IMMUNOCHEMICAL RELATIONSHIPS BETWEEN BACTERIA BELONGING TO TWO SEPARATE FAMILIES: PNEUMOCOCCI AND KLEBSIELLA MICHAEL HEIDELBERGER* and WOLFGANG NIMMICH Department of Pathology, New York University Medical Center, NY 10016. U.S.A. and thc Institute of Medical Microbiology and Epidemiology, University of Rostock. Rostock, GDR (Recriaed 18 February 1975) Abstract-The cross-reactions of more than 60 capsular (K) type-specific polysaccharides of Klr~hsiellu in antipneumococcal sera of 26 specific types and in several anti-Salr~~o/iellu sera are described and discussed, many for the first time. Quantitative data on the extent of cross-reactivity are given in many instances. From these, and from a study of residual fractions of antibody in the supernatants, it has been possible to confirm the presence of non-reducing end-groups of wglucuronic acid in KZ. K8 and KZO. Such groups were also correctly predicted for K9 and K59 and will most probably be found upon chemical examination of K15, K23, K27, K30, K33, K45. K51 and K55. Non-reducine - end-groups of L-rhamnose were correctly predicted for K47 and K56 and will likewise be found in K17 and K19. Other correlations of chemical structure and immunological specificity are pointed out, as well as limitations of the methods used. INTRODUCTION Knowledge of the molecular basis for the frequent cross-reactions which occur between microorganisms belonging to widely disparate families is not only of ,-practical interest, but also has a broader theoretical significance, for it is capable of revealing strict rela- tions between the chemical constitution of certain microbial antigens and their immunological speci- ficities. Several such aspects of cross-reactivities have already been reviewed [14]. The bacterial polysaccharides provide a rich source of antigenic material, and since they are often the principal antigenic determinants of the parent micro- organisms, the corresponding antisera are also fre- quently available. Hence the cross-reaction of a poly- saccharide of known chemical structure with anti- bodies to a polysaccharide of uncertain or unknown structure may yield information as to one or more sugars contained in the unknown and even as to the positions at which the sugars are linked. Conversely, cross-reactivity of a polysaccharide of unknown struc- ture with antibodies to a polysaccharide of known composition and linkage may be equally informative. If the cross-reactions are massive, qualitative tests could suffice, but if only a small proportion of the antibody is involved, quantitative estimations are necessary, as differences in titer might easily be mini- mal. Quantitative analyses have the further advantage of yielding information as to which fractions of the usual complex mixture of antibodies are cross-reac- tive, and which fractions remain in the supernatant fluid. The data so obtained must, however, be inter- preted with caution, for it has been shown that, in *Aided by grants from the National Science Founda- tion, USA. certain instances of cross-reactivity, part of the preci- pitable antibody behaves as if it were multivalent with respect to the cross-reactive antigen and part as if it were univalent. being coprecipitated only in the presence of the multivalent portion [4h]. Precipi- tation and reactivity are therefore not synonymous. An additional and occasionally observed difficulty is that a second cross-reactive antigen may serve to coprecipitate soluble antigen-antibody complexes remaining from the first precipitation. Instances illus- trative of these difficulties will be given. The present review encompasses the cross-reactivi- ties of some 60 of the 80 or more capsular, type- specific polysaccharides of KIrhsiellu in a series of antipneumococcal sera chosen for their availability. Most of these antisera were raised in horses before it was known that the tendency of this animal to pro- duce macromolecular anticarbohydrate [5] diminished its usefulness for supplying quick and effi- cient type-specific antibodies for the cure of pneumo- coccal pneumonia [SI. This very property, however, appears important for the tendency of equine anticar- bohydrate to be more cross-reactive than that of the rabbit, which is usually IgG. Accordingly, most of the antisera used in this study were those of equine origin which were not discarded when the purpose for which they were produced became obsolete. Some data are also given for antisera raised in rabbits and these are designated by the letter R. Limited data are also included on cross-reactions in a few anti- Salriioriella sera. MATERIALS AND METHODS Klebsirllrr K polysaccharides were mainly isolated by Nim- mich [I I. I?]. Samples of K 1 and K3 were also obtained from Prof. S. D. Henriksen and Dr. Jorunn Eriksen; K6 67 68 MICHAEL HEIDELBERGER and WOLFGANG NIMMICH from Prof. R. W. Wheat; KI 1 from Dr. S. Stirm; and K5, deacetylated K5, K18, 20, 21, 24 and 62 from Prof. G. G. S. Dutton. Antipneum~~cal (anti-Pn) sera were supplied by the Bureaus of Laboratories of the New York City and New York State Departments of Health. Antipneumococcal type-specific rabbit globulins were given by E. R. Squibb & Sons; anti-Salmorrella sera by Dr. Anne-Marie Staub. Anti-Pn sera containing more than a few pg/ml of antibody to the pneumococcal groupspecific C-polysaccharide were precipitated with this substance before use and carry the letter C after the number of the animal. Qualitative testsare recorded on a scale of - to + + -t +. on which + + is roughly equivalent to 8-2Opg of anti- body nitrogen per ml (by actual comparison) and + + + + to 100 and up. The tests were usually carried out with 0.5-1 mi of antiserum, depending upon its content of anti- body, in tubes immersed in icewater. Polysaccharide (0.05 mg) was first added, followed by 0.05 mg more if pre- cipitation became evident at the interface. The contents of the tubes were thoroughly mixed, and the rack of tubes, usually 4 rows of 10 tubes in each, was placed in a cold room for 6-8 days. After immersion in ice-water. readings were taken. Quantitative analyses at several levels of antigen were made of the heaviest reactions (Table 4) and an additional 01 or 0.15 mgiml was added to the remain- ing tubes, since a large excess of antigen was frequently required for maximal cross-precipitation. After 68 days second readings were taken, after which the sera were used for a second series of tests. In these, the second polysac- charide was added to the remaining tubes of two rows in order to lessen the possibility of negative tests due to inhibition by the excess of one or the other of the sub- stances first tested. Results of such second (or occasionally, third) tests are recorded in parentheses. Strong cross-reac- tions usually appeared rapidly, whether or not the sera had been negative or weakly reactive in earlier tests. A reading af-+.