[Back to Number 11 ToC] [Back to Journal Contents] [Back to Biokhimiya Home page]

The Molecular Mechanism of Morphogenesis: a Theory of Location DNA

A. M. Olovnikov1

1Institute of Biochemical Physics, Russian Academy of Sciences, ul. Chernahovskogo 5-94, Moscow, 125319 Russia; E-mail: am@olovnikov.msk.ru

Submitted September 17, 1996.
A molecular mechanism whereby cells read positional information during morphogenesis and regeneration is proposed. This mechanism enables the cells to translate the linear genomic information into three-dimensional shapes of the body. The mechanism uses a new DNA fraction called location DNA; this term replaces the former "egoistic" DNA. Domains formed by this DNA and packaged by means of lipid-containing bridges are selectively unpackaged in the gradient of an inductor whose concentration is positively correlated with generation of free radicals in the cell. Free radicals selectively destroy the lipid bridges; individual bridges display various degrees of resistance to oxidative destruction. Therefore, domains of location DNA are selectively decompacted and activated so that the position information can be read out. The epigenetic memory on the achieved state of cellular determination is based on triplexation, that is, the formation of a triplex consisting of a signal RNA molecule and a nascent double-stranded DNA molecule during lagging DNA strand synthesis or DNA repair synthesis. A telomeric element (the chronomere) of postmitotic neurons, due to DNA end underrepair synthesis, allows the organisms to measure the flow of biological time. The length of the chronomere indicates the biological age of the body.
KEY WORDS: positional information, epigenetic memory, determination, biological time, free radicals, junk DNA, lipids, telomere, triplex.

The entire Ocean can only be viewed from a satellite: individual waves are not discernible but underwater streams and other deep secrets are clearly visible. In this work, I attempted to perform such a satellite flight. The whole body of literature on biological morphogenesis is impossible to review, but a new glance into still unresolved enigmas of this exciting process in an attempt to provide a new interpretation for them may be interesting.

Morphogenesis is the central event in the individual development of every living creature. In multicellular organisms, the essence of morphogenesis is to build spatially ordered structures (tissues and organs) whose coordinated functioning would be based on the activity of optimally determined and differentiated cells.

Of all possible facets of the morphogenetic saga, I have selected here a single but the key story: how does the body manage to translate the linear records of one-dimensional information contained in DNA into a complex three-dimensional structure.

On Terminology Used in This Paper

For convenience, I shall first consider old and new notions and terms suitable for discussing the mechanism proposed here. Major, generally accepted notions of biological morphogenesis will be used: morphogens or inductors, determination, and competence.

The terms inductor and morphogen are used as synonyms. The primary inductor (primary morphogen) is the substance capable of eliciting the determination of competent cells and the formation of the anlage; the axis of the anlage is oriented along the gradient of the inductor distributed over the area where morphogenesis proceeds. The secondary inductor (secondary morphogen) is a substance produced by a group of cells undergoing the same determination and released toward adjacent cells in order to induce their determination (provided that both cell groups are within the reach of the same primary inductor).

Competence is the ability of cells to undergo determination when affected by an inductor. Competence is the preparedness of cells to accomplish a series of specific determinations when the tissue is affected by a morphogenic gradient. The cells are considered to be competent if they can undergo determination and morphogenetic events governed by the gradient of an inductor. The mechanism of competence is still unknown.

Determination is the transition of a competent cell to a state capable of performing a series of differentiations. Determination, as a rule, proceeds from the whole to its parts: first the whole anlage (or rudiment) of the future embryo undergoes determination, and then its parts. The difference between determination and differentiation is most clearly demonstrated by imaginal disks of Drosophila. These are small islands of cells in the larval body that give rise to organs of the imago, and each disk is the source of a strictly determined organ. Determination of imaginal disks occurs during embryogenesis, whereas differentiation only occurs in the pupa. Determination of imaginal disks is generally irreversible and is maintained throughout the entire period of larval development. Upon artificial cloning (from fly to fly), these cellular islands can be cultivated for unlimited time periods. Exogenous administration of ecdysone, a natural hormone that causes the pupa formation and metamorphosis, immediately elicits differentiation processes in determined cells of imaginal disks. Each ecdysone-stimulated disk proceeds according to its own scenario of development through of a series of diverse (but typical of the eukaryotic world) events, including the formation of competent cells, induction-related interactions, and execution of appropriate organogeneses. In humans, a similar mechanism regulates morphogenetic events related to the formation of secondary sexual traits: a hormone permits the continuation of morphogenetic processes that have been started but not finished during embryogenesis.

This work will deal with the mechanism of determination rather than differentiation of cells; the latter will be discussed only in passing. The significance of competence for interpretation of positional information by cells forming complex morphological structures (patterns) will be considered.

I propose that several novel structures exist in the eukaryotic genome. The first is the fatomere or determomere. This structure controls the determination of cells and is therefore responsible for the fate of the cell group that shapes a certain morphological structure. The word fatomere originates from fatum, the fate. A particular type of telomeric fatomere carries the so-called chronomere on its end. The chronomere serves for estimating the course of biological time during ontogeny.

To solve the problem of reading the positional information, eukaryotic genomes use the morphons, the specific structures, which are domains of the large fraction of DNA which, according to the hypothesis discussed here, performs an entirely novel function. This type of DNA, the so-called location DNA is considered in the second part of this communication.


Examples of Different Morphogens

The progress in studies of various morphogens revealed both low-molecular-weight (e.g., retinoic acid) and polypeptide inductors in animals (reviewed in [1, 2]). In plants, the majority of morphogeneses can be induced by hormones; 6-benzylaminopurine is the substance that most often determines the plant morphogenetic pathways and therefore can be considered as an analog of the primary morphogen that directs the axis of anlage in animals. Auxins, such as derivatives of phenoxyacetic and naphthylacetic acids, facilitate dedifferentiation, thereby creating conditions for switching between various morphogeneses in plants.

The term morphogenetic field denotes cooperative activity of cells in a group (often less than one hundred) that display similar responses to developmental signals; each cell responds to these signals in accordance with its own topographical position in a cell population sharing the same competence.

The concept ascending back to Driesch and suggesting that the fate of a part (e.g., a group of cells) is the function of its position in the whole (e.g., in the anlage) is of special importance for studies of biological morphogeneses. This theory stresses the crucial role of the position of cells as a factor determining their developmental fate and an important role of morphogenic gradients in development. These ideas [3, 4] formulated by Driesch, Morgan (who was initially an embryologist and then became a geneticist; he proposed the term telomere often used in this paper), and Child were further developed by Wolpert who suggested that a certain mechanism should help individual cells to obtain, by means of morphogenic gradients, information on their position in the cell group [5]. Each cell interprets this kind of information according to its own genetic constitution and developmental history [6]. Morphogenic gradients are believed to carry this positional information. How cells receive, read, and interpret this information is unknown. To answer this question means to elucidate the mechanism of translation of one-dimensional (linear) records of genetic information into three-dimensional structures. This is exactly what I intend to do in this paper.

The abundance of facts gained in studies of biochemical bases of development is depreciated by this unresolved question of prime importance: how exactly the cell can perceive its spatial position and how linear genomic information is then translated into the three-dimensional shape of the embryo.

The state of competence is known to be canceled if it is not realized in due time; in this case, it is replaced by another competence, that is, the potential capability of cells to undergo certain other determinations (which are allowed only when the cells are placed in a new gradient of an appropriate primary inductor).

I suppose that the physical cause of this loss of competence is the loss of the ability of chromosomes to amplify specific sequences (fatomeres) due to transition of chromosomal templates to a highly compacted state (e.g., upon heterochromatization of a telomere containing a series of fatomeres).

The fatomeric (or determomeric) model proposed here and described in much detail below postulates that the positional information (the memory of the cell about its position in the coordinates of the anlage) is stored as epigenetic memory on particular gradients that have sequentially (over the preceding development) arranged the axes along which the initially homogeneous cell mass was heterogenized. The mechanisms of this epigenetic memory and the role of fatomeres in morphogenesis constitute the first part of the proposed model of morphogenesis. The second part deals with location DNA whose postulated existence makes unnecessary the concept of egoistic DNA.

In the overwhelming majority of embryonic regulation processes that build up normal embryonic structures even after microsurgical ablation, translocation, and addition of cells, the whole organism forces each cell to follow a particular pattern of the developmental fate [7, 8]. The hypotheses described here may replace the currently available models designed to explain morphogenesis (see [8] for a critical analysis of various models of biological self-organization). The well-known Turing's model suggests that stable structures can emerge upon interaction of two substances displaying different diffusion rates. Variants of this idea often emerge in attempts to explain the creation of morphological patterns [9], although the supporting evidence is scarce. This situation can be simply due to the absence of an acceptable alternative hypothesis. The list of biological examples where this mechanism actually operates is short [10, 11], despite many attempts to discover such a process.

Role of Mechanical Tensions in Morphogeneses

Complexes of embryonic axial organs are extended along the lines of strongest mechanical tensions in tissues because of active cell migration in these directions [12]. Repeated cycles of these migrations and new determinations in sequentially arising morphogenic gradients, cell differentiation limited by a given determination, and new kinds of migration provide the basis for any morphogenesis. The permanent (while morphogenesis is underway and the organism develops) generation of new positional information and its subsequent interpretation gradually increase the degree of specialization of tissues and organs. For this reason, the mechanism responsible for interpretation of positional information is the key mechanism of development.

The primary inductor is delivered to competent, cytologically polarized cells having nonrandom spatial orientation and induces the expression of fatomeres. Mechanical tension is of great importance for the correct functioning of fatomeric mechanisms (see below). I assert here that tensions result in correct orientation of the morphogenic gradient in relation to competent, polarized cells; in addition, tensions actually synchronize cells in a transitory nonmitotic state, which is absolutely necessary for the proper functioning of the mechanism proposed here.

