{Reprinted from THE JOURNAL OF EXPERIMENTAL MEDICINE, October 1, 1959, Vol. 110, No. 4, pp. 571-584] Printed in U.S.A. SIMULTANEOUS PRODUCTION OF TWO CAPSULAR POLYSAC- CHARIDES BY PNEUMOCOCCUS I. PROPERTIES OF A PNEUMOCOccUS MANIFESTING BrnaRy CAPSULATION* By ROBERT AUSTRIAN, M.D., anp HARRIET P. BERNHEIMER, Pu.D. (From the Department of Medicine, Stale University of New York College of Medicine at New York City, Brooklyn) Prates 32 To 35 (Received for publication, June 4, 1959) The concept that a pneumococcal cell may produce more than one capsular antigen is not new. The possible occurrence of the phenomenon was considered as early as 1928 by Griffith (1), who described, in his original report of the transformation reaction, a strain of pneumococcus agglutinated specifically by antisera to two different pneumococcal types. In the same year, the serologic cross-reactivity between pneumococcus Type III and a strain classified later as pneumococcus Type VIII was recognized (2, 3). It was shown subsequently by Goebel (4), however, that the immunologic relationship between these two capsular types resulted from the presence in their respective capsular poly- saccharides of a common chemical subgroup rather than from the production by each type of several discrete capsular antigens, one of which was common to both. Similar immunologic cross-reactivity has been demonstrated among a number of other pneumococcal types (5), but in no instance has it been shown conclusively that an organism of this species produces more than one molecular type of capsular polysaccharide. In 1953, Leidy, Hahn, and Alexander (6) reported on the production in vitro of new types of Hemophilus influenzae obtained by transforming capsulated or non-capsulated organisms of this species with deoxyribonucleates (DNA) derived from bacterial cells of heterologous capsuar type. Organisms reacting with each of two unitypic anticapsular sera were isolated and, although defin- itive experiments are not described, it is possible that these cells were pro- ducing capsular polysaccharides of two molecular species. Analogous experi- ments performed with pneumococcus have led to the isolation of a variety of phenotypes manifesting binary capsulation. In this and in subsequent reports, some of their biological, biochemical, and genetic properties will be set forth. * This investigation was supported in part by research grant E-1018(C2) from the In- stitute of Allergy and Infectious Diseases, United States Public Health Service. §71 572 PRODUCTION OF CAPSULAR POLYSACCHARIDES. I Materials and Methods Nomenclature of Pneumococcal Variants.—In certain of the pneumococcal strains to be described, there is lack of complete correspondence between phenotype and genotype. For this reason, complexities arise in the description of these strains. For the present, only pheno- typic designations will be employed and terms descriptive of genotype will be introduced Jater. Fully capsulated phenotypes are indicated by the letter S followed by a Roman numeral designating capsular type; e.g., SHI is a capsulated strain of pneumococcus Type Ill. Be- cause of certain biochemical relationships to be described subsequently and for convenience, non-capsulated variants and intermediate capsular variants agglutinable by antibody to somatic antigens, including those described by Taylor (7) under the term SIII-1, will be designated by the expression S—x,, the subscript designating the capsular type from which the mutant strain was derived and the strain’s designation; e.g., S—1t, is Strain 4 of a non-capsu- lated phenotype derived from pneumococcus Type IIT. Phentotypes manifesting binary capsulation will be indicated by the letter S followed by the two Roman numerals descriptive of the strain’s capsular components separated by a hyphen; e.g., SI-IIT is a pneumococcus characterized by the production of both Type I and Type III capsular polysaccharides. Strains of Pneumococcus.