Breast Cancer Shows the Peak of Pressure in the 20-29 to 30-39

VOL. 52, 1964 GENETICS: GIBSON AND SONNEBORN 869

Breast cancer shows the peak of pressure in the 20-29 to 30-39 periods with a decline in pressure to 40-49 and then with a very low ratio of increase (1.49, 1.40, 1.32, and 1.29) in the four remaining age groups. The degree to which the amount of pressure toward tumorigenesis follows the amount of reproductive activity is striking. Conclusion.-The relationship between the rate of increase or of decrease in cancer incidence at various sites, in successive age periods, may be used to give us suggestive information concerning the nature of the host response to tumorigenic pressures of different types and derivation. It should be emphasized that the present presentation and analysis of data is neither extensive nor exhaustive. It is intended merely to show that characteristic responses of each site exist for each sex and to call attention to differences which need further study and analysis. * The present paper is based on a statistical procedure presented before the Academy some years ago. It utilizes more recent data on incidence rather than those on mortality which were the basis of the earlier communication. 1 Cancer in Connecticut 1935-51, Conn. State Dept. of Health (1955); Cancer in New York State 1941-60, Bureau of Cancer Control, N.Y. Department of Health (1962). 2 It is recognized that other methods may be used, for example, the cohort type of tabulation and analysis. The present communication is, however, intended to raise questions rather than to answer them and must of necessity be brief. It will therefore be confined to one method of presentation of -the data. 3 It should be recognized that data which include cancer of all sites are composite. They represent a mixed population of competitive trends, the composition of which may vary from year to year. It is used here merely as an illustration of the method involved.

IS THE METAGON AN m-RNA IN PARAMECIUM AND A VIRUS IN DIDINIUM?* BY IAN GIBSONt AND T. IM. SONNEBORN

DEPARTMENT OF ZOOLOGY, INDIANA UNIVERSITY Communicated August 10, 1964 Certain stocks of Paramecium aurelia (syngens 1, 2, 4, and 8) contain in their cytoplasm various symbiotic organisms termed lambda, kappa, mu, etc., and the presence of these organisms results in specific "killer" phenotypes.' The maintenance and multiplication of some of the symbionts have been found to depend on a gene or genes in the nucleus of the host paramecium.' In stock 540 (syngen 1) either of two unlinked dominant genes (M, and M2) is necessary for the maintenance of mu.2 These genes act via an intermediate, the metagon.3 Active metagons can be introduced into paramecia by exposing them to cell-free extracts of metagon- bearing paramecia.4 Activity of these extracts is associated with the ribosomal fraction and with the RNA component of this fraction.4 Sonneborn5 discovered that kappa could be introduced into another ciliate, Didinium nasutum, by feeding them killer paramecia, and that kappa was thereafter maintained and multiplied even when the didinia were fed nonkillers lacking the gene essential for kappa maintenance. Downloaded by guest on September 27, 2021 870 GENETICS: GIBSON AND SONNEBORN PROC. N. A. S.