+ 2 would be slightly greater than + + if read in the same series on the same day, but might be considered + + at another time, a failing common to all qualitative tests. Quantitative analyses were carried out as in earlier papers [7-91. Antigen and antibody were mixed at 0°C and the tubes were capped and immersed at least to the level of their contents in a 0°C bath (Forma Scientific, Inc.). After 4-14 days, depending upon the rapidity of Roc- culation, if any, the tubes were centrifuged at OT. Superna- tants were saved for eventual analysis, and the inverted tubes were allowed to drain in a cold box [8] before the rims were wiped and the precipitates dislodged and washed in the tubes at 0°C with ice-cold saline. As cross-reactions may be sensitive even to small differences in temperature (for example, [IO]), the contents of the tubes should be kept as close to 0" as possible. All values are reported as micrograms of antibody nitrogen precipitated at 0°C per ml of antiserum, even though in some instances of massive cross-reactivity as little as 0.05 ml was actually used. Analyses were run in duplicate. with a serum blank, at two or more levels of polysaccharide. Only maximal values are given in the tables. RESULTS AND DISCUSSION Several anti-Pn sera which were known to cross- react heavily in other instances (e.g. [l]) failed to give tests stronger than + + with any of the KZ~bsjel~a K polysaccharides tested and are omitted from Table 4. The sera were: anti-Pn XI613 (- to f); anti-Pn XI1296C [+ +, (+ +)with K10, K581; anti-Pn RXIII Squibb, (K44+ +); anti-Pn XIV635C [K7, K12+, K41, K5O+ +, K57(+)]; anti-Pn XVI594C (K22+, K36++ = 16pgN); and anti-Pn RXVII (- to +). Anti-Pn XXVIII5lOC, which precipitated 35 pg N with K52, is also omitted as it reacted only + with K2, K26, K54 and + f with K8. Klebsiella K type polysaccharides 1-59, 62-64, 66, 72, 75 and 81 were tested. Their composition is given in Table 1, which is adapted from [ll, 121. Structures, insofar as presently known, are shown in Table 2, with corresponding data for pneumococcal capsutar Table 1. Constituents a of capsular polysaccharides (K) of Klebsiella tested for Cross-reactivity GlcA, gal, glc GlcA, gal, man GlcA, gal, rham GlcA, glc, man 2, 24 GlcA, glc, rham 17, 44 GIcA, glc, fuc I, 54PyA GlcA, gal, glc, man KSPyA, 15, 25, 27PyA, 51 20, ZIPyA, 29PyA, 42PyA. 43, 66 9, 47, 52, 81 4, SPyA, lPyA, 10. lIPyA, I3PyA, 26PyA. 28, 30PyA, 31PyA, 33PyA, 3SPyA. 39,46PyA, SO, 59, 62 GlcA, gal, glc, rham 12PyA, 18, 19, 23, 36PyA, 41. 4SPyA. 5SPyA GlcA, gal, glc, fuc 16, 58PyA 40, 53 GlcA, glc, man, rham 64PyA GlcA, glc, man, fuc 6PyA GlcA, gal, glc, man, rham 14PyA GalA, gal, man 3PyA, 49. 57PyA, 75 GalA, glc, rham 34, 48 GalA, gal, rham, fuc 63 PYA, gal, rham 32 PYA, gal, glc, rham 3-~oxy-~-g~yce~o-~ntulo~nic acid, Unknown keto acid, gal, glc GlcA, gal, man, rham 56, 12 gal, glc 38 (13) 22. 37. K37 has 4-04 1 -carboxyethyl)-~-glcA Abbreviations used in this and subsequent tables: gal, galactose; glc, glucose; man, mannose; rham. rhamnose; fuc. fucose; ara. arabinose; galA, galacturonic acid; glcA. glucuronic acid; PYA. pyruvic acid; DPA, 3-deoxy-L-glycero- pentufosonic acid; galN. galNAc, galactosamine and its N-acetyl derivative; gIcN, gfcNAc, glucosamine and -NHAc; manN, manNAc; fucN, fucNAc. In Tables 2 and 3, sugars with a capital first letter. as Glc. indicate non-reducing end-groups. Immunochemical Relationships of Pneumococci and Klebsiella 69 Table 2. Known partial and complete structures of Klebsiella K polysaccharides tested / K Type / 1 2 3 5 7 8 9 11 18 20 21 24 28 37 38 [D-Gk-(I--$-, most sugar linkages u-; D-glcA, L-fuc dn I 3)-i,-glc-/$-( I -+ 4)-o-man-B-( I -+ 4)-~-glc-a-( I J' D-GIcA-Z- D-Gal. D-man; galA -+ D-man t' OAc 3t /I 4)-D-gICA-P-( I -+ 4)-D-glC-B-( 1 + 3)-D-man-B-( 1 I I CH,CCOOH f f. 3)-~-glcA-P-( 1 -+ 2)-~-man-( 1 - ~)-D-$!c-( I t 3)-D-gaI-B-( 1 T -+ I I1 D-Gal 3)-~-gal-a-( I f 3)-D-gk-p-( I -+ 3)-D-gkA-p-( I -+ 3)-D-gal-Y-( I IT4 t D-Gal-%-=PYA -a-~-gal-(3 t l)-/?-~-GlcA --OAc 1 -+ 3)-~-man-a-(l-+ 2)-o-man-z-( 1 - 3)-0-gal-p-( 1 I 4 COOH I\ I 1 /\/ -Z-D-Gal c \ /\ 70 MICHAEL HEIDELBERGER and WOLFGANG NlMMlCH 3)-i,-gal-/l-( I -+ 4)-~-rham-r-( I L-R ham-O-( I -+ 4)-1>-gkA-&- 3)-u-gal-( I -+ ?)-L-rham-( I -+ 4)-u-glcA-( f -+ 3)-D-ga!-( 1 -4 4)-~-rham-( I f I I- n-Gal- 47 6)-~-gl:-/l-( I -+ 4)-D-gICA-E-( I ---t ~)-L-~uc-( I 52 54(A3Sl) i J)-~-glc-P-( 1 -+ 3)-~-gat-&-( I -+ v L-R ham+ 56 CH,. C COOH f 6/ 3)-D-gal-P-( I-d)-~-galA-a-( I -+ 2)-~-man-r-( 1 T4 i r>-Man-l-r- 3)-D-gk-P-( 1 -+ 3)-~-gal$'-( 1 -+ 2)-~-man-z-( 1 -+ 3)-o-man-x-( I b Ac b AC D-ClcA-)I- [' 57 59 64 ,... glcA --+ man, 0-glC, D-man, L-rham 3)-~-glc-P-( 1 --+ 3)-~-rham-x-( 1 4 Z)-~-rham-r-(l-+ 3)-~-rham-r-( 1 3/ '4 v CH, .C .COOH Z)-~-rharn-z-( 1 --+3f-t-rham-z-( 1 -+ 4)-o-glcA-j-( 1 -+ 2)-~-rham-r-( 1 --+ 3)-~-rham- [' 72 (35) (36) (80) f 2-( 1 + 3)-D-gal-P-( 1 - Partial substitution. Sugars are pyranoses unless otherwise indicated. Repeating unit is double that given, with lateral substituents only on alternate half-units. XA = 4-0-( I-carboxyethy1)-0-glucuronic acid. ' Not uniquely determined. polysaccharides, the principal determinants of pneu- mococcal type specificity, in Table 3. Qualitative and quantitative data are summarized in Table 4 and its footnotes. ~ru.s.~-~eu(,~iof7s in mti-Pn I Although seven of the K polysaccharides contain galacturonic acid, only two, K49 and K63, cross-react heavily (Table 4). In these, then, the sugar acid belongs to the D-SeTieS,* and is linked, at least in part, as lateral non-reducing end-groups and/or attached at the 1 and 3 positions, as in Pn SI [37]. Footnote a (Table 4) shows that both substances precipitate * For an examplc of differences in the specificity of 1)- and L-forms of the same sugar cf. [79]. a portion of the same fraction of anti-Pn I; also that ketha gum, in which ~-grilA was first identified as a result of the gum's massive cross-reactivity in anti-Pn I [57]- removes all of the fraction of anti-Pn I precipitable by K49 and K63, K57, which contains ~-galA linked 1.3.4-1321. did not precipitate anti-Pn I. Two strong anti-Pn I rabbit sera were tested with K49 and K63; one gave no precipitate with either; the other, RI937, with about 2000pg anti-Pn I nitrogen per ml, gave 267 pgN with K49 and 2 with K63. Cross-reaeiions iri anti-Pn 11 The ty~-sp~ificity of Pn TI is separable into several partial specificities: an immunodominant one Immunochemical Relationships of Pneumococci and Klrhsirlla 71 Table 3. Constitution of pneumococcal capsular polysaccharldes of types used c Pn type I I1 I11 IV V VI VI1 VI11 IX X galA - galN. galA -+ glcN, gilA( I -+ 3)-glcN-( I -+ 3)-gdlA i,-GalA. probably partly 0-acetylated. glc. 3)-L-rham-x-( I - 3)-L-rham-r-( 1 -+ 3)-L-rham-P-( I - 4)-~-glc-z-( I 112 r,-glc-x- I T6 II-GIcA-Y- f (3-~-glcA-/j-( I -+ 4)-~-glc-P(?)-( 1 u-gal, u-galNAc. manNAc. fucNAc. PyA possibly 2,3- on gal 2-AcNH-2,6-dideoxy-~-talose D-gal, D-glc, - 2)-~-glcA-P-( I - ~)-L-~ucN. D-glcA-/-( I -+ .i)-L-fucN-( 1 + 4)-1>-gl~. 0 .. 2)-~-gal-a-( 1 4 3)-~-glc-a-( 1 + 3)-~-rham-r-( I -+ 3)-ribitol- 1 or 2. or 4 or 5-0 P '0- ONa D-GIcNAc-( I . - 4)-D-&-B-. f. [ f I,4,6-~-galNAc D-Gal-P-( 1 - 3)-~-rham-( 1 4)-D-gkA-P-( 1 + 4)-D-gk-P-( 1 ---t 4)-D-glC-X-( I --* 4)-D-gal-( I -D-gkA-a-( I + 3)-~-glc, glcA-( 1 --t 3)-glcNAc, manNAc-( 1 --t 3)-glc-( I 4 3)manNAc gal(f), galN, glcN, ribitol, PO,. 3)-D-gal-( 1 4)-D-gk-Z-( 1 + 6)-D-gIC-( 1 4 4)-D-gal-( 1 T" I XIA(XL1ll) 1 HOCH,~CHOH.CH,O.P(ONa)O .. 0 XIV xv XVIII XVIIIA XIX xx XXII XXIII xxv XXVII XXVIII 3)-o-glcyAc-( I 3)-D-gal-( I -+ I --t 6)-~-glcNAc-( 1 - 4)-D-gIc-( 1 T4 D-GlC-( I P-D-Gal-( 1 + 4)-D-gk-( 1 z-D-Gai-( 1 gal. glc, galN, glcN, glycerol, PO,. 4 3)-~-gal-( 1 + 4)-D-gk-a-( 1 -+ 6)-D-gk-( 1 -+ 3)-~-rham-( 1 + 4)-D-gk-( 1 HOCH 2 ' CHOH ' CH,OP(OH)O- 111 -0Ac probably on D-gal or alternative with single glc and isomaltosyl residues exchanged D-gal, D-glc, rham, glcNAc, glycerol, PO,, 2: 3.5: 1 : I : I : 1 ; no -0Ac. glc. rham, D-manNAc, PO,, 1 :2: 1 : 1 (49); gal, glc, rham. galN. glcN, PO, gal. ~IC. glcN, PO4 D-GIcA-x-( 1 -+ 3)-u-glc. ~i-gal-p-( I -+ 3)-~-ara. o-gal-/l-( 1 - 5)-~-ara. - 4)-rham D-gk-/j-( 1 -+ 4)-I>-gk, erythritol, Po,. ,+ 2)-~-rham-( I-, -+ 4)-D-gk-(l-. I,-gd gal. galA, 2 aminosugars gal. glc, rham. glcN. Po,, PyA glc. rham, glycerol, PO, (75) (46) (47) (5 I. 52) (53) (46. 54) (46) 72 MICHAEL HEIDELBERGER and WOLFGANG NIMMICH Pn type XXIX 1;;- ~,-galNAc-/??-( I - 6)-u-gal(f)-/I'?