ABSTRACT THE PRECIPITINOGENIC RELATIONSHIPS AMONG MYCOBACTERIA OF BOVINE, PORCINE, HUMAN, AND SOIL ORIGINS by Calvin Armstrong Davenport The precipitinogenic relationships of 49 representative mycobacteria of human, bovine, porcine, and soil origins were compared by use of an agar slide immunodiffusion technique. ' The organisms studied included the classical pathogens and sap- rophytes, and "atypicals" of human, animal, and soil origins. Antigenic differences exist among species and groups, and in addition, among strains within a group. No one precipitin- ogen was found to be common to all mycobacterial strains. From two to six precipitinogens were detected in homol- ogous'antigen-antibody systems, and in all tests, the homol- ogous systems reacted to give equal or greater numbers of precipitate bands than the heterologous systems. The Group III strains of bovine and porcine origins were closely related but not identical to M, avium. A new grouping of "atypicals, based on antigenic relationships to known mycobacterial species, is presented. THE PRECIPITINOGENIC RELATIONSHIPS AMONG MYCOBACTERIA OF BOVINE, PORCINE HUMAN, AND SOIL ORIGINS BY Calvin Armstrong Davenport A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Microbiology and Public Health 1963 ACKNOWLEDGEMENTS The author wishes to express his sincere appreciation and thanks to Drs. V. H. and W. L. Mallmann for their con- tinued interest and guidance throughout this study. Thanks also, to Dr. James Ray, Mr. Robert Walker and all members of the Bovine Tuberculosis Research Project at Michigan State University for their cooperation in the use (of limited space and equipment. Gratitude is extended to the U. S. Department of Agri— culture who provided the grant that made this study possible. ii TABLE OF CONTENTS Page INTRODUCTION . . . . . . . . . . . . . . . . . . . . 1 LITERATURE REVIEW . . . . . . . . . . . . . . . . . 2 PreCipitin Reaction . . . . . . . . . . . . . . 2 Influence of Physico-chemical Factors On Precipitation . . . . . . . . . . . . . . . 5 Immunodiffusion . . . . . . . . . . . . . . . . 7 Studies of Mycobacteria . . . . . . . . . . . . 12 Need for Identification and Classification 12 Applications of Immunodiffusion to the Study of Mycobacteria 17 MATERIALS AND METHODS . . . . . . . . . . . . . . . 25 Cultures . . . . . . . . . . . . . . . . . . . . 25 Antigens . . . . . . . . . . . . . . . . . . . . 31 Antisera . . . . . . . . . . . . . . . . . . . . 33 Electrolytes Used for Preparing Diffusion Media 34 Media for Diffusion Studies . . . . . . . . . . 36 Gels for Preliminary Studies by the Tube and Petri Dish Methods 36 Gels for Preliminary Studies by the Agar Slide Method 37 Immunodiffusion Test Procedures . . . . . . . . 37 Tube Method 38 Petri Dish Method 39 Agar Slide Method 40 iii The Concentration of the Reactants Staining of PrecipitatiOn Lines Method for Calculating Antigenic Relatedness of Reference Strains . RESULTS AND DISCUSSION OF PRELIMINARY OBSERVATIONS RESULTS OF STUDIES OF PRECIPITINOGENIC RELATIONSHIPS Homologous Antigen-Antibody Reactions Precipitinogenic Relationships of the 12 Reference Strains Precipitinogenic Relationships of Mycobacteria of Human, and Soil Origins Group II Bovine, Group III Group IV Pseudochromes Bovine—skin Isolants A Comparison of the Relationships of "Atypicals" of A11 Runyon Groups DISCUSSION . SUMMARY . . BIBLIOGRAPHY Porcine, Presented by Groups . iv Page 41 42 43 46 54 54 57 68 68 68 72 75 77 79 84 94 96 LIST OF TABLES Table Page 1 Atypical strains from which culture filtrate antigens were prepared . . . . . . 26 2 Known laboratory species from which culture filtrate antigens were prepared . . 28 3 Organisms used for production of antibodies . 29 4 The rmmber of precipitation lines detectable in double-diffusion tests conducted by the tube, petri dish, and agar slide methods . 47 5 The number of precipitation lines detectable in various diffusion media in tests conducted by the agar slide method . . . . 49 6 The number of precipitins detected in individual rabbit antiserum by homologous antigen-antibody reactions . . . . . . . . 52 7 The number of precipitation lines formed by homologous and heterologous antigen- antibody reactions . . . . . . . . . . . . 56 8 The antigenic "relatedness" of selected mycobacteria reported in per cent of their total antigens that are related . . . 66 9 The number of precipitation lines formed by reactions of culture filtrates of Group II strains in reactions with reference antisera . . . . . . . . . . . . 67 10 The number of precipitation lines formed by culture filtrates of Group III strains in reactions with reference antisera . . . . . 70 11 The number of precipitatiOn lines formed by culture filtrates of Group IV strains in reactions with reference antisera . . . . . 74 12 The number of precipitation lines formed by culture filtrates of pseudochromes in reactions with reference antisera . . . . . 76 Table Page 13 The number of precipitation lines formed by culture filtrates prepared from various isolants of bovine skin-lesions in reactions with reference antisera . . . 78 14 The number of strains within each Runyon Group that possessed one or more precipitating antigens which were related to antigens found in 12 reference strains . . . . . . . . . . . . . 80 15 A regrouping of atypical strains based on their antigenic relationships among known mycobacterial species . . . . . . . . 82 vi INTRODUCTION Many unsuccessful approaches have been made to the prob- lem of differentiating mycobacterial species. Primarily, the mycobacteria have been classified according to their morphologic characteristics and virulence for experimental animals; such classification systems have numerous short- comings. The many strains and variants of acid-fast bacilli constantly being isolated from various sources and the con- sequent difficulties in attempting to group them emphasize the need for new criteria for their identification. Recent advanaces in the interpretation of the immuno- diffusion tests offer an excellent means of studying the precipitinogenic relationships of acid-fast bacilli. Suf- ficient information on these relationships may provide the foundation for a reliable system of identification and clas- sification of these organisms. Moreover, some adaptation of gel diffusion tests may eventually serve as an aid for detecting mycobacteriosis in infected individuals. In this thesis are presented observed precipitinogenic similarities and dissimilarities of acid-fast bacilli iso- lated from man, cattle, swine, and soil, as interpreted from the results of slide immunodiffusion tests. LITERATURE REVIEW The Precipitin Reaction Kraus (1897) observed that a precipitate formed when the sera of animals immunized with typhoid or cholera bacilli were mixed with filtrates of broth cultures of the homol- ogous organism. This marked the discovery of the precipitin reaction and led to the further observation that most proteins and many polysaccharides may elicit antibody responses when injected into suitable animals. Under proper physico-chemical conditions, antigen and antibody react to form a precipitate in two distinct stages (Crowle, 1961). The first stage is a rapid and invisible combination of antigen and antibody molecules to form a spe— cific but readily reversible complex. The second or aggre- gation stage develops slowly and marks the initial formation of the visible precipitate. The antigen-antibody complexes, or lattices of these complexes, become too large to remain either invisible or soluble (Crowle, 1961). The first stage of antigen—antibody complexing results from an attraction of the two reactants due to one or sev- eral of five types of forces (Crowle, 1961): (l) attrac- tion by van der Waals forces, (2) dipole-dipole attraction, (3) dipole-ion attraction, (4) positive—negative chemical group attraction, and (5) hydrogen bonding. Visible pre- cipitates result from combinations of certain groups on a given reacting molecule. If the molecule is disrupted into relatively complex fractions, each possessing two or more reactive sites, visible precipitation may still occur. If the molecule is disrupted into simple fragments possessing no more than one reactive site, combination may be possible but visible aggregation will not occur. Fragments will not be able to form the lattices necessary for developing vis- ible precipitates. Animals may produce "incomplete" or "blocking" antibodies, which, like the simple fractions, combine with antigens but fail to cause aggregflfion (Crowle, 1961). Combinations of reactant sites without aggregation, termed "incomplete pre- cipitation" (Crowle, 1961), may interfere with specific pre— cipitation by deviating one of the reactants in a mutual reaction area (Oudin, 1948). Recent studies (Ouchterlony, 1958) have suggested that in most precipitation tests both antibody and antigen are impure. Untreated antigenic material such as a culture filtrate is usually a mixture of molecules carrying various groups. After injection into an animal, each determinant group may induce the formation of corresponding antibodies. If a culture filtrate is fractionated, the fractions obtained may be chemically pure but not neccessarily immu- nologically pure. If the individual particles of a specific fraction carried only one kind of reactive site, the frac- tion would be immunologically pure. But since each particle is a molecule of considerable size and complicated structure, other assumptions must be made: certain particles may carry several serologically related groups, some particles may carry several serologically non-related groups, and other particles may carry both related and non-related determinant groups. Furthermore, these groups may be in a hidden, par- tially hidden, or an exposed position on the particle. During immunization, antibodies may be produced that correspond to each type of originally exposed determinant, or to one which became exposed during the immunization. The degree of anti- body response will depend on the predominance, exposure, and antigenicity of the determinant groups (Ouchterlony, 1958). Studies (Ouchterlony, 1958; Oudin, 1952) of precipita- tion systems have established this classification: (1) simple, in which a single antigen reacts with an antibody; (2) complex, in which several antigens cross-react with a given antibody; and (3) multiple, in which several antigens and several antibodies react simultaneously. The Influence of Physico—chemical Factor§_on Precipitation Antigen-antibody combinations can occur in the absence of electrolytes, but the rate and intensity of precipitation is enhanced by their presence (Aladjem and Lieberman, 1952). The optimal range of electrolyte concentrations for precip- itation tests varies and should be determined for each antigen-antibody system studied. Studies have shown that in liquid media, horse antitoxin and diptheria toxin (Boyd, 1956), precipitate best at a NaCl concentration of between 0.05 and 0.25 M. Chicken precipitins require 1.5 M NaCl for 1., 1957). optimal precipitation in liquid media (Goodman §t_ Yet, the same chicken sera precipitate the same antigens much better at 0.85 M NaCl in gel media whether these are con— ducted in agar, gelatin, or cellulose acetate (Crowle, 1961). Minute quantities of the salts of cadmium and lanthanum have been found to enhance or inhibit specific precipitation (Crowle, 1958c). The optimal temperature for antigen-antibody combination varies from system to system. Boyd (1956) reported that incubation temperatures of 15 C to 40 C permitted optimal precipitation for most antisera obtained from warm—blooded animals, and that 10 C was best for antibodies produced in the frog. Rheins _t_§1, (1956) observed that certain spe- cific precipitates of antigens and antibodies of rabbit origin that formed in a gel medium at 4 C disappeared on warming to room temperature. Precipitation tests in semi- solid media must be conducted at a constant incubation tem- perature to prevent formation of artifacts resembling pre- cipitate bands (Oudin, 1946, 1948; Crowle, 1961). Studies by Grabar (1959) suggest that the ideal pH for a particular precipitation system is within the range of 6.5 to 8.2. At a pH lower than 6.5 nonspecific precipita- tion of serum proteins often occurred, and at a pH greater than 8.2 specific precipitation was hindered and certain antigen-antibody complexes tended to dissociate. Other factors affecting specific precipitation are the presence of lipids in the reactants, the presence of pro- teins foreign to the desired reaction, and the relative proportions of antigens and antibodies (Crowle, 1961). Max- imum precipitation in minimum time results when both react- ants are used in optimal proportions. It was reported (Oudin, 1948) and confirmed (Becker and Munoz, 1949; Munoz and Becker, 1950) that there is a straight-line relationship between the movement of a pre- cipitate band and the square root of the time. This move- ment is also influenced by the size, shape, and molecular weight of the particles (Crowle, 1961); the initial con— centrations of reactants (Munoz and Becker, 1950; Oudin, 1948; Ouchterlony, 1958); and the gravitational field (Crowle, 1961). Immunodiffusion Immunodiffusion, a term which replaced the phrase "agar diffusion precipitin test," refers to serologic tests con- ducted in or on semi-solid media (Crowle, 1961). The first use of an immunodiffusion technique was made by Bechhold (1905). Anti—goat serum of rabbit origin was mixed With 1% gelatin in test tubes. After the mixture was gelled, goat serum was poured onto the gel. Precipitates were observed that differed from the types that appear from inorganic substances. Two heavy, distinct precipitates were described but the possibility that these could have