NUCLEAR AND CYTOPLASMIC FACTORS CONTROLLING ENZYMATIC CONSTITUTION S. SPIEGELMAN! Modern physiological research has emphasized the predominant role of enzymes in controlling cel- lular processes. On the. basis of overwhelming evi- dence we accept as established the fact that a cell can do whatever it does by virtue of the enzymes it contains. Genetics, on the other hand, has pro-. vided us with another set of units, genes, which are critical in determining cell properties. It is cus- tomary to synthesize the results of these two bio- logical disciplines in terms of a relationship between their fundamental units. Thus, it is commonly as- sumed that the genotype determines the phenotype by virtue of genic control over enzyme synthesis. On this basis, a mutant would differ from the normal type because the introduction of an atypical gene would result in a modification of the cytoplasmic enzymatic constitution. The question now naturally arises, what is really meant by such statements as “genes control en- zymes” in-terms of the mechanism and extent of such control, It is clear that the gene must in some manner influence one or more of the steps in the chain of reactions leading to the transformation of protein into some specific enzyme. Before one could profitably speculate; therefore, on the mechanism of gene-enzyme relationships, it would seem essen- tial to obtain some information on how proteins are transformed into enzymes and what connections such transformations have with other metabolic processes. Some of the basic issues of the problem may be posed in terms of the following questions: (1) How are enzymes formed in cells? (2) What mechanisms are available for the main- tenance of normal enzymatic constitution? (3) How far and in what directions can the en- zymatic make-up of a cell with a given genome be modified experimentally? (4) How does this enzymatic variation depend on the genome? The present paper will attempt to summarize and interpret the results of experiments with adaptive enzymes in yeast designed to obtain information on these and related problems. ENZYMATIC ADAPTATION The phenomenon of enzymatic adaptation may be simply stated in the following terms: a popula- tion of cells placed in contact with-some substrate acquires, after the lapse of some time, the.enzymes necessary to metabolize the added substrate. The *Some of the work reported here was greatly facilitated by a grant from -the Donner Foundation. removal of substrate leads to the disappearance gi the enzyme system it evoked. So widespread is thik phenomenon that it is now customary to folloy Karstrom (6) in designating as “adaptive” thos enzymes that are produced as a specific response ty the presence of the homologous substrate. Such en. zymes are differentiated from the . “constitutive” ones, which are presumed to be synthesized by the cell at all times regardless of the presence or absence of their homologous substrates. It has been noted previously (19, 20) that the reality of the qualitative difference implied’ by such a division of enzymes into two classes may well be questioned, This point will be returned to later and discussed briefly in the light of some more recert data. For the moment it is sufficient to point ou that the existence of the phenomenon of enzymatic adaptation, by allowing for the controlled variatior of enzyme content with the aid of substrate, pro vides a unique opportunity for an experimental ‘analysis of some of the problems raised in the in. troduction. Whether or not enzymatic adaptation can be used as a tool for the analysis of the gene-enzyme problem depends on the possibility of performing these enzyme- inducing experiments against a con. trolled genctic background. The realization of this possibility was afforded only recently through the fundamental work of Winge and Laustsen (35, 36, 37) in Denmark and of the Lindegrens (9, 10, 11) in this country on the genetics and life cycle of the yeasts. Previous papers (29, 27) have dealt with an¢ defined the conditions necessary for the analysis o: the adaptive process against a constant genome These results emphasized the importance of en: ploying known diploid strains in which the haplo phase could be suppressed whenever genetic con stancy was desired during the course of an experi ment. Haplophase cultures are uncontrollably vat iable with respect to both biochemical and morphs logical properties. Such cultures were, however, ex tremely useful since they permitted the demonstre tion (28) that capacity to form enzyme in respons to substrate is, like other genetically controle’ characters, subject to mutational changes. Mule tions resulting in both the loss and gain of sut characters were observed in haploid clones. Tur GENETICS OF Enzymatic ADAPTABILITY If a better understanding of the nature and tent of the genic control over enzymatic adaptabi ity was to be attained, it was essential to study t [ 256} FACTORS CONTROLLING ENZYMATIC CONSTITUTION 257 mechanism of its transmission through crosses of various kinds. Winge and Laustsen (37) were the frst 10 provide infomation of this nature. They used 5. validus and S. mandshuricus, both of which possess the ability to ferment melibiose. Hybrids were created between these forms and others incap- gule of fermenting this disaccharide. Their results Jed them to state that “the presence of a specific enzyme is dominant to its absence in all the in- stances studied.” ; These pioneering experiments. of Winge and Laustsen were of primary importance in opening, up the possibility of a more direct experimental ap- proach {o the gene-enzyme problem. Usually ge- netic experiments study the ‘transmission of such characters as color, shape, or the presence or ab- sence of complete structures. These, of course, in- yolve enzyme activity. However, they usually are the end results of many enzyme-controlled reac- tions. Any conclusions, drawn from such data, about the nature of the relation between gene and enzyme are necessarily indirect. Here for the first time the transmission of a well-defined enzyme, rather than the products of enzymatic activity, was being ana- lyzed genetically. The necessity for extending these results in several directions was immediately ob- vious. It seemed desirable to perform experiments which would characterize the phenotype of the two haploids used in the mating, as well as to examine the segregation of the fermentative prop- erty in the haploid spores derived from the hybrid. Another point these experiments raised is the fol- lowing; enzymes are cytoplasmic constituents, and hence the dominance observed could arise from the fact that the enzyme was carried into the cytoplasm of the hybrid as a result of the fusion of the two mating haploids. Certain features of the mating method employed by Winge and Laustsen made it difficult to analyze some of these questions. It consisted of placing the two spores in contact. with each other and allowing the resulting hybrid to grow. The resulting clone was then tested for the desired character. Since the two haploids are consumed in the process, examina- tion of their phenotype is impossible. The use of the hybridizing procedure developed by Lindegren and Lindegren (10) permitted the fur- ther genetic analysis of the inheritance mechanism, In this method, hybrids are produced by mixing haplophase cultures. Since only part of the culture is needed for the mating, the remainder can be used to determine the characteristics of the parent strain. Portions of the clone-culture are also available for back-crossing or mating to other clones of interest. A genetic analysis of melibiose inheritance was made by Lindegren, Spiegelman, and Lindegren (12), using this hybridization procedure and em- ploying S. cerevisiae, which cannot form melibiase, and S. carlsbergensis, which can. In all, 175 proge-. nies of the interspecific and of related progenies were examined. Table 1 gives some typical results. All haploid segregants from S. carlsbergensis could form melibiase, whereas all spores derived from S. cerevisiae failed to do so, indicating a homozygous condition for the presence and absence of this char- acter in the two strains. Wybrid I, which was formed by crossing a carls- bergensis haploid with a cerevisiae one, was pheno- typically positive, confirming the dominant nature of the fermentative capacity. When asci from such a mating were dissected, three different types were found, Out of 6 asci, from each of which all four Tasir 1, FERMENTATION OF ME Liniose BY PROGENY OF Various CROSSES BETWEEN S. CEREVISIAE AND S. CARLSBERGENSIS Four spores were removed from cach ascus und designated arbitrarily as A, B, C, D. Plus and minus signs indicate, respectively, capacity to form mctibiase and inability to do so. Pheno- Pheno- | Phenotype of Haploid lof ype of aN Niploid Copu- | type of Scgregants lagts Diploid; A B S. carlsbergensis + a a S. cerevisiae : _ ae Hybrid I +xX- + + + + + + + + = + + 07-0 ~~ Hybrid IV +xX- + ++ - - spores were recovered and tested, 3 behaved like the original S. carlsbergensis, yielding all four seg- regants as positive; 2 asci yielded 3 positives and 1 negative; and 1 ascus produced the 1:1 ratio of positives to negatives expected of a heterozygote segregating a single dominant gene. Essentially the. same results are obtained from any hybrid involv- ing the mating of one of the original carlsbergensis segregants to a negative haploid, the origin of the latter apparently not influencing the results. When a hybrid (Hybrid IV of Table 1) is formed by mating a positive spore derived as a segregant from Hybrid I with one of the negative haploids from S. cerevisiae, the resulting diploid, although phenotypically positive like Hybrid I, always yields a 1:1 ratio of positives to negatives. The peculiar and unexpected distribution of phenotypes observed tu occur in Hybrid I may be explained in the fol- lowing three ways: (1) Only one gene actually segregates, but the expected 1:1: Mendelian ratio is obscured by mutations of some of the negative spores to positives.' (2) Only one gene actually segregates, but the expected 1:1 Mendelian ratio is obscured by cytoplasmic components originating from the positive spore, which can form the en- zyme in the absence of the gene. 258 S. SPIEGELMAN (3) Two or more genes, either one of which can mediate the formation of the melibiase, are seprezated, The first explanation secms to be quite definitely ruled out by the fact that no S. cerevisiae haploid was ever observed to mutate to mclibiose fermen- tation, even under conditions of long-continued and intensive selection for such mutants. As further evi- dence against this hypothesis we may cite the 1:1 ratios always observed with hybrids similar to IV, in the segregants of which the same tendency towards such mutations should obtain, Either one of the last two explanations would adequately ex- plain the data, However, as will be shown in the following section, the accumulated evidence strongly supports the eyptoplasmic-factor hypothesis. Subsequently, similar experiments were done (7, 8) involving another adaptive enzyme, galactozy- mase. For these experiments, S. bayanus, which cannot adapt to ferment galactose, and S. cerevisiae, which can, were used. These experiments were com- plicated by illegitimate diploidization, which made aumeertain the origin of sporulating diploids re- covered from the mixture of the two haplopbases since they need not have resulted from a hybrid mating. Fortunately, another genetic marker was present to mitigate this situation. S. bayanus pro- duces large cylindrical cells in both the haplophase and the diplophase, whereas the haploid cells of S. cerevisiae are srnall and round. The hybrid between these two was large and cylindrical, showing that the bayanus type cell is dominant. The segregation of cylindrical versus round cells from such hybrids was regularly Mendelian, thus permitting the dis- tinction between legitimate and illegitimate diploids. The results of the segregation of the first inter- specific hybrid (analogous to Hybrid I of Table 1) again gave asci all of whose spores possessed fer- mentative capacity. The fact that the same spores which segregated the cell-type character in a per- fectly normal Mendelian fashion did not segregate the enzyme served to emphasize the peculiarity in the inheritance of the enzymatic character. Again, as in the case of melibiase inheritance, either the cytoplasmic hypothesis or the multiple-gene hy- pothesis could explain the data obtained with the le- gitimate hybrids and their segregants. Tue Exrent or Gene Conrrot oF ENZYME FORMATION — In the introductory paragraphs it was pointed out that one of the basic issues of the gene-enzyme problem is the extent of the control exerted by a gene over the enzyme whose synthesis it is presumed to determine. It is clear from the very existence of the phenomenon of enzymatic: adaptation that this control is not unlimited. It is evident, for example, that the possession by S. cerevisiae of the gene re- quired for the formation of the enzyme galactozy- mase does not alone guarantee that this enzyme will be found in the cell. The presence of the specific substrate, galactose, is also required. ‘The question rer ains whether intervention by the appropriate vene is necessary every Lime a mole. cule of enzyme is formed. In view of the second hy- pothesis offered for the segregation experiments described in the previous section, the possibility clearly exists that such direct genic intervention need not be necessary. It was noted in an earlier paper (19) that a care. ful analysis of the kinetics of adaptation could pro- vide data which would be