Portugaliae Acta Biologica - 1949
A Cytophysiological Theory of the Gene, Gene Mutation
and Position Effect
J. A. Serra
Secção de Zoologia-Anthropologia, Faculdade de Ciências. Universidade, Coimbra
After a detailed discussion of gene action and gene evolution and of the effects of chromosome rearrangements, we are now in a better position to approach the problem of gene nature and the mutation process on a sound basis. The chief part of this chapter will be devoted to a discussion of the chemical composition of chromonemata and genes and the changes corresponding to gene mutation. A brief interpretation of the causality of gene mutations and chromosome rearrangements will also be attempted.
A preliminary to chromosome rearrangementsand there is also strong reason to admit that the same holds true for gene mutationis the production of chromonema breaks. These breaks may be produced in any phase of the mitotic cycle or during the resting stage, as is demonstrated by many experiments with different kinds of cells. However, the resting stage and the diverse mitotic phases may have different sensitivities to breakage, this being due to several causes which act either upon the chromosomes as a whole or, separately, upon the matrix and the chromonemata. Different materials and methods of observation give different results in this respect. Two distinct processes interact to produce certain types of rearrangements: breakage and reunion. Each visible chromosome or chromatid break is in itself the end result of two processes also: the breakage of the chromonema and the failure of restitution, by this latter term being designated the re-formation of the original chromonema by union of the broken ends. There is substantial evidence which points out that, at least in some materials, the greater part of the breaks which occur are never visible by virtue of they restituting (Lea 1946).
For each visible and lasting rearrangement, therefore, it is necessary, at least: 1) that a break be produced; 2) that the break fails to restitute; and 3) that the broken end either a) in rearrangements involving at least two breaks, unites with another broken end different from the original one, or alternatively b) in a one break rearrangement (fragment) that the break "heals". Any factor capable of influencing one or more of these processes may also influence the frequency of the observed rearrangements. The causal analysis of the breakage process, as inferred from the visible rearrangements, is of necessity also very complicated and only its general lines are as yet more or less understood. In the following we will attempt to make a brief mention of the conclusions more important for our purpose of discussing the nature of genes, mutations and position effects.
Radiation induced breaks
The most detailed knowledge on the process of breakage has been obtained from radiation experiments (reviews in Timofeeff-Ressovsky 1937, 1940, Muller 1940, Zimmer 1943, Lea 1946, Serra 1949). We will refer first to the qualitative aspect of the effects of radiation on the cell nucleus.
Two orders of effects on the nucleus may be distinguished: matricial and chromonematic. The first includes all changes in the externam appearance of the chromosomes, especially stickiness, and in the nucleus as a whole. The chief change is the delay in division. These effects appear in the literature also under the name of physiological or primary effects. The other order of effects are the chromonematic, also called structural or secondary, which pertain to the production of breaks in the chromonemata and all their consequences. The designation of primary and secondary effects must be avoided because confusion about relations of causality between the two orders of effects may result.
The two kinds of effects, matricial or physiologic and structural or chromonematic, very probably have a great degree of independence, that is are not causally interlinked, though both may be correlated through the same agent inducing the two kinds of changes. Proofs in favour of this assumption of a great degree of independence are the following: 1. According to the data as yet obtained, chromonema breaks may be induced in any phase of the mitotic cycle or in resting stage, while for the matricial effects the existence of a sensitive phase has been demonstrated. The phase of greater sensitivity in what concerns the delay in division is located at the end of the resting stage or until the middle of prophase, according to the materials (Henshaw and colls., Carlson 1945, refs. and discussion in Lea 1946). If the irradiation is administered after the chromosomes attained an advanced prophase stage, when the matrix has alreay reached an high degree of condensation, as a rule the division is finished, while if the same dose is given during the sensitive period the division is stopped until eventually, after a shorter or longer time, recovery is attained. The delay in division is always accompanied of a greater or lesser degree of matrix stickiness. Irradiating advanced phases of division, usually only stickiness, not a delay, results. On the other hand, the chromonematic effects may also show an apparent period of greater sensitivity in anaphase, but this is a consequence of the fact that breaks induced in previous phases will reveal in anaphase, due to the movements of the chromosomes and to a mild peptisation of the peripheric nucleoproteins.
2. There is the possibility of the matricial effects being completely or almost completely compensated for if relatively low doses are used, the division after a recovery period showing no effect of a previous irradiation, except in some cases a low degree of stickiness. On the contrary, the breaks induced in the chromonemata do not recover, either restitute or become broken ends, which afterwards may heal or unite with other ends. The breakage of a chromonema is an all-or-none process, while matricial effects are gradual and capable of recovery.
3. Matricial effects, delay in division and stickiness may appear after very small doses of a dozen or so roentgens, given at the suitable sensitive phase, while for an apparent effect upon the chromonemata as a rule greater doses are required, of the order of a thousand roentgens. However, there is strong evidence that for a break only an ionising particle suffices, but restitution and union stongly influence the end result.
4. The most important distinction between the two orders of effects consists in the proportionality between dose and effect, which for not too great doses is always verified for the breaks while for the matricial effects, due to recovery and to the existence of a sensitive phase, the result obtained stongly depends upon the time of irradiation.
It seems, therefore, that the usual assumption that the two orders of effects are almost independent, is justified. This does not mean, of course, that complete independence is a fact. Matrix stickiness, if sufficiently stong, will induce the formation of matricial bridges in anaphase, which may cause chromonema breaks or even hinder the separation of sister chromotids. Usually, however, for the habitual doses of radiation the matrix bridges will have no consequences on the chromonemata, the bridges as a rule being broken at the right point between the two sister chromatids (Carlson 1941). On the other hand, it is possible that an intense stickiness process of the matrix may reach the chromonemata and cause a break before the anaphasic movements of the chromosomes, simply by the chemical changes which bring about the stickiness. The actual realization of this possibility, however, rests to be proved and seems improbable for mild degrees of stickiness.
It appears very proable that both the delay in division and the stickiness, which generally appear together, are caused by the same basic process which changes the deree of polymerisation of thymonucleic acid and of the matricial nucleo proteins. It is known that radiation depolymerizes thymonucleic acid (Greenstein and Jenrett 1941) and the surface stickiness of chromosomes may well be attributed to this depolymerization (Darlington 1942, Serra 1942). A delay in division could perhaps be due to a small degree of polymerization being caused by the enzymes, or the conditions necessary for such a polymerization having been transitorily damaged or impared by the radiation. It is also possible to admit an effect upon the enzymes controling the transformation of ribo- into thymonucleic acid (J. S. Mitchel 1942, cit. Lea 1946) but this seems unlikely since thymonucleic acid may be synthesized from phosphorus sources other than ribonucleic acid, as is demonstrated by radioisotope studies (refs. Hevesy 1948). The delay could also be due to a difference in permeability of the nuclear membrane, or even to a damage of the nucleolus (L. H. Gray and colls, cit. Lea 1946). So many alternatives are possible that it is not easy to prefer anyone and the only thing which seems assured is the final effect upon the matrix, probably by means of thymonucleic acid but also possibly by means of the histones. The computation of a sensitive volume of effects is as yet without a biological counterpart and therefore has a limited importance.
We will turn now to the chromonema effects, the breaks. The chief results concern the proportionality between dose and number of breaks, and the differential effects of several kinds of radiations and particles. Simple breaks which are revealed by the production of fragments or by chromosome or chromatid breakage were found in several organisms to be simply proportional to the dose of X rays, neutrons and α-particles (refs. in Lea 1946). When the rearrangements involve more than one break, the effect is proportional to 3/2 power of the dose, except for the «minute» rearrangements (involving a length of 1% or less of the chromonema) which are simply proportional to the dose. The 3/2 power of the dose is due to each non-minute rearrangement implying at least two breaks, and a part of them restituting or causing lethality.
The existence of proportionality between dose and the yield of minute rearrangements has been interpreted (Muller 1940, 1941) as signifying that this type of rearrangements is caused by a single ionisation, as should happen also with gene mutations. However, if appears that from the actually existing data the only legitimate conclusion is that the two breaks involved are caused by a single ionising particle (Lea 1946). The «effective» path of ionisation of the X rays usually employed in these experiments is of the order of 0.3µ, which is greater than 1% of the length of a Drosophila mitotic chromosome. Spiralisation is another means by which it can easily be explained that a single ionising particle may bring about two breaks located at a short distance (Buck 1939). The conclusion is that minute rearrangements do not seem to present any special problem which can not be explained by one or both of the two alternative explanations.
Another important point on chromonema breakage is the relative efficiency of several radiations and ionising particles. By considering the breaks primarily produced, that is the visible breaks plus those which restitute, Lea (1946) concludes that the relative efficiencies, as measured in Tradescantia for a same dose of r or v, are about 0.09 for X rays of 0.15-1.5 Å of wavelength, 0.21 for neutrons and 0.10 for αparticles. As the density of ionisation increases from X rays to neutrons to helium nuclei tracks, this finding has been interpreted as signifying that a densely ionising track of about 15-20 ionisations, produced in about 0.3 µ, is necessary to cause a break. α-particles ionise too densely and lose efficiency, while mean X rays ionise too sparsely and by this have also a lower efficiency than neutrons. We will discuss below the significance of this result for an understanding of the structure of the chromonema and the gene.
