs from S19 in that it has two side chains which are lacking in S19 (40). In summary, in most of the reciprocal cross- reactions between Pn types of which the struc- tures of the capsular polysaccharides are known, at least two sugars or their immunologi- cally partially equivalent amino analogs are pre- sent in common. This is not, however, a rigid requirement. In four of the pairs listed, a single immunodominant sugar, or a, presumably equiv- alent combination, suffices to make the cross- reactions reciprocal. This seems reasonabfe. In only two instances, 2,6 and 6,8, probably due to unusually favorable steric factors, does recipro- cal cross-reactivity appear to be due to a single non-immunodominant sugar in common. Ac- cordingly, reciprocal cross-reactivity, like the nonreciprocal reactions, has an orderly structur- al basis. Where the chemical structures of the capsular polysaccharides are known, it should now be possible to predict the likelihood, but not the certainty, of reciprocal cross-reactivity. The use of more sensitive methods than those in the current exploratory studies might make such prediction more certain. For example, as early as 1939 use of capsular sweliing and agglu- tination (96) in rabbit anti-Pn sera revealed the following reciprocally cross-reactive pairs: 2,20; 33; 8J9; 10,20; 10.29; 13.17; 13,29; 18,28; 20,31; 22,31; and 23,32. In reference 58, the pairs 8J9; 10,20; 10,29; 16,28; 20,31; and 23,32 are recorded. No information on the complete structure of any Pn polysaccha~de was available at that time. ACKNOWLEDGMENT Data in the tables from papers cited by reference number arc from work carried out under grants from the National Science Foundation which terminated in 1976. Since 1976, polysaccharides or antisera or both have been received with gratitude from the New York City Department of Health Laboratories, kindness of P. May and Y. C. Faur; the New York State Department of Health Research Labora- tories, kindness of K. Amiraian; and from R. Austrian, J. Henrichsen, and G. SchSman. I am also indebted to Charles Serhan, Corcy Mallett, and Andrew Zakarian for technical assistance and to the secretariat of the Department of Patholo- gy of the New York University Medical Center for typing the manuscript. LITERATURE CITED 1. Men, P. Z., M. Hdddkrger, I. A. Rappaport, rad C. M. Wud. 1978. Homologous and cross-reactive precipitates in antipneumococcal sera raid in mules. Immunology 35;105-113. 2. Men, P. Z., and B. Prescott. 1978. Immunochemical studies on a Mycoplasma pneumoniuc polysaccharide fraction: cross-reactions with type 33 and 32 antipneu~ VOL. 41, 1983 CROSS-REACTIONS OF PNEUMOCOCCAL TYPES 1243 coccal rabbit sera. Infect. Immun. 20421-429. 3. Avery, 0. T. 1915. A further study on the biologic classification of pneumococci. J. Exp. Med. 22804-819. 4. Barker, S. A., M. Heidclberger, M. Stncey, and D. J. Tipper. 1958. Structure of the immunologically specific polysaccharide of Pneumococcus type XIV. I. Chem. Soc. 3468-3474. 5. Bennett, L. G.. and C. T. Bishop. 1977. Structure of the type XXVII Streptococcus pneumoniae (pneumococcal) capsular polysaccharide. Can. J. Chem. 55:8-16. 6. Bennett, L. G., and C. T. Bishop. 1980. Structure of type XXXIll Streptococcus pneumoniae (pneumococcal) cap- sular polysaccharide. Can. J. Chem. 582724-2727. 7. Brown, R., and L. K. Robinson. 1943. Chemical and immunological studies of the pneumococcus. VI. The soluble specific substances of new types and sub-types. 1. Immunol. 47:7-13. 8. Campbell, J. H., and A. M. Pappenheimer, Jr. 1966. Inhibition of precipitin reactions with oligosaccharides isolated from hydrolysates of S3 and S8. Immunochemis- try 3213-222. 9. Chntterjn, B. P., S. hrhysstha, and C. V. N. Rao. 1976. Structure of Pneumococcus type XXII capsular polysac- charide. Indian J. Chem. 14B:914-918. 10. Chaudhari. A. S., C. T. Bishop, and R. S. Fielder. 1972. Structural studies on the specific type VI1 pneumococcal polysaccharide. Carbohydr. Res. 25161-172. 11. Chittenden, G. J. V., W. K. Roberts, J. G. Buchanan, and J. Baddiky. 1968. The specific substance from Pneumo- coccus type 34 (41). The phosphdiester linkages. Bio- chem. J. 109597-602. 