Dr. Britten is on the staff of the Department of Terrestrial Magnetism, Carnegie Institution of Washington, Washington DC 20015, and Dr. Davidson is on the staff of Rockefeller University, New York NY 10021.

Introduction:
Elements of the Model:

Gene:
Producer gene:
Receptor gene:
Activator RNA:
Integrator gene:
Sensor gene:
Battery of genes:

Integrative Function of the Model:
Figure 1: (Types of Integrative System within the Model).
Figure 2: (Overlapping Batteries of Genes).
Fraction of the Genome Utilized for Regulation:
Figure 3: (The Minimum Amount of DNA that has been Observed for Species).

Experimental Justification of the Elements of the Model:
Integration of Physically Unlinked Producer-Gene Activity:
Table 1: (Several of the Functionally Linked Enzyme Systems Present in Liver).
Table 2: (Distribution of Various Enzymes in Tissues of One Organism).
Table 3: (Estrogen Effects on the Uterus).
Evidence for the Existence of Sensor Elements:
Evidence Suggesting the Existence of Activator RNAs:
Table 4: (Nucleus-Confined Apparently Heterogenous RNAs of Unknown Function).
Large Changes in Transcription of Redundant Sequences:
Regulatory Genes Known in Higher Organisms:
DNA Sequence Repetition:
Evolutionary Implications of the Model:
Inactivity of New Genetic Material:
Formation of New Integrative Relations:
Divergence within Repeated Sequence Families:

Summary:
References and Notes:
Reference no. 8:
Additional References:
Other Sites:
Feedback: 

Little is known, however, of the molecular mechanisms by which gene expression is controlled in differentiated cells. As far as we are aware no theoretical concepts have been advanced which provide an interpretation of certain salient features of genomic structure and function in higher organisms. We consider here experimental evidence relating to these features. (i) Change in state of differentiation in higher cell types is often mediated by simple external signals, as, for example, in the action of hormones or embryonic inductive agents. (ii) A given state of differentiation tends to require the integrated activation of a very large number of noncontiguous genes. (iii) There exists a significant class of genomic sequences which are transcribed in the nuclei of higher cell types but appear to be absent from cytoplasmic RNAs. (iv) The genome present in higher cell types is extremely large, compared to that in bacteria. (v) This genome differs strikingly from the bacterial genome due to the presence of large fractions of repetitive nucleotide sequences which are scattered throughout the genome. (vi) Furthermore, these repetitive sequences are transcribed in differentiated cells according to cell type-specific patterns.

In this article we propose a new set of regulatory mechanisms for the cells of higher organisms such that multiple changes in gene activity can result from a single initiatory event. These proposals are presented in the form of a specific, relatively detailed model at the level of complexity which appears to us to be required for the genomic regulatory machinery of higher cells. We make no attempt to arrive at definitive statements regarding these proposed mechanisms; obviously evidence is not now available to support any model in detail. Our purpose in presenting an explicit theory is to describe the regulatory system proposed in terms of elements and processes which are capable of facing direct experimental test. It is hoped that our relatively detailed commitment will induce discussion and experiment, and it is expected that major modifications in concept will result.

Undoubtedly important regulatory processes occur at all levels of biological organization. We emphasize that this theory is restricted to processes of cell regulation at the level of genomic transcription.

We begin by describing our usage of certain terms and their role in the model, and then present the model itself. We then consider relevant experimental observations and certain testable implications of the model. Finally, some general implications of the model for evolutionary theory are mentioned.

Elements of the Model

The following definitions are intended only to clarify the usage of certain terms in our discussion of this model.

Gene: A region of the genome with a narrowly definable or elementary function. It need not contain information for specifying the primary structure of a protein.

Producer gene: A region of the genome transcribed to yield a template RNA molecule or other species of RNA molecules, except those engaged directly in genomic regulation. We are using this term in a manner analogous to that in which the term "structural gene" has been used in the context of certain bacterial regulation systems (5). Products of the producer gene include all RNAs other than those exclusively performing genomic regulation by recognition of a specific sequence. Among producer genes, for example, are the genes on which the messenger RNA template for a hemoglobin subunit is synthesized, and also the genes on which transfer RNA molecules are synthesized.