helpful in deciding the relative role of the gene during the formation of an enzyme. , The usual description of gene action assumes that the gene mediates directly the production of the enzyme it controls. From this point of view, every replication of every enzyme molecule would require the intervention of the appropriate gene. On this basis one would ascribe the increase in enzyme in the presence of substrate to the stabilizing influence of the substrate on the enzyme. We may picture this mechanism by the following reaction diagram, Gi { k PF 60) Were P, is the immediate protein precursor, whose transformation yields enzyme, the activity of which is measured. The velocity constant of the transfor- mation from P, to E, is k, and its magnitude is de- termined by the gene controlling the reaction. The enzyme FE, is very unstable, however, and reverts to P, quickly, the velocity constant of the back reac- tion, k’, being very much larger than that of the forward one. Under such conditions only very small amounts of enzyme would accumulate in the cell. We now assume that substrate S stabilizes E,, and that, in the presence of excess substrate, E,S is formed predominantly. This effectively lowers the value of k’. Reaction diagram (1) and the above assumption predict, then, that in the presence of substrate the enzyme activity should increase with time according to the following equation: E=P(1—e*) (2) where P is the total amount of enzyme finally formed. According to (2), adaptation curves should always be concave to the time axis, and the ob- served rate of Increase in enzyme activity should be greatest at the onset and decrease continuously until maximal activity is reached, In the course of examining the appearance of various adaptive enzyme systems (galactozymase, maltase, melibiase, hydrogenlyase), well over 1200 adaptation curves have been obtained. In no case does the activity curve_resemble the course pre- FACTORS CONTROLLING ENZYMATIC CONSIITUTION dicted by equation (2). In all instances, the initial part of the curve is characterized by a rising rate uf enzyme formation. This is then followed by a declining rate portion, when presumably the pro- tein being transformed into enzyme becomes limit- ing and is finally exhausted. These facts would rule yut the simple mechanism suggested by reaction diagram (1). The increasing rate of enzyme formation with increasing amount of enzyme suggested an obvious modification of the mechanism detailed in diagram (1). Retaining all the properties ascribed to the first mechanism, we add the additional one that the enzyme is part of a cytoplasmic self-duplicating mechanism, such that, once the enzyme is formed, further formation of enzyme molecules can proceed without the intervention of the gene. A mechanism of this sort would result in an autocatalytic trans- formation of P into E, and would predict that en- zyme activity would increase with time according to the following relation: Po ~ 1 -+4- en kt where a is a constant depending on initial conditions and the other symbols have. their usual meanings. Table 2 reproduces representative results for two 8 (3) TADLE 2. KINEvTICS OF ADAPTATION The calculated values were determined by equation (3) on the assumption of an autocatalytic process in which the velocity of enzyme synthesis is a function of the amount of enzyme present. : Melibiase Activity . Galactozymase Activity Calcu- Calcu- Minutes lated Observed] Minutes lated Observed 60 12 8 60 4 2 90 26 28 90 HN” 10 120 56 51 120 26 25 150 99 98 150 48 48 180 150 149 180 87 91. 240 218 216 240 152 156 300 237 238 300 181 i8t enzymes when the observed values of enzyme activ- ity in an adapting culture are compared with those calculated with the aid of equation (3). There is no doubt that the data lend support to the cytoplasmic self-duplication hypothesis. CytopLasmic DUPLICATION OF AN ENZYME IN THE ABSENCE OF THE GENE An important and critical. prediction stemming from the self-duplication mechanism is one that is explicitly stated in the second of the three explana- tions offered previously for the atypical segregation of enzyme activities; viz., once the process of enzyme synthesis is started it should be able to 259 proceed in the absence of the gene that initiated it. A test of this prediction is feasible only In a case where the initiating gene can be climinated. An, op- portunity of performing this experiment was offered by the existence of such hybrids as IV in Table 1, which regularly segregate the potentiality for en- zyme formation. Under such conditions one could be relatively certain that only two out of every four spores carried the gene responsible for the fermen- tative capacity. It is evident from the discussion of the phenome- non of adaptation and the role of substrate in it that the case of exhibiting any self-duplicating cytoplasmic system involved in enzyme synthesis would be greatly facilitated by the presence of sub- strate, This suggested that a test of this hypothesis could be made by comparing the segregability of fermentative character in the presence and absence of substrate. Such experiments were carried out by Spiegel- man, Lindegren, and Lindegren (30). A hybrid Taste 3. SEGREGATION UNDER NorMat Conpitrons oF Autuiry To Form MELriasE FROM A HyBRID Smmirar ro Hyuriw IV or Taste 1 Ascus Number 8 9 i0 11 12 13 14 15 16 17 Spores QO oO PLP Ld tttt | > tiblti litle bi lee i t+ tltb i ttt i TABLE 4, SEGREGATION IN THE PRESENCE OF MELIBIOSE ~~ oF THE Anitity TO Form MELIBIASE Ascus Number Spores f 0 ON Gi Be Gd BD eet ttt+tt+ | > t+tt+++ | bt++t++++] 0 l++++++ | 9 ‘similar to IV of Table 1 from S, cerevisiae X S. carls- bergensis was employed, This hybrid was formed in the usual manner and allowed to sporulate, the asci dissected and tested for the character. The re- sults are shown in Table 3. It is clear that the regular 1:1 Mendelian ratio is obtained, characteris- tic of a heterozygote segregating a dominant gene. It should be noted for later reference that, in handling asci 8 and 9, the agar in which the dis- sected spores were planted contained melibiose, 260 S. SPIEGELMAN In these experiments, the cells came into contact with melibiose for the first time in the test for adapt- ability after segregation had already taken place. Using exactly the same haploids, the same cross was carried out in the presence of meclibiose. Segre- gations were also allowed to occur in the presence of this sugar. The data on the phenotypes of the haploid segregants obtained in this manner are summarized in Table 4. With the exception of ascus 7, identical heterozygotes treated with melibiose yielded four adaptable spores from each ascus. The segregation results obtained in the absence of substrate (Table 3) prove that only 2 spores from each tetrad in asci 1-6, inclusive, of Table 4 contain the specific gene responsible for the initia- tion of the adaptation towards the fermentation of melibiose. Despite this, all four spores from these tetrads produced haplophase cultures which fer- mented melibiose. It would thus appear that with the aid of substrate we were successful here in ob- scuring what is normally a simple Mendelian segre- galion and duplicated the abnormal segregations earlier noted with the interspecific hybrid between carlsbergensis and cerevisiae. It might be argued that, since all steps were car- ried out in the presence of melibiose, selection of adaptable mutants from haploids originally unable to ferment melibiose may have occurred. Several specific facts would appear to rule out this possi- bility: (1) during the testing of many haploid segregants from S. Cerevisiae, all of which are nega- tive, no mutation to an adaptable type has ever been observed, whether melibiose is present or not; (2) the same is true of negative haploids derived from heterozygous hybrids; and (3) asci 8 and 9, whose segregants were planted on melibiose, yielded the standard 1:1 ratio. We are thus led to suppose that the cultures from two spores of each tetrad from the first 6 asci in Table 4 lacked the gene and were able to ferment melibiose only owing to the enzyme-forming factors present in the cytoplasm. If this is correct, it might be expected that removal of the melibiose would lead not only to the disappearance of the melibiase but also, if the cytoplasmic factors were sufficiently labile, to the eventual loss of readaptability in those two out of every four spores of asci 1-6 that lacked. the gene. Such experiments were performed, and, to exclude the complications of mutation away from adapta- bility, nondividing cultures, suspended in M/15 KH,PO, containing glucose, were used. Portions of all 24 adapted cultures originating from the first 6. asci were treated in this way in the absence of meli- biose. In varying periods of time, ranging from 7 to 20 days, all these suspensions lost the ability to evolve significant amounts. of CO, anaerobically from melibiose. Twenty-four hours after a suspen- sion showed no significant traces of enzyme activity, a sample was removed and incubated with melibiose aerobically to test for readaptability. At the same _ time its ability to ferment glucose was also. ex. amined. This was done to avoid (esting cells whose physiological condition was scriously impaired by the long and vigerous shaking in a relatively un. favorable environment. With cells unable to ferment glucose, the inability to ferment mclibiose would be difficult to interpret. Three suspensions of the ori- ginal 24 were eliminated on this basis. The data on the asci all four of whose clones survived the treat- ment are given in Table 5. The removal of the mcli- TABLE 5, READAPTABILITY OF Sports OuTAINED BY MATING IN THE PRESENCE OF MELiniosE AFTeR ITAVING Lost ALL MELIBIASE Activity Ascus Number Spores o CA ke ho me t++i+] > ++ Pi-rr a lil biose leads to the disappearance of the cytoplasmic factors responsible for the appearance of the en- zymatic character in all four segregants and the reappearance of the expected Mendelian ratio of 1:1 on readaptation, This appearance of enzyme in the cytoplasm of cells apparently not carrying the gene could have been due to a passive transfer and subsequent reten- tion of the enzyme which was induced by substrate. in the cytoplasm of the parent diploid. That this was not the case was shown by experiments in which por- tions of all clones were allowed to fall to low enzyme-activity values (QNco, between 1.8 and . 10.1) in the absence of melibiose. Aliquots were then removed and incubated with melibiose, and evidence for regeneration of activity followed at intervals. In all cases, including those that eventu- ally lost their ability to adapt and therefore pre- sumably lacked the gene, marked increases in ac- tivity were observed. Activity values greater than 100 were attained in every instance. Furthermore, all the clones were carried in standard media with melibiose and were tested at weekly intervals. At the end of three months they could all without ex- ception ferment melibiose at rates equal to or greater than the ones they started with. This period is equivalent to over 2000 cell gencrations. It is evident from these results that the enzyme is actu- -ally synthesized in the cytoplasm of these cells in the absence of the genes. In view of these experiments, the atypical scgre- gation values of the interspecific hybrid, and the kinetics of the adaptive process, it seems difficult to avoid adopting the cytoplasmic. hypothesis, All - three types of results can-easily be understood in these terms. FACTORS CONTROLLING ENZYMATIC CONSTITUTION 261 We may therefore explain the effect of melibiose on the inheritance of the corresponding enzyme as follows: bectause the mating is performed in the presence of melibiose, the cytoplasm of the hap- loid carrying the gene is packed with mcelibiase. Since both copulating haploids contribute cytoplasm equally to the diploid, it starts out with some en- zyme and builds up. more, since it has the gene also. Because sporulation occurs in the presence of mcli- blose and because the sporulation period is char- acterized by growth and considerable storage, the enzyme molecules are stabilized and possibly in- creased in amount: Each of the four haploid segre- yants derives its cytoplasm from the diploid hy- brid, and it follows that cach will have enzyme molecules in its cytoplasm no matter what its genetic constilution chances to be. Finally the enzyme molecules are stabilized and are duplicated in the cytoplasm of the clones derived from the spores which did not carry the gene as long as they are kept in contact with substrate. The fact that Hybrid IV normally gives typical Mendelian segregation of the’ capacity to form enzyme, whereas Hybrid I does not, may be ex- plained by the fact that the carlsbergensis cyto- plasm has been twice diluted with cerevisiae cyto- plasm. The cytoplasmic factors that obscure the Mendelian picture of Hybrid I have thus apparently been diluted in Hybrid 1V to the point where they can no longer play critical roles in determining potentiality for melibiase formation. The cytoplasmic hypothesis received further sup- port from subsequent experiments (7, 8) involving galactozymase. Results analogous to those just de- scribed with melibiase were obtained with the S. cerevisiae X S. bayanus pedigree. Some asci which yielded four adaptable spores were shown to be heterozygous for the gene required to initiate galac- tozymase production, since on. growing the four segregant clones in the absence of substrate two of them irreversibly lost the capacity to’ adapt to galactose, Another series of experiments offers indirect sup- port for the importance of cytoplasmic factors