Polarized cells (those displaying a nonrandom orientation of intracellular organelles shared by all cells in the group), when placed in a morphogenetic field, become polarized because of mechanical torsion, compression, stretch, and other forces generated by molecules of the extracellular matrix, cytoskeleton, and other structures. A morphogenetic field usually acts on up to hundreds of cells [13]. Most secondary morphogens are probably proteins, in contrast to low-molecular-weight primary inductors such as retinoic acid. Trans-retinoic acid is the most extensively studied inductor [1]. Secondary morphogens can be produced by a group of cells (the starting group) that has already been determined by the primary inductor. The secondary inductor diffuses into adjacent cells of the same competence that occur in a common morphogenetic field with all these cells (that is, within the range of activity of the primary inductor gradient). The secondary morphogen stimulates its recipient cells to express another fatomere controlling the establishment of a state of new determination. Both types of morphogens spread along lines of the strongest mechanical tension in the embryonic tissue.

An alternative mechanism whereby cells receive positional information without the involvement of secondary morphogens will be discussed together with the role of location DNA in the second part of this paper. Generally, secondary morphogens are required for informing cells about the presence of their neighbors.

The number of different primary inductors may be small because tissues of different competence can have fatomeres which are expressed when affected by the same inductor. For example, the same retinoic acid induces different processes of morphogenesis in different competent tissues of various animal species [1]; certain tissues are stimulated to form a bud of the wing, whereas other form central nervous structures.

Fatomeres. The Problem of Genome Instability and Fatomeres

In most eukaryotes, the genome remains nearly the same over the whole period of individual development. However, there are important exceptions to this common rule.

In particular, all normal dividing somatic cells of multicellular eukaryotes lose a certain number of nucleotides from absolute ends of telomeric DNA in each mitosis. This is due to the so-called effect of end underreplication or marginotomy. In addition to end underreplication, DNA end underrepair synthesis may occur (the essence of these processes and the history of this issue are discussed in [14-22]).

Another example of somatic changes in genome is the diminution of chromatin, that is, the developmental loss of part of the genome and its rearrangement found in ascarides, cyclopes, and several other species.

Trypanosomes and certain other protozoa can rearrange their genomes in order to avoid the host's immune attack; in trypanosomes, this rearrangement involves the telomeric region, whereas amplification of repeated sequences and genes of surface antigens is so strong that it makes a major contribution to changes in the length of polymorphic chromosomes of the parasite [23].

Differential amplification of certain nucleotide sequences may occur in a cell affected by some pathological processes, e.g., cancer [24].

The somatic rearrangement of immunoglobulins, somatic amplification of ribosomal and some other genes, the capacity of polyploidization, and selective polytenization of individual fragments of chromosomes in certain animals, as well as transpositions and other phenomena have long indicated that certain elements of the eukaryotic genome could be labile [25, 26].

In light of this, it seems less unlikely that biological morphogenesis, at a certain stage, uses amplified DNA in addition to the main genome. The possibility of these somatic changes is fully consistent with results of experiments where adult animals were grown from nuclei isolated from somatic cells (e.g., Gurdon's experiments) or plants were grown from somatic cells; indeed, developmental changes in DNA postulated here actually involve small extrachromosomal copies rather than their genomic templates.

Heterochromatization of Telomeres and Fatomeres

An increase in copy number of ribosomal genes occurs simultaneously with heterochromatization of nucleolar genes, which probably indicates a regulatory coupling of these processes [27]. The telomeric region in many objects consists of heterochromatin, and individual segments of the subtelomeric region may be compacted to different extents. It seems probable that certain fragments of telomeres are decompacted during expression of fatomeres, whereas adjacent sequences are maximally compacted to prevent unscheduled expression of fatomeres whose expression has not been started during the given stage of development (the relationship between the work of fatomeres and location DNA will be described below in sections devoted to location DNA). Many studies have found ori in subtelomeric regions. This phenomenon finds a natural explanation in this context: these structures may be used for local expression of fatomeres whose amplified copies may exist as free entities beyond the continuous chromosomal DNA. If all fatomeres were linear molecules (like the chronomeric fatomere, a variant described below), they should be expected to shorten from the free end of DNA due to DNA end underrepair synthesis. Unrepaired breaches inevitably appear near the end of the linear DNA molecule [17-21] and cause linear DNA shortening.

I postulate that the primary inductor not only determines, by its gradient, the direction of the axis along which corresponding morphological structures will be located, but also initiates an increase in the level of free radicals in cells. This process can be mediated by not only intranuclear events. It probably involves specific functions of mitochondria. At this stage, a mitochondrion may approach the nucleus and export in it the free radicals produced by mitochondrial mechanisms. A short but intense burst of generation of free radicals can cause many nicks that are repaired everywhere in the genome, except for the most distant nucleotides of the telomeric DNA. These fragments cannot be repaired because of steric hindrance, inevitably preventing repair at the very end of the template; this has been postulated and discussed in detail in earlier publications [17-21]. Recent literature contains data on intense nicking of DNA during embryogenesis; such phenomena are documented for cell nuclei sampled at early stages of development [28]. It is conceivable that nicking is a direct consequence of a burst of free radical generation postulated in this paper. The mechanism of DNA end underrepair synthesis should shorten the termini of all linear DNA molecules, even those found in nondividing cells. The formation of circular fatomeres is the best defense against underrepair (however, this does not apply to chronomeres for reasons described below).

Mechanisms of Amplification of Fatomeres

Prokof'eva-Bel'govskaya has summarized all data on heterochromatin available by that time and concluded [26] that amplification of heterochromatin can occur in embryogenesis and play a major regulatory role. The model proposed here is consistent with this hypothesis and suggests that fatomeres perform the role of developmentally amplified heterochromatin.

The first variant of amplification can be considered as an analog of chromosomal polytenization. A copy of the fatomere is synthesized by DNA-polymerase and persists, for example, in the form of a copy positioned side-to-side near its template. This is what can be called "lateral" or "amplitenic" fatomere (provided that little attention is paid to the copy number, which can be small; even a single copy may exist). Such a lateral fatomere, when carried beyond the chromosomal DNA, has no centromere and cannot be reliably inherited by daughter cells. However, it does not need to be inherited if the competent cell undergoes the whole process of determination in one cell cycle.

The mechanism of cell differentiation, which works after amplification of fatomeres and is based on DNA triplexation, will not be described here in much detail (however, the significance of triplexation will be discussed below in relationship with epigenetic memory in determination).

The second variant of amplification differs from the first variant only in that the nascent copy of a fatomere leaves the chromosome in the form of circular DNA, which can be stored for a long time in nondividing cells until a primary inductor affects the cell. By this mechanism, the cell can remain competent in relation to the effect of the particular morphogen for a long time.

The third variant is similar to the first two, but the fatomere is copied through the stage of reverse transcription.

Amplification of Segments of Genome and the Fatomere

The telomeric sequence (TTAGGG)n, which marks the end of the telomere, is spread (probably because of ancestral chromosomal fusion) over not only along the whole telomere length, but also in certain parts of chromosomal arms. Moreover, these interstitial repeats were shown to be involved in genomic amplification [25], probably involving fatomeres.

Telomerase and Chronomere

The absence of telomerase-binding sites from the chronomere, which measures the biological time by the extent of end underrepair synthesis, could probably allow it to work even in telomerase-containing embryonic cells. Nevertheless, the absence of telomerase from adult somatic tissues, e.g., neurons, increases the reliability of measuring the biological time by the chronomere because telomerase could otherwise erroneously elongate the chronomere. In addition, this allows the chronomere to use certain combinations of nucleotides that would be prohibited in the presence of telomerase.

Fatomeres and the Problem of Equivalence of Genomes in Somatic Cells

The central hypothesis of developmental genetics postulates that cell differentiation does not involve genetic changes. However, the expression of a fatomere is a change.

It should be noted that different species use fatomeres for their morphogenesis in different ways. Certain species use directly chromosomal templates. This may be especially characteristic of species that use diminution of chromatin in order to expose an appropriate chronomere on the end of a chromosome. Other species (the majority of eukaryotes) use amplified fatomere products.

Normally, the fatomere template should be safely compacted in heterochromatin and never transcribed over almost the entire ontogeny; transcription only occurs at its most important moment (the cell determination) in embryonic morphogenesis.

While discussing the results of experiments with transplantation of nuclei from animal somatic cells to enucleated oocytes, DiBerardino [29] stressed that oligopotency of nuclei (e.g., in amphibians), that is, their ability to undergo a few determinations, is characteristic even of amphibian red blood cells. These findings indirectly suggest that fatomere copies are tools of morphogenesis, and fatomeres do not work as true chromosomal linear terminal templates because they would be irreversibly shortened after many divisions of animal cells.

Evolution and Fatomeres

The entire course of evolution in eukaryotes (its part that concerns telomeres) should specifically involve the addition, removal, and transposition of fatomeres if they are regulators belonging to a high level of hierarchy that monitor determinations and morphogenetic transformations.

Chromosomes fuse and undergo fragmentation in different ways during chromosomal evolution [30]. Certain telomeres are removed from the end of the chromosome, whereas other telomeres that occupy intrachromosomal positions are exposed. A chronomere can play its role in morphogenesis only after it has been moved to the absolute end of chromosomal DNA.

Polymorphism of Telomeres and Fatomeres

The termini of telomeres in humans and many other species contain repeated sequences TTAGGG (reviewed in [31]). This repeat is located distally from the complex of subtelomeric nucleotide sequences found in many telomeres. The organization of sequences adjacent to the telomeric TTAGGG repeats in primates displays a strict species-specificity. This was demonstrated in experiments with repeats that displayed different hybridization patterns with subtelomeric DNA in a range of primate species and humans [32-35]. Different subtelomeric sequences found, e.g., in primates, contain: 1) sequences uniquely found in the telomere; 2) intratelomeric moderate repeats; 3) very frequent repeats composed of Alu and LI-like elements. Elements that contain ori can be used for producing extra copies of fatomeres.

