The microstructure of colonies of the Connaught BCG strain*

Bull. Org. mond. Sante} 1970, 43, 699-705Bull. Wid Hlthi Org. The Microstructure of Colonies of the Connaught BCG Strain* L. gULA 1 It has previously been shown that there is a certain correlation between the biological properties of BCG strains-properties on which their immunogenicity and allergenicity depend-and the macroscopic appearance of the growth of these colonies on solid and liquid media. To investigate this phenomenon in greater detail, the author examined the microstructure of colonies of the Connaught BCG strain grown on both solid and liquid media. Colonies were fixed in agar, embedded in paraffin, sectioned and stained by the Ziehl-Neelsen technique. A striking finding was the alternation of acid-fast and non-acid-fast zones in colonies grown on bovine-serum agar or Ogawa egg medium; the strata nearest the surface of the solid media were usually more acid-fast than were the deeper strata. Colonies grown in gula's liquid medium, on the other hand, showed no such stratification and were equally acid-fast at all points. These differences may be the result of genetic factors or of the different nutritional conditions provided by solid and liquid media. The BCG strain developed many years ago by Calmette and Guerin from a virulent strain of Mycobacterium bovis originally isolated by Nocard in 1902 from the udder of a tuberculous cow is still the best material for the production of a potent antituberculosis vaccine. But the only confirmation of Calmette's hypothesis that his strain was a virus fixe is that, after many thousands of passages on artificial media in laboratories all over the world, its virulence is attenuated and it has under- gone such a profound genotypical change that it is highly unlikely to revert to its original virulence for laboratory animals or man. On the other hand, the degree of residual virulence and certain general biological characteristics of BCG strains-i.e., their microscopical morphology, immunogenicity, allergenicity, and nutritional re- quirements-differ greatly from one strain to another, as Guld et al. (1968) recently demonstrated again. As the biological properties may be connected with the morphological pattern of growth on solid or liquid media, the author examined the micro- structure of the colonies of BCG and MP strains * This study was supported by a grant from the World Health Organization. ' Chief, WHO International Centre for the Diagnosis of Tuberculosis, Tuberculosis Research Institute, Srobirova 48, Prague 10, Czechoslovakia. used for the production of antituberculosis vaccines in Prague. The results obtained so far have shown considerable variations in the microscopical struc- ture of the colonies grown on Lowenstein-Jensen, Ogawa or agar media (gula, 1969). Among the vaccine strains under study, the BCG strain used by the Connaught Laboratories in Toronto for vaccine production showed the most interesting microscopical growth pattern. The microstructure of the colonies of this strain is described in the present paper. MATERIAL AND METHODS A strain of BCG from Connaught Laboratories, Toronto (lyophilized BCG vaccine, Batch No. 192-1), which will be referred to throughout this paper as the Connaught-Toronto strain, was cultured on Ogawa medium and bovine-serum-agar medium in Petri dishes. The latter was prepared by mixing 100 ml of concentrated gula's bovine serum liquid medium (gula, 1963) with 1.5% Difco agar base. The cultures were incubated at 37°C for 1 month. A second series of experiments was made with the same strain maintained by successive monthly passages for 2 years on Sula's liquid medium (Connaught-Prague strain). These strains were also cultured on Sauton medium for 10 days so as 2593 - 699 L. §ULA to obtain a bacterial mass large enough to allow their lipid content to be measured and correlated with their acid-fastness. For this purpose, the culture was filtered and the resulting bacterial mass was dried to constant weight under P205 in an exsiccator and extracted in a Soxhlet apparatus, first with alcohol for 24 hours and then with ether for a further 24 hours. Sections of the colonies grown on Ogawa and bovine-serum agar were prepared for microscopical examination, using the following technique. The cultures in the Petri dishes were flooded with a 10% formol solution for 2 hours. The solu- tion was then poured off and the cultures were washed with saline. After drying, blocks of agar and Ogawa medium with the selected colonies were cut out. In the meantime, warm 2% agar had been prepared in other Petri dishes into which the blocks were placed and immediately covered with liquefied agar so as to obtain complete adhesion of the agar to the blocks. After solidification, the agar blocks were cut out and treated like any other histological specimens, i.e., subjected to dehydration with 5 x 96% alcohol, acetone, benzine (petroleum ether) and finally embedded in paraffin. In order to investigate the damaging effects of organic sol- vents on acid-fastness, some sections were examined in gelatin. As no substantial differences were found between the staining properties of the paraffin and gelatin preparations, the former technique was used throughout the whole study. All sections were stained by the Ziehl-Neelsen, haematoxylin-eosin, Gram, Hallberg and Nyka techniques (Hallberg, 1941; Nyka, 1963). The colonies grown in submerged culture in Wula's liquid medium were similarly treated for the pre- paration of sections, by pipetting off the medium and embedding the granular colonies in liquefied agar. The sections were examined as described above. RESULTS This paper presents only the results obtained with the Ziehl-Neelsen staining technique on colonies cultured from the Connaught-Toronto strain. On the whole no substantial differences were observed between the microscopical appearance of the colonies grown on Ogawa egg medium and on bovine-serum agar. This was equally true for the Connaught-Toronto and Connaught-Prague strains. They produced a typical granular deposit after 1 month's incubation at 37°C. Therefore the in vitro environment presented by the facultative anae- robic conditions prevailing in the depths of the liquid medium did not seem to have affected the growth pattern of the Connaught-Prague strain. Furthermore, there were no substantial differences in the macroscopical appearance of the growth of these strains cultivated on Sauton medium. The surface pellicle showed a texture of eugonic growth typical of other BCG strains also (Fig. 1). A typical colony on Ogawa medium is pale orange, large (1 cm-2 cm in diameter), flat, and roughly circular with a central nipple surrounded by a smooth or firmly granular zone from which raised serpentine wrinkles radiate outwards to the edge of the colony, which is again smooth or granular (Fig. 2). In some colonies the wrinkles are closely attached to the central knob (Fig. 3). In others, the whole centre was smooth and granular with a finely lobulated border zone. The line of demarcation between the medium and the outer edges of the colonies was not clearly dis- cernible because of a lack of orange pigment in the young actively growing cells on the periphery of the colonies. Apart from the colonies spread out on the sur- face of the medium, there were small, smooth, glistening, dome-like colonies with narrow borders standing out distinctly against the medium. One of these small protuberances is depicted on the lower edge of the central colony in Fig. 3. The flat colonies were firmly fixed to the surface of the medium and were not easily removed with a bacteriological loop, whereas the small dome-shaped colonies were easily removable. Sections of the colonies showed no deep growth into the substrate of Ogawa medium. The growth pattern on bovine-serum agar differed according to the size of the inoculum. With small inocula (10-5 mg-10-6 mg) colonies of two types were found. First, there were small colonies with a smooth surface and no particular structure that was visible macroscopically. These colonies presumably grow from the individual rods contained in the inoculum. Secondly, there were larger colonies which could have grown from larger clumps. These colonies had a very picturesque texture, their surface consisting of thick, roughly interwoven wrinkles separated by deep clefts. Again, as in the colonies grown on Ogawa medium, there was a central protuberance from which the wrinkles radiated to the outer edge (Fig. 4). This peculiar surface texture results from dense cording of the bacilli. When removed from the substrate, some of the colonies were like empty shells of which the surface was a network of 700 MICROSTRUCTURE OF COLONIES OF CONNAUGHT BCG STRAIN firmly interwoven, serpentine-like cords; others with the same surface structure had a solid granular centre. Again, as on Ogawa medium, no subsurface growth was observed on the agar medium. When a large inoculum was used for planting on both Ogawa and bovine-serum agar in Petri dishes, a flat lacy pellicle with sharply defined borders was formed on the surface (Fig. 5). As in the isolated colonies grown on Ogawa egg medium, protuberances and rough and fine wrinkles also were present. The structure of the colonies grown in the depths of gula's liquid medium resembled that of virulent mycobacteria. Large inocula (1 mg wet weight of bacterial mass) produced pellicles with typical wrinkles and folds both on the surface and deep under it, and produced a yellow discoloration of the medium. Small inocula (10-5-10-6 mg) grew in isolated granular colonies, leaving the medium entirely clear and colourless. However, the most peculiar pattern was observed on microscopical examination of sections of the colonies, according to the culture medium used. Bovine-serum agar The