Chemical production of breaks
Breaks have been produced under the action of chemicals used to induce mutations. The only chemicals as yet used which have demonstrated to be capable of an efficiency comparable to that of radiations are the mustards (Auerbach and Robson 1944-47, Gilman and Philips 1946, Horowitz and cols. 1946, Gustaffson and McKey 1948, Tatum 1946, McElroy and cols. 1947, Darlington and Koller 1947). Other compounds which have been found to have a more or less marked mutagenic action are allylisothyocianate (Auerbach and Robson 1944), uretane (Oehlkers 1943), potassium thyocianate (Stubbe 1940), sulfamide (Chevais and Thomas 1943), methylxanthines (Fries and Kilman 1948), phenol (Haddorn and Niggle 1946), an phenanthrene carcinogenic compounds (Demerec 1949). As was to be expected, radioactive phosphorus compounds have also been found to be capable of provoking breaks (Arnason and colls 1948) but their effect must be due to γ and β emission and therefore fall under the heading of radiation effects.
From the diversity of the chemicals used with positive mutagenic action very probably all mutagenics may also have a breakage action it follows that the limiting factor in what concerns the action of chemicals upon the chromonemata is the existence of barriers against they reaching the chromonemata proper. This is also shown by the experiments of Haddorn and Niggli (1946) which obtained a mutagenic effect by treatment in vitro of gonads of Drosophila larvae with phenol, which is not a mutagenic agent (or is only a very weak one) if orally given. A chromonema presents the following barriers if a product is administered orally: 1. the intestine epithelium; 2. internal organs like the liver in Vertebrates, which «detoxicate» the organism by transformation or storage of the product; 3. the barrier of the passage from storage organs to the blood or lymph; 4. the passage from the body fluids to the gonads and within these to the germ cells proper; 5. within these cells, a barrier formed by the cytoplasm and the nuclear membrane and nucleoplasm; 6. the barrier of the chromosome matrix, which operates chiefly during the mitotic stages, but also during interkineses, since generally some matrix nucleoproteins remain on the chromonemata, and in their absence the pelicle formed at the interfact nucleoplasm-chromonemata must play a similar role; and finally, and probably much more important than all the other barriers together, 7. the protection of the peculiar metabolism of the chromonema itself, which probably has a very low desintegrative turnover rate (Serra 1947c) and in synthetic metabolism receives from the cytoplasm already relatively complicated polypeptides which form a kind of protecting zone and exclude other compounds from coming into contact with the chromonemata.
These barriers must have been acquired in early life forms, as without the stability of the composition of hereditary elements no true reproduction would be possible. In the last analysis, the stability of the hereditary units should be a result of the high complexity of its constituents, the proteins. According to this view, what is truly relevant for the mutagenic action of a chemical is the possibility of it reaching the chromonemata. Probably what is chiefly tested by the mutagenic experiments, besides the reactivity of the compound with cell constituents, is its «penetrability» and not properly any special «mutagenic property»
To judge from the action of the mustards, it seems inescapable to conclude that the action of chemicals upon the chromonemata is not specific. Mustards form in polar solvents a stongly reactive ethylenonion, capable of alkylating α-amino and α-imino groups of aminoacids and polypeptides, as well as phosphate, sulfhydril, phenol and other groups of several important compounds existing in the cells (Gilman and Philips 1946). With a so varied and strong action they disorganize the cytoplasm and penetrate the barriers, attaining the chromonemata. Of course, if another compound manages to penetrate the cell and reaches the chromonemata but is little reactive, its mutagenic action will be less marked than that of strongly reactive compounds such as the mustards. However, it is to be expected that, to a certain extent, reactivity and penetrability will be correlated, except in the case of liposoluble compounds, whose greater penetrabilty in the cells is well known.
The most important conclusion to draw from these studies of chemical mutagens and chromonema breakage agents, in what concerns the nature of the gene, is that widely different compounds exert similar actions, which, therefore, must not depend upon a specific reaction. The action of the radiations must also, in the last analysis, be similar to that of the mutagenic compounds. The energy necessary for a chromonema break finally has the function of breaking chemical bonds existing between neighboring pieces or units of the chromonema.
Chromonema breaks also occur spontaneously. Possibly, the chief cause of spontaneous breakage is chromatin agglutination (Resende 1947) by the mechanism of matrix bridge formation. It is also possible to suppose, however, that sometimes a sufficient amount of energy is liberated by exoenergetic reactions at the immediate vicinity of the chromonema, and thermal oscilation may also be operative in causing chromonema breaks. As in the case of mutations, it seems that cosmic rays will not be of a marked importance in causing «spontaneous» breaks.
All the agents which give rise to chromosome breaks are also capable of causing alterations in the hereditary materials supposed to be gene mutations. In view of the similarity in the production of both types of chanes, we will discus here only the chief differences in what concerns the qualitative and quantitative aspects.
As yet no specific means of producing a given mutation has been found. Radiations, mustards and other mutagenic chemicals, thermal shocks and also the processes which give rise to spontaneous mutations produce unpredictable results in what concerns the quality of mutation, that is the locus chaned. The only way as yet found to give a relatively high amount of mutations in a certain chromosome region is to provide a mechanism of rupture of that region (for instance in the case of maize McClintock 1941-1944). A difference in the relative proportions of certain types of mutations in barley between X rays and neutrons has also been described (Gustaffson and McKey 1948), but this probably results from the fact that certain types of rearrangements, especially those involving several breaks, may be more frequent when neutrons are used, due to the denser ionisation caused by protons compared with photo- and comptonelectrons.
In view of the lack of decisive results on the qualitative aspects of gene mutation, only the quantitative ones give indications about the nature of the mutative process. For the study of this aspect of mutations the chief methods used are the ClB and the attached-XX for the determination of sex linked recessive lethals, and more rarely also the second for visible mutations. In discussing the significance of the results, the particularities of the methods employed must be born in mind, although the parallelism found between visible and lethal mutations in what concerns their frequency in relation to radiation dose is in favour of the conclusions obtained with the ClB and attached-XX methods having a general validity.
As for the chromonema breaks, mutations have been found to be proportional to radiation dose; in contrast, however, the efficiency of several radiations in producing gene mutations is about the same for γ and X rays of several wavelengths (except very soft X rays, which are somewhat less efficient) but seems to be lower, with about a 60-80% efficiency, for neutrons (Zimmer and Timofeeff-Ressovsky 1938, Kaurmann 1941, Fano 1943, 1944, Giles 1943). If this result could be taken at its face value, it would mean that each mutation is due to a single ionisation, neutrons being less efficient by virtue of producing too dense ionisations, and this is indeed the general conclusion found in the literature, (see Timofeeff-Ressovsky 1937, Muller 1941 and Lea 1946 for discussions of previous results).
Gene mutations should therefore be different from chromosome breaks, since these latter necessitate of several ionizations to appear. However, this conclusion must be carefully weighted, in view of important incertitudes upon which it is based. Besides some hesitations regarding dosage equivalencies, these incertitudes concern chiefly the methods by which the «mutations» are detected. In reality, the recessive lethals which generally are used to compare with the dose include a great proportion of chromosome rearrangements, among them small deficiencies and position effects. In a case it has been found that about 57% of the recessive lethals were in fact rearrangement effects, including small duplications (Lea and Catcheside 1945). Despite this high proportion of detected rearrangements (the smallest, undetected, should yet have increased the proportion) the overall result of a proportionality between dose and effect has not been visibly impared, which demonstrates the lack of sensitivity of these methods and therefore stongly limits the significance of the conclusions based on them. It could be admitted that the ionisation caused the lethal at the same time that induced the break, that is the lethals should be a consequence of the breaks themselves (Lea 1946). However, this interpretation seems unlikely, in view of the numerous cases now known in which reversion of chromosome rearrangements, associated with recessive lethal effects, to the original structure, caused also disappearance of the lethals (see also Fano 1947). We can be sure that the great majority of the lethal effects associated with rearrangements other than deficiencies are in fact due to position effects, as was discussed in Chapters 3. and 4.
On the other hand, as says also Lea (1946), the calculation of a mean number of about 17 ionisations for a break is valid for the case of Tradescantia chromosomes, but in other cases a different number may be necessary. In this respect, the diameter of the chromosome may be of importance, thin chromosomes necessitating less energy for a break (Giles 1943). Other conditions may yet be more important, particularly the state and the amount of the matrix, the metabolism of the cell, and so on. These conditions may modify not only reunion as also the breakage itself, by buffering or not the effects of the ionisations.
At a time it was supposed that a distinction between the causality of gene mutations and rearrangements could be secured by experiments with ultraviolet irradiation. This hope has not been fulfilled. Ultraviolet ordinarily causes principally «gene mutations», while rearrangements rarely appear. This, however, may be due to the energy being first absorbed by nucleoproteins or simple proteins and then transferred from these to the chromonemata, for a mutation to be produced. Only in small localized spots should an amount of energy sufficient for breakage be concentrated by absorption, and the rearrangements would be, at most, of the order of these spots. The chromosome as a whole would not be broken, because the ultraviolet acts upon the matrix, causing an incipient stickiness which hinders the breakage of the chromosome, though not of the chromonema. Facts in favour of this interpretation are: first, the action spectrum of the ultraviolet, which is similar to the absorption spectrum of nucleoproteins (or of simple proteins, in some cases), and second the finding that ultraviolet administered after X-ray irradiation reduces the frequency of breakage frequency in Drosophila (Kaufmann 1946). However, the effects of the ultraviolet are as yet little understood and their interpretation is only tentative.