Receptor Gene: A DNA sequence linked to a producer gene which causes transcription of the producer gene to occur when a sequence-specific complex is formed between the receptor sequence and an RNA molecule called an activator RNA. We do not, in this model, wish to specify a mode of action for the receptor gene - that is, the nature of the molecular events occurring between the DNA, histones, polymerases, and so forth, present in the receptor complex. This model is concerned primarily with interrelations among the DNA sequences present in the genome.

Activator RNA: The RNA molecules which form a sequence-specific complex with receptor genes linked to producer genes. The complex suggested here is between native (double-stranded) DNA and a single-stranded RNA molecule (6). The role proposed for activator RNA could well be carried out by protein molecules coded by these RNAs without changing the formal structure of the model (7). Decisive evidence is lacking in higher cells, and we have chosen the simpler alternative (8). As the discussion of the evolutionary implications of this model will indicate, however, the probability of formation of new batteries of genes in evolution appears to differ greatly between these two alternatives.

Integrator gene: A gene whose function is the synthesis of an activator RNA. The term integrator is intended to emphasize the role of these genes in leading, by way of their activator RNAs, to the coordinated activity of a number of producer genes. A set of linked integrator genes is activated together in response to a specific initiating event, resulting in the concerted activity of a number of producer genes not sharing a given receptor gene sequence.

Sensor gene: A sequence serving as a binding site for agents which induce the occurrence of specific patterns of activity in the genome. Binding of these inducing agents is a sequence-specific phenomenon dependent on the sensor gene sequence, and it results in the activation of the integrator gene or genes linked to the sensor gene. Such agents include, for example, hormones and other molecules active in intercellular relations as well as in intracellular control. Most will not bind to sensor gene DNA, and an intermediary structure such as a specific protein molecule will be required. This structure must complex with the inducing agent and must bind to the sensor gene DNA in a sequence-specific way.

Battery of genes: The set of producer genes which is activated when a particular sensor gene activates its set of integrator genes. A particular cell state will usually require the operation of many batteries.

Integrative Function of the Model

The concerted activation of one or more batteries of producer genes is considered to underlie the existence of diverse states of differentiation. Examples of two basic aspects of the proposed integrative function appear in Fig. 1.

Fig 1. Types of integrative system within the model. (A) Integrative system depending upon redundancy among the regulator genes. (B) Integrative system depending upon redundancy among the integrator genes. These diagrams schematize the events that occur after the three sensor genes have initiated transcription of their integrator genes. Activator RNAs diffuse (symbolized by dotted line) from their sites of synthesis - the integrator genes - to receptor genes. The formation of a complex between them leads to active transcription of the product genes P_A, PB, and P_C.

In Fig. 1A, the control pattern depends on the existence of redundant receptor sequences in the receptor gene sets of the the three producer genes. Inclusion of a particular producer gene in each of the batteries calling on it depends on the presence of the appropriate receptor gene adjacent to the producer gene. Thus, in the case where there is only one integrator gene per sensor as in Fig. 1A, there will be as many copies of a given receptor gene sequence as there are producer genes in a battery.

In the case shown in Fig. 1B, however, redundancy is present between the integrator genes of different integrator sets. A particular producer gene, in this example, is included in each of several batteries calling on it by virtue of the inclusion of the same integrator gene adjacent to each of the appropriate sensor genes. Here there will be as many copies of a given integrator gene as there are batteries that call on its producer gene. For certain commonly required genes, for example those used in the fundamental biochemistry of each cell, this could be a very large number indeed.