The Growth of Skeleton in Animals

Why are elephants bigger than mice and why are their skeletal proportions different? Modern molecular developmental biology provides no answer to these childish questions. The same level of ignorance is characteristic of our understanding of the nature of differences in the rates and end results of growth in different species. Indeed, high doses of growth factors have never produced a mouse that would be similar in size to a rat, to say nothing of an elephant. Nevertheless, the problem of scaling, that is, the control over the length ratios between different parts of the skeleton and other parts of the body, is important both in theoretical and practical aspects [8]. The developmental and evolutionary factors that affect the size ratios between different parts of the skeleton in animals are still unknown [36]. However, additive effects of a complex of factors may be important for differential growth of individual bones: 1) mechanical tension that stretches the bone and modulates its accelerated growth [37], including the species-specific rate of growth of the most rapidly elongated bones that cause mechanical tension of the skeleton and muscles; 2) size ratios between cell populations belonging to a certain anlage during embryogenesis (embryologists even described the competition for undetermined cell mass between anlages); 3) differences in local innervation of bones can account for considerable differences in the supply of neuropeptides and other neurotropic factors to these bones (the qualitative spectra of these factors may also be different); 4) different sets of genes activated in different rudiments of what will be the skeleton strongly determines the ability of these cells and their offspring to respond to various stimuli, such as mechanical tension, that play important roles in the morphogenesis and allometric growth [8].

Mitoses and the Function of Fatomeres

The regulation modes described by Driesch are only characteristic of these areas of tissues that already contain the fields of cellular tensions as distinct topological structures: the regularly positioned strands and tapes of elongated, polarized cells [38]. Therefore, inductors act on stretched and therefore temporarily nondividing cells. For this reason, fatomeres cannot be occasionally lost during DNA replication.

Cell division is an obligatory condition of embryogenesis; however, mitoses may never be involved in shaping the anlage. Dividing cells can occupy random positions relative to each other. For example, the bud of the brain has a zone of dividing cells, but this is not the site of the main morphological transformations. Dividing cells gradually leave this zone and begin their collective migrations, during which they ordinarily do not divide. The cell migration is usually directed toward the maximum mechanical tension of the substrate. In addition, migration can be directed by local compression by adjacent cell groups, etc. [8, 12, 38]. What is behind all these events and what is the goal of these intricate cell migrations?

The suggested answer is as follows. Morphogenetic migrations create the following necessary conditions: 1) migration brings a group of competent cells closer to the source of the next primary morphogen (this is a commonly shared concept); 2) active migrations of cells are, as a rule, incompatible with mitosis; therefore, cells arranged in the gradient of a primary inductor are temporarily indivisible. The latter condition may be important for unification of the state of their chromosomal machinery. The necessity of these conditions forces the cells to perform a multitude of migrations that may seem unnecessary. The body geometry owes much to mitotically silent cells.

On Expression of Fatomeres

A primary inductor (e.g., retinoic acid in many animals) causes copying of chromosomal templates and the formation of extrachromosomal fatomeres in the inductor-competent cells. This means that the cell competence is its preparedness to express appropriate fatomeres in response to the inductor. This state is mediated by local unpackaging of heterochromatin, in which the given genomic fatomere is encoded. This is why competent cells can undergo a series of divisions without losing the state of competence. However, after a local amplification has occurred and the extrachromosomal fatomere has begun its short life, mitoses are prohibited because they would interfere with estimation of positional information.

Extrachromosomal fatomeres, e.g., those controlling the formation of fingers, must be activated in a specific sequence that closely matches the local concentration of a primary morphogen (whose gradient acts on the group of competent cells) around each cell. The cell whose fatomere has been already activated acquires the ability (owing to its tripler RNA) to rearrange by triplexation the packaging of its chromatin in the appropriate chromosomal region. The process of triplexation, which will be considered in more detail in the next section, is essentially the formation of a triplex containing a RNA molecule and a nascent double-stranded DNA molecule that has not yet adopt its final double-helical B-form. It is important that the formation of such triplexes may occur not only upon the lagging strand synthesis, but also during DNA repair synthesis. The latter is exactly the type of synthesis that can and must occur even in nondividing competent cells located in the area covered by a primary inductor, provided that it stimulates the cell to produce considerable amounts of free radicals, thereby rapidly increasing the amount of single-stranded chromosomal DNA breaks that require repairing. The reason for increasing the level of free radicals in cells experiencing the action of a morphogen will be explained below in the section dealing with location DNA.

Experiments showed [1] that retinoic acid, a primary inductor, can form specific structures in a strict conformity with local concentrations of the inductor in a number of animal species. This occurs during regeneration and embryonic morphogenesis [1]. The corresponding cell group is usually labeled as the zones polarized activity (ZPA). These cells generate the gradient of a primary inductor. ZPA are consecutively formed in various parts of the embryo and induce the formation of regularly shaped structures along all three axes of the embryo.

The formation of paired structures, such as the right and left hands, is an important issue. This process requires a single set of fatomeres rather than a double set.

Where and How Is the Body Plan Recorded in the Genome? The Body Plan and the "Genetic Homunculus"

The body plan is stored in the genome as a genetic homunculus consisting of a system of fatomeres and acceptors of their signals. For example, each bone of the human skeleton is created under control of its own fatomere. The same applies to morphogenesis of other systems and organs. There is probably enough genetic material in the genome to perform this function because elements other than exons account for approximately 97% of the genome [39], and the function of this DNA fraction (called junk DNA) remains unexplained.

The formation of the body plan is a fundamental problem of developmental biology [1]. If fatomeres are the substrate of the body plan, then major macroevolutionary events should be expected to concern primarily the fatomeres and changes in the time (heterochrony) and place (heterotopy) of their expression in the body. The genomes of humans and chimpanzees are approximately 2% different, and the major difference concerns certain sequences of subtelomeric DNA [32-35]. It is conceivable that these differences mainly concern the fatomeres.

On the Mechanism of Triplexation and Epigenetic Memory

The mechanism of triplexation can be used for creating the epigenetic memory during cell determinations and differentiations. Triplexes can serve as initial landmarks for chromosome site-labeling or marking, e.g., before differential methylation of labeled DNA segments, and for marking chromosomal regions without covalent modification of nucleotides. In particular, triplexation can be used for labeling the repeats that should be methylated in the genome; such repeats have been found in various species.

DNA methylation occurs with a certain delay relative to the synthesis of Okazaki fragments [40]. This is an especially important feature that requires further explanation. Its meaning may be as follows. The delay of DNA methylation relative to DNA replication is necessary for triplexation (that is, the formation of a complex between the tripler RNA and the nascent double-stranded DNA resulting in a triplex of RNA-dsDNA) to have enough time to proceed. Triplexation (and the following, triplexation-related distribution of chromatin proteins in the marked chromosomal region) should occur before the nascent lagging DNA strand undergoes methylation according to the pattern of its template.

Triplexation and Signals from Fatomeres

One fatomere can have one or more transcribed domains (DNA loops). As a rule, these transcripts become tripler RNA species and place epigenetic marks in appropriate sites of chromosomes through the mechanism of triplexation. These epigenetic signals may be positive (activating the chromosomal segment) and negative (inactivating these sites). In all cases, these marks are inherited by offspring cells because their own RNA molecules are transcribed on marked regions; some of these RNA molecules act, in turn, as tripler RNAs and are responsible for local maintenance of the epigenetic state adopted by chromatin (epigenetic memory).

Morphogenesis, DNA Redundancy, and the Mechanism of Triplexation

While speaking about chromosomal evolution, it should be noted that the so-called redundancy of genomic DNA can, to a certain extent, be explained by the necessity for the genome to possess (in addition to known regulators) numerous repeated sequences that serve as targets for RNA transcribed on fatomeres. These targets are the most important elements of cellular determinations because they allow fatomeres to control the activation and compaction of chromosomal regions. Since the time of discovery of different classes of repeated sequences in eukaryotic DNA, numerous attempts have been made to understand the meaning of their existence in the eukaryotic genome. These repeats do not encode proteins; many researches suggested that they are molecular garbage or, at the best, egoistic DNA. These repeats seemed to perform no useful function; this resulted in the concept of DNA redundancy (alternatively termed the C-value paradox). However, the repeats were suggested to be involved in the cascade regulation of expression of structural genes [41]. Nevertheless, the precise role of repeats, especially those occurring with moderate frequencies, and the role of unique segments of noncoding DNA remain unclear.

As pointed out above, the proposed mechanism of triplexation is based on the assumption that tripler RNAs coded by fatomeres find their complementary segments in appropriate sites of chromosomes and hybridize with them during replication of the DNA lagging strand, as well as in DNA repair synthesis, to form marker triplexes (probably short-lived). This means that the presence of a considerable number of repeated sequences (which have at least several sites complementary to the corresponding fatomeres) in eukaryotic genomes is a necessary condition for the postulated process of triplexation to occur.

The greater perfection in locating the targets of fatomeric tripler RNAs is achieved by genomes in evolution, the more economic and rational may be the use of these targets. Primitive organisms may use a separate set of targets for each determination, whereas higher forms can use the same target in a number of combinations, which makes the genome more compact.