As a general conclusion, it may be said that no sharp distinction, in what concerns their causality, is actually possible between gene mutations and chromosome breaks, though the former seem to require a lower amount of energy than the second. This apparent difference may be caused by the materials tested rather than the processes involved: when chromosomes provided with a matrix, or spermatozoa, also with a very condensated matrix (the whole head is a kind of general matrix in which the chromonemata are embedded), are irradiated or treated with mutagenic agents such as the mustards, a greater energy is necessary, while in stages where the chromonemata are with little matrix a lower amount of energy will be needed. Despite these incertitudes, an important conclusion may already be drawn, namely that both the production of mutations and of breaks finally consist in a physico-chemical process in which chemical bonds are broken; this conclusion is independent of any assumption regarding the nature of the bonds which must be broken. Gene mutations and chromosome breaks alike, have in common at least this property of necessitating the breakage of some chemical bonds. From a consideration of the energy necessary for this bond breakage, some further conclusions are possible concerning the nature of the process involved; this is attempted below, together with an interpretation of chemical data on chromonema composition.
Chromosomes are composed of chromonemata and matrix. Matrix is all that is desintegrated in telophase and more profoundly during the great growth period of the oocytes, when only the pure or almost pure chromonemata remain (see Serra 1947c). In ordinary chromosomes probably about 10-90%, according to the phase of mitosis, is of matrix and even in interphase nuclei and salivaries about 10% must be the minimum amount of matrix (in salivaries only in the bands). The matris is chiefly composed of nucleoproteins, formed of thymonucleic acid and histones (or protamines, in certain cases). Other proteins and ribonucleic acid probably exist in small amounts, chiefly as components of the nucleoplasm, which is included in the matrix gel. The nucleoplasm is a sol of proteins, lipids and some ribonucleic acid (sometimes also thymonucleotides) and nucleoli have proteins, basic and non-basic, ribonucleic acid and lipids (discussion in Serra 1947c).
However important the questions concerning the composition of the matrix may be for an understanding of the causality of mutations and chromosome breakage, for the nature of the gene what interests more is the composition of the chromonemata. Unhappily, very little is known about this. It is certain that the pure chromonema, where are the genes, is composed of proteins which wither are non-basic or have a basic character less marked than the histones. If the genes are nucleoproteins, the chromonemata must also have nucleic acid, but this is uncertain and possibly it will be verified that in reality the chromonemata in their purest state, that is when totally deprived of matrix, do not possess nucleic acid or possess it only in a very small amount.
Are the genes nucleoproteins?
As this question of the existence of the nucleoproteins in chromonemata and genes is very important, we will discuss it in more detail. It has been assumed that the gene has a nucleoprotein nature. This is based primarily on the fact that in their conspicuous state the chromosomes owe their staining properties to the thymonucleoproteins they have. In the second place the circumstance that cristallysable viruses and cytoplasmic granula have ribonucleic acid has by analogy induced many genecists and cytologists to admit that genes also are nucleoprotein in nature.
The experimental facts, however, may have a different meaning. In effect, by a careful Feulgen test no thymonucleic acid is discovered in certain oocyte chromonemata, and colorations by basic stains, which usually reveal the presence of ribonucleic acid, also fail to demonstrate the existence of acidic compounds in these compounds (at least without thymonucleic acid) may be explained simply by admitting that this is due to the presence of matrix in the bands.
The case of the viruses may also be similarly explained. The isolated part of the virus-cell complex, in such viruses as the tabacco mosaic and other simple viruses, in reality is only the minimal unit which, introduced into a host cell induces that other similar units be formed by the cell. These units behave rather like trypsin molecules when in presence of trypsinogen, inducing rapidly the formation of more of their kind. Under another aspect the viruses also recall the cytoplasmic granula capable of self-reproduction, and very probably represent in their majority the end-point of parasitic evolution by simplification.
The interpretation of the property of virus self-duplication has recently been challenged (Cohen 1947) chiefly on the basis of the simpler viruses do not possessing enzymes and deriving their nucleic acid by a seemingly non-autocatalytic process (see also discussion of Cohen's paper by Spiegelman); instead of self-duplication, the virus would induce only duplication of its kind by the cell. Of course, this is the last term of self-duplication, an end-point in which not only the nutrients as also the enzymes are supplied by the host; the dilemma between self-duplication and simple duplication is in this case similar to that between the viruses being living or non-living.
More important for the case of the nature of the gene is the composition of the viruses. It seems that at least a small amount of nucleic acid accompanies the virus proteins, although some analyses indicate that certain viruses (a colliphage, according to Kalamanson and Bronfenbrenner, cit. Hoagland 1943) may be simple proteins. Supposing that indeed the viruses all possess at least come nucleic acid, the very small amount found in certain of them (for example 1.1% in the Newcastle disease, and 5% in the influenza, viruses see Knight 1947) in relation to the quantity of protein, is highly suggestive of the nucleic acid, either of the ribose or of the desoxyribose type or both, being a kind of structural contaminant, that is a compound essential for the host cell structures in contact with which the viruses are reduplicated (microgranula of the cytoplasm, or possibly other structures) and which pass to the virus protein during the process of reduplication. If this interpretation is true, it would be possible to free the virus from nucleic acid, without affecting its power of inducing a reduplication of its units in the host cell. In practice, this has been found impossible or very difficult, owing to the nucleotides being firmly bound to the proteins, very probably according to a certain pattern along the polypeptide chains. However, the separation may perhaps result with some viruses. An indication about this point may be obtained by studying if there is correlation between the composition of the viruses in what concerns their nucleic acids and their site of origin within the cells: viruses wholy or partially originating within the nucleus of the host cell could contain desoxyribonucleic acid, while those synthesized in the cytoplasm should not have this acid. Of course, this refers only to the simplest viruses, not to those of higher organization.
The only conclusion at present valid is that the viruses have nucleic acid, but the host cell has also apparently the same nucleic acid, and therefore only the protein part of the molecule is truly specific. The actually existing data do not favour the interpretations about the specificity of the nucleic acids found in viruses; it seems that virus nucleic acids are similar to the cell nucleic acids (STANLEY and KNIGHT 1941, KNIGHT 1947). This is another proof that the nucleic acids probably are connected with the synthesis of the viruses but not with their specificity.
Therefore, even if we admitted that the viruses constitute a valid analogy for the genes—which is very discutable and would necessitate of being carefully weighted—it could not he concluded that the nucleic acids are inconditionally necessary in order that a protein may induce duplication of its kind in the cell. In fact, nucleic acids may have a structural role and serve also as energy activating metabolites, but do not seem to confer specificity to the duplicating unit.
The facts about the composition of oocyte chromosomes and the salivaries are rather in favour of the genes being simply protein in nature. The data on this point, however, are as yet meagre but the fact that a complete disappearance of the Feulgen test and of the affinity for basic dyes is obtained naturally in oocytes and artificially by means of nucleases in salivaries, while at the same time the protein skeleton of the chromosomes remains intact (SERRA 1944-1947 must be considered as very important for a decision on this point On the other hand, the numerous data on the importance of nucleic acids for the growth and multiplication of the cells, obtained since the modern precursor work of BRACHET (1933-1947) and of Caspersson, Hyden and colls. (CASPERSON 1936, 1941, HYDÉN 1943, THORELL 1947) prove that nucleic acids are important for the synthesis of proteins in the cell. The thymonucleoproteins of the chromosomes (peripheric or matrix nucleoproteins, according to our designation' are concerned chiefly with the individualization and the mechanics of the chromosomes in the cell, not with the reduplication of the basic structures of the chromonemata, which very probably occurs in the resting stage, not during mitosis (SERRA 1942-1947). It is possible, however, that a residual laver of nucleoproteins may play a role in the reduplication of the basic chromonemata structures.
In this respect, the results of Mirsky, Pollister and Ris (MIRSKY 1947, MIRSKY and POLLISTER 1943, MIRSKY and RIS 1947) are of interest. By a mild extraction with 1M NaCl, nucleohistones are easily extracted from resting stage nuclei plus some rare mitotic chromosomes, and the residue is found to be formed mainly of protein with about 10-15% of nucleic acid. While the nucleic acid of the fraction extracted with 1M NaCl is almost wholly desoxyribonucleic, that of the non-extractable residuum is almost wholly ribonucleic (MIRSKY and RIS 1947). However, in certain cases the amount of ribonucleic acid is very low (0.15% in salmon sperm and 3% in thymus gland chromosomes — MIRSKY 1947). The non-extractable residuum very probably corresponds to the chromonemata, plus a more or less great part of components from the nucleoplasm and rests of the matrix more firmly bound to the chromonemata.