been caused by separate antigen—antibody systems was not mentioned. Oudin (1946) made the interpretations that established immunodiffusion in its present form as an analytic technique for mixed antigen-antibody systems. Employing a method of simple diffusion, he placed an antigen solution over a solid immune serum-agar mixture that was contained in a thin—bore tube. An antigen-antibody precipitate band formed and ap- peared to migrate down the tube as more antigen diffused into the antibody-agar mixture. At equivalent concentrations of reactants, one band formed when a single antigen-antibody system was present, and multiple bands formed when more than one antigen-antibody system was present. Oudin's studies (1948, 1952) led to these observations: (1) positive correlation exists between precipitate inten- sity and antibody concentration; (2) sudden changes in in- cubation temperature may cause the appearance of artifacts which usually have characteristics that differentiate them from primary precipitates; and (3) cross—reacting antigen can influence the formation of precipitate by the homologous antigen. The Oudin tube technique has been applied to the study of many antigen-antibody systems. Becker and Munoz (1949) used this method to study the multiplicity of antigens in ragweed pollen extracts, and Telfer (1953) used it to study the antigenic differences existing between the larval, pupal, and adult forms of the Cecropia moth. Shortly after the application of simple or single-dif— fusion to antigenic studies, techniques of double-diffusion in two dimensions were developed independently by Ouchterlony (1948) and Elek (1948). Ouchterlony, studying the toxin- producing ability of various strains of Corynebacterium diphtheriae poured 1% agar prepared with saline into petri dishes and allowed it to harden. wells were cut in the agar and filled with antigen and antibody. Both antigen and anti- body diffused into a common reactant area to form precipitate bands. The gel technique of Elek (1948) has been called the "Double-diffusion Gradient Test." It differs from Ouchter- lony's method in that no wells were cut in the agar. Two filter paper strips impregnated with the reactants were placed on the agar at right angles to each other to form an L. Chen and Meyer (1955) utilized this technique for study- ing antigens of Pasteurella pestis and Pasteurella paratuber— culosis. No antigen specific for P. pestis was detected, but two antigen—antibody bands were common to P. pestis and g, paratuberculogis, indicating that two identical or closely related antigens were present in both species. Modifications have increased the usefulness of immunodiffusion techniques. The inhibition plate technique (Bjorklund, 1952) made use of the principle that pretreating the diffusion medium with 10 a sufficient amount of a component from a complex immuno- logic system completely inhibited the subsequent appearance in the medium of precipitates corresponding to this partic— ular component. Furthermore, the pretreatment of the agar did not interfere with the diffusion and precipitation of other components in the system. In experiments designed to compensate for the large volumes of reactants required by Ouchterlony's double-dif— fusion plate technique, Preer (1956) developed micro-gel dif- fusion methods. The use of modified micro-immunodiffusion methods by Mansi (1958) and Crowle (1958a, 1958b, 1960) have proved that micro—techniques are of greater sensitivity than the agar plate method. Immunoelectrophoresis, a combination of agar-diffusion with electrophoresis, has increased resolution of precipitate bands formed by complex antigen-antibody systems in agar media (Crowle, 1961; Grabar, 1959; Grabar and Williams, 1955). Techniques for drying and staining improved the definition of bandsilehotographs made for permanent records (Crowle, 1961). Immunodiffusion tests are useful in resolving multiple precipitating systems into their individual components, and in comparing two antisera against the same antigen or two ll antigens against the same antiserum. The resolution of multiple precipitating systems into individual components is based on the assumption that specific precipitates in semisolid media permit the diffusion of unrelated antigens and antibodies. The number of bands developing in a mul— tiple precipitating system represents the minimum number of antigen-antibody systems present, not the maximum number. A band formed by one system may mask other bands (Becker, 1953; Ouchterlony, 1958; Oudin, 1948). When two identical antigens are compared by immunodiffusion tests on a single plane, the precipitation lines developing from individual components fuse to form a continuous band called the "Re— action of Identity." Diffusion of two serologically related antigens against an antiserum containing specific antibodies results in a precipitation pattern consisting of a contin- uous arc with a spur extending above the precipitation line and is termed "Reaction of Partial Identity." The compar- ison of two serologically unrelated antigens with an anti- serum containing homologous antibodies results in the forma— tion of precipitation lines that cross each other. This pattern is termed "Reaction of NOn-Identity" (Ouchterlony, 1958). 12 Studies of Mycobacteria The Need for Identification and Classification. Four- teen species of the genus Mycobacterium are listed in the Seventh Edition of Bergey's Manual for Determinative Bac- teriology (Breed §t__l,, 1957): M, phlei, M, smeggatis, l3 . fortuitum, M, marinum, M, thamnopheos, M, platypoecilus, l3 . ulcerans, M, tuberculosis, M, bovis, M, microti, M, avium, l3 . paratuberculosig, M, leprae, and M, lepraemurium, These species are differentiated primarily on their virulence for laboratory animals and their ability to grow on non-living media at various temperatures. The existence of other strains and variants of the genus Mycobggterium has been established (Mallmann, Mallmann, and Robinson, 1961; Parlett and Youmans, 1956, 1958; Runyon, 1960; Xalabarder, 1961), yet the significance of these atyp- ical organisms remains controversial. Some "atypicals" are pathogens; others may be mere saprophytes. Studies by Runyon (1960) of "atypicals" of human origin led to a classification system based primarily on the rate of growth and the presence or absence of pigmentation when the organisms were grown in the light or dark. Runyon's group designations and characterizations are: Photochromogens 13 (Group I) develop a yellow pigment after a short exposure to light; Scotochromogens (Group II) develop a yellow or orange pigment when grown in the dark which deepens when grown in the light; NOn-photochromogens (Group III) and Rapid-growers (Group IV) do not produce pigment under any growth conditions, but are differentiated by variations in their growth rates. Organisms in Groups I, II, and III require one to two weeks for visible growth and organisms in Group IV need only two to four days. Runyon (1960) reported that the tendency of the "atyp- icals" to form strands or cords was proportional to the roughness of the cultures. All Group I strains and Group IV rough strains formed cords, but strains of Groups II and III rarely exhibited cording. Runyon‘s virulence studies (1960) indicated that "atyp- icals" of human origin do not produce progressive disease in guinea pigs, rabbits, or birds. In mice, strains of Group I and some strains of Groups III and IV produced lesions, and sometimes death. On the basis of the frequency with which an organism was isolated from a patient, from resected lung tissue and from sputum, and the absence of other disease agents, Runyon (1960) concluded: Group I strains, with few exceptions, are 14 agents of disease in man; Group II strains are generally non— pathogenic in man; and strains of Groups III and IV are some— times pathogenic for man. Similar problems prevail in the field of animal tuber— culosis (Johnson, Balsden, and Frank, 1961; Mallmann, Mall- mann, and Robinson, 1961; Ranney, 1961). Mallmann §t_§l, (1961) reported the isolation of large numbers of acid-fast bacilli from bovine and porcine tissues that differ from the classical pathogens. Studies of these "atypicals" from the standpoint of their growth and cytochemical characteristics, and their ability to sensitize and infect animals, substan- tiated earlier observations that one set of characteristics is insufficient for the identification of any one mycobac- terium. Furthermore, compilations ofthe characteristics did not always establish conclusive identification. The use of the Runyon classification system was not adequate for dif- ferentiating "atypicals" from animals. Mallmann _§ ML. (1961) isolated a group of organisms that was intermediate between Group I and Group II in its pigment response to light ex- posure. Other tests indicated that this group, designated as pseudochromes, was most closely related to Runyon Group III, and may be more accurately designated as pigmented Group III. "Atypicals" of bovines and porcines were heterogeneous, 15 highly variable, and adaptable. Increased virulence of Group III organisms by animal passage suggested that some possess a markedly higher virulence potential than has been reported for the "atypicals" from human beings. Tuberculosis, or tuberculosis-like disease, in domestic animals may be caused by a variety of acid—fast organisms. The problems of grouping mycobacteria on the basis of morphologic characteristics and virulence for laboratory animals are well known (Negré 1947; Chapman, 1960; Guy and Bernard 1960; Runyon, 1960; Xalabarder, 1961). Strains may increase in virulence by animal passage or lose virulence by continued growth on artificial media. Cultures of the same strain may vary somewhat in their growth and pigmentation characteristics, depending on such an insignificant factor as the size of the inoculum (Mallmann §£_§;,, 1961). Recognizing the shortcomings of classification by mor— phologic characteristics and virulence tests only, investi— gators have employed many serologic methods for the study of mycobacteria. A variety of crude antigens, such as cell extracts, whole cells (heat-killed or viable), and Old Tuberculin have been employed to stimulate antibody production in suitable animals. The antisera obtained have been allowed to react with crude l6 antigens properly prepared for precipitation (Meynell, 1954; Schaefer, 1940, 1947), agglutination (Bando, 1962; Furth, 1926), complement—fixation (Vardeman and Larsen, 1961; Toda, 1956), hemagglutination (Boyden, 1951; Middlebrook and Dubos, 1948) and hemolytic tests (Buehler and Rheins, 1959). None afforded a method for the systematic classification of the mycobacteria. Other workers (Coghill, 1931; Creighton and Anderson, 1944; Creighton, Chang, and Anderson, 1944; Heidelberger and Menzel 1934, 1937; Pepys, Augustin, and Paterson, 1959; Seibert, 1949, 1958; Seibert, Crumb, and Seibert, 1950; Seibert and Soto-Figueroa, 1957) approached the problem by attempting to chemically isolate specific antigens from mycobacteria. Two polysaccharide antigens and three pro- tein antigens have been isolated from the mycobacterial cell, but no one has been successful in isolating a series of anti- gens from which a serological classification could be estab- lished. This may be due, in part, to the difficulties en- countered in the isolation and purification of antigenic stubstances, to the low antibody response in animals inoc- ulated with mycobacterial antigens (Parlett and Youmans, 1956), and to the adaptability and mutability of the tubercle bacillus (Darzins, 1958). 17 Applications of Immunodiffusion to the Study of Myco— bacteria. Successful analyses of complex antigenic mixtures by immunodiffusion led to its application to the studies of acid-fast bacilli. Parlett and Youmans (1956) studied the antigenic relationships among 42 strains of mycobacteria by means of the Oudin tube test and a modified Ouchterlony plate method. Employing unheated culture filtrate antigens, and antibodies elicted following the subcutaneous inocula- tion of rabbits, they found that the modified Ouchterlony method was superior to the Oudin test for determining anti- genic patterns. The presence or absence of cross-reacting antigens and antibodies permitted a division of the various strains of mycobacteria into four antigenic groups. Group I consisted of known virulent and avirulent M, tuberculqsis, atypical acid-fast organisms from human beings, an attenuated strain of M. bovis, and a strain of M, gyigm. Four different precipitating antigens were common to all strains placed in this group. Group II was composed of several strains of M, bovis and one strain of M, avium. The filtrates of the group contained two precipitating antigens, both of which were identical to two of the four antigens found in organisms of Group I. 18 Group III consisted mainly of saprophytes. Filtrates of these strains contained four antigens which were iden— tical for the strains within the group, but unrelated to the antigens of Groups I, II, and IV. Group IV consisted of a single strain of an atypical, chromogenic acid-fast organism originally isolated from man. The filtrate of this strain contained two precipitating antigens unrelated to those produced by organisms in Groups I, II, and III. A common antigen among all strains of mycobacteria was not detected. The classical pathogens, M, tuberculosis, M, ayigg, were closely related, and none of the antigens of the saprophytic strains cross—reacted with the antibodies spe- cific for the human, bovine, or avian species. In a study of mycobacterial antigens, Rheins, Burrell, and Birkeland (1956) employed two procedures for eliciting antibodies in rabbits: (1) a single intravenous injection of whole cells of mycobacteria was made with M, bgyig (Ravenel), M, tuberculosis (H37Rv), M, phlei, and the Bacil- lus of Calumette and Guerin (BCG), and the blood collected at intervals; (2) three injections of 2.0 mg of BCG were made at three-day intervals, and the blood collected one week following the last injection. Agar plate diffusion 19 tests were conducted with the antisera, with Old Tuberculin as the source of antigens. The antisera collected following either procedure for eliciting antibodies contained a low titer of a single antibody. Studies of Parlett and Youmans (1956) suggest that the intravenous injection of whole mycobacterial cells elicited antibodies in rabbits which was markedly less effective for detecting antigen by gel—diffusion methods than the antiserum obtained when cell products or culture filtrates were in- corporated in a water-in—oil emulsion, and several subcu- taneous inoculations in rabbits made. Furthermore, Old Tuberculin and P P D (Purified Protein Derivatives) were poor indicator antigens in gel—diffusion tests (Parlett and Youmans, 1958). Seibert and Soto-Figueroa (1957) utilized a tube double- diffusion technique to study the reactions of tuberculopro- teins and polysaccharides with antibodies (rabbit origin) elicited by BCG. Proteins were extracted from unheated cul- ture filtrates, and polysaccharides from heated culture fil- trates. Different protein extracts formed two or more bands with the anti—ECG serum, which indicated that the extracts consisted of mixtures of antigens. The polysaccharide extract produced two distinct bands with the anti-BCG serum. 