FURTHER STUDIES OF GENE-COKTROL SYSTEMS IN MAIZE Barbara McClintock Continued study of the Spm system of control of gene action in maize has pro- duced additional evidence of the system's versatility. A control system of this type is capable of regulating the action of many genes during development. In view of the large number of genes present in the nuclei of higher organisms, and the need for coordination of their action during development, it is reasonable to assume that evolutionary processes would have initiated superregulatory mecha- nisms to accomplish that end. The Spm system serves as a model of the mode of operation of one type of superregulatory mechanism. Such a system can activate or inactivate particular genes in some cells early in development, and activate or inactivate other genes later in develop- ment. It can turn on the action of some genes at the same time that it turns off the action of others. It can adjust the level of activity of a particular gene in different parts of an organism. Some evidence to support these statements has been presented in previous Year Books and other publications, and addit,ional evidence will be reported here. At the time they were first detected, controlling elements were distinguishable from genes because, unlike genes, they could be transposed from one location to another in the chromosome complement. Thus the operator element of a regulatory system may be inserted at the locus of a gene, whose action is then subject to control by the system to which the operator element belongs. Because the detected controlling elements of a system may reside at different' locations in the chromosome complement, they have been likened to episomes in bacteria. This comparison is justified, since it is known that the activity of bacterial genes may be subject to control by episomes. In maize, however, the operator element of either the AC (Activator) or the Spm (Suppressor-mutstor) system may be- come integrated into a chromosome in such a way that it no longer undergoes transposition. In its fixed position it continues to respond to signals from the regulator, and its responses are made evident by changes in action of the gene with which the int'egrated operator ele- ment is associated. At present, there is no direct evidence for or against episomal origin of the components of known gene- Control systems in maize, It is conceiv- able, nevertheless, that some superregu- latory mechanisms in higher organisms may have originated through incorporn- tion and adaptation .of the gene-control components of episomes. The AC and Spm gene-control systems are not considered to play a vital role in coordinating gene action during develop- ment of the maize plant, at least as far GENETICS RESEARCH UNIT 487 as the known states of the component were control systems so suddenly revealed elements are concerned. If they did play and at so many different gene loci? It was such a role, it is doubtful that they could concluded that regulatory elements, dis- have been analyzed so readily; for then tinct from the genes, must have been the selected mutants of the regulator present in the nuclei of the maize plants elements, which have provided much before the breakage-fusion-bridge-cycle evidence about the modes of operation of experiment was conducted, and that this the systems, would have so altered coordi- cycle, in some yet unknown manner, had nation as to affect adversely the growth induced modifications in the p!e-existing of plant or kernel. Adverse effects have elements. The evidence was so compelling not been noted. Thus, the controlling that it was later decided to test the elements of the examined systems may conclusion. If it was correct, the breakage- represent foreign, nonessential, episome- fusion-bridge cycle should be able to like components that have been inte- induce a known type of regulatory ele- grated into the maize genome; or, on the ment from some unknown element in a other hand, they may be true chromo- nucleus whose ancestor nuclei, in their somnl components of present-day maize, past history, through many plant gener- whatever their evolutionary origins and ations, had never given evidence of its histories may have been. Some evidence presence. The experiment, reported in supporting the second consideration will Year Book 50, was successful. There was be reviewed briefly. no doubt that D&type regulation could be My study of gene-control systems in induced, independently, in a number of maize was initiated with an experiment different nuclei where the cycle was in conducted some years ago, which uti- effect. It was also evident that each lized the chromosome type of breakage- induction was the consequence of a single fusion-bridge cycle. The unexpected