The idea we get from the available cytological evidence and the results of these analyses is that the extracted «chromosomes» are composed of chromonemata to which is linked, or which include between the bundles of their skeleton structures, a part of the nucleoplasm and to which are also. linked matrix or peripheric nucleoproteins. Of these three fractions—pure chromonemata, nucleoplasm and peripheric nucleoproteins—only the first must have a stable composition, while the two other fractions are variable according to the physiology of the cell. For example, in sperms or in rapidly dividing tissues the fraction of peripheric nucleoproteins will be greater while the nucleoplasm fraction will be at a minimum, corresponding to chromosomes in a condensed state. On the contrary, in cells engaged in active cytoplasmic elaborations, such as those of glands in general, liver, kidney, etc., but not in mitosis, the nucleoplasm fraction will be the most conspicuous. On the other hand, the nucleoplasm fraction will be highly variable in composition, according to the particular organ in study. Since the nucleoplasm is the intermediate between chromonemata or resting stage «chromosomes» and the cytoplasm, its composition must reflect the nucleus-cytoplasm metabolic changes. At least, lipids, nucleic acid (chiefly ribose type) and proteins are to be found in it (see SERRA 1947c, Table I). The nature of the metabolic changes between nucleoplasm and chromonemata will also alter the composition of that part of the nucleoplasm which is extracted with the chromosomes by the methods of MIRSKY and coils. The nucleoplasm firmly bound will appear as being an integrant part of the chromosomes, and so a variable composition, according to the tissues, will be found.
The conclusion to be drawn from the cytological and chemical facts at present known is that, while the composition of the essential part of the chromonemata must be maintained practically unaltered, the chromosome as a whole has a composition variable according to the tissue and the physiology of the cell. The pure chromonemata are not to be found in ordinary glandular tissues and even in sperm heads there are also peripheric nucleoproteins and other non-essential components. The fact that sperms have only a very small amount of ribonucleic acid and that the extraction residuum of chromosomes of other tissues has also a very small quantity of desoxyribonucleic acid (MIRSKY 1947) is strongly in favour of concluding that pure chromonemata either do not have nucleic acids or only possess a small amount of them at some points in a quantity completely insufficient to account for the genes which must form about 50% of the pure chromonema. That is, if nucleotides are present in the chromonemata they are a kind of prostetic groups which exist only at certain points not necessarily at every gene.
This conclusion plainly concords with the interpretation of the cytochemical tests (see SERRA 1947c) and it is the convergence of both the quantitative analyses and the chemocytological reactions which gives weight to the interpretation we present. The existing data, though not furnishing a final demonstration, are strongly in favour of the chromonema, which contains the genes, being not nucleoproteic, but simpy proteic, in nature. On the other hand, in certain stages namely when haptogenes are being synthesized and when the genes are being reduplicated, nucleoproteins chiefly ribonucleic are more or less firmly bound to the chromonemata (constituting the fraction we have called nucleoplasmic) and during division an overcharge of thymonucleic histones comes also to cover the chromonemata. The two apparently contradictory conclusions, that the genes are simple proteins as components of the chromonema and nucleoproteins when working, are resolved by a synthesis at an higher, dynamic, level: the genes are simple proteins as components of the) chromonema in its simplest or structural composition, and are nucleoproteins when they function as phenogenic units or reduplicating units.
The proteins of the chromosomes
Since the pioneer work of MIESCHER, KOSSELL, STEUDEL and others (see KOSSELL 1928) at the end of the last century and the beginning of the actual, only recently has the question of the proteins of the cell nucleus been again approached with modern methods. Two chief lines of attack have been 1 followed: isolation of compounds from whole nuclei, and isolation of bodies similar to chromosomes obtained from more thoroughly broken cells. Both approaches are indeed necessary for a final understanding of the composition of chromosomes, since, as is implied in what we said above, the true composition of any part of the cell actively engaged in metabolic changes may be obtained with analytical methods only by studying it under different conditions and by trying to study concomitantly other cell parts with which it intimately integrates in a metabolic system. The results of both lines of attack are not yet in a state which allows an integration in a comprehensive picture, even if only approximate, of the dynamic composition of the cell, nucleus. Hence, in dealing with the important question of the proteins of the chromosomes, only the results of the line of attack which tries to work directly with the chromosomes may be considered at present.
Two kinds of proteins, which by their names or by explicit statements, were thought to be specific of the chromosomes, have been isolated from cell nuclei: chromosomin (STEDMAN and STEDMAN 1943, 1947) and chromosins (MIRSKY and POLLISTER 1943). The second designation, chromosins, has now been dropped (see MIRSKY 1947, discussion) as its authors realized that indeed they were dealing with nucleohistone preparations which only superficially may bear a relation with the specific chromosome proteins. The designation of the other protein, chromosomin, has been continued, apparently implying that the nucleus, after lipid extraction, besides nucleic acid plus histone (or protamine) contains almost nothing else than chromosomin (STEDMAN and STEDMAN 1947a). However, the basis for such a statement seems, at least as yet, not to be sufficiently grounded, since the amount of chromosomin is obtained by difference and it has not been completely demonstrated that chromosomin is a definite protein and not a more or less complex mixture. The solubility in dilute alcalies is common to many compounds.
We have already discussed elsewhere (SERRA 1947c) the qualitative composition and the approximative amounts which may be expected to occur in the several parts of which a cell nucleus is formed, so it is not necessary to treat here of this question again in detail. Also, we have referred above to the conclusion that «chromosomes» isolated from resting stage nuclei are complex formations, including a greater or lesser part of metabolites which necessarily come to them through the nucleoplasm, and therefore may be considered to belong both to the chromosomes and to the nucleoplasm. Besides, ordinary resting stage nuclei have also nucleoli and free nucleoplasm (that is nucleoplasm not included within, or linked to, the chromonemata) and a nuclear membrane is also present. This complexity must be born in mind by the biochemist who tries to draw conclusions regarding the localization of the compounds or mixtures he extracts from a certain nuclear formation.
Even the most simple material, sperm heads, has also some other formations besides the nucleus, namely a little cytoplasm, the acrosome and in many cases an intermediary piece. The nucleus is formed of chromonemata, peripheric nucleoproteins (matrix) and some rests of nucleoplasm, and possibly a part of the nucleolus in certain cases, plus a membrane. The chroruonemata must form only a very small part of the sperm's head, the chief constituent in volume being the matrix, composed chiefly of nucleohistone, or in certain fishes nucleoprotamine. If the other constituents are neglected, and ordinarily this may be done, the product isolated will be chiefly the matrix nucleoproteins and the chromonemata proper will pass almost unnoticed. This is the ground for earlier claims that histone is the protein of the genes. The reality, however, is much more complicated; from the evidence of chemocytological tests, it is concluded that the proteins of pure chronionemata (that is chromonemata at their minimal complication, without, or with the least amount, of matrix, and without nucleoplasm bound to them) are not markedly basic.
This is the only conclusion sufficiently grounded on direct analytical facts. However, from what is known concerning the great diversity of genie action and the complexities of the enzymatic systems of the cell, some extrapolations may be attempted. It is to be expected that the pure chromonemata shall consist, not of a definite protein but of several proteins or polypeptides, corresponding to the diversity of genic action. Biochemists who succeed in isolating nuclear proteins must not be hasty in giving it a unique name, which may mislead to the conclusion that they are dealing with only one definite kind of molecule. These oversimplifications habitually only lead in the end to a frustration of what were perhaps exagerated hopes, and to a dislike of the subject by subsequent workers of the same field.
On the other hand, to the chromonemata of the resting stage nuclei, nucleoplasm (with the meaning we gave above) and matrix components are linked. We propose, therefore, that in future the proteins to be isolated from nuclei must preferably be classified in at least four groups: proteins from the free nucleoplasm (nucleoplasmins); matrix proteins (matricins or chromatins); enchylematic proteins, from the nucleoplasm linked to the chromonemata (enchylemins); and the chromonematic proteins proper (genonemoproteins or tenoproteins). A possible 5.th group would be joined when it would be possible to isolate the nucleolus. In proposing this scheme we have no doubt about the difficulty of reaching a state in the preparatory methods enabling a close localization of a given compound within the nucleus; it is even possible that the scheme be impracticable, but nevertheless it serves to show the complexity of the subject.
The physical state of the chromonema proteins
According to their state, proteins may be divided into globular and fibrous, these latter being formed by parallel bundles of polypeptide chains disposed in fibres, while the former form particles of a definite size and globular shape. Except silk fibroin, all the other protein fibres in their natural state have their backbone polypeptide chains more or less folded according to definite, though as yet not fully understood, patterns which give rise to the regular period found in. their X ray diagrams. The folding of the polypeptide chains of the globular proteins probably is also specific for each kind, otherwise it would be difficult to explain the formation of characteristic crystals. Between the two classes, fibrous and globular, stand the viruses like the tobacco mosaic which, in a not too diluted solution, easily form «liquid crystals» or crystalline bundles of their long molecules.
The state of the chromonema proteins seems also to be intermediate between fibrous and globular. However, this is very different from the case of the long virus particles associating in solution to form bundles, and the intermediate state results here not from the properties of the individual particles but from the organization of the chromosome, which obviously never is in a state of solution. The chromosome may become optically invisible, for instance in salivary glands, and reappear following changes in the cell fluid, according to there is present or not calcium or other gelifying ions. It must be assumed that also in ordinary resting nuclei the chromonemata persist always individualized as protein bundles, but can become extraordinarily swollen by inclusion of more or less hydrated nucleoplasm. The end point of this process is the filling of the entire nucleus (excepting the nucleolus, of course) by the chromonemata swollen to a kind of very diluted gel or almost a sol, whose liquid phase is almost continuous. Also in this property, the nucleus must differ from ordinary gels by the chromonemata do not forming cross linkages or net-point bonds at random, but only between neighboring homologous points within a same chromosome, while a distinction among different chromosomes must always be maintained. This is a postulate which results from the genetic individuality of the gene materials and their identic arrangement on the chromonema from generation to generation. Work on the colloidal state of several types of resting nuclei, which is as yet lacking, could substantiate this claim.