20 Immunoelectrophoretic studies (Burtin, 1959; Burtin and Kourilsky, 1959) revealed that four different protein anti- gens and two polysaccharide antigens are present in M, tuberculosis. Parlett and Youmans (1958) studied the antigenic rela- tionships among 98 strains of mycobacteria and 4 fungi by the agar-plate diffusion method. The organisms studied, predominantly from human beings were: 20 strains of M, tuberculosis; 54 atypical mycobacteria isolated from man; 4 M, bgyigj 2 M, ayigg; 10 saprophytic mycobacteria; 7 mis- cellaneous mycobacteria, including M, ulcerans, M, balnei, M, marinum and M,fortuitum; and 4 species of fungi: NOcardia pelliteriae, NOcardia braziliensis, Penicillium notatum, and Candida albicans. Antisera (rabbit origin) were prepared from 43 strains of mycobacteria and the 4 species of fungi. Subcutaneous injections of water-in-oil emulsions of concentrated culture filtrates were made weekly for six weeks and after a 10-14 day waiting period, blood was obtained by heart puncture. Viable mycobacterial cell suspensions served as a source of antigens in the immunodiffusion tests. The number of pre— cipitins detectable for any one strain varied from 2 to 6, but no single antigen was common to all of the mycobacteria. 21 There were no cross-reactions between the 4 fungi and the mycobacteria. On the basis of the precipitation bands that were formed, the 98 strains of mycobacteria could be divided into 8 antigenic groups. Parlett and Youmans (1959) conducted studies of the specificity and sensitivity of a gel—diffusion tube test to determine its usefulness in detecting antibodies in human sera specific for mycobacterial antigens. Sera were obtained from tuberculous and non—tuberculous hospitalized patients. Gel double-diffusion tube precipitation tests were performed with concentrated culture filtrates as sources of antigens. From 465 patients hospitalized for various diseases other than tuberculosis, ten specimens were positive. These ten were assumed not to be true biologic false positives, but were accounted for by technical factors, or undiagnosed or healed tuberculosis. Of the serum from patients with extrapulmonary tuberculosis, 43.9% were positive; of sera from persons classified as having inactive tuberculosis or a history of "cured'I tuberculosis, 56.4% were positive; of sera from patients from whom only atypical acid-fast bacilli had been isolated 74.3% were positive; of sera from cases of moderately advanced active pulmonary tuberculosis 73.5% were positive; of sera from cases of far advanced active 22 pulmonary tuberculosis 84.2% were postive; and in nine cases of primary tuberculosis in infants, 10 to 120 months of age, all of the sera were positive. The specificity of the gel-diffusion tube technique was demonstrated and also the need for a more sensitive diag— nostic method. Sushida g£_§l, (1961) utilizing the Ouchter- lony plate method to determine antibodies in the sera of tuberculous persons, found that the plate technique also lacked the sensitivity required for a diagnostic test. Parlett (1961) recognized the influence of serum anti- gen on precipitation tests conducted with serum from pa- tients with chronic disease. He reported that when an anti- gen resides in the tissues, as happens in mycobacterial in— fections, varying amounts of antigen may enter the circulation. He assumed that there is an inverse relationship of serum antigen to serum antibody concentration; therefore, in chronic disease, serum would contain free antibodies only in the cured individual. At other times, it is probable that some antigen will be present in the serum, either free or as part of an antigen—antibody complex. ;g_yit£g_tests for antibodies are performed by adding antigens to the serum. With serum which contains significant antigen, this addition would shift the antigen-antibody ratio further away from the 23 equivalence ratio and into an area in which antigen-inhibition may occur (Parlett, 1961). (Parlett (1961) tested the sera of 373 tuberculous patients for the presence of serum anti- gens, and antibodies. Results showed that 40.8%»were pos- itive for both antigen and antibody; 3.7% were negative for both; 47.6% were positive for antibody only; and 7.7%»were positive for antigen only. It may be necessary that diag- nostic tests for the presence of serum antibodies elicited by mycobacteria will require various antigen concentrations. The gel double-diffusion test is more promising from the standpoint of specificity and sensitivity than hemagglutina- tion and comlement-fixation tests. Testing each serum with four concentrations of antigens, Parlett examined serum specimens from 1,379 non—tuberculous persons and from 1,452 persons in all stages of tubercular infection and receiving various therapeutic treatments. Of sera from the non-tuber- culous persons 6.5%.were positive; and from tuberculous patients 82.3% of the sera positive. Kniker and LaBorde (1962) and Kniker and Heiner (1960) made antigenic analyses of 20 strains of mycobacteria re- presentative of the major mycobacterial groups. A culture filtrate of each strain was placed on an ion—exchange column of DEAE - cellulose and followed with gradient elution. 24 From each filtrate 11 to 16 fractions were collected and the antigenic composition of each fraction was determined by a micro-immunodiffusion technique. In each culture fil- trate 20 to 30 antigens were present. Antigens common to the 20 organisms were detected and moreover, 2 fractions were found that contained 1 to 5 antigens specific for each organism or its group. The separation of different fractions of the culture filtrates by elution might have resulted in the differentia— tion of antigens (in immunodiffusion tests) that could have been undetectable without previous fractionation. There is also the possibility that fractionation resulted in the dis- sociation of large antigenic molecules into fragments. Such fragments may have possessed two or more functional groups, and would therefore be capable of combining with antibodies to form visible precipitates. MATERIALS AND METHODS Cultures Fifty representative strains of mycobacteria were studied. These strains, with their hypersensitivity and infectivity characterizations,* are listed in Tables 1 and 2. The strains from which antisera were prepared are listed in Table 3. The 5 atypical strains of human origin and the 8 known laboratory strains were obtained from the Tuberculosis Unit, Communicable Disease Center, Atlanta, Georgia. The remain- ing 37 strains were "atypicals" isolated from soil, cattle, and swine in 1960-1962 by members of the Bovine Tuberculosis Project, Michigan State University, East Lansing, Michigan. *Studies were made by Mallmann and Robinson (1961, 1962). Infectivity studies were made by intradermally in- oculating 0.1 m1 of suspensions containing 0.1 mg wet wt of the organisms into guinea pigs. The development of a lesion of any size was considered evidence of virulence. Tuberculin sensitivity tests were performed by two pro- cedures: (1) 30 days post-inoculation of guinea pigs intra- muscularly, quantities of 0.1 ml of avian O.T. diluted 1:20, 0.1 ml of mammalian O.T. diluted 1:10, and 0.1 m1 of Edward's PPD-B were injected intradermally. Reactions were recorded after 48 hours. (2) Approximately 50 days post-inoculation, guinea pigs inoculated intradermally in the infectivity studies were tuberculin tested as in (1) above. 25 26 Table l. Atypical strains from which culture filtrate antigens were prepared prepared Group Source Strain Hypersensitivity Infectivity in guinea pigs for guinea pigs I man R-P4 mammalian virulent R-P8 mammalian virulent I I man R-P 15 nondefinitive avirulent soil X27-l no sensitivity avirulent X28-1 no sensitivity avirulent X37-1 no sensitivity avirulent cattle 368E-1 no sensitivity avirulent I I I man R-P 39 avian virulent cattle 50B- 0 avian/battey virulent 62D- 0 avian/battey virulent B lOZE-O avian virulent 107E- 0 nondefinitive virulent 83F- 0 nondefinitive virulent 94F- 0 nondefinitive virulent 98F-0 avian virulent B 124F- 0 nondefinitive virulent B66F-l no sensitivity avirulent swine 152A 1-1 avian virulent 93C-0 avian virulent 15 1C—1 avian virulent 172C1-l mammalian virulent 186C-1 avian virulent 19 3C2-1 mammalian virulent 198C- 1 avian virulent 228C—1 mammalian virulent 2421-1 avian virulent 244 I- 1 avian virulent 350-1 avian virulent 352-1 avian virulent Table 1. --Continued 27 Group Source Strain Hypersensitivity InfeFtiVitY in guinea pigs for guinea pigs IV man R-P3805 nondefinitive virulent cattle B117B- 0 no sensitivity avirulent B368D-l no sensitivity avirulent B292 E-l no sensitivity avirulent B28E-l no sensitivity avirulent M177F-O no sensitivity avirulent Pseudo- swine 152A2-1 avian virulent chromes cattle 63A- 0 avian/battey virulent 58A-1 avian virulent 52H- 1 avian virulent 65F-O avian virulent 77F-1 non-definitive virulent 28 Table 2. Known laboratory species from which culture filtrate antigens were prepared Species Hype r sensitivity in guinea pigs Infectivity for guinea pigs K . bovi s , Ravenel . bovi s , BCG KIK . tuberculosis, H37Ra tuberculosis, H37Rv avi um fortuitum phlei I: ls I: I: I: I smegmati s mammalian mammalian mammalian mammalian avian no sensitivity no sensitivity no sensitivity virulent virulent vi rulent yirulent virulent virulent avirulent avi rulent 29 Table 3. Organisms used for the production of antibodies ‘izpzzizzsigig? .Ttsiizrggs AIL £93.13: Ravenel mammalian virulent All; m avian virulent _l\£ 2.13.131 no sensitivity avirulent 11/1; fortuitum no sensitivity virulent R-P 8 I mammalian virulent R-P l 5 I I nondefinitive avirulent R-P39 III avian virulent 107E- l I I I nondefinitive virulent 193C2-l III mammalian virulent R-P3803 I V nondefinitive virulent 58A-l Pseudo- nondefinitive virulent chrome 30 Atypical strains are grouped according to the Runyon system of classification. In addition, the coding system of the tuberculosis research group of Michigan State Uni- versity has been retained for culture designations. The letter "R" preceding atypical strains of human origin refers to cultures originally identified by Dr. E. H. Runyon. The letter "X" preceding a number indicates that the strain was isolated from soil. Letters preceding numbers of cultures isolated from bovine and porcine tissues refer to the methods uSed for initial isolation: ”B" refers to the pentane-enzyme method; "M" refers to the pentane method followed by a disinfectant treatment, and no letter preceding the culture number re- fers to a culture isolated by the sodium hydroxide method. The letter designations following the numbers of atyp- icals of animal origin indicate the tissues from which the strains were isolated; A = cervical lymph nodes, B = thoracic lymph nodes, C = mesenteric lymph nodes, D = carcass lymph nodes, E = lung, F = skin, H = Peyer's patches, I = lesions. The number following the letter refers to the year in which the culture was isolated; 0 = 1960, l = 1961, and 2 = 31 Tests1 of the ability of ten strains of atypicals to infect calves showed that one strain, 152Al-1, was virulent. Strains SOB-0, 62D-0, 107E-0, 93C-0, 172C -l, 186C-l, 193C -1, l 2 Bll7B-0, and 52H—1 were avirulent for calves. Antigens Strains of mycobacteria used for antigen production were seeded into Dubos Broth Base (Difco) with 1% glucose to ob- tain actively growing cultures. These cultures were seeded into the medium of Wong and weinzirl as modified by Fregnan, Smith, and Randall (1961). The latter medium was dispensed as follows: 2 m1 and 5 m1 into 30 m1 screw-cap culture tubes, 10 ml in to 50 ml screw-cap culture tubes, 50 ml into 250 m1 Erlenmeyer flasks, and 500 ml into 1 gal glass bottles. Tubes containing 2 m1 of modified wong and weinzirl's medium were seeded with 2 ml of Dubos Broth cultures. When abundant growth developed, the 4 ml culture was seeded into a tube containing 5 m1 of modified WOng and weinzirl's medium. After heavy growth appeared in this tube, the 9 m1 culture 1Calves were injected with 1.0 to 10.0 mg wet wt of cul- ture. Only those producing progressive disease as determined by histopathological studies were considered virulent. 