NUCLEAR AND CYTOP1,AShllC FACTORS CONTROLLING ENZYMATIC CONST i`r`U`f ION S. SPIEGELMAN' 3Ioderii physiological rcscarch has cniphasizcd tlie predominant role of cnzymcs in controlling cel- lular processes. On thc basis of oveinhclming evi- dcnce we accept as established the fact that a cell can do whatever it does by virtue of the cnzynies it contains. Genetics, on the other hand, has pro- vitlcd us with anothcr set of units, genes, which are critical in determining cell properties. It is cus- toiliaty to synthcsix the results of thcse two bio- logical disciplincs in terms of a relationship between tiicir lundamrntal units. Thus, it is conimonly as- sumed that the genotype determines the phenotype by virtue ol gcnic conttol over enzyme synthesis. On this hasis, a miif ant ~voultl cliffcr from the normal type 1)ccauhe thc introduction of an atypical gcnc uoultl rcsult in a modification of thc cytoplasmic cn~ymatic constitution. The question now naturally arises, what is rcally 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 sonic information on how proteins are transformed into enzymes and what connections such transformatjons have with other metabolic processcs. Some of thc basic issues of the problem may be posed in terns of the follo~ving questions: How are enzymes formed in cells? What mechanisms are available for the main- tenance of normal enzymatic constitution? How far and in what directions can the en- zymatic make-up of a cell with a given genome be modified experimentally? How does this enzymatic variation depend on the renome? 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 Llonner Foundation. reinoval of sulj:;(r;itc lrads to the tlisnpprarance the cnzymc system it cvokcd. so witlcsprcatl is [hi: phcnomenoii thxt it is now custamary to folio,, Itscn wcre of primary importance in opening the possibility of a more direct experimental ap pro.ich lo the gene-enzyme prolilem. Usually ge- netic e\pcriments study the transmission of such ,-]]arxters as color, shape, or the prcscnce or ab- ynce of complete structures. These, of course, in- ) olve enzyme activity. However, they usually are the end icsults of many enzyme-controlled reac- lions. Any conclusions, drawn from such data, about til: nature of the rclntion bctwccn gcne and enzyme IILCCS~~I ily indirect. Here for the first time the :r.lnmission of a \cell-tlefined cnzyme, rather than tile products of enLyniatic activity, was being ana- Ig~ttl genetically. The necessity for extending these results in scveral directions was immediately ob- vious. It seemed desirable to perform experiments r\hich would characterize the phenotype of the t\\o haploids used in the mating, as well as to ,wnine the segregation of the fermentative prop- zrly in the haploid spores derived from the hybrid. .\nother point thcse experiments raised is the fol- lowing; enzymes are cytoplasmic constituents, and hence the dominance observed could arise from the fact that the enzyine 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 v,as 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 (IO) permitted the fur- iher 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. corlsbergensis, which can. In all, 175 proge- nies of the interspecific and of related progenies Phcnotype ol IInploid Sc~rcganls A 14 C D were c?iaminc.tl. Talde 1 gives somc typical results. All haploid scgrcgnnts from S. cclr'l.~bc:rscli.~i.~ could form melil)iase, whereas all SpCJrL5 tlcrivcd from S. ccrci~isiue failctl to do so, indicaiing a honi(,xygous contliticin for the presence and absence of this char- actcr in the two strains. IIybrid I, which was fornied by crossing a carls- bergensis haploid with a ccrcvisiae one, \vas phcno- typically positive, confirming the doininant nafure of the fermentative capacity. \Vhcn asci from such a mating wcre dissected, thrce different tYiJcS were found. Out of 6 asci, from each of which all four TABLE 1. hR>!J.NThTioN OF ~IEl.lI~IOSE 1iY ]PROC17.NY OF VARlOUS CROSSES DI:TWF.EN s. ClLKI:VISIAE AND s. CAH1SUEIICT:P;SIS Four spores were rciiioved from cnch ascus and tlcsigna1pd arbitrarily as A, D, C, D. Plus and minus signs indicate, respectively, capacity to forrn nirlil~iase and inability to do so. -~ __ IIybrid I I +x- Hybrid IV 1 +x- ++++ ++++ +++- ++-- ++ -- spores were recovered and tested, 3 behaved like the original S. carlsbcrgeds, yielding all four seg- regants as positive; 2 asci yielded 3 positives and 1 negative; and 1 ascus produced the 1:l ratio of positives to negatives espccted of a heterozygote Gegregating a single dominant gcne. 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 I\' of Table 1) is formed by mating a positive spore derived as a segregant from Hybrid 1 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 to 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 gcne 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. 25s S. Si'lEGBl.M.4 N (3) Two or more genes, either one of which can mctEiatc !lie forination of the inclil)iasc, are segregated. The first explanation S~CIIIS to be qllitc definitely ruled out by the fact that no S. ccrcohiae haploid \vas ever observed to mutate to melibiose fermen- tation, even under contlitions of long-continued and intensive selection for such mutants. As further evi- dence against this hypothesis we may cite the 1:l ratios aln ays observed with hybrids similar to IV, in the segregants of which the same tendency towards such inuta~ions should obtain. Either one of the last two esplnnations would aticyuately ex- plain the (lata. IToiwver, as will be shown in the following scction, the accumulated evidence strongly supports the cyptoplasn~ic-factor hypothesis. Subscquently, similar cspcrimcnts were done (7, 5') involving another adaptive enzyme, galcxtozy- mase. For these experiments, S. bayanus, which cannot adapt to ferment galactose, and S. cerczkke, which can, were used. These eqxriments wcre com- plicaied by illegitliinate diploidization, which itiade uncertain the origin of sporulating diploids re- covered from the mixture of the two haplo9haszs since they need not have resulted from a hybrid mating. Fortunately, another genetic marker was present to mitigate this situation. S. li~p~u~ pro- duces large cylindrical cells in both the hapiophase and the diplophase, whereas the haploid cells of S. ccrcvisiae are small and round. The hybrid between these two was large and- cylindrical, showing that the bayanus type cell is dominant. The segregation of c~~lin~rical 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 nielibiase inheritance, either the cytoplasmic hypothesis or the multiple-gene hy- potlicsis could explain the data obtained with the le- gitiniate hybrids and their segregants. TIIE EXTENT OF GENE CONTROL OF ENZYME FORMATION In the introductory paragraphs it was pointed out that one of the basic issues of the gene-enzyme problem is the extcnt 