12. Daoust, V., D. J. Carlo, J. Y. Zeltner, and M. B. Perry. 1981. Specific capsular polysaccharide of type 45 Sfrepto- coccus pneumoniae (American type 72). Infect. Immun. 321028-1033. 13. Dss, A., M. Heidelberger, and R. Brown. 1976. Identifica- tion of D-galacturonic acid in the specific capsular poly- saccharide of pneumococcal type XXV. Carbohydr. Res. 48:304-305. 14. DPJ, A., J. D. Higginbothnm, and M. Heidelberger. 1972. Oxidation of the capsular polysaccharide of pneumococ- cal type IX by periodate. Biochem. J. 126:233-236. 15. Duke, J. L., I. J. Gddstein, and J. A. Cifonelli. 1974. Structural studies of capsular polysaccharide of type XI1 Diplococcus pneumoniae. Carbohydr. Res. 37:81-88. 16. Estrada-Parra, S., and M. Heidelberger. 1%3. The specif- ic polysaccharide of type XVIII pneumococcus. 111. Bio- chemistry 2:1288-1294. 17. Heidelberger, M. 1960. The immunological specificity of type I1 pneumococcus and its separation into partial specificities. 11. J. Exp. Med. 111:33-43. 18. Heidelberger, M. 1%2. Immunochemistry of pneumococ- cal types 11, V and VI. IV. Cross-reactions of type V antipneumococcal sera and their bearing on the relation between types 11 and V. Arch. Biochem. Biophys. l(Suppl.):169-173. 19. Heidelberger, M., A. C. Ahenberg, and W. 2. Hassid. 1954. Glycogen, an immunologically specific polysaccha- ride. J. Exp. Med. 99:343-353. 20. Heidelberger, M., 2. Dkhe, W. B. Neely, and M. Wol- from. 1955. Immunochemistry and the structure of lung galactan. J. Am. Chem. Soc. 77:3511-3514. 21. Hcidelberger, M., W. F. Dudman, and W. Nimmich. 1970. Immunochemical relationships of certain capsular poly- saccharides of Klebsiella. pneumococci. and Rhizobia. I. Immunol. 104:1321-1328. 22. Heidelberger, M., S. Estrada-Pnrra, and R. Brown. 1964. The specific polysaccharide of type 18A pneumococcus. Biochemistry 3:1548-1550. 23. Heidelberger, M., and E. A. Kabat. 1937. Chemical stud- ies on bacterial agglutination. Ill. Reaction mechanism and quantitative theory. J. Exp. Med. 65885-902. 24. Heidelberger, M., E. A. Kabat, and M. M. Mayer. 1942. Further study of the cross-reaction between the specific polysaccharides of types 111 and VI11 pneumococci in horse antisera. J. Exp. Med. 7535-47. 25. Heidelberger, M., E. A. Kabat, and D. L. Shrivastava. 1937. Quantitative study of the cross reaction of types 111 and VI11 pneumococci in horse and rabbit antisera. 1. Exp. Med. 654874%. 26. Heidelberger, M., and F. E. Kendd. 1935. The precipitin reaction between type 111 pneumococcus polysaccharide and homologous antibody. 111. A quantitative study and a theory of the reaction mechanism. J. Exp. Med. 61563- 591. 27. Heidelberger, M., and F. E. Kendnll. 1937. A quantitative theory of the precipitin reaction. IV. The reaction of pneumococcus specific polysaccharides with homologous rabbit antisera. J. Exp. Med. 69647660. 28. Heidelberger, M., and P. A. Rebers. 1960. Immunochemis- try of the pneumococcal types 11. V. and VI. 1. The relation of type VI to type I1 and other correlations between chemical constitution and precipitation in anti- sera to type VI. J. Bacteriol. 80:145-153. 29. Heidelberger, M., N. Roy, and C. P. J. Ghudemnns. 1%9. Inhibition by aldobiouronates in the precipitation of pneu- mococcal type I1 and 111 systems. Biochemistry 84822- 4824. 30. Heidelberger, M., and J. M. Tyler. 1964. Cross-reactions of pneumococcal types. Quantitative studies with the capsular polysaccharides. J. Exp. Med. 120:711-719. 31. Heidelberger, M., J. M. N. Willers, and M. F. Michcl. 1%9. Immunochemical relationships of certain slrepto- coccal group and type polysaccharides to pneumococcal capsular antigens. J. Immunol. 102:1119-1127. 32. Higginbotham, J. D., and M. Heidelberger. 1972. The specific capsular polysaccharide of pneumococcus type IV. Carbohydr. Res. U:165-173. 33. Higginbotham, J. D., M. Heidelberger, and E. C. Gotsch- lich. 1970. Degradation of a pneumococcal type-specific polysaccharide with exposure of group-specificity. Proc. Natl. Acad. Sci. U.S.A. 67:139-142. 34. Jansson, P. E., B. Lindberg, and U. Lindqvist. 1981. Structural studies of the capsular polysaccharide from Streptococcus pneumoniae type 4. Carbohydr. Res. 95:73-80. 