—SI phenotypes—SVI, ID, 1415S, IP: strains of the SI genotype isolated originally from human sources and carried in the laboratory for several years. SII phenotype--IIIA66: a strain of the SIII genotype carried in the laboratory for many years. S—1n phenotypes— S—rn,, Si: strains of a non-capsulated variant of pneumococcus Type II (R36A) which carry different mutated Type IIT genomes acquired by transformation with DNA from variants of Strain IITA66. S—1in,: the same non-capsulated mutant of pneumococcus Type II transformed with DNA of a variant of the Type III Strain SV3. (The three preceding strains are Ephrussi- Taylor’s SIII-1, SITI-ib, and SIII-1c (8) and were made available to us by her and by Dr. Arnold Ravin.) S—n1,1 4 non-capsulated variant of strain TILA66. Preparation of DNA (Transforming Principles) and Technique of Transformation Re- actions.—The methods used were those described by MacLeod and Krauss (9). Preparation of Anticapsular Sera.—Vaccines of unitypic pneumococci and of pneumo- cocci producing two capsular polysaccharides were prepared from cultures grown at 37°C. for 12 to 16 hours. After standing for 1 hour at room temperature following the addition of neutralized formalin to a final concentration of 1 per cent, the organisms were collected by centrifugation and resuspended in a volume of normal saline solution one-tenth that of the original culture. The suspension was heated at 65°C. for 30 minutes and stored at 4°C. Rabbits were immunized by the injection intravenously of 1 cc. of vaccine daily for 5 days and were bled on the 10th day following each course of vaccine. Precipilin Tests.—Precipitin tests were carried out in small test tubes. 0.2 cc. of each of several serial dilutions of antigen was layered over 0.2 cc. of serum. The tubes were incubated for 1 hour at 37°C. and the reactions read after the tubes had stood overnight at 4°C. Precipitin reactions in agar gels were carried out according to the technique of Ouchterlony as modified by Halbert (10). Virulence and Protection Tests.—Tests of virulence were performed by the injection intra- petitoneally into mice of the CFW strain of 1 cc. of an appropriate dilution of a blood broth culture of the strain to be tested which had been incubated for 16 hours. Counts of viable bac- teria] units were obtained by dilution and plating. In protection tests, mice were injected intraperitoneally with 1 cc. of the serum to be tested 24 hours before being infected with pneumococci as in tests of virulence. Control animals received normal rabbit serum. All tests R. AUSTRIAN AND H. P. BERNHEIMER 573 were continued for 10 days and results were scored in terms of survival or death. The cause of death was determined by culture of the animal’s blood at autopsy. In experiments in which the enzyme which hydrolyzes specifically Type HI capsular polysaccharide! was employed, 2 mg. of enzyme were injected intraperitoneally. Tests of Phagocytosis by Human Polymor phonuclear Lewkocytes.—The technique employed was one described previously (11), When the enzyme hydrolyzing Type III polysaccharide was used, it was present in a final concentration of 2 mg./cc. EXPERIMENTAL Phenotypes Manifesting Binary Capsulation-—When cells of pneumococcal Strain S—i, are transformed in the presence of DNA from any of the Type I pneumococcal strains listed above, organisms of two different phenotypes can be recovered: cells of the SI phenotype and cells of the SI-III phenotype. TABLE I Relationships of Pneumococcal Strains Participating in T ransformation Reactions Giving Rise to Variants Producing Two Capsular Polysaccharides Original capsular type of strain manifesting the SI-I11 phenotype following transformation Parent Type ITI strain of variant with S-y11 genome Source of SI genome sI IITA66 SVI SII IILA66 SVI, ID, 1418, IP IlISV3 SVI, ID SII TITA66 SVI, ID, 1418, IP SV IITAG6 ID In similar fashion, SI-III strains may be obtained by transforming cells of lines S—131,, S—11r, and S—111,. From these experiments, it can be seen that mutant Type III cells, the capsular genomes of which were derived originally from two different Type III strains, ITIA66 and ITISV3, can be transformed to the SI-III phenotype with DNA from several SI strains. In addition, it is evident that non-capsulated cells derived originally from capsulated strains of different capsular types, i.e. Type II and Type III, may manifest binary capsu- lation when endowed with the proper genetic factors. It has been possible also to transform non-capsulated cells derived from Type I and from Type V pneu- mococci to the SI-III phenotype. The phenomenon of binary capsulation, therefore, may be expressed by cells derived from a variety of capsular types when they are endowed with the appropriate genetic factors derived from any of several Type I and mutant Type III strains (Table 1. 