most important phenomenon observed in the colonies grown on this medium was the zonal growth composed of acid-fast and non-acid-fast layers. When the colony illustrated in Fig. 4 and 5 was studied in sections, non-acid-fast layers were seen in the central part of the colony and the peri- phery was covered by a thin layer of acid-fast bacilli (Fig. 6). Fig. 7 shows part of the acid-fast and non-acid-fast zone, containing many acid-fast rods, both isolated and interwoven in various ways, as well as acid-fast granules. All these acid-fast elements are embedded in a homogeneous non- acid-fast mass, called by French authors substance cyanophile (Calmette, 1936). The same zonal phenomenon is demonstrated more clearly in the small colonies growing in iso- lation on the bovine-serum agar (Fig. 8) which have a thin, strongly acid-fast layer adjacent to the surface of the medium. In the middle of the colony is a non-acid-fast core surrounded by an acid-fast zone which, in turn, is separated from a weakly acid-fast zone by a narrow colourless band. A similar stratification of acid-fast and non-acid-fast layers is seen in the upper part of the colony. With greater magnification, the internal non-acid-fast layers are seen to be mostly of granular texture, the external acid-fast strata being composed of acid-fast rods and granules. Fig. 9 shows the vertical section of three interlinked colonies of which one colony has a central non-acid-fast core and the two others are almost completely non-acid-fast. A higher magni- fication of the first more acid-fast colony is given in Fig. 10. The microscopical picture of the lacy surface pellicule grown on agar medium using a large inoculum, which is demonstrated macroscopically in Fig. 5, is also very interesting. The colony is not firmly attached to the surface of the medium but is loosely connected with it by means of irregular short stems resembling the rim of a cog-wheel (Fig. 11). At a higher magnification, the base of the cogs is sharply demarcated from the central part of the non-acid-fast membrane by a thin blue line. As in the previous pictures, the inner parts of the colony consist mostly of a non-acid-fast homogeneous mass permeated with loosely connected acid-fast bacilli (Fig. 12). Ogawa egg medium A similar stratification of acid-fast and non-acid- fast zones was observed on vertical section of the nippled colonies grown on Ogawa egg medium. The cogs and blue line are not as clearly discernible as in Fig. 12, but the division between acid-fast and non-acid-fast strata is again well demarcated (Fig. 13). It is interesting to note that the wrinkles separating the central and peripheral portions of the colony grow out of the non-acid-fast upper stratum of the colony and contain only isolated strands of acid-fast rods. The central part of the wrinkle is almost free from acid-fast rods and granules (Fig. 14). gula's liquid medium The most conspicuous difference in the micro- scopical texture disclosed in the sections of the sur- face pellicle and in the granular colonies grown in the depth of Sula's liquid medium is the complete absence of zonal stratification. On solid Ogawa or bovine-serum agar the wrinkles are mostly composed of a non-acid-fast substance, except for the parts which have direct contact with the medium. On the other hand, the wrinkles found on the surface of the pellicle grown on Sula's liquid medium are completely acid-fast, being composed of fully deve- loped acid-fast bacilli. The same phenomenon was observed in sections of granular colonies from a deep growth of myco- bacteria in the liquid medium. These colonies (illustrated in Fig. 15 and 16) are composed only 701 L. §ULA FIG. 1. Surface pellicle of 1-month-old Connaught BCG strain grown on Sauton medium. Typical crumpled membrane rising up the walls of the culture flask. FIG. 2. One-month-old, nippled, spreading colony grown on Ogawa medium. Finely granular centre surrounded by a circle of wrinkles (x 2). FIG. 3. One-month-old, flat, spreading, smooth colonies grown on Ogawa medium. On the periphery of the central wrinkled colony, one small, smooth, secondary non-spreading colony (x 2). FIG. 4. Non-spreading, well-circumscribed, 6-week-old colonies. The texture of the colonies grown on bovine-serum agar resembled the eugonic pellicle on Sauton medium (x 2.2). FIG. 5. Lacy pellicle, 6- weeks-old, grown on bovine-serum agar (x 2.2). FIG. 6. Vertical section of the colony depicted in Fig. 4, stained by the Ziehl-Neelsen technique. The greater part of the colony is non- acid-fast (x 20). In the original microphotograph in colour, the black parts of the colony are shown to be acid-fast and the white parts non-acid-fast. This applies also to Fig. 7 and Fig. 8. FIG. 7. Same colony as in Fig. 6, magnified 200 x. The section contains typical acid-fast bacilli and granules embedded in substance cyanophile, which forms the greater part of the