Organisms displaying similar morphological features often have genomes of different sizes. In addition, patterns of distribution of noncoding sequences may be similar in taxonomically distant species but different in closely related species (the so-called Drosophila-like and Xenopus-like types of distribution of repeats). An explanation to this phenomenon may be that the choice of the mode of the genome marking is a problem of evolutionary tactics and is not directly related to the level occupied by the animal on the evolutionary scale. The chosen pattern of chromosome marking during determination is probably related to the tempo of progress at the corresponding stages of embryonic and postembryonic morphogenesis but not to the final result achieved.

It should be stressed that the length of the segment of the tripler RNA involved in triplexation could be relatively small (probably, less that 20 nucleotides). This is a minimum required to achieve the necessary specificity. Relatively short acceptor sites of DNA required for hybridization with these signal ribotriplers emitted by fatomeres are located on chromosomes in strictly specific positions. Each of these acceptor sites is used for marking one border of the gene cluster that will be involved in the current cell determination. Each cluster requires a minimum of two acceptor sites to mark the start and end of the region. Accordingly, there should be two types of tripler RNA. The starting triplex is then used for spreading the sequential collective binding of appropriate chromatin proteins along the given DNA segment. The propagation of the wave of this binding is limited on the opposite end of the chromosomal region by the second (terminating) triplex. The significance of distribution of many different repeats within each chromosomal region is to ensure the particular configuration of chromatin dictated by messages (tripler RNA species) from fatomeres. Other potential acceptor sites are found within any chromosomal region; these sites are also involved in sequential cell determinations and will form their own triplexes in the next cell cycles. The more rational the construction of the eukaryotic genome, the smaller the number of repeats are required to ensure the precise chromosomal configurations that should be adopted during the entire morphogenesis.

Triplexation, Epigenetic Memory, and Cancer

Position effects, genomic imprinting, and some other phenomena of epigenetic memory could and should be based on triplexation, the existence of which is suggested here. The presence of many repeats having ori on chromosomal arms is required not only for making the repeats mobile but also for memorizing the signal from fatomeres by triplexation. The presence of an ori allows the acceptor of a tripler RNA to coordinate its replication with the fatomere transcription without interference of other ori.

Thus, triplexation is used by fatomeres as a means of control over the cell epigenetic memory. This process can be used for all modes of genome marking (covalent and noncovalent marking of cytosines; the latter occurs through reorganization of the protein spectrum of chromatin). In both cases, triplexes delimit the borders of the region within which DNA methylation or demethylation will participation. In this context, note that cancer is both a genetic (certain mutations are known to be carcinogenic) and epigenetic disease. Disorders of epigenetic memory in cancer can be manifested by the loss of the cell's achieved state of determination (or in achieving another stage of determination by the given cell line) with participation of fatomeres.

On the Role of Triplexation in Maintenance of the State of Cell Differentiation

Triplexation, a process used by fatomeres for creating and maintenance of epigenetic memory on the cell determination achieved may be also used for maintaining the epigenetic memory about the achieved level of cell differentiation. In the latter case, triplexation may involve the so-called 3´-UTR regions or pre-mRNA (that is, its noncoding 3´ end).

Triplexation and Coordination of Transcription and Replication in Chromosomes

Tripler RNA species acting as signals should hit their chromosomal targets at the moment of their replication; therefore, the precise coordination of transcription and replication schedules of appropriate DNA sequences is of considerable importance. This coordination can probably be maintained under variable conditions when the source (a fatomere) and the target (regulatory sequences accepting the tripler RNA) occur in the same arm of the chromosome. The species-specificity of the position of genes along chromosomes (the basis of chromosomal mapping) may be largely due to the necessity of matching the two schedules: the sequence and time of transcription and replication of DNA.

About "Egoistic" or "Junk" DNA and the Meaning of Existence of Transposones in Eukaryotic Genomes

The significance of existence of a considerable number of mobile elements in eukaryotic genomes is probably determined by the involvement of transposones in the transfer of acceptor sites capable of triplexation with regulatory RNA molecules (transcribed from fatomeres) across the genome. Because of this, one and the same fatomere can regulate (by local compaction or decompaction) different regions of genome in different species or even in one species. Therefore, transposones, in addition to the ability (of some of them) to be involved in mutations, allow the organisms to create new determinations and morphological structures. Transposones can make the existing determination more variable by recruiting a new structural gene, enhancer, etc., into the old determination. This statement bears a clear analogy with the model proposed by Britten and Davidson [41] long ago. It is important to stress the nature of similarity and differences between the two models. They are similar in that the mentioned acceptor sites of the genome are involved in concerted ensembles of the same genomic regions, each time in a new combination (like different symphonies written by using the same set of symbols). This is exactly the role played by different gene types in the model of Britten and Davidson [41]: receptor genes, integrator genes, etc. However, this terminates the list of similar features and opens the list of dissimilarities. The model of Britten and Davidson is a model of regulation of cell differentiation. The model proposed here is almost not concerned with the problem of differentiation; rather, it only deals with determinations. However, this is not the main difference. What is essential is that I propose here for the first time a new molecular mechanism of epigenetic memory working in cell determination: the triplexation mechanism.

Chromosomal Evolution and Telomeres

Fatomeres are always found in telomeres. No fatomeres exist in paracentromeric, short telomeres of telocentric chromosomes. They may be absent from short arms of acrocentric chromosomes. Acrocentrics often fuse via short arms to produce a metacentric chromosome without disturbing morphogenesis. This precludes the presence of chronomeres in these sites because they need a free end of chromosomal DNA to work in neurons.

Acceptors of Signals from Fatomeres and Homeosis Genes

Practically all animals have a family of genes called homeosis genes or Hox. In addition to certain other developmental genes, Hox genes play key roles in shaping the body in various species, from spiders to humans [44].

There is strict correlation between the order of expression of Hox genes in chromosomes and the topographical position of their expression along the main anteroposterior axis of the animal body [42, 43]. Hox genes can regulate the activity of genes occupying hierarchically lower positions and may be involved in demarcation of relative positions of certain parts of animal bodies during morphogenesis; however, these genes are not specialized for the formation of a specific structure.

Nucleotide sequences encoding homeodomains and/or repeats on the flanks of homeosis genes could be expected to meet the criteria for being potential acceptors of tripler RNA molecules. In this context, of special interest is the finding that the topography of specific homeosis genes in a chromosome corresponds to the topography of the body segments encoded by these genes. Like a sequence of pages in a book, each of these Hox genes appears to contain two independent instructions: 1) information coded in its own nucleotide sequence, and 2) information on its two (upstream and downstream) neighbors. This may help achieve not only a possible simplification of sequential unpackaging of condensed chromatin. With such a relayed (page-to-page) repackaging of chromatin, the control over the correctness and specificity of repackaging can and must be effected by the repackaging genes themselves (thereby liberating fatomeres from the responsibility for this kind of control). In this way, the sequence of pages in the book is maintained by pagination and binding rather than by additional written instructions. In turn, this opens the possibility of a considerable economy of DNA because the same template can now encode several consecutively positioned segments. In identical embryonal cells, fatomeres enable the expression of different gene groups without having any instruction on the specific information contained in these genes. Similar to a good manager, the fatomere ensures favorable conditions for specialized genes to work without interfering with details of their operation.

On the Role of Fatomeres in Evolution of Genomes and Redundant DNA. Chromosomal Evolution, Robertsonian Translocations, and Fatomeres

Robertsonian translocations, that is, chromosomal rearrangements involving the transfer of arms between chromosomes (or a complete chromosome onto another chromosome) is a widely used mechanism of chromosomal evolution [30]. Intraspecies polymorphism of Robertsonian translocations is characteristic of many mammals. The Robertsonian fan, an extreme variant of such rearrangements, is found in cases of wide variability of diploid chromosome numbers in various specimens of the same species, provided that the number of long chromosomal arms does not change. The total number of chromosomes can increase because of the appearance of single-armed (acrocentric and telocentric) chromosomes or decrease by chromosomal fusion. Therefore, fusion of two telomeres in long arms is a rarity in intraspecific chromosomal polymorphism, whereas centromeric fusion (internalization of the centromere of single-armed chromosomes) occurs frequently. From the point of view of the theory of chronomere, the Robertson's law is explained as follows: telomeres of long chromosome arms should be kept on free ends of chromosomes because internalization of a telomere would preclude the functioning of a chronomere in it.

Genomic Sites that Accept the Tripler RNAs from Fatomeres

The so-called taxon-printing, or the use of restrictase analysis of DNA repeats revealed the presence of short (less than 300 bp) species-specific nucleotide sequences in genomes of various species (lizards, fishes, and rodents) [44, 45]. All specimens of a particular species have the same set of these short DNA segments; however, different sets are found in many closely related species. This is consistent with earlier reported data on species-specificity of a certain fraction of DNA repeats, which is often sullied as egoistic, junk, and parasitic because its functions are unclear. The alternative concept suggests that these are regulatory elements of the genome that receive protein signals from the cytoplasm of the fertilized egg and activate the appropriate structural genes [45].

It is conceivable that many species-specific DNA repeats are involved in signal transduction from fatomeres (by means of triplexation). There is evidence that certain fractions of species-specific repeats are similar to tRNA pseudogenes (reviewed in [45]). A possible explanation is that tRNA genes were ancestral genes from which the first fatomeres have evolved.

On the Regulation of Regulators

This concept suggesting that triplexation is the basis of epigenetic memory probably resolves the old problem of "regulation of regulators". The vicious circle is broken because the persistent activity of certain higher regulatory systems, which could control hierarchically subordinate regulators, is no longer necessary.


The full regeneration of UV-irradiated (to block mitoses) tentacle hypostome in Hydra was demonstrated in [46]. This is consistent with the hypothesis that a temporary extramitotic state of a cell population is required for short-term activity of fatomeres. A similar explanation can be given to experiments with Xenopus embryos, in which cell cycles were arrested but morphogenesis proceeded over a long time.

Why Do Species Differ in the Effectiveness of Regeneration?