1 The enzyme preparation was made available through the generosity of Dr. Alan W. Bernheimer, Department of Microbiology, New York University College of Medicine. 874 PRODUCTION OF CAPSULAR POLYSACCHARIDES. I SIT-ITT cells were observed first in experiments in which strain S—,.;, had been exposed to DNA from a capsulated Type I strain. In quellung prepara- tions, it was noted that the transformed cells showed capsular swelling when exposed either to Type I or to Type III antiserum. The size of the refractile capsular halo in Type TI antiserum was approximately that seen about cells of the normally capsulated Type JII phenotype but the refractile zone about cells in Type I antiserum was distinctly smaller than that about capsulated Type I cells and at times could be seen only with difficulty. Variation in the size of the two capsular components in relation to the phase of bacterial growth was not observed. When the cells manifesting binary capsulation were exposed to a mixture of Type I and Type III antisera, the capsular swelling was not recognizably greater than that which followed contact with Type III antiserum alone. In smears stained by the Gram technique, such cells showed no consistent variation in size or in shape which permitted their distinction from singly capsulated variants. SI-ITI cells isolated from the peritoneal cavity of the white mouse did show elongated forms often in short chains, morphologic charac- teristics not manifested by the organism when grown in vitro. On solid medium, SI-III cells give rise to clones morphologically indis- tinguishable from those formed by pneumococcus Type III, a not surprising finding in view of the significant amount of Type IIT polysaccharide formed by the SI-IIT cell. fevidence Establishing the Existence of the SI-III Cell_—The observation that more than 90 per cent of cells from clones of the SI-III phenotype gave a positive quellung reaction in both Type I and Type III anticapsular serum provided strong presumptive evidence for the existence of a cell producing two capsular polysaccharides or one so constituted chemically as to possess the serologically reactive groups of Type I and of Type III polysaccharide. To distinguish between cells manifesting binary capsulation and mixtures of SI and SIII cells, several types of experiment were performed. Table II shows the results of growing SI, SIII, mixtures of SI and SIII cells and SI-III cells in broth containing either Type I or Type III anticapsular serum. Only those cells producing simultaneously two capsular polysaccharides are agglutinated by both unitypic antisera and yield cultures with limpid supernatant fluids in the presence of each antiserum in so called “thread tests.” Similar evidence distinguishing cells with binary capsules from unitypic cells has been obtained from mouse protection tests. The evidence in Table III indicates that all animals infected with cells with binary capsules were protected by either unitypic capsular antiserum whereas none was protected by either ? Capsular typing sera were made available through the courtesy of Dr. H. D. Piersma, Lederle Laboratories Division, American Cyanamid Company, and Dr. E. F. Roberts, Wyeth Laboratories, American Home Products Corporation. R. AUSTRIAN AND H. P. BERNHEIMER 575 antiserum when infection was carried out with mixtures of Type I and Type Til cells. Evidence for the Production of Two Capsular Polysaccharides by SI-III Cells. — SI-IIJ cells grown by continuous passage in broth containing either Type I or Type III anticapsular serum give rise to mutant cells producing in amounts detectable by the quellung reaction only Type III or Type I capsular poly- saccharide. These mutants derived from SI-III cells suggest that separately variable factors controlling two capsular components are present in the SI-III cell but