-( I - 3)-u-gal-p!-( I -+ 6)-~-gal(f)-[f!-( I -+ .. 0 o Not uniquely determined - Table 4. Qualitative and quantitative Data on Klrhsidlo K Polysaccharides K Anti-Pn 11057C 11513C [117'92C polysacch. Homologous: 1024 4040 Mx) 72 75 XI - + - IV608C V555 VI6XIC VllY37C VIIIIWX - X627C XV628 X64 770 Immunochemical Relationships of Pneumococci and Klehsiellu 13 due to lateral non-reducing end-groups of 1,-glucur- onic acid [I, 58,591 linked r-1.6- to D-glucose [60]. a minor specificity arising from consecutive residues of L-rhamnose linked 1.3- in the main chain, and another due to 1.6- or 1,4- linked D-glucose. End- groups of D-glucose, such as those in glycogen and certain dextrans, may also fit into sites on antibodies designed for D-ghCUrOniC acid [59. lo]. Of the 52 K-substances tested which contained glcA, 20 reacted extensively with anti-Pn 11. One of the earliest pneumococcal cross-reactions noted, that of the K2 polysaccharide in anti-Pn I1 [61], is now known to be caused by the presence, as lateral substi- tuents in the repeating units (Table 2). of multiple [62] non-reducing end-groups of D-gkA [ 151. Such end-groups have also been found in K8 [ 191. K9 [ZO] (in which they were predicted on the basis of its heavy cross-reaction), K20 [23] and K59 [33]. These serolo- gical reactions have the advantage over chroma- tographic methods of identifying the glcA as ~-glcA; but anti-Pn I1 does not distinguish between D-&A and 4-O-methyl-~-glcA. Multiple lateral non-reducing end-groups of r,-glcA will probably be found in the remaining heaviest precipitators of anti-Pn 11. namely. Maximal precipitation at 0°C; quantitative Analyses calculated to 1.0 ml L Pard Para 5 K XV111495C XIX631C XX616C XXllSh6 XXlllY12C XXVS13C XXVllhhXC XXVIIISIOC' XXIXh42C' Ty A T!B .;ciillcnhc.rg polysacch. 2200 2x0 755 x7x 11s I n6 277 735 3x9 TY 730 332 3ini I 2 3 4 5 6 7 8 9 10 11 12 13 14 I5 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 41 48 49 50 51 52 53 54 55 56 57 58 59 62 63 64 66 72 75 81 Footnotes to Table 4 on pp. 8-9. 74 MICHAEL HEIDELBERGER and WOLFGANG NlMMlCH Footnotes to Table 4 Sera with the designation C were absorbed with pneumococcal group-specific C-polysaccharide; sera not marked with C contained little or no anti-C. .' K49 supernatants + K63 gave 62 jig N; further data as well as pptn. in anti-Pn 1 704. in 1561. K63 supernatants + K49 gave IO1 pgN [56]. Serum absorbed with ketha gum [57] failed to ppt. with K49 or with subsequently added K63. K2 supernatants + K9 pptd. 208pg N; K9 supernatants gave 1025pg N with K? 1561. KI5 supernatants + K2 gave 8oOpg N: + K23 gave 313pg N. after which K25 pptd. 70pg N; K15 supernatants + K25 gave 106pg N, after which K23 pptd. 280pg N. ' K70 supernatants gave 819pg N with K2. 46pg N with K 8; after pptn. with K20 and KX. K23 gave 287 pg N. K23 supernatants + K25 gave 188pg N. K25 supernatants + K8 pptd. 86pg N; + K9, 664pg N; + K15. 523pg N, after which K2 gave 800jrg N. K27 supernatants + K2 pptd. 51 I pg N and reacted weakly with K8. K30 s~ipernatants + K2, K9. K55 gave 17. 123. 600pg N, resp. Serum absorbed with K30 and K55 gave Y33jcg N with K51; the cross-reactive ppts. totaled 3283 jig N. K34 supernatants pptd. 777 pg N with K9 [56]. *' K35 superii~~t~~nts gave 275pg N with K36; K36 sup~rnatants pptd. 56pg N with K35; the combined K35, K36 and K36. K35 supernatants gave a further 196pgN with K25. K44 supernatants pptd. 13Hpg N with K72, 245pg N with K35; serum absorbed with K44 and K72 gave 350pg N with K8. K45 supernatants gave 935pg N with K2. ' K48 supernatants pptd. K35 as did intact serum and gave 610 pg N with K4, possibly owing to copptn. of soluble K48-antibody complexes; K48, K35 supernatants gave 240pgN with K8. K51 supernatants gave no ppt. with K2 but pptd. lobpg N with K72. K55 supernatants + K51 gave 840yg N. K72 supernatants gave 128pg N with K48, 630 with K4, 418 with K36. ' Deacetylated K5 gave 57pg N; K5. deAcK5 supernatants pptd. 55, 54pgN with Pn S VIII; intact serum gave 1 l2pg N. When the KS, S VIII and deAcK5. S VI11 supernatants were mixed no further pptn. occurred. Subsequent addition of the polysaccharide of Fornes aiinosus, which also contains fi-linked u-glc and n-glcA residues, but not cellobiouronosyl [64], gave 83pg N; intact serum gave 115. Anti-Pn 111 absorbed with deAcK5 gave 22pg N with deacetylated Arthrobactrr NRRL B-1797 polysaccharide which contains D-gal, D-&. u-glcA. PyA [65]. Because of this reaction and a much stronger one in ar.ti-Pn VI11 it may be considered to have cetlobiouronosyl groups; intact anti-Pn I11 gave 60. K25 supernatants + S VI11 gave 96pg N. K39 supernatants reacted with K25 and gave 21 pg N with K5; supernatants from the second pptn. gave 28pg N with Pn S VIII. Anti-Pn 111 792C absorbed with S VI11 gave no ppt. with K39. * K1 supernatants + K 13, K63 gave 48, I79 pg N, resp.; K I, K 13 supernatants + K29 pptd. 50 pg N. KI 3 supernatants gave 212pg N with K63. as in intact serum. K29 supernatants pptd. 42, 17, traces, 203 jig N with Kl, K13, K42, K63, resp. From [54], in which additional data are to be found. 'K37 supernatants from which an average of ea. 4OOpg N had been pptd. gave 552pg N with K32, 656 out of 928 with deiijif'uvylated Pn S IV, and reacted with K29, K63. K42 supernatants which had pptd. an average of 46 pg N gave 37, 64, 69pg N with K1, K13, K29, resp. The 42, K13 supernatants pptd. 323pg N with K37; the K42, K29 supernatants gave 327 pg N. K63 supernatants gave 698 pg N with K32 [56]. + + with K27. ' K2 supernatants pptd. 30pg N with Pn S 11; intact serum gave 63 at level used [69]. K8 value from [56]. Serum which had pptd. 450jig N with E. coli K85 [71] gave 53pg N with K8. K9 supernatants + degraded gum arabic gave 48 pg N 1561; intact serum gave 283 1697. K I5 supernatants + K8, K9 gave 0. 9 pg N. resp. K25 supernatants gave 3 pg N with K8. K27 supernatants pptd. 127 pg N with K25. K36 supernatants + K8 gave 7pg N. K45 supernatants + K59, K9 gave 37, 31 pg N resp.; supernatants from K45, K9 + K8 pptd. 8pg N. K62 supernatants + S 11, K8, K27 gave 36, 19, 453pg N, resp; serum absorbed with K62 and K8 gave 530pg N with K25. 'K12 supernatants + S I1 gave 5pg N; intact serum pptd. 21 [7]. K12 supernatants pptd. guar gum much as did intact serum. Anti-Pn VI which had pptd. 65pg N with guar gave 45pg N with K12. K45 supernatants pptd. 14pg N aith K47. K50 su~rnatants + Kl2 gave 45jig N, as in intact serum at the ievel used: + S 11, l2pg N. + K47. + + f ; supernatants from pptn. with K12 gave IOpg N with K45. "' Supernatants gave I1 pg N with S XIV; intact serum pptd. 35 191. Deacetylated K5 pptd. l8pg N. Serum absorbed with K5 or Ki8 pptd. as much N with K18 or K5 as did intact serum; the doubly pptd. sera gave 208. 48pg N with S 111, oat glucan: intact serum gave 203, 63pg N at the levels used. K39 supernatants gave 15pg N with L~~~I~IJJCCT srarkryi out of 41 [66. 671; + K5 gave 10pg N; supernatants from this pptd. 97, IoOpg N with S Ill, S XlX, the batter as in intact serum. K44 supernatants pptd. S 111, S XIX as did intact serum. Supernatants from the K44, S 111 absorptions gave no ppt. with K5. ' K21 supernatants + K64 gave 31 pg N out of 41 [72]. K50 supernatants + K64 pptd. 24pg N. KI supernatants pptd. l2pg N with K39, 53pg N out of 104 E731 with streptococcal group L substance [74J K31 supernatants + KI gave 19pg N; + S XIV, 48pg N out of 60 [9]; + oxidized-reduced E, coli K85, 57pg N out of 70 [71]. Supernatants from the pptns. with K1 and oxd.