colony. FIG. 8. Vertical section of a 1-month-old colony grown on agar medium. Typical acid-fast and non-acid-fast strata (x 22). FIG. 9. Vertical section of two non-acid-fast colonies and one acid-fast colony with non-acid-fast centre (x 14). One-month-old culture grown on bovine-serum agar. FIG. 10. Acid-fast colony from Fig. 9, magnified 27 x. The central non- acid-fast core, formed by substance cyanophile, is scattered with very fine acid-fast granules. FIG. 11. Vertical section through the basal parts of the colony depicted in Fig. 4. The colony is loosely connected with the surface layers of the agar medium by means of strongly acid-fast cogs (x 32). FIG. 12. The same section as in Fig. 11, magnified 150 x. Between the cogs, a blue line showing where the production of substance cyanophile begins. FIG. 13. Vertical section through a 1-month-old, spreading colony grown on Ogawa medium (x 77). The colony consists of two different layers, of which the lower is formed by fully acid-fast strands, the upper one being non-acid-fast and scattered with individual rods, granules and small strands of acid-fast bacilli. FIG. 14. Details of the section depicted in Fig. 13. The section is made on the borders between the flat, fine granular layer and the base of the wrinkles, which are very poor in acid-fast rods (x 32). FIG. 15. Section through a 1-month-old granular colony grown in gula's liquid medium (x 32). The colony consists of loosely connected, densely packed cords. Non-acid-fast forms are not found. FIG. 16. Same section as in Fig. 15, magnified 94 x. Characteristic cords of varying width and length composed of fully acid-fast granular rods. 702 I2 3 1*.s''w.........a - ;t 7 4', 1 1 4 6 8 :P *9 .. i: :;t i.:- : W ! d,-Y, 4~~~~~~~~~~~~~~~~~~~~~~ g 2~~~~~~~~~~~~~~~~~~ 13 14 ....6.. t:.: M -T ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ V .~~. . 1 MICROSTRUCTURE OF COLONIES OF CONNAUGHT BCG STRAIN of acid-fast bacilli mostly longitudinally oriented into the cords which are typical of fully virulent myco- bacteria. Lipid content of the Connaught-Toronto strain As acid-fastness is directly correlated with lipid content, the bacterial mass of this strain obtained after 12, 16 and 30 days of incubation at 37°C was investigated, with alcohol and ether as organic solvents. Extraction was carried out in the Soxhlet apparatus for 2x 24 hours. The results are shown in the table. For comparison, equivalent data are given in the lower part of the table for a strain isolated in Prague in 1947 from the Danish BCG vaccine batch 725 and cultured and analysed in the same manner. FATS AND PHOSPHATIDES CONTENT OF CONNAUGHT-TORONTO AND DANISH a BCG STRAINS Time of incubation Fats Phosphatides Total(days) (%) (%) (%) Connaught-Toronto BCG strain 12 21.69 0.34 22.03 16 21.51 0.46 21.97 30 26.83 0.21 27.04 Danish BCG strain a 7 25.01 1.36 26.37 14 30.28 1.73 32.01 21 36.21 3.59 39.80 30 36.25 1.90 38.15 u Strain isolated from Danish vaccine batch No. 725. It is evident that the major part of the lipids is produced as early as the first week of incubation. Further, the content of phosphatides dissolved in alcohol does not substantially increase during further incubation. Strangely enough, the production of lipids in deep cultures of the Connaught-Toronto strain is very low in spite of the fact that such bacilli are fully acid-fast. DISCUSSION The morphological characteristics of different BCG strains have been studied extensively by Pierce & Dubos (1956). Basically two different kinds of growth-spreading and non-spreading colonies- were observed in the cultures of these BCG strains on albumin-oleic-acid agar or on norite agar. The spreading colonies resembled those of virulent mycobacteria. The avirulent strains produced mostly non-spreading colonies. The proportions of these two forms did not change under standard conditions of growth, but when a surface-active substance such as Tween 80 was added to the medium the proportion of non-spreading colonies increased considerably. The amount of Tween 80 necessary for changing the characteristics of the colonies from spreading to non-spreading forms was in direct proportion to the degree of attenuated virulence of the BCG strains under study. Other studies on the structure of mycobacterial colonies, including those of four different BCG strains, were carried out by Mizuno (1953). On studying the growth of the colonies on cover- slips under the microscope, he found certain diffe- rences between the BCG strains, particularly with regard to the textures of the cords, some ofthese being very short and thick, and others varying remarkably in breadth and length. The non-sectioned colonies were stained with heated Fonte carbol crystal violet, processed further with Lugol's iodine solu- tion, and decolorized with 2% hydrochloric acid. No reference was made by Mizuno to any differences in the staining properties of the BCG strains