do not exclude the possibility that each mutant is associated with a qualitative variation in the composition of one of two polysaccharides. A second type of evidence is derived from the exposure of SI-ITI cells to the enzyme which hydrolyzes Type III capsular polysaccharide. After such treat- TABLE II Appearance of Supernatant Fluid of Pneumococcal Cultures Containing Unitypic Anticapsular Serum Serum Pneumococcal type or types in culture Anti-SI Anti-SHI SI SIT SI + SI SI-ITI SI-ITI + SI SI-III + Sil Beta CAA Hy Medium: neopeptone broth plus 10 per cent anticapsular rabbit serum; L = limpid supernatant fluid. T = turbid supernatant fluid. ment, SI-III cells are no longer agglutinable by Type HI anticapsular serum and they fail to give a quellung reaction with Type ITT antibody. Their aggluti- nability by Type I anticapsular serum remains unimpaired, however, and the treated cells give still a positive quellung reaction with Type I antibody. It has been shown that the action of the enzyme hydrolyzing Type II pneumo- coccal polysaccharide is more highly specific than antibody to Type III poly- saccharide. In addition, the action of this enzyme has been demonstrated recently to be that of an endoenzyme rather than that of an exoenzyme attack- ing the end-groups of a macromolecule (12). The experiments suggest strongly, therefore, that the SI-III cell produces normal Type III polysaccharide or a substance resembling it very closely and one that is probably distinct from the Type I component of the capsule. In precipitin reactions carried out in an agar gel by the modified technique of Ouchterlony (10), the precipitates formed when Type III anticapsular 576 PRODUCTION OF CAPSULAR POLYSACCHARIDES. I antibody is allowed to react in the same plate with SI-III cells, SIII cells and Type III polysaccharide give rise to a line of continuity (Fig. 1); and, in like manner, a line of continuity is produced when the reactants are Type I anti- capsular antibody, SI-III cells and Type I polysaccharide (Fig. 2). Precipitin tests performed with partially purified polysaccharides from SI-III cells show that the Type TT polysaccharide can be precipitated by Type ITI anticapsular serum without significant removal of the Type I com- TABLE III Protective Effect of Type I and Type III Antiserum tn Mice Infected with SI-III and with Mixtures of SI and STII Pneumococci Strain Amount of culture injected Serum injected D/S* ce, SLIII 10-6 Anti-SI 0/6 10-6 Anti-SIII 0/6 10-6 Normal 6/0 10-6 None 6/0 10-7 None 5/1 10-8 None 5/1 Viable units/cc. (undiluted culture): 4 x 108 Strains Amount of culture injected Serum injected D/S* cc. SI + SHI 10-6 Anti-SI 6/0 10-6 Anti-SIII 6/0 10-6 Norma] 6/0 10-8 None 6/0 107 None 6/0 10-3 None 4/2 Viable units/cc. (undiluted culture): 3 & 108 (equal numbers of SI + SII) *D/S = died/survived, ponent of the capsule and conversely, specific precipitation of the Type I capsular component is not accompanied by removal of the Type III component from solution. A 200 cc. culture of SI-HI cells was grown overnight after which time glucose was added to a final concentration of 1 per cent. During an additional 6 hours’ incubation, the lactic acid formed was neutralized by the addition of 3 x NaOH. The cells were then removed by centrifugation and 1 volume of 95 per cent ethy! alcohol was added to the separated super- natant fluid. The suspension was stored overnight at 4°C. after which time, the precipitate formed was collected by centrifugation and dissolved in 10 cc. of normal salt solution. This solution served as the unfractionated capsular antigen. Aliquots thereof were allowed to R. AUSTRIAN AND H, P, BERNHEIMER 577 react with appropriate amounts of Type I or Type I! anticapsular rabbit serum or with normal rabbit serum. All volumes were adjusted to give the same quantity of the original antigen in precipitin tests. The results are shown in Table IV. On the bases of the independent variability of its capsular components, of the susceptibility of the Type III capsular component to the enzyme hydro- TABLE IV Independent Precipitability of Each Capsular Component of the SI-III Phenotype by Unity pic Capsular Antiserum Polysaccharide precipitated after treatment