While the state of the proteins during the synthetically active period of resting stage is only inferred from very indirect evidence, the state of the proteins of the visible chromosomes have been a little more studied, although the results are actually poor. The chromosomes are always very extensible, not only the metaphasic (or prophasic-anaphasic) chromosomes which have contracted from much longer distended chromonemata, as also the salivary, and oocyte chromosomes which are already in a state of almost the maximum naturaL distension. Frog oocyte «lampbrush» chromosomes have been stretched to a length of 514µ representing an increase of 450% length before breakage, and the deformation caused by an elongation to about 125% was not permanent, the chromosomes returning to their primitive aspect even after repeating many times this amount of stretching (DURYEE 1937). Salivary chromosomes of Chironomus may also be stretched to several times their length without rupture, an elongation to 3-4 times the primitive length being generally withstood (PFEIFFER 1941, BUCK 1942). These elongations cause permanent deformation; the chromosome resting afterwards visibly longer, while a stretching of about 125% is possible without any visible deformation (the first figure according to PFEIFFER, the second after BUCK; the methods followed and the pretreatments of the chromosomes may, of course, markedly influence the results).
The great extensiveness of the chromosomes cannot be accounted for by the visible helicoidal coiling or so called spiralisation. In effect, while an explanation of the elasticity of ordinary mitotic or meiotic chromosomes could yet be attempted on such a basis (though this in the end probably would not give a sufficient interpretation), salivary and oocyte chromosomes show only rests of a slight spiralisation, which could not explain their great extensiveness and elasticity. Therefore, another interpretation is necessary and this can not be other than the folding of the protein fibrils at a sub-microscopic, including the molecular, level. The data on this sub-microscopic foldings are yet only indirect, obtained from experiments on the stretching of the chromosomes and from observations in polarized light; in this respect, electron microscope photography has given unclear pictures, whose interpretation presents many difficulties.
The nucleic acid chains of the sperm head are to a great extent oriented according to the length of the head (SCHMIDT 1937, 1941) but in salivary bands only a minor fraction of the nucleic acid may be oriented parallel to the chromosome axis (CASPERSSON 1940); the degree of orientation of the nucleic acid particles of the salivary bands may be increased by dehydrating treatments. These results, however, have no bearing on the composition of the chromonemata, since the nucleic acid of the bands is either totally or for the most part a matrix product (with the sense we give to this term, synonymous of peripheric nucleoproteins). That really the thymonucleoproteins of the bands play the role of a deposit on the chromonemata is shown by the fact that their remotion leaves a continuous skeleton of protein, and the pictures obtained with the electron microscope support this view. The bands have a density of structure so great in relation to the interbands that they appear as completely opaque to the electron beam (see for instance the photographs in PEASE and BAKER 1949, and SCHULTZ, DUFFEE and ANDERSON 1949). The most probable interpretation of this fact is that the bands have approximately the same amount of material as the interbands plus a charge of basic thymonucleoproteins, which is the interpretation given on fig. 3e of SERRA 1947c.
|* Note added in proof: SCHULTZ, MacDUFEE and ANDERSON (1949) have also studied salivary chromosomes with the electron microscope but do not claim to have seen the genes. Globular bodies were observed to compose the bands, of a size more or less correspondent to that found by, earlier authors. It seems that the careful attitude of SCHULTZ et al., of drawing no far reaching conclusions about the band structures and the genes, is the only possible until more details are known about these chromosomes.|
The results of BUCHHOLZ (1947) on pachytene chromosomes of maize accord with this interpretation. Spheroidal particles of from 85 to 660µ of diameter are seen embeded in, or attached to, a continuous, much less dense, chromonema thread; the author rightly interprets these particles as being rather gene-products than the gene-elements themselves. Probably the particles are of matrix nucleoproteins in process of «covering» all the chromonema and which finally give the continuous matrix seen at the end of pachytene and afterwards. These particles are not chromomeres in the sense given to this term by the cytologist working with the ordinary microscope, since the visible chromomeres are of the order of 1µ or more. Particles with about this size are seen in the photographs of BUCHHOLZ (especially in phot. 5) and the figures suggest that the thread is folded in one or more loops of about the size of the chromomere, at a microscopic or only slightly submicroscopic level; on and around the thread a lot of nucleoprotein particles of about the size found in the interchromomeres (that is about 100 mµ) are observed.
The more important part of the structure of the chromosome, that of the chromonema, continues yet to be studied in the future; this will be possible only by observing chromonemata after remotion, as far as possible, of the chromatin matrix, and especially by selecting materials where this is easy to perform or is already naturally done. Although the observation of details will be difficult due to lack of contrast caused by the existence of an almost uniform density, the observation of the submicroscopic structure of the chromonema will be very important. The particles associated with the chromonemata must be interpreted in the light of what is known about the components of the chromosomes in resting stage and mitosis, and may be of the nucleoplasm or of the matrix. It is clear that the gene, as a unitary particle, will not be found directly by electron microscopy of pure chromonemata; what can perhaps be observed, is the minimal chromonema unit or nemamere.
When it is attempted to summarise these data on the physical state of the chromonema proteins, it is apparent that they only allow the drawing of some rather modest conclusions. Fortunately, however, a few more extrapolations of considerable importance are possible from genet ico-cytological facts. The chromosome as a whole is a kind of fibre of a special kind, in which the proteins are disposed as if they were semi-globular and semi-fibrous units, in a sequence as that of the genes. The proteins are folded in a visible helix (spiralisation) and more or less similar foldings exist" also at sub-microscopic levels, down to the molecular looping which must be similar to the folding of globular proteins, rather than to the more or less straight disposition of the polypeptide chains in fibrous proteins. This is concluded from the great extensiveness of the mitotic, salivary and so called lampbrush chromosomes.
The polypeptide bundles of the chromonemata must not be continuous from one end to the other of the chromonema, the polypeptide chains are not uninterrupted. In protein fibres also, some kind of interruption probably exists, corresponding to the greater periods found in X-rays diagrams, but the case of the chromosomes certainly is different. The essential unit of the chromonema is the nemamere (see page 418) the minimal segment which can be deleted or otherwise rearranged; at an higher level there is also the cross-over unit, which is no more than an assembly of nemameres forming an elementary chromomere so protected in the pachytene, in part by matrix nucleoproteins, that cross-over breaks are very unlikely to occur within it. If the chromosome was a true fibre, with only the long periods shown by protein fibres and which apparently do not give rise to breakage zones, a great amount of energy would be required to break a chromonema, much more than probably is available at a small spot in crossing-over and in spontaneous and experimental breakage.
As this point is important, we will develop it a little more A rough calculation of the number of polypeptide chains of each mitotic or meiotic chromonema (SERRA 1947, 1949) is possible from a consideration of the thickness of the thread and the mean distance of 9.8 Å found between two chains in X-rays data of protein fibres (ASTBURY 1939, 1942). A thread just at the limit of ordinary microscope visibility, with a thickness of 0.1-0.2µ, formed of parallel chains, will be composed of about 104-4·104 chains. Due to the hydration of the proteins and to the chains being looped, these numbers may reduce to 1/2 - 1/4 if it is supposed that the proteins form bundles or groupes more or less separated by hydration layers, the number of chains may perhaps be only 103, but finally, even the most conservative estimate will not go below this level. It can be said, therefore, that without any doubt the most thin chromonema is formed, according to its diameter, of about 104-103 polypeptide chains. Therefore, the chromonema never is a single polypeptide chain as, it seems, some implicitly admit, but on the contrary, is a kind of statistical aggregate of elementary single chains. This conclusion is important and from it several corollaries derive, as will be discussed in the following.
First of all, if the polypeptide chains were uninterrupted throughout the whole chromosome, the energy necessary to break a chromonema should be about 104 times that for a peptide linkage. As the CO-NH bond corresponds to an energy of 48.6 kcal/ mole (PAULING 1940), it follows that in order to break a chromonema about 486,000 kcal/mole would be necessary. It seems highly improbable that such an amount of energy is available at every cross-over point in the perfectly physiological conditions of normal pachytene, and even in experimental produced breaks this amount would not be easily obtained in a small spot. Let us discuss the case of the radiation-produced breaks. For mean wavelength X-rays the energy corresponding to one ion pair (in the air) is 32.3 eV or, at the equivalence of 23.05 kcal/mole = 1 eV/mole, about 749 kcal/mole. From the results of LEA and CATCHESIDE (see LEA 1946) it is known that in Tradescantia in mean about 17 ionisations are necessary to break a chromonema. This corresponds to an amount of energy of 12,733 kcal/mole, which is much less than the 486,000 kcal/mole necessary to break the 104 polypeptide chains of a chromonema. Of course, if instead of 17 we admitted that a single ionisation is capable of furnishing the energy necessary for a break, the remainder. of the energy being absorbed in the matrix and the immediate neighborhood of the chromonema, it would yet be more difficult to imagine how a smaller amount of energy would. suffice to break a chromonema.