32 was added to 10 ml of wong-Weinzirl's medium. Similar step- wise increases in amounts of inoculum were employed until four bottles, each containing 500 ml of this medium were seeded with each strain. The cultures were grown as pellicle cultures in the bottles for 8-20 weeks. Bfichner funnels with Reeve Angel filter paper no. 202 were used to remove clumps of cells, and the filtrate was sterilized by Berkefeld filtration under vacuum. The filtrate was stored at 4 C and sterility tests were conducted by seeding duplicate 1 ml quantities of the filtrate into tubes of Brain Heart Infusion Broth and mod- ified WOng and weinzirl's medium. Tubes containing the latter medium were incubated at 37 C for 3 weeks, and the tubes with Brain Heart Infusion Broth were incubated at 37 C for 3 days. When necessary, the filtrate was refiltered through Berkefeld filters. The sterile filtrates were concentrated about 10X at room temperature by the pervaporation method of Jennings (1953). Dialysis tubing with a flat width of 2.88 to 3.14 inches and a wall thickness of 0.0016 inches was sterilized; by autoclaving at 121 C for 15 min. The filtrate was poured aseptically into the tubing and the filled sac was placed in front of an electric fan set at medium speed. When 33 concentrated, the filtrate was stored at 4 C and sterility tests were conducted as above. Antisera A reference antiserum, 1yophilized M, tuberculggig (H37Ra) antiserum (rabbit origin) was obtained from Difco Laboratories, Detroit, Michigan. For use in precipitation tests, the 1yophilized antiserum was reconstituted with physiological saline. Antibodies were elicited from cul- ture filtrates of eleven strains of mycobacteria: M, ayigm, M, legi, M, fortuitum, M, bgyig (Ravenel), R—P8, R-P15, R-P39, R-P3803, 58A, 107E-l, and 193C2-l. Three Dutch Belted rabbits were used for each of the first nine strains listed, and two rabbits each for strains 107E-1 and 193C2-l. The method employed for stimulating antibody production was essentially the same as that used by Parlett and Youmans (1956). Falba and paraffin oil (viscosity 125/135) were placed individually in screw—cap tubes and sterilized in the hot air oven at 180 F for 2 hrs. A.water-in-oil emul- sion was prepared by mixing with a syringe two parts of con- centrated culture filtrate, two parts paraffin oil and one part melted Falba (Freund, 1951; Freund and Bonanto, 1944). At intervals during mixing, one drop of the mixture was 34 placed on water (at room temperature) to test its stability (Crowle, 1961). When the drop remained intact, mixing was discontinued and the emulsion stored at 4 C. One week preceding inoculations, 15 m1 of blood were drawn from each rabbit by cardiac puncture. Aliquots of each serum specimen were stored at 4 C and -20 C to be used subsequently as controls. Subcutaneous inoculations of 0.75 ml of the water-in- oil emulsions were made into one site once weekly for six weeks. Ten days after the last injection, blood was col- lected, and aliquots of each serum speciment were stored at 4 C and —20 C. Electrolytes Used for Preparing Diffusion Media The electrolyte used to prepare the medium employed for all final studies was phosphate-saline, 0.15 M, pH 7.2. The formula for this buffer and those for others used in preliminary studies follow: Phggphate—Saline, 0.15 M, pH 7.2 Monopotassium Phosphate Solution, 0.15 M . . . . 150 ml Disodium Phosphate Solution, 0.15 M . . . . . . 350 ml Sodium Chloride Solution, 0.15 M . . . . . . . . 500 m1 35 Barbital, ionicity,0.15,*MpH 7.4 Sodium Barbital . . . . . . . . . . . . . . Sodium Chloride . . . . . . . . . . . . . . 1 N Hydrochloric Acid . . . . . . . . . . . Distilled water . . . . . . . . . . . q.s. TRIS,¥ionicity40.15Lf pH 7.4 2-Amino-2-(Hydroxymethyl)-l, 3-Propanediol . . . . . . . . . . . . . . 1 N Hydrochloric Acid . . . . . . . . . . . Sodium Chloride . . . . . . . . . . . . . . Distilled water . . . . . . . . . . . q.s. Ethylenediamine-Acetic Acid (ETDA), ionicitv 0.15,* pH 7.4 Ethylenediamine . . . . . . . . . . . . . . Glacial Acetic Acid . . . . . . . . . . . . Distilled water . . . . . . . . . . . q.s. Phosphate, ionicity 0.15,? pH 7.4 Disodium Phosphate . . . . . . . . . . . . . Monosodium Phosphate Monohydrate . . . . . . Distilled Water . . . . . . . . . . . q.s. Phosphate, 0.01 M, pH 7.5 1 part of 0.01 M Monopotassium Phosphate 5 parts of 0.01 M Dipotassium Phosphate 6.98 9 6.00 g 2.70 1000.00 9.30 74.00 7.00 1000.00 m1 m1 ml ml 15.80 g 23.80 g 1000.00 12.80 2.62 1000.00 m1 ml 36 Isotonic Sodium Chloride, ionicity 0.15,* pH variable Sodium Chloride . . . . . . . . . . . . . . 8.80 g Distilled Water . . . . . . . . . . . q.s. 1000.00 ml Media_for Diffusion Studies Different diffusion media were used in preliminary tests conducted by the glass-tube, petri dish, and glass-slide methods. Purified agar, Difco certified, was the gelling agent and the various media were prepared in the following manner: To 100 m1 of an electrolyte solution contained in a flask, 1 m1 of 1-100 merthiolate and a weighed quantity of agar were added, and the flask placed in a flowing steam. When the agar had dissolved, the solution was filtered through Reeve Angel filter paper no. 202 and immediately used for the preparation of diffusion tubes or plates. Gels for Preliminary Studies by the Tube and Petri Di§M_ Methods. Gels containing 1% and 2% agar were prepared as above in 0.01 M phosphate buffer, pH 7.5 for studying anti- gen-antibody systems by the glass-tube and Petri dish methods. *Ionicity values (based on electric conductivity tests) taken from p. 302 in Immunodiffusion by A. J. Crowle (1961). 37 Gels for Preliminary Studies by the Agar Slide Method. Preliminary studies to determine the best diffusion medium for the development of precipitation bands were conducted on agar slides by varying the agar concentration, pH, ionic strength, and the buffering system. Electrolytes used to prepare 1% and 1.5% agar media, pH 7.4, with an ionic strength of 0.15 were: Sodium chloride, barbital, tris (hydroxymethyl) aminomethane (TRIS), ethylenediamine-acetic acid (ETDA), phosphate, and phosphate—saline. Other media employed in studies by the slide method were: 0.7% agar gels prepared in 0.01 M phosphate buffers of pH 7.2 and 7.4, respectively; 0.7% agar gel prepared in 0.01 M phos- phate-saline, pH 7.2; and 1% agar gel prepared in 0.15 M phosphate-saline, pH 7.2. immunodiffusion Tag; Procedures In preliminary studies the glass tube, Petri dish and agar slide methods were employed. The latter method was used in the final studies of precipitinogenic relationships. Concentrated modified WOng and weinzirl's medium and normal rabbit serum served as controls. The reactants were con- centrated culture filtrates and rabbit antisera. 38 Tube Method. Soft-glass tubing with an inside diameter of 3 mm was cut into 10 cm lengths. The tubes were washed with Tide and put through one tap water rinse, five distilled water rinses, and a final rinse in 95% ethanol. Each tube was heat—sealed at one end and internally coated with a thin film of 0.1% agar prepared with ditilled water. The film was applied by inverting the tubes in a beaker contain— ing sufficient agar to cover them, and autoclaving at 15 lbs pressure for about 15 min. When the temperature decreased to 75 C, excess agar was removed from the tubes by inverting and shaking them individually. The tubes were placed in plastic bags, dried overnight at 42 C, and stored at 4 C. Before the tube immunodiffusion tests were performed, the other materials and culture filtrates were warmed to prevent premature solidification of the agar. Sterile Kahn tubes, culture filtrates and 2% agar medium were placed in a 50 C water bath. Graduated 0.2 m1 pipettes and capillary pipettes were warmed in a 45 C incubator. A 2:1 ratio of antigen to agar mixture was prepared in a Kahn tube and then added, by means of a capillary pipette, to a depth of 2.5 to 3.0 cm in the agar—coated tubes. When the antigen—agar layer had gelled, a 0.5 cm layer of 1% agar was added. When the second layer had solidified, undiluted 39 antiserum was added and the tube placed upright in a wooden rack.1 The tubes were covered and some were incubated at 37 C for 7 days, others at 25 C for 14 days. At 12-hr intervals, the reactions were observed by holding the tubes against a black background and passing a strong fluorescent light obliquely through them.2 Petri Dish Method. Glass Petri dishes were washed and rinsed as described above in the tube method. To form a 4 mm layer, 8 ml of diffusion medium were pipetted into each dish and allowed to solidify. The dishes were immediately placed in portable humidity chambers3 at 4 C for 6-12 hr. Then, with paper templates as guides, circular reactant wells were cut with cork borers and the agar-discs removed with a 16-gauge needle. wells employed were: 7 mm wells placed 8 mm apart, 6 mm wells placed 7 mm apart, and 4 mm wells lRacks were made by drilling holes, 0.5 cm in diameter and 4.0 cm deep, in 2 x 4'x 8-inch softwood blocks. 2The light source was lamp model M209, made by the Dazor Mfg. Corp., St. Louis, Mo. 3Moistened gauze was placed in polyprolene pans with tightly fitting covers. The pans were made by the Dynalab Corp., 625 Goodman St., Rochester, N. Y. 40 placed 5 mm apart. The wells were sealed with 0.3% agar prepared with distilled water, and the dishes stored in humidity chambers for at least 1 hr before the reactants were added. The wells were filled by means of tuberculin syringes. The dishes were incubated for 10 days in moist chambers at constant temperature of 28 C and 37 C. Reactions were ob- served daily by holding the dishes in front of the difffused blue-filtered light from a microscope illuminator.l Agar Slide Method. Glass slides of two sizes, 3.25 x 4.0 in. and 2.75 x 2.75 in. were cleaned, rinsed and dried as in the previously described methods. They were then coated with 0.3% agar prepared with distilled water. All agar was removed from one side of each slide with a lint- 1ess cloth before drying at 37 C. Each slide was placed with the coated-side up on a level surface, and covered with a uniform 2 mm layer formed by pipetting 10 m1 of diffusion medium onto it. These agar blanks were stored in moist chambers at 4 C for 6-12 hr. 1Lamp Model 370, 100 watts, 115 volts, made by the American Optical Co., Brooklyn, N. Y. 41 Circular reactant wells were cut with cork borers; re- actant trenches were cut with thin plastic rulers of the same lengths as the desired trenches. The most common size and arrangement of the reactant depots was one or more 86 mm x 3 mm antigen trenches extending the length of the 3.25 x 4.0 in slide, with circular antibody wells, 6 mm in diam- eter, placed parallel to and 7 mm from the trenches. All reactant depots were sealed with 0.3% agar prepared with distilled water. The reactants were added by means of tuberculin syringes and the slides incubated in moist cham- bers at 28 C for 5 days. Observations were made at 8-hr intervals by holding the slides so that diffused blue light from a microscope illuminatorl passed obliquely through the medium. The Concentration of Reactants Small quantities of M, bovis (Ravenel) and M, gyigm_anti— sera R-P39 and 107E—0 antisera were placed in sterile dial- ysis tubing. They were concentrated 2X by two methods: (1) a filled—tube was placed in a concentrated solution of Carbowax 40002 at room temperature for about 2 hr, and (2) a similarly lSee Footnote 1. P. 40 2Obtained from the Union Carbide Co., Charleston 30, W. Va. 42 filled-tube was concentrated by pervaporation at 4 C until the volume decreased one-half. The antisera that had been concentrated by dialysis against Carbowax were employed in immunodiffusion tests with and without being further dialyzed against phosphate buffer, pH 7.2. The antisera concentrated by pervaporation were used in the tests without further treatment. Portions of the culture filtrates that had been formerly concentrated 10X were further concentrated to 20X by per- vaporation at 4 C. Preliminary tests were performed by the agar slide method with concentrated antisera and culture fil- trates concentrated 20X. Immunodiffusion tests were conducted employing numerous combinations of filtrate-antiserum concentrations. An- ti-sera concentrated two-fold, undiluted, and diluted 1:2 and 1:4 with physiological saline were diffused against culture filtrates that were concentrated 20X, 10X, and 5X. Stainingyof Precipitiation Lines The reactant depots on the agar slides were rinsed and filled with distilled water. Moistened filter paper was immediately placed on top of the agar, and the slide dried at room temperature. The filter paper was carefully removed 43 and the slide soaked overnight in phosphate—saline buffer, pH 7.2. It was then placed in distilled water for 15 min and finally stained for 10 min in 0.1% thiazine red R pre- pared with 1% acetic acid. The agar background of the precipitation lines was de- stained by dipping the slide in 70% ethanol containing 1% acetic acid. After drying at room temperature, the slide was examined for stained precipitates. Method for Calculating Antigenic Rgiatedness of Reference Strains By modifying an equation employed by Colwell (1963), the degree to which the reference strains were antigenically interrelated could be determined. In taxonomic studies by Adansonian analysis, Colwell used this equation: S = ; where S = the similarity of two organisms; NS = the number of positive features shared; and Nd = the number of features positive for one organism and negative for the other. In adapting the equation to this study, the following symbols are employed: 44 SXy = the percent of the total antigens in strains X and Y that are related. Xr = the number of antigens in X related to those in Y. Yr = the number of antigens in Y related to those in X. Xt’ Yt = the total detectable antigens in strains X and Y , respectively. The antigenic relatedness of two strains, then, may be represented by the percent obtained when the following equa- tion is employed: Xr + Yr Sxy = x + Y . 