-red. E. coli K85 gave 50, 45pg N, resp., with S XIV. K39 supernatants pptd. 30, l4pg N with KI. K31, resp.; serum absorbed with K39 and K31 gave 23pg N with S XV; intact serum gave 60 [73]. K59 supernatants pptd. with K1 as did intact serum. K2 supernatants + depyruvylated Khizohiirm rilelibti B (dp Rh.m.B) gave 26 pg N out of I I I [54]. K22 supernatants pptd. with K2,. K33. dp Rh.m.B. gave 44. traces. 19pg N resp. K33 supernatants + K2 gave 18pg N; + K22, 48 jig N; supernatants from this second pptn. gave 4pg N with K43. K43 supernatants gave no ppt. with K2. K39 supernatants + S VI1 gave no ppt., + dextran N236 gave 35 pg N, (intact serum pptd. 47 pg [47]); serum absorbed with K39 and dextran gave no ppt. with S VIII. K43 supernatants failed to ppt. with K57, S XIA; the latter [46a] gave 278 pg N with intact serum [73]. K57 supernatants + oxidized-reduced Sp~robo~offt~c~s aeetylphos- phoglucogalactan gave 688pg N; intact serum gave 1200 1771; + Liponiycc,s lipu/iws [66], pptd. 72pg N out of 173 ~671. ' K34-supernatants + K48 gave 171 pg N [56]. K48 supernatants pptd. 48 pg N with K34; + S VIII, 66pg N out of 124 [9. 681. Immunochemical Relationships of Pneumococci and Klchsiella 75 'K22 supernatants gave 1Opg N with K24; with S 11, 2Opg out of 40[9]; serum absorbed with K22 and K24 gave 16pg N with S 11. K39 supernatants + K22 pptd. 16jcg N, as in intact serum, but pptd, only 24 jig N with s 11. "K17 supernatants pptd. 40. 34pg N with K47. K52. resp. K39 supernatants gave 50pg N with K47, + + & with K17. K47 supernatants + K52 gave 30pg N; K52 supernatants + K47 pptd. 40jig N [56]. K53 supernatants + "K10 supernatants gave as much ppt. with K14, K56 as did intact serum: ahsorption with KIO and K14 left I II pg N lor K17, 47 for K56; pptn. with KIO and K56 left 114pg N for K17. K14 supernatants pptd. 26jig N with K49. after which K47 gave I87pg N. K17 supernatants + streptococcal group B gave 2pg N; intact serum pptd. 209 [73]. K17 gave + + + in a potent Squibb anti-Pn XXIll rabbit serum. K34 supernatants + K64 gave 71 pg N; serum which had pptd. 39pg N with streptococcal group F zl [68] gave 34pg N with K34. K47 supernatants from an earlier bleeding gave no ppt, with streptococcal group B; intact Serum gave 97pg N at the level of B-substance used 1511; the serum, absorbed with B, gave 82pg N with K47 [%I. K49 supernatants pptd. K47 as did intact serum and gave 70pg N with K64; serum absorbed with K49 and K64 yielded 35pg N with K34. cj: also [56]. K56 supernatants gave Ill pg N with K17, 112 with K19. 176 with K47. K64 supernatants pptd. 31 pg N with K34. K47. K52 gave 25. 9pg N. resp. 'Supernatants which had averaged 47pg N with K63 gave 150pg N with S XXV. YK5 supernatants + K32 gave 60pg N. K6 supernatants pptd. 18pg N with K7; with Rhizobium rrfolii UNZ29, 18pg N out of 78 [54]. K7 supernatants failed to ppt. with K6 but gave 62pg N with K32. K32 supernatants [54]. K50 supernatants + K56 gave 64pg N; + K5 pptd. 40pg N. K56 supernatants + K32 pptd. 60pg N; supernatants from this gave 7 pg N out of 80 with deacetylated Rhirohiurii trifolii TAI [54]. 'Serum from which carob mucilage had pptd. an average of 32 pg N gave 3 pg N with K52. K4. 15. 23, 25*. 27. 30. 33. 45. 51, and 55, although all contain u-glc as well. K4. dissolved with the aid of brief exposure to alkali, precipitated much more antibody (1Wpg N) than did the solution listed in Table 4 and [56], poss- ibly owing to removal of a hindering -0Ac. Footnote ', Table 4, shows that K2 precipitates most of the fraction of antibody reactive with K9; the reverse is not true, possibly because the end- groups of D-glcA in K9 are in the 6-anomeric form, while in Pn SII, the antigenic determinant of Pn 11, and in K2:they are the r-form. Precipitation of K9 in anti-Pn I1 is probably reinforced by the tandem 1,3-linked residues of L-rhamnose similar to those which give rise to a minor partial specificity of Pn I1 [SS]. However K8, in which the end-groups are E-, precipitates less antibody than K20, in which they are 6- and there is no rhamnose. In K2, K8, K9 and K20, the end-groups are not linked to D-glucose. as in SII, so that spatial and conformational relations must be modifying factors in addition to the chemical nature and linkage of the adjoining sugar. Footnote also shows that K15 precipitates about one-half of the antibody reactive with K2 as well as a large part of the fraction of antibody precipitable by K23 and K25. Similar information is given for other anti-Pn II- precipitating K polysaccharides in footnotes 'vd and '. With respect to K44, a rather weak reactor, the data are explicable if K44 has few or no end-groups Of D-glcA, or if such end-groups are partially blocked by -0Ac. The fraction of anti-I1 precipitated is one- half of that reactive with K72, which contains no &A. Its string of L-rhamnose residues (Table 2) is undoubtedly responsible for the cross-reaction and it is therefore likely that two or three such consecutive residues will be found in K44, which contains glcA, glc, and rham. Prior precipitation with K72 scarcely * Niemann and Stirm (personal communication) report that K25 has D-GIc-(I- and -2)-~-glcA-(l--, either of which could account for the cross-reactivity, especially as 1.2-linked sugars may function as somewhat hindered end- groups (see anti-Pn V and VI). affects the cross-reaction of K48 (footnote '). Since K48 contains galA and no glcA, but does have glc and rham, the last is either hindered by a substituent or linked otherwise than 1,3-. Evidently the reaction is mediated by D-glucose. K51 cross-reacts more massively in anti-Pn I1513 than any of the others tested, precipitating 67% of the antibodies. Of the large number of polysacchar- ides of varying origin which react specifically with this serum, only that of Brachychiton exceeds this pro- portion, precipitating 75% in the bleeding presently in use. K30 and K55 are the next best, and it would be interesting to know whether or not in these two, and in K51, the glcA is linked a-1 + 6 to D-glc, as in Pn SII. Such linkage would not necessarily be expected, since ~-GlcA-r-(l-+ 2)-~-rham has been iso- lated as a partial hydrolytic product of Brachychiton gum [70]. A solution containing 6000pg anti-Pn I1 nitrogen per ml, made from Squibb's purified, freeze-dried, rab- bit anti-Pn 11, gave 36pg N with K2 and 122pgN with K30, also the following qualitative tests: with K8 k,KIS(++k),K23(++++).K33 (+++). A New York State Department of Health Bureau of Laboratories rabbit anti-Pn I1 serum, bleeding of 11/10/45, with 3100pg anti-Pn 11 nitrogen per ml, gave + + + with K2,23,33, f with K8,25, + + with K9, 15, + + f with K4, 27, 45, + + + f with K30, + + + + with K51, 55. Cross-rerrctions in unti-Pn I11 arid -VI11 Multiple residues of cellobiouronic acid in the main chain of the repeating unit are obviously the cause of the cross-reaction of K5 [17,63] and will undoub- tedly prove to be the reason for the heavier precipi- tation by K39 which includes more than one-half of the antibody precipitated by K5 (footnote '). K5 and its deacetylated derivative precipitate the same amount and the same fraction of anti-111, and this is a portion of the antibody reactive with S VI11 and with other polysaccharides containing, or probably containing, multiple cellobiouronosyl residues (footnote '). The behavior of K25 is atypical, since it does not precipi- tate anti-Pn VI11 or remove much of the antibody 7 0 MICHAEL HEIDELBERGER and WOLFGANG NIMMICH precipitated by S VI11 from anti-Pn 111. K66 also pre- cipitates anti-Pn 111 but not anti-Pn VIII. The cello- biouronic acid units of Pn S Ill are linked through the 3-position of glcA and all linkages in S VI11 are 1-4. In anti-Pn VIII, K5 and K18 apparently precipi- tated different fractions of antibody (footnote "), but the amounts were small. K18 did not precipitate anti- Pn I11 and