February 11, 1952 The Genetics of Bacteria (B72-C3) Annual Progress Report submitted to the Microbiological Institute, Netional Institutes of Health Public Health Service, Bethesda 14, Maryland by Joshua Lederberg, Associate Professor of Genetics Department of Genetics, University of Wisconsin This report covers the period from April 20, 1951 to February 1, 1952 A. Summary In Salmonella typhimurium, a new mechanism of genetic exchange has been discovered: transduction. Many cultures engender a filtrable agent (FA) characterized by its genetic activity. FA transmits individual charecteristics to individual cells, in contrast to the linkege complexes exchanzed in sexual recombination. All of the traits studied (nutritional requirements; fermentations of sugars; resistence to streptomycin; and flagellar serotype) were transducible. For the most pert, laboretory mutetions of S. typhimurium were studied. In addition, the natural differences of S. typhi and S, typhimurium were exchanged resulting, for example, in the serological "hybrid" of tne antigenic formula: IX, XII, it--, (somatic antigen of typhi; flagella of tyvhimurium). Detailed studies of the genetic and kinetic properties of the TA have been initiated. It may be related to varticles about .1 micron in diameter visualized with the electron microscope in partially purified active preparations. In Escherichia coli, the sexual process underlying genetic recombination has been studied for some time. Many new interfertile strains have been found. For the first time, evidence of snecific compatibility relation- ships, ieé., of sexual or heterothallic differentiation, hes been uncovered. Direct evidence of the sponteneity of drug-resistance mutation has been secured by a new method: replica plating. ~ 2. THE GENDTICS OF BACTERIA B72-C(3) B. Full Stetement of Progress le Genetics of Salmonella. The last report presented first evidence for a process of genetic exchenge in Salmonella typhimurium, This report also gave a brief review of previous work leading to these studies. At this time, it can be seen thet certain of the conclusions of the 1951 report were incorrect, par- ticularly those which attempted to rationalize genetic exchange in Salmonella in terms of "reduced cells" or gametes. The mechanism of recombination in Salmonella appears to be fundamentally different from the sexual processes of 3. cold. To emphasize this distinction, the term "transduction" will be enplied to the Salmonella system. Most, if not all, strains of S. typhimurium are lysogenic, carrying one or more letent phages, acting on other typhimurium or other serotypes, Under av-propriate conditions involving either attack by a latent phage from another strain, or the (rather obscure) activation of its own latent phage, almost all strains have been found to liberate a filtrable agent "TA" with remarkable genetic properties. Sterile, cell-free preparations of FA are cenable of transferring individual traits from the donor strain to other susceptible strains. For the most part, we have relied upon nutritional requirements (obtained as ultreviolet-light induced mutants, and isoleted with the help of the penicillin method) as the genetic "markers" for these experiments. We have also used fermentative differences (both natural and mutetional), resistance and susceptibility to streptomycin and the natural serological differences between serotypes (S$. typhi and S. typhimurium). Since the frequency of transduction for a particular character is only about 1076 te 107? of the treated cells, selective media must be used to detect the ~ 3+ chenzes. For nutritional changes this simply involves pintings on synthetic ager medium. Fermentative or resistance transductions were detected by platings on sugar- or streptomycin-containing agar. Flagellar antigenic changes were selected by Gard!s technique (inoculation of semisolid agar containing specific antiserum). For technical reasons, the genetic changes involving nutritional requirements have been emphasized for gQuantitetive studies. Severel of the auxotrophic mutants are sufficiently stable that spontaneous variations could not be detected; for some, the spontaneous reversions did occur frequently enough to require corrections for their rates. Perhaps the most important distinctive feature of trans- duction (aside from the filtrability of the egent) has to do with the un- correlated behavior of different markers. In several different detailed experiments in wiich the FA-donor differed from the Fa-recipient strain in several respects (e.2., nutrition; fermentation of xylose; fermentation of galactose; resistance to streptomycin), each of the individual traits wes subject to transduction, but independently of all the others, Thet is, Gee, all of the streptomycin-resistant transduced cells remained like the parent cells in respect to fermentation and nutrition, and so on. This is in marked contrast to the unrestricted exchange of several markers in the non-filtrable BH. coli system, We have therefore defined "transduction as a genetically unilateral exchange, in contrast to the unior of equivalent elements in fertilization", In the lest report, we were uncertain whether the Salmonella FA might not consist of some sort of "reduced ce]1" or "gamete", but the transfer of single elements is inconsistent with such en interpretation. We were also concerned about the persistence of occasional,possibly dorment, cells of the donor strain in the FA filtretes. Ye have since found thet filtration through UF sintered Pyrex discs con- sistently removes all contaminating cells without impairing transductive -~ben activity. The