An way to escape this conclusion would be to suppose that the chromonema has only about 10 or at most 102 polypeptide chains, but this seems highly improbable. The necessary conclusion, therefore, is that the high amount of energy which would be necessary to break the peptide linkages of a chromonema excludes that this is composed of uninterrupted polypeptide chains. Another, qualitative, argument (MULLER 1941) reinforces this conclusion. If the breakage process opened peptide bonds, COO and NH3+ groups would result, which could unite only to re-form the primitive links if the opposite groups were available, that is COO uniting with NH3+ groups. Therefore, reunion of broken ends should not be a random process, which is contrary to observation: unless a healing process intervenes, it seems that the breaks reunite at random, limited only by space and time relations. To escape this argument it could be supposed that the «surface» of breakage of the composite chromonema, with its 104 chains, could present both COO- and NH3+ groups which would attract the opposite groups and would combine, if a small torsion of each chain is possible. It is to be expected, however, that in mean not all of the spatial relations could be satisfied and a fraction of the COO- and NH3+ groups would not reconstitute the peptide linkage. A weak spot with a tendency to rupture would thus result.. No such spots of preferential rupture seem normally to occur and therefore the assumption is improbable. The final conclusion, based chiefly on the quantitative argument, is that the chromonema is composed of sub-units which are not linked to the others by peptide linkages and therefore the peptide chains do not run uninterruptely from one end to the other of the chromonema. This conclusion is the only which also accords well with the genetical facts.
Details of the state of the proteins in each chronionema minimal unit or nemamere are lacking. The lower extensiveness of the salivary bands as compared to that of the interbands has been interpreted as signifying that in the bands, which would correspond to the genes, there are polypeptide chains in a state approaching that of 13-keratin, with a minimal amount of looping (discussion in SERRA 1942) and this admission has been made the basis of a model of chromosome with distended polypeptide chains in the genetically active zones of the chromonema (SERRA 1942, 1944). If it is admitted that the polypeptide chains of the reduplicating genes are synthesized in contact with those already existing in the chromonema, it would be difficult to visualize this synthesis if the polypeptide chains were folded in complicated patterns. On the contrary, distended chains would constitute an obviously simpler templet-see below, Fig. 5.
The hypothesis of the genetically active zones having distended polypeptide chains explains the lower extensiveness of the salivary bands and seems to give a simpler model of the reduplicating unit. However, it is only a possible hypothesis, not a necessary one. In fact, the lower extensiveness of the salivary bands can also be explained as being due, at least in part, to the matrix nucleoproteins hindering the stretching, and the difficulties about reduplication of the chromonema chains may be overcome by admitting that already relatively long blocks of polypeptide chains reach the chromonema through the nucleoplasm and their linking in a specific order is accomplished at certain points of the chromonema. It is only in these small spots that the chains are distended. Probably it is also at these points that a small amount of nucleic acid of the ribose type exists to serve as a kind of negative or templet for the basic groups of the polypeptide chains. Of course, the amount of nucleic acid would increase during gene reduplication and production of haptogene, as said above.
A decision between the two alternatives, of almost wholly distended and only partially distended polypeptide chains in the genetically active loci or chromomeres, may be obtained only by the study of details of lampbrush and salivary chromosome structure to which the matrix is removed. Actually, the second hypothesis, of the polypeptide chains being distended only at certain points of the chromonema and looped in the rest, appears as the most probable. The chromonema model we developed before (SERRA 1942, 1944) must be fied to fit this second alternative.
Another point to modify in our former scheme concerns the existence of an almost equal amount of alternating protein and nucleic acid (thymonucleic) chains. In the preceding discussion we have concluded that the actually existing data rather favour the hypothesis of the existence of only a small amount of ribose nucleotide chains at certain points of the chromonema; this amount increases when the chromonema structures reduplicate. Afterwards histone thymonucleoproteins deposit on certain points, the chromomeres, andthis achieves the differentiation of elementary chromomeres and gene individualization. An overcharge of hystone thymonucleoproteins is also necessary for chromosome division.
After the preceding detailed discussion of the composition of the chromonema it is now easy to deal with the composition of the gene in a brief way, as the chief properties of the' genes may be inferred from those of the chromonernata. The relevant genetic and cytological facts about the gene have been reviewed in detail above and therefore we are excused; of dealing again with them. The subjects of gene structure and gene mutation are so intimately related that forcedly, some points on mutation must be discussed also in this section, though a special heading will be devoted to mutation, below.
The gene as a physico-chemical unit
One of the most notorious trends in discussions about the nature of the gene and of gene mutation is to consider the gene as a well defined particle, like, a molecule or even an, atom. Calculations of the size of the gene give values down to some mµ; for instance, the most thin salivary bands obtained by digestions with trypsin and precipitants of nucleic acid have a thickness of 0.2-0.1 µ. The dimensions obtained for the target diameter in radiation experiments are of the order of 5 mµ and calculations from the position effect give a value of 3-5 mµ (see page 434). These dimensions would compare to those of viruses and protein molecules. For instance, an egg albumin molecule has a size of 9×3 mµ and an hemoglobine molecule is 15×3 mµ. We have criticized above (chapter 3, page 435) some of the estimates of gene size and the conclusion is that in reality calculations which do not have in due regard the spiralization of the chromosome at the microscopic level and the looping, of the chromonema at the submicroscopic level are not demonstrative.
Calculations based upon radiation effects also are not conclusive in what concerns the size of the gene, although they may be of importance for a comprehension of the processes by which radiations act upon living matter. Similar calculations have given in the case of virus inactivation sensitive volumes which agree with the known size of the viruses (see discussion in LEA 1946) and this success has strengthened the view that the same methods should also give a valid estimate of gene size. However, this may not be true and indeed there are good reasons to believe that it is not. Besides the argument mentioned by MULLER (1941a) there are those we referred to above when dealing with the interpretation of radiation effects, namely that great doubt exists about the nature of the cytogenetic changes upon which the calculations are based being really gene mutations and not minute rearrangements. The most probable conclusion is that the «gene size» obtained from radiation experiments, if any, has only the meaning of a sensitive volume within which a single ionisation (or perhaps a small group of ionisations) causes an happening with a cytologic or genetic consequence.
Two chief possibilities about the true nature of this sensitive volume offer themselves: it may be that some molecules or particles must receive the energy of the ionisations and the energy is then transferred to the chromonema or else, instead of a particle, we should rather consider a «sensitive surface» or sensitive cross-section, which very probably is the link between two consecutive nemameres or chromonema blocks. Ultraviolet action spectra is in favour of the first possibility and particularly of the particles which absorb the energy being nucleoproteins, while the facts about the structure of the chromonema favour the second alternative. We will see below that a kind of synthesis between the two possibilities seems to explain the mutation process.
The conclusion about the question of the gene being a kind of physico-chemical unit like a molecule is that in reality the gene represents a biological, not a simply physico-chemical entity, and therefore is an unit at an higher level of organization; that is, the gene is composed of several physico-chemical units but these have an unitary biological action. The gene is, therefore, an unit in the functional sense and a composite in the structural sense.
In the course of our discussion we have repeatedly refer red to these theories and therefore here only the general conclusions about them will be exposed. To consider the chromosome as one unit is certainly valid at an organizational level above that of the gene, but this does not resolve the case of the gene. The chief point to explain when we pass from the gene to the chromosome level is the phenomenon of position effect. In chapter 3. we have tried to give an answer to the question of position effect without a recourse to the architectural theories. Even if it is supposed that the hypothesis we propose and others which maintain the concept of the gene as an unit will in the end be found incapable of fully explaining the position effect, this failure would not be: a proof of the extreme architectural theories, since these in their turn do not furnish such an explanation, also. In fact, what the architectural theories achieve is a transference of the problem to another level, without resolving it it is necessary to explain why a change in the architecture of the chromosome gives rise to a change in genetical properties.
An analogy with what happens at the atomic level seems-unjustifiable, by the same reason that an analogy between the cell and a molecule is also not valid; in the last analysis such comparisons reduce to the well known case of false analogies, obtained without due regard to the phenomenological level. An example of architectural hypotheses is that developed by KOLTZOFF (1939) according to whom the genonema is a polypeptide chain along which certain radicals are disposed in a certain order. Any change in the arrangement of the atoms of these chains could be a mutation, the gene itself being formed by one of the side groups along the polypeptide chain. If the chromonema, as we have said above, is a composite of many polypeptide chains, a mutation of this kind would imply the simultaneous change of many side groups, and this appears improbable.
The modern architectural theory of Goldschmidt, such as is exposed in GOLDSCHMIDT 1946, already considers the successive levels of organization of the gene and the chromosome, but continues to assimilate gene mutations simply with chromosome rearrangements, which seems to be an oversimplification. In reality, there is room for both categories of phenomena, according to the organizational level. The two extremes of known position effects are: 1. the long-range effect of heterochromatin, which can not be assimilated to a gene mutation; and 2. the small-range effect of the separation of chromonema blocks composing a gene, which are relocated apart, this being indeed very similar to a gene mutation of the ordinary inactivation-hypomorphic kind.
A gene concept, valid at our present scale of a very limited cytogenetical knowledge, must include elements of the architectural theories but the central point of the unit character of the gene must also be preserved. It is necessary not to transfer the problems to other levels which do not seem more promising, without first trying to resolve them at the basic level, in this case by maintaining the unitary gene concept at the same time that position effect is explained.