re- event involving some component in an sults of this experiment, reported at the individual cell. After induction, the regu- time, should again be emphasized here. latory activity of the induced element was In the self-pollinated progeny of a number registered in the descendent cells. of plants that had commenced develop- Additional evidence for considering ment with a chromosome pair undergoing that the examined types of controlling the breakage-fusion-bridge cycle, it was elements are derived from components startling to observe that many unrelated normally present in maize chromosomes genes, whose action had previously been comes from studies of induction of change normal with respect to expression in plant in action of a presumed "wild-type" gene tissues, were now exhibiting marked by particular alleles of the gene. This differences in activity in different parts induction effect is termed "paramuta- of plants. The patterns of change in the tion" by R. A. Brink in studies of alleles somatic tissues, regardless of the cellu- of the R gene, located in chromosome 10, lar function with which a particular modi- and is termed "conversion" by E. H. Coe fied gene was concerned (that is, plastid in studies of alleles of the B gene, located development, chlorophyll synthesis, an- in chromosome 2. With regard to either thocyanin synthesis, or capacity of the R or B, the initially recognized allele that cell for growth or division), reflected some induced change in action of a presumed mechanism of control of action of these normal allele would have been placed in genes during development. the "mutable gene" category had its Studies were then begun to examine behavior been analyzed in earlier years. these gene-control mechanisms. Later, Indeed, genes under the cont'rol of the AC? when the components of a control system Spm, and Dt systems were at first were recognized, it was possible to recon- referred to as mutable genes, since they sider the basic question posed by the exhibited phenotypes to which this term results of the initial experiment: Why had been applied in many early genetic 488 CARNEGIE INSTITUTION OF WASHINGTON investigations. It seems most likely that induction of change in gene action of "wild-type" alleles, and the subsequent capacity of some of the changed alleles to induce similar changes, resides in con- trolling elements normally associated with the genes. The resemblance between effects brought about by the inducing alleles, described above, and those produced by elements of known control systems is further illustrated in studies conducted by Dr. Brink and his students with the R alleles. The initially recognized inducing allele, R"`, gives rise to a distinctive pattern of pigmented areas in a nonpig- mented background in the aleurone layer of the kernel. The frequency of occurrence of change from inactive to active expres- sion of the R gene in the cells of the endosperm during its development may be determined readily by examining the numbers of such pigmented areas in the kernel. There is a dominant modifier that markedly increases the frequency of occurrence of these events without alter- ing the time during development when they take place. Initially, this modifier was placed approximately six crossover units distal to the locus of R*i. Its effect resembles in detail that of a dominant modifier element in the Spm system, reported in Year Book 57, and like that element it undergoes transposition. Thus, accumulating evidence obtained in studies of the R and B alleles now makes it reasonable to consider that "paramuta- tion" and "conversion" reflect alterations induced by and in components of gene- control systems normally present in -maize. It would be instructive to deter- mine, therefore, whether or not the individual components of these systems could be identified through transposition and insertion elsewhere in the chromo- some complement. Modij'ied States of aim--2 In Year Book 61 the mode of control by the Spm system of gene action at aim-* was reviewed. As was stated there, the regulator element, Spm, resides close to the locus of the A, gene. When this element is in an active phase, the A1 gene is active, but during an inactive phase of the element gene action is inhibited unless an active Spm element is present else- where in the chromosome complement. When an active Spm is present, release of gene action from the control of the Spm system occurs in some somatic cells, and is accompanied by the production of mutants of the A1 locus, the classes of which were described last year. When the Spm element associated with the aim-* locus is inactive, and no additional active Spm is present in the nuclei of plant or kernel, such mutants are not produced. In this