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 phenomeiion 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 galxtozy- mase does not alone guarantee that this enzyme will be found in the cell. The prr::cnce of the specific sul,sl rate, plactosc, is also rcquired. The question rei' 4ns nhether intcmrntion ljj the appropriate gcnc is necessary eveiy time :i mdc- cute of enzyme is formed. In view of the second hy- pothesis offered for the segregation exqm-inients tlcscri1)ed in the prcvious section, the possibility clearly exists that such direct genic intcrvcntion need not be necessary; It was noted in an earlier paper (19) that a care- ful analysis of the kinetics of atlnptafion could pro- vide data which would be helpful in tlecitling the relative role of the gene during the formation of art en xy ni e. The usual dew-iption of gene action assiimes that the gene mediates directly the production of the enzyme it controls. From this point of view, every replication of cvery 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 staliilizing influence of the substrate on the enzyme. We may picture this mechanism by the follolving reaction diagram. GI 1 Ilere 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 E, 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. `Cinder such conditions only very small amounts of enzyme would accumulate in the cell. ?Ve now assume that substrate S stabilizes E,, and that, in the presence of excess substrate, E,S is formed predoniinant~y. This effectively lo~ers the value of k'. Reaction diagram (I) and the above assumption predict, then, that in the presence of substrate the enzyme activity should increase with time according to the foIlo~~~ng equation: E = p(1- f+) (2) where ? is the total amount of enzyme finally formed. According to (21, adap~ation curves should always be concave to the time axis, and the ob- served rate of incrcnse in cnzynie activity shotild be greatest at the onset and tlccreasc continuously until maximal activity is rcached. 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- ~icicd by eqiiation (2). In all instances, the initial prt of thc curvc is charxterizcd by a rising rate "f enLyrllc fonnstion. This is then followed by a dc.clining rate portion, when presumably the pro- [[.in bcing transformed into enzyme becomes limit- ing and is finally exhausted. These facts ~oultl rule c,ut the simple niechanism suggested by reaction 90 120 150 180 240 300 dingrarn (1 )I The increasing rate of enzyme forin~tion with 90 11 - IO incrcnsing amount of enzyme suggested an obvious ril(,tiification of the mechanism detailed in diagram { 1 ). I~et~inin~ all the properties ~icribcd to the fir5t nicclianisin, we add the atltlitionnl one that the Iuymc is part of a cytoplasmic self-duplicating tj;i.clinnism, such that, once the enzyme is formed, iurther formation of enzyme molccules can proceed without the intervention of the gene. A mechanism of this sort would result in an autoca~~lytic trans- forination of P into E, and would predict that en- zyme activity would increase with time according to the following relation: - P E=l 1 +- e-kt (3) !There a is a constant depending on initiat conditions nntl the other symbols have their usual meanings. Table 2 reproduces representative results for two TAnLE 2. Klh'ETICS OF hDAPTATION Thc mlciilated valiies were dctern~ined by equntion (3) on the assumption of an autocnfalytic process in which the velocity of enzyme synthcsis is a function of the amount of enzyme present. _I___. - hlelibiase Activity 1 Galactozymase Activity Calcu- I Calcu- hfinutes lated Observed Minutes lated Ohserved 26 56 99 1.50 218 237 28 51 98 149 216 238 120 215 25 150 48 48 1 so 87 91 240 152 156 300 181 1s1 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. CYTOPLASBZIC DUPLICATION OF AN ENZYME IN THE ABSENCE OF THE GENE An important and critical prediction stemming from the self-dup~ication 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 proccctl in the nb:.c.nce of the gene thnt inifintctf it. A test of this prctliction is feasible mly In a CA.~ where the initiating gene can be eliminated. An, op- portunity of performing this e~p~~riin~nt was offered by the existence of such hy1)ritIs as IV in Table 1, which replarly segregate the potentiality for en- zyme formation, Under such conditions one could be relatively certain that only two out of every four spores carried thc gene responsible for the frrrnen- tative capacity. It is evident from thc discussion of thc phenome- non of adaptation and the role of su1)stratc in it that the ease of exhibiting my ~elf-(l~ipli~n~in~ cytoplasmic system involved in enzymc synthesis would be greatly facilitxfcd by the presence of sub- strate. This suggested that a test of this hypothesis could be made by comlming the scgregahility of fermentative character in the presence and absence of substrate. Such experiments were carried out by Spiegel- man; Lindegren, and 1,indegren (30). A hybrid TADLE 3. SEGREGATION UNIIER Eu'ORMAL CONDlTlONS OF SIMILAR TO IIYljXID Iv 0P.TAnI.E 1 ABILITY TO FORU MELIBLASE FROX A HYBRID - .- Spores IA B C D Ascus Number 8 9 10 11 12 13 14 15 16 17 - - c f + + -I- + - - TABLE 4. SEGREGATION IN TEE PRESENCE OF hfELIDI0SE.- . OF TBE ABILITY TO FORM ?VfELIBIASE Spores B c D Ascus Number + + t f + + + + + + + + + + + + 4- + + 4- + + + f + -' - + - similar to IV of Table 1 from S. cerevisiae X S. cads- bergemis 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 gcne. 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, In thcse experiments, the cclls came into contact with melibiose for the first time in the test for adapt- ability after scgrcgation had already taken place. Using exactly the same haploitls, the same cross was carried out in the presence of melibiose. Scgre- gations were also allowed to occur in the presence of this sugar. The (lata on the phenotypes of the haploid segregants obtained in this manner are suinmarixd in Table 4. With the exception of awus 7, identical heterozygotes treated with melibiose yielded four adaptable spores from each ascus. The segregation results obtained in the absence of sillstrate (Tatite 3) prove that only 2 spores froni each tetrad in asci 1-6, inclusive, of Table 4 contain the specific gene rcsponsiblc for the initia- tion of the adaptation towards the ferme~it~~ion of melihiose. Despite this, all four spores from thcse tetrads produced haplophase cultures which fer- nirnted melibiose. It would-thus appear that with the aid of substrate we were successful here in ob- scuring what is normally a simple Mendelian scgre- gal ion and duplicakd the abnormal scgregations earlier noted with the intcrspccific hybrid between carlsbergensis and cerevisiae. It might be argued that, since all steps were car- ried out in the preseiice 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: (I) during the testing of many haploid segregants