therefore probably does not contain cello- biouronosyl residues. Possibly the reaction in anti- VI11 is caused by -+4)-p-~-glcA( 1- residues. Tests of the supernatants from precipitates by K39 showed that a corresponding portion of the antibody precipi- table by S 111 had disappeared. A part of the fraction reactive with Lipoinyces starkei [66,67] was also in- volved, but not the fraction which precipitates with S XIX [SI. This is believed to react because of 1,4- linked glucosyl residues [68]. The supernatants from K44 precipitated S 111 and S XIX as did intact serum: after the reaction with S 111, K5 gave no precipitate. Cross-reactions in anti-Pn I V K1 reacts partly with the portion of anti-Pn IV reactive with K13 and K63; serum absorbed with K1 and K63 gave 5Opg N with K29 (footnote ". How- ever, K13 supernatants precipitated K63 as did intact serum, possibly owing to coprecipitation of soluble K13 anti-Pn IV complexes. K29 supernatants showed little overlapping with K1 and K63 but much with K13 and K42. K13, K29, and K42 all contain galac- tose and pyruvic acid (Table I), as does Pn S IV [40], but the structures of the three Ks are unknown. K1 contains no sugars in common with S IV, but if its fucose has the D-configuration and is linked 1,4- it might fit into anti-Pn IV spaces designed for 1,4- linked D-galactose. K63 has no pyruvic acid (unpub- lished data), but contains galactose and fucose. Its structure is not known. K32, also of unknown structure, was the heaviest reactor in anti-Pn IV608C (42% of the antibody, foot- note h, and precipitated every anti-Pn IV tested [54,56]. Like S IV. it consists partially of gal and PYA. K37 also precipitated anti-Pn IV strongly (foot- note '). It contains D-galactose, D-glucose, and 4-O-(l- carboxyethy1)-D-glcA [27]. There seems to be no obvious reason for its cross-reactivity, which involves a portion of the fractions of antibody precipitated by K32 and depyruvylated S IV. The antibody reactive with K42 is partly a portion of that precipitated by K1, K13, K29, and K37. E. R. Squibb and Sons' puri- fied rabbit anti-Pn IV, at 4850pg N with PnS IV, gave 1610pg N with K32 and 44pg N with K63. tains 2-substituted D-gIcA. like S V, which also fails to precipitate anti-Pn I1513 1691. Cross-reactions in anti-Pn V As in studies with other polysaccharides [69], many of the Ks which react strongly in anti-Pn I1 precipi- tate less heavily in anti-Pn V, a difference ascribed to the +2)-~-gkA$-(- linkages of S V [41], which, in certain conformations, might leave positions 3, 4 and 6 of the glucuronic acid available for reacti- vity with antibody designed for the unsubstituted molecule. However, some of the heaviest precipitators of anti-Pn 11, such as K20, K30, K33, K48, K51 and K55 show little or no reactivity in anti-Pn V. K62 precipitates anti-V but not anti-11, and perhaps con- C,.os.s-rc'actioIls ill anri-Pn VI K50 was the strongest reactor in anti-Pn VI681C, followed by K12, which removed much of the anti-Pn VI precipitable from this serum by S I1 (footnote I), K12 will probably be found to contain L-rham linked l,3-, as the cross-reaction between Pn I1 and -VI has been traced to the presence of multiple residues of l,3-, iis the cross-reaction between Pn IJ and VI has of both types [7]. (See also Table 3). Footnote ' also shows that u-gal, as non-reducing end-groups, or linked 1,2- as in S VI, is not involved in this cross- reaction. K50 appears not to react with the fraction of antibody which precipitates with multiples of 1,3-, linked L-rham. but it is possible that soluble K50 anti- Pn VI complexes might have coprecipitated with K12. K45 precipitates a portion of the antibody reactive with K47, the structure of which is known. However, there are three sugars in common. Cross-reactions in anti-Pn VI1 K15 precipitates a portion of the fraction of anti- body cross-reactive with Pn S XIV (footnote '"), which also contains D-gal and D-glc [75], as does S VI1 (Table 3). The limited cross-reactivity (+ + k) of K9 may be due to residues of 1,3-linked L-rham common to S VI1 and K9. Precipitation by K24 and K47 is difficult to explain, (cf: Tables 2 and 3), while that of K52 may be due to its non-reducing end-groups of D-gal. Possibly the two aminosugars of S VII, D- galactosamine and D-glucosamine, give rise to specifi- cities which cross-react with D-gal and D-gk. Cross-wortions in anti-PnIX K50 is the strongest reactor in this antiserum, fol- lowed by K64. After precipitation with the former, K64 still gave 24 pg N (footnote "), as well as yielding 31 pg N from supernatants from which K21 had pre- cipitated 11 pg N. Intact serum gave 41 pgN with K64[71]. The structure of K50 is not known, but K21 has -+3)-~-glcA-r-( 1 + 3)-~-man-(+25). The internal D-glcA+ man of K64 is probably also linked 1,3 between the sugars 1351. S IX has an internal D-gkA-a-( 1 -+ 3)-~-glc [45b] and this would account for the cross-reactions of K21 and K64 if the 3, 4 and 6 positions of the mannosyl residues could fit into antibody spaces designed for glucose, as has already been postulated for many of the cross-reac- tions of Lipomyces [67]. In K21 the glcA is also branched at C4, while in K24, which does not precipi- tate anti-Pn IX623C, glcA is linked 1,2,4. K2 has non- reducing end-groups of D-gkA attached r-( 1 + 3) to 1,Clinked D-man in the main chain and reacts + + in anti-Pn IX. Cr0,s.s rcY7crions in anti-PI1 x KI, K31. K39 and K59 are the best precipitators of the available anti-Pn X. S X is said to contain galac- tofuranose. galactosamine, glucosamine, ribitol, and phosphate [46], but its structure is not known, nor have those of the first three K substances been eluci- dated, other than that KI has non-reducing end- groups of D-gk. Footnote '' shows that K1, K31, and K39 precipitate much of the same fraction of anti- body, and that this is the portion which reacts with Pn S XV and with the group-specific polysaccharide Immunochemical Relationships of Pneumococci and Klchsirllu 77 of streptococcal group L, composed of D-gal, rham. pgalNAc and D-glcNAc [74]. Although no singlc sugar is shared by all. these five substances and S x will eventually be found to contain at least one immunologically equivalent sugar and linkage. On the other hand, K59. S XIV, and oxidized-reduced E. coli K85 precipitate mainly a different fraction. Reactions rated at + + were given by K15, K22. K24 and K42. Of these, only the structure of K24 is known [25J Cross-reuctioris iri uriti-Pn XV The best reactors in this antiserum were K2, K22. K33 and K43, the last being the most potent. All four of these must contain a similar immunologically reactive structural feature in their repeating units, as they precipitate much of the same fraction of anti-Pn XV (footnote "). This is also the portion reactive with the depyruvylated extracellular polysaccharide of Rhi- zobium meliloti B [54], which, in its intact form, con- tains end-groups of D-gal substituted at 4 and 6 with PYA, -+3)D-gal-B-( I-, and D-gk linked in various ways and partly acetylated [76]. S XV is said to con- tain gal, glc, galN, glcN, glycerol, and PO, [46], but .its structure is not known. Cross-reuctiorls in uriri-Pn XVIII Only K43 and K57 showed noteworthy behavior in this antiserum, both giving massive precipitates of the same fraction of antibody (footnote '). The super- natants from the smaller reaction with K43 failed to precipitate with K57 or Pn S XIA, indicating that K43 had reacted with all of the multivalent cross- reactive antibody and that only antibody univalent with respect to these substances remained (cf: p. OOO). The precipitate with K57 also consisted mainly of antibody cross-reactive