study of aim-2, an occasional kernel on a testcross ear exhibited an atypical phenotype with respect to ex- pression of the Al gene. Plants were grown from some of these kernels, and each was subj'ected to testcrosses in order to explore the cause of the modified gene expression. It could be determined by this means that the atypical phenotypes arose through modification of either the regulator element, Spm, or the operator component residing at the A, locus. Modified states of Spm at the alme locus. There are several readily recognized types of modification of the Spm element. One type changes. its phase of activity; another type alters the time of its trans- position during development; a third markedly alters its capacity to effect responses of the operator element that result in release of gene action from the control of the Spm system. The third type of modification, discovered initially in studies of aim--l and azm--l, was designated Spmw and was described in Year Book 56. With aim--l and the class I states of a*"`-*, the inhibitory effect of Spmw on gene action at either of these loci is pronounced, but its capacity to induce responses of their operator elements that release the genes from Spm control is very much weakened. Few mutants of this type are produced. The type of control of A1 gene action GENETICS RESEARCH UNIT 489 when an Spmw is present at the al--2 kernels, and continued with their progeny. locus is that expected. The gene is active, It was learned that the modified pheno- since the inhibitory effect of Spm on gene type in 4 of these kernels could be action at aim-* is the reverse of that on ascribed neither to an event that had gene action at ulm--l and azm--l. In the removed Spm from the locus of ulmm2 nor kernels, pigment of medium intensity is to change in Spm action. These 4 modified distributed over the aleurone layer. The states of almm2 differ from the original weakened capacity of Spmw to induce state mainly in the frequency of occur- modification of gene action is expressed rence of each of the types of AI gene in the kernel either by an absence of expression that arise in conjunction with mutant areas or by the presence of one or release of the gene from control of the several very small areas showing a mutant Spm system. The remaining 3 altered . phenotype. Thus, Spm" differs from states show no evidence of the presence standard Spm (which in this account will of Spm at the modified al"`--2 locus. Each, be designated Spm*) in capacity to effect however, responds in its own way to an mutation-inducing responses of the oper- active Spm element located elsewhere in ator element of the system. the chromosome complement. Two of Two isolates of aim--2 in which the these states promise to provide new associated Spm# element had mutated to evidence about kinds of control of gene Spmw were examined this past year. The expression during development. Spmw mutants differ in frequency of One of them illustrates the manner in reversion to Spma, one reverting frequent- which anthocyanin pigment of different ly and the other rarely. Previous studies intensities may be produced in individual of Spmw mutants, conducted with plants cells of a tissue. Such differences may be having either aim--l or u~-l, had shown due to "cross-feeding" between adjacent that, when Spmw and Spma are both cells that, individually, either cannot present in the nuclei of a plant or kernel, synthesize pigment or can synthesize very Spml exerts a dominant effect. Spm* is little. The phenomenon is suggested by not altered by the association, however, the distribution of pigmented cells in but can be recovered quite unchangedin kernels having one of the two mentioned the progeny of plants having both the states of almw2 and an active Spm element. Spm elements. Each of the isolates of These kernels have pigmented areas in a al m-2 with Spmw located close to it colorless or nearly colorless background. responds in this same manner to Spma Some areas are not solidly pigmented, but located elsewhere in the chromosome are bounded by a band of pigmented cells. complement, and the response is similar The intensity of pigment in the cells to that seen when Spma resides at the composing these ring-shaped bands is not locus of aim--?. Moreover, tests conducted uniform. Many of the bands have an with one of these isolates have indicated inner row of deeply pigmented cells. The that Spmw is recovered unaltered in the cells on both sides of this inner row progeny of plants in which an indepen- exhibit a decreasing gradient of intensity dently located Spma element is also of pigment, fading gradually into colorless present. or nearly colorless cells. The origin of the Altered states of the operator element at