from S. Eerevisiae, all of which are nega- tive, no mutation to an adaptable type has ever been obscrved, whether nielibiose 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 enzy~ne-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 tvo 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, nondiv~din~ cultures, suspended in M/15 KH,PO, contain in^ 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 aerol)ically to test for rc;trl;iptallility. At the sane time its ahility to fcrnmcnt glucose vas also ex- amined. This wis done io avoid tcsting cclls who9 I~h~~sioI~~ical c~n(~itj~n was scriously impaired by the long and vigorous shaking in a relatively un- favorable environment. With cells unalde to ferment glucose, the inability to ferment niclibiosc wuld difficult fo intcrpret. Three suspensions of the ori- ginal 24 were eliminated on this b,asis. The data on the asci all four of whose doncs survivccl the treat- ment are given in Table 5. The removal of thc mcli- biose leads to the tlicappcarance oi the cytop?;wnic factors responsible for the appearance of the en- zytnatic character in all four segregants and the reappearance of the expected filendclian ratio of 1 : 1 on reaclnptation. This appearance of enzyme in the cyioplasin of cells apparently not carrying the gene could have been due to a passive transfer and subsequent reten- tion of the enzyme which \vas intluced by substrate in the cytoplasm of the parent diploid. That this ius not the case was shown by experiments in which por- tions of all clones were allowed to {all to loa enzyme-activity values (QNC02 betwen 1 .S and 10.1) in the absence of melibiose. Aliquots were then removed and incubated with melibiose, and evidence for regeneration of activity followed at intends. In all cases, including those that eventu- ally lost their ability to adapt and thelefore 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 wekly intervals. At the end of three months they could all without es- ception fernlent melibiose at rates equal to or greater than the ones they started with. This period is equivalent to over 2000 cell generations. 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 scgrc 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. FA C7'0 Iy sccing whether the induction or ;In at1;iptivc enzyme rc- sulted in any loss of activity in a constitutive one. For this purpose a culture \-;as atlaptcd to galactose, L-y-U , I20 24 0 360 tlhutee FIG. 1. Interaction bctivccn galactozyrnase and glucozy- mase during adaptation to galnctose in the absence of an exogenous source of nitrogcn. Curve (A) is a control, dcmon- strating stability of glucozymnse in R nonadapting culture. and at intervals samples were removed for siniul- taneous determinations of glucozyniase 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 aclapta- 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 rnca~urements 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 ~~FQW serious doubt on the basis for the di\tinction ninrlc in the lilrralure 011 cnzynie \.ariation bctwccn the two types of enzynics. It cni- _. i)hn:izs once niorc that the only tlifrclcncc fhat exists i~>-(lcg~~g 01 st`iljility and utilimtion. The impo~~ancc of thcG-findings resides in the fact that 0 240 200 /"" 0 Glucoryrnaee > -. -.: 460. 120. r. Y 2 60. E C W *Oi / 120 24 0 360 flinutes FIG. 2. Protective action of exogenous nitrogen source on glucozyniase activity of an adapting culture. I , . . . ,--, , , . . , . , \o--- -. 0 0 (02) 500 1000 1500 fiinutee FIG. 3. Comparison of galactozymase stability in absence .~~ of substrate under acrobic and anaerobic conditions. they naturally make it possible to extend the re- sults obtained with adaptive enzymes to enzyme synthesis and maintenance in general. Tim RELATION OF ENZYME STABILITY TO THE SYNTIIESIS OF OTHER ENZYMES The finding that the formation of a new enzyme by a cell affects other enzymes suggested that the \diolc 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. An 0:)LiOllb LI)IMY~UCI~CC or thi> is illat thc sta- bility of a 1,arliciil:ir cnzynic ;it any given time w0:ild dci~cntl on the rntc of ciii:yinc turnover at that nioriicnt. ~fJn~i)1C!~c st;iliilization of cnzyii1;itiC cun:ititulioll CULII~ ~11~5 CUIICC;\~IIJI~ 1)~ ~II lai~~etl if all metabolic activity could he stoppcd. `This con- clusion was test ed with gal;ictclzyinase, since it was felt !hat if this was true fur ;in adaptive cnzymc, the inL:iI)ility in the abscnce of metabolic activity would 5Llsgcst that the inability of a cell to maintain a particular enqme \\bile metaboli~ing another sub- strate is probably not due primarily to an "inherent instability" in this enzyme but rather to competi- Live interaction between enzyme-forming systems. THE INI-IIBITION OF ENZYhlE l?ORMAT1ON It has been pointed out previously that any pro- cctlure that seriously interferes with cell viability or prcvcnts the cell from nietaliolizing rcsults in a loss of enzymatic adaptability. This only tells us that adaptation, like other cellular synthetic pro- cesws, requires a properly functioning metabolic cycle to supply the energy for the proper reactions. It tells us nothing of the nature of the linkage be- twcn the adaptive process and the metabolic re- nc t ions. One method of approaching this problem was to attempt to dissociate the process of adaptation from the over-all metabolic cycle. In such cells, where enzyme formation is prevented without af- fecting the over-all metabolic rate, one could hope to analyze the nature of this link. Dissociations of this kind between synthetic processes and metabo- lism have been accomplished with the aid of various compounds. Thus the ability of NaN, to-prevent the utilization of metabolic energy for synthesis is quite gencral, having been demonstrated for such diverse processes as cell division, embryonic development, regeneration, and carbohydrate and ammonia as- similation. It was of some interest to determine whether enzymatic adaptation behaved in a manner similar to these synthetic processes with respect to inhibition by azide. Since azide is a powerful inhibitor of aerobic nietal)olism, it was trivial to find that it inhibited aerobic formation of enzyme. Putting azide into the atlapting mediuni is cquivalent to establishing an- wroliinsis, :rnd \vc hnvc nlrcntly not ctl llint niincro- hic cvmlitioiis, if ins[ itulcrl froin tlic I)cgiiining, nrc cflcctive inhibitors of cnzynic formation. Of grptcr interest were the experiments (22) in which the effect of azide was tested on anaerobic synthesis of enzyme subsequent to a period of aerobic incuba- tion with substrate. Fig. 5 describes such an experi- ment. Cells were incubated aerobically with galac- tose until the gnlactozyni;i.;e xtivity rc*;iched a QNC02 value of 40, artcr which ;in:!cm)l)in,4s was instituted and inculmtion ivith the sulistrntc con- t inued. Control susi)erisions