This is the central problem of regeneration [47]. Why humans do not regenerate their legs, whereas Sabella, a polychaete and champion in regeneration, can even restore its head and other parts of the body? The function of fatomeres is essential for regeneration, and its success under otherwise equal conditions depends on the method of fatomere decompaction. Regeneration is facilitated by easy compaction of heterochromatin containing fatomere templates.

Aging and the Location of Chronomeres

Fatomeres of brain neurons possess a buffer sequence on their free ends, which measures the biological time and is therefore called chronomere. In postmitotic cells of adult brain, these chronomeres are protected against nucleases and may probably be shortened only by the earlier postulated [17-21] mechanism of DNA end underrepair synthesis. Having lost their chronomere, neurons may no longer receive signals imparted by the DNA sequence located behind the buffer (chronomere). In contrast to other fatomeres, chronomeres may have a nearby situated sequence that functions as a modulator gene whose signals stimulate the work of other genes. The loss of the modulator upon aging decreases the total activity of many dependent genes.

The main feature of chronomeres in the brain is that they regulate the events occurring not only in their host cell but also in the whole body. Chronomeres that slowly shorten in neuroendocrine cells may serve as a clock emitting signals (mediated by neuroendocrine products) which initiate changes between particular stages of ontogeny, e.g., the change from deciduous to permanent dentition, the term of puberty, etc.

Antioxidants and some other geroprotective agents capable of decelerating the process of DNA end underrepair synthesis, when other conditions are equal, can affect the life span in humans and animals because their administration results in a more economical use of the length of chronomeres.

The length of a chronomere in brain neurons mainly depends on two factors. The first factor is the astronomic time elapsed after the moment of terminal differentiation of the cell. The amount of damage (and the number of repairing acts) increases with time, including instances of absolutely irreparable damage done to DNA end breaches. The second factor is the concentration of damaging agents (such as free radicals) and the activity of the damage-protecting antioxidant system.

Conclusions before Discussing the Role of Location DNA in the Eukaryotic Genome

The subtelomeric or other heterochromatized DNA populated by telomeres is a factor of cell determinations and switches the activity of large intrachromosomal blocks in response to signals imparted by fatomeres by means of tripler RNAs. The considerable role of telomeres in cell determinations requires the following comment on the problem of replicative aging of cells in vitro.

Under normal conditions, beyond the stage of cell competence, each fatomere should be stored in a tightly packaged state in order to prevent its hazardous unscheduled activation. The packaging may be done by means of certain sequences flanking the fatomere. This means that a fatomere that loses its end flank from the side of the absolute end of the telomere can cause unscheduled activation of chromosomal regions which were previously shielded from transcription. This opens the way to the corresponding unscheduled, spontaneous events of cell differentiation. This is exactly what may cause the differentiation of aging cells described in [48]. Having a long history of doubling, fibroblasts from various animal species exhaust their cell doubling limits (therefore, their telomeric DNA is considerably shortened due to end underreplication), and various unscheduled differentiations begin. Senescent cells whose Hayflick's limit is close to zero, display a number of changes in the expression of secretory proteins [49], which provides indirect evidence for possible involvement of changes in subtelomeric DNA. In addition to observations of cell behavior upon the exhaustion of their Hayflick's limit [50], the necessity of maintaining telomeres compacted is consistent with the so-called telomeric position effect, that is, a reversible repression of foreign genes inserted near the end of a chromosome [51]. Subtelomeric DNA was suggested to play a role of genes with an unknown regulatory function. For example, there is a hypothesis [52] that the loss of underreplicated chromatin may change the expression of certain subtelomeric genes controlling a senescent program and delay the growth of cells.

According to the interpretation of differentiation of aging cells [48] proposed here, the cells whose Hayflick's limit is exhausted [50] perform unscheduled determination, which is caused by euchromatization of the fatomere deprived of its compacting flank. The determination, in its turn, permits some new cytophenotypes.

Recording the signal coming from fatomeres in the form of ribotriplers is mediated by triplexation, a new mechanism proposed in this paper. The essence of this mechanism is the formation of a triplex between RNA and nascent double-stranded DNA. This process can occur near a replisomal loop of nascent DNA lagging strand (and also during DNA repair synthesis). Under these conditions, DNA is capable of forming triplexes with complementary segments of RNA because it adopts an intermediate conformation rather than the B-form with which RNA is unable to form a triplex under physiological conditions [53]. This effect is partly due to SSB proteins binding to a particular site of DNA template from which a replica is being synthesized.

The triplex formed on this site of double-stranded DNA then binds to the appropriate protein called triplerin. This process marks the flanks of any chromosomal region that will be later activated (a positive signal) or inactivated (a negative signal). Each ribotripler, or tripler RNA, consists of two domains. The first domain is complementary to a distinct site of the genome. This domain, called variable, has different sequences in different ribotriplers. The second (constant) domain binds chromatin proteins (triplerins). Ribotriplers (triRNA, or tripler RNA), having built the start and end triplexes at both ends of a certain chromosomal region, enable a strict location of all required patterns of chromatin modifications.

Triplexation may be promising for biotechnological applications.


Location DNA and the Positioning Mechanism as the Basis of Reading the Positional Information by Cells

How is positional information read before expression of fatomeres? The expression of fatomeres is already an act of interpretation of information. However, this information should first be obtained, that is, the cell should be able to use certain means of reading the information contained in the concentration gradient of a morphogen. How do cells manage to do so?

To answer this question, I postulate the existence of a mechanism, which is a key feature of morphogenesis. I call it location or positioning mechanism. It will be considered below, taking a limb formed under the influence of retinoic acid as an example.

Many lines of evidence suggest that positional control actually exists [1, 8]. To take only one example of an infinite multitude, consider experiments with regulation of the body size in haploid and pentaploid tritons. In pentaploid specimens, cells are larger but the whole body and individual organ sizes are the same as in their haploid counterparts. This is due to a smaller number of cells in each tissue of pentaploid tritons.

An interesting feature is that not all embryonic tissues are capable of undergoing Drieschian embryonic regulations, but only those parts of tissues in which the fields of cell tension have already emerged as distinct topological structures: appropriately positioned strands and tapes of extended, polarized cells [38].

Experiments reviewed in [1] showed that competent cells can respond to an inductor gradient (within the range of its physiological concentrations) by forming a strictly defined set of anlages located in a standard order relative to each other. What is the mechanism capable of responding by an adequate multicellular structure to any particular concentration of a morphogen?

Can fatomeres and other genomic structures essential for morphogenesis be somehow hidden in the three-dimensional structure of chromosomes until a certain "developing" concentration of the inductor (an analogy to the photographic developer) allows the appropriate DNA sequences to be activated? This formal analogy can, to a certain extent, illustrate the approach described in this section.

Let us assume that a fatomere required for the formation of the first anlage is compacted most effectively in a chromosome, whereas the second fatomere required for the emergence of the next anlage is compacted to a lesser extent. Assume then that not only proteins but also lipids are used for compaction. Certain lipids are easily degradable by oxidation. Assume that different packaged domains of heterochromatin containing fatomeres are packaged with involvement of different lipids and lipid--protein complexes.

Further, assume that the inductor gradient causes cells to generate free radicals in amounts directly depending on the concentration of the inductor (such as retinoic acid) in the environment. The minimal concentration of free radicals in cells will be found at a considerable distance of the source of the inductor. Therefore, only the least resistant lipid bridges sealing the chromatin will be destroyed in these remote (distal) cells. This particular domain, which may contain a certain fatomere, is then decompacted. The highest concentration of free radicals is found in the vicinity of the source of the inductor gradient; therefore, another domain (containing its own fatomere), whose lipid binder (seal) displays a greater resistance to free radicals, will be decompacted in cells occupying this position. Therefore, at this site both fatomeres (those more and less effectively sealed by lipids) will be accessible to polymerases. In the vicinity of the source of the morphogenic gradient, both fatomeres will behave like genies released from their lipid jars. Assume that two fatomeres together encode an anlage that will appear in the region of a high concentration of the inductor. However, the fatomere activated alone after breaking the lipid seal (at a low concentration of the inductor) will cause an entirely different anlage located far from the first rudiment. Therefore, the two anlages will emerge at two different sites in relation to the source of the morphogenic gradient: exactly where they have been captured by the image-developing effect of free radicals.

Oxidative destruction can damage various cell components. However, lipids are known to be the most vulnerable to this damage. Carbohydrates and especially nucleic acids are much more resistant to oxidation in comparison to unsaturated lipids. The cause is known: lipid oxidation proceeds by the free-radical chain mechanism, which does not occur in nucleic acids and proteins. It is also known that free radicals can attack DNA and cause single-strand breaks (nicks) and breaches but no chain reaction will occur in this polymer. In addition, DNA is protected by repairing systems. Active oxygen species that cause free-radical oxidation of many molecules are especially effective in oxidizing unsaturated phospholipid residues. Free-radical oxidation of cellular lipids is a normal process. However, excessive production of active oxygen species causes a profound degradation of unsaturated lipids and is hazardous to cells. Antioxidant defenses protect cells from this damage. Generally, cells seem to possess regulatory systems that allow them to either increase or decrease the level of free radicals in various compartments. It is conceivable that the concentration of free radicals can change in the immediate vicinity of chromosomes.

Lipid Bridges of the "Stereogenome"

Lipid (or lipid--protein) bridges may be involved in packaging certain distinct domains of chromatin and form complexes with only nontranscribed DNA, which encodes neither protein nor RNA. Nevertheless, this nontranscribed DNA is a coding DNA species because it contains information on when lipoprotein bridges must appear and how the given chromosomal segment must be compacted in order to keep it ready to read positional information during individual development. In this sense, the three-dimensional configuration of the chromosome (stereogenome) contains an additional amount of genetic information which is necessary for development and cannot be received by cells by means of reading the nucleotide sequence by polymerases. The eukaryotic genome seems to be redundant because cells use a number of means to read the information contained in its DNA. Cells receiving, e.g., positional information at a certain stage of development use an alternative mode of reading the template, whereby the genome is not only a bank containing one-dimensional information but also works as a kind of analog device tentatively called stereogenome. Its functioning is explained in this section.