ring-shaped bands, in which some cells a1 m-2. Kernels exhibiting atypical pheno- exhibit pigment intensities approaching types were selected with the intent of that produced by the normal A, gene, finding among them examples in which may be explained by cross-feeding be- the regulator element, Spm, but not the tween adjacent cells that differ geno- operator element had been removed from typically. The shapes of the pigmented the locus of ulm--2. Tests were initiated bands indicate that the cells enclosed by with plants derived from 7 selected each band are descendants of one in 490 CARNEGIE INSTITUTION OF WASHINGTON which the genotype was modified, and the modifications appear to have been con- trolled by the Spm system. The second of these two modified states of alme was mentioned in Year Book 61 during a discussion of alme that con- sidered potential isolates having an operator element of the Spm system but no Spm regulator element associated with the locus. This state was detected, initially, in a kernel that exhibited an atypical phenotype characterized by some lightly pigmented areas in a colorless background. The intensity of anthocyanin pigmentation differed among the cells of an area: some were more intensely pigmented than others, and some ap- peared to have no pigment, but none was deeply pigmented. As was reported last year, there was no evidence of the presence of Spm in the plant derived from this kernel. The results obtained in initial tests conducted with this plant and in more extensive tests of its progeny will be summarized here. It was learned that the phenotype described above appears in kernels in subsequent generations as long as the altered state of almm2 has not passed through a plant generation in which an active Spm element is also present in the nuclei. When an active Spm is introduced by appropriate crosses, the phenotype of the kernels that have received the altered al+-2 locus from one parent and an active Spm from the other parent, and also the phenotype of the plants derived from these kernels, is similar to that produced by the original state of alme2. When test- grosses were conducted, however, with the plants that had received a newly introduced Spm, some of the kernels on the resulting ears exhibited a decidedly modified phenotype, which appeared only among kernels that had received the altered state of aim--2 but no Spm. These kernels had mottled areas, within which many of the cells were intensely pig- mented. The presence of intense pigmen- tation in the mottled areas allowed recognition of confluence of these areas in a manner that gave rise to a clearly defined pattern of pigment distribution over the aleurone layer. Microscopic examination of the kernels showed that the distribution and intensity of pigment in individual cells within an area very much resembled those produced by the kind of cross-feeding between adjacent cells described above. Among the kernels on these ears that had aImB2 but no Spm, a wide range in expression of maximum pigment intensi- ties was noted, although within an individual kernel all the mottled areas showed the same range of intensit,y of cell pigmentation. Among the different ker- nels the maximum intensity ranged from very dark to rather faint. Kernels with strikingly enhanced pigmentation in the mottled areas appeared on ears produced by crosses conducted with each of 11 tested plants that commenced develop- ment with a newly introduced Spm. They did not appear on ears produced by similar crosses conducted with 26 sister plants that had not received Spm. There- fore it is suspected that Spm induced, in many cells of the germ line, a particular type of modification of the operator element of the altered state of aim--2. This change enhanced the capacity of the A1 gene to contribute to pigment formation in cells of a mottled area after Spm had been removed from the nucleus in the meiotic process. In some respects this phenomenon resembles the pammuta- genie effect described earlier. Study of aim-2 has added to our appreciation of gene-control mechanisms in that it suggests cooperation between two or more regulatory mechanisms. With the original state, activation and inac- tivation of the A1 gene are controlled by Spm. Another regulatory mechanism must operate, however, when the gene is active. This fact was revealed early in the study of ulm--?, by the `restricted distri- bution of anthocyanin in plant tissues described in Year Book 61. It is also strikingly revealed by the very distinctive pattern of arrangement of mottled areas GENETICS RESEARCH UNIT in the kernels just described. It is again manifested by the distribution of pigment in the cob: large areas of deep pigmen- tation often appear in cobs of plants