Systems of the type portrayed in Fig. 1A might be most useful in the case where the producer genes to be integrated direct the synthesis of enzymes whose function is tightly coordinated physiologically, for example, the ten enzymes of the urea synthesis system. Where the system is needed, all genes would be needed. The system portrayed in Fig. 1B is a more powerful integrative system since it can govern a larger diversity of producer genes. The number of receptor sequences governing each producer sequence is probably small since transcription of a producer gene sequence is not likely to be activated from a great distance along the DNA strand. There is no reason a priori, on the other hand, to restrict the number of integrator genes per integrator set, except for the requirement that the integrator genes not be so distant that there is a high probability of their being separated by translocation.

In this model, regulation is accompanied by sequence-specific binding of an activator RNA and not be sequence recognition on the part of histones. The latter seem clearly to be the general inhibitors of transcription in the genome, but evidently these general inhibitors do not possess sufficient diversity to be considered as sequence-specific regulatory elements themselves (9, 10). We have assumed that, unless otherwise specified, the state of the higher cell genome is histone-mediated repression and that regulation is accomplished by specific activation of otherwise-repressed sites, rather than by repression of otherwise active sites.

Figure 2 combines the elements and systems we have thus far described.

Fig. 2. This diagram is intended to suggest the existence of overlapping batteries of genes and to show how, according to the model, control of their transcription might occur. The dotted lines symbolize the diffusion of activator RNA from its sites of synthesis, the integrator genes, to the receptor genes. The numbers in parentheses show which sensor genes control the transcription of the producer genes. At each sensor the battery of producer genes activated by that sensor is listed. In reality many batteries will be much larger than those shown and some genes will be part of hundreds of batteries.

Sequential patterns of gene activation, as in development, could result if certain sensors respond to the products of producer genes. In addition, the protein of a newly effective sensor assembly is, in the model, a product of a previously activated producer gene. Stabilization of a cell type in a given state of differentiation might also be explained in this way. Living systems continuously adjust their activities in accordance with their internal state, and it is evident that a requirement for sensors sensitive to feedback control by certain producer-gene products exists as well.

Fraction of the Genome Utilized for Regulation

Broadly speaking, genome size increases with the grade of organization of eukaryotes, as first pointed out by Mirsky and Ris in 1951 (11, 12). The wide range of genome sizes often observed among closely related creatures obscures the correlation. Organisms with large genomes presumably have a requirement for genomic information similar to that of their relatives with smaller genomes. This implies the evolutionary multiplication of the genome of ancestors possessing the minimum amount of DNA to effect each grade of organization. It is thus useful to consider the minimum amount of DNA observable at each grade of organization. Figure 3 shows the minimum genome size (13) for some major steps in evolution between viruses and the higher chordates.

Fig. 3. The minimum amount of DNA that has been observed for species (13) at various grades of organization. Each point represents the measured DNA content per cell for a haploid set of chromosomes. In the cases of mammals, amphibians, teleosts, bacteria, and viruses enough measurements exist to give the minimum value meaning. However for the intermediate grades few measurements are available, and the values shown may not be truly minimal. No measurements were unearthed for acoela, pseudocoela and mesozoa. The ordinate is not a numerical scale, and the exact shape of the curve has little significance. The figure shows that a great increase in DNA content is a necessary concomitant to increased complexity of organization. 

Quite possibly, the principal difference between a poriferan and a mammal could lie in the degree of integrated cellular activity, and thus in a vastly increased number of producer genes (16). Much of the DNA accumulating in the genomes toward the upper end of the curve in Fig. 3 might then have a regulative function. The model also suggests that a large amount of DNA could be devoted regulatory function: consider integrator and receptor sequences which are not redundant. In this case a battery of producer genes would require a distinct integrator gene for each producer gene. Producer genes occurring in several batteries would require receptor genes corresponding to each battery. The resulting multiplicity of integrator and receptor genes might result in a much larger quantity of DNA in regulatory sequences than in producer sequences. It is likely that an ever-growing library of different combinations of groups of producer genes is needed as more complex organisms evolve. An effective way of storing the information specifying these combinations in the genome is to make use of sensors responsive to the activator RNAs of other integrative sets. Thus we propose that a higher level of integrator gene sets is accumulated. each of these, when activated, could specify a very large program of producer gene activations by specifying the activity of a network of other sensor-integrator sets. Thereby many batteries of genes of the sort shown in Fig. 2 could be activated.