The cyto-physiological theory of the gene
We will now expose a concept of the gene which tries to include the conclusions we have referred to in the preceding discussions. As the treatment of the matter has been rather detailed and apologetic, it is now possible to follow a more dogmatic method, the reasons for the statements we make here having already been sufficiently discussed at one or other point of the preceding chapters.
The physiological approach
The typical gene has a unitary action. In certain cases, however, the action of a certain chromosome segment may appear almost unitary, although in it several individual genes; are separable by crossing-over. This may be the result of two opposite processes, either of one of incipient diversification following the establishment of a repeat, or of a process, of incipient inactivation due to heterochromatization. In both cases the lack of a well marked unitary action is the result of the segment of the chromosome being in a state of rapid evolution in what concerns its genes. When not otherwise: explicitly stated, it is presumed that the genes under discussion are typical, with a unitary action, that is genes in a state of evolutionary balance with the rest of the genotype and the ambient.
The unitary action of the gene must correspond to the elaboration of a certain haptogene, which intervenes in phe nogenesis. On the other hand, the quantity of haptogene produced is variable according to the allele in question and the respective quantum number (see page 467). The quantum number is the index of the amount of haptogene produced both below and above the gD level of phenotypic saturation effect. The amount of haptogene produced by a certain chromonema segment may be affected both by gene mutation. and by mimic position effect. For the production of a given quantum level it is irrelevant how the activation or inactivation has taken place.
The discrete Q+1 levels of haptogene production must correspond to the existence of Q possibilities of inactivation. If it is assumed that inactivation at the elementary level (that is at the level of the nemamere or the minimal chromonema block capable of genetic activity) is an all-or-none process, then Q represents the number of these minimal segments or active-zones in a gene, as we have termed them in a former work (SERRA 1944). Reasons for assuming that inactivation of each nemamere is an all-or-none process are: 1. each nemamere must be small and the region where the haptogene is produced very probably is only fraction of the nemamere and therefore even smaller, perhaps only a segment of the order of a hundred Å or so, with the active groups or the specific composition; 2. to the well known marked by quantic nature of the mutations, as is shown by the respective phenotypes, must correspond also quantic, discrete, levels of inactivation which necessarily are brought about by jumps, that is by all-or-none processes. We will assume, therefore, that Q represents both the physiological quantum level of maximum genetic activity and the cytological number of nemameres or the maximum of active zones. For a given level q of activity of the gene, the number of nemamers continues to be Q but those in activity are only q: the nemameres inactivated are then Q-q.
Another interesting question concerns the relation between the quantity of haptogene produced and the action of the gene. The inferences about this point are only of an indirect nature, obtained from examples of dosage. For relatively small doses the phenotypic effect is more or less proportional to dosage, but only those cases where the dose could be increased during a relatively great interval, until the saturation effect is reached or passed, are really demonstrative. In these cases, dosage-effect curves are exponential, or better, sigmoid in shape, since the proportional effect is found only at the beginning of the curves. The effect is approximately exponential until an inflection point is reached where the rate of increase starts decreasing. The exponential nature of the curves and other facts known about the phenogenesis of biochemical characters point out that the gene-controlled processes are, also in the case of growth and morphogenesis, catalytic reactions. The saturation effect is a consequence of it being produced a greater quantity of catalyst than was necessary to obtain the maximum effect, a mechanism which serves to assure the obtention of the maximum effect in the heterozygote, habitually corresponding to the most viable or normal phenotype.
The quantity of haptogene produced must be simply proportional to the number of active zones or nemameres which function, that is not inactivated by mutation or position effect, and the amount produced by each active zone probably is approximately the same for all the nemameres in activity and may be designated by h. On the other hand, the quantity of catalyst produced must also be simply proportional to the amount of haptogene, since probably this enters directly into the composition of polypeptides which function as apoferments in phenogenesis (see scheme page 464). It is only at the level of phenogenesis proper, when the catalyst comes into action, that an exponential effect is obtained.
|* Details on phenogenic and growth sigmoid curves may be found in SERRA 1949 and in a work to be published.|
According to this interpretation, the q number of a gene corresponds not only to the number of nemameres in activity, as also is an index of the amount of haptogene produced, which is h·q. The quantity of catalyst produced is also proportional to q, being a·hq while the final phenogenetic effect f is more complicated. This effect is for low q numbers approximately exponential f=b·eahq but for higher numbers a saturation is attained, so that the general curve is a sigmoid of the type encountered in. growth problems, of the form f =c (1-e-ahq)k or log f = log c+k·log (1-e-ahq). In all these equations f is the phenotypic effect, h is the amount of haptogene produced by each nemamere (haptogene constant) q is the quantum number or number of nemameres in activity, e is the basis of natural logarithms, a, b, c, and k are constants I characteristic of each gene, generally small integral positive numbers*.
The cytochemical approach
In typical cases, the phenogenic unit must correspond to a cytological unit during crossing-over. There must exist, therefore, certain conditions which establish the correspondence between the two kinds of units. The simplest supposition is to admit that the gene as a phenogenesis unit would correspond to a chromomere. We have discussed in detail this correspondence and concluded that only in special cases does it holds true. It is necessary, therefore, to try another explanation of the correspondence found in typical cases between the cross-over and the phenogenetic units.
Two hypotheses offer themselves as a basis of explanation: 1. In the critical stage of pachytene, when cross-over takes place, each segment of chromonema corresponding to a phenogenetic unit or gene would be individualized by deposition of matrix nucleoproteins, each phenogenetic unit being in this manner delimited from its neighbours. Cross-over breakage should take place between these units, but not, or only very rarely, within them. The deposition of the matrix nucleoproteins, individually upon each phenogenetic unit, could be due to a reaction of the haptogene with the histone-nucleotides of the nucleoplasm, forming a complex which in pachytene would deposit upon the corresponding chromonema unit segment. To each kind of haptogene should correspond a different complex and therefore the segment with a certain kind of haptogene would be individualized from its neighbours forming different haptogenes.
2. The second alternative hypothesis is that the bonds which unite each gene-segment or phenogenic unit to its neighbours are different from those which within each gene unite the elementary chromonema blocks or nemameres. Such an hypothesis would also, at first sight, correspond to data of another order, obtained in irradiation experiments. In fact, it has been claimed, as we have seen above, that for a chromosome breakage a densely ionising track with a mean of about 17 ionisations is necessary, while gene mutation would be due to a single ion pair (see page 494). If this distinction could be interpreted at its face value, it would mean that the bonds within the gene are weaker than the bonds between the genes, a conclusion at variance with the facts known on crossing-over and which therefore is to be rejected. To overcome this difficulty, as a supplementary hypothesis it could be postulated that the energy must be absorbed by the matrix nucleoproteins before it can affect the chromonema; if such is the case, those points with a greater amount of nucleoproteins, that is precisely the gene-units, should require a lower amount of energy. This Supplementary hypothesis would explain also the results of ultraviolet irradiation. It would imply, also, that heterochromatic regions should be more sensitive to irradiation effects than euchromatic ones, which do not corresponds to the best evidence obtained in Drosophila melanogaster (KAUFMANN 1946). However, the fact that euchromatin and heterochromatin in the sperm head must have almost the same charge of peripheric nucleoproteins, and therefore present no distinction in this respect when sperms are the irradiated material, explains the similarity found in breakage frequency of both: kinds of chromatin. Only the irradiation of stages where a distinction between the nucleoprotein charges of the two chromatin kinds is possible, for instance of salivary glands primordia, would allow that a conclusion be reached on this point.
Nevertheless, there are facts of another order which are, against the hypothesis that intra-genic bonds are weaker than inter-genic ones. In effect, if such a distinction existed, it would be expected that the evolution of the genetic systems: should be difficult, a diversification within a segment which once was a phenogenetic unit being highly improbable. This is contrary to the existence of well characterized cases of, incipient gene formation to which we repeatedly referred above and indeed would hinder considerably the evolution of living forms. In this domain a greater souplesse is the indispensable attribute of nature, if the existence of the innumerable alleys of evolution are to be explained. Not only greater blocks of the chromonemata as also minute ones must be susceptible of diversification.
As the two alternative hypotheses have some advantages and disadvantages, it seems preferable to adopt a combination of both. The bonds which unite the nemameres within a gene must be of the same nature as those which unite consecutive genes, but when the chromosome is in prophase, and particularly in pachytene, the intragenic bonds become more strong than the intergenic ones. This probably is realized through the deposition of the matrix nucleoproteins and can. consist in a combination of the haptogene, or of the active zones themselves, with the nucleoproteins; the regions between the genes, where active zones and haptogenes do not exist, would remain weaker. Alternatively, it may be supposed that the intragenic bonds are maintained throughout pachytene with their normal strength but the intergenic ones are weakened by some process, particularly during the division of the chromonemata, since cross-over is contemporary to chromonema division in pachytene. Probably the deposition of the matrix and the division process both play a role.
According to what is known about the crossing-over process, only a small amount of energy is necessary to break an intergenic link in pachytene. As the exchange process takes place in perfectly physiological conditions and is of great biological significance for the genetic variability of the species, it must be based upon a safety mechanism which assures a greater amount of energy than habitually is necessary. Probably this is obtained rather by the intergenic bonds being weakened than by means of higher amounts of energy becoming available at the level of the chromonema, in late pachytene.