carrying the original state of al*--2 and an active Spm element. They are not due to a heritable modification of almm2, as is shown by the results of tests and is also indicated by the fact that kernels above these areas have the same phenotype as other kernels on the ear. A heritable change would be reflected both in the cob areas and in the kernels overlying them, since the aim+ locus in the cells of all is descended from one that was present in a common ancestor cell. Extension of Spm Control of- Gene Action These studies, over a period of years, have uncovered a number of examples of gene control by foreign systems that have not yet been reported because of inade- quate identification. Whether or not the system controlling a particular gene was AC could be determined rapidly, because precise methods of testing were available. Early methods for testing whether or not a system might be Spm were much less direct and, as applied to some genes, were considered too indirect to be efficient. Therefore some of the isolates, after they had been found not to represent AC control, were temporarily dropped from study. When it was determined that wxme8 was under the control of the Spm system, it became possible to construct tester stocks that could directly reveal Spm control of gene action at other loci. Five suspected examples of control of gene action by this system could now be tested with wxm-*. Two, isolated some years ago, involve the C1 locus in chromo- some 9 and are designated elm--5 and c~"-~. Two others, recently isolated, in- volve the CZ locus in chromosome 4 and are designated c2"`-l and c~~--~. The last involves the Pr locus in chromosome 5. This past year, the required combina- tions of gene markers for such tests were 491 Plants that were clmA5 Shl wx/cl shl WX*-~ in constitution were crossed with plants that were homozygous for cl, shl, and wx and carried no Spm. Other plants, Clm-5 wx/cl wx in constitution, were crossed with plants that were cl wxm--8/c1 wx and had no Spm. Among the kernels on the ears produced by these crosses, response of wxm-8 to Spm was correlated with the presence of elm--j in the kernel. These initial tests placed Spm close to the locus of clmV5 in the examined plants. They also indicated that the origin of some of the stable mutants produced by Cl m-5 was associated with transposition of Spm to a new location in the chromosome complement. Further confirmation of Spm control of gene action at clm-5 was obtained from tests conducted with one plant in which Spm underwent change from the inactive to the active phase late in development. On the ears of this plant, some kernels that had received both clme5 and wxmm8 gave no evidence of the presence of Spm except in one sector, produced by descendants of the cell in which Spm had changed to the active phase. A well defined area in the aleurone layer exhib- ited the types of pigmented spots that are produced by change in gene action at n-5. and only in the underlying cells, 2eriv'ed from the common ancestor cell did the starch in individual cells 0; clusters of cells display a phenotype produced by mutation at wxmv8. It may be added here that wxmA8 has also been useful in analyses of change in phase of activity of Spm at the almV2 locus. Responses of the two loci to phase of activity of Spm are similarly correlated, Further Studies of Topographical Relations of Elements of a Control System In Year Book 61, the origin of a two- element system of control of gene action from an apparently single-element con- trol system was outlined with respect to the AC system in its relation to bzm-* and available only in plants having ~~m-5. one of its altered states, Bz~. Initially, 492 CARNEGIE INSTITUTION OF WASHINGTON the regulator element, AC, was present at or very close to the locus of the bronze gene, which is involved in the biosyn- thetic pathway leading to anthocyanin formation in plant and kernel. In studies, with b,~-~, of the relation between trans- position of AC away from the bronze locus and initiation of Bz gene action, one instance was found in which Bz expression appeared in conjunction with transpo- sition of AC to a position in chromosome 9 between the locus of the Bz mutant and that of WZ, approximately 10 crossover units distal to Wx. In one type of test- cross, pollen from a plant that was I Sh Bz(standard) wx Ds/C sh Bz(mutant) AC Wx in constitution was placed on the silks of ears of plants that were homozy- gous for c, sh, Bz, and wx and had no AC. On one of the resulting ears, a single nonpigmented kernel was present in which changes in Wx gene action, from a low level to a high level of gene expres- sion, had occurred in some cells during endosperm development. In order to examine the system responsible