The present paper describes the behavior of mu in Didinium and the nature of the metagon in Paramecium and Didinium. The evidence indicates that the metagon replicates in Didinium and that metagon RNA hybridizes specifically with DNA from paramecia bearing an M gene. Materials and Methods.-(1) Cultures: The various stocks of Paramecium aurelia used have been described elsewhere,3 as have the methods of culturing the cells in test tubes and larger mass cultures.4 Stock 540 (genotype M1 M1 M2 M2) bears mu particles and is a killer. Branches of it exist free from mu, and these are no longer killers but sensitives, containing, metagons, however. Stocks 513 and d-200 (genotype ml ml m2 m2) are sensitive and lack both metagons and mu. Stock d-200 is largely isogenic with stock 540 except for the m genes; it was derived by back-crossing the progeny of a hybrid (540 X 513) eight successive times to stock 540. Four stocks of Didinium nasutum were used: 1, from Twin Lakes, Indiana; 1*, derived from 1 by growth at 310C long enough to free it from kappa that had been introduced into it;5 2, from General Biological Supply House, Chicago; and 3, from Yellowood Lake, Indiana. Didinium nasutum feeds and thrives on live paramecia. Test-tube and larger cultures of didinia were grown by adding packed paramecia in quantities and with a frequency adjusted to yield the growth rates and population densities appropriate to the various experimental needs. Routinely, 0.3 to 3.0 gm/liter (wet weight) of packed didinia could usually be obtained. (2) Preparation of ribosomal extracts from the two ciliates: The solutions, conditions of homog- enization, and the details of differential centrifugation were essentially the same as used pre- viously4 for Paramecium, with the following modifications. The microsomal pellet obtained at 105,000 X g was treated with sodium deoxycholate and respun at the same force to yield a ribosomal pellet. Ribosomal extracts from both Paramecium and Didinium were suspended in a solution consisting of 0.01 M Tris, 5 X 10-3 M MgClo, and bentonite 1 mg/ml. The suspension was dialyzed against this same solution overnight and was then stored at -20°C. (3) Preparation of RNA: Ribonucleic acid was extracted from the microsomes using either of two techniques: sodium lauryl sulfate and phenol (technique a)4 or guanidine hydrochloride (technique b).7 These methods gave nucleic acid preparations in which neither DNA nor protein could be detected.6 After precipitation with ethanol, the RNA was dissolved in SSC (0.015 M sodium citrate, 0.15 M sodium chloride) at a final concentration of 2 mg RNA/ml. This solution could be stored at +4°C and used over a period of up to 2 months. (4) Method of detection of metagon activity: The two methods of detecting metagon activity have been described.4 Both involve infection of metagons into cells about to lose mu because of loss of metagons. Method a, the method of 11th-fission cells, employs cells 11 fissions past the change at autogamy from genotype M1 ml m2 m2 to mI ml m2 mi. The number of metagons per cell of the latter genotype decreases until by the 11th fission about 50% of the cells have none, and therefore no mu, and the others have only one or a few metagons but a full complement of 103 or more mu. They are therefore destined to yield only one or a few descendants with a metagon, and only these can maintain mu. The method is therefore to infect with metagons at the 11th fission and examine for presence of mu the eight animals produced after three more fissions. The increased number that possess at least one metagon and therefore mu is a measure of successful infection. Most critical is the number of groups with six to eight of the eight cells possessing mu, for uninfected controls seldom, if ever yield, so many (see Table 1). Method b, the method of RNase treatment, destroys the metagons in M1 M1 M2 M2 mu-bearers by exposure to RNase (0.5 mg/ml) for about 12 hr; mu is then lost at the next fission unless reinfection with metagons occurs earlier, in which case both metagons and mu persist in the clonal progeny. (5) Nucleic acid hybridization: Paramecia were concentrated and the RNA was extracted as above. Electrophoresis on cellulose acetate paper was then carried out under ionic conditions in which a metagon-containing band, a, appeared.7 This was eluted from the paper and the eluate was concentrated to 1 ml by dialysis against polyethylene glycol. This method allowed partial purification of metagon RNA. Hybridization was accomplished using the DNA agar method of Bolton and McCarthy8 with certain modifications. Details will be published in a future paper. DNA was extracted (for method see Gibson7) from Paramecium, Didinium, Tetrahymena, Aerobacter aerogenes (the bac- terium used as food for Paramecium), and purified mu,6 incubated with agar at 1000C, treated with ribonuclease 10 wg/ml in 2 X SSC at 250C for 1 hr, and washed with bentonite solution Downloaded by guest on September 27, 2021 VOL. 52, 1964 GENETICS: GIBSON AND SONNEBORN 871