with oxidized-reduced Sporo- bolomyces acetylphosphoglucogalactan [77] and partly of the fraction which precipitated Liporiiycrs lipfirus [67]. Possibly an acetylated sugar residue is involved. Cross-reuctioris in uriti-Pn XIX K34 precipitated a portion of the same reactive fraction of antibody as did K48 (footnote '). The structures of K34, K48 and S XIX are not known, but glucose is the only sugar in common and is prob- ably responsible for the cross-reactions, as the super- natants from K48 gave only one-half of the antibody precipitated from intact serum by S VI11 (footnote Cross-reactions iri unti-Pn XX S XX contains galactose, glucose, and glucosamine [46]. All of the K substances which precipitate + + Or more in this antiserum contain glucose and all but K24 contain galactose as well. K22 and K39 react with a portion of the fraction of antibody precipitated by S I1 (footnote '). Since they do not contain L-rham, as does S 11. this is possibly an indication that they might have their u-glucose linked 1,6- and/or 1.4- [lo, 591. Cross-reactions iri unti-Pn XXII Not much explicit information can be derived from this series. Footnote " shows that K17 precipitates and ref. [68]). a portion of the same fraction of antibody as K47 and K52, both of which contain 1,4-linked D-glcA as well as L-rham bound 1,3,4- in the former and 1,4- in the latter. I,4-linked rhamnose and variously linked glcA are believed to occur in S XXII [SO]. K53 also contains rhamnose and precipitates a portion of the same antibody. Closs-rc~tr~riori.~ in uriti-Pn XXIII The heavy precipitation of K47 in this antiserum has already been noted [54,56] and prompted the prediction that K47 would be found to have non- reducing end-groups of L-rhamnose, as proved to be true [28]. A similar prediction for K56 was also well- founded [31] and it is equally evident that such end- groups will be found in K17 and K19. possibly also in K14 and K64. The small precipitation by K10 is due to another component. These relationships are detailed in footnotes ' and w. Cross-reuctioris in ariti-Pn XXV The anomalous, heavy cross-reactions of K49 and other substances in this antiserum have already been commented upon, as well as the reaction of K63, which appears to be entirely due to antibody ([56,71] and footnote "). An anti-Pn XXV raised in a mule, through the kindness of Dr. Gerald M. Ward, New York City Department of Health Laboratories, Otisville, New York, with 610pg anti-S XXV N per ml, gave 13 pg N with K34, which did not precipitate antiserum 513, and reacted only + with K63. It failed to precipitate K33 or K49. A New York State Department of Health rabbit anti-Pn XXV was negative with K30, K49 and K63, & with K33. Cross-retrctions iri unti-Pn XXVII The sugars in S XXVII are gal, glc, rham and glcNH, [46], but their order and linkages are un- known. A pyruvylated sugar is a principal antigenic determinant [54]. K6 and K7 precipitate mainly the same fraction of antibody (footnote 3. K32 evidently reacts with a different portion of the antibody than do K5, K7 and K56, while K5, K50, and K56 scarcely overlap. The pyruvylated sugar in K5 is D-man [17], in K7 [lS] and K56, D-glc [31], attached as a ketal at 4- and 6-. Cross-rructions in urlri-Pn XXVIII and anti-Pn XXIX Only K52 reacted markedly with anti-Pn XXVIIISIOC @. OOO). The principal structural features of K52 have been determined (Table 2) but those of S XXVIII, which contains glucose, rhamnose, gly- cerol, and PO, [46], are not known. An unusual feature of S XXIX is its content of both D-galactofuranosyl and D-galactopyranosyl residues [MI. It also contains galNAc and ribitol-PO, ([46] and Table 3). One can only guess at the reasons for the limited cross-reactions recorded in Table 4. Cross-rructions iri anti-Salmonella seru Since no massive precipitations occurred in the limited number of available antisera, the reactions recorded in Table 4 are mainly qualitative. All of the 7s MICHAEL HEIDELBERGER K substances which precipitate + + or + + + con- tain at least two sugars in common with the corre- sponding 0 antigen. K24. like the antigen of group A, contains 1,2- linked D-man, as does K28 [26]. Un- fortunately, little is known of the structures of the other reactive Ks, except for K7, which precipitates ariti-parntyphi A for no obvious reason. K28 and K38 are said to have non-reducing end-groups of P-D-glC, but the former precipitates anti-parutyphi A and anti- S. ser$tc~riherg (antigen I?) while the lattcr fails to react with anti-para A and precipitates anti-se~ftenberq rather weakly. KI is also said to contain end-groups of 11-glc but reacts + + only in anti-par-utyphi B, of the three sera tested (Table 4). CONCLUSION The present compilation of new and previously obtained data provides insights relating the chemical constitution and structure of many KIrhsiella K poly- saccharides to their immunological specificity. Both the qualitative and quantitatiye serological tests are relatively simple and rapid. Positive cross-reactions vary from the immediate, even at 0°C. the tempera- ture preferred cf. for example, [ IO]), to approximately 2 weeks for completion, and, when clear-cut interpre- tation is possible, yield structural information that could take months to obtain by purely organic chemi- cal means. Examples are: 1. Although the structure of Pn S I is only partially known (Table 3), its content of D-galA permits the prediction that in K49 and K63, which react strongly with anti-Pn I, the galA will be found to be in the D-form and &*east partly in linkages similar to those in S I. 2. The strong precipitation of K2, K8, and K20 in anti-Pn I1 confirms the assignment, based on chemical studies, of non-reducing end-groups of D- glcA to the repeating units of these substances (cf. Tables 2 and 4). 3. The presence of such groups in K9 and K59 was predicted from their heavy precipitation of anti- Pn I1 and later confirmed chemically [20,33]. 4. It is probable that non-reducing end-groups of D-glc might fit into sites on molecules of anti-Pn 11 homologous for D-gkA. This explanation of the reac- tivity of glycogen [lo, 781, certain dextrans [59] and synthetic polyglucoses [78] in anti-Pn I1 somewhat weakens the prediction (p. 75) that end-groups of D-glcA would be found in K15, K23, K27, K30, K33, K45, K51 and K55. However, K1, K18, K28, K38 and K54, which are known to possess end- groups of D-glC, react not at all or only very weakly in the anti-Pn I1 used. Moreover, glycogen and a large number of dextrans precipitated much smaller amounts of antibody than did the first series of K- substances in the same antiserum. 5. Presence of cellobiouronosyl residues in K5 was suspected during structural studies E171 and con- firmed by the cross-reactivity of K5 in anti-Pn 111 and -VI11 (Table 4) and by its precipitation (footnote ') of a portion of the same fraction of anti-Pn 111 as was reactive with S VIII. Although similar reason- ing was used to identify a glcA- and glc-containing aldobiouronic acid isolated from the slime elaborated by Sphaerotilus naturis [SI]. it now appears that a and WOLFGANG NIMMICH polysaccharide such as that of the fruit bodies of the mold Foriirs U~II~OSLIS. which contains the residues -D- glcA-fl-( 1 -+ 4)-i,-glcA and -r>-glcA-P-( 1 ---t 3)-~-glc [64] can also precipitate anti-Pn I11 and -VI11 (un- published results). 6. The very heavy precipitation of anti-Pn XXIII by K47 [54] showed that it must contain non-reduc- ing end-groups of L-rham. A prediction of similar end-groups in K56 was made as a result of lower, but still strong reactivity in the same antiserum. Both deductions were verified ([28,31] Table 2). 