continuctl to increase their cnzymc cuntcnt in a minner tlciiictcd by the Iirst thrce points ml Lhc da;;hctl part of the curve. If at any point -in the tlcvcloprncnt of this enzyme activity azide is intruducetl in conccntratiun of 5 X ILI or higher, no further enLyine is formed 160 4 I S 2 4 6 0 Houre FIG. 5. Effect of NaNI (5 IO-', hl) on anaerobic forrna- Lion 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 aliscnce of oxygen pro- viding a small amount of glucose is furnished. This utilization of the energy obtained froin 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 oiily thc synthesis of somr eiiLyrnr, Init :11w its tlis:i1)1"':i~;iti"C, 110 1n;il kr wl~it sul)s[ra(c the ccll was nwtnliolizing nor how fast it was doing this. Fig. 6 gives the results of an eqerinient 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 galactozyniase activity disappears quite rapidly. If at niiy time in this process, howcvcr, ,wide is atltlcti, further disappearance of this cnzyme ccases iniinct1i;itcly and thc c ,zymc activity rcninins in- dclinitcly at thc level rcachcd whcn the azide was adtlcd. 'I'hc fact that the samc compounds or con- ditions that inhibit enzyme formation also inhibit rate. J.Io\ccvcr, ivhcrws the con1 rul cxhihitctl a relatively rapid uptake ;:nd c:iclli~ng~ of inol-gniiic phoqha te, the cspcri men t a1 d it1 not. C h cm i cal an- alyscs fur total, inorganic, and oi.ganic phosphatcs, and radic~a4vc clainination of these iractions, con- firmed thcsc rcsults. Azide is able to prcvcnt the 5- 2 4 6 0 Houre FIG. 6. Complete stabilization by NaN: (5 x lo-' 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. TIIE hfEC1IANISM OF AZIDE INIIIBITION OF ENZYME SYNTIIESIS AND ITS RELATION TO PIIOSPI-IORYLATION- The ability of azide to prevent the forniation of cnzyme 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 NaN3 on phosphorus metabolism, using radioactive phosphorus (P") 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 linown amount of tracer P and allow the cells to fer- ment glucose anaerobically, 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 les 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 FIG. 7. Effect of NaN: (2.5 x 10.' M) on P turnover dur- ing anaerobic glycolysis. Ordinate of Iower 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 NaN1 (open circles) and in the prexnce of NaN: (2.5 x lo-' 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 KaN, 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 cnergy- rich organic phosphate bonds, Further experiments wcre perfornied (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 inorgacic P into FACTORS CONTROLLlNC ENZYMATIC CONSTITUTION 267 :! e c.lrI)ohydr;ite cycle in two places: (1) phos- i,:ioroI~ais of glycogen; (2) the couplcd oxidation ph~qdiorylation of glyceraldehyde phoiphate to di-~liosphoglyceric acid. 'rile involvement of the first stcp sccmed un- !,~clg, since no polysxcharide is synthesixd by )ca~ cells in the prcwice of aide (38). Since it ,,cnlcd most probable that the aide was affecting the .ciond stcp, experiments \\ere devised to cxam- ;nc the bcliavior of the glycolytic cycle at this level in the presence of uide. It is well known that i(&;icetic acid (MA) is a strong inhibitor of the cnzj me (triose-phosphate oxidase) controlling this ,;c.p. It was rea\oned that if aide modified the i~ho~phorylative mechanism at this stcp it might be expected that a parallel change in the sensitivity of the fermentation to IAA might appear, and such I\L~ actually found to be the case. A concentration of 1.4A (2 X lo-' M) which was sufficient to in- hibit fermentation completely within 10 minutes in the absence of uide took 180 minutes to bring the rate down to zero in the presence of 5 X ?.I NaN,, and failed to affect the rate at all for the ht 70 minutes. Lower concentrations of azide prc- trcted against IAA inhibition to a lesser extent. Iligher concentrations continued to lengthen the period of protection until inhibitory concentrations of uide 1% ere reached, whereupon the protective action against IAA disappeared. It should also be noted that uide is conipletcly unable to protect fcrinentation against the inhibitory action of fluor- ide, which poisons an enzyme controlling an en- tirely different step. SUCLEOPROTEIN AS A CONTROLLING ELEMENT IN ENZYME SYNTHESIS .hy attempt 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- t\\ cen phosphate metabolism and enzynie 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 bad to be obtained on what happened to the varioiis organic ph(JSphate compounds while a protein or enzyme was 1Jcing syrithcsized. Experiments were dc?ipietI (25) to ohlain such information, again using ra(1iu;ictive phosphate. In thesc experiments the phosphate in the various fractions of the cells was tagged with tracer before the bcginning of the experiment so that subscqucnt movcmcnt of the P from one frac- tion to another could be follo\wd. The most consistcnt corrclation bctween phos- phorus mcta1)olism and protein or enzyme forma- tion was found in the flow of phosphate jroiit the nucleoprotein fraction (NP). This latter is the residue phosphate remaining after siiccessive ex- tractions with water, cold trichloroacctic acid, al- cohol, and hot alcohol-ether (3: 1). The behavior of the phosphate in this fraction was followcd under various Conditions, employing in the following manner. Cells were grown in the usual media at 30' C. in the prescnce of Pa' (activity, 5 X lo5 cts./min./mg. P). This resulted in coiiiplete cciuililiration of the lahclcd phosphorus in all fractions. After 48 hours thcse cells were har- vested, washed three times in unlrtljeled hl/l5 KH21?04, resuspendcd in unlabcled M/15 KH,P04 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 ycre with- drawn at intervals for activity men~iircinents. 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 content) as well as total P31 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 (KP) fraction even though the major portion of the re- maining phosphate mas 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 per-iods 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 the t1:it;i in Fig. 8. In this c.\pcrimcnt cells wrc sus- p~~ii(1ed in physiological s;iIinc containing: (a) glu- cow, (b) glucose C (NH,),SO,, (c) glucose + (SlJ,),SO, 4- NaN,, (d) glucosc -1- (IVIL)2S04 -!- tlinitiophcnol. 'The aniount of (NI-I,),SO, was equivalent in nitrogel] to 50% of thc nitrogen con- tent of the yeast. ?'he conccntrations of NaN, and tlinitrophcnol \yere 5 X antl 5 X lo-' rcsl'cc- tivcly, sufiicicnt to completely inhibit enzyme for- mat ion. It \vi11 be noted that with glucose alone there was no change in activity, whereas when an~monia was present, with conscqucnt budding, the nucleoprotein P diopped to 3S70 of its original total activity, Phphata Acllvlly. 