In accordance with these statements, let us assume that lipids, in cooperation with proteins of chromatin, play a key role of binders or bridges maintaining a highly compacted state of specific regions of the chromosomal DNA fraction that will be involved in reading the positional information by competent cells. Assume that when these bridges are destroyed by oxidation, these particular DNA segments are rapidly unpackaged.

Assume also that certain chromosomal modules (that is, DNA segments folded with proteins and lipids to form three-dimensional superdomains) encode the formation of strictly specific morphological structures. Each of these modules will be designated as a morphon. The sum of all morphons can be called the genetic homunculus. Certainly, three-dimensional chromatin structures are not similar in shape to a wing or hand; however, each fragment of these parts of the body is represented in the genome by compacted domains of heterochromatin called a morphon. For example, one chromatin module encodes the ability to determine the formation of an anlage of one bone of the skeleton. The formation of the rudiment of the second bone can be encoded by another morphon, which could be located in a chromosome near the first morphon, and so forth. Bones located near each other in the skeleton are most probably encoded by adjacent morphons in a chromosome. This would be helpful in unpackaging and repackaging chromatin. However, this close neighborhood is not an obligatory condition for the information-reading mechanism proposed here. Three-dimensional units of chromatin, that is, morphons, contain sequences of fatomeres and acceptor sites for receiving instructions from ribotriplers (see the preceding section on the role of ribotriplers, triplexes, and triplexation). These instructions govern the emergence of epigenetic memory on the achieved state of cell determination and on the creation of a specific anlage (e.g., that of the second phalanx of the third finger of the hand).

Another important question is how exactly the difference in the shape of the bone rudiments is achieved after the positional information has already been read. Here we should take into account the following.

The order of positions of anlages emerging in the area covered by a concentration gradient of an inductor is by itself a direct effect of the sequence of unpackaging different morphons. Oxidative destruction of lipids is the basis of the natural technology by which cells read their positional information.

The orientation of every anlage is governed by mechanical tensions in the tissue. These tensions determine both the precise direction of the inductor spreading and the position of the main axis of the future anlage whose width and thickness are probably determined in the same way, that is, by lines of mechanical tensions in the morphogenetic field.

The length of each developing anlage will be regulated later, exactly when those genes begin to work whose activation is allowed under the given cell determination. In other words, it is controlled by the given morphon, its fatomeres, etc.

I suggest that the above-mentioned competition between anlages is based on association of cells, which begin reading their positional information, into separate compartments; each compartment (anlage) contains cells communicating through gap junctions and has a certain individual concentration of free radicals.

Gap junctions located near the borders of the compartment gradually close, thereby isolating the future bud from other cells. An increased activity of the anlage in recruiting neighbor cells into its own compartment can strongly affect the size of the future organs.

Concerning the shape of the emerging anlage, its length, width, thickness, and orientation, these parameters change with the growth of the anlage. For example, the middle parts of the longest anlages become thin because rapid growth stretches the structure. These features are not coded simultaneously with reading the positional information and are products of further development. In terms of mechanics, any growing biological object has regions of volume expansion and outward thrust (like in an unevenly heated body) resulting in a constant field of mechanical tensions closely related to a deformation field [54]. As a result, complex tissue structures need not be preprogrammed in every minute detail. On the genetic level, the final shape (and intermediate shapes) are coded by, e.g., the initial position of anlages in the embryo relative to each other (by reading the positional information), the growth rates of particular elements of the construction, marking the positions of stiffening ribs and the least breakage-resistant parts, etc. However, all these events are based on the crucial process of morphogenesis: the reading and interpretation by cells of their positional information.

The Stereogenome, DNA Repeats, and Evolution

The chromosomal location of morphons, the modules determining the formation of a discrete embryonic morphological structure such as an anlage of a bone, is an important trait of the chromosome. This trait is no less important than the location of a finger on the hand rather than on the ear.

Transposition of a morphon from its standard place to another environment can place it in the vicinity of particular DNA sequence that can force the morphon to adopt a new mode of packaging in heterochromatin. In this new environment, the transposed morphon will be packaged by new lipid binders, which may have different spectra of unsaturated lipids, phosphate residues, etc.

This would eventually change the behavior of this morphon in its reading the position information: the morphon can start unpackaging at a previously unprogrammed concentration of the inductor, resulting in dramatic alterations in morphogenesis (some of them will be lethal, whereas others may cause considerable evolutionary disorders).

The postulated genomic modules (morphons) may contain unique, moderately and frequently repeated sequences, but they usually do not contain "housekeeping genes". Morphons are located in heterochromatin. The phylogenetic phenomenon of qualitative and quantitative changes in particular repeats may be exactly a manifestation of evolutionary work intended to create and modify morphons. Changes in the time and/or space of expression of morphons responsible for anlage positioning and carrying important regulatory elements of cell determination (such as fatomeres) will inevitably be clearly manifested during biological evolution. The literature on molecular evolution contains ample evidence that various families of repeated sequences are propagated, their distribution maps are changed, transposones are involved in genomic reshuffling, etc. [55]. The theory described here suggests that all these changes can be understood as events related to the fate of morphons, alterations in the set of fatomeres, and modification of packaging of a certain morphon with lipid seals of different composition and properties. In other words, the apparent evolutionary mess in what most authors describe as egoistic DNA is actually a highly expedient process. In most cases, it is clearly intended to change the mode of reading the positional information, that is, to change morphological traits. However, what biological evolution is if it is not the evolution of shapes? What are differences between an elephant and a mouse or a mouse and a biochemist? Certainly, the shape and behavior are different; the latter is largely determined by the former (the architectonics of the body and brain). In addition, changes in the shape often lead to physiological changes. If phylogeny, during which basic systems (such as genetic code or bioenergetics) undergo few changes, is based on the play of morphological traits, then everything related to reading the positional information is of particular significance in evolution. That is why morphons (morphological modules of a genome) are especially important in eukaryotes.

In light of this, the fraction of DNA formerly considered as egoistic DNA appears to play a new role. It becomes what can be called location DNA, that is, DNA required for creating morphons as key elements of the mechanism enabling cells to read their positional information.

Are Secondary Morphogens Needed?

The functioning of the positioning mechanism suggested here and based on the use of location DNA, morphons, and their lipid binders, makes it unnecessary to use the relay-race of secondary morphogens at the stages when they otherwise (without the proposed mechanism) would have to produce consecutive activation of fatomeres and their acceptor sites during morphogenesis. Nevertheless, this does not exclude the possibility that secondary morphogens [2] play an important role facilitating intercellular interactions, which is especially important in regeneration.

Nontranscribed DNA and Its Lipid Binders

Nontranscribed DNA is a key factor of morphogenesis. A new function ascribed to this large fraction is to create morphons. The morphon is a large segment of DNA compacted in heterochromatin and containing transcribed DNA (that codes the cell determinations and the information on the composition of a particular anlage) and a large fraction of nontranscribed DNA. The latter is needed for structural purposes, especially for compaction of the particular genomic module and sealing it with specific lipid bridges. These lipid-containing binders of different morphons must be different in terms of resistance to free-radical-induced destruction and, as pointed out above, are important elements of the "morphon-based reading machine". This machine reads the positional information whose interpretation leads to what Driesch summarized by the expression: the property of the part is a function of its position in the whole.

Presentation of their lipid binders by morphons subjected to selective oxidative destruction by free radicals in the inductor gradient is expedient only when the hypothesis proposed here (that the contents of free radicals generated by a competent cell correlates with the local concentration of the embryonic inductor) holds. Destruction of lipid bridges, which are different between domains and display different resistance to free-radical attack, unpackages a morphon. This heterochromatin domain carries, in addition to location DNA (nontranscribed), certain transcribed sequences, such as fatomeres.

The nontranscribed fraction of a morphon (location DNA) is part of highly compacted heterochromatin and causes a configuration of the whole morphon, which could be selectively unpackaged upon reading the position information, by simply "burning" the lipoprotein fetters of the genome. The expression of the transcribed part of a morphon is impossible until it is unpackaged and released from lipoprotein binders (sometimes numerous).

It is important to note that lipid bridges of various modules of heterochromatin can include lipids, proteins, and even probably short amplified extrachromosomal copies of certain sites of heterochromatin, and/or riboproteins that can ensure precise placement of lipid binders. By the way, the ability of triplexation to precisely mark chromosomal regions could be used in development for the same purpose, that is, for precise positioning on the chromosome of complexes responsible for synthesis of lipid binders.

All these processes can ensure the necessary heterogeneity of complexes formed by lipids and morphons, which is absolutely necessary for achieving different degrees of resistance of different lipid seals that must be broken at various concentrations of free radicals.

Thus, the concept presented here postulates that the three-dimensional structure of chromosomes (in the nontranscribed but coding DNA of an eukaryotic genome) contains a specific kind of additional information that allows the translation of ordinary (transcribed) genetic information into the three-dimensional structure of the embryo. The cell evaluates its own positional information because lipid binders are removed by oxidative destruction caused by free radicals. The concentration gradient of a morphogen (such as retinoic acid) induces free-radical lipid oxidation in cells; the intensity of oxidation depends directly on the local concentration of the morphogen. If different DNA domains (morphons) are compacted in the chromosome by lipid binders of different compositions, their breakage can be selective, depending on the local contents of free radicals, which, in turn, can depend on the local concentration of the morphogen. This results in selective decompaction of certain chromosomal modules (morphons) in accordance with the spatial position of the cell. Under these conditions, morphons will be unpackaged in competent cells occurring in the inductor gradient in a strict order (in time and space). This unpackaging will permit the expression of the transcribed fraction of the morphon DNA. This is the essence of the mechanism of positioning or receiving and reading the positional information suggested in this paper.