greeter filtrability (through Mandler diatomaceous earth candles) of bacteria in FA preparations compered to ordinary cultures is of questionable relevance to the present prcblem, although it may still be significant in other connections (cf. "l-forms"), Most of the FA preparations usedin our present work show no evidence of the persistence of the donor cells in any organized form. Since individual traits only are transferred, transduced types are usually readily distinguishable from both the donor and the recipient strain. Therefore, the sterility of the FA preparations (although well verified) is not critical for the validity of their effects. Some thirty different markers, in several different strains, have been tested for their transmission by FA, and gave comparable results in each case, There can be no doubt that the transductive system applies to most or all of the genetic material of Salmonella. Reproducible, linear assays for TA have been develoved, based upon the yield of prototrophs from suspensions of auxotrophic cells plated with the test samples on minimal agar. This has allowed studies of the stability and purification of FA, and of its adsorption on ta susceptible cells, Concentrated preparations assaying close to 10° units/ml have been obtained by frectionel centrifugation end vrecipitation with alcohol and ammonium sulfate. Such preparations are visibly opalesecent, and show, under dark- field and electron microscopy numerous granules uniformly about .1 micron in diameter. The association of transductive activity with these granules is only a working hypothesis, but the size estimate is consistent with that secured by centrifugation and filtration experiments with gradocol membranes. The preparations are not yet, in our judgment, sufficiently pure to warrant direct chemical analysis. FA has proved to be resistant to several enzymes, including trypsin and desoxyribonuclease, The last point is the chief difference between the Salmonella transduction and the "transforming" -5- systems of the pneumococcus and Hemophilus, but may te a reflection of greater structural organization. The adsorptive properties of FA ere related to the "XII" sometic antigen of the salmonellae of grouns B end D. Most of the serotypes tested from these groups adsorbed FA with high efficiency; rough Se typhimurium, or bacteria from other groups lacking the XII antigen fail to bind FA, Whether the adsorption of §. typhimurium FA by other serotypes is followed by genetic transduction has teen tested only for §. typhi, The differential traits most readily available were streptomycin resistance (from a S" S, typhimurium), fermentation of rhamnose and of arabinose, and the serological differences. The resistance and fermentative differ- ences were individually transducible along the same pattern as between S. typhimurium strains, Monospecific anti-IX sera could not be obtained in sufficiently high titer for selective experiments involving the som-tic antigens (owing to the complexity of the XII antigen common to typhi and typhimurium), The differences in the flageller antigens afforded an excellent opvortunity to test the possibility thet transduction would yield sig- nificant, new, "hybrid" types. 5S. typhi is diagnosed as IX, XII; di—, S. typhimurium is IV, V, XII; i-1,2,;3.. . » Barring the exceptional occurrence of artificial "j" phases, S, typhi cells are immobilized in Semisolid agar containing d-antiserum., In about half of a score of trials, however, inocula of S typhi cells exposed to S. typhimurium FA have resulted in a motile (nord) phase, Following purification, these have been proven to be a new "hybrid" serotype: IX, XII; i--—--, (We ere indebted to Dr. P, R. Edwards of the P.H.S. Communicable Disease Center for confirming tnais diagnosis, and for providing many experimental materials). This type hes not yet been encountered outside the laboratory. It is amply clear, ~6- however, that transduction may well be an active mechanism for the continued evolution of new Salmonella types by the recombination of existing entigens,. Quentitetive adsorption studies have been conducted only within §. typhimurium, Adsorption of FA occurs promptly: suspensions of 10? cells/ml become saturated within fifteen minutes at 37° Gc, No cofactors or pre- treatments of the cells have been found necessary for adsorption (contra certein phages, or the preumococcus transforming principle). The efficiency of transduction appears to be limited by the saturation of the susceptible cells, The experiments do not permit us to evaluate how many particles are adsorbed per cell, It is not necessary to conclude that one adsorbed particle excludes another, but this is a possible explanation for the Single-ness of trensduced effects, The saturation of cells expesed to excess FA has been verified by their refrectoriness to transduction by a second FA preparation capable of mediating additions] effects. The Saturation level corresponds to the conversion of about one per 500,000 cells with respect to a Single trait. The proportion of cells trensduced for all other traits is certainly much higher, but no effects are dis— cernible unless the donor and recipient cells are different. The most consistent picture is that each FA perticle is associated with a Single genetic potentiality. If a cell adsorbs a perticle of a type relerant to later experimental tests, it will have a certein (so far unde termined) chance of being counted as a transduction. It is tempting to identify the FA particles as "naked genes", but the fate of comperable speculations on filtrable viruses serves as a warning. In any event, the particles