(1 mg/ml in 2 X SSC). Following incubation of the DNA with RNA containing a known amount of metagon activity and treatment of the mixture with ribonuclease solution and bentonite as above, it was placed in a column, and five 5-ml fractions (unhybridized RNA) were removed with 2 X SSC at 60'C, then a further five with 0.01 X SSC at 750C (hybridized RNA). The fractions in each group were pooled and concentrated to give 3 ml of RNA solution. This was then diluted with 2 X SSC (or in some cases-e.g., with fractions from Didinium-remained undiluted), and a 1-ml sample was used to test 20-50 cells by procedure a or b. The remainder of the solution was stored at -20'C and used if further tests were necessary. The percentage of cells infected gave a relative measure of the amount of metagon activity present. Results.-(1) Fate of mu and metagons in strains of Didinium: One didinium of each of the four strains was fed one killer paramecium (stock 540) and thereafter an excess of sensitive paramecia (stock 513) for 6 months, about 1,000 cell generations. At the end of this period and earlier, samples of the progeny were examined for the presence of mu, and extracts of other samples were tested for metagon activity. Mu was abundantly present in all individuals examined in strains 1, 1*, and 3; but it was never found in individuals of strain 2. The method of examina- tion was to look at squashed cells by phase contrast microscopy. Ribosomal and RNA extracts of all four strains regularly exhibited metagon activity by both tests a and b. Sample results with strain 2 of Didinium are given in Table 1. Thus, although strain 2 of Didinium could not maintain mu, it maintained the metagon, as did the other strains. TABLE 1 INFECTION OF METAGONS INTO DIDINIUM (STRAIN 2) FROM PARAMECIUM AND INFECTION BACK INTO PARAMECIUM Numbers of mu-bearing paramecia in groups of 8 Stocks of paramecia used as Type ot Didinium descended atter 3 fissions from single 11th-fission initial food of Didinium extract used animals 0 1 2 3 4 5 6 7 8 Stock 540 (MK) (metagons Ribosomal 33 0 1 0 0 1 4 5 10 and mu) RNA 15 0 0 0 1 0 3 1 5 Stock 540 (Sens.) (metagons, Ribosomal 46 0 0 0 0 2 3 5 26 no mu) RNA 19 0 2 1 0 0 0 3 7 Stock 513 (Sens.) (neither Ribosomal 29 7 6 6 3 2 0 0 0 metagons nor mu) RNA 29 3 9 11 1 0 0 0 0 None None 39 11 8 10 3 1 0 0 0 After feeding the didinia one paramecium of each of the stocks indicated, they were fed paramecia of stock 513 for 6 months. Then extracts (ribosomal or RNA) of the didinia were tested by method a (11th-fission paramecia) for metagon activity. Controls (last line of table) were not exposed to extracts from didinia. The numbers in the body of the table are the numbers of sets of eight paramecia which included 0 to 8 individuals with mu. Sens. = sensitive. MK = mate killer. (2) Do didinia possess metagon activity and mu before feeding on killers? From the start, a branch of each of the four strains of Didinium was set aside to be kept free from paramecia bearing either metagons and mu or metagons alone. These branches were therefore cultured on paramecia of stock 513. The cultures were examined and tested after 3 weeks, 3 months, and 6 months. Direct observation revealed no mu. The standard tests with extracts showed no metagon activity. Thus, unless supplied with mu and metagons from paramecia, didinia do not possess them. (3) Does mu depend upon the metagon in Didinium? Three lines of evidence were adduced to discover whether, in Didinium as in Paramecium, the maintenance of mu depends upon the presence of metagons. Exposure of didinia (stocks 1 and 3), bearing metagons and mu, to RNase resulted in the loss of both metagons and mu, the latter disappearing after the first fission. This implies the prior loss of Downloaded by guest on September 27, 2021 872 GENETICS: GIBSON AND SONNEBORN PROC. N. A. S.