in the transmission of enzymatic adaptability, by mak- ing the polygenic hypothesis offered as an alterna- tive seem very unlikely. In these experiments, the same S. carlsbergensis haplophase employed previ- ously was crossed with a haplophase strain lack- ing the ability to adapt to either galactose or meli- biose. Most of the asci obtained from this cross yielded a 1:1 ratio for the capacity. to form galacto- zymase as well as a 1:1 ratio for the ability to form melibiase, and, furthermore, both characters segre- gated completely: independently. However, a few of the asci*Obtained produced four spores, all of which exhibited the capacity to form both galacto- zymase and melibiase. This latter result would be difficult to understand on a genic basis but would be expected if in such asci the proper Mendelian ratio for the two characters was being obscured by the transfer of cytoplasmic components. The general picture which thus emerges of enzyme formation and the relation of the gene to this process may be described as follows: enzymes may be produced by cytoplasmic components which possess the capacity for sclf-duplication. In such instances the sole function assignable to the gene is the initiation of the enzyme synthesis. The initia- tion could be effected by virtue of a low but ever- present capacity of the gene to mediate the produc- tion of a few of the cytoplasmic self-duplicating units involved in the formation of the enzyme. The presence of the gene would thus retain indefinitely for the cell the capacity to adapt by forming a large number of enzyme molecules when it comes in con- tact with the proper substrate. The introduction of a cytoplasmic component as critically determining in enzyme formation natural- ly raises a whole host of new problems. Of primary interest are the following two questions: (1) What is the nature of this cytoplasmic com- ponent and what relation does it have to the homologous gene and enzyme? (2) How does it intervene in the formation of the enzyme? Actually when the findings on melibiase were re- ported (30, 19) it was suggested that the enzyme itself was the self-duplicating system. Subsequently, Lindegren (7) identified the enzyme with the self- duplicating cytoplasmic unit which he called the “cytogene.” It is clear, however, that the evidence thus far presented does not permit a decisive con- clusion as to whether the cytoplasmic self-duplicat- ing unit is the enzyme itself or something mediating its formation. It soon became apparent that an extension of these results in an attempt to answer questions re-_ garding the identity of this cytoplasmic unit and: its role in enzyme formation must be preceded by the acquisition of some insight into the physiologi- cal and biochemical details of the process. Without further information of this kind on the pathway or mechanism of enzyme formation it scemed im- possible to think fruitfully about these aspects of the problem. An adequate description of the mechanism of enzyme formation must necessarily provide answers to the following questions, among others: (1) What is the relation of enzyme formation to over-all cellular inetabolism? (2) What is the source of protein for the forma- tion of an enzyme? (3) What is the source of energy required for its synthesis? THE RELATION or ENzYME ForMATION TO CELLULAR METABOLISM Earlier workers had already established the fact that enzymatic adaptation requires physiologically 262 S. SPIEGELMAN functioning cells. Attempts to obtain adaptation with nonviable cells, or with cells whose physiology has been seriously interfered with by various re- agents, have uniformly met with failure. The sole exception was the report by Abderhalden (1) that he had obtained adaptation to galactose fermenta- tion with dried dead yeast cells. This experiment has never been successfully repeated, and his failure to check the possibility that the adaptation may have been due to the growth of a few surviving cells throws grave doubt on the validity of his conclu- sions. Von Euler and Nilsson (34) reported that adaptation to galactose would not occur when the cells were suspended in ordinary phosphate solu- tion containing galactose, and they maintained that the addition of “Z” factor was necessary. These observations were in direct. contradiction to the earlier experiments of Dienert (5), who observed adaptation with washed cells suspended in ordinary phosphate. Our own experience agrees with that of Dicnert. All our adaptations were carried out with thorough- ly washed cells suspended in phosphate solutions of substrate. These experiments establish the impor- tant fact that enzymes can form in yeast cells in the absence of an external source of nitrogen. It must be noted, however, that the maximal level of enzymatic activity reached under such conditions is about half or less that attained when the adapta- tion is carried out in the presence of an exogenous source of nitrogen. Yeasts are cells which possess both an aerobic and an anaerobic metabolic pathway, either of which can be used for purposes of growth. The first prob- lem that was experimentally analyzed was to see whether both of these pathways could be utilized for enzyme formation. Evidence on this question was conflicting. Stephenson and Yudkin (32) state that oxygen has no influence on the adaptive proc- ess. On the other hand, Schultz, Atkin, and Frey (17), who also investigated the adaptation of yeast to galactose fermentation, reported a very striking effect of oxygen. This question was reinvestigated (20) with the aid of five known diploid strains. Two of these strains were unable to initiate the forma- tion of galactozymase if they experienced only an- aerobic contact with galactose,. while the others could form the enzyme anaerobically: However, the rate of anaerobic adaptation in these latter cases “was only 1/40 of that. attained in the presence of oxygen, and furthermore, this anaerobic adapta- bility could easily be abolished by dissimilation for a few hours. This same relative inefficiency of anaerobic incubation with substrate in forming the homologous enzyme was also observed with maltose — and melibiose. It might be supposed that anaerobic inadaptabil- ity could be explained by the inability of the cell to use anaerobic metabolism for enzyme synthesis, This was tested with one of the strains, which was completely incapable of anaerobic adaptation. The changes in enzyme activity under anaerobic condi- tions were followed subsequent to periods of aerobic incubation with galactose. It was found that if sufficient aerobic contact with the substrate was allowed, so as to attain an enzyme activity equal to or greater than a QNoo, of 15, further enzyme was formed during a subsequent period of anacrobic contact with the substrate. These experiments made it quite evident that the anacrobic metabolic pathway could also be used for enzyme formation, The condition that appar- ently had to be met here was that sufficient enzyme be formed first acrobically so that the. cell would be in a position to utilize the cnergy content of the galactose molecule anaerobically at a rate adequate for the synthesis of more enzyme. We thus have here the interesting biological situation that the substrate not only stimulates the formation of the enzyme but in addition acts as the ultimate source of energy for the synthesis. The critical role of an energy supply from the metabolic cycle for the adaptive process is ilus-' trated by several other facts. It has been shown (31) that for these strains, in common with others (see 33), the endogenous reserves are not ferment- able and can only be used in the presence of oxygen. Such cells, when placed under anacrobic conditions in the absence of a fermentable substrate, do not metabolize. It is therefore not surprising that the capacity to form enzymes anacrobically is so poor. The same viewpoint suggests that aerobic adapta- tion occurred with such ease because with oxygen present the cell could draw on the energy coming from the oxidation of the endogenous reserves for the synthetic activity. To confirm this, experiments were performed (19) to examine adaptation times {time required to reach a QNoo, value of 100) when galactose was added at different levels of the endog- enous respiration. It was found that little difference in adaptation times was encountered as long as the galactose was added during the early period of the endogenous material before the declining-rate phase was attained. Subsequent to this, adaptation times increased sharply. The surprising fact to emerge, however, was that adaptation, although delayed, still took place when the galactose was added after all metabolic activity had ceased as measured by oxygen uptake and CO, output. These experiments seem to indicate that adaptation can occur after all the oxidizable reserves have been exhausted and there is no apparent source of energy. This puzzling situation was clarified, however, by the finding (21) that the galactose itself is oxidized by some enzyme system other than the fermentative one, which forms later under its stimulation. It is evident from these experiments that the presence of a functional aerobic or anaerobic met- abolic system is necessary for the formation of the enzyme, . FACTORS CONTROLLING ENZYMATIC CONSTITUTION 263 Tre SouRCE OF PROTEIN ror ENZYME ForMATION AND THE INTERCONVERTIBILITY OF TENZYMES It has already been noted that the presence of an exoxenous source of nitrogen raises the maximal attainable level of enzyme activity. This seemed to indicate that the cell could use externally supplied nitrogen for the purpose of increasing its activity. However, the fact that enzyme activity can appear in a nitrogen-free environment implics that there exist endogenous sources of protein, which are avail- able for transformation into enzyme protein. It has yenerally been supposed that most cells contain very little “storage protein” and that at any given moment most of the protein is functional, enzymatic or otherwise. There existed the possibility, there- fore, that when a cell is forced to form a new enzyme it may draw upon existent enzymes as a source of protein. Presumably an interaction of this kind would be most pronounced between unstabil-, ized labile enzymes—i.e., between members of the so-called adaptive group of enzymies. This possibility was tested (23) in the following manner: one group of cells was fully adapted to galactose and another to maltose. Then a compari- son was made of the ability of galactose-adapted cells to form maltase when incubated with an equal mixture of galactose and maltose in the presence and absence of an external source of nitrogen. A similar comparison was made of the ability of maltose-adapted cells to form galactozymase. If such an interaction between enzymes of a cell exists, TABLE 6. COMPETITIVE INTERACTION BETWEEN ADAPTIVE ENZYMES: - Maximal Activities (Q¥co,) Without External With External E Nitrogen Nitrogen nzyme Cells Cells Cells Cells adapted to jadapted to {adapted to jadapted to galactose | maltose | galactose | maltose Galacto- , . zymase 163 101 232 227 Maltase 86 182 _ 229 248 . inducing to maximal activity one enzyme of a cell and stabilizing it with its substrate should influence the ability of another enzyme to establish itself in the same cell. The results of a typical experiment, which are given in Table 6, leave no doubt that such an interaction exists. The presence in a cell of maltase depresses considerably the attainable level of galactozymase, and vice versa. Both inhibitions are effectively abolished by the addition of a nitro- gen source, which apparently permits the cell to form new enzyme without. drawing upon existent proteins. Another way of studying this phenomenon of interaction is to actually follow what is happening to the existent enzymes of accll while a new cnzyine is being furmed. Thus if one starts out with a cul- ture fully adapted to maltose (QNeo, = 240) and adapts to yvalactose one finds that as the yalacto- zymase activity increases there is a sharp drop in maltase activity so that within four hours practically all of it is gone. A test of the generality of this type of interaction between cytoplasmic enzymes was made by secing whether the induction of an adaptive cnzyme re- sulted in any loss of activity in a constitutive one. For this purpose a culture was adapted to galactose, Ne cas 200] Ng. >» = a — 3, 1004 OBB 2 _g 120. al Olucozymase g o R604 ® Galactozymase pow Y 40) Qo 320 240 360 Ninutes Fic. 1. Interaction between galactozymase and glucozy- mase during adaptation to galactose in the absence of an exogenous source of nitrogen. Curve (A) is a control, demon- strating stability of glucozymase in a nonadapting culture, and at intervals samples were removed for simul- tancous determinations of glucozymase and galacto- zymase activities, The results of an experiment of. this type are depicted in Fig. 1. The same results, are obtained here as in the case of interaction be- tween two adaptive enzymes. To ascertain whether a question of nitrogen supply is involved here also, a control experiment was run in which the adapta- tion was carried out in the presence of nitrogen. Fig. 3 gives the results. It will be noted that the adaptation rate is increased and the maximal enzyme activity level attained is greater; and, even more significantly, parallel measurements in glu- cozymase activity during the course of the adapta- tion show no change. Thus these experiments provide some information on the question of the protein source during adapta- tion. If an external source of nitrogen is present this will be employed. In its absence, however, a cell will use existent cellular enzymes as a source of protein to form the enzyme being induced by substrate, The fact that a constitutive enzyme was ob- served to disappear during the formation of an adaptive one throws serious doubt on the basis for 264 the distinction made in the literature on enzyme variation between the two types of enzymes. It em- ~phasizes once more that the only difference that exists is-in- degree of stability and utilization, The importance of these findings resides in the fact that 240 p99 enn? 200, “ @ Glucozymase 2 © Galactozymase “27-1604 = Vv < 120] oY & z 604 < nw 40] . 