7. Other instances of reactivity in different anti- pneumococcal sera are discussed in the text, but the interpretation of many promising cross-reactions remains unclear because of lack of knowledge of the structure of the polysaccharide of the pneumococcal type which stimulated the antibody. 8. Comparison of the presently known composi- tions of the Klebsiella K and pneumococcal capsular polysaccharides shows that : (a) many K substances contain D-mannose, which has not yet been found in any pneumococcal type-specific substance; (b) K substances apparently never contain aminosugars, while many Pn S do, some showing as many as three. The extent of participation of aminosugars in the overall specificities of the pneumococcal types con- taining them is still uncertain and requires investiga- tion. 9. In a current extension of earlier studies on cross- reactivities of Klebsiella [ 161 the present investigation has served as a guide to numerous hitherto unsus- pected relationships between the K types. Ack,iowledgements-The technical assistance of W. P. Grosvenor and C. T. Fang is gratefully acknowledged. I. 2. 3. 4. 5. REFERENCES Heidelberger M. (I 960) Structure and immunological specificity of polysaccharides, Fortschr. Chem. ory. Nut- Stoft 18. 503-536. Heidelberger M. (1964) Some recent correlations of constitution and immunological specificity. New Pers- pectiues in Biology. (Edited by Sela M.), pp. 215-224. Elsevier, Amsterdam. Heidelberger M. (1972, 1973) Relatioris chimiyues et im- riiunologiques eiitre polposidc,~ hucthriens (Edited by Bordet P.), Chap. IV and supplement, 273-288 o. Im- munologie, Flammarion, Paris. (a) Heidelberger M. (1973) Immunochemistry of bac- terial polysaccharides. Research in Immunochemistry mid Immuriohiology (edited by Kwapinski J. B. G.. as- sociate editor Day E. D.), Vol. 3. pp. 1-40, Univ. Park Press, Baltimore; (b) Heidelberger M., Kabat E. A. and Mayer M. (1942) A further study of the cross-reaction between the specific polysaccharides of types I11 and VI11 pneumococci in horse antisera. J. ~sp. Med. 73. 3547. Elford W. J., Grabar P. and Fischer W. (1936) Ultra- filtration studies with normal horse sera. Biochem. J. 30, 92-99; Goodner K., Horsfall F. L., Jr. and Bauer J. H. (1936) Ultrafiltration of Type I antipneumococcal sera. Proc. Soc. erp. Biol. Med. 34. 617-619; Heidel- berger M., Pedersen K. 0. and Tiselius A. (1936) Ultra- centrifugal and electrophoretic studies on antibodies. Naturc. Lond. 138. 165; Heidelberger M. and Pedersen K. 0. (1937) The molecular weight of antibodies. J. exp. Med. 65. 3934 14. Immunochemical Relationships of Pneumococci and Klehsicllu 79 6. Horsfall F. L.. Jr., Goodner K.. MacLeod C. M. and Harris A. H. (1937) Anti-pneumococcus rabhit serum as a therapeutic agent in lobar pneumonia. J. Am. rnc~l. Ass. 108. 1483-1490; Heidelberger M., Turner J. C. and So0 Hoo C. M. (1937-8) Preparation and administration of globulin from rabbit antipneumo- coccus sera. Proc. SOC. exp. Biol. Med. 37. 734736. 7. Heidelberger M. and Rebers P. A. (1960) Immunoche- mistry of the pneumococcal types 11. V, and VI. J. Bact. 80. 145- 153. 8. Heidelberger M. and Rebers P. A. (1958) Cross-reac- tions of polyglucoses in antipneumococcal sera -~vI. .I. Am. chem. Soc. 80. I 1G118. 9. Heidelberger M. and Tyler J. M. (1964) Cross-reactions of pneumococcal types. J. c~xp. Med. 120. 71 1-719. 10. Heidelberger M., Aisenberg A. C. and Hassid W. Z. (1954) Glycogen, an immunologically specific polysac- charide. J. c'zp. Med. 99. 343-353. 11. Nimmich W. (1968) Zur Isolierung und qualitativen Bausteinanalyse der K-Antigene von Klebsiellen. Z. Med. Mikrobiol. Immunol. 154. 117-131. 12. Nimmich W. (1971) Uber die spezifischen Polysacchar- ide (K-Antigene) der Klebsiella-Typen K73-K80. Acta biol. med. germ. 26, 397403. 13. Lindberg B., Samuelsson B. and Nimmich W. (1973) The Klebsirlla type 38 capsular polysaccharide. Carho- hydr. Res. 30. 63-70. 14. Barker S. A., Brimacombe J. S., Eriksen J. L. and Stacey M. (1963) Capsular polysaccharide of Klebsiella pneu- monia type A (strain 1265). Nature, Lond. 197. 899-900. 15. Gahan L. C., Sandford P. A. and Conrad H. E. (1967) Structure of the serotype 2 capsular polysaccharide of Aerobacter aerogenrs. Biochemistry 6, 27552767. 16. Eriksen J. (1965) Cross-reactions in the Klebsiella group. 10. Structure of the capsular polysaccharide of type 3 (C). Acta path. microbiol. scand. 64. 347-361. 17. Dutton G. G. S. and Yang M.-T. (1973) Structure of the'i?apsular polysaccharide of Klebsiella K-type 5. Can. J. Chem. 51. 1826-1832. 18. Dutton G. G. S., Stephen A. M. and Churms S. C. (1974) Structural investigation of Klebsiella serotype K7 polysaccharide. Carbohydr. Res. 38. 22S237. 19. Sutherland I. W. (1970) Structure of the Klebsiella aer- ogenes type 8 polysaccharide. Biochemistry 9. 2180- 2185. 20. Lindberg B., Lonngren J., Thompson J. L. and Nim- mich W. (1972) Structural studies of the Klebsirlla type 9 capsular polysaccharide. Carbohydr. Res. 25. 49-57. 21. Stirm S. (1975) Carbohydr. Res. 41, 241-255. 22.. Dutton G. G. S. et a/. unpublished results. 23. Choy Y.-M. and Dutton G. G. S. (1973) Structure of the capsular polysaccharide of Klebsiella K-type 20. Can. J. Chem. 51. 301S3020. 24. Dutton G. G. S. and Choy Y.-M. (1972, 1973) Capsular polysaccharide from Klebsiella type 21, Carbohydr. Rrs. 21. 169-172; Can. J. Chem. 51. 198-207. 25. Choy Y.-M., Dutton G. G. S. and Zanlungo A. M. (1973) Can. J. Chem. 51. 1819-1825. 26. Lindberg B. et a/ (1975) Carhohydr. Res., 42, 95-105. 27. Lindberg B., unpublished results. 28. Bjorndal H., Lindberg B., Lonngren J., Nimmich W. and Rose11 K. G. (1973) Structural studies of the Kleb- siella type 47 capsular polysaccharide. Carhohydr. Res. 29. Bjorndal H., Lindberg B., Lonngren J., Mesziros M., Thompson J. L. and Nimmich W. (1973) Structural studies of the capsular polysaccharide of Klebsiella type 52. Carbohydr. Res. 31. 93-100. 30. (a) Sandford P. A. and Conrad H. E. (1966) Structure of the Aerobircter uerogqcwc's A3 (SI) poljsoccharide-I. Re-examination using improved procedures for methy- lation analysis. Biochemistry 5. 1508-1517; (b) Conrad H. E., Bamburg J. R., Epley J. D. and Kindt T. J. 27. 373-378. (1966) Sequence analysis and hjdrolysis studies. Bio- chemistry 5. 2808-281 7. 31. Choy Y.-M. and Dutton G. G. S. (1973) Structure of the capsular polysaccharide of Klrhsicllu K-type 56. Can. J. Chem. 51. 3021-3026. 32. Lindberg B.. unpublished results. 33. Lindherg B. et trl. (1975) Crrrholij.tlr.. Rcs. 42, 83-93. 34. Dutton G. G. S., unpublished results. 35. Barker S. A,, Foster A. B., Siddiqui I. R. and Stacey M. (1958) Structure of an acidic polysaccharide elabor- ated by AerohactcJr aerogenrs. ,Yarurc. Loiid. 181. 999. 36. Choy Y.-M. and Dutton G. G. S. (1974) Structure of the capsular polysaccharide of Klehsic4la K-type 72. Occurrence of 3.4-0-(1-carboiyethylidene)-~-rham- nose. Ctrn. J. Chcwi. 52, 684 - 687. 37. Guy R. C. E.. How M. J., Stace) M. and Heidelberger M. (1967) The capsular polysaccharide of type I pneu- mococcus-I : Purification and chemical modification. J. hiol. Chmt. 242. 5106-51 I I. 38. Lindberg B., unpublished results. 39. Reeves R. E. and Goebel W. F. (1941) Chemoimmuno- logical studies on the soluble specific substance of pneumococcus: structure of type 111 polysaccharide. J. 40. (a) Higginbotham J. D. and Heidelberger M. (1972) The specific capsular polysaccharide of Pneumococcus type IV. Carbohydr. Res. 23. 165-173; (b) Higgin- botham J. D. and Heidelberger M. (1973) Oxidation of the capsular polysaccharide of pneumococcal type IV by periodate. Carbohydr. Res. 27. 297-302; (c) Lew J. Y. and Heidelberger M., unpublished results. 