Experimental Justification of the Elements of the Model

There are five important classes of elements in this model: sensor genes, integrator genes, activator RNA, receptor genes, and producer genes. Is this degree of complexity really necessary ? The particular set of elements we have postulated may of course not be the required ones. Five, however, is the minimum number of classes of elements which can carry out the following formally described process: (i) response to an external signal; (ii) production of a second signal; (iii) transmission of the second signal to a number of receptors unresponsive to the original signal; (iv) reception of the second signal; and (v) response to this event by activation of a producer gene and its transcription to provide the cell with the producer gene product. In the following sections we examine evidence that such a description is applicable to gene regulation in higher organisms, and explore evidence that suggests the existence of the elements of the model.

Integration of Physically Unlinked Producer-Gene Activity

We have assumed that a given state of differentiation depends on the coordinated activity of a number of biochemical systems. Each of these systems will probably contain a number of components. As an example, Table 1 lists some of the enzyme systems operating in  one cell type, mammalian liver.

The data considered so far provide instances of the type of pattern which our model is designed to interpret, but they do not indicate the extensiveness of the producer-gene batteries called forth in given conditions of differentiation. Table 3 lists some of the effects of estrogen on the uterus, an estrogen target tissue.

Evidence for the Existence of Sensor Elements

There are many chemically defined agents that have the evident property of inducing large-scale changes in the producer-gene activity of specific target tissues. These agents now include steroid hormones, polypeptide hormones, several plant hormones, several vitamins, and several embryonic inductive agents (20). Frequently, the responsible agents also produce an alteration in the spectrum of RNAs being transcribed in the target tissues, as indicated by data obtained with RNA-DNA hybridization and studies in vitro of chromatin template activity (20); and these agents have been identified in the nuclear apparatus of the target cells (20). The most intensively studied system is perhaps estrogen response (Table 3). All of the above-mentioned forms of evidence exist for this hormone (20).

In addition, Maurer and Chalkey (21) have isolated from calf endometrial chromatin a protein that binds 17b-estradiol. The binding is stereospecific, noncovalent, and strong ( the Michaelis constant, K_m, for binding is 2 x 10 M ^-8 ); and the responsible protein appears not to be a histone. It does not bind sterols as closely related as 17a-estradiol or diethylstilbesterol. Such a protein, in combination with the specific external agent for which it is the receptor, must interact with the genome in a sequence-specific way, since this interaction results in the activation of only a certain group of genes. Consider a system in which the genomic binding sites are simply adjacent to all the producer genes activated by the external chemical agent. Such a system would appear to possess a limited integrative function which might be utilized for certain small gene batteries. However, the binding of an external agent to a sequence-specific site on the genome could lead to the activation of a large number of distant producer genes. This is exactly the role the sensor elements of this model carry out. Implicit in the available data on hormone action are genomic elements performing some of the functions of the producer and integrator genes of the model.

Evidence Suggesting the Existence of Activator RNAs

Many of the properties attributed to the RNA in our model are actually those of a certain class of RNA molecules already described extensively; yet no known function has so far been attributed to to this class of RNAs. The activator RNA molecules of the model have the following properties that can be tested. (i) They will, in the main, be confined to the nucleus, that is, they are not precursors of cytoplasmic polysomes. (ii) When observed in their functional role, they would be found in chromatin, bound to DNA, in a sequence-specific manner. (iii) They are often the product of the redundant fraction of the genome. (iv) They include sequences not present in the polysomes carrying producer-gene templates, that is, most or all cytoplasmic polysomes.