2 4 6 noum FIG. 8. Flow of nuclcoprotcin phosphatc in cclls suspcndcd in (A) filucosc, (B) glucosc $. (hN,)SO,, (C) ducosc + (h'I1,)SO. + NaN., (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) ijgcnts that arc cffcctivc in inhiliiting el-tzymc fctrm:ition arid 1JKJLcin syntltesis a1r;o prcvcnt fl(J\V of I' from the SI' fraction. 'To the must be ;itlticd the fu11tl;1111~11lal ob- servations of Caspcrsson and his colla1)c~r;itors on yeast (2), as well as on many otllcr cells, which point to a rigid connection bctwccn nucleic-acid metabolism antl protein synthcsis. A ~1:ol)osl:u PU~C~~~~~ or: ~ucr,l:olwo.lI:l~ 1N The data prcscntcd in the previous section leave little tloubt that nucleoprotein mctabolism is es: scntial for the synthcsis of cnzymcs and proteins. `The question that naturally arks is wl-,at role the 1111~l~oprot~in plays in thcse synthctic processes. Is it a matter of energy supply, or substrate, or specificity? No definite answcr is available. I-Iowcvcr, modern biochemical research (13), which has crnphacizcd' the rnle of organic pliosp1r;ite 1)nntls as smlrces of cwrgy for syn~lit:~ ic ;tct ivilics, pr-ovitlm a rnuntla- lion ii1)on which inny 1JC linsctl ;L r~xmmhlc hy- po thesis of 11 I iclcoprot c in ,function in c nzy nr form 3- tion. Of particular value hcrc is the incrhnnisni of complex polysaccharide synthesis, the elucidation of which we owe to the brilliant work of the Cork (3) and their collaborators. Two things arc re quired in the formation of a compound like glyco- gen. One is, of coiirse, thc basic 1iczol;e iinit. The other is the encrgy neccwary to form the lionds linking thcse units into the complex polysncchnride. The fundamental contribution of the Cork was to show that in the synthcsis of a glycosidic bond, glucose-1-phosphate rather than glucose is the re- actant involved. This phosphate ester of glucose (Cori estcr) already contains in the (C-0-P) link the amount of energy required for the forniation 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 iitolecular stnrctrire of me of the react ants. 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) wXich have more than sufficient energy to form a glycosidic link if there actually existed sqme mechanism for "feeding" it directly into the reac- tion, No such mechanism exists, however, and EN zshi E SYN ,r I i ESI s tl,c,rcforc tlic ciicru content of these conipountls L,lnnot he used for this pur1)o.w. Clearly, in xldi- r;on to the purely qliantitative aspcct, thcrc is ,,hat may be called the "specificity" of the bond 1211crgy. Thus, the actual natu~e of the bond, and cuinpouiid carrying the cncrgy, will detcrmine i:s ,uitability for driving a p;irticular reaction. The lll(~dgy geneiated by the ~`cataholic \~hecl" and :r,~ppl in such energy accuniulators as adcno4nc- i:;phosphatc or creatine-phosphate cannot be used .is such in driving all the vaiious synthetic mecha- nims of anabolism. The energ)r_containcd in such cc~nipounds must first be transferred to others, nhich cm then act as specific energy donors for p~rticular SJ nthctic reactions. `fliis C(J1lCCpt unifics and sirnplifics the problem 01 I)ioIogic:il synthesis, since it avoids scpai ating the problem of synthesis into one involving the rc- xtants and anothcr concerned with the source of the "coupled" driving energy. This principle tells 11s that the solution of one neccssarily leads us to rhc solution of the other aspect of the publem, >incc ~hcy arc one and the same. I:r0i11 this point of view and thc establishcd hiiiort~iiicc of phospho~ylatcd compounds in syn- rhctic rcactions, it is not surprising to find that nucleoproteins are controlling elements in enzyme synthesis. We may further plausibly suggest that tIic`.;e phozPliorus-containinS proteins are the spe- cific enerp donors vhich make possible reactions !cu3ing to protein and enzyme synihesis. EFFECTS OF NUCLEOPROTEIN FRACTION EXTRACTS ON ADAPTATION One must be cautious in drawing any final con- clu4ons about the precise role of nucleoprotein either from the preceding data or from the discus- 4on. Even granting that nucleoproteins are directly in\ol\ed in the synthesis of proteins, it does not iollow that they necessarily intervene in the final 5tcps leading to the conversion of such protein :iiolecules 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- niatic specificity by a nucleoprotein component. The most critical experiment one could offer in this direction vould be one analogous to the already classical investigations of Avery and McCarty on pneumococcus transformation; Le., the induction of a particular enzyme with a nucleoprotein com- ponent in a cell lacking the homologous gene. No such experiments have yet been successfully 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 / oI,scrve slxcific cffccts of wch fr;rcl ions on at1;ipt- ing cclls. Various types or frac:ion;ttion wre t rictl, and I slinll hue rq)rjrt Ixicfly 011 tlic procc(1urc that yicltled the most act ivc pIqxir;itions. Cclls wre cstractcd with KaIICO~, at pII 9.0 for 3 hours at 30° C. with constant stii-ring. On re- moval of the cells the hupcrnatnnt was ntljustctl to 1311 3.5 and the result in:: precipitate wn;hctl 2nd rcdissolvcd. This was follo\vcd by Lwo su~JAYliIcnt precipitations. 'rhe activity of thcse fractions was tcs:cd in the following manner. Unadapted cclls \vcre ruspcntlcd in buffer at pH 7.8, to which had bcen atlded `the adapting substrate, biotin and p~~itoih~ni~ acid, S-i/cc., and 1 nig. of nitrogen per cc. in the form of (NJT,),SO,. `rhc last three coiniminds vicre added to insurc that a `source of nitrogen \vas avail- able and utilizable. It \;.as espcrinicntally found that the presence of these substances provided the optimal conditions for testing activity, since certain preparations which were inactive in their abscnce were cstrenicly active \vhci1 they were ;idtlcd. It was found that, when cells a(1:q)tiiig to galac- tose were incubated with such NP fi-actions prc- pared from galactose-adaptcd .cells, the adaptation time could be cut down from 1SO minutes to 20 minutes. Similar preparations from normal un- adapted or maltose-ada;-Jtcd cells poc:wwl no 5 ti rn ul a t o ry ql ;I r: t 0; ,c -:I (1 ;! j , I in g suspensions. Jn a i.irnilar rrimficr !r;:(.:iori,. frorn maltose-arlaptct! cell:, I:C:C only ~a[JdJlc of :,?irrIu- lating adaptation to maltose, and the formation of this enzyme was not affected by the addition of extracts from galactose-adapted or nomnl cells. That the enzyme itself is being transferred tb the unadapted cells in this fraction seciiis vcry 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. hIore tlccisive, howver, is the fact that neither could confer the spccific activity on active glucozyma5c extracts ohtainctl from unadapted cells. i'icithcr, therefore, containctl in detectable amounts an apocnzymatic or co- enzymatic component of the specific cnzymc in- volved. The conditions for obtaining