he studied. The microscopical texture of the sectioned spread- ing and non-spreading BCG colonies was not investigated by either Mizuno or Pierce & Dubos, who limited their observations to the well-known fact that non-spreading colonies do not form cords. The present study on the microscopical structure of the mycobacterial colonies is the continuation of research work undertaken by previous workers (Kahn & Nonidez, 1936) who investigated the micro- scopical structure of different mycobacteria such as H37Rv, Myco. bovis, Myco. avium, Myco. smegma- tis, etc. BCG strains were not included in these experiments. Thus our present studies on the microscopical structure of the Connaught-Toronto strain are the first to demonstrate profound morphological and staining differences between the spreading and non-spreading colonies of this BCG strain. The metabolic process resulting in the differently stained microscopical zones is not known. The explanation is advanced here that, in the spreading colonies grown on Ogawa medium, the capacity of the strain 5 703 L. §ULA to synthesize the lipid substances responsible for acid-fastness has not been fully developed. This may be due to a deficiency in the enzymatic systems responsible for the synthesis of lipids, or because the nutritional substances supplied by Ogawa medium were insufficient for the BCG strain to develop the full synthetic capacity necessary for the production of surface lipids. The second hypo- thesis seems to be the more likely as the surface and deep cultures in gula's liquid medium supplied only fully acid-fast bacilli. This indicates that the liquid medium allows more efficient synthesis of all meta- bolites for actively growing cells. On the solid medium, the nutritional and metabolic conditions are different. First, the bacilli which are in intimate contact with the medium have much better conditions for growth. They have better access to the nutri- tional substances contained in the medium than the cells growing over them, to which the nutrients must be transmitted by diffusion through the basal layers of the colony having direct contact with the medium. Also, the provision of free space to the growing cells is very important for their morphology. Colonies grown in the solid medium contain only short coccobacilli, whereas the same strain grown in the liquid medium produces long, bent, granular rods typical of the mycobacteria encountered in sputum or pathological lesions. As growth on the solid medium continues, the distance between the surface layers of medium and upper parts of the colonies is increased and the nutrition of the actively growing bacilli becomes even more difficult. In this way the active multiplication of the cells and the synthesis of acid-fast substances are delayed, and the viability of the bacilli is diminished. Some may even die before reaching maturity. Thus it seems probable that, for growth to continue, some substances released after the death of a certain number of bacilli are again utilized for newly growing mycobacteria which otherwise would lack nutrients from the medium. This might explain the peculiar stratification of colonies, which display acid-fast and non-acid-fast zones such as are never seen in the small granular colonies grown in the depth or on the surface of the liquid medium. Once the sub- stances necessary for the multiplication of myco- bacteria have accumulated, the colony starts to grow again, forming a new acid-fast layer on top of the previously formed non-acid-fast layer. But even when the medium contains all the nutrients necessary for mycobacterial growth, their utilization depends on the enzymatic equipment of the mycobacteria and on the structure of their surface directly exposed to the medium. The cell walls do not play an active part in the transport of the nutrients from the medium, which proceeds passively through the external surface layers of the cells, directed only by the physical laws of diffusion (Mandelstam & McGuillen, 1968). But their further transport through the cell membrane to the interior of the cytoplasm is a highly selective process requiring the energy provided by certain fatty sub- stances in which the amino acids are incorporated. Only this compound is capable of passing through the cell membrane (Ames, 1968; Gitores & Roth, 1968). Another explanation of the stratification pheno- menon might be the damaging effects of organic solvents (alcohol, benzine, xylol) used for making the sections, which might affect young actively growing and antigenically not fully endowed cells in different ways. This possibility cannot be complete- ly excluded, although, in the Connaught-Toronto BCG strain, the same stratification phenomenon was observed with the modified technique described in this paper (i.e., fixing the colonies in gelatin, without using any organic solvents) as with the method in which the colonies are embedded