with normal rabbit serum Dilution of antigen | Serum 1:32 | 1:64 12128) 1: 236 1:512/1: a 2048/1: 4096/1:8192, C AntiSt | +++) +4 +4 + 0 fo 0 | 0 anism Loree] eee fat | ek | | 0 | 0 Normal 0 | 0 0 | 0 | Polysaccharide precipitated after treatment with type I anticapsular rabbit serum Dilution of antigen Serum 1:32 | 1:64 | 12128 1:256 1:512)1:1024/1:2048|1:4096 1:8192 Cc Anti-SI tk | 0 | 0 | 0 0 0 0 0 0 0 Anti-SI |++++) +++ tt ++) + += | 0 | 0 | 0 0 , Normal 0 | Polysaccharide precipitated after treatment with type III anticapsular rabbit serum Dilution of antigen Serum 1:32 | 1:64 j tens 1:256 1:512{1:102411:204811:409611:8192 Cc | Anti-SI ++ + + + 0 0 0 0 0 0 Anti-SHI 0 0 0 0 0 Q 0 0 0 0 Normal 0 lyzing Type III pneumococcal polysaccharide, of the serologic cross-reactivity of its cells with Type I and Type III polysaccharides in precipitin tests in agar gels, and of the independent precipitability of its two capsular components, SI-III cells may be considered to produce two capsular polysaccharides. Phagocytosis of SI-III Cells by Human Polymorphonuclear Leukocytes.—In Table V are shown the results of exposing SI-IIT cells to human polymorpho- 578 PRODUCTION OF CAPSULAR POLYSACCHARIDES. I nuclear leukocytes in a smooth-walled system. Several facts of interest are demonstrable in these experiments. First, the persistence of the antiphagocytic effect of the Type III capsular component is demonstrable even in the presence of Type I anticapsular serum. That the Type I antiserum is fully effective is shown in the last line of the table. Second, the relatively small antiphagocytic effect of the Type I capsular component becomes apparent when the SI-III cells are exposed to the enzyme depolymerizing Type III polysaccharide to- gether with normal rabbit serum. The efficacy of Type I antibody is definitely demonstrable in the presence of the same enzyme, however, by a further in- crease in phagocytosis. The Relative Importance of the Two Capsular Components to the Virulence of SI-ITT Pneumococci.—-Table VI shows the effect of several reagents upon the TABLE V Phagocytosis of SI-III Pneumococci by Human Polymorphonuclear Leukocytes in the Presence of Various Sera and of the Ensyme Hydrolysing Type III Polysaccharide Strain Serum Enzyme (Rgf gent PMN con. Average No.of prt SI-III Anti-SI Absent 8 0.36 Anti-SIII 7 100 10.0 Normal « 4 0.08 Anti-SI Present 96 8.2 Anti-SIII “ 96 8.0 Normal “ 32 0.92 SI Anti-SI Absent 88 8.0 outcome of infection with SI-ITE cells in mice. It may be seen that each capsular polysaccharide contributes to the virulence of the infecting organism. Of interest is the role played by the Type I capsular component. Although the number of animals infected is small, it may be noted that the only animal infected with pneumococci manifesting binary capsulation and not receiving antiserum to survive was one treated with the enzyme depolymerizing Type IIT polysaccharide. Not shown in the table but noteworthy also is the fact that animals treated with a single injection of enzyme at the time of infection survived twice as long after inoculation as did untreated controls. Similar results have been observed in analogous experiments performed with cells of the binary capsular SIII-V phenotype. These observations are in accord with those derived from the study of phagocytosis of these strains. They suggest that the smaller Type I capsular component is able to protect the population of SI-III cells from complete elimination during the period that the enzyme is still hydrolyzing effectively the Type III capsular component but is unable by R. AUSTRIAN AND H. P. BERNHEIMER 579 itself to shift the balance of infection decisively in favor of the pneumococcus. Later, when the enzyme is no longer effective, regeneration of the Type III capsular component provides more effective protection against phagocytosis and permits rapid progression of the infection to a fatal outcome. Properties of Antiserum Prepared with Vaccines of SI-IIT Pnewmococct.