unifoi-inly active preparations have not yet been coinplctely dcter- mined. Since the discovery of this phenomenon about 5 months ago, 32 separate prcparations have been made. Of these, 21 have eshibitcd specific activity in increasing rate of enzyme synthesis. The range of activities observed in the active prepara- tions has been large, extending from Soy to 0.001~ per cc. for' obtaining maximal stimulation. Noth- ing is lrnown at present about the actual active component in this mixture. In view of our ibqorance of the nature of the act ivi ;y f o .,*;a r (1:; 2 70 S. Sl'Il!GE LMA N nitivc com~mncnt, it is clear that nothing can be szid about i(1rntifyjng it with nuclcopro~cin. The intcrcsting and suggestive fact to cmcrgc from thcx cxpcrimcnts is that it is possible to cxtract in the nuclcoprotein portion a component pcculiar to atlaptcd cells, which can specifically stimulate the formation of the same cnzynie containcd in the cells from which the fraction originated. Until fur- ther information is obtained on its Iiocl~cniical identity we may call this active component by the neutral tcnn "adaptin." A TIIEOKY OF GENE ACTION As was pointed out in the introduction, the pri- mary purpose of the prcsent investigntion is to acquire infoimation that niay lead to an adequate concept of gene action. It is proposed in the prcsent 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 niechnnism to be described must be regarded only as a tent a t' ive working hypothesis- -its uscfulness 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 relating to genes and enzyme fornia- 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;ha\S as their primary function the indefinite retention for the cell of the potcizfiality 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 cytoplwnic :clf-tluplicntin~ tnit will in large part detcrminc the nnturc of the hy- pothcsis dcviscd to c?:l)lain the irirch:inis~n of genic , control of cnzymatic cnndtut ion--i);irticuIarly since the same hypothesis must afford sonic under- standing of the rclations bctwcn such units and the genes. It would be hazardous at prcxnt to :~!tcnipt to offer a definitive Jori~iulation of what we mean by a "self-duplicating" unit. However, one attrilmte such a unit is likely to possrss is the ability to trans. form and accumulate energy within its ow1 molcc- ular structure which can he used for the synthesis of similar units. At any ratc, it is relatively casy to show that the growth kinetics of such "enc`rgy a~~~niulat~r~~' is of the self-tluplicnting 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. Presumalily thesc protcins could, in addition to aiding the foiination of other proteins, drive ~~r~~~cin-synthcsizing reac- tions towards the foimation of units like them- selves. In view of our interpretation of the role of nucleoprotein in protein synthesis, nnd of the above discussion, it seems reasonable provisionally to identify the self-duplicating plasmagenes which mediate cnzyme synthesis with the nucleoproteins. Such a hypothcsis ~vould be in harmony \vith the findings that all accepted self-duplicating cntities 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 cmpli,zsized that stating that thcse cytoplasmic self-duplicating units are nucleoprotein in nature does not imply that all nucleoprotcins 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 consequcnces. It i~ould 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 nielibiase synthesis w~as not ohtained 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- :ni.r.;Ion of the capacity to form cnzyme cCm occur in 12;r almnce of any rneasurable enzyine activity. 1:rLrni a thcoretical point of v+w, the nucfeo-' j,rut~in concept of the plasmagene more or less ,i:,-t;tics its relation to the gene. In view of the ,:c>unicd siinilarity bctirccn the two, it seems al- j.:o?t nwxary to conclude that the self-dupli~iting r;ililcoproiein in the cytoplasm, which merfiatcs the j,,:mniion of an enzynic, is derived li-om the gene. ()ne iinniedi;rte value of this conclusion resides in lllc iact that it provirlcs us with an espcrimentally ::nnlyzablc and testable entity, which can bridge rfic fi"p betwen the gcne in the nuclcus and the iii;:y,ne in the cytoplasm. \!re are thus led to propose the following concept or gzne action. Gencs cnt~tirz~rafly produce at various rntcs more or less complete replicas of themselves, irjiich enter the cytoplasm. These replicas or plas- fingenes are nucleoprotein in nature .and possess CY TOPLASM I. Pt, E,S,,+P' +M +Sf > 2PI, E, s, FIG. 9. A mechanism for gene control of enzyme synthesis. The symbols are G, for gene, PIz for plasmagene, Pr for enzyme precursor, E, for enzyme, SI for substrate, and M for cytoplasmic material used in duplication of plasmagenes. The constant (k) is a reaction-velocity constant, measuring rste of production of plasinngencs from gcnes. All double arrows indicate a self-duplicating process; broken-%e arrows denote decay to inactive protcin (IP) of the elements from which the arrows point. Where fo~ard-and-b~ckward reac- tions are denoted, the longer of the two arrows indicates the major tendency of the reaction. to varying degrees the capacity for se~f-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 tcrms of one gcnc G, and its corrcspontling enzyme E,. All the tloiible arroivs tlcnote ~el~-d~~i~li~~~i~i~ re- actions. Gcnc G, cc~n tirrunlly proOuces its plasma- genes (PI,) at a rate clenotctl by k. The 1)l:isnia- gcne by its very nature must possess hetcrocatalytic potcntialities; i.e., in addition to hcing auto- synthetic it must possess the capacity for catalyz- ing the synthcsis of units (enzymes) other than itself. Consequently, once it is in the cytoplasm several things may happen to plasmagene (1'1,). If it is successful in obtaining the proper iiiatcrial (M) in the cytoplasm it will duplicate itself. It may, on the other hand, combine with prccursar protein (Pr) arid convert it to El, resulting in the fori~iat~on of the Pl,E, coni;)lex. Since little enzyme is found cxperimcntally in the absence of substrate, one must assume that this complex is highly unstzble and quickly breaks up into its two coinponents. A pIasr-.sgcne once formed cannot, of course, exist indefinitely, particularly in a population of other such units scti\dy comptting for the material of which it is coriiposctf. The re- action leading from P1, to IP (inactive protein) in the figure describes this fact. It \yill be noted that in accord with the fact described in prcvioris sec- tions the same is assumed to be true for the E, formed. By IP or inactive protein \ye nican merely that the plasniajyne has broken doivn to a protein unit which has lost the capacities for self-duplica- tion and enzyme format.ion, 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 PI, production by GI wcre extrzmely high or the stability of the enzyme or enzyme- plasmagene complex very great. h'either condition is apparently satisfied in the cases of the enzymes reported on here. MThen 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 pk%- magene-enzyme (Pl,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 de~nition of a self-duplicat- ing entity is the concept that, should such a unit undergo a modification at any given moment, all subsequent repIicas would bear this niod~~cation. Thus when PI, .exists alone it. duplicates only PI,, 2 72 S. SI'lIiCBl,AlAN but \\lien ~uhtrntc is atltlctl and the stalilc PI,E,S, canil);nnt ion rcsults this is now duplicated, which is indicntcd in the (Ii,igi`irn by the double airow. 