The hybridization experiments of McCarthy and Shearer (23) (Table 4) were performed at relatively low concentrations of nucleic acid and at short incubation times. Therefore, the RNAs they describe are the products of the redundant fraction of the genome. The presence of sequences specific to the nucleus and their absence from the cytoplasm is indicated by competition experiments. Furthermore, the nuclear RNAs contain sequences binding as much as five times more DNA than the cytoplasmic RNA at empirical saturation of the DNA with RNA. These and other data of Table 4 show that the RNAs are already known which might fulfill the functions we have assigned to activator RNAs.

Large Changes in Transcription of Redundant Sequences

It is a striking fact that very large changes in the spectrum of RNAs deriving from repetitive sequences are observed when the state of differentiation alters. This knowledge is derived from RNA-DNA hybridization experiments carried out at relatively low concentrations of nucleic acid and short annealing times, so that the reaction of RNA with any but the repetitive sequences in the genome is precluded. The spectrum of RNAs present or or in the process of being synthesized in different tissues (4, 24), both in hormone response (25) and in embryonic development and differentiation (26, 27), has been investigated with competition procedures. In these experiments RNA from a cell type in one state of differentiation is used to compete with RNA from a cell type in another state state of differentiation for binding sites in the repetitive fraction of the DNA. This type of analysis has shown that different families of repetitive genomic sequences are represented in the RNA of cells in diverse states of differentiation. Changes as large as 100 percent (apparent complete lack of homology) in the measured RNAs have been observed - for example, in successive stages of the the embryogenesis of Xenopus (27). It is not particularly obvious why such changes should be detected, since the populations of producer genes active in each state of differentiation might be expected in general to be strongly overlapping. One possible explanation would be that much of the pulse-labeled RNA monitored in these studies is the rapidly-turning-over product of different regulatory genes such as the integrator genes of this model.

Regulatory Genes Known in Higher Organisms

The model suggests that a sizable portion of the functional genes in differentiated cell types may be regulatory genes (integrator and receptor genes). If this is so, it might be expected that, despite the difficulty of detecting such genes with classical genetic procedures, a certain number of apparent regulatory mutations would be known in higher organisms. The distinguishing characteristic of such regulatory loci would be pleiotropic effects on the activity of a number of producer genes, particularly with reference to a pattern of integration on the part of the latter. A number of good cases of this genre actually exist, particularly for drosophila and maize. A notable example is the Notch series of x-chromosome deficiencies (28), some of which are sharply localized. Notch mutants display a very large variety of developmental abnormalities - all affecting early embryonic organization - for example, failure to form a complete gut, failure of mesodermal differentiation to occur, overly large neural structure, and subnormal ectodermal skin production. Their effects are clearly pleiotropic. The multiplicity of the actual primary failures of these mutants is unknown. That is, no comparison can be made of the number of diverse producer genes affected simultaneously, as opposed to the array of sequential effects that follow the initial primary effects. Nonetheless, the effect of the Notch genes on the organization of the embryo is consistent with what would be expected of mutations in integrator gene sets. Many similar cases are known in which specific organizational lesions result from simple mutations affecting a small region of the genome (29). Studies with drosophila imaginal disk cell determinations and transdetermination carried out by Hadorn and his associates (30) also demonstrate the existence of an apparatus in the genome for specifying integrated patterns of activity in the various cell types deriving from the disk cells. In experimental imaginal disk systems, highly exact specification of the patterns of producer gene activity is heritable through many cell divisions and is separated in time from producer gene function per se (that is, manifest differentiation).

Genes are known to act in maize which display control over producer genes and are located in the genome at sites distant from the producer genes that they control (31). In addition, McClintock and others have demonstrated the presence of other control sites adjacent in the genome to the same producer genes as those controlled by the distal regulatory elements (31). Control of the expression of the producer gene is accomplished through the interaction of the distant regulatory gene with the contiguous regulatory gene. This point has been demonstrated by insertion of the contiguous regulatory genes at different sites in the genome, near known genes, which then respond to the same control system governed by the distant regulatory unit. An example is the system termed Ac-Ds. Here the distant regulatory element Ac (which behaves as an integrator gene of this model) can be made to govern producer genes in other chromosomes such as the gene series for synthesis of anthocyanin pigment. Establishment of Ac control over the pigment synthesis system is accomplished by transposing the contiguous regulatory element responsive to Ac (Ds) to the loci of the anthocyanin producer genes (Ds thus behaves like a receptor gene of this model). In several ways, these and other data presented by McClintock (32) would seem to fit a model such as that presented here.