— Rabbits immunized with vaccines of pneumococci manifesting binary capsu- lation form antibodies to each of the capsular components of such cells. The relative amount of antibody to each of the capsular antigens may vary during the course of immunization. Antisera to SI-III cells cause capsular swelling of SI, SILI, SI-IIT, and SIII-V cells. Shown in Table VII is the protective effect TABLE VI Effect of the Enzyme Hydrolyzing Type III Polysaccharide and of Unitypic Antisera on the Virulence of SI-III Pneumococci Strain Amoung of culture Serum injected Enzyme injected D/S ce. mg. SLIT 10-* None 2 3/1 10-® Anti-SIII 2 0/4 10-6 Anti-SIIT None 0/4 10-6 Anti-SI “ 0/4 40-6 None “ 3/0 10-7 “cc “ 2/0 SI 10-& “ 2 4/0 STI 10-& “ 2 0/4 10-8 Anti-STII None 0/4 Viable units/cc. (undiluted culture): SI-III = 2 x 10°, SI = 2 X 108; SIIT = 3 x 108. of such antiserum in mice infected with cells of each of the aforementioned phenotypes and with mixtures of SI and SII1 cells. Recorded also in the table is the absence of protection against infection with pneumococcus Type V. Ana- logous results are observed in experiments concerned with the phagocytosis of pneumococci of several capsular types in the presence of Type I-III antiserum (Table VITI). Precipitin tests in an agar gel (Fig. 3) disclose several facts. First, with Type I cells and Type I capsular polysaccharide, the anti-SI-III serum forms a continuous band of precipitate, and with Type III cells and Type III capsular polysaccharide a second and distinct band is formed. Second, these lines of precipitate are continuous with those formed by the antibody reacting with each of the capsular components of the SI-III cell. Third, there is no band of precipitate formed to indicate the presence of a single species of antibody TABLE VII Protective Effect of Type I-III Antiserum in Mice Infected with Pneumococci of a Variety of Capsular Ty pes Strain or strains Amount of cpiture Serum injected D/S ce, SI-III 10-6 Anti-SL-III 0/5 SI 10-8 “ 0/5 SIII 10-8 “ 0/5 SI + SII 1o-8 ‘ 0/5 SIIL-V 10-8 “ 0/5 SV 10-6 “ 5/0 SII + SV 10-6 “ 5/0 SI-Ii1 10-6 Normal 5/0 SI 10-6 “ 5/0 SIII io-6 “ 5/0 SI + SIT 10-6 “ 5/0 SIiI-V 10-® “ 5/0 SV 10-& “ 5/0 SIII + SV 10-6 “ 5/0 SL-TiI 10-7 None 2/0 SI 1077 “ 2/0 STII 10-7 “ 2/0 SI + SII 10-7 “ 2/0 SII-V 107 “ 3/0 SV 10-77 “ 2/0 SIII + SV 10-7 “ 2/0 Viable units/cc. (undiluted culture): SI-III = 5 & 108; SI = 3 x 108; SITI = 2 x 108; SI + SITI = 2 x 108; SIII-V = 2 x 10°; SV = 2 x 108; SIII + SV = 2 x 103. TABLE VIII Phagocytosis of Pneumococci of Several Types by Human Polymorphonuclear Leukocytes in the Presence of Type I-III Anticapsular Rabbit Serum * MN ini No.” Strain Serum PDneumococel = | preumocoee! ped PMN per cent SI-III Anti-SLIII 84+ 8.6 SI “ 60 9.4 SIH “ 92 8.1 SUI-V “ 72 7.4 SV “ 0 0 SLIT Normal 0 0 SI “ 0 0 SHI “ 0 0 SUI-V \ “ 0 0 SV | “ 0 0 580 R. AUSTRIAN AND H. P. BERNHEIMER 581 reacting with both capsular components or the presence of a capsular com- ponent reacting with antibody in a fashion different from that of unitypic cells. The results are consistent with those presented earlier which indicate that cells manifesting binary capsulation produce two species of capsular polysaccharide each of which is closely similar if not identical with that of the corresponding unitypic pneumococcus. In precipitin tests performed in tubes, Type I polysaccharide and Type III polysaccharide are each precipitated by antiserum prepared with a vaccine of TABLE IX Independent Precipitability of Capsular Antibodies from an Antiserum Prepared with a Vaccine of SI-III Pneumococet Polysaccharide precipitated by anti-SI-III serum before absorption Dilution of antigen ! | 1 Antigen* tal ttl 1:32, 1:64! 12128! 1:26 1512) 1:1024 } 1:2048 | 1:4096 | 1:8192 | C sI decledinn +44) 44) 4+ + | foot a} 0 sit [etattthebettttte) bed] te) He] oF] 0 0} 0 Polysaccharide precipitated by anti-SI-III serum after absorption with SI cells Dilution of antigen Antigen 1:8. 1d 1:32 : 1:64! 1128! 1254 1: sn 1:1024 | 1:2048 | 1:4096 | 1:8192 | C i | | ' —_ SI af ol 0, 0 0. 