120 240 360 Minutes Fic. 2. Protective action of exogenous nitrogen source on glucozymase activity of an adapting culture. > 1604 ow 3 % 1204 2 aol > 18 ° ° o (N2) a . HH 40] NN : Nese ———~a == (02) Ly —— ae 500 1000 1500 Minutes __ Fie. 3. Comparison of galactozymase stability in absence of substrate under aerobic_and anaerobic conditions. they naturally make it possible to extend the re- sults obtained with adaptive enzymes to enzyme synthesis and maintenance in general. Tur RELATION OF ENZYME STABILITY TO THE SYNTHESIS OF OTHER ENZYMES The finding that the formation of a new enzyme by a cell affects other enzymes suggested that the whole question of enzyme stability and maintenance must be looked at from a different point of view. Whether a particular enzyme survives in the cyto- plasm during the course of an enzymatic change would, from this point.of view, depend not only on its inherent instability but perhaps even more ' significantly on the ability of the units involved in its synthesis to compete with other such units for proteins. . S. SPIEGELMAN An opvious consequence of this is that the sta- bility of a particular enzyme at any given time would depend on the rate of cnzyme turnover at that moment. Complete stabilization of enzymatic constitution could thus conceivably be attained if all metabolic activity could be stopped. This con- clusion was tested with galactozymase, since it was felt that if this was true for an adaptive enzyme, the instability of which is well known, its yenerality would be more certain. It will be recalled that, under anacrobic conditions, no detectable metabolic activity can be observed in ycast suspension in the absence of a fermentable substrate. The experiment was therefore performed of testing the stability of galactozymase in oxygen and in nitrogen in the absence of substrate. The results are given in Fig. 3. Except for a slight drop at the beginning, com- plete stabilization of the enzyme was attained in the absence of metabolic activity. The same sus- pension in the same period of time lost almost all of its activity when allowed to metabolize its endogenous reserves. In Fig. 4, the ability of a suspension to retain ‘its galactozymase activity is compared during 2 160] Oy (A) 5 7 , 160 2 >= — 3 120 “~~ & ea B 80 276. 3 ° > a“ 8 | G 40), 0 a ¥ v T T 2 4 6 8 Hours Fic. 5. Effect of NaNa (5 X: 10M) on anaerobic forma- tion of galactozymase. and the suspension continues to ferment the galac- tose at the rate attained at the time of the addi- tion. Azide is also effective in preventing cells from utilizing the energy of other fermentable substrates for adaptive purposes. Schultz, Atkin, and Frey (17) have shown that adaptation to maltose fer- mentation can occur in the absence of oxygen pro- - viding a small amount of glucose is furnished. This utilization of the energy obtained from the glucose to form maltase is completely prevented by azide (unpublished experiments). It is also of interest to recall that assimilation of ammonia by yeast cells was shown by Winzler, Burk, and du Vigneaud (39) to be prevented by the same con-~ centrations of azide. Enzymatic adaptation thus appears to behave in a manner similar to other synthetic processes in that it is possible by the addition of azide to prevent its occurrence without affecting the measured over-all metabolism. It was pointed out previously that the ability of a cell to maintain a particular enzyme is probably closely linked with the intensity with which other enzymes are being synthesized. If azide _ can effectively prevent enzyme synthesis in gen- eral, it follows that this compound should prevent not. only the synthesis of some enzyme, but also its disappearance, no matter what substrate the cell was metabolizing nor how fast it was doing this. Fig. 6 gives the results of an experiment testing this possibility. A galactose-adapted culture was allowed to consume glucose at maximal rate under anaerobic conditions. As shown by the broken line, the galactozymase activity disappears quite rapidly. 266 S. SPIEGELMAN If at any time in this process, however, azide is added, further disappearance of this enzyme ceases immediately and the «.zyme activity remains in- definitely at the level reached when the azide was added. The fact that the same compounds or con- ditions that inhibit enzyme formation also inhibit 21605 2 NaN: % 120 | Z| ° ° ° o “\ ig _— 40. ss ¥ T ¥ == ¥ 2 4 6 8 Fic. 6. Complete stabilization by NaN (5 & 10° M) of galactozymase in cells metabolizing glucose. enzyme loss makes it difficult to avoid accepting competitive interaction between cytoplasmic en- zyme-synthesizing units as playing a critical role in determining cellular enzymatic constitution, True MECHANISM oF AZIDE INHIBITION OF ENZYME SYNTHESIS AND Its RELATION TO PHOSPHORYLATION The ability of azide to prevent the formation of enzyme without disturbing over-all metabolism obviously represented a system the analysis of which could provide a clue to the link between the metabolic energy cycle and the synthesis of enzymes. An attempt was therefore made to determine the mechanism of this inhibition. On the assumption that P-bond energy as gen- erated by the glycolytic system forms the primary source of energy for cell function and growth, ex- periments were undertaken (26) to examine the effect of NaN, on phosphorus metabolism, using radioactive phosphorus (P2?) as a tracer. All the experiments were done under anaerobic conditions. The procedure in these experiments was to suspend cells in an inorganic phosphate medium containing a known amount of tracer P and allow the cells to fer- ment glucose anacrobically, removing samples at intervals for radioactive and chemical analysis, The results of a typical experiment of this kind are given in Fig. 7. In this experiment 2 X 10° M NaN, was used. It is seen from ,the upper curve that the presence of this amount of azide did not disturb the ability of the suspension to metabolize, _since both consumed glucose at precisely the same rate. JIowever, whereas the control exhibited a relatively rapid uptake and exchange of inorganic phosphate, the experimental did not. Chemical an- alyses for total, inorganic, and organic phosphates, and radigactive examination of these fractions, con- firmed these results. Azide is able to prevent the — 2 22 Moms Glucose /cc. > —o o—oNals 0, 180 20 300 Minutes Fic. 7. Effect of NaNs (2.5 X 107° M) on P turnover dur- ing anaerobic glycolysis. Ordinate of lower curve represents ratio of specific activity of P inside the cell to that of the P outside. Upper curves describe consumption of glucose in the absence of NaNs (open circles) and in the presence of NaNs (2.5 < 10% M) (half-shaded circles). accumulation and formation of organic phosphate bonds which normally accompany the metabolism of carbohydrates. These results suggest that the capacity of NaN; to prevent cellular utilization of metabolic energy for enzyme synthesis (as well as other synthetic processes) resides in its ability to dissociate carbo- hydrate metabolism from the generation of energy- rich organic phosphate bonds. Further experiments were performed (24) with the purpose of determining the site of the un- coupling of phosphorylation from carbohydrate metabolism in the presence of azide. According to the classical glycolytic mechanism, the formation of organic phosphate bonds from inorganic phos- phate occurs by the entrance of inorgaric P into FACTORS CONTROLLING ENZYMATIC CONSTITUTION 267 he carbohydrate cycle in two places: (1) phos- vhorolysis of glycogen; (2) the coupled oxidation ind phosphorylation of glyccraldehyde phosphate to di-phosphoglyceric acid. The involvement of the first step scemed un- Hikely, since no polysaccharide is synthesized by yeast cells in the presence of azide (38). Since it seemed most probable that the azide was affecting the second step, experiments were devised to cxam- tne the behavior of the glycolytic cycle at this level in the presence of azide. It is well known that jodoacetic acid (IAA) is a strong inhibitor of the enzyme (triose-phosphate oxidase) controlling this siep. It was reasoned that if azide modified the phosphorylative mechanism at this step it might be expected that a parallel.change in the sensitivity of the fermentation to IAA might appear, and such was actually found to be the case. A concentration of JAA (2 % 10-* M) which was sufficient to in- hibit fermentation completely within 10 minutes in the absence of azide took 180 minutes to bring the rate down to zero in the presence of 5 X 10° M NaN,, and failed to affect the rate at all for the first 70 minutes, Lower concentrations of azide pro- tected against IAA inhibition to a lesser extent. Higher concentrations continued to lengthen the period of protection until inhibitory concentrations: of azide were reached, whereupon the protective action against IAA disappeared. It should also be noted that azide is completcly unable to protect fermentation against the inhibitory action of fluor- ide, which poisons an enzyme controlling an en- - tirely different step. NUCLEOPROTEIN AS A-CONTROLLING ELEMENT IN ENzyME SYNTHESIS Any atlempt at elucidating the mechanism of a biological synthesis must of necessity concern itself with the problem of the immediate donor of energy and substrate in the synthetic reaction. The ex- periments thus far described on the relation be- tween phosphate metabolism and enzyme adapta- tion have dealt with the formation of organic phos- phate bonds. They demonstrated that inhibition of the generation of these bonds is always accom- panied by an inhibition of enzyme formation. This tells us only that organic phosphate bonds are re- quired. It does not tell us which particular ones are critical nor does it tell us how they are used. The insufficiency of this knowledge is pointedly em- phasized by experiments with dinitrophenol, which, as was independently shown by Monod (15), is also capable of preventing enzyme adaptation without inhibiting over-all oxidative metabolism. However, this compound does not interfere with organic phos- phate esterification nearly so effectively as azide. Suppression of only 20% is obtained with dinitro- phenol concentrations capable’ -of preventing enzyme adaptation. Clearly some information had to be obtained on what happened to the various organic phosphate compounds while a protein or enzyme was hcing synthesized. Experiments were designed (25) to obiain such information, again using radivactive phosphate. In these experiments the phosphate in the various fractions of the cells was tagged with tracer before the beginning of the experiment so that subsequent movement of the P from one frac- tion to another could be followed. The most consistent correlation between phos- phorus metabolism and protein or enzyme forma- tion was found in the flow of phosphate from the nucleoprotein fraction (NP). This latter is the residue phosphate remaining after successive ex- tractions with water, cold trichloroacetic acid, al- cohol, and hot alcohol-ether (3:1). The behavior of the phosphate in this fraction was followed under various conditions, employing P?, in the following manner. Cells were grown in the usual media at 30° C. in the presence of P*? (activity, 5 X 105 cts./min./mg. P). This resulted in complete equilibration of the labeled phosphorus in all fractions. After 48 hours these cells were har- vested, washed three times in unlabeled M/15 KH,PO,, resuspended in unlabeled M/15 KH,PO, with 4% glucose, and allowed to ferment the car- bohydrate under completely anaerobic conditions. No budding or increase in protein nitrogen is ob- served in such suspensions. Samples were with- drawn at intervals for activity measurements. It was found that within four hours about one-half of the total P content of the cells. The total activity was found that (except for 1 or 2%) this loss in activity could be completely accounted for in the acid-soluble fraction which forms about 50% of the total P content of the cells. The total activity (P*? content) as well as total P°1 of the nucleo- protein fraction had actually increased slightly (8%) during this period, indicating flow of phos- phate into this fraction. These data clearly showed that rapidly metabolizing but nondividing cells did not lose phosphate from the nucleoprotein (NP) fraction even though the major portion of the re- maining phosphate was being rapidly equilibrated. Since activity of the phosphate in the acid-soluble fraction of such cells was about one-fourth that of the NP phosphate, they were favorable material for the further study of exchanges between the two fractions. Allowing such cells to ferment carbo- hydrate for longer periods of time (up to six hours) again left the total activity of the NP fraction un- changed, although the specific activity was de- creased slightly owing to dilution by the flow of low-specific-activity phosphate from the acid-solu- ble fraction. The entire behavior of the NP fraction was changed, however, when such cells were induced to form new protein either by adding ammonia or by