41. Barker S. A,, Bick S. M., Brimacombe J. S., How M. J. and Stacey M. (1966) Structural studies on the cap- sular polysaccharide of pneumococcus type V. Carbo- hydr. Res. 2. 224233. 42. Rebers P. A. and Heidelberger M. (1959, 1961) The specific polysaccharide of type VI pneumococcus-I: Preparation, properties, and reactions. J. Am. chem. Soc. 81. 2415-2419; 11: The repeating unit. J. Am. cheni. Soc. 83. 3056-3059. 43. Chaudhari A. S., Bishop C. T. and Fielder F. S. (1972) Structural studies on the specific type VI1 pneumococ- cal polysaccharide. Carhohydr. Res. 25. 161-172. 44. Jones J. K. N. and Perry M. B. (1957) The structure of the type VI11 pneumococcus specific polysaccharide. J. Am. chrm. Soc. 79. 2787-2793. 45. (a) Higginbotham J. D.. Das A. and Heidelberger M. (1972) Immunochemical studies on the capsular poly- saccharide of pneumococcal type IX. Biochem. J. 126. 225-231; (b) Das A., Higginbotham J. D. and Heidel- berger M. (1972) Oxidation of the capsular polysac- charide of pneumococcal type IX by periodate. Bio- 46. Shabarova 2. A,, Buchanan J. G. and Baddiley J. (1962) Composition of pneumococcus type-specific substances containing phosphorus. Biochim. hiophys. Acta 57. 146-148; (a) Kennedy D. A., Buchanan J. G. and Baddiley J. (1969) The type-specific substance from Pneumococcus type XI A (XLIII). Biochem. J., 47. Estrada-Parra S. and Heidelberger M. (1963) The spe- cific polysaccharide of type XVIII pneumococcus-111. Biochemistry 2. 1288-1294. 48. Heidelberger M., Estrada-Parra S. and Brown R. (1964) The specific polysaccharide of type XVIII A pneumococcus. Biochemistry 3. 1548-1 550. 49. Miyazaki 1. and Yadomae I. (1971) Polysaccharides of type XIX Pneumococcus-11: The type-specific poly- saccharide and its chemical behaviour. Carhohydr. Res. 16. 153-159. 50. Rao C. V. N. and Chatterjee B.. unpublished results. 51. Heidelberger M.. Davie J. M. and Krause R. M. (1967) Cross-reactions of the group-specific polysaccharides bid. Chrll7. 139, 51 1-519. chrt~l. J. 126. 133-236. 115. 37-45. 80 MICHAEL HEIDELBERGER and WOLFGANG NIMMICH of streptococcal groups B and G in anti pneuniococcal sera with especial reference to type XXlII and its deter- minants. J. Immun. 99. 794-796. 52. Jones J. K. N. and Perry, M. B., unpublished results. 53. Das A. and Heidelberger M., unpublished results. 54. Heidelberger M., Dudman W. F. and Nimmich W. (1970) Immunochemical relationships of certain capsu- lar polysaccharides of Klebsiella. pneumococci and Rhizobia. J. Immun. 104. 1321-1328. 55. Rao E. V.. Watson M. J., Buchanan J. G., and Baddiley J. (1969) The type-specific substance from Pncumococ- cus type 29. Biochrni. J. I1 1. 547-556. 56. Heidelberger M. and Nimmich W. (1972) Additional immunochemical relationships of capsular polysac- charides of Klebsiella and pneumococci. J. Imniun. 57. Heidelberger M., Tyler J. M. and Mukherjee S. (1962) The cross-reactivity of ketha gum and pneumococcal type I-Short cut to a constituent of a polysaccharide. Inimurioloyy 5, 666672. 58 (a) Heidelberger M. and Adams J. (1956) The immuno- logical specificity of type I1 pneumococcus and its sep- aration into partial specificities. J. c'.~p. Med. 103. 189- 197; (b) J. exp. Med. 111, 3343. 59. Goodman J. W. and Kabat E. A. (1960) Immunoche- mical studies on cross-reactions of antipneumococcal sera-I: Cross-reactions of types I1 and XX antipneu- mococcal sera with dextrans and of type I1 antipneu- mococcal serum with glycogen and Friedlander type B polysaccharide. J. Immun. 84, 333-346. 60. Heidelberger M., Roy N. and Glaudemans C. P. J. (1969) Inhibition by aldobiouronates in the precipi- tation of pneumococcal type I1 and I11 systems. Bio- chemistry 8. 4822-4824. 61. Heidelberger M., Goebel W. F. and Avery 0. T. (1925) The soluble specific substance of a strain of Friedlan- der's bacillus-I. J. exp. Med. 42. 701-707; 11. Avery 0. T., H$;le?berger M. and Goebel W. F. (1925) Chemical and immunological relationships of pneumo- coccus type I1 and a strain of Friedlander's bacillus. J. exp. Med. 42. 709-725. 62. Heidelberger M. and Kendall F. E. (1935) The precipi- tin reaction between type 111 pneumococcus polysac- , charide and homologous antibody-111: A quantitative study and a theory of the reaction mechanism. J. exp. 63. Heidelberger M. and Dutton G. G. S. (1973) Cross- reactions of additional extracellular polysaccharides of Klebsiella. J. Immun. 111. 857-859. 64. Axelsson K. and Bjorndal H. (1970) Polysaccharides elaborated by Fomes annosus (Fr.) Cooke.4: A water- soluble acidic polysaccharide from the fruit bodies. Acta. chem. scand. 24. 713-714. 65. Knutson C. A,, Pittsley J. E. and Jeanes A. (1971) Extracellular polysaccharides produced by Arthro- bacrer species. Am. Chem. SOC. Abstract of meeting, Mar-Apr. 66. Slodki M. E. and Wickerham L. J. (1966) Extracellular polysaccharides and classification of the genus Lipo- mycrs. J. yen. Microbiol. 42. 381-385. 109. 1337-1344. Med. 61. 563-591. 67. Heidclherger M. and Slodki M. E. (1972) Cross-reac- tions of polysaccharides of Liponiyces in antipneumo. coccal and other antisera. Carhohydr. Res. 24. 401-407. 68. Heidelberger M., Willers J. M. N. and Michel M. F. (1969) Immunochemical relationships of certain strep- tococcal group and type polysaccharides to pneumo- coccal capsular antigens. J. Immun. 102. I1 19-1 127. 69. Heidelberger M. (1962) Immunochemistry of the pneu- mococcal types 11. V and VI-IV: Cross-reactions of type V antipneumococcal sera and their bearing on the relation between types I1 and V. Arch Biochem. 70. Hirst E. L., Percival E. and Williams R. S. (1958) The structure of Brachychiton diuersifoliuni gum (Sterculia caudara). J. chem. Soc. 1942-1950. 71. Heidelberger M., Jann K.. Jann B., Brskov F., Brskov I. and Westphal 0. (1968) Relations between structures of three K polysaccharides of Escherichia coli and cross-reactivity in antipneumococcal sera. J. Bact. 95. 72. Heidelberger M. (1963) Cross-reactions of glucosecon- taining polysaccharides in antipneumococcal sera -VU: Precipitation of type IX antisera by isolichenin and other reactions. J. Immun. 91. 735-739. Biophy~. SUPPI. 1. 169-173. 24 15-24 17. 73. Unpublished data from this laboratory. 74. Karakawa W. W.. Wagner J. E. and Pazur J. H. (1971) Immunochemistry of the cell-wall carbohydrate of group L hemolytic streptococci. J. Immun. 107. 554- 562. 75. Barker S. A,, Heidelberger M., Stacey M. and Tipper D. J. (1958) Immunopolysaccharides-X: Structure of the immunologically specific polysaccharide of Pneumococcus type XIV. J. chem. SOC. 3468-3474. 76. Bjorndal H., Erbing C., Lindberg B., Fihraeus G. and Ljunggren H. (1971) Studies on an extracellular poly- saccharide from Rhizobium meliloti. Acta chem. scond. 77. Heidelberger M. and Slodki M. E. (1968, 1970) Pre- dicted and unpredicted cross-reactions of an acetyl- phosphogalactan of Sporobolomyces yeast. J. exp. Med. 128. 189-196; 132. 1105-1106. 78. Heidelberger M., Jahrmarker H., Bjorkland B. and Adams J. (1957) Cross reactions of polyglucoses in antipneumococcal sera-111: Reactions in horse sera. J. Immun. 78. 419426. 79. Heidelberger M. and Ashwell G. (1968) Inibizioni im- munologiche con L- e D-ramnosio. Atti Acad. naz. Lin- cui. Series 8. 44, 695-696. 80. Nimmich W. and Munter W. (1975) Ein neuer Kleb- siella Serotyp K 81. Z. allg. Mikrobiol. (in press). Cur- vall M., Lindberg B., Lonngren J. and Nimmich W. (1975) Structural studies of the Klebsiella type 81 cap- sular polysaccharide. Carhohydr. Res. 42, 7>82. 81. Heidelberger M., Gaudy E. and Wolfe R. S. (1964) Im- munochemical identification of the aldobiuronic acid of the slime of Sphaerotilus natans. Proc. natn. Acad. 25, 1281-1286. Sci. U.S.A. 51. 568-569.