DNA Sequence Repetition

The existence of repeated sequences in higher organisms led us independently to consider models of gene regulation of the type we describe here. This model depends in part on the general presence of repeated DNA sequences. This model suggests a present-day function for these repeated DNA sequences in addition to their possible evolutionary role as the raw material for creation of novel producer gene sequences. The apparently universal occurrence of large quantities of sequence repetition in the genomes of higher organisms (14) suggests strongly that they have an important current function.

The quantity of DNA in repeated sequences, the frequency of repetition (that is, number of times a given sequence is present per genome), and the precision of the repetition show great variation among species. Frequencies from 100 to 1,000,000 have been observed, and the quantities of DNA involved range from 15 to 80 percent of the total DNA. The usual relation between repeated sequences is not that of a perfect copy (33), but the sharing of most of the nucleotides in a sequence extending for at least a few hundred nucleotides. Repeated sequence families in the DNA are observed with degrees of similarity varying from perfect matching to matching of perhaps only two-thirds of the nucleotides. Expression of families of repeated sequences by transcription into RNA shows tissue specificity (as mentioned above) in spite of the fact that the individual families contain these widely divergent sequences.

In the cases studied there is good evidence that the repeated sequences are scattered throughout the DNA. For example, in bovine DNA, 75 percent of all fragments about 5000 nucleotides long contain a segment of repeated DNA (34). When the fragment size is reduced to about 500 nucleotides, only 45 percent contain repeated sequences. Therefore, the typical bovine DNA fragment of 5000 nucleotides is a composite of lengths of repeated sequences and nonrepeated sequences. For longer fragments (20,000 or so nucleotides), there is suggestive evidence (14) that more than 95 percent contain repetitive sequences. Therefore, for bovine DNA ( and probably that of other organisms) repeated sequences are intimately interspersed with nonrepeated sequences, throughout the length of the genome. This is precisely the pattern required in our model if repeated sequences are usually or often regulatory in function.

Evolutionary Implications of the Model

Any evolutionary changes in the phenotype of an organism require, in addition to changes in the producer genes, consistent changes in the regulatory system. Not only must the changes be compatible with the interplay of regulatory processes in the adult, but also during the events of development and differentiation. At higher grades of organization, evolution might indeed be considered principally in terms of changes in the regulatory systems. It is therefore a requirement of a theory of genetic regulation that it supply a means of visualizing the process of evolution.

Inactivity of New Genetic Material

A characteristic of this model is that DNA sequences are inactive in transcription, unless specifically activated. Thus the genome of an organism can accommodate new and even useless or dangerous segments of DNA sequence such as might result from a saltatory replication (35). Initially these sequences would not be transcribed, and thus would not be subject to adverse selection. Only by inclusion in integrated producer gene batteries (through translocation of receptor genes) would their usefulness as producer genes be tested.

Formation of New Integrative Relations

A peculiar combination of conservatism and flexibility is supplied by the model system. Preexisting useful batteries of genes will tend to remain integrated in function. At the same time, there is the potentiality of formation of new integrative combinations of preexisting producer genes. These combinations would be the result of translocations, principally among the integrator gene sets. Less often, new producer gene batteries would result from events in which receptor genes are translocated into positions contiguous to other producer genes.

We visualize many of the integrator genes and receptor genes as being members of families of repeated DNA sequences. It is known that new repeated sequence families have originated periodically in the course of evolution (35). The new families of repeated sequences might well be utilized to form integrator and receptor gene sets specifying novel batteries of producer genes. Thus saltatory replications can be considered the source of new regulatory DNA. All that is required for regulatory function in this model is sequence complementarity (translocation of members of the same repetitive sequence family to integrator and receptor positions). Almost any set of nucleotide sequences would suffice. The likelihood of utilization of new DNA for regulation is thus far greater than the likelihood of invention of a new and useful amino acid sequence, since for the latter case great restrictions on the nucleotide sequence exist.