forcing the synthesis of a new enzyme. The results obtained in a typical experiment are exemplified by 268 S. SPIEGELMAN the data in Fig. 8. In this experiment eclls were sus- pended in physiological saline containing: (a) glu- cose, (b) glucose + (NH,),SO,, (c) glucose + (NH1,)250O, + NaNs, (d) glucose +, (NII,)2S0, + dinitrophenol. The amount of (NH,).SO, was equivalent in nitrogen to 50% of the nitrogen con- tent of the yeast. The concentrations of NaN, and dinitrophenol were $5 X 10° and 5 X 10-4 respec- tively, sufficient to completely inhibit enzyme for- mation. Tt will be noted that with glucose alone there was no change in activity, whereas when ammonia was present, with consequent budding, the nucleoprotein P dropped to 38% of its original total activity, ~ so] a]. . @ (t 8 40 $ Ssh Spay it Be ft Hucleoproteln Phoophate Activity. “> 304 § £ 204 2 (s) = 10] 2 4 6 Hours Fic. 8. Flow of nucleoprotein phosphate in ccils suspended in (A) glucose, (B) glucose + (NH.):SO., (C) glucose ++ (NIT)SO. + NaNs, (D) glucose + (NH.)SO. + di- nitrophenol. indicating a flow of phosphate from this fraction. It is evident that the azide and, to only a slightly lesser extent, the dinitrophenol were able to prevent this utilization. Except for the fact that the transfer of less phosphate was involved, the same phenome- non was observed when cells were induced to form . a new enzyme. Thus, in an experiment in which cells were adapted to maltose, a 34% drop in activity of the nucleoprotein phosphate was observed. Again azide and dinitrophenol in the above concentra- - tions prevented both the formation of the enzyme and the transfer of phosphate from the nucleopro- tein fraction. These findings provide us with the following cor- relations between protein or enzyme syntheses and the transfer of phosphate from the nucleoprotein fraction: (1) Rapidly metabolizing cells, which are not synthesizing new protein, do not transfer phosphate from the NP fraction, (2) Synthesis of new protein or enzyme is al- ways paralleled by a marked. transfer of phosphate from the NP fraction. , (3) Agents that are effective in inhibiting enzyme formation and protcin synthesis alsg prevent flow of P from the NP fraction. To these must be added the fundamental ob- servations of Caspersson and his collaborators on yeast (2), as well as on many other cells, which point to a rigid connection between nucleic-acid metabolism and protein synthesis, A Provosep FunctION OF NUCLEOPROTEIN IN ENZYME SyNTIIESIS The data presented in the previous section Icave little doubt that nucleoprotein metabolism is es: sential for the synthesis of enzymes and proteins. The question that naturally arises is what role the nucleoprotein plays in these synthetic processes. Is it a matter of energy supply, or substrate, or specificity? No definite answer is available. However, rnodern biochemical research (13), which has emphasized’ the role of organic phosphate bonds as sources of enerpy for synthetic activitics, provides a founda. tion upon which may be based a reasonable hy- pothesis of nucleoprotein function in enzyme forma- tion. Of particular value here js the mechanism of complex polysaccharide synthesis, the elucidation of which we owe to the brilliant work of the Coris (3) and their collaborators. Two things are re- quired in the formation of a compound like glyco- gen, One is, of course, the basic hexose unit. The other is the energy necessary to form the bonds linking these units into the complex polysaccharide. The fundamental contribution of the Coris was to show that in the synthesis of a glycosidic bond, glucose-1-phosphate rather than glucose is the re- actant involved. This phosphate ester of glucose (Cori ester) already contains in the (C-O-P) link the amount of energy required for the formation _of the glycosidic bond. This makes it unnecessary to involve some other compound as an energy donor " ’ for the purposes of driving the reaction towards synthesis. The unique feature here is the conver- sion of an energy-requiring synthetic step into a spontaneous reaction by supplying the necessary energy in the molecular structure of one of the reactants, . There seems little doubt that this finding can profitably be accepted as a model on which may be based efforts towards the elucidation of other com- plex biological syntheses, Several important con- sequences immediately flow from the adoption of this point of view. It appears that the quantitative ‘energy requirement for a particular synthetic reac- tion is not the crucial issue in determining its mech- anism. There are many phosphorylated com- pounds (e.g., di-P-glycerate, adenosinetriphosphate) which have more than sufficient energy to form a glycosidic link if there actually existed some mechanism for “feeding” it directly into the reac- tion. No such mechanism exists, however, and FACTORS CONTROLLING ENZYMATIC CONSTITUTION 269 therefore the energy content of these compounds cannot be used for this purpose. Clearly, in addi- tion to the purely quantitative aspeet, there is what may be called the “specificity” of the bond energy. Thus, the actual nature of the bond, and the compound carrying the energy, will determine its suitability for driving a particular reaction. The energy generated by the “catabolic wheel” and trapped in such energy accumulators as adenosine- triphosphate or creatinc-phosphate cannot be used as such in driving all the various synthetic mecha- nisms of anabolism. The energy_contained in such compounds must first be. transferred to others, which can then act as specific energy donors for particular synthetic reactions. This concept unifies and simplifies the problem of biological synthesis, since it avoids separating the problem of synthesis into one involving the re- actants and another concerned with the source of the “coupled” driving energy. This principle tells us that the solution of one necessarily leads us to the solution of the other aspect of the problem, since they are one and the same. Fram this point of view and the established importance of phosphorylated compounds in syn- thetic reactions, it is not surprising to find that nucleoprotcins are controlling elements in enzyme synthesis, We may further plausibly suggest that these phosphorus-containing proteins are the spe- cific energy donors which make possible reactions leading to protein and enzyme synthesis. Errecrs OF NUCLEOPROTFIN FRactTION Extracts oN ADAPTATION One must be cautious in drawing any final con- chsions about the precise role of nucleoprotein. either from the preceding data or from the discus- sion, Even granting that nucleoproteins are directly involved in the synthesis of proteins, it does not follow that they necessarily intervene in the final steps leading to the conversion of such protein molecules into enzymes. Clearly, some demonstration of specific influence by the nucleoprotein on enzymatic constitution would be necessary, before one could draw con- clusive inferences of direct determination of enzy- matic specificity by a nucleoprotein component. The most critical experiment one could offer in this direction would be one analogous to the already classical investigations of Avery and McCarty on pneumococcus transformation; i.e., the induction of a particular enzyme with a nucleoprotein com- ponent in a cell lacking the homologous gene. No such experiments have yet been succeésfully: concluded with yeast. However, some results of a very preliminary. nature have been obtained re- cently, which bear on this question and warrant mention here. It was reasoned that, if the nucleo- protein fraction (NP) was specifically concerned with enzyme synthesis, it should be possible to observe specific effects of such fractions on adapt- ing cells, Various types of fractionation were tried, and I shall here report briefly on the procedure that yiclded the most active preparations, Cells were extracted with NalICO, at plI 9.0 for 3 hours at 30° C. with constant stirring. On re- moval of the cells the supernatant was adjusted to pH 3.5 and the resulting precipitate washed and redissolved. This was followed by two subsequent precipitations. The activity of these fractions was tested in the following manner. Unadapted cells were suspended in buffer at pH 7.8, to which had been added “the adapting substrate, biotin and pantothenic acid, Sy/cc., and 1 mg. of nitrogen per cc. in the form of (NIT,),SO,. The last three compounds were added to insure that a source of nitrogen was avail- able and utilizable. It was experimentally found that the presence of these substances provided the optimal conditions for testing activity, since certain preparations which were inactive in their absence were extremely active when they were added, It was found that, when cells adapting to galac- tose were incubated with such NP fractions pre- pared from galactose-adapted cells, the adaptation time could be cut down from 180 minutes to 20 minutes. Similar preparations from normal un- adapted or maltose-adapted cells possessed no stimulatory activity towards galacto.e-adapting suspensions. In a similar manner fraction: from maltuse-adapted cells were only capable of stimu- lating adaptation to maltose, and the formation of this enzyme was not affected by the addition of extracts from galactose-adapted or normal cells. . That the enzyme itself is being transferred to the unadapted cells in this fraction seems very un- likely from several experiments. The fractions possessed no enzymatic activity by direct test. No- . maltase activity could be detected in the NP frac- tion from maltose-adapted “cells, nor could the similar fraction coming from galactose-adapted cells influence added galactose. More decisive, however, is the fact that neither could confer the spccific activity on active glucozymase extracts obtained from unadapted cells. Neither, therefore, contained in detectable amounts an apocnzymatic or co- enzymatic component of the specific enzyme in- volved. The conditions for obtaining uniformly active preparations have not yet been completely deter- mined. Since the discovery of this phenomenon about 5 months ago, 32 separate preparations have been made. Of these, 21 have exhibited specific activity in increasing rate of enzyme synthesis. The range of activities observed in the active prepara- tions has been large, extending from 50y to 0.001y per cc. for obtaining maximal stimulation. Noth- ing is known at present about the actual active component in this mixture. — In view of our ignorance of the nature of the 270 S. SPIEGELMAN active component, it is clear that nothing can be said about identifying it with nucleoprotein. The interesting and suggestive fact to emerge from these experiments is that it is possible to extract in the nucleoprotein portion a component peculiar to adapted cells, which can specifically stimulate the formation of the same enzyme contained in the cells from which the fraction originated. Until fur- ther information is, obtained on its biochemical identity we may call this active component by the neutral term “adaptin,” A Trreory oF Genz ACTION As was pointed out in the introduction, the pri- mary purpose of the present investigation is to acquire information that may lead to an adequate concept of gene action. It is proposed in the present section to suggest a mechanism of gene action based on the experimental: findings described here, as well as on previously available information. It need hardly be emphasized that the mechanism to be described must be regarded only as a tentative working hypothesis-—its usefulness to be assessed in terms of its success in unifying the diverse data on hereditary phenomena and its fruitfulness as a guide to future experiments. We may summarize the important findings and conclusions reJating to genes and enzyme forma- tion by the following statements. (1) Under normal conditions the transmission of characters (enzymes and the products of their activities) follows the classical Mendelian laws derived from. the assumption that the controlling units are self-duplicating entities, genes, located on chromosomes in the nucleus. - (2) The existence of a particular gene in the nucleus of a cell does not guarantee that the cor- responding enzyme will be found in the cytoplasm, as evidenced by such phenomena as cellular dif- ferentiation and enzymatic adaptation. Genes, there- fore;haye as their primary function the indefinite retention for the cell.of the potentiality for enzyme formation. Certain recessive genes (e.g., the meli- biase locus in S. cerevisiae) do not possess the capacity for effectively performing this function. (3) The actual formation of an enzyme in the cytoplasm is mediated directly by a cytoplasmic unit (plasmagene) which possesses the capacity for self-duplication in the presence or absence of the corresponding gene. (Following Wright, 40, 41, the term “plasmagene” is adopted for the cyto- plasmic self-duplicating entity postulated here:) (4) The presence of the homologous substrates greatly accentuates the capacity of these self- duplicating plasmagenes to produce enzyme. (5) Competitive interactions exist among the cytoplasmic enzyme-forming units. (6) Nucleoproteins are involved in the syn- thesis of enzymes. Obviously, :the view adopted concerning the identity of the cytoplasmic self-duplicating unit will in Jarge part determine the nature of the hy- pothesis devised to explain the mechanism of genic control of enzymatic constitution—particularly since the same hypothesis must afford some under- standing of the relations between such units and the genes. It would be hazardous at present to attempt to offer a definitive formulation of what we mean by a “self-duplicating” unit. However, one attribute such a unit is likely to possess is the ability to trans. form and accumulate energy within its own molec- ‘ular structure which can be used for the synthesis of similar units. At any rate, it is relatively casy to show that the growth kinetics of such “energy accumulators” is of the self-duplicating or auto- catalytic type. Of all the proteins, then, those which would be most likely to be self-duplicating are those which are involved as energy donors in protein or enzyme synthesis. Presumably these proteins could, in addition to aiding the forination of other proteins, drive protcin-synthesizing reac- tions towards the formation of units like them- selves. In view of our interpretation of the role of nucleoprotein in. protein synthesis, and of the above discussion, it seems reasonable provisionally to identify the self-duplicating plasmagenes which - mediate enzyme synthesis with the nucleoproteins, Such a hypothesis would be in harmony with the findings that all accepted self-duplicating entities have been found to be linked with nucleic-acid- containing compounds; among such entities may be mentioned genes, plastogenes, viruses, and_ the pneumococcus “transforming principle.” It must be emphasized that stating that these cytoplasmic self-duplicating units are nucleoprotein in nature does not imply that all nucleoproteins are capable of self-duplication. Identifying the cytoplasmic unit with the nu- cleoprotein, rather than with the enzyme as had been done in earlier publications, has several theo- retical and experimental consequences. It would be expected that cells not possessing the initiating gene for a particular enzyme could still retain capacity for synthesis of this enzyme, even in its absence, provided an adequate number of the ap- propriate nucleoprotein units were present. It will be recalled that experiments with melibiase are con- sistent with this point of view. In some of the clones lacking the gene, irreversible loss of poten- tiality for melibiase synthesis was not obtained until about 20 hours subsequent to the disappear- ance of all measurable enzyme activity. Thus, for a considerable period of time, these cells retained the capacity for the synthesis of this enzyme in the absence of any evidence for its presence in the cytoplasm. The experiments cited with galacto- zymase are also suggestive of this interpretation, since they demonstrated that cytoplasmic trans- FACTORS CONTROLLING ENZYMATIC CONSTITUTION 27k mission of the capacity to form enzyme can occur in the absence of any measurable enzyme activity. From a theoretical point of view, the nucleo- srotcin. concept of the plasmagene more or less dictates its relation to the gene. In view of the presumed similarity between the two, it scems al- raost necessary to conclude that the sclf-duplicating nacleoprotcin in the cytoplasm, which mediates the formation of an enzyme, is derived from the gene. One immediate value of this conclusion resides in the fact that it provides us with an experimentally, analyzable and testable entity, which can bridge the gap between the gene in the nucleus and the enzyme in the cytoplasm. We are thus Jed to propose the following concept of yene action. Genes continually produce at various rates more or Iess complete replicas of themselves, which enter the cytoplasm, These replicas or plas- magenes are nucleoprotein in nature ‘and possess CYTOPLASM NUCLEUS iP - ats opi, +Pr [i : PLE, == PLE, IP — +S; PI, E,Seeee MS's op €, s, Fic. 9. A mechanism for gene control of enzyme synthesis. The symbols are G: for gene, Ph for plasmagene, Pr for enzyme precursor, E, for enzyme, S; for substrate, and M for cytoplasmic material used in duplication of plasmagenes. The constant (k) is a reaction-velocity conslant, measuring rate of production of plasmagenes from genes. All double arrows indicate a self-duplicating process; broken-line arrows denote decay to inactive protein (IP) of the elements from which the arrows point. Where forward-and-backward reac- tions are denoted, the longer of the two arrows indicates the major tendency of the reaction. to varying degrees the capacity for self-duplication. Their presence in the cytoplasm controls the types and amounts of proteins and: enzymes synthesized. These plasmagenes, like all self-duplicating entities, compete with each other for protein and energy, and the outcome of such competitive interactions then determines the enzymatic constitution of the cytoplasm. Inherent in this concept is the possibil- ity of changing the ultimate result of this. competi- tion by varying the conditions: (e.g., substrates available) under which it takes place. The various reactions and the role of substrate in the process are detailed in Fig. 9 in terms of one gene G, and its corresponding enzyme F,. All the double arrows denote self-duplicating re- actions. Gene G, continually produces its plasma- genes (Pl,) at a rate denoted by k. The plasma- gene by its very nature must possess heterocatalytic potentialities; ie, in addition to being auto- synthetic it must possess the capacity for catalyz- ing the synthesis of units (enzymes) other than - itself. Consequently, once it is in the cytoplasm several things may happen to plasmagene (PI,). If it is successful in obtaining the proper material (M) in the cytoplasm it will duplicate itself. It may, on the other hand, combine with precursor protein (Pr) and convert it to E,, resulting in the formation of the P1],E, complex. Since little enzyme is found experimentally in the absence of substrate, one must assume that this complex is highly unstable and quickly breaks up into its two components. A plasmagene once formed cannot, of course, exist indefinitely, particularly in a population of other such units actively competing for the material of which it is composed. The re- action leading from Pl, to IP (inactive protein) in the figure describes this fact. It will be noted that in accord with the fact described in previons sec- tions the same is assumed to be true for the E, formed. By IP or inactive protein we mean merely that the plasmagene has broken down to a protein unit which has lost the capacities for self-duplica- tion and enzyme formation, and, in the case of en- zyme, into a protein which no longer possesses en- zyme activity. Thus far we have described the reactions that take place in the absence of substrate. It is clear that little enzyme would be found in the cell, unless the rate of Pl, production by G, were extremely high or the stability of the enzyme or enzyme- plasmagene complex very great. Neither condition is apparently satisfied in the cases of the enzymes reported on here. When substrate S, is added, how- ever, it will combine with E, and two things may result. It is well known that the addition of sub- strate to enzyme stabilizes it against inactivation. Hence the presence of substrate would decrease the _ rate at which free E, is converted to inactive pro- tein. More critical, however, is the possibility that S, would combine with E, while the latter is still united with Pl,, thus resulting in a Pi,E,S, com- plex. The substrate now not only would stabilize the enzyme but also could stabilize the unstable plas- magene-enzyme (P1,E,) combination. Such stabi- lizations of unstable complexes by the addition of a third component are quite common in organic chemistry. Inherent in the very definition of a self-duplicat- ing entity is the concept that, should such a unit undergo a modification at any given moment, all subsequent replicas would bear this modification. - Thus when Pl, exists alone it duplicates only Pl,, 272 S. SPIEGELMAN but when substrate is added and the stable P1,E,S, combination results his is now duplicated, which is indicated in the diagram by the double arrow. This scheme provides, therefore, a concrete mccha- nism whereby substrate can modify competitive in- teractions bétween plasmagenes., What is essential- ly accomplished by substrate is the creation of a new self-duplicating unit, which duplicates not only the plasmagene but also the enzyme corre- sponding to the substrate added. It is clear that this scheme would fit both the genetic and the kinetic data discussed in this paper. From an experimental point of view this mecha- nisin suggests that the enzymes of a cell should be found associated with the nucleoprotein components, in agreement with the data recently accumulated on the enzymatic constitution of the nucleoprotein- containing cytoplasmic granules. More important, it offers a possible experimental solution to a very irritating dilemma which attends any effort at identifying a particular biologically active molecule from a population of chemically similar ones. This is well illustrated by the present-day situation of the pneumococcus transforming principle. Only a relatively small number (about 1 in 10,000,000) of all the molecules in an active preparation of desoxyribonucleate is actually capable of inducing the specific transformation. The biological test for the presence of the active principle is so much more sensitive than any available chemical tests that positive identification of the chemical identity of the active molecules is at present impossible. Thus, while there Is little doubt that desoxy- ribonucleic acid is an essential component of the principle, the possibility that, for example, a pro- tein component is associated with it has not been excluded. It is obvious that any further efforts on our part to identify by chemical means the nature of the “adaptin” in the nucleoprotein fraction will very likely encounter precisely the same difficulties and arrive at the same impasse. The above theory sug- gests, however, that the biological specificity of the “adaptin” may be employed as an analytical tool by forming the PIES complex. Thus the corre- sponding enzyme and substrate could conceivably be used as fractionating devices in separating a particular plasmagene from a host of others, which, while they might be chemically similar, would not combine with any but their own enzymes and sub- strates. A further advantage which is suggested is that the various stages in the fractionation and iso- lation of a particular PHES complex can be followed enzymatically if one works with a well-defined and easily measured enzyme. : From a more general point of view, the unique feature of the above theory of gene action is that, while supplying a link. between the gene and the enzyme, it at the same time predicts that cells with identical genomes need not possess identical en- zymatic constitutions. Whether. a particular char- acter (enzyme) will be transmitted from one cell gencration to another in a Mendelian fashion will thus derend on the relative rates of duplication of the controlling cytoplasmic units as compared with their rate (k of the diagram) of production from the genome. If the latter is quantitatively determin. ing, Mendclian inheritance will be observed. If the former is determining, the Mendelian picture will be obscured to varying degrees depending on the self-duplicating capacity of the plasmavenes. It is also clear how substrate could so intensify cyto- plasmic inheritance of a particular enzyme as to completely obscure the segregation of the corre- sponding gene, As a tentative working hypothesis, this theory has the advantage of providing a unified point of view from which such diverse and apparently con-’ 'tradictory phenomena as classical Mendelian ge- netics, cytoplasmic inheritance, cellular differentia- tion, and enzymatic adaptation may be analyzed, The basic problem of cancer involves explaining the appearance of a sudden Acritable change in so- matic cells, analogous in scveral ways to enzyme adaptation or cellular differentiation. It is, there- fore, not surprising that cancer investigators were one of the first groups of biological workers to support strongly the suggested existence of a cyto- plasmic hereditary unit. An entity of this kind, by being self-duplicating, provides them with another level at which a mutation can take place and be subsequently transmitted via the cytoplasm from one cell-generation to the next, More or less similar views have been proposed by geneticists. Wright (40) in particular empha- sized several difficulties in trying to explain either growth or differentiation in terms of the classical Mendelian concept of the gene. Thus, the assump- tion that every time a new protein molecule is formed during growth the gene on the chromosome must intervene as a kind of model implies that growth must proceed linearly from a_ relatively minute portion of the cell. The kinetics of cell growth follow an autocatalytic law and so are not consistent with this. Wright therefore suggested that perhaps “duplicates or partial duplicates of genes reach the cytoplasm when the nuclear mem- brane disappears in mitosis and that these can pro- duce duplicates in turn, and so on, permitting ex- ponential increase.” To explain the fact that cyto- plasmic inheritance is rarely observed, he assumed that the self-duplicating capacity of these free genic replicas is subject to decay. Those that retain this capacity indefinitely he called “plasmagenes.” Again, in connection with cellular differentiation Wright (42) pointed out that the heritable stabil- ity of the differentiated state is more easily under- stood if we assume the existence of self-duplicating cytoplasmic components (plasmagenes), which can undergo controlled mutations. Stimulated by the fundamental observations of Sonneborn (18), Dar- lington (4) also postulated the existence of a FACTORS CONTROLIING ENZYMATIC CONSTITUTION cptepli asmic self-duplicating unit, which he called the “plasmagene”’ and which he assumed controls heredity at the “molecular level.” It may be of importance for reasons of c favity to emphasize certain differences between the plas- magene concept developed here and those employed by Wright and Darlington. The plasmagene as defined in the present paper Is a more or less complete gene- replica, which possesses to a vary- ing extent the capacity to self-duplicate. It is not a special or unique cytoplasmic component in the sense that it is outside normal physiological prac- esses, It is an integral part of the enzyme-synthe- sizing system and is the normal link by means of which genes can effect control over protein forma- tion in the cytoplasm. Whether or not plasmagenes are “molecular” is not pertinent to their definition. It seems probable, however, that they would per- form their synthesizing functions on the surfaces of relatively large particles (cytoplasmic granules), which could provide the necessary protein and energy-rich groups. REFERENCES ~ . AsperALDEN, E. Versuche tiber den Einfluss der Ziich- tung von ITlefe auf Galaktose auf die Vergirbarkeit dieses Kohlenhydrates durch diese. Fermentforschung. 