100 t t and since X = Y , r r 2X xy + Xt Yt For example, by letting X = M, avium and Y = R-P15, and substituting values for the related and total antigens as listed in Table 6, it is seen that: 45 s 2 l . xy = 'g—fI—ji 100 20 Or 20% of the total antigens found in M, avium and R-P15 are related. RESULTS AND DISCUSSION OF PRELIMINARY OBSERVATIONS Relative Sensitivities of Methgds Table 4 presents a comparison of the relative sensitiv- ities of the tube, Petri dish, and agar slide methods for studying the precipitation lines formed by the homologous antigen-antibody systems of M, syism and M, tuberculgsis, H37Ra. In repeated tests performed by the agar slide method, six ;precipitation.lines were detected in the reaction area between the M, syigm_culture filtrate and M, syism_antiserum, and five lines were distinguishable in reactions of M, PEPEET culosis, H37Ra, filtrate and the homologous antiserum (Difco). In tests conducted by the tube and Petri dish methods, the homologous systems of M, syism_formed five precipitation lines, and M, tubsiculqsis, H37Ra, formed four. Varying the methods by replenishing the antiserum layer of the tubes and refilling the reactant depots of the agar slides and Petri dishes served to intensify the lines of precipitation that developed, but did not increase the num- ber of lines. 46 47 Table 4. The number of precipitation lines detectable in double-diffusion tests conducted by the tube, petri dish, and agar slide methods J ___—,—__ Lines formed by the homologous antigen - antibody systems of: Method M . M; tuberculosis -—' avmm H37Ra Tube, 0. 7% Agar* 5 4 1. 0% Agar 5 4 Petri Dish, 0. 7% Agar 5 4 l. 0% Agar 5 4 2.. 0% Agar 5 4 Agar Slide, 0. 7% Agar 6 5 1% Agar 6 5 2. 0% Agar 6 5 *The diffusion media for tests conducted by all methods were pre- pared with 0. 01 M phosphate buffer at pH 7. 4. 48 In addition to the sensitivity of the agar slide method it was the simplest method to perform, and the reactions on slides were most easily observed and photographed. The agar slide method was therefore employed for the study itself. Reactions in Various Media The number of precipitation lines that developed in various diffusion media on agar slides is summarized in Table 5. An identical number of lines formed in all media: 6 lines from the homologous antigen-antibody systems of M, syigm and 5 lines from M. tuberculosis, H37Ra. A 1% agar gel prepared with 0.15 M phosphate-saline, pH 7.2, was selected as the medium for studying the antigenic relationships of the mycobacteria. The Influence of Reactant Concentrations After selecting the diffusion medium and the method for studying the precipitinogenic relationships, tests were per— formed to determine the necessity of concentrating the anti- sera and further concentrating the culture filtrates beyond 10X. At various concentrations of the antigens of M, syism, M, ssyis (Ravenel), R—P39, and 107E-0 with their homologous antibodies it was found that: 49 Table 5. The number of precipitation lines detectable in various diffusion media in tests conducted by the agar slide method Homologous antigen- antibody systems of: D'ff ' d. . 1 usmn me ium £ tuber- _M_. avium culosis, H37Ra Physiological Saline, pH variable, Agar 1% 6 5 Physiological Saline, pH variable, Agar 1. 5% 6 5 Barbital, u =0. 15, pH 7. 4, Agar 1% 6 5 Barbital, u =0.15, pH 7. 4, Agar l. 5% 6 5 TRIS, u=0.15, pH 7. 4, Agar 1% 6 5 TRIS, u =0. 15, pH 7. 4, Agar l. 5% 6 5 ETDA, p30.15, pH 7. 4, Agar 1% 6 5 ETDA, No.15, pH 7. 4, Agar 1. 5% 6 5 Phosphate, 0. 01 M, pH 7. 4, Agar 0. 7% 6 5 Phosphate, I4 =0.15, pH 7. 4, Agar 1% 6 5 Phosphate, u =0.15, pH 7. 4, Agar l. 5% 6 5 Phosphate- Saline, o. 01 M, pH 7. 2, Agar o. 7% 6 5 Phosphate- Saline, 0.15 M, pH 7. 2, Agar 1% 6 5 Phosphate- Saline, 0.15 M, pH 7. 4, Agar 1% 6 5 Phosphate- Saline, 0. 15 M, pH 7. 4, Agar 1. 5% 6 5 50 (l) The two-fold concentration of antisera dif- fused against the 10X and 20X concentrated filtrates enhanced the intensity and broadness of precipitate bands. It did not increase the number of bands. (2) No precipitation lines developed when 1:2 and 1:4 dilutions of the antisera from R-P39 and 107E—0 were diffused against 5X, 10X, or 20X concentrations of their homologous filtrates. (3) In contrast to the six lines that developed in homologous reactions employing undiluted or 2X concen- trations of antisera containing antibodies elicited by M, syism_and M, sgyis_(Ravenel), only four precipitation lines formed when antisera diluted 1:2 were used. Further dilution of the M, ngis_(Ravenel) and M, syism_antisera to 1:4 resulted in the formation of one and two precip- itation lines, respectively. (4) Except in those instances in which the antigen wells were replenished during the immunodiffusion tests, fewer numbers of precipitation lines formed from the homologous reactions of the 5X concentrated filtrates than the 10X. 51 (5) Filtrates concentrated 20X did not increase the number of precipitation lines that had been observed when the 10X concentrated filtrates were used. 0n the basis of the above results undiluted antisera and culture filtrates that were concentrated a maximum of 10X were used for studies of the precipitinogenic relationships of the mycobacteria. However, when comparing the reactions of two antisera with one filtrate, or two filtrates with one antiserum, dilutions of the reactants were sometimes employed. Antibody_Responss in Individusi_Rabbits Tests of the reactivity of antisera collected from rab- bits showed wide variations in the precipitin response of individual rabbits to injections of the same antigenic prep- aration. The results of these tests are presented in Table 6. Of nine culture filtrates that were each injected into three rabbits, only the filtrate prepared from M, syisg. elicited identical precipitin responses in all. Six of the nine filtrates elicited identical responses in two rabbits only, and two filtrates elicited a different precipitin response in each of the three rabbits injected. Of two filtrates that were injected into only two 52 Table 6. The number of precipitins detected in individual rabbit antiserum by homologous antigen-antibody reactions* Culture No. rabbits Maximum no. of precipitins detected filtrate injected in individual rabbit antiserum %v%' 3 6 6 5 1A m 3 6 6 6 M: hlei 3 6 4 6 M: fortuitum 3 6 6 5 R-P8 3 2 4 4 R-P15 3 3 4 2 R-P39 3 2 2. l lO7E—0 2. 2 l 193C2-l 2 3 3 R-P3803 3 6 5 4 58A-l 3 1 l 2 *All tests were conducted on agar slides in l. 0% agar prepared with 0.15 M phosphate-saline buffer, pH 7. 2 (incubated at 28 C). 53 rabbits each, only one elicited identical precipitin re— sponses in both rabbits. The antisera employed in the study of precipitinogenic relationships were those that contained the greatest number. of detectable precipitins in the preliminary tests of homol- ogous antigen-antibody systems. RESULTS OF PRECI PITINOGENIC STUDIES HOmologous Antigen-Antibody Reactions In diffusion tests of 12 mycobacterial strains employing homologous culture filtrate antigens and antisera, from 2 to 6 precipitation lines were observed (Table 7). Reactions of the homologous antigen-antibody systems of M. MisRavenel (Fig. l), M. M, M. ghie_i (Fig. 2), M, fortuitum (Fig. 3), and R-P38OS each developed six pre— cipitation lines. This was evidence that each organism possessed a minimum of six precipitinogens. The culture filtrate of M, tubercuiosis (H37Ra) reacted with the homologous commercial antiserum (Difco) to produce five lines of precipitation. Yet, the same culture filtrate produced six precipitation lines when reacting with the heterologous M, sgyis_(Ravenel) antiserum. Repeated immuno— diffusion tests designed to demonstrate that six or more antibodies were present in the Difco-prepared M, tuberculgsis (H37Ra) antiserum were conducted. The dialysis of the cul— ture filtrate and its use in tests in varying proportions and concentrations of both reactants, failed to demonstrate more than five antibodies existing in the Difco antiserum. 54 55 Since the presence of six precipitinogens in M, tuber— culosis (H37Ra) was established in tests with M, bovis (Ravenel) antiserum, the detection of only five with the reconstituted Difco antiserum indicated that the former was a somewhat superior reactant for immunodiffusion studies. This could possibly be due to variations in the methods of preparing antiserum or to individual differences in the response of rabbits immunized. Nevertheless, the Difco antiserum was employed as one of the reference antisera and was diffused against all culture filtrates. Two strains of human origin, R-P8 (Group I) and R—P15 (Group II) formed four lines of precipitation with the homologous antiserum. With each of the strains, two of the four lines were of moderate intensity and the other two were of low intensity. Three lines, all of moderate intensity (Fig. 6) were produced by the homologous filtrate-antiserum systems of 193C2-l, a Group III strain originally isolated from mesenteric lymph node lesions of a pig. The least reactive homologous antigen-antibody systems produced but two lines of precipitation. These were the systems of a Runyon Group III strain of human origin, R-P39; a Group III strain of bovine (lung) origin, 107E-0; and a 56 Table 7. The number of precipitation lines formed by homologous and heterologous antigen-antibody reactions antiserum 2% era ,. r: a a i ’v‘ Q) a) O H H H 94 E > O O A H H 04 H O (d .—1 ”J E H g, Q, 2" T o 3... . . :4 u U) H V E .H S D g g v H a) O '5' f3 “3 .3 .33 "‘ v o I o '9 o :3 a: > .c: 8 00 E} S; I N 22 H 3 .13 u I‘- (0 ‘H [1] U | ' ' :1: ' ‘ ' ' ' C ox ' 00 v EIEI Elzlzlmmm—«Hmm M. bovis, Ravenel 6 5 l 0 2 2 2 0 0 1 2 l _I\_/1. tuberculosis, 6 5 1 0 2 2 2 0 l 2 l H37Ra M. avium 1 1 6 1 2 1 1 1 1 3 1 M: phlei 0 0 l 6 2 l l 0 0 l 0 0 ll. fortuitum 2 2 2 2 6 l l 0 0 2 2 l R-P8 (Grp I) 2 2 1 1 1 4 2 0 0 2 1 0 R-P15 (Grp II) 2 2 l l l 2 4 0 l 2 l 1 R-P39 (Grp III) 0 O l 0 0 0 0 2 0 0 O 0 107E-0 (Grp III) 0 0 l 0 0 0 l 0 2 l 0 O 193C2-l (Grp III) 1 l 3 l 2 2 2. 0 1 3 1 l R-P3805 (Grp IV) 2 2 1 0 2 l l 0 0 l 6 0 58A-l l l 0 0 1- 0 l 0 0 l 0 2 (pseudochrome) _l\£ tuberculosis, 6 5 l 0 2 2 2 0 0 l 2. 1 H37Rv M bovis. BCG 3 3 1 0 1 1 1 0 0 1 1 1 R-P4 (Grp 1) 0 0 0 0 0 2 0 0 0 0 0 0 Msmegmatis 002220100000 57 pseudochrome, SSA-l, originally isolated from a bovine cer- vical lymph node. The two lines of precipitation. that developed from the homologous reactions of these three strains were of rather low intensity. Replenishing reactant wells enhanced the in- tensity but did not increase the number of bands observed. grecipitinggenic Relationships of the 12 Reference Strains The number of precipitation lines formed in the homol- ogous and heterologous reactions of 12 strains of mycobacteria are summarized in Table 7. Included are classical strains, two saprophytes,members of the four Runyon groups, and a pseudochrome. Filtrates of the Ravenel strain of M, bgyi§_and the H37Ra and H37Rv strains of M, tuberculosis gave identical antigenic patterns when diffused against the 12 antisera. There were no cross—reactions of the three filtrates with the antiserum prepared from M, phlei, R-P39, or lO7E-O. Each of the filtrates of M, pgyig (Ravenel) and M, tuberculosis (H37Ra and H37Rv) reacted to produce one line of precipita— tion with the antisera of M, avium, 193C -1, and the pseudo— 2 chrome, SBA—l; two lines each were produced in reactions with M, fortuitum, R-P8, R—PlS, and R—P3803; six lines 58 developed with the M, bgyig (Ravenel) antiserum, and five lines with the Difco M, tuberculosis (H37Ra) antiserum. The reactions of the six antigens in the filtrates of M. 1191551 (Ravenel) and M. tuberculfiig (H37Ra and H37Rv) were compared; when M, bgyi§_(Ravenel) antiserum was placed in the center well and diffused against the three filtrates contained in three equidistant depots surrounding the anti- serum well, six continuous lines of precipitation developed. This indicated that a minimum of six identical antigens were present in the Ravenel strain of M, bgyig, and the H37Ra and H37Rv strains of M, tuberculosis (Fig. 4). One of the six precipitating antigens in the above three strains was identical to an antigen found in M, avium; and two of the six antigens were identical to antigens found in the atypicals of human origin, R-P8 (Group I) and R-P15 (Group II). Mycobacterium tuberculosis (H37Ra and H37Rv) and M, bgyi§_(Ravenel) shared one antigen of identity and one antigen of partial identity with M, fortuitum and R—P3803 (Fig. 5). The culture filtrate of M, phl§i_possessed two antigens that were related to those found in M, fortuitum (Fig. 7), 59 Fig. 1. Homologous antigen-antibody reactions of_M_. bovis (Ravenel) l = Conc. culture medium B = M_. bovis antiserum 3 = Conc. M_._ bovis culture filtrate N = Normal rabbit serum Fig. 2. Homologous antigen-antibody reactions of M hlei l, 2, 3, =_M_. phlei antisera Ag = Conc. M. phlei culture filtrate Fig. 3. Homologous antigen-antibody reactions of M_. fortuitum 1 = Conc. culture medium A = M fortuitum antiserum 2 = Conc. M. fortuitum culture , — N = Normal rabbit serum filtrate Fig. 4. Reactions of culture filtrates of classical mammalian species withM. bovis (Ravenel) antiserum 1 = M bovis (Ravenel) antiserum A = Conc. culture medium B = Conc. M bovis (Ravenel) culture filtrate C = Conc. _M_. tuberculosis (H37Ra) culture filtrate D = Conc. M. tuberculosis (H37Rv) culture filtrate Fig. 5. Homologous and heterologous antigen-antibody reactions C = M bovis (Ravenel) antiserum A, D = Conc. M bovis culture filtrate B = Conc. M fortuitum culture filtrate E = Conc. R-P3805 culture filtrate 3. 