'I'his schcnic provitlcs, tho cfore, a concrcte niccha- niwi uhcrcl)y subqtrate can modify competitive in- teractions bctwcn plasmagenes. What is eisential- ly accornplishcd by substrate is the creation of a new self-duplicating unit, which duplicates not only the plamagcne but also the enzyme corre- spmtling to the wbstrnte arltled. It is clear that this scheme \\oultl fit both the genetic and the kinetic data di~cu~scd in this paper. From an cap imcntal point of view this niecha- nisin suggests that the cnzymcs of a cell should be found aqsocinted ivith the nuclcoprotcin components, in agreeincnt n ith the data recently accun~ulated on the cnzyinatic con\titution of the nuclcop-olein- containing cytoplasmic granules. MOIC important, it offers a possible expeiiincntal solution to a very ii ritating dilemma nhich attends any effort at itlciitifying a particular bio1o:;ically active molecule from a population of clirniically similar ones. This is iwll illu~tiatcd by the picsent-day situation of the ~~ncu~nococcus ti ansfoiming piinciple. Only a relaiivcly small number (about 1 in 10,000,000) of all the niolecules in Zn active preparation of desoxyrihonucleate is actually capable of inducing the specific transformation. The biological test for the picsence of the active principle is so much more 5ciisitive than any available chemical tests that pokitive identification of the cheinical identity of the active molccules 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 thcory sug- gests, however, that the biological specificity of the "adaptin" may be employed as an analytical tool by forming the PES 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 enzymcs and sub- strates. A further advantage which is suggested is that the various stages in the fractionation and iso- lation of a particular PIES coinplex can he followcd 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 transmiltctl frorji onc cell grncr:rtion to another in R hlcntlclian I;i::liion \yj\\ thiis [ley-nd on the rclntive rates of tlu~~lic;rtion 01 the controlling cytoplasmic units as conil);~retl \yith thcir rate (k of the diagram) of production from thc genome. If the latter is quantiIatively dctcrniin. ing, XTcndelinn inhcritance will lie obscrvctl. If the forincr is tletcrniining, the Xlcntlcliaii picturc \~il] be obsctlrcd to varying dcgrces tlepcntling on {lie self-rluplicating capacity of the placniaprs. It is also clear how sL1)strate could so intensify cyto- plasmic inheritancc of a particular cnzyme as to completely obscure thc scgrcgntion of the corre- SlXJIl d i 11 g gene. As a tentative working hypothesis, this ihcory has the atl\witage of providing a iinifictl point of view from which such diverse and apparent ly con- tradictory phenomena as classical Mcnddian p- netics, cytoplasmic inheritance, cclliilar difrcrentia- tion, and enzymatic adaptation may be analyzed. The basic problcm of cancer involves esplaining the appcarance of a sudden licritnhle change in so- nintic cclls, analogous in several ways to eiizgnie adaptation or cellular differentiation. It is, there- fore, not surprising that cancer investigators were one of the first groups of biological worl~ers to support strongly the suggested existence of a cyto- plasmic hereditary unit. An entity of this kind, by being self-diiplicating, provides them with another level at which a mutation can take place and be suhseciucntly 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 cnipha- sized several difficulties in trying to explain either growth or cliffercntiation in terms of the clnssical hfendelian concept of the gene. Thus, the assunip- 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 intlcfinitely he called "plasniagenes." Again, in connection with cellular tliffcrcntiation Wright (41) pointed out that the heritable stabil- ity of the differentiated state is more easily untlcr- 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 273 cj~~,,$~mic ~elf-(luplicating unit, which he called i~~d:~snxig:cne" and which he assumed ccmtroIs hcrc,tIity at the "niolecular Icvel." It may be of iii~~~ortance for reasons of clarity ci~q)hasizc certain Mcrences Lctwccn the plas- 1IIn;;cne concept tlevcloped here and those cmploycd 1)~ jf`right and Darlington. The plmmngcne as ~~~,~nr(~ in the present paper is a inore or less `~qdete gene-replica, which possesscs to a vary- ing extent the capacity to self-duplicate. It is not a ;pi.ciaI or unique cytoplasmic co~li~#iicnt in fhe sense that it is outside normal physiological proc- &+es. It is an integral part of the enzyme-synthe- &in: system and is the normal link by means of ydiich gcries can effect control over pi-olein iorma- [i#R in the cytoplasm. Whether or not plasmagenes ;ire "molecular" is not pertinent to their definition. ft semis probable, however, that they would per- lorin their synthesizing functions on the surfaces of rclatively large particles (cytoplasmic granules), which could provide the necessary protein and rntxgy-rich groups. RETEREXCES I. AGDEREALUEN, E. Versuche uber den Einffuis der Zuch- 2. 3. 4. 5. 6. 7. 8. 9. 1 0. 11. 12. 13. 14. 15. tung von IIefe auf Galaktose 3uf die Ver$irbarkeit diescs Kohlenhydrates durch diese. Fernientforschung. 8: 12-55. 1925. CASI*I.RSSOX, T., and BRAXDT, K. Nucieotidumyitz und \i`adis:um bei Pm~hefe. Protoplasma 35 : 507. 1040. CORT, G. T., Smhssox, hil. A,, arid CORT, C. F. The mechnnism of forination of starch and glycogen. Fed. Pruc. Amer. SOC. Exp. Biol. 4: 234-241. 1945. DARLINGTON, C. 0. Heredity, development and infec- lion Nature 154: 164-165. 1944. DIIXERT, F. Sur la fermentation du galactose et sur l'accoutuniance des lcvures B ce sucre. Ann. fnst. P`uteur KARSTR~M, H. Enzymatische Adaptation bei Mikroor- ganismen. Ergebn. Enzyniforsch. 7 : 350-376. 1938. LIXDEGREN, C. C. Mendelian and cytoplasmic inheri- hnce in yeasts. Ann. Missouri Bot. Garden 32: 107-323. 1945. LISDFCEEN, C. C. Yeast genetics. Bact. Rev. 9: 111- 170. 1945. LTKDEGREN, C. C., and LINDEGREN, G. Seiccting, in- breeding, recombining and hybridizing commercial yeasts. J. Bact. 40: 405-419. 1943. LINDEGREN, C. C., and LINDEGREN, G. A new method of hybridizing yeasts. Proc. Nat. Acad. Sci. 29: 306-308. 