8: 42-55. 1925. 2. Caspersson, T., and Brawor, K. Nucleotidumsatz und Wachsium bei Presshefe. Protoplasma 35: 507, 1940. 3. Cort, G. T., Swanson, M. A., and Cort, C. F. The mechanism of formation of starch and glycogen. Fed. Proc. Amer. Soc. Exp. Biol. 4: 234-241. 1945. © 4. Dartincton, C..O. Heredity, development and infec- {ion Nature 154: 164-165. 1944, 5. Diewert, F. Sur'la. fermentation du galactose et sur _ iccourumance des levures a ce sucre. Ann, Inst. Pasteur - + 139-189, 1900. 6. eaten H, Enzymatische Adaptation bei Mikroor- ganismen. Ergebn. Enzymforsch. 7: 350-376. 1938, 7, Linpecren, C. C. Mendelian and cytoplasmic inheri- tance in yeasts. Ann. Missouri Bot. ‘Garden 32: 107-123. 1945. 8 Linprcren, C. C. Yeast genetics. Bact. Rev. 170, 1945, 9. Luvvecren, C. C., and Linpecren, G. Selecting, in- breeding, recombining and hybridizing commercial yeasts. J. Bact. 40: 405-419. 1943. 10. Linvecren, C. C., and Linpecren, G. A new method of hybridizing yeasts. Proc. Nat. Acad. Sci, 29: 306-308. 1943, 11. Liypecren, C. C., and LinvecReN, G. Segregation, mu- tation and copulation in Saccharomyces cerevisiae, Ann. Missouri Bot. Garden. 30: 453-464. 1943. 12. Linvecren, C. C., Sprecenman, S., and Linpecren, G. Mendelian inheritance. of adaptive enzymes in yeast. Proc. Nat. Acad. Sci. 30: 346-352. 1944, 9: 111- 13. Lipmann, F. Metabolic generation and utilization of . phosphate bond energy. Advances in Enzymology 1: 99- 162. 1941. 14. Monon, J. Recherches sur la-croissance des cultures bacteriennes, Actualités Scientifiques et Industries. 210 pp. Hermann & Co. Paris, 1942, 15. Monop, J. Inhibition de Padaptation enzymatique chez 16. . 17, 19, 20. 21, 22, 23. 24, 25. 26, 27. 28. 29. 30. 31, 32 33, 34, 35, 36, 273 B. coli en présence de 2-4 dinitrephenol. Ann. Inst. Pasteur 70: 318-384, 1944, Monop, J. Sur Ja nature du phénoméne de diauxie. Ann, Inst. Pasteur 71: 37-40. 1945, Scnurtz, A, §. Atkin, L., and Ferry, C. N. Influence of oxygen on the fermentation o: maltose and galactose, J. Amer, Chem. Soe, 62: 2271-2272. 1940, . SONNEBORN, T, M. The dependence of the physiological action of a gene on a primer and the relation of primer to gene. Amer. Nat, 79: 318-339, 1945, StIEGELMAN, S. The physiology and genctic significance of enzymatic adaptation, Ann. Missouri Bot. Garden 32: 139-163. 1945. SrixceLMAN, §, The effect of anaerobiosis on adaptive enzyme formation. J. Cell. Comp. Physiol. 25: 121-131. 1945, SriecruMAN, S. The physiological properties of adap- tive enzyme formation. J. Bact. 49: 108. 1945. SeirceLMan, S$, Inhibition of enzyme formation. Fed. Proc. Amer. Soc, Exp. Biol. 5: 99. 1946. SrreceLMAN, S., and Dunn, R. Competitive inlerac- tions between enzyme forming systems. J. Cell. Comp. Physiol., in press. SPIECELMAN, S., and Kamen, M. D. The site of un- coupling of phosphorylation from carhohydrate metab- olism in the presence of NaNy. Fed. Proc. Amer. Soc. Exp. Biol. 5: 99. 1946, SPIEGELMAN, S., and Kamen, M: D. Genes and nucleo- proteins in the synthesis of enzymes. Science, in press. SPIEGELMAN, S., Kamen, M. D., and Dunn, R. Mech- anism of azide inhibition of synthetic activity and its relation to phosphorylation. Fed. Proc. Amer. Soc. Exp. Biol, §: 100. 1946. SPIEGELMAN, §., and Linproren, C. C. A comparison of the kinetics of enzymatic adaptation in genetically homogeneous and heterogeneous yeast populations, Ann. Missouri Bot. Garden 31: 219-233. 1944, SPIEGELMAN, S,, and Linpecren, C. C. The relation of sporulation and the range of variation of the haplo- phase to populational adaptation. J. Bact. 49: 257-269, 1945, SPIEGELMAN, S., LixpEGREN, C. C., and Henccocx, L.. Mechanisms of enzymatic adaptations i in genetically con- trolled yeast populations. Proc. Nat. Acad. Sci. 30: 13- 23. 1944, SPIEGELMAN, S., LinpecrREN, C. C., and Linpecren, G. Maintenance and increase of a genetic character by a substrate-cytoplasmic interaction in the absence of the specific gene. Proc, Nat. Acad. Sci. 31: 95-102, 1945, SpiecELMAN, S., and Nozawa, M, On the inability of intact yeast cells to ferment their carbohydrate reserves. Arch, Biochem. 6: 303-322. 1945, STEPIIENSON, M., and Yupxkin, J. Galactozymase con- sidered as an adaptive enzyme. Bio-chem. J. 30: 506-514. 1936, Stizer, T. J. B., and Stannaro, J. N. The metabolic systems involved in dissimilation of carbohydrate reserves in baker’s yeast. J, Gen. Physiol. 19: 461-477. 1935, v. Ever, H., und Ninsson, R.- Uber die Galaktosever- gurung durch Hefe nach Yorbehandlung mit dieser Zuckerart. Z. physiol. Chem, 143: 89-107. 1925, Wince, O., and Laustsen, O. On two types of spore germination and on genetic segregations in Saccharomyces demonstrated through single spore cultures. C. R. Lab. Carlsberg Ser. Physiol. 22: 99-116. 1937. Wince, O., and Laustsen, O. Artificial species hybridi- 274 S. SPIEGELMAN zation in’ yeast. C, R. Lab, Carlsberg Ser. Physiol. 22: 235-244, 1938, 37, Wince, O., and Laustsen, O. On fourteen new yeast types, produced by hybridization. C. R. Lab. Carlsberg Ser, Physiol. 22: 337-352. 1939, - 38. Winzter, R. J. Azide inhibition of anacrobic assimila- tion of glucose by yeast and its application to the de- termination of fermentable sugar. Science 99: 327-328. 1944, : 39. Winztrr, R. J., Burk, D., and pu Viewraup, V. Riotin in fermentation, respiration, growth, and N as- similation by yeast. Arch. Biochem. 5: 25-47. 1945, 40. Wrictt, S. The physiology of the gene. Physiol. Rev. 21: 487-527. 1941, so 41, Wricut, S. Genes as physiological agents, Amer. Nat. 79: 289-303. 1945, DISCUSSION ZAMENIIOF: The Delaporte method, mentioned by Dr. Spiegelman for the obtaining of his “nucleo- protcin fraction,” was devised originally to obtain desoxyribose nucleic acid; the first step of this method is to get rid of ribose nucleic acid by ex- traction with dilute sodium bicarbonate solution; the extract contains ribose nucleic acid but no desoxyribose nucleic acid. After extraction, the metachromatic granules disappear from the cells but the Jatter retain their full vitality as long as their desoxyribose nucleic acid content has not been touched. Dr. Spiegelman uses in his method just this first extract and therefore the chances are that his extract contains ribose nucleic acid and no cesoxyribose nucleic acid. If_this proves to be correct, then his active principle will be basically different from the one involved in the transforma- tion of pneumococci. If his principle proves to be indeed ribose nucleic acid, then it is perhaps worth mentioning that ac- cording to the recent work of Brachet (Enzymo- logia, 1945) there may. be two kinds of ribonucleic acid in yeast: one “granular,” which is thrown down when centrifuging at 50,000 r.p.m., and one “free,” which remains in supernatant. While the “granular” fraction is always quantitatively con- stant and may be involved in some important un- known functions, the quantity of the free fraction can be easily decreased by growing yeast on phos- phate-free medium; it is to this “free” fraction that Brachet ascribes many metabolic functions which may make it similar to Dr. Spiegelman’s ac- | tive principle. . _ Dr. Spiegelman’s substance is probably quite depolymerized, since he incubates it for 3 hours at 30° at pH 9.5. However, one has to be careful with such a statement. In Boivin’s recent work on trans-: formation of Z. coli, he uses as much as 1 mg./ml. of his desoxyribose nucleic acid obtained at pH 3.5 (and therefore depolymerized). Had this pub- lication appeared before the work of! Avery and McCarty, one would think that a large amount of depolymerized substance is necessary to induce transformation; actually, as we know from the Avery and McCarty experiments, the amounts needed are of the order of 10°* mg., but the sub. stance needed must be highly polymerized: evi- dently in Boivin’s work sume very small amounts escaped depolymerization, Couren: Before the genetic implications of Dr, Spiegelman’s exciting experiments and theory are discussed, I should like to make a plea for a more temperate use of the term “nucleoprotein.” ‘The nucleic acids exist in most organized cell struc. tures. In actual practice, especially in one in. volving an isoelectric precipitation of intraccllular material at pH 3.5, it becomes very difficult to ob- tain material which docs not contain nucleic acid, But fractions of this sort contain, in addition to nucleic acid and protein, phospholipids, sterols, neutral fats, enzymes, several types of antigens, vitamins, carbohydrates, ete. Thus almost every type of substance is found in these fractions, In other words, depending on your point of view, these materials might be called Npoproteins, elycopro- teins, enzymes, ctc. To call the fraction which Dr, Spiegelman has obtained “nucleoprotein” is to imply a causal relationship between his phenome- non and nucleic acid for which there is no evidence at present. As a matter of fact, the data directly relating nucleic acid to specifically inheritable phenomena are very sparse in general. The reac- tions described by Dr. Spiegelman stand on their own feet and gain not at all in interest by allu- sions to substances which may or may not be involved in the curious effects described. Monon: Since the inspiring results and inter- pretations reported by Dr. Spiegelman are based on experiments with yeasts, it may be of interest to see how they agree with what is known about enzymatic adaptation in bacteria. In the first place it should be pointed out that the interactions in the formation of different specific enzymes reported by Spiegelman agree very well with the data obtained on bacteria. In fact, it has been shown that such interactions are of very wide occurrence among carbohydrate-attacking enzymes, and that they account for the phenomenon of diauxis. The basic facts concerning this phe- nomenon have been briefly mentioned in Lwoff’s paper. From a rather extensive study of the occur- rence and mechanism of diauxis, the essential con- clusion to be recalled in connection with this dis- cussion seems to be that all the adaptive and constitutive carbohydrate-attacking enzymes in bacteria may depend on a common mechanism of synthesis, or, more precisely, on a common pre- cursor. A similar situation, as shown by Spiegel- man, appears to exist with glucozymase, galacto- zymase, and melibiozymase in yeast. (It should be noted, however, that the mutual-exclusion effect of different specific enzymes seems to be much more pronounced with bacteria than with yeast, FACTORS CONTROLLING ENZYMATIC CONSTITUTION 275 since in the latter it occurs only in the absence of an external supply of nitrogen. Furthermore, this is in ¢-cord with the fact that an external nitro- ven source is required for the adaptation of bacteria but not for the adaptation of yeasts.) From the results on bacteria and yeasts, the prob- jem naturally arises of the nature and degree of specificity of the common precursor (which I have called preenzyme) and of the mechanism of its substrate-induced transformation into a variety of specific enzymes. The data on bacteria seem to show: (a) that the mutual-exclusion effect ex- pressed in diauxis results from a competitive inter- action of the swbsfrates for the preenzyme; (b) that the rate of adaptation docs not depend on the de- gree of saturation of the pre-existing enzyme. Con- sequently, it was concluded that the formation of the enzyme occurs as the result of a substrate- precuzyme combination, and does not depend on the substrate-enzyme combination. Clearly these conclusions do not agree with Yud- kin’s “mass action” scheme of enzymatic adapta- tion, and it is quite noteworthy that Spiegelman, on the basis of entirely different experiments with another material, should also have been led to dis- card this theory. Oo The same conclusions may also appear, at first glance, to disagree with Spiegelman’s scheme of substrate-Induced adaptation, since here, as in Yudkin’s scheme, the rate of adaptation may seem to depend on the substrate-enzyme combination, However, the main object of Spiegelman’s .scheme is to explain gene-controlled adaptation and the role of the plasmagene. It implies no hypothesis as to- the nature or specificity of the precursor. I believe that with a little further specification Spiegelman’s scheme could be made to agree well with the con- clusions mentioned above. One need simply assume that the formation of the PIES complex results from the_reaction Pr + P! + S = PIES. The ¥ate~of. formation of new enzyme—that is, the rate of