0 ol 0; 0 0 o| 0 SHI setae esl tH+t t+) +4] + + ++ | +| +} of 0 Polysaccharide precipitated by anti-SI-III serum after absorption with SIIT cells Dilution of antigen Antigen 18 1: tal bs? i: a 1:123 1:256 6 1: st 1:1024 | 1:2048 | 1:4096 | 1:8192 | C | — SI +t+4+++ tibet Salen + | + | 0 0} 0 SII a 0 5 60] 0 0) 0 i i * Saline solution of purified Type I or Type IIT ate polysaccharide in a concentration of 1 mg./cc. diluted as indicated. SLIII cells. That two distinct anticapsular antibodies are present in such an antiserum can be demonstrated by the independent precipitability of each. When an antiserum prepared with a vaccine of SI-III cells is absorbed with either capsulated Type I or Type III cells, the corresponding anticapsular antibody is removed without significant alteration of the titer of antibody precipitating the other capsular component (Table TX). Analogous experiments in which precipitin reactions were carried out in an agar gel demonstrate the independent removal of one capsular antibody without alteration in the reaction of the other with appropriate antigens (Figs. 4 to 6). 582 PRODUCTION OF CAPSULAR POLYSACCHARIDES. I Properties of Deoxyribonucleates of Pneumococct of the SI-III Phenotype.—To determine the nature of the genetic determinants of the capsular components of the SI-ITT cell, DNA was prepared from cultures of such cells in the usual manner. When cells of a non-capsulated pneumococcus (R36NC) derived originally from a strain of capsular Type II were exposed in the transforming system to DNA from SJ-III cells, organisms of three different phenotypes were recovered: S— yz, SI and SI-III. No SIII cells were observed. The results indicate that there has been no major alteration in either of the genetic units concerned with the formation of the two capsular components of the cell mani- festing binary capsulation during their residence within that cell and that each is able to engender expression of that character controlled by it when it is reintroduced as the sole genetic factor concerned with capsular formation into another pneumococcal cell. DISCUSSION Through the agency of the transformation reaction, it has been possible to isolate pneumococci manifesting binary capsulation of several phenotypes. Examination of cells of the SJ-ITI phenotype in a number of experimental systems reveals that organisms of this type produce two distinct capsular polysaccharides which are closely similar if not identical to those produced by cells of the corresponding unitypic phenotypes. The evidence available suggests that the two types of polysaccharide are intermingled in the capsule in a random fashion. Study of cells of the SIII-V phenotype yields results of a similar nature. Binary capsulation in pneumococcus provides a new basis for serologic cross-reactivity in this species. Cells manifesting binary capsulation react with unitypic antiserum to each of their capsular components and stimulate the formation of antibodies which will react with cells bearing either of the capsular antigens produced by cells of the binary capsular phenotype. Whether or not pneumococci manifesting binary capsulation exist in nature remains an un- answered question. Examination by genetic and serologic techniques of several groups of pneumococcal strains the capsular antigens of which show cross- reactivity has failed to result in the recognition in any of the strains of separable capsular antigens. It would appear more likely that the serologic interrelation- ships of these strains are based upon the presence of common chemical sub- groups in their single capsular polysaccharides rather than upon their production of multiple types of capsular polysaccharide. For this reason the nomenclature of pneumococcal types proposed by Eddy (13) seems preferable to that proposed by Kauffmann and Lund (14). These experiments with pneumococcus show many similarities to those performed with H. influenzae by Leidy, Hahn, and Alexander (6), and it ap- pears probable that binary capsulation may be manifested by organisms of the R. AUSTRIAN AND H. P, BERNHEIMER 583 latter species. In each, however, it would appear to represent a product of the laboratory. This fact notwithstanding, binary capsulation provides a useful object for the study of a variety of biologic phenomena. SUMMARY Pneumococci manifesting binary capsulation have been produced through the agency of transformation reactions. Such cells produce two capsular poly- saccharides which are closely similar if not identical with those produced by corresponding singly capsulated strains. Binary capsulation in pneumococcus provides a new basis for serologic cross-reactivity in this species. BIBLIOGRAPHY 1. Griffith, F., The significance of pneumococcal types, J. Hyg., 1928, 27, 113. 