Changes in the integrator systems make possible the origin of new functions and possibly even of new tissues and organs. In other words, the model supplies an avenue for the appearance of novelty in evolution by combining into new systems the already functioning parts of preexisting systems.

Divergence within Repeated Sequence Families

Individual sequences may differ from others in a family as a result of many base changes. We presume that binding of activator RNA to the receptor genes will occur for a degree of sequence homology far short of perfect complementarity. However, at some degree of divergence, binding would be lost, and a producer gene would fail to be activated as a part of its previous battery. Eventually, the process of divergence might yield regulatory DNA in which the original patterns of repetition are no longer observable. In this way, non-repeated (unique) regulatory DNA could arise, leading to the situation discussed earlier with respect to the fraction of the genome utilized for regulation.

The possibility of increasing sequence divergence among integrator and receptor genes suggests a novel evolutionary mechanism. The divergence of regulatory sequences can be expected to be reversible. If the degree of complementarity required for binding between activator RNA and receptor sequence is fairly low then a reasonably good probability would exist for a subsequent base change to restore the complementarity lost by an earlier change. Intermediate degrees of transcription of certain producer genes will probably result since sequences with a degree of complementarity near some critical value will bind only part of the time. Natural selection could then reversibly affect the integration of individual producer genes into batteries. The possibility for smoothly changing patterns of integration among many sets of producer genes supplies a mechanism for direct adjustment by natural selection of the organization of systems of cellular activity. In other words, the model implies that selective factors can influence the integrative configurations in which an organism uses its genes.

The families of repeated sequences that appear and remain in the genome of a species affect the rate at which newly integrated systems of producer genes will arise. Thereby, the rate of evolution is affected. It follows that the rate of evolution will be acted on by natural selection.

The issues raised in considering the evolution of the regulatory systems themselves are of a magnitude which is really out of reach in this brief discussion. However, the model offers interesting and surprising predictions. The properties of the model regulatory system suggest that both the rate and the direction of evolution (for example, toward greater or lesser complexity) may be subject to control by natural selection.

A theory for the genomic regulation systems of higher organisms is described. Batteries of producer genes are regulated by activator RNA molecules synthesized on integrator genes. The effect of the integrator genes is to induce transcription of many producer genes in response to a single molecular event. Current evidence suggesting the existence of elements of this model is summarized. Some evolutionary implications are indicated.

Reference no. 8:
J. H. Frenster, "A Model of Specific De-repression within Interphase Chromatin", Nature vol. 206, no. 4990, pp. 1269-1270 (June 19, 1965).

0. Blumenthal T, Evans D, Link CD, Guffanti A, Lawson D, Thierry-Mieg J, Thierry-Mieg D, Chiu WL,
Duke K, Kiraly M, and Kim SK, "A Global Analysis of Caenorhabditis elegans Operons". Nature vol. 417, no. 6891, pp. 851-854 (June 20, 2002).

1. Frenster JH, Allfrey VG, and Mirsky AE, "Metabolism and Morphology of Ribonucleoprotein Particles from the Cell Nucleus of Lymphocytes", Proc. Nat. Acad. Sci. U.S.A., vol. 46: pp. 432-444 (April, 1960).

2. Frenster JH, Allfrey VG, and Mirsky AE, "In Vitro Incorporation of Amino Acids into the Proteins of Isolated Nuclear Ribosomes", Biochimica et Biophysica Acta vol. 47: pp. 130-137 (1961).

3. Frenster JH, "Constraints on Isolation of Mammalian Chromosomes", Exp. Cell Res. vol. 9, 235-238 (1963).

4. Frenster JH, Allfrey VG, and Mirsky AE, "Repressed and Active Chromatin Isolated from Interphase Lymphocytes", Proc. Nat. Acad. Sci. U.S.A., vol. 50, no. 6, pp. 1026-1032 (Dec. 1963).