12919 "3 m. 91.1.; (howl) Fig. 1 Fig. 2 Fig. 3 Fig. 4 Fig. 5 61 Fig. 6. Homologous antigen-antibody reactions of 193C -1 . . . 2 (Group III, Sw1ne Origin) B =Antibodies elicited by 193CZ-l N = Normal rabbit serum 1 = Concentrated culture medium 3 = Concentrated culture filtrate of 193C2-l Fig. 7. Homologous and heterologous reactions of culture filtrates of M. phlei and M_._ fortuitum with reference antisera Antigens: 4, 6 = Concentrated culture filtrate OfM fortuitum 5 = Concentrated culture filtrate of M hlei Antisera: A = R-P 8 = R-P 15 = R-380s — 58A-1 ‘ bovis, Ravenel av1um rmemommuow I ll ’3'5'33'3 F”: 3 M tuberculosis, H37Ra (”Difco”) Fig. 8. Precipitates formed by cultures of three Group 111 strains in reactions with reference antisera Reference antisera: P 8 = R-P 8 (Group 1, human origin) P15 = R-P15 (Group II, human origin) P 3803 = R-P 3805 (Group IV, human origin) 58A-1 = 58 A-l (Pseudochrome, bovine origin) Rav = M bovis, Ravenel Av = M avium Phl =M. phlei For = M. fortuitum P 39 = R-P 39 (Group 111, human orig—1;.) 107 = lO7E-O (Group III, bovine origin) 193 = 193C2-l (Group III, swine origin) Ra =M tuberculosis, H37Ra ("Difco") Antigens: Conc. 186C-1 culture filtrate Conc. l98C-l culture filtrate Conc. 172C1-1 culture filtrate oooooo seoooo I _.1 813 3:77:71 TRATE". I fl" assess [fl ‘178C-l FILTRATE @osoee 1___LT2;CI:L_ .FJfL TRA [if 63 both giving reactions of partial identity when compared with antigens of the latter species. One antigen of M, phlgi_also gave reactions of partial identity with one of the six anti- gens possessed by M, gyigm_and with one of the three antigens possessed by 193C2-1. One antigen of M, phlgi_was related to antigens found in R-P8 (Group I) and R-P15 (Group III); when this antigen was compared with each of the antigens in the latter strains, reactions of partial identity were ob— served for both. The antigens of M, gyigfl_reacted with one or more of the antibodies elicited by each of the reference strains except the pseudochrome, 58A-l. Unlike M, tuberculosi§_and M, bgyig, M, gyigg_was related to M. phlei, R-P39, and lO7E-O. Strain 193C2-1, a Group III of swine origin, was re- lated to all of the reference strains except R-P39, the Group III strain of human origin, and was most closely re— lated to M, gyigm, The three antigens detected in the homol- ogous antigen-antibody reactions of 193C2-l were all related to ML_§yigm, A reaction of identity was observed for one, and a reaction of partial identity for the other two. The Group III strains, R—P39 (human origin) and lO7E-O (bovine origin), were relatively non—reactive. Each pro- duced but two precipitation bands in their homologous 64 antigen—antibody reactions. One antigen of each strain was related to an antigen possessed by M, ayigg, Strain 107E—0 also possessed one antigen that was related to R-PlS, and one that was related to 193C2-l. The atypical strain of human origin, R—P4, contained two antigens that were related to antigens in the other Group I strain isolated from man, R-P8, but was not related to the other 11 reference strains. In contrast, R-P8 was related to 8 reference strains, and not related to strains R-P39, lO7E-O, and SBA—l. The filtrate of BCG, though reacting with the same ref- erence antisera as did the filtrate of the Ravenel strain of M, bpyig, was far less reactive then the latter. The fil- trate of BCG reacted with the Ravenel antiserum to form but three precipitation bands, whereas, six bands were developed by reactions of the homologous antigen-antibody systems of M, Mggig, Ravenel. The BCG strain was also less reactive with other reference antisera than was Ravenel, and had only one antigen related to M, gyigg, M, fortuitum, R-P8, R-PlS, 193C2-l, R-P380s, or 58A-l. Mycobacteriug,§megmatis, contained two precipitating antigens that were related to antigens in filtrates of M, phlei, M, avium, and M, fortuitum, and one antigen related 65 to an antigen in R—PlS. M, smegmatis was not related anti- genically to these reference strains: M, bovis (Ravenel), M, tuberculosis (H37Ra), R—P8, R—P39, lO7E—O, 193C2—l, R-PBBOs and 58A-l. The antigenic "relatedness" of 11 selected mycobacteria is shown in Table 8. The values listed in the table were determined as previously described. Of the total precip- itating antigens found in M;_bgyi§_(Ravenel) and R—P8, as well as, in M, ngi§_(Ravenel) and R-PlS, 40%1were related. Mycobacterium bovig (Ravenel) was found to be more closely related to these two strains than to the other 10 selected strains in Table 8. Mycobacterium avium and 193C -1 were more closely re— 2 lated than were any other two strains of the group with 67% .relatedness. Mycobacterium phlei was more closely related to M, fortuitum than to other strains; 33% of their total antigens were related. On the other hand, M, fortuitum was related most closely to 193C -1. The strains, R—P8, R-PlS, 2 lO7E-O and 58A-l, also were more closely related to 193C2—l than to other strains. The Group IV strain, R-P3803 (human origin), was related to the same extent to M, bovis (Ravenel) and M, fortuitum; 33% of the total antigens of R—P380s and each of the latter two strains were related. 66 cm mm mN 0 Ln N I’V8S NN ON ON mm m M SOBEd'H Ow NN Orv hm 5m 3. NN N N I‘ZDS6I mm [D N O'HLOI O mN 8Ed’H 0 mm ON hm mm om ON ON C ‘14 SIcI'H Goflomon msowofioaoa 93 manomoummn pad ficOOH n *X O ON hm om ON ON O V‘ 8d'H mN mm 3. ON ON mm. M M 'W: umnnuo; O O NN ON ON mm Pa 191113 '76 o O NH he mN mN ON ON mN mm NN Ow Ow mm NH *X Iona/veg ‘ smoq W AoEOMSUOOSom n3 mm .mgoo. pom—damn mum €23 mcowflam H33 .393 m0 “Coupon 5 pouuomon mfiuouomnflootwa Umuooaom mo :mmofipou—dHoH: UmCowflGm 0:8 .w 033. 67 Table 9. The number of precipitation lines formed by culture filtrates of Group 11 strains in reactions with reference antisera antiserum a?" A r: a a T: 8 O C: E t: g. 04 8 0 3 E S H 94 H "4 2 '3 "" :3 94 2‘ H O O ,c: - on H E." H U 0 v V U U) ’5 “V E "" .8 O 0 v V H U) '8 g C‘. ,8 “5 g 2 I: V I: 0* O | g :3 Culture 3 g d E E: 5‘ 3 oo .—. to [1'1 UN m "f <19 - *‘ cu m cu m <: ‘” filtrate , a? , EB . . . , , . g g . oo 91 2| 2| Elzlzlmmmaamm R-PlS (human origin) 2 2 1 1 1 2 4 0 0 2 1 1 368E-l (bovine origin) 0 O 2 4 3 l 1 0 0 O l 0 XZ7-l (soil) 0 O 0 3 O O O 0 0 O O O X28-l (soil) 0 O O 1 0 0 O 0 O 0 O O X37-l(soi1) 0 01422200000 68 Precipitinogenic Relationships of Mycobacteria of Human, Bovine, Porcine and:§oil Qgigins Presented by Groups Group II. The number of precipitation lines formed by culture filtrates of Group II strains in reactions with the reference antisera is shown in Table 9. All Group II strains possessed one or more antigens related to those of M, pglgi, The antigens of X27-l and X28-l reacted only with antibodies elicited by M, phigi and produced 3 and l precipitation bands, respectively. Reactions of the filtrate of 368E—l with the reference antisera resulted in the development of the following pre- cipitation bands: 4 with M, pMigi, 3 with M, fortuitum, 2 with M, ayigm, one each with R-PB, R-P15 and R-P3803, and no hands with the remaining reference antisera. Strain R-PlS, one of the reference strains of human origin, was the only member in Group II that was related antigenically to M, bgyig (Ravenel), M, tuberculosis (H37Ra), 193C2-l, or 58A-l. Group III. The results of immunodiffusion tests in which the filtrates of 22 Group III strains were diffused against the antisera prepared from the 12 reference myco- bacteria are shown in Table 10. 69 One of the 22 strains, that failed to sensitize or cause an intradermal lesion in guinea pigs, B66thq‘was the only mem- ber of the group antigenically related to 58A—1. Seven strains of swine origin (152A -1, 93C—O, lSlC-l, l 186C-l, 193C2-l, l98C-l, and 228C-l) and one strain of bovine origin (BlOZE-O) were related to all but three of the reference strains; the non-related reference strains were R-P39, lO7E-O, and SBA-l. Three strains of swine origin that were avirulent for calves, 93C-O, 186C-1 and 193C -1, and a strain of the same 2 origin that was virulent for calves, 152Al-l, were related to the same 9 reference strains. These four strains also developed similar antigenic patterns in reactions with the reference antisera, showing that the antigens present in the virulent strain were not distinct from those possessed by the avirulent strains of similar origin. Two strains of cattle origin, BlOZE—O and 83F-O, were the only strains in Group III that were more closely related to R-P3803, a Group III of human origin, than to one of the other reference strains. Most of the strains of cattle and swine origins in Group III did not possess antigens that were related to 70 The number of precipitation lines formed by culture filtrates of Table 10. Group III strains in reactions with reference antisera Aoaofifloovdommv 7233 t: 9A: mowmmé a: 9:8 Tmomz a: 9.2 amt: EH 39 omnfim 5 95V 29.x : 95v 3-x 3 fl. EEm fl: antiserum Ed 3d 42 38:: SEE . m 8030.893 .2 Hoco>mm .325 Q filtrate Culture Human origin: R-P39 Bovine origin: SOB-0 62D-0 2 B102E-O 107E-0 83F-0 94F-0 98F-0 O 0 B124F-0 B66F-l Porcine origin: 1 l 1 152A1-1 93c-0 lSlC-l 0 172c1-1 186C-1 193C2-1 198C -1 228C-l 350-1 352-1 2421-1 2441-1 71 the Group III of human origin, R-P39, the Group III of bovine origin, lO7E—O, or the pseudochrome of porcine origin, 58A-l (Fig. 8) . The strains originally isolated from cattle varied con- siderably in their antigenic relationships. Though eight of the nine strains contained at least one antigen related to those possessed by M, ayipm, only 62D-O was related more closely to M, gyigm_than to one of the other reference strains. Six of the isolants from cattle were related to some extent to one or more of the strains, M. fortuitum, R-P8, 193C2-l and R-P3805; five strains were related to M, phlgi, and five to R-P15. All of the Group III isolants from swine were closely related but not identical to M, ayipm, The relatively non- reactive filtrate of 350-1 reacted only with the avian anti- serum, and formed one precipitation band. The antigens of 186C-l developed four bands with M, gyigm antiserum, four with R—P8, three with R—PlS, two with lO7E—O, and one band each with M, bovis (Ravenel), M, tuberculosis (H37Ra) and R-P3803. Strain 193C2-l possessed three precipitating antigens. As described in the results of homologous and heterologous 72 reactions of selected mycobacteria, two of these antigens gave reactions of partial identity, and one antigen, a re- action of identity with antigens of M, avium. Unlike the Group III strains of bovine origin, the Group III swine isolants contained either the same number of antigens related to M, avium as to other reference strains, or they contained a greater number related to M, avium. Seven swine isolants were more closely related to M, avium than to the other 11 representative strains. One strain, 186C-l was related equally to M, avium and R-P8; strain 2441-1 was related equally to M, avium and lO7E-O; and 2421—1 was related more closely to another Group III swine isolant (193C —1) than to M, avium. 2 In general, the 22 Group III strains of both swine and bovine origin contained antigens that were more related to antigens of M, avium than to those of other representative strains. Onlytflmzfiltrate from the relatively non—reactive strain, SOB-O, failed to react with M, avium antiserum. Group IV. Reactions of seven Group IV strains (a ref- erence strain of human origin, R-PBBOs, and six strains of bovine origin) are shown in Table 11. Antigens of these strains failed to react with the antibodies elicited by 73 107E—O or 58A-l, and only some of the antigens of R-P3803 formed precipitates with the antibodies elicited by M, tuberculosis (H37Ra), M, ppygg (Ravenel), or M, fortuitum. The filtrate of R-PBBOs also developed one precipitation band in reactions with the antisera of M, gyipm, R—P8, R-PlS and 193C2-l. Only two Group IV isolants of bovine origin reacted with the antibodies elicited by R—P380s, and these two, Bll7B-O and B368D—l, each had but one antigen related to antigens possessed by R—PBBOs. Strain l4lF-l, the only bovine isolant in Group IV capable of inducing avian hypersensitivity or intradermal lesions in guinea pigs, formed three bands with M, gyipp_ antiserum and failed to react with other antisera. Even though M. avium was the only strain among the 12 represent— ative mycobacteria tovflfixfl114lF—l was related, 3 of the 7 strains in Group IV were not related to M, ayipm, Only one strain in the Group was related more closely to M. fortuitum than to the other reference strains. The filtrate prepared from this strain, B368D—l, developed three lines of precipitatiOn Mdth.the antiserum of M, fortuitum, and one line each with the antibodies induced by M, pplgi, R-P8, R-PlS, and R-P3BOs. 74 The number of precipitation lines formed by culture filtrates of Table 11. Group IV strains in reactions with reference antisera antiserum A050p£oopdommv Hu ,C: 0 w H m N m dfféftfnroror‘é‘gnr 2 2| Elm 2| 2| 2| m m a: 3 93 ad Group I From man 2 1 1 l 1 1 2 1 0 O 1 1 Group II From man 1 1 1 1 1 1 l 1 O O l 1 From cattle l O 0 1 1 l 1 0 O 0 1 From soil 3 O O 1 3 1 l 1 O 0 0 0 TOTALS 5 1 1 3 5 3 3 3 O 0 l 2 Group III From man 1 O O 1 0 O O O l 0 O O From cattle 9 4 4 8 5 6 6 5 6 6 From swine 12 8 8 12 8 10 10 10 1 1 9 8 TOTALS 22 12 12 21 l3 l6 16 15 3 4 15 14 Group IV From man 1 1 l 1 O 1 1 1 O O l 1 From cattle 6 0 O 3 4 2 4 3 l 0 TOTALS 7 1 1 4 4 3 5 4 l O l 3 Pseudochromes" From cattle 5 1 l 1 2 1 O 2 1 From swine 1 0 O l 1 l O l O O 0 TOTALS 6 l 1 2 4 4 2 3 1 0 2 1 TOTALS OF ALL ATYPICALS 58A-1 (pseudochrome) O OOH H Or—‘O Ol—l 81 identity or partial identity, nor was any consideration given to the number of precipitation lines formed in a re» action. The table presents then, an estimate of the rela— tionships of the groups of "atypicals" with selected myco- bacterial strains. It does not show the degree of the relationships, yet it accurately reflects the wide—range of antigens possessed by strains within each Runyon group. A number of striking comparisons are summarized in the Table: Of 41 atypicals of all groups and sources, 31 or 76%.were related to the classical avian bacillus and but 16, or 39%.were related to the mammalian strains. A surprisingly large number (27 to 28 strains) were related to M, phlei, M, fortuitum, R-P8, and R-P15. Only 3 to 5 "atypicals" were related to the Group III strains, R-P39 (of human origin), and the pseudochrome, 58A-l. In Table 15 the "atypicals” have been grouped with known mycobacterial strains with which they were most closely related antigenically. 