metagons, but direct tests for metagons were made 10 days later; none could be found. So, as in paramecia, loss of mu is correlated with loss of metagons. The question was also tested by feeding to didinia (strains 1 and 3) paramecia containing mu but no metagons. Such paramecia are obtained as in method b (see Methods, §4). Between RNase treatment and the first fission, when the paramecia bear viable mu but no metagons, they were used as food for the didinia. Thereafter the didinia were fed paramecia (stock 513) lacking both mu and metagons. Mu was looked for in the didinia after 1 and 10 cell generations and after dividing for 3 months; tests for metagons were made after 3 months. Neither metagons nor mu could be found. Thus, again, didinia. did not maintain mu in the absence of metagons. Finally, the same procedure of feeding didinia RNase-treated paramecia containing mu but no metagons was followed, but this time the didinia used were already carrying metagons. Such didinia are obtained by feeding them paramecia of the sensitive branch of stock 540. In this experiment, the didinia acquired mu and maintained it during subsequent generations. Hence, in Didinium as in Paramecium, the maintenance of mu depends upon the presence of metagons. (4) Do didinia possess a gene comparable to the M genes of Paramecium? Three lines of evidence are pertinent. (1) When an M gene is present in paramecia, metagons are regularly formed; but they are not formed in didinia except after introduction from paramecia (Results, §2). (2) When an M gene is present in paramecia, loss of metagons by RNase treatment is followed by their reappearance; in didinia as in m1m1m2m2 paramecia (Gibson, unpublished), loss of metagons by this treatment (Results, §3) is not followed by their reappearance, the loss being permanent. Thus, in both (1) and (2), gene M shows the capacity to initiate the generation of metagons, a capacity not shown by Didinium. This indicates that didinia do not possess a gene comparable to M, an indication supported by the third line of evidence, from DNA-RNA hybridization, set forth below (Results, §5, ¶3). (5) Nature of the metagon in Didinium and Paramecium: The persistence of metagons in didinia for 1,000 cell generations after cutting off the supply from paramecia implies that the number of metagons increased in didinia by a factor of 21Y°° or roughly 103° without loss of activity. Since this increase occurred in the apparent absence of the only known3 generator of metagons, the M genes, and was dependent upon the presence of metagons for its initiation and for its continua- tion, the metagon seems to have increased by replication like an RNA virus, with a doubling time of about 4 hr (1,000 generations in 6 months). These paradoxical conclusions about the same object, the metagon-a gene-product in one organism, a virus in another-led us to use the technique of DNA-RNA hybridization to explore the relation of the metagon to the DNA's of the two organisms. The most striking result (Table 2) is the very high metagon activity recovered from RNA that had been hybridized with the DNA of M1M1M2M2 paramecia (stock 540). Of the RNase-treated paramecia exposed to this RNA, 90 per cent proved to have been infected with metagons. Only 5 per cent were infected when the DNA used for hybridization came from m1m1m2m2 paramecia (stock d-200). As stocks 540 and d-200 are largely isogenic, except for the M-m locus, the results are prima facie evidence for the conclusion that the DNA's of the alleles at this locus are the basis of the differential binding of the metagon RNA. This conclusion is reinforced Downloaded by guest on September 27, 2021 VOL. 52, 1964 GENETICS: GIBSON AND SONNEBORN 873

TABLE 2 METAGON ACTIVITY OF RNA DISSOCIATED FROM DNA AFTER HYBRIDIZATION WITH DNA FROM VARIOUS SOURCES Mean % of metagon activity in Source of DNA hybridized material Didinium strain 1 (metagons, no mu) 0.67 Didinium strain 2 (no metagons or mu) 1.2 Mu 1.3 Tetrahymena 1 Paramecium (metagons, no mu), stock 540, (genotype MIMIM2M2) 90 Paramecium (neither metagons nor mu), stock 513, (genotype MImOmW272) 5 Aerobacter aerogenes 1.2 For methods, see text. The percentage metagon activity is the percentage of RNase-treated paramecia (stock 540) that retained mu after exposure to the various dissociated RNA's (see method b for detection of metagon activity). by preliminary results, to be supplemented and reported later, which indicate that the RNA's released from hybridization with the DNA of other isogenic stocks differing chiefly in the m alleles yield characteristically different percentages of infection, but always much less than when the DNA from MlM1M2M2 paramecia is used for the hybridization. The full meaning of these characteristic differences remains to be explored, but they at least indicate that the m genes are not total deletions and that the metagon