adaptation—would then depend on the product of the concentrations of Pr, Pl, and S, and also in part on the affinity of Pr for S. Such a scheme would agree with all the data available at the pres- ent time on enzymatic adaptation in both bacteria and yeast, and it might lead to a number of conclu- sions which could be experimentally tested. For instance, it can be predicted from it that the syn-’ thesis of enzymes depending on a common precursor might be affected by two types of mutation: (a) Mutations involving a change in a “specific” plasmagene, which would affect a single enzyme of the group. (b) Mutations involving a change in the struc- ture of the precursor, which might affect simultane- ously several or all of the enzymes of the group. It may be stated here that a preliminary study of a number of bacterial mutations affecting carbo- hydrate-attacking enzymes appears to confirm that mutations of type (b) may indeed occur. Wowever, no definite conclusions as to the validity of these views can be drawn at present frum experiments with bacteria. Further confirmations will have to be ‘sought with a material on which straight genetic tests can be performed. Rriner: As a matter of fact, it is possible to ob- tain direct evidence for the inhibitory effect of con- stitutive upon adaptive enzymes, using direct meas- ‘urements of adaptation rates instead of growth ex- periments. For example, we have shown this for the case of glucose and galactose. if washed cells are adapted in the presence of galactose alone, the initial rate of adaptation is about twice as high as when glucose is added, measuring this rate of adap- tation by the increase of anacrobic fermentation rate on galactose in a given period of time. In view of this, it seemed likely that there should be competition between galactose and other substrates than glucose. It also scemed likcly that the intensity of substrate competition should be strongest for substrates whose configurations are close to that of galactose, and become progressively weaker as less similar substrates are introduced. This possibility was tested by studying the effect of these substrates on the breakdown of the adap- tive galactozymase, We found an appreciable de- crease in adaptive fermentation rate when adapted cells were shaken aerobically in phosphate buffer for 4 to 8 hours, and a very much higher rate of break- down when glucose was added during this treat- ment. According to our suppositions, substrates like alcoho] and pyruvate should have an effect inter- mediate between glucose and buffer alone. It was found that the curves for alcohol and pyruvate co- incided with each other, and fell just about halfway between the other two curves, in accordance with the hypothesis. - With regard to the question of rate-limiting reac- tions raised by Dr. Monod, I should like to direct attention to the theoretical demonstration by Bur- ton, a number of years ago, that the over-all rate of a chain of reactions depends not only on the rate of the slowest reaction, but on the rate of other components of the chain as well. Burton indicated that the effect of a slow reaction on the over-all rate depended on the rates of the adjacent reactions and on the presence of other’slow reactions in the chain. In view of this, it is conceivable that the introduction of one or more new (adaptive) reaction steps might change the characteristics of a chain without necessarily introducing a step slower than any of those already present. This point deserves more careful investigation, We have some evidence that the adaptive steps in the case of galactozymase constitute a rather short branch close to the beginning of the Meyerhof scheme for glucose. It seems to be most economical biologically if an adaptive substrate is switched into existing metabolic channels as early as possible, in- 276 S. SPIEGELMAN stead of requiring the adaptive formation of a large set of new enzymes, This appears to be the case with galactose. The experiments will ‘ear a good deal of repetition before we are entirely satished with them. As far as they go, however, they indicate that galactose is first phosphorylated, and that the re- sulting galactose phosphate is acted upon by an isomerase, which converts it to glucose phosphate. There are some indications that even unadapted, ‘cells can carry out the phosphorylation, and that the isomerase is the adaptive enzyme. In any event, galactose would seem to enter the Meycrhof scheme after at most two steps. In connection with Dr. Spiegelman’s experiments with adaptin, it may be of interest to mention some experiments performed in my laboratory about a ycar ago, whose interpretation seemed rather ob- scure at the time. Suspensions of S. carlsbergensis were incubated acrobically with galactose and buf- fer according to our usual technique for adaptation experiments. Samples were taken at intervals, the cells removed by centrifigation or filtration, and the supernaiant liquid added to fresh unadapted cells in the presence of fresh galactose. The time required for adaptation to begin in the fresh cells was de- termined manometrically in the usual way. It was found that supernatant obtained before 45 minutes of previous incubation had no effect, but that super- nafant obtained after 45 minutes shor{cned the time of adaptation, the effect increasing with the Jength of the incubation period. Supernatant after 180 minutes cut the time down from the control value of 90 minutes to about 45 minutes. It was supposed at the time that the supernatant might contain intermediates of galactose fermenta- tion which would act as fermentable substrates for the unadapted cells (e.g., pyruvate). However, an’ investigation of the compounds which could be de- tected in cells- and supernatant during various phases of the adaptive process made this seem quite unlikely. In the light of Dr. Spiegelman’s experi- ments, it appears possible that our active super- natants contained enough adaptin to produce the effects which were observed. -. The question which was raised concerning the demonstration that adaptin itself contains no en- zyme may be answered by referring to some unpub- lished experiments by Spiegelman. It was shown that an apozymase prepared from unadapted cells would not ferment galactose when supplemented with a cozymase preparation (boiled yeast juice) made from adapted cells. On the other hand, an apozymase from adapted cells would ferment galac- tose with cozymase preparations from unadapted or — from..adapted cells. It is clear from these results that unadapted--cells already contain all the co- factors necessary for galactose fermentation, while they do not contain the enzyme. The addition of adaptin to a Lebedew juice from unadapted cells should therefore’ result in’ fermentation if any enzyme were present in the adaptin. The negative result. conclusively demonstrates the absence of enzymatic activity. SPIEGELMAN: I shall in what follows consider some of the questions raiscd in the preceding dis. cussions. One can hardly fail to agree wholcheart. edly with Dr. Cohen’s plea for caution in’ using chemically defined terms for biological agents of undetermined chemical constitution. In reporting the facts of our investigations here I have spoken of the “nucleoprotein fraction,” by which I meant to imply only that the active principle or process can be localized in that fraction. No conclusive chemical identification with nucleoprotein was made, It is in this spirit that such chemically neu- tral terms as “adaptin” and “nlasmagene” were employed for the biologically active agents. No ob- jection can be raised, however, against speculations concerning the biochemical roles and nature of these agents, as long as care is taken to distinguish theory from fact. Our work on the chemistry of the active fraction is still in a primitive stage, especially when com- pared with the remarkable advances recently made in Avery’s laboratory on’ the pneumococcus trans- forming principle. Under the circumstances, there- fore, I cannot agree with Dr. Zamenhof that one can be relatively certain that the preparation con- tains no desoxyribose nucleic acid. There is little doubt that ribose nucleic is present, but it is cer- iainly impossible to say now whether it is either the active principle itself or a component of it, Dr. Pontecorvo raised the interesting question why more instances of cytoplasmic inheritance are’ not observed in the higher plants and animals if genes normally function via self-duplicating cyto- plasmic units as is postulated in the theory pro- posed here. I don’t think it is merely a matter of coincidence that as soon as attention was focused on the mech- anism of transmission of characters in the lower organisms (protozoa, rusts, yeasts) cytoplasmic components almost immediately began to exhibit their capacity to dominate the nuclear hereditary units. The “distance,” from the point of view of cell generation and extent of differentiation, be- tween an individual and its gametes is almost vanishingly small in these lower forms. Thus, for example, a diploid yeast cell goes through only two divisions, one equational and one reductional, to: produce its haploid gametes. Here both types of gamete, and not merely the “female type,” inherit quite directly the cytoplasm of the individual from which they arose, by a process which probably leaves its plasmagene population relatively undis- turbed. If the cytoplasm possessed self-duplicating units, this type of gametogenesis would certainly tend to augment and intensify any capacity pos- - sessed by these units to overshadow the transmis- sion of the nuclear heredity-determining units, FACTORS CONTROLLING ENZYMATIC CONSTITUTION 277 Compare this situation with the relatively com- plicated series of differentiations which accompany either spermatogenesis or odgenesis in the higher organisms. The cytoplasm of the resulting gamctes is highly specialized as a result of these processes, and the final plasmagene population is quite dif- ferent from the initial one, A rigid selection of the proper plasmagenes must have occurred during differentiation to form a cell so uniquely suited to perform its biological function..In addition to this, it must be remembered that the germ line is isolated as such relatively early in the embryogenesis of higher plants and animals. Consequently, any plasmagene that is to disturb the Mendelian mech- anism must survive in the cytoplasm for a rela- “ively long period extending from early in em- bryonic development to “sexual maturity. The chances of any but the “proper” (i.e., necessary for the onset of differentiation in the zygote) plas- magenes surviving this long waiting period and “then the subsequent sclection accompanying differ- entiation must be relatively small, although’ not completely impossible as demonstrated by the exist- ence of maternal cytoplasmic effects. Presumably, the survival of the “proper plasmagenes” is en- couraged by providing the appropriate environ- mental conditions and substrates. It is not difficult to understand why the higher pants and animals would tend to develop mecha- nisms leading to rather rigid control over the plas- ‘aagene populations of their gametes, and hence to suppression of the tendency for such cytoplasmic units to determine the transmission of characters. The adult, gamete-producing individuals in the higher plants and animals are the result of a long, complicated series of differentiation reactions deli- cately synchronized in space and time. The dis- turbance of any one of these steps could lead to the death of the organisms. Insuring the uniformity of the starting zygotic cytoplasm by suppressing all unnecessary, and therefore possibly harmful, cyto- plasmic elements in the gainctes would provide an important factor of safety for the developmental process. Obviously, where developmental processes leading to the adult iurm are either trivial or al- together nonexistent, as in the single-celled forms, control over gamete plasrnagene population would confer relatively little selective advantage. Dr. Stern raised the question of the relation of the gene dosage effect experiments to the results and theory of gene action presented here. Whether or not a gene dosage effect with a particular gene will be observed will depend on whether the rate of production of the plasmagene from the gene, or the rate of plasmagene self-duplication, is determining, If the Jatter is low, either because of an inherently poor capacity for autosynthesis or because of com- petitive conditions in the cytoplasm, then increas- ing the dosage of this gene in the nucleus will in- crease the number of the corresponding plasma- genes that will be found in the cytoplasm. On the other hand, no dosage effects will be ohserved with any gene whose plasmagene possesses a self-dupli- cating capacity high enough to quantitatively over- whelm the production rate of the gene, in so far as determining the number of plasmagenes in the cytoplasm is concerned. It should be emphasized that the theory de- veloped here is one concerned primarily with the mechanism of gene action and not of gene trans- mission, The present theory assumes that the latter is strictly Mendelian and is supplementary to the established rules of transmission. Inherent in the proposed mechanism of gene action is the possibil- ity of explaining why in certain instances the trans- mission of a character may not follow that of the gene, We sometimes call such cases instances of non-Mendelian inheritance. Actually, since classi- cal Mendelian theory refers only to nuclear gene transmission, such phenomena do not represent violations of Mendel’s laws.