82 ET gm 373$ AmvH-~mm Amvs-omm A3783 3:62: Aazhomfi Amvamoowfi Lax- om: 3705 $3.03 €3.39 grams A-:-m©om Eoémm €3.52...“ @30st £3.53 33.3on 3?? 33mm 3\£o-mom H: ITSx :7wa Trémx T:-mwom 3 Smé : E: 37m NE: 3-x H pofiflouucoz owmnmmwop 853.393 .2 51.3 fl a fl. wwmflhhmmllmwmww MW“: Gmmmmm .m a .m d .o .m 4.. Hmfimzo @330“ romofio v.83 mfimoarfim o5. £033 on magnum 550an m 30on Ha: ouomno 9» E 5505“. wqoam mmEmcoflmHou oficowficm .323 no woman mawmuum Hmoarfim mo wnadouwou < .mfi 3an. 83 Sat/minnow on” A; .>o3.mb\c.mw>.mn n\m .ozficflop unocu G int/m.” .m .GNSQEEMEH E ”mwa mocwdm 5 pomoaokwp >fi>33maomnom>£ m0 693 9.3. N .mfimfim Esoax v.38 no 02$ 0“ poumfimu >stvo one? ”:23 magnum mo pomomaoo 925 H QSOHOH A5 Huht. moaouao 3304.3 cm: -mmm 3?me 379,32 -owsomm Iméwmm T563522 3 TBS Emommmé 7:533 THAMES Z Tmmomm >H 033:“on III .II “I: I .II mwmogoumng fl @5050 poumfiou GoZ icoZ 8.93223 2 51AM E 8:15 2 pcmglm fl co>csm mm MAM .Q .U .m .< HmQHmCO p330.” 3o 03 who? mfimoarfim 9t £033 3 mammuum GBOGM . podcflaou I: .3 2an DISCUSSION It is well to bear in mind that the results observed could have been tempered by many factors. Employing con- centrated culture filtrates in diffusion reactions and for eliciting antibodies in rabbits is hardly ideal. Not only should one expect the various antigens present in a fil- trate to exist in varying concentrations, but also, the filtrate preparation itself to contain a wide variety of metabolic and autolytic products of bacterial cells. These products may have varying affinities for combining with the determinant groups of different antigens, thereby altering antigenic specificity or completely blocking the antigen. In this study no attempt was made to purify the fil- trates used to elicit antibodies in rabbits. waever, partial purification of M, tubercu10§i§_(H37Ra) filtrate by dialysis, and diffusion of the dialyzed product against Difco M, tuberculosis (H37Ra) antiserum showed that dialysis of this filtrate had little effect. Slightly sharper pre— cipitation lines were developed by the dialyzed filtrate, but the number of lines forming remained the same as before dialysis. Since no added information was gathered other culture filtrates were not dialyzed. 84 85 Another factor that may alter immunodiffusion results is the method employed to elicit antibodies. The low anti- genicity of the mycobacteria prompted the adoption of a procedure, for this study, that is a form of hyperimmuni- zation, Crowle (1961) reported that " . . . hyperimmuni— zation tends to induce production of a range of antibodies which make the antiserum likely to cross-react even with distantly related antigens." No direct evidence was sought in this study to show that the form of hyperimmunization employed did not cause non-specificty of the antibodies. HOwever, each of the 49 filtrates of mycobacterial strains reacted with at least one antiserum, and no single antiserum reacted with all of the 49 filtrates. The results, then, suggest that the anti- bodies possessed at least moderate specificities. Another factor influencing the quality of antisera-- and thereby, immunodiffusion results--is the variation of individual animals in response to injections of antigens. In this study, there was much variation in the number of antigens that were detectable when the same filtrate was diffused against antisera obtained from different rabbits that had been injected with identical filtrates. The anti— serum for each rabbit was tested individually; only those 86 which developed the largest number of precipitation bands in homologous reactions were used for studying the precipitino- genic relationships. A final limitation of immunodiffusion techniques is that during periods of incubation of a plate, precipitation lines are formed at varying rates and have different relative in— tensities; a dense line developing at a rapid rate may some- times mask the development of a slower—developing line. To minimize such effects, reactants were employed at varying relative concentrations, and observations were made at 8 hr intervals during incubation of the agar diffusion slides. When interpreting results, it is well to realize that shortcomings of the aforementioned sort are inherent in im- munodiffusion studies. In this study, the results of immunodiffusion tests of 49 concentrated culture filtrates with the 12 reference anti— sera suggest that marked antigenic differences exist among various strains of mycobacteria. These variances not only exist among strains of different origins and different Runyon groups, but also among strains isolated from similar sources with similar morphological and growth characteristics. Hew- ever, certain definite patterns of antigenic relationships were noted with many strains. 87 The observation of six identical precipitinogens in the Ravenel strain of M, bovis, and the H37Ra and H37Rv strains of M, tuberculggis, substantiates, in part, an earlier finding of Parlett and Youmans (1958). These authors found six identical precipitating antigens in H37Ra and H37Rv strains, but reported that only four of these anti- gens were possessed by the four strains of M, bovis which they studied. However, their study did not include the Ravenel strain. The non—pathogenic species, M, smegmatis and M. phlei, and the bovine and soil isolants that were avirulent for guinea pigs were not related to M, bgyi§_(Ravenel) or M, tuberculosis (H37Ra). If this pattern were consistent with avirulent strains only, the relationship would immediately have useful applications. But it was observed that 15 strains, virulent for guinea pigs (of bovine, swine, and cattle origins) were also unrelated to H37Ra and Ravenel. Yet, the criterion for determining virulence in guinea pigs had been the production of any size intradermal lesion at the site of inoculation, and not necessarily, progressive disease. If one assumed that the production of an intra— dermal lesion alone is an inadequate test for virulence, it would be of interest to set a more rigid criterion for 88 virulence and compare the relationships of the avirulent yg, the virulent atypical strains with the classical human and bovine tubercle bacilli. The atypical strains of soil, bovine, and porcine origins were, with but one exception, related more closely to one or more of the strains, M, ayigm, M, legi, and M, fortuitum, than to M, bgyi§_or M, tuberculosis. The strains in Runyon Group II and the majority of the pseudochromes were more closely related to M, phlei; the Group III strains of swine origin were invariably more closely related to M, ayigm; and the strains in Groups III and IV of bovine origin were re- lated more closely to either M, gyigm, M, legi, or M, fortuitum. The atypical isolants from man, however, were quite dif— ferent in their antigenic relationships. In general, they too, were related to M, ayigm, M, phlei, and M, fortuitum, but in addition, three of the five atypical human isolants each had two antigens related to_M, bovis and M, tuberculosis. This finding gives added support to the conviction that the atypicals of animal and soil origins markedly differ from the atypical human isolants. The nine bovine—skin isolants were either members of Runyon Groups III and IV, or they were pseudochromes. Though 89 somewhat related to M, ayigm, M, pglgi, and M, fortuitum, and R-P8, these strains were relatively non—reactive with the 12 reference antisera. Whether this could be due to a general lack of antigenicity of the skin isolants or to a lack of relationship to the reference strains, cannot be assessed as there was no antiserum prepared from a skin isolant or from M, ulcerans, a species associated with skin infections of man. It is noteworthy that this study, like a previous study by Parlett and Youmans (1958), failed to reveal that a com- mon precipitinogen was possessed by all strains of mycobac- teria. This finding, then, precludes the grouping of the strains on the basis of antigens possessed in common by more than one species of a certain genus. The shortcomings of other systems for classifying atyp- ical mycobacteria, in part, prompted this study. From the results presented, it is clear that the many variations in the biochemical, morphological, and growth characteristics of the "atypicals" are accompanied by a similar lack of harmony in the precipitinogenic relationships. The extreme adaptability and mutability of the mycobacteria continue to be apparent from all approaches of study. It is probable that a knowledge of the antigenic relationships will never, 90 in itself, solve all of the problems of classifying the mycobacteria. Such information,nevertheless, when eval- uated with data obtained from other approaches of study, may well provide for a more meaningful classification system than presently exists. The results reported in this study showed that "atyp- icals" of soil, porcine and bovine origins were related to M, phlei, M, fortuitum, and to the classical mammalian and avian strains. However, with few exceptions, each "atypical" was more closely related, antigenically, to one known strain than to others. On this basis alone, it has been possible to place the "atypicals" in six new groups (Table 13). The'hammalian-like" strains were placed in Group A; the "avian-like" strains, in Group B; the "phlei-like" strains, in Group C; the "fortuitum-like" strains, in Group D; the . strains related equally to two or more known strains, in Group E; and the strains non-related antigenically to the known strains, in Group F. Group A consisted of but three strains, all of mammalian origin (one bovine and two human), and each was from a dif- ferent Runyon Group (I, II, and III). Two strains were virulent and the other, avirulent for guinea pigs. The 91 three isolants each induced a different type of hypersensi- tivity in guinea pigs (mammalian, avian/battey, and non- definitive). Group B consisted of 18 strains: a single Group III human isolant, 2 Group III bovine isolants, 12 Group III swine isolants, 2 Group IV bovine isolants, and a pseudo- chrome of porcine origin. Of the 18 strains, 16 were vir- ulent and induced hypersensitivity in guinea pigs; 12 strains induced avian, 3 mammalian, and 1 induced avian/ battey' sensitivity. Group C consisted of 10 strains: 3 soil isolants and a bovine isolant from Runyon Group II; 2 bovine isolants from each of the Runyon Groups, III and IV; and 2 pseudo- chromes of bovine origin. Four strains in this group were virulent and caused hypersensitivity in guinea pigs, (2 in- duced avian and 2, non-definitive sensitivity). Group D consisted of 5 bovine isolants: 3 from Runyon Group III and 2 from Group IV. Only the 3 Group III strains were virulent and induced hypersensitivity in guinea pigs; one strain induced avian and the other two induced non- definitive sensitivity. Group E consisted of 3 strains: a Group III of bovine origin, a Group IV of human origin and a pseudochrome of 92 bovine origin. The 3 strains were virulent and induced hypersensitivity in guinea pigs; the pseudochrome induced avian and the other two, non-definitive sensitivity. Group F consisted of R—P4, a Group I strain of human origin, and 63A—O, a pseudochrome of bovine origin. Both strains were virulent and induced hypersensitivity in guinea pigs; the Group I isolant induced mammalian sensitivity, and the pseudochrome induced avian/battey sensitivity. The two strains were relatively non-reactive with the ref~ erence antisera; R-P4 was related only to R-P8, and 63A-O was related only to R-P39. This grouping is not presented to suggest that it would be either a practical or a meaningful system for classify— ing mycobacteria, but rather, to show the heterogeneity that exists within the atypical strains of varied origins. The strains placed in the various groups in Table 14 shows definite patterns of similarities in their morphological (Runyon Grouping), virulence, and hypersensitivity charac- teristics, but no clear-cut positive correlation of these characteristics with the new grouping is noted. It is clearly shown here that atypical isolants of cat- tle, swine, and soil origins have greater antigenic simi- larities with a known saprophyte (M, phlei), than has been 93 reported for atypical human isolants. In a study of 98 strains, primarily of human origin, Parlett and Youmans (1958) reported, "The clear-cut lack of antigenic similarity between the 'atypicals' and the saprophytic mycobacteria demonstrates that this simple double-diffusion method [Agar plate] will provide an easy method for the differentiation between these two groups." It is clear, and to be expected, that the above does not apply to "atypicals" of bovine, porcine, and soil ori- gins. In soil one should expect to find some saprophytes having characteristics that are identical to those of M, pglgi, and others with characteristics that are dissimilar in varying extents to those of M, phlei, Also, pathogenicity and virulence are such relative factors in infection, and are interrelated with