The proposed concept of location DNA makes unnecessary the commonly accepted notion of the existence of redundant or egoistic DNA, which is now assumed to be location DNA.

Morphon and Fatomere. The Possibility of Short-Term Existence of Fatomere as Circular DNA

A morphon is unpackaged in the presence of a relatively high concentration of free radicals. Therefore, the immediately expressed linear form of a fatomere could be subjected to extremely rapid shortening due to DNA end underrepair synthesis because end breaches in linear DNA molecules can never be fully repaired [17]. In light of this, it is possible that fatomere expresses as a circular DNA. The possibility of the existence of some regulatory elements in a circular form was discussed, although in a different aspect, in [56]. The circle has no ends and is protected from DNA end underrepair synthesis until the cell acquires the ability to transcribe information from this circular amplification product under normal conditions (in terms of the concentration of free radicals, which must decrease soon after the cells have left the area controlled by the morphogenic gradient).

The Role of Past History of Cells Occurring in a Morphogenetic Field

The number of cells occurring in a morphogenetic field rarely exceeds one hundred [13], whereas the mean concentration of trans-retinoic acid, a common animal morphogen, near cells is approximately 25 nM (at least this is true of chick embryonic tissues [57]). The characteristic time of the emergence and fading of competence and the ability to respond to morphogens in amphibians is measured by tens of hours [58].

The visually indistinguishable cell types involved in the formation of a limb, when occurring in a common morphogenetic field at equal distances from the source of the morphogen can, in accordance with their previous history, become, e.g., a cartilage or be a source of fibroblasts of connective tissue [59]. These phenomena are explained by the previous history of the cells: different cell types prepare different morphons (having different types of lipid seals) for obtaining their positional information. Therefore, the same concentration of free radicals in two different cell types will destroy different lipid binders, unpackage different DNA domains (morphons) and, as a consequence, will permit further differentiations of dissimilar cell lines.

On Orientation of Mechanical Tensions

The orientation of tensions changes in the morphogenetic field simultaneously with the formation of limbs, a process often taken as an example of well-studied objects of morphogenesis. The early rudiment of the primate hand is known to be positioned at a nearly right angle to the body axis. After the formation of the humeroulnar articulation, the forearm, upper arm, and hand are rotated ventrally so that the palm of the hand turns toward the body. The hand then rotates by 90° around its long axis, so that the elbows are turned caudally. This is an additional piece of evidence for the importance of permanent changes in mechanical tensions [8, 47] and probably changes in channels formed by cells, along which morphogens spread by diffusion. The phenomenology of the process is that the formation of limbs occurs in the proximodistal direction, and distal structures are consecutively added to the first anlage (which is formed in the vicinity of the source of the inductor). The addition of these distal structures can by itself change the direction of local mechanical tensions produced by supporting elements of the tissue, which could simultaneously serve as the system coordinating and redirecting (during development of the bud) the local flows of the inductor (e.g., retinoic acid) gradient. Note that local mechanical tensions can affect (through the system of links between the extracellular matrix, cytoskeleton, and nucleoskeleton, which is directly connected to chromatin) mechanical tensions in units consisting of large chromosomal regions. This, in turn, can influence a rapid, large-scale repackaging of the genome in its units preprogrammed to be sensitive to such an external mechanical stimulus. Artificial incisions in tissues could probably interfere with the normal development of this process [12, 60]. When competent cells experience reorientation of axes of mechanical tensions after they have undergone the positioning breaks of lipid bridges, they probably undergo unscheduled breakage of certain remaining lipid bridges in the genome. This is possible because of the continuity of links between cytoskeleton, nucleoskeleton, and extracellular matrix.

Thus, concentration-dependent, selective breakage of lipid binders, unpackaging of the morphon, and therefore the reading the cell's positional information create favorable conditions for the formation of anlages. In zones with high concentrations of the inductor (and free radicals), that is, in the vicinity of the source of a morphogen, all lipid binders (even the most resistant of them), which seal morphons located in cells of the same competence, break. On the opposite end of the morphogenetic field, that is, at minimal concentrations of free radicals, only the weakest lipid binders can be broken, for example, those of the last morphon. A cell found in the middle of this distance may undergo the activation of approximately a half of all morphons. However, all of these are nonrandomly selected, strictly predetermined morphons and their fatomeres. The appropriate subpopulations of cells create anlages coded by a distinct group of morphons.

The act of transcription of fatomeres (and the corresponding act of determination) may be delayed to the time point when the morphogen no longer acts on cells, and the levels of free radicals are normal in cells whose morphons have already been unpackaged. This delay is advantageous because transcription and other processes can develop without the disturbing interference of free radicals.

It should be noted that the mechanism of unpackaging morphons by breaking lipid binders is alone, without additional mechanisms, capable of achieving an optimal level of the morphogenetic use of the whole mass of available competent cells. If the cells are few, the inductor gradient will be steep but morphons will be unpackaged by fee radicals following the same scheme as in gently sloping (not steep) gradients. The size of the anlages formed in flat gradients will be greater because the same binder type will be burned in a greater number of cells (however, this process will take a longer time). Therefore, more cells will be included in the anlage. The total structure of the future organism will not change, that is, the phenomena will be exactly the same as observed by Driesch in his classical experiments.

The relative sizes of all anlages of the future wing are not considerably different at the moment of their emergence [61]. This could be expected because morphons are unpackaged in a discrete mode in the primary inductor gradient.

Diminution of Chromatin and Morphons

In the approach discussed in this paper, the sense of genomic rearrangement upon diminution of chromatin in ascarides and cyclopes is to prepare morphons to read the positional information. It is assumed that these animals cannot perform a large series of consecutive repackaging in long, gametically inherited chromosomes, and the morphons cannot be prepared to reading the positional information by mechanisms common to other animals. Therefore, ascarides are forced to rearrange the genome and create short chromosomes, in which the processes of domain repackaging do not interfere. Cyclopes discard several medial DNA fragments from various chromosomes, probably in order to bring closer the sequences that will create a required morphon and/or serve as acceptors for signals delivered from fatomeres.

Heterochromatin and Lipid Binders

Lipid molecules alone cannot find appropriate nucleotide sequences in the genome in order to seal the packaged morphon. Therefore, they need help offered by proteins and DNA. The simplest way to resolve this problem has been pointed out above: the local amplification (polytenization) of appropriate sites. These regions should be present in heterochromatin, which can bind to lipids that serve as binders for morphons. These lipid-containing bridges are required only within a single cell cycle in which the positional information is read; therefore, there is no need to ensure that the amplified heterochromatin linked to lipid binders inherits its localization in future cell divisions.

The Morphogenetic Cycle

As pointed out by Spemann, morphogenesis is essentially a chain of sequential inductions, that is, interactions between an inductor and competent cells [8]. The essence of the morphogenetic cycle repeated in different tissues during various types of organogenesis is as follows: 1) selective burning of lipid binders in accordance with location of the cell in the inductor gradient; 2) unpackaging the morphons and amplification of fatomeres; 3) transcription of fatomeres and recording their signal during triplexation; 4) repackaging of chromatin with creation of new morphons and their lipid binders for the next act of reading the positional information; and 5) translocation of cells relative to a new source of an inductor in order to accomplish the next morphogenetic cycle.

Mechanical tensions could facilitate the standardization of the state of all competent cells of a given morphogenetic field at the same stage of packaging of their chromatin, as well as prevent mitoses, which are inappropriate at this stage.

During morphogenesis, a shape is first created without differentiations in a manner similar to that of an artist who draws a contour before filling it with colored details.

A Brief Conclusion

Taking all facts together, this study suggests that the fraction of DNA formerly treated as egoistic exists in the genome for directing embryonic development. During this period, this DNA pool forms morphons sealed with lipid binders. The morphons are responsible for reading the positional information. In adult organisms, this DNA is rarely used. For example, birds need it to make feathers, and various organs require this DNA for regeneration.

The approach proposed here seems to make it unnecessary to use the theory of dissipative structures for explaining self-organization during biological morphogenesis, as it is often done in the literature.

This theory suggests that the main part of the nontranscribed fraction of the genome is used in morphogenesis, specifically for reading the cell's positional information and translating the DNA linear records into the three-dimensional structure of the embryo. Therefore, studies focused on other stages of development yielded misleading conclusions that this fraction of DNA is never used. According to its function, I propose to call this DNA fraction location DNA. The information contained in it can be read by well-investigated cellular mechanisms that use the template principle in various modes. I suggest that Nature has designed an unusual but simple method to perform an unusual task of reading the positional information. The essence of this method is to create regions of chromatin sealed with lipid bridges in chromosomes. The bridges display different resistance to destructive effects of free radicals whose level increases in proportion to the concentration of an inductor (e.g., retinoic acid) in the environment of the cell. As the concentration of free radicals in cells increases, the lipid bridges are destroyed one by one and liberate their chromosomal modules (morphons) containing information on anlages and cell determinations. This becomes the basis for the mechanism that helps cells to evaluate their position in the inductor gradient. Despite the continuity of the inductor gradient, there are discrete differences between cells that have unpackaged different chromosomal modules.

High rates of DNA nicking in early embryogenesis provide indirect evidence for the involvement of free radicals in processes occurring at this period [28]. This effect can account for a rapid increase in the frequency of recombination of minisatellite DNA in early development [62] and the sister-chromatid exchange [63, 64]. Many sites of DNA damage are repaired, but the nicks themselves can be produced by high concentrations of free radicals.