1943. LINDKREN, c. C., and LINDLGREN, G. Segregation, mu- tation and copulation in Saccharooayces cerevisine. Ann. Missouri Bot. Garden. 30: 453-463. 1943. LINDZGREN, C. C., SPIEGELMAN, S., and LINDEGREN, G. hhdelian inheritance of adaptive enzymes in yeast. Proc. Nat. Acad. Sci. 30: 346-352. 1944. LIPMANN, F. Metabolic generation and utilization of phosphate bond energy. Advances in Enzymology 1 : 99- 162. 1941. MONOD, J. Recherches sur la croissance des cultures bactcrienncs. ActualitG Scientifiqucs et Industries. 210 pp. Hcrmmn & Co. Paris, 1942. MONOD, J. Inhibition de Sadaptation enzyrnatiaue chez -- 11: 139-159, 1900. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 3 5. 36. 13. coli en pr~`wnce (le 2-1 rlir,~trrt~Ji~nol, Ann. lnst. P;istrur 70: 319-384. 1044. TO NOD, J. Sur la nature du IJ,inotrrFne tlc &:liixic. Ann. Inst. P:atcur 71 : 37-10. 191.5. SC~IULTZ, A. S., ATKIN, I,., and FYLY, C. N. Innurnre of oxygen on tlie fcrinvniaticm .,I i:nftose :1nr1 g:s1;tctosc. J. Amer. Chem. SOC. 62: 2271-2272. 1940. SONNLIICJRN, T. M. Tlic dcpcnrlcnce of the p1iysiolo;:ical action of a gcne on a primer and the relation of primer to gene. Anxr. Nnt. 79: 318-339. 1945. SIW;C.I.LMAN, S. The physiology and gcnetic sipifirnnce of enzyrnntic adaptation, Ann. Xliswuri Bot. G;rrtlen SIWCLXXIAN, S. The eflcct of anneroliiosis on adaptive enzyme formation. J. Cell. Comp. Physinl. 25: 121-131. 194s. SIWXILAIAN, S. The physiological propcrties of nd:rp- tive enzynie formstinn. J. Ract. 49: 10%. 1915. SIWGELMAN, S. Inhibition of eiizyinc formation. Fed. Proc. Amer. SOC. ESP. Bid. 5: 99. 1916. SPIEGEI.XIAN, s., and DUNN, R. Competitive inlcrac- tions betxveen enzyme forming systems. J. Cell. Comp. Physiol., in press. SPIEGI:I.PIAN, S., and KAXEN, hl. D. The site of un- coi~pling of p!iosphoryl:ition from r:irlioh?dr;ttc mrtnb- olism in the prcxncc of NaN,. Fed. Proc. Amer. SOC. Exp. BioI. 5: 99. 1946. SPIEGELMAN, S., and KiAbrw, hl: D. Genes and nucleo- proteins in the synthesis of enzymes. Science, in pres. SPIEGELMAN, S., KAMEN, M. D., and DUNN, R. Mech- anism of azide inhibition of synthctic activity and its relation to phospIiory1;ition. Fed. Proc. Anier. Soc. Esp. Bioi. 5: 100. 1916. SIWGEL~~~AN, S., and ~,IXDIIC.REN, C. C. A compsiison of the kinetics of enzymatic adaptation in genetically homo~eneous and heterogeneous yeast populations. Ann. hlissouri Bot. Garden 31 : 219-233. 1914. S~~~CELMhN, s., and LINDECREN, c. c. The rehtion of sporulation and the range of variation of the hnplo- phase to populational adaptation. J. Bact. 49 : 251-269. 194s. Sx~IEGELPfAN, s., LINDEGREN, c. c., and HEnccofK, L. Mechanisms of enzymatic adaptations in genetically con- trolled yeast populations. Proc. Kat. Acad. Sci. 30: 13- 23. 1944. SPIEGELMAN, S., LWDEGREN, C. C., and LIXDECREN, 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. SPIEGELMAN, S., and NOZAWA, M. On the inability of intact yeast, cells to ferment their carbohydrate reserves. Arch. Biochem. 6: 303-322. 1945. SmrxwpsoN, M., and YUDEIN, J. Galaclozyniase con- sidered as an adaptive enzyme. Bio-chcm. J. 30: 506-514. 1936. STIER, T. J. B., and STANNARD, J. N. The metabolic systems involved in di~imilation of carbohydrate reiervtls in baker's yeast. J. Gen. Physiol. 19: 461-477. 1935, v. EULER, H., und NITSON, R. - tlbw die Galaktosever- gsrung durch Hefe nach Vor~eliandlun~ mil dieser Zuckerart. 2. physiol. Chem. 143: S9-107. 1925. WINCE, O., and LAUSTSEN, 0. 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 LAWSTSEN, 0. Artificial species hybridi- 32: 139-163. 1945. 271 s. SP16G E I2 A N ??!ion in yc:lcr plasmagcncs must have occurred (luring f1i;;crcntiation to form a ccII so uniquely suited to ;,?:form its biological function.. Jn atltlition to this, it niut bc reIiicnih*rcd that the germ line is isolated .2$ such relatively early in the embryogenesis of h$cr plants and animals. Consequently, any $:imagene that is to disturb the Mendelian niech- ani5111 must survive in the cytoplasm lor a rela- ;ively long period .-c$cnding froin early in em- I,rlonic dcvclopment to~~~sex-ual maturity. The chmccs of any but the "pro~xr') (Le., necessary for the onset of differentiation in the zygote) plas- nlnpes surviving this long waiting period and :hen the subsequent selection accompanying differ- cat intion must be rclativcly small, although not ~~m~dctcly impossible as denionstiatcd by the csist- ence of maternal cytoplasmic effects. Presumably, the survival of the "proper planiagcnes" is en- couraged by providing the appropriate environ- mental conditions and substrates. It is not difficult to understand why the higher plants and animals would tend to develop mccha- :-;is lding to rather rigid control over the plas- :`ne populations of their gametes, and hence to -u?p:csion 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, conplicated series of differentiation reactions dcli- tately synchronized in space and time. Thc dis- iurhnce of my 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- plnsmic clcmcnts in the g;trnctc:s ~\nultl provitlc an import;tnt factor of safcty for the (lc~c11 11mcnlal proccss. O1J\'i(IUSly, where tlc\cl0i1n1~~1t a1 procc..isrs lcatling to the adult ,~mn arc cithcr tri\ri:il or al- togcthcr n;jnexistcnt, `xs in the single-ccllcd forms, control over gamcte plasmagcnc population would confer rclatively little selcctivc atlvaptage. Dr. Stcrn raised the question of the relation of the gcnc dosage effect csperirncnts to the rcsults and thcory of gcnc action pi-e.;cntcd here. \\hether or not a gene cloyage dfect with a Iiarticiilar gene will he obscivcd will depend on whethcr the rate of production of the plasmagene from the gene, or the rate of playmagcnc self-duplication, is determining. If the latter is low, citlier because of an inherently poor capacity for autosynthesis or ~JrcaUse of com- petitive conditions in the cytopin?m, thcn increas- iag the dozage of this gene in the nuclcus will in- crease the number of the correspnding plasma- genes that will be found in the cytoplasm. On the other hand, no dosage effccts ivill lie olwrved with ;my gci~ whose pl~i~inngciie pos~csscs n sclf-diipli- cnting capacity high cnciiigh to qiiaiit itativcly ovcr- whelm the prodiiction rate of the gcne, in so far as dcterinining 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 mechanicni of gene aciio~ 2nd not of gene ft'Q72S- nzission. The prcwnt theory a.;su~nes that the latter is strictly llenrlclian and is SUllll~eJ11ciital-y to the established rules of transnnission. Inherent in the proposed mechanism of gene action is the possibil- ity of explaining why in certain instances the trans- mission of a cJ/nrncter may not follow that of the geize. \Vc soinetinncs call such cam instnnces of non-Llendelian inhcritance. Act~ially, since classi- cal ATcndelian thcory refers only to nuclear gene transmission, such phenomena do not represtint violations of Mendel's laws.