hybridizes specifically with the M-m locus. In marked contrast with the results using DNA's from paramecia are those obtained with DNA's from Didinium, Tetrahymena, Aerobacter, and mu (Table 2). The preparations recovered after attempts to hybridize metagon RNA with them yielded no more than 1.3 per cent "infection" under conditions in which RNA did bind to these DNA samples. These values are not significant; they lie within the range (1-2%) obtained with controls, i.e., the percentage RNase-treated cells which yielded mu-bearing descendants in the absence of exposure to metagon preparations. In other words, the standard RNase treatment fails to destroy all metagons in up to 2 per cent of the treated M1M1M2M2 test cells. These facts appear to warrant the conclusion that the metagon RNA derived from Paramecium is complementary to the DNA of the gene M and to some extent of the gene m, but not to any DNA from Didinium or the other sources. This com- plementarity suggests that the metagon derived from Paramecium is the product or "messenger" of the gene M. Failure of the metagon to hybridize with Didinium DNA indicates not only the absence of an M-like gene in Didinium, but also that the increase of metagons in Didinium occurs by a different mechanism, presumably replication. Elsewhere, evidence will be presented showing that the metagon, long after introduction into didinia, exhibits the same specificity for hybridization with gene M, thus indicating that the presumptive replicating form of the metagon in Didinium still possesses RNA complementary to the M genes. As far as our evidence goes, therefore, the two paradoxical conclusions with which we were confronted both appear to be correct: the metagon seems to be both an RNA product of gene M and capable of replication. Discussion.-The major feature of our results is their indication that two well- known properties of different RNA's-origin from DNA and replication-can also be properties of one and the same RNA, the metagon. This is so novel and the- oretically important that alternative possibilities of explaining away one or the other property need to be considered. Downloaded by guest on September 27, 2021 874 GENETICS: GIBSON AND SONNEBORN PROC. N. A. S.

Can all increase of metagons be accounted for by origin from genes without resort to replication? The M paramecia present no difficulty; the evidence indicates that their M genes do produce metagons. The m paramecia present a problem. The kinetics of loss of metagons by dilution in the course of fissions after M is replaced by m at fertilization (Methods, §4) is consistent with the assumption of slow replication of metagons with a doubling time of about 40 hr.'0 The kinetics data have not been examined in relation to the assumption of slow production of metagons by m genes, probably because this possibility seemed excluded by the failure of extracts of m paramecia lacking mu to exhibit metagon activity.4 However, it is conceivable that very low activity would escape detection or that m genes are repres- sible, making metagons only when some are already present. Didinia present a more formidable problem. The hybridization data (Table 2) indicate that they lack DNA complementary to the metagon, i.e., any gene like M or m, either in their own genome or as replicates of genes derived from ingested paramecia. That didinia lack functioning M genes obtained from paramecia is also shown by the failure to find metagons in didinia after feeding them RNase-treated M paramecia before their first fission (Results, §3). 11 Thus the data do not appear to allow for metagon production by genes in Didinium. Can then all increase of metagons be accounted for by metagon replication without resort to their production by genes? Here didinia present no problem; metagons appear to increase in them only by replication. And ordinary m paramecia are irrelevant because they lack metagons, while those briefly possessing metagons are indecisive: if metagons do actually increase slowly in them, the available evidence neither supports nor opposes interpreting the increase as due to replication. But the M paramecia present three difficult problems. (1) Metagon activity reap- pears after its loss during RNase treatment.9 Clearly not a single active metagon es- capes the treatment (which does not directly affect mu), for one active metagon suf- fices to maintain a full complement of mu,3 yet mu disappears before metagon activity can again be detected. To explain this reappearance by replication, it would have to be assumed that one or more metagons were only temporarily inactive and that activity can be regained in M paramecia, but not in didinia, for active metagons do not arise in didinia after feeding them RNase-treated M paramecia at the stage when the assumed temporarily inactive metagons would have to be present (Methods, §4). These are obviously