There is evidence that trans-retinoic acid (the most active and well-studied agent in animal morphogenesis) uses the following mechanism for regulating the levels of free radicals in cells. Retinoic acids regulate the activity of protooncogene bcl-2, which is involved in prevention of oxidative destruction of lipids [65-67].

Histone H1 displays a high affinity toward acid phospholipids, which is consistent with the important role of chromatin-binding lipids (see [68] for references).

A piece of indirect evidence for the possibility of activation of fatomeres after the loss of telomeric DNA upon end underreplication (see the above discussion of data reported by Bayreuther et al. [48]) is provided by the following observations. Decondensation of chromatin and the appearance of extrachromosomal circular DNA were found in the last mitoses of fibroblasts aging in vitro [69]. The function of this kind of DNA may be to prevent the DNA end underrepair of fatomeres.

In conclusion, I must note that the molecular model of regulation proposed by Britten and Davidson [41] provides no solution to the main problem of developmental biology: why different groups of genes are activated in identical embryonic cells. This paper provides an answer to this old question. The missile of morphogenesis is propelled by lipid fuel.

This work was supported in part by grant No. 96-04-13127 from the Russian Foundation for Fundamental Research.


1.Meiden, M. (1995) Ontogenez, 26, 419-429.
2.Mikhailov, A. T. (1988) Embryonic Inductors [in Russian], Nauka, Moscow.
3.Driesch, H. (1892) Z. F. Wissenschaft. Zool., 53, 160-178.
4.Morgan, T. H. (1906) J. Exp. Zool., 2, 495.
5.Wolpert, L. (1969) J. Theor. Biol., 25, 1-47.
6.Wolpert, L. (1989) Development (Suppl.), 3-12.
7.Driesch, H. (1915) Vitalism: History and System [Russian translation], MGU, Moscow.
8.Beloussov, L. V. (1987) Biological Morphogenesis [in Russian], MGU, Moscow.
9.Meinhardt, H. (1989) Development (Suppl.), 169-180.
10.Kondo, S. A., and Asai, R. (1995) Nature, 376, 765-768.
11.Meinhardt, H. (1995) Nature, 376, 722-723.
12.Beloussov, L. V., Saveliev, S. V., Naumidi, I. I., and Novoselov, V. V. (1994) Int. Rev. Cytol., 150, 1-34.
13.Crick, F. H. C. (1970) Nature, 225, 420-422.
14.Olovnikov, A. M. (1971) Dokl. Akad. Nauk SSSR, 201, 1496-1499.
15.Olovnikov, A. M. (1972) Vestnik Akad. Med. Nauk SSSR, 12, 85-87.
16.Olovnikov, A. M. (1973) J. Theor. Biol., 41, 181-190.
17.Olovnikov, A. M. (1992) Izv. Ross. Akad. Nauk. Ser. Biol., No. 4, 641-643.
18.Olovnikov, A. M. (1995) Izv. Ross. Akad. Nauk. Ser. Biol., No. 4, 501-503.
19.Olovnikov, A. M. (1995) Izv. Ross. Akad. Nauk. Ser. Biol., No. 4, 504-507.
20.Olovnikov, A. M. (1995) Ontogenez, 26, 254-256.
21.Olovnikov, A. M. (1995) Ontogenez, 26, 332-334.
22.Olovnikov, A. M. (1996) Exp. Gerontol., 31, 443-448.
23.Henriksson, J., Aslund, L., and Pettersson, U. (1996) Parasitol. Today, 12, 108-114.
24.Szymanska, J., Tarkkanen, M., Wiklund, T., Virolainen, M., Blomqvist, C., Askoseljavaara, S., Tukiainen, E., Elomaa, I., and Knuutila, S. (1996) Gene Chromosome Cancer, 15, 89-94.
25.Khesin, R. B. (1984) Instability of the Genome [in Russian], Nauka, Moscow.
26.Prokof'eva-Bel'govskaya, A. A. (1986) Heterochromatic Regions of Chromosomes [in Russian], Nauka, Moscow.
27.Kingham, K. I., Duckett, J. G., Glyn, M. C. P., and Leitch, A. R. (1995) New Phytol., 131, 543-556.
28.Vatolin, S. Yu., Okhapkina, E. V., Zhdanova, N. S., and Serov, O. L. (1996) Dokl. Ross. Akad. Nauk, 348, 834-837.
29.DiBerardino, M. A. (1987) Amer. Zool., 27, 623-644.
30.Birshtein, V. Ya. (1987) Cytogenetic and Molecular Aspects of Evolution of Vertebratres [in Russian], Nauka, Moscow.
31.Greider, C. W. (1996) Annu. Rev. Biochem., 65, 337-365.
32.Yunis, J., and Prakash, O. (1982) Science, 215, 1525-1530.
33.Goodman, M., Tagle, D., Fitch, D., Bailey, W., Czelusniak, J., Koop, B., Benson, P., and Slighton, J. (1990) J. Mol. Evol., 30, 260-266.
34.Royle, N., Duncan, M., and Jeffreys, A. (1994) Nature Genetics, 6, 52-56.
35.Thoraval, D., Asakawa, J., Kodaira, M., Chang, C., Radany, E., Kuick, R., Lamb, B., Richardson, B., Neel, J. V., Glover, T., and Hanash, S. (1996) Proc. Natl. Acad. Sci. USA, 93, 4442-4447.
36.Vandermeulen, M. C. H., and Carter, D. R. (1995) J. Theor. Biol., 172, 323-327.
37.Frankhauser, G. (1955) Analysis of Development (Willier, B. H., Wiess, P. A., and Hamburger, N. P., eds.) Saunders, Philadelphia, pp. 126-150.
38.Beloussov, L. V. (1979) Ontogenez, 10, 120-129.
39.Koop, B. F. (1995) Trends Genet., 11, 367-371.
40.Kirnos, M. D., Aleksandrushkina, N. I., Vlasova, T. I., and Vanuyushin, B. F. (1995) Mol. Biol. (Moscow), 29, 736-746.
41.Davidson, E. H., and Britten, R. J. (1973) Quart. Rev. Biol., 48, 565-613.
42.Carroll, S. B. (1995) Nature, 376, 479-485.
43.Lewis, E. B. (1978) Nature, 276, 565-570.
44.Fedorov, A. N., Grechko, V. V., Slobodyanyuk, S. Ya., Fedorova, L. V., and Timokhina, G. I. (1992) Mol. Biol. (Moscow), 26, 464-469.
45.Mednikov, B. M. (1995) Ross. Khim. Zh., 39, 30-38.
46.Wolpert, L., Hicklin, J., and Hornbruch, A. ( 1971) Symp. Exp. Biol., 25, 391-400.
47.Karlson, B. M. (1986) Regeneration [in Russian], Nauka, Moscow.
48.Bayreuther, K., Francz, P. I., Gogol, J., and Kontermann, K. (1992) Ann. N. Y. Acad. Sci., 663, 167-179.
49.Shay, J. W., West, M. D., and Wright, W. E. (1992) Exp. Gerontol., 27, 477-492.
50.Hayflick, L., and Moorhead, P. S. (1961) Exp. Cell. Res., 25, 585-621.
51.Zakian, V. A. (1989) Annu. Rev. Genet., 23, 579-604.
52.Wright, W. E., and Shay, J. W. (1995) Trends Cell Biol., 5, 293-296.
53.Soyfer, V. N., and Potaman, V. N. (1996) Triple-Helical Nucleic Acids, Springer-Verlag, New York.
54.Martynov, L. A. (1982) Matematical Biology of Development (Zotin, A. I., and Preshov, E. V., eds.) [in Russian], Nauka, Moscow, pp. 135-154.
55.Charlesworth, B., Sniegowski, P., and Stephan, W. (1994) Nature 371, 215-220.
56.Romanchikov, Yu. A. (1991) Usp. Sovr. Biol., 111, 643-653.
57.Thaller, C., and Eichele, G. (1987) Nature, 327, 625-628.
58.Beloussov, L. V. (1982) Matematical Biology of Development (Zotin, A. I., and Preshov, E. V., eds.) [in Russian], Nauka, Moscow, pp. 102-111.
59.Newman, S. A. (1989) Trends in Genetics, 4, 327-332.
60.Beloussov, L. V., Luchinskaia, N. N. (1995) Biochem. Cell. Biol., 73, 555-563.
61.Wolpert, L., Lewis, J., and Summerbell, D. (1975) Cell Patterning, Ciba Foundation Symp., Vol. 29, London, pp. 95-119.
62.Armour, J. L., and Jeffreys, A. J. (1992) Curr. Opin. Genet. Develop., 2, 850-856.
63.Bennett, J., and Pedersen, R. A. (1984) Mutat. Res., 126, 153-157.
64.Patkin, E. L., Kustova, M. E., and Dyban, A. P. (1994) Cytogenet. Cell Genet., 66, 31-32.
65.Hu, Z. B., Yang, G. S., Li, M., Miyamoto, N., Minden, M. D., and Mcculloch, E. A. (1995) Leukemia, 9, 789-798.
66.Korsmeyer, S. J., Yin, X. M., Oltvai, Z. N., Veisnovack, D. J., and Linette, G. P. (1995) Biochim. Biophys. Acta (Mol. Basis Dis.), 1271, 63-66.
67.Shimizu, S., Eguchi, Y., Kosaka, H., Kamiike, W., Matsuda, H., and Tsujimoto, Y. (1995) Nature, 374, 811-813.
68.Rytomaa, M., and Kinnunen, P. K. J. (1996) Biochemistry, 35, 4529-4539.
69.Macieira-Coelho, A. (1995) Mech. Age Dev., 82, 91-104.