anpear to be too large to correspond to individual genetic units: either they are more complex than our experiments have go far revealed, or the perticles are vehicles for a much smaller active unit. This renort is abstracted from a manuscript by N.D. Zinder ard J, Lederberg Genetic Exchange in Salmonellat already submitted for publication, 7 we. 2. Mechanism of Development of Becterial Resistance to Antibiotics. Desnite considerable evidence favoring the spontaneous mutation theory of "drug" resistance of bacteria, ‘the question has continued to be mooted whether an antibiotic might not actually induce directed mutations for resistance. So long as resistant mutants could not readily be isoleted except by exposing bacteria directly to the selective agents, only rather indirect and abstruse (biometric) evidence was available, and judging from many informal discussions, this has not been entirely convincing to meny bacteriologists. A recent advance in metnods--replica pleting and indirect selection—-has resolved this problem to some extent by meking possible the isoletion of resistant mutants without direct exposure of the selected bacteria to the drug. Replica~plating is a method for "copying" or "printing" the pattern of microbial growth from an initial agar plate to a series of other media, A sheet of velveteen (previously steam-sterilized) is fastened to a support with circuler cross-section slightly smaller then a Petri plete. The initial plate carrying the bacterial growth (colonies or otherwise) is pressed down on the velveteen, transferring an imprint of each colonly to the fabric. Plates of various agar media can then be pressed on the same fabric, and each of them will be imprinted with a replice of the original growth. Upwards of 200 colonies on a plate can be transferred in one operetion to a series of other plates with accurate registration of their position. The applications, for example, to the more efficient Getection of nutritional variants, or determination of antibiotic spectra, ere nearly self-evident. For the present application, the iritial plate carries a uniform film of growth from a large, spread inoculum, rether than single colonies, If, for example, mutations of Hi. coli resistant to streptomycin have -~B- occurred spontaneously during growth on this plain medium, the resistant cells should be distributed as families or clones of varying size, This prediction was readily verified, for replicas to a series of streptomycin- ager plates resulted in the develonment of resistant colonies at super- imposable positions, corresponding to the locations of the mutent clones on the initisl plate. The clonal occurrence of resistant cells is not consistent with the idea that resistance is directly induced by the streptomycin. The experiment can be extended to give an even more rigorous con- clusion. The initial plate has not been exposed to streptomycin, but the sites of many of the resistant clones have been revealed on the replica plates. Inocula taken from the indiceted sites are considerably enriched in the proportion of mutant cells (from 100 to 1000-fold). ‘The enriched inocula were replated at a higher dilution so that a few resistant cells were included. After growth on the second initial plete, replicas were egain made on streptomycin ager, and the resistent sites again located. The reneated enrichment by 100~-fold resulted after a few cycles in inocula of which a considerable fraction was resistant, so that pure resistant cultures were isolated from single colonies. At no point has the indirect selection line been exposed to the streptomycin; the replicas alone were SO exposed, and these were used only to locate the mutant clones. The pure resistant cultures were stable, resembling in every respect the mutants obtained by direct selection. ‘The method was also applied to phage-resistence in HE, coli, and should be of general utility. These data are embodied in "Replic Plating and Indirect Selection of Bacterial Mutants", by J. Lederberg and EM, Lederberg, 1952, J. Bact., In Press (Merch 1952), ~9- Note on physiological mechanism of streptomycir-resistance: Other workers have reported (Smith, Oginsky, and Umbreit, 1949, J. Bact. 58: 761-767) that mutants of E, coli selected for streptomycin resistance show profound changes in their aerobic metabolism, so that the resistant forms are unable to be benefited in their growth by aeretion. It might be thought that this modification was a direct effect of exposure to streptomycin, similar to the acriflavine~induced non-aerobic mutants of yeast discussed by Ephrussi, rather than a consequence of the resistance mutation. This possibility would be of great genetic interest, and would be accessible to study be means of indirect selection. We soon found however thet we could not reproduce the published findings: each of several resistant selections in severel strains of 3. coli was benefited by aeration to the same marked degree as its sensitive parent. The same held for strains provided by the workers cited. We have since learned that other investigators likewise failed to confirm the reported findings. ~10- 3. Genetic Recombination in Escherichia coli. The methods, experiments, and reasoning that have led to the conclusion that a sexual phase operates in H. coli have been presented at length in several publications (cited in bibliography in support of Research Plan) as well as in previous progress reports, and will not be repeated here. This process is under intensive study from many aspects: two new developments are of sufficient maturity and general interest to be recounted in this report. @. Strain crosses.