2. Sugg, J. Y., Gaspari, E. L., Fleming, W. L., and Neill, J. M., Studies on im- munological relationships among the pneumococci. I. A virulent strain of pneumococcus which is immunologically related to, but not identical with typi- cal strains of Type IIL pneumococci, J. Exp. Med., 1928, 47, 917. 3. Harris, A. L., Sugg, J. ¥. and Neill, J. M., Studies on immunological relation- ships among the pneumococci. II. A comparison of the antibody responses of mice and of rabbits to immunization with typical Type HI pneumococci and to immunization with a related strain, J. Exp. Med., 1928, 47, 933. 4. Goebel, W. F., Chemo-immunological studies on the soluble specific substance of pneumococcus. IJ. The chemical basis for the immunological relationship between the capsular polysaccharides of Types III and VIII pneumococcus, J. Biol. Chem., 1935, 110, 391. 5. Merch, E., Serological studies on the pneumococci, London, Humphrey Mil- ford, Oxford University Press, 1943. 6. Leidy, G., Hahn, E., and Alexander, H. E., In vitro production of new types of Hemophilus influenzae, J. Exp. Med., 1953, 97, 467. 7. Taylor, H. E., Additive effects of transforming agents from some variants of pneumococcus, J. Exp. Med., 1949, 89, 399. 8. Ephrussi-Taylor, H., Transformations allogénes du pneumocoque, Exp. Cell Research, 1951, 2, 589. 9, MacLeod, C. M., and Krauss, M. R., Stepwise intratype transformation of pneumococcus from R to S by way of a variant intermediate in capsular poly- saccharide production, J. Exp. Med., 1947, 86, 439. 10, Halbert, S. P., Swick, L., and Sonn, C., The use of precipitin analysis in agar for the study of human streptococcal infections. II. Ouchterlony and Oakley technics, J. Exp. AMfed., 1955, 101, 557. 11. Austrian, R., Morphologic variation in pneumococcus. I. An analysis of the bases for morphologic variation in pneumococcus and description of a hitherto undefined morphologic variant, J. Exp. Med., 1953, 98, 21. 12. Torriani, A. M., personal communication. 584 PRODUCTION OF CAPSULAR POLYSACCHARIDES, I 13. Eddy, B. E., Nomenclature of pneumococcic types, Pub. Health Rep., 1944, 59, 449, 14. Kauffmann, F., and Lund, E., Memorandum on the nomenclature of the pneu- mococcus group, Juternat. Bull. Bact. Nomenclature and Taxonomy, 1954, 4,125. EXPLANATION OF PLATES Pirate 32 Fic, 1. Reaction of Type HI anticapsular antibody (A) with Type IIT cells (B), Type L-Ill cells (C), and Type Uf capsular polysaccharide (D). Fic. 2. Reaction of Type I anticapsular antibody (A) with Type I cells (B), Type LISI cells (C), and Type I capsular polysaccharide (D). THE JOURNAL OF EXPERIMENTAL MEDICINE vot. 110 PLATE 32 (Austrian and Bernheimer: Production of capsular polysaccharides. I) Pilate 33 Fic. 3. Reaction of Type LIL anticapsular serum (A) with Type LIIL cells 03), Type L cells (C), Type | polysaccharide (D), Type IL cells (E) and Type If poly- saccharide (F). 2 THE JOURNAL OF EXPERIMENTAL MEDICINE VoL. 110 PLATE 33 (Austrian and Bernheimer: Production of capsular polysaccharides. 1) PLate 34 Vic. 4. Reaction ef Type I-IT] anticapsular serum absorbed with Type IL pneu- mococci (A), with Type LHI cells (B), Type ET cells (C), Type I polysaccharide (D), Type Hf cells (FE) and Type HI polysaccharide (F). THE JOURNAL OF EXPERIMENTAL MEDICINE VoL. 110 PLATE 3+ (Austrian and Bernheimer: Production of capsular polysaccharides. 1) PLATE 35 Fic. 5. Reaction of Type LIIT anticapsular serum absorbed with Type I pneu- mococei (A), with Type I-1Lf cells (B), Type [cells (C), Pype | polysaccharide (2), Type LL cells (2), and Type ILL polysaccharide (F). Pic. 6. Reaction of Type LILI cells (A) with Type 1-Iif antiserum (B), Type I-L antiserum absorbed with Type 1 pneumococci (C), Type I antiserum (D2), Type 1-411] serum absorbed with Type lI pneumococci CE), and Type T antiserum (IF). Reactions between CA) and (F) failed to develop sufficiently to be visible in the photo- graph. THE JOURNAL OF EXPERIMENTAL MEDICINE VoL. 110 PLATE 35 (Austrian and Bernheimer: Production of capsular polysaccharides. 1}