for control of Wx gene action in this kernel, a plant was grown from it in the summer of 1961. When pollen collected from this plant was stained with an I-K1 solution, half the grains stained a red-brown (wx) but, unexpectedly, the other half stained a deep blue as if to indicate the presence of a normal Wx gene. It was then suspected that only the endosperm of the kernel that gave rise to the plant had received the modified Wx gene, and that the normal Wx gene had been delivered to the zygote. To verify this suspicion, a few testcrosses were conducted with the plant. It was crossed reciprocally with an AC-tester plant, homozygous for C, Sh, Bz, wx, and Ds (standard location). It was also crossed to plants that were homozygous for C, sh, bz, and wx, either with or without AC in their nuclei, and to a plant whose constitution was c Xh bz wx/c sh bz wx, no AC. The kernels produced by these crosses revealed that the zygote had, in fact, received the modified Wx gene and that its action was under the control of the AC system. Moreover, AC was found to be located very close to this modified Wx gene, which was then designated wxm-g because it represented the ninth detected change in the Cold Spring Harbor cultures whereby the action of the normal Wx gene had become subject to a foreign control system. Results of these initial tests and of others conducted during the past year show that return to a high level of Wx gene action is usually associated with removal of AC from the locus of wxmMg. The results of two types of testcross conducted with plants carrying WX*-~ are given in tables 1 and 2. They demonstrate TABLE 1. Phenotypes of Kernels on Ears Produced by Reciprocal Crosses of Plants Having the Constitution c, w.z*-9 /cl wz with AC-Tester Plants That Were Homozygous for Cr, WZ, and Ds (Standard Location) and Had No AC 9 = heterozygote employed as ear parent; c? = heterozygote employed as pollen parent. Pigmentation of Aleurone Layer Level of Expression of Waxy Gene in Starch of Endosperm Uniformly Colorless Areas in Pigmented Pigmented Background (No AC) (AC) Totals Q 8 Q 3 High level throughout (germinal mutation) 8 11 12 11 42 Low level with sectors of high level 0 0 1834 986 2820 Low level throughout 5 3 0 0 8 Null level (w.2 allele) 1818 1003 30 64 2915 Totals 1831 1017 1876 1061 5785 GENETICS RESEARCH UNIT 493 TABLE 2. Phenotypes of Kernels on Ears Produced by Reciprocal Crosses of Plants Having the Constitution I wwzmeD/ c, wz with AC-Tester Plants That Were Homozygous for 4, WZ, and Ds (Standard Location) and Had No AC The colorless kernels received I, whereas the pigmented kerneIs received the allele cl from the heterozygous parent. 9 = heterozygote employed as ear parent; 3 = heterozygote employed as pollen parent. Pigmentation of Aleurone Layer Level of Expression of Waxy Gene in Starch of Endosperm "oiformly Ul Colorless -' rrgmented (No AC) Colorless Areas in Pigmented Background (AC) Totals High level throughout (germinal mutation) 17 1 1 2 2 0 23* Low level with sectors of high level 1212 117 0 0 418 33 1780 Low level throughout 3 0 0 1 1 0 5 Ku11 level (wz allele) 479 51 1214 110 27 0 1881 Totals 1711 169 1215 iii iii ii 3689 * Six plants derived from these kernels were examined; 4 had no AC, and 2 had AC unlinked to markers in chromosome 9. the location of AC before transposition, and the relation of transposition of AC to release of Wx gene action from control by the AC system. Among the total of 9474 kernels recorded in tables 1 and 2,13 kernels with the wxmVg phenotype showed no evidence of the presence of AC. Two of them appeared on ears produced by crosses conducted with the original plant carry- ing wxmmg. During the past year, plants were grown from these two kernels and each plant was tested for the presence of AC in its nuclei and also for the response of the derivative of wxmmg to introduced AC. The tests showed that neither plant had AC, and that the low level of gene expression characteristic of derivatives of WX~-~ remained unchanged as long as AC was absent. When AC was introduced, however, each responded to it. Mutations occurred in individual cells, whose de- scendent cells exhibited a high level of Wx gene action. The time during develop- ment when the mutations occurred reflected the dose of AC present in the nuclei of the kernels: the higher the dose of AC, the later the time of occurrence of mutations. The genie marker constitution of the chromosome 9 carrying WX~-~ in these two plants indicated that the removal of AC from the immediate vicinity of the original WX~-~ locus could not be attributed to crossing over. Thus, WX~-~ provides another example of origin of a two-element system of control of gene action from an apparently one-element system.