ad hoc assumptions. (2) During conjugation, all gamete nuclei bearing an M gene transmit the capacity to form metagons, but nuclei bearing an m gene never do.2 To explain this by metagon replication requires assuming that at least one metagon is carried by every gamete nucleus that possesses an M gene, while none is carried by any m-bearing nucleus. Metagon segregation would have to parallel gene segregation at meiosis in heterozygotes. (3) The metagon hybridizes specifically with both M and m genes, but only M genes are clearly correlated with metagon-forming capacity. Hybridization with M is readily inter- pretable in terms of replication by assuming that the metagon is bound to gene M as an episome. But then, by the same reasoning, one would expect it also to be bound to gene m. Yet metagon-forming capacity is never detectably transmitted by gamete nuclei bearing any m gene, even those showing the greatest in vitro hybridization with the metagon (Results, §5), and even when the paramecia forming the m gametc nuclei are heterozygotes possessing both M and m genes and therefore Downloaded by guest on September 27, 2021 VOL. 52, 1964 GENETICS: GIBSON AND SONNEBORN 875

also metagons. This discrepancy between observations and expectation on the episome form of the hypothesis that metagons reproduce only by replication, while it may later be resolved, at present remains unexplained. In this connection it should be emphasized that M paramecia have never lost the capacity to produce metagons during the 9 years they have been under observation. Unless the M gene can be dissociated from that capacity, the essential requirements for identifying the metagon as an episome is missing. The preceding attempts to explain all of the facts only by genic production of metagons and only by metagon replication both met with difficulties. Such difficulties are not encountered, and no additional assumptions are needed on the hypothesis that the metagon replicates after infection into Didinium and arises as the m-RNA of M genes in Paramecium. There are, however, some strange features of the metagon, such as its peculiar high proportion of G + C (Gibson, unpublished). Nevertheless, while recognizing the need for much further evidence on both properties, we tentatively adopt the m-RNA-replication hypothesis as fitting best the con- siderable evidence already available. The possibility of the conjunction of these properties in a single RNA arose with the demonstration that viral RNA replicates and the evidence that m-RNA is the primary product of genic action. The question has been not whether RNA can be produced by genes or whether RNA can replicate, but why until now any given RNA has shown one or the other feature but never both. An opportunity to attack this problem-perhaps by relating it to a specific polymerase'3 which is blocked in Paramecium but not in Didinium-appears to be provided by comparing metagon behavior in Paramecium and Didinium. If our interpretation is correct, as now seems likely, then the metagon would conform to Wright's'4 conception of a plasmagene in modern terms, an m-RNA capable of controllable replication; and the release of control when in a foreign organism would have obvious implications for the origin of RNA viruses that play so important a role in modern oncology. Summary.-The ciliate Didinium normally does not contain metagons or mu, but can acquire them by eating paramecia which contain them. They then persist and multiply in didinia (followed for 1,000 cell generations). Although such multiplication of mu in Paramecium depends on the continuous presence of an M gene, it occurs in Didinium even when fed paramecia lacking this gene (as well as metagons and mu). In Didinium, as in Paramecium, mu persists and multiplies only in the continuous presence of the metagon. Exposure of didinia carrying metagons to RNase is followed by permanent loss of metagons, whereas such treatment of paramecia carrying an M gene yields only brief transient loss of metagon activity. Metagons can be extracted from didinia, as from paramecia, in the ribosomal and RNA fractions. Detector paramecia can be infected with metagons extracted from either organism. Regardless of the source, the metagons multiply little, if at all, when infected into m paramecia and are quickly diluted out in the course of several fissions. Attempts were made to hybridize the metagon RNA with DNA extracts of M and m paramecia, Didinium, mu, Tetrahymena, and Aerobacter. Tests for metagon activity of the RNA released from the DNA were negative except for RNA hybridized with DNA from Paramecium. Much more activity was shown by the RNA hybridized with DNA from M than from m paramecia. Metagon RNA thus appears to be complementary to DNA of the M locus, confirming earlier re- Downloaded by guest on September 27, 2021 876 BIOCHEMISTRY: MACH AND TATUM PROC. N. A. S.