in paraffin. As regards the non-acid-fast colonies, this phenomenon might be explained by the hypo- thesis that some genetic defect occurs in certain BCG cells. Owing to the lack of the respective operons, certain fatty substances responsible for acid-fastness are not produced. No such colonies have yet been discovered in liquid media. Further- more, the surface pellicle grown on Sauton medium inoculated with the Connaught-Toronto strain was fully acid-fast and did not show any stratification when stained by the Ziehl-Neelsen technique. How- ever, the lipid content is much lower in the Connaught strain than in the Danish BCG strain, as is shown in the table. It remains to be investigated whether the dif- ferences encountered in the growth and lipid con- tent are specific for this strain or whether they are characteristic of other BCG strains as well. Some preliminary observations made on the current strain used for BCG vaccine production in Copenhagen, which is commonly recognized as an immuno- genically very potent strain, seem to indicate that this behaves differently and that colonies of this strain grown on Ogawa or bovine-serum-agar medium are composed mostly of fully acid-fast layers con- taining only acid-fast bacilli. 704 MICROSTRUCTURE OF COLONIES OF CONNAUGHT BCG STRAIN 705 Thus the current systematic investigation of the microscopical structure of colonies of different BCG strains may lead to a better understanding of their metabolic pattern and hence to the selection of the immunogenically most potent strains, con- ferring a high degree of immunity to tuberculosis. ACKNOWLEDGEMENTS Thanks are expressed to Dr S. Landi, Head of the Tuberculosis Department, Connaught Medical Research Laboratories, University of Toronto, Toronto, Canada, for providing the Connaught BCG strain; to Mr A. Winter for the photographs; and to Mrs Z. Cvejnova, of the Tuberculosis Research Institute, Prague, for her technical assistance. RtSUMti MICROSTRUCTURE DES COLONIES DE LA SOUCHE CONNAUGHT DE BCG La notion d'une correlation entre les proprietes bio- logiques des souches de BCG et les aspects morpholo- giques de leur croissance sur milieu liquide ou solide a incite a etudier plus en dMtail la microstructure des colonies d'une souche de BCG. Une souche fournie par les Laboratoires Connaught, de Toronto (Canada), a 6te mise en culture de surface sur milieu d'Ogawa a l'euf et sur gelose au serum bovin ainsi qu'en culture profonde en milieu liquide de gula. Apres un mois d'incubation a 37°C, on a pr6lev6 des tranches de milieu solide portant les colonies destinees a etre examin6es, on les a incluses dans la g6lose liquide a 2% puis enrob6es dans la paraffine. Les colonies obtenues en milieu liquide ont subi le meme traitement apres elimination a la pipette de l'exces de liquide. Les coupes ont ete etudiees apres coloration par la technique de Ziehl-Neelsen. Sur milieu d'Ogawa, on obtient des colonies plates, plus ou moins circulaires, presentant une protuberance centrale, une zone interne lisse ou granulaire et une zone externe a surface ridee. Sur gelose, la souche donne des colonies surelevees, avec une protuberance centrale et une surface rugueuse formee d'epais cordons separes par de profondes scissures. Sur milieu de Sula, la croissance du BCG s'opere, selon la taille de l'inoculum, sous la forme de petits amas granulaires ou de membranes. A l'examen microscopique, le fait le plus remarquable est l'altemance de zones acido-r6sistantes et de zones non acido-resistantes dans les colonies prelevees sur gelose ou sur milieu d'Ogawa. En g6neral, les couches directement en contact avec le milieu solide sont plus acido-r6sistantes que les portions centrales de la colonie. En revanche, cette stratification par zones n'apparait pas sur les coupes de colonies obtenues sur milieu de Sula, dont les parties superficielles ou profondes sont egale- ment acido-resistantes. Ces differences d'affinite pour les colorants peuvent etre dues i des facteurs genetiques ou 'a la diversite des conditions de nutrition des mycobact6ries r6alisees dans les milieux solides et dans les milieux liquides. REFERENCES Ames, G. F. (1968) J. Bact., 95, 833-843 Calmette, A. (1936) L'infection bacillaire et la tuberculose, Paris, Masson, p. 33 Gitores, G. F. & Roth, J. R. (1968) J. Bact., 96, 1742-1749 Guld, J., Waaler, H., Sundaresan, T. K., Kaufmann, P. C. & ten Dam, H. G. (1968) Bull. Wld Hith Org., 39, 829-836 Hallberg, V. (1941) Acta med. scand., 108, 12-17 Kahn, M. C. & Nonidez, J. F. (1936) Amer. Rev. Tuberc., 34, 361-382 Mandelstam, J. & McGuillen, K. (1968) In: Biochemistry ofbacterial growth, Oxford, Blackwell, p. 15 Mizuno, D. (1953) Acta Sch. med. Univ. Kioto, 30, 260-269 Nyka, W. (1963) Amer. Rev. resp. Dis., 88, 670-679 Pierce, C. H. & Dubos, R. J. (1956) Amer. Rev. resp. Dis., 74, 667-682 gula, L. (1963) Bull. Wld Hlth Org., 29, 589-606 Sula, L. (1969) Rozhl. Tuberk., 29, 351-355