5. Littau VC, Allfrey VG, Frenster JH, and Mirsky AE, "Active and Inactive Regions of Nuclear Chromatin as Revealed by Electron Microscope Autoradiography", Proc. Natl. Acad. Sci. U.S.A., vol. 52, no. 1, pp. 93-100 (July, 1964).

6. Frenster JH, "Ultrastructural Continuity Between Active and Repressed Chromatin", Nature vol. 205, no. 4978, pp. 1341-1342 (March 27, 1965).

7. Frenster JH, "Nuclear Polyanions as De-Repressors of Synthesis of Ribonucleic Acid", Nature, vol. 206, no. 4985, pp. 680-683 (May 15, 1965).

8. Frenster JH, "A Model of Specific De-repression within Interphase Chromatin", Nature vol. 206, no. 4990, pp. 1269-1270 (June 19, 1965).

9. Frenster JH, "Localized Strand Separations within Deoxyribonucleic Acid during Selective Transcription", Nature, vol. 208: no. 5013, pp. 894-896 (November 27, 1965).

10. Rose HG, and Frenster JH, "Composition and Metabolism of Lipids within Repressed and Active Chromatin of Interphase Lymphocytes", Biochimica et Biophysica Acta, vol. 106, no. 3: pp. 577-591 (December 2, 1965).

11. Frenster JH, "Correlation of the Binding to DNA Loops or to DNA Helices with the Effect on RNA Synthesis", Nature vol. 208, no. 5015, p. 1093 (December 11, 1965).

12. Frenster JH, "Mechanisms of Repression and De-Repression within Interphase Chromatin", Presented at the Annual Meeting of the Tissue Culture Association as part of the Symposium: "The Chromosome: Structural and Functional Aspects", (Dawe CJ and Yerganian G, eds.), Williams and Wilkens, Baltimore: 1965, and also published in: In-Vitro, vol. 1, pp. 78-101, (1965).

13. Herstein PR, and Frenster JH, "Mated Models of Gene Regulation in Eukaryotes", "Embryonic and Fetal Antigens in Cancer", vol. 2, pp. 5-7, (Anderson NG, Coggin JH, eds.), National Technical Information Service, U.S. Dept. Commerce, Springfield, VA., 1972.

14. Frenster JH, and Herstein PR, "Gene De-Repression", New Eng. J. Med. 288: 1224-1229 (June 7, 1973).

15. Frenster JH, "Selective Control of RNA Helix Openings during Gene Regulation", Cancer Res., vol. 36, pp. 3394-3398 (September, 1976).

16. Frenster JH, "Selective Gene De-Repression by De-Repressor RNA", "Eukaryotic Gene Regulation", Volume 1, pp. 131-143, 1980, (Kolodny GM, ed.), CRC Press, Boca Raton, FL, USA.

17. Frenster JH, "Single-Cell Analysis of DNase I-Sensitive Sites during Neoplastic and Normal Cell Differentiation within Human Bone Marrow", Ann. N.Y. Acad. Sci. 567: 334-336 (Aug. 4, 1989).

18. Frenster JH, "Nuclear RNA Species Activate DNA Transcription Within Chromatin", FASEB Journal, Vol. 13, No. 7, A1506 (April 23, 1999).

19. Frenster JH, "Uni-Polar Clustering of Lymphocyte DNA Templates Toward Neoplastic Target Cells Within Hodgkin’s Disease Lymph Nodes", 2002.

20. Frenster JH, "Yeast  RNA  Re-Programming  of  Already-Active  Mammalian Chromatin", 2002.

21. Frenster JH, and Hovsepian JA, "RNA Feedback Mechanisms during Eukaryotic Gene Regulation", 2002.

22. Frenster JH, "Ultrastructural Probes of Active DNA Sites, and the RNA Activators of DNA", 2002.