suits suggesting that the metagon is m-RNA of M genes. Failure of the metagon to hybridize with DNA from Didinium indicates that this organism has no gene comparable to the M genes of Paramecium. Attempts to account for all of the facts by assuming that the metagon increases only by replication or only by production from host genes meet with difficulties. The metagon appears to replicate like an RNA virus in Didinium and to arise as the m-RNA of M genes in Paramecium. Its possible relation to Wright's concept of the plasmagene and to the origin of RNA tumor viruses is recognized. Note added in proof: A full and clear summary of the background work on the metagon has just appeared: Beale, G., in Cellular Control Mechanisms and Cancer, ed. P. Emmelot and 0. Muhilbock (Amsterdam: Elsevier Publ. Co., 1964), pp. 8-18. * Contribution no. 747 from the Department of Zoology, Indiana University. The experiments with metagons and mu were carried out by Ian Gibson. Aided by grant to T. M. Sonneborn AT(11-1)-235-10 of the Atomic Energy Commission. t Postdoctoral fellow, Public Health Service genetics training grant. Present address: Department of Zoology, University of Washington, Seattle 5. 1 Sonneborn, T. M., Advan. Virus Res., 6, 229 (1959). 2 Gibson, I., and G. H. Beale, Genet. Res., 2, 82 (1961). 3 Ibid., 3, 24 (1962). 4 Ibid., 5, 85 (1964). 6 Sonneborm, T. M., in preparation. 6 Gibson, I., in preparation. 7 Gibson, I., Proc. Roy. Soc. (London), in press. 8 Bolton, E., and B. McCarthy, these PROCEEDINGS, 48, 1390 (1962). 9 Gibson, I., and G. H. Beale, Genet. Res., 4, 42 (1963). "Reeve, E. C. R., and G. J. S. Ross, Genet. Res., 4, 158 (1963). 1" Called to our attention by Barbara McManamy. 12Jacob, F., and E. L. Wollman, Sexuality and Genetics of Bacteria (New York: Academic Press, 1961), chap. 16, pp. 319-324. 13Haruna, I., K. Nozu, Y. Ohtaka, and S. Spiegelman, these PROCEEDINGS, 50, 905 (1963); Weissman, C., L. Simon, P. Borst, and S. Ochoa, Synthesis and Structure of Macromolecules, Cold Spring Harbor Symposia on Quantitative Biology, vol. 28 (1963), p. 99; Baltimore, D., H. J. Eggers, R. M. Franklin, and I. Tamm, these PROCEEDINGS, 49, 843 (1963). 14 Wright, S., Am. Naturalist, 79, 289 (1945).

ENVIRONMENTAL CONTROL OF AMINO ACID SUBSTITUTIONS IN THE BIOSYNTHESIS OF THE ANTIBIOTIC POLYPEPTIDE TYROCIDINE* BY BERNARD MACH AND E. L. TATUM LABORATORY OF BIOCHEMICAL GENETICS, THE ROCKEFELLER INSTITUTE Read before the Academy April 27, 1964 Studies on the biosynthesis of tyrocidine, a bacterial decapeptide, have provided the first example of the biosynthesis of a free polypeptide by mechanisms different from those involved in the biosynthesis of proteins." 2 When several aspects of protein biosynthesis were studied in comparison with the biosynthesis of tyrocidine, it was demonstrated that (1) the enzymatic mechanisms involved in the incorpora- Downloaded by guest on September 27, 2021