35. Jones, J. K. N., and M. B. Perry. 1957. Structure of the type VI11 pneumococcus specific polysaccharide. J. Am. Chem. Soc. 792787-2793. 36. Kenne, L., B. Lindberg, and J. K. Madden. 1979. Structur- al studies of the capsular antigen from Srreptococcus pneumoniae type 26. Carbohydr. Res. 73:175-183. 37. Kenne, L., B. Lindberg, and S. Svcasron. 1975. Structure of the capsular polysaccharide of pneumococcus type 11. Carbohydr. Res. 40:69-75. 38. Kennedy, D. A., J. G. Buchanan, and J. Baddiley. 1%9. The type-specific substance from Pneumococcus type 11A (43). Biochem. J. 115:3745. 39. Knecht, J. C., G. Schiffman, and R. Austrian. 1970. Some biological properties of pneumococcus type 37 and the chemistry of its capsular polysaccharide. J. Exp. Med. 132475487, 40. Lee, C.J., and B. A. Frerer. 1980. Structures of the cross- reactive types 19 (19F) and 57 (19A) pneumococcal capsu- lar polysaccharides. J. Biol. Chem. 255:6847-6853. 41. Lee, C.-J., and K. Koizumi. 1981. Immunochemical rela- tions betwen pneumococcal group 19 and Klebsiella cap- sular polysaccharides. J. Immunol. 127:1619-1623. 42. Leontein, K., B. Lindberg, and J. Lanngren. 1981. Struc- tural studies of the capsular polysaccharide from Strepto- coccus pneumoniae type 12F. Can. J. Chem. 59:2081- 2085. 43. Lew, J. Y.. and M. Heidelberger. 1976. Linkage of pyruvyl groups in the specific capsular polysaccharide of pneumo- coccus type IV. Carbohydr. Res. 52255-258. 44. Lindberg, B., B. Lindqvist, J. Liinngrcn, and D. A. Powell. 1980. Structural studies of the capsular polysaccharide from Streptococcus pneumoniae type 1. Carbohydr. Res. 45. Lindberg, B., J. Liinngrcn, and D. A. Powell. 1977. 78:111-117. 1244 HEIDELBERGER INFECT. IMMUN. Structural studies on the specific type 14 pneumococcal polysaccharide. Carbohydr. Res. 58177-186. 46. Lyd, N. W., and H. R. Odell. 1939. Production and standardization of diagnostic antipneumococcus sera. Am. J. Hyg. 29103-106. 47. Ohno, N., T. YadomPe, and T. Miynzpki. 1980. Structure of the type-specific polysaccharide of Pneumococcus type XIX. Carbohydr. Res. W297-304. 48. Perry, M. B., D. R. Bundle, V. Dnoust, and D. J. Carlo. 1982. The specific capsular polysaccharide of Streprococ- cus pneumoniae type 15F. Mol. Immunol. 19:235-246. 49. Perry, M. B., V. Dnoust, and D. J. Carlo. 1981. Specific capsular polysaccharide of Streptococcus pneumoniae type 9V. Can. J. Biochem. 59524-533. 49a.R.o. C. V. N., and M. Heidelberger. 1966. Capsular polysaccharide of pneumococcus type IX. J. Exp. Med. 123:913-920. 50. Rm, E. V., J. G. Bucbnnn, and J. Baddiley. 1966. The type-specific substance from Pneumococcus type 10A (34). Biochem. J. 100:811-814. 51. Rao, E. V., M. J. Watson, J. G. Buchannn, and J. Baddiley. 1%9. The type-specific substance from Pneu- mococcus type 29. Biochem. J. 111547-556. 52. Rebers, P. A., and M. Heidelberger. 1961. The specific polysaccharide of type VI pneumococcus. 11. The repeat- ing unit. J. Am. Chem. Soc. 83:3056-3059. 52a.Richards, 1. C., M. B. Perry, and D. J. Carlo. 1983. The specific capsular polysaccharide of Streptococcus pneu- moniae type 20. Can. J. Biochem. Cell Biol. 61:178-190. 53. Roy, N., and C. P. J. Ghudemsns. 1968. The specific substance from Diplococcus pneumoniae type 34 (US. type 41). Location of the 0-acetyl groups. Carbohydr. Res. 8214-218. 54. Shabuova, 2. A., J. G. Buchannn, and J. Bsddiley. 1962. Composition of pneumococcus type-specific substances containing phosphorus. Biochim. Biophys. Acta 57:146- 148. 55. SN, S., C.4. Lee, D. Carlo, and J. Henrichsen. 1981. Immunochemical characterization of the cross-reactivity of pneumococcal group 9 capsular polysaccharide types 9N, 9A, 9L. and 9V. Infect. Immun. 31:371-379. 56. Tykr, J. M., and M. Heidelberger. 1968. The specific capsular polysaccharide of type VI1 Pneumococcus. Bio- chemistry 7:1384-1392. 57. Watson, M. J., J. M. Tykr, J. G. Buchman, and J. Baddiley. 1972. The type-specific substance from Pneu- mococcus type 13. Biochem. J. 1M45-54. 58. Welch, H., E. K. Bormnnd, and F. L. Mickle. 1939. Preparation of diagnostic antipneumococcus sera. Am. 1. Public Health 2935-42.