host resistance and the size of in— oculum. It is highly conceiveable, then, that cattle and swine, in their natural habitats, could be exposed to such huge numbers of saprophyte-like mycobacteria that these could develop lesions. SUMMARY The precipitinogenic relationships of 49 representative mycobacteria of human, bovine, porcine, and soil origins were compared by use of an agar slide immunodiffusion tech- nique. Concentrated culture filtrate antigens from 49 strains were employed; 12 of them were used to prepare ref- erence antisera. The 12 strains included the classical path- ogens and saprophytes, and "atypicals" of human and animal origins. Antigenic dissimilarities exist between species and groups, and in addition, between strains within a group. No antigenic differences were observed between the six precip- itinogens possessed by each M, bgyis (Ravenel) and M, Egbgrr culosis (strains H37Rv and H37Ra). They were related to other known species as follows: one was identical to an antigen of M, ayigm; one was related and one identical to antigens of M, fortuitum; and none were related to M, phlei and M, smegmatis. In all tests, the homologous antigenic preparations and antisera reacted to give equal or greater numbers of precip- itate bands than did heterologous systems. The Group II strains of bovine and soil origins were closely related to M. phlei, but had no antigens in common 94 95 with M, tuberculosi§_or M, bovis. In contrast, a Group II strain of human origin possessed two antigens identical to those in the latter strains, and one antigen related to M, pilla- Very few of 22 strains in Group III (predominantly of bovine and porcine origins) were related antigenically to either of the Group III reference strains of human and bovine origins (58A-l). Eight of nine strains of bovine origin contained at least one antigen related to M, £3133, Only one of these eight, however, was more closely related to M, gyigm_than to other reference strains. In contrast, all Group III strains of swine origin were closely related, but not identical, to M, gyigm, The pseudochromes and the strains in Group IV were even more heterogeneous antigenically than the Group III organisms. Few Group IV strains and pseudochromes were related to.ML bovis and M, tuberculosis; most of them were closely related to M, phlei. A new grouping of "atypicals,' on the basis of their antigenic relationships to known mycobacteria, has been pre- sented. This grouping does not, however, solve the problem of classifying these organisms. BIBLIOGRAPHY Aladjem, F., and M. Lieberman. 1952. The antigen-antibody reaction. I. The influence of sodium chloride concen— tration (”1 the quantitative precipitin reaction. J. Immunol. ,§2:117-130. Bando, K. 1962. Specific agglutination of coloured col- lodion particles. [In Japanese, English summary]; Kitano Hosp. J. Med. Z;(2)93-100. Bechhold, H. 1905. Strukturbildung in Gallerten. Z. Phy- sik. Chem. 5g;185-199. Becker, E. L. 1953. Antigen-antibody reaction in gels. Federation Proc. l§§7l7-722. Becker, E. L., and J. Munoz. 1949. Multiplicity of anti- gens in ragweed pollen extract as demonstrated by the technique of Oudin. Proc. Soc. Exptl. Biol. Med. ,1;: 287-289. Bjorklund, B. '1952. Specific inhibition of precipitation as an aid in the antigen analysis with gel diffusion methods. Proc. Soc. Expt. Biol. Med. 12,319-324. Boyd, W. C. 1956. Fundamentals of Immunology. 3rd ed. Interscience Publ., New York. 776 pp. Boyden, S. V. 1951. The adsorption of proteins on erythrocytes treated with tannic acid and subsequent hemagglutination by antiprotein sera. J. Exptl. Med. 2;;107-120. Boyden, S. V. 1958. The immunological response to anti- gens of the tubercle bacillus. Progr. Allergy. 53149- 214. Breed, R. S., E. G. D. Murray, and N. R. Smith [Ed.]. 1957. Bergey's Manual of Determinative Bacteriology. 7th ed. Williams and Wilkins Co., Baltimore. pp. 696-707. 96 97 Buehler, E. V., and M. S. Rheins. 1959. Serologic rela— tionships of Old Tuberculin — sensitized sheep erythro- cytes and trypsinized human erythrocytes. Am. Rev. Tuberc. Pulmonary Dis. 225622—630. Burtin, P. 1959. Recherches immunochimiques sur 1es antigenes solubles du bacille tuberculeux. II. Etude immunochimique de deux polysaccharides presents dans 1es filtrats de culture du bacille tuberculeux. Ann. Inst. Pasteur. ,21:325-339. Burtin, P., and R. Kourilsky. 1959. Recherches immunol- ogiques sur les antigenes et haptenes solubles du bacille tuberculeux. I. Denommbrement et identification des constituants precipitant avec les immunserums specifiques. Ann. Inst. Pasteur. 215148-170. Carpenter, P. L. 1956. Immunology and Serology. Saunders Publ., Philadelphia. 351 pp. Chapman, J. S. [Ed.] 1960. The anonymous mycobacteria in human disease. Charles C. Thomas Co., Springfield, Illinois. 173 pp. Chen, T. H., and K. F. Meyer. 1955. Studies on the immu- nization against plague. X. Specific precipitation of Pasteurella pestis antigens and antibodies in gels. J. Immunol. 14;501-507. Coghill, R. D. 1931. The nucleic acid of the timothy hay bacillus. J. Biol. Chem. 29,57-63. Colwell, R. R. 1963. Correlation between Adansonian anal- ysis and DNA base composition in bacterial classification. Proceedings of the 63rd Annual Meeting of the Am. Soc. for Microbiol., Cleveland, Ohio. Creighton, M. M., and R. J. Anderson. 1944. The chemistry of the lipids of the tubercle bacillus: LXVIII. The lipids of cell residues from the preparation of tuber- culin. J. Biol. Chem. 154:581-585. 98 Creighton, M. M., L. H. Chang, and R. J. Anderson. 1944. The chemistry of the lipids of the tubercle bacillus. J. Biol. Chem. 154:569-580. Crowle, A. J. 1958a. A simplified micro double—diffusion agar precipitin technique. J. Lab. Clin. Med. 5g;784- 787. Crowle, A. J. 1958b. Enhancement by cadmium of double- diffusion precipitin reactions. J. Immunol. 81,194- 198. Crowle, A. J. 1960. Four modifications of the micro agar diffusion precipitiation test. J. Lab. Clin. Med. 555 593-604. Crowle, A. J. 1961. Immunodiffusion. Academic Press. New York. 333 pp. Darzins, E. 1958. The bacteriology of tuberculosis. Uni— versity of Minnesota Press, Minneapolis. 488 pp. Elek, S. D. 1948. The recognition of toxicogenic bacterial strains ig_vitro. Brit. Med. J. 1,493-497. Fregnan, G. B., D. W. Smith, and H. M. Randall. 1961. Bio— logical and chemical studies on mycobacteria. Relation- ships of colony morphology to mycoside content for Myco- bacterium kansasii and Mycobacterium fortuitum. J. Bacteriol. 825517-527. Freund, J. 1951. The effect of paraffin oil and mycobac- teria on antibody formation and sensitization. Am J. Clin. Pathol. 21,645-658. Freund, J., and M. V. Bonanto. 1944. The effect of paraf— fin oil, lanolin—like substances, and killed tubercle bacilli on immunization with diphtheria toxoid and Bag: terium typhosum. J. Immunol. 48,325-334. Furth, J. 1926. On the serological relationship of acid- fast bacteria. J. Immunol. 12,273-292. 99 Grabar, P. 1959. Immunoelectrophoretic analysis. Methods Biochem. Anal. I1:1-38. Grabar, P., and C. A. Williams. 1955. Méthode immuno- electrophorétique d'analyse de melanges de substances antigeniques. Biochim. Biophys. acta. 11367—74. Guy, L. R., and S. Bernard. 1960. Laboratory techniques for the differentiation of mycobacteria. pp. 21-30 in J. S. Chapman [Ed.], Anonymous mycobacteria in human disease. Charles C. Thomas Co., Springfield, Illinois. Heidelberger, M., and A. E. O. Menzel. 1934. Protein fractions of the human strain (H-37) of tubercle ba- cillus. J. Biol. Chem. 104:655-665. Heidelberger, M., and A. E. O. Menzel. 1937. Specific and non-specific cell polysaccharides of a human strain of tubercle bacillus, H—37. J. Biol. Chem. 118:79-100. Jennings, R. K. 1953. Antigens in scarlatinal erythogenic toxin demonstrable by the Oudin Technique. J. Immunol. 19:181-186. Johnson, W. H., L. A. Balsden, and A. H. Frank. 1961. Bovine tuberculosis. J. Am. Vet. Med. Assoc. 138: 239-243. Kniker, W. T., and D. C. Heiner. 1960. The separation of antigens of Mycobacterium tuberculosis by ion-exchange chromatography and gel diffusion (Abstr.). Am. J. Dis. Child. 199,604-605. Kniker, W. T., and J. B. LaBorde. 1962. Isolation of pure mycobacterial antigens (Abstr.). Annual Meeting Ab- stracts, National Tuberculosis Association. New York. pp. 5—6. Kraus, R. 1897. Wiener Klin. WOchenschr. 19,736. (As cited by Smith, Martin gt $1., 1948). Mallmann, V. H., W. L. Mallmann, and P. Robinson. 1961. Relationship of atypical bovine and porcine mycobac- teria to those of human origin. Mich. State Univ., East Lansing, Michigan. In press. 100 Mallmann, V. H. and P. Robinson. 1961, 1962. Identifica— tion studies of acid-fast isolants. Mich. State Univ., East Lansing, Michigan. Unpublished data. Mansi, W. 1958. Slide gel precipitin tests. Nature 181: 1289. Meynell, G. G. 1954. The antigenic structure of Mycobac- terium tuberculosis var. hominis. J. Pathol. Bacteriol. 61:137-150. Middlebrook, G., and R. J. Dubos. 1948. The specific serum agglutination of erythrocytes sensitized with extracts of tubercle bacilli. J. Exptl. Med. 88:521-528. Munoz, J., and E. L. Becker. 1950. Antigen-antibody re- actions in agar. I. Complexity of antigen-antibody systems as demonstrated by a serum-agar technique. J. Immunol. ‘§§:47-58. Negré, L. 1947. Etude de la virulence des Bacilles Tuberculeux Jeunes. Ann. Inst. Pasteur 135713-724. Ouchterlony, O. 1948. MM vitro method for testing the toxin-producing capacity of diphtheria bacteria. Acta. Pathol. Microbiol. Scand. 26:516-524. Ouchterlony, O. 1958. Diffusion-in-gel methods for im- munological analysis. Progr. Allergy 551-78. Oudin, J. 1946. Méthode d'analyse immunochimique par precipitation specifique en milieu gélifié. Comp. rend. Acad. d. sc. 222:115-116. Oudin, J. 1948. L'analyse immunochimique qualitative; méthode par diffusion des antigénes au sein de l,im— muneserum precipitant gélosé. Ann. Inst. Pasteur 1;:30-51. Oudin, J. 1952. Homogeneity of proteins and polysaccharides: agar diffusion techniques. Methods Med. Res. 5:335-378. Parlett, R. C. 1961. Antigenic analysis and antibody ti- tration by gel double—diffusion techniques. Symposium on Research in Leprosy, Leonard wood Memorial (Am. Leprosy Foundation), Washington, D. C., pp. 64-69. lOl Parlett, R. C., and G. P. Youmans. 1956. Antigenic rela- tionships between mycobacteria as determined by agar— diffusion precipitin techniques. Am. Rev. Tuberc. Pulmonary Dis. 11:637-649. Parlett, R. C., and G. P. Youmans. 1958. Antigenic re1a~ tionships between 98 strains of mycobacteria using gel- diffusion precipitation techniques. Am. Rev. Tuberc. Pulmonary Dis. 11:450—461. Parlett, R. C., and G. P. Youmans. 1959. An evaluation of the specificity and sensitivity of a gel double-diffusion test for tuberculosis. Am. Rev. Resp. Dis. 89:153-166. Pepys, J., R. Augustin, and A. B. Paterson. 1959. Common antigenic components of mycobacterial extracts. Tu- bercle 49:163-172. Preer, J. R. 1956. A quantitative study of a technique of double-diffusion in agar. J. Immunol. 11:52-60. Ranney, A. F. 1961. Bovine tuberculosis eradication. Dis. of the Chest 3251-7. Rheins, M., R. G. Burrell, and J. M. Birkeland. 1956. Tuberculous antibodies demonstrated by agar diffusion: I. Specificity and incidence of agar-diffusion antibodies in rabbit sera. Am. Rev. Tuberc. 14:299-238. Runyon, E. H. 1960. Anonymous mycobacteria in human disease. pp. 3-12 in J. s. Chapman [Ed.] Anonymous mycobacteria in human disease. Charles C. Thomas Co., Springfield, Illinois. Schaefer, W. B. 1940. Recherches sur la spécifité 1es protéides des bacillus tuberculeux. Identification sérologique des souches bovines et sérodiagnostique de la tuberculose bovine chez 1'homme. Ann. Inst. Pasteur 64:517-529. Schaefer, W. B. 1947. Identification sérologique des bacilles tuberculeux du type bovin. Ann. Inst. Pasteur 11:1157-1165. 102 Seibert, F. B. 1944. The chemistry of tuberculin. Chem. Rev. 84:107-127. Seibert, F. B. 1949. The isolation of three different pro- teins and two polysaccharides from tuberculin by alcohol fractionation. Their chemical and biological properties. Am. Rev. Tuberc. 88:864101. Seibert, F. B. 1958. The interplay of an immune substance with tuberculo-polysaccharide and its antibody in tuberculosis. J. Infect. Dis. 103:52-60. Seibert, F. B., C. Crumb, and M. Seibert. 1950. The com- parison of the filtrate fractions of certain bacillary extracts of the H37, H37Rv, and H37Ra strains of tubercle bacilli. J. Am. Chem. Soc. 18:2678. Seibert, F. B., and E. Soto-Figueroa. 1957. Study of tuber- culin protein and polysaccharide antigens by gel—diffusion techniques. Am. Rev. Tuberc. Pulmonary Dis. 18:601-607. Smith, D. T., D. S. Martin, N. F. Conant, J. W. Beard, G. Taylor, H. J. Kohn, and M. A. Poston. 1948. Zinsser's textbook of bacteriology, 9th ed. Appleton-Century- Crofts Publ., New York. 992 pp. Sushida, K., K. Hirano, M. Mikami, C. Koyama, and H. Takeuchi. 1961. Demonstration of antibody in the serum of tuber- culosflspatients by the Ouchterlony method (Abstr.). Am. Rev. Resp. Dis. 84:160. Telfer, W. H. 1953. Antigenic analysis of insect blood. Federation Proc. 18:734-738. Toda, Y. 1956. On the isolation of serologically active proteins from the tubercle bacillus. Ann,Tuberc. Z; 1-13. Vardeman, T. H., and A. B. Larsen. 1961. Bovine tuber- culosis - studies of complement - fixation antigens. Am. J. Vet. Res. 88:204-208. 103 wong, S., and J. weinzirl. 1936. An inexpensive medium for the growth of Mycobacterium tuberculosis. Am. Rev. Tuberc. 885577-583. Xalabarder, C. 1961. The so—called problem of unclassified mycobacteria. A review. Am. Rev. Respir. Dis. 8:1-15. \ "\ " ' , v ‘ ‘ x . 5 r“ > .. ‘ L ~, " ' ' l ‘ ‘ N ,, . - . y . 1 12.39 '17" ‘l g "5;.“ ’ :. :‘r" L: ré-mfw -- T: I" 1V4" 4 1'. - - I H TY "7n{11117111111311(Edujfyfflmmlljffllmfl"5