Let us now proceed to discuss the nature of the intergenic and intragenic bonds, which are supposed to be of the same nature. We have already seen above that they cannot be peptide linkages or other high energy bonds. Even the very weak hydrogen bond, with an energy of 5 kcal/mole, seems to require an higher level of energy than is available at the breaks, in radiation experiments. If we assume that an ion pair is responsible for the breakage of the intragenic bonds, then only 749/5 or about 1.5×102 of these bonds could exist in each intragenic link (and intergenic also, if they are identical). If, however, the energy of the ionisations which in mean are necessary to break a Tradescantia chromosome is all employed in breaking a chromonema, then about 2.5×103 hydrogen bonds could be broken.
Hydrogen bonds are formed between CO and NH neighbouring groups of the polypeptide chains and have been postulated to play an important role in maintaining the structure of the globular proteins (PAULING 1940). Very probably they play also an important role in the case of structural proteins of the cell. Other types of bonds have also been postulated to occur in cytoplasmic proteins, for instance saline bonds between positive and negative groups, disulphide bonds, etc. (FREY WYSSLING 1938, SERRA 1942). These, however, all are short range forces and could not be operative between two broken chromosome ends at a distance of the order of 1µ or so. For such a distance only long range forces may come into action. Polymerization forces, which could perhaps act at relatively long distances, seem to be out of question in the case of the chromonema proper, because this is definitely allways formed of discrete units. It seems safe to conclude that for such distances as those which prevail between broken ends, only long range forces of the coulomb type may be operative. Attraction forces between two particles, of the Van der Waals type, depend upon the existence of permanent or induced dipole moments. Calculating the number of dipole moments, each of the order of magnitude of 2 debyes, as are usually met with in organic molecules, by means of the formula E=m/r3 (see page 436) we find, for a value of r = 1µ and E of 749 kcal/mole (corresponding to an ion pair), that about 1012 dipole moments are necessary. If each end group of the polypeptide chains of the chromonema forms a dipole, 1012 end groups would be necessary, that is about 108 groups for each of the 104 chains which form a minimal chromonema. If we take 102 as the mean weight of an aminoacid residue and suppose that each bears one or two polarisable groups, this number of 108 polarisable groups in the terminal part of the poly peptide chains would require a very long chain, with about 106 aminoacid residues. Accordingly, the polypeptide chains at the surface of contact between two consecutive nemameres should be bent in a manner so as to expose a very great number of polarisable groups whose dipoles could in part be, induced by the approximation of similar broken ends. It seems unlikely that such bent long chains could exist at the nernamere's contacts and the right conclusion is that other causes are operative in bringing together broken ends when they happen to come within less than about 1µ, of one anothers. Even if it was assumed that the distance at which the forces of attraction act is lower than 1µ, say about 0.2µ, and that the number of polypeptide chains in the chromonema is of the order of 106 instead of 104, the number of polarisable groups at the nemamere's contacts should yet be of the order of 102-104, which does not sound unreasonable but would yet necessitate of the end parts of the polypeptide chains being bent or looped in special patterns to expose a relatively great number of polarisable groups.
Although these calculations are subject to considerable incertitude, it seems that from the above considerations it may reasonably be concluded that forces other than those developed between the broken ends of the chromonemata are operative in bringing together those ends. The evidence obtained in the case of sperms is in favour of the broken ends remaining so until the sperm enters the egg, when its nucleus acquires a swollen mitotic stage. The habitual result of X‑raying in the resting stage is also the appearance of breaks at the following anaphase. This points out that at least a part of the breaks which occur in resting stage do not restitute, despite the fact that broken ends were close together. All the facts tend to suggest that the reunion process is strongly influenced by the matrix nucleoproteins. Very probably the energy necessary to break a chromosome is chiefly absorbed in altering the state of the nucleoproteins and, according to the degree of this alteration of the matrix gel, the chromosome may or may not be broken. The energy required to break the chromonema proper must be low when compared with that necessary to bring about the matrix changes. Conversely, the attractive forces which act at the broken ends probably also concern in a great measure the matrix nucleoproteins. It seems likely that a kind of «halo» of partially depolymerized nucleoproteins extends from recently broken and non‑healed ends, which would facilitate the reunion by a gelification process. The forces acting between particles which are in process of forming a gel are in part long range forces (see discussion in. FERRY 1948). In the case of the matrix nucleoproteins it appears that in part they are polymerisation forces chiefly from the thymonucleotides, as is demonstrated by the fact that when an advanced gel state is attained the matrix nucleoproteins of a chromosome no more have the possibility of fusion with those of other chromosomes, while, on the contrary, depolymerisation of the matrix brings about stickiness and fusion.
The probable conclusion, therefore, is that the reunion of broken ends is first determined by the capacity of fusion of the matrix nucleoproteins which act at relatively great distances, of about 1 µ; afterwards, when the chromonema broken ends are closer, at distances of the order of 0.1 µ, other forces, yet long range ones, of the nature of Van der Waals forces, may come into action; finally, at shorter distances, of some Angströms, the chemical bonding forces of electrovalent, hydrogen and other types come into action. The overall result is a relatively weak link of the chromonema units, probably with an energy of the order of 102-105 kcal/mole, but this link is greatly reinforced by the action of the gel formed by matrix nucleoproteins and also, when the matrix is absent, as happens in oocyte chromosomes and certain resting stage nuclei, probably by the interaction of the chromonema with the nucleoplasm, this forming a kind of pelicle with rod shaped and punctiform corpuscles at the surface of the chromonema (SERRA 1947c). The links are thus sufficiently reinforced to give a relatively great stability to the chromonema as a whole.
Conclusions on the structure of the chromonema and the gene
From the foregoing discussion the following picture of the chromonema and the gene is obtained (FIG. 5)
1. According to its thickness, each chromonema is composed of about 104-106 polypeptide chains, which in the main run parallel to the chromosome axis, but are always more or less folded. Probably in some points, namely the genetically active regions, where the haptogene or primary gene product is produced, the polypeptide chains are almost distended, while the rest of the chromonema presents their chains folded.
|FIG. 5. At the left: Scheme of the composition of a gene, formed by 3 nemameres. Ordinarily the number of nemameres of a gene must be between 3 and 10. At the points h the polypeptide chains, whose number is at least 104-106, are distended; it is at these points that the haptogene is formed. At the two sides of the gene are linking regions, with folded chains. At the right: A detail of the zone h, where the haptogene is produced. Plain lines represent polypeptide chains, dotted a nucleic acid chain. It is supposed that in each nemamere only one of these regions h is present.|
2. The minimal composition of the chromonema is of proteins of a non-basic type, probably with a small amount of ribose type nucleotides at certain points, perhaps at the active zones. This minimal composition only realizes in certain cases, when the matrix is totally absent, namely during the growing period of oocytes and probably in certain resting stages. Generally the chromosome is composed of the chromonemata plus matrix thymonucleoproteins (with histone type proteins) and products of interaction with the nucleoplasm, these latter possessing ribonucleotides besides other types of compounds (proteins, lipoproteins, phosphatides, lipids, etc.).
3. According to its length, each chromonema is composed of a succession of minimal structural units which may be called nemameres. One, or generally more than one, of these units form a phenogenetic unit or gene.
4. The links between the nemameres are relatively weak, probably like cohesion forces with an energy of the order of 102-103 kcal/mole. These links can not be peptide nor other high energy bonds. Probably they are chiefly of the nature of Van der Waals forces resulting from permanent and induced dipole moments at the terminal groups of the 104-106 polypeptide chains which form each chromonema. These weak links, however, are reinforced by interactions with the nucleoplasm and with the matrix nucleoproteins, which protect the chromonema against rupture. In order to break a chromonema it is necessary not only that the breakage agent reaches the chromosome as also that it alters the means which protect the chromonema against rupture.
5. The links between the nemameres which form a gene and those between adjacent genes have been postulated to be of the same nature. What individualizes a gene is the production of one kind of haptogene and this results in the formation of an haptogene-matrix complex which separates the phenogenic unit from its neighbours during certain stages of meiosis, and particularly in pachytene. In other stages, and especially in the salivary chromosomes, this distinction, is not valid and inter- and intra-links of the genes are equal In this latter case it is only possibly to recognise the gene limits by analysing the genetic consequences of the breakage at certain points; a morphological inspection of such chromosomes do not allow that a distinction be made between genes and inter-genes.
6. Each gene or phenogenic unit is formed of a number of nemameres which is equal to the maximum number of active zones. This number is designated by Q, the genic quantum number. Each active zone produces an approximately equal amount h of haptogene (haptogene constant). Inactivation and activation of each nemamere or active zone are postulated to be all-or-none processes. The number of nemameres which are in activity in each gene may be from o to Q; the gene has therefore Q+1 possible states. The total quantity of haptogene produced is h·q., in which q is an integral number corresponding to the active quantic number (from o to Q) that is of the active zones, or nemameres which have not been inactivated.
7. It is supposed that the haptogenes are relatively, simple polypeptide compounds which are further elaborated in the nucleoplasm and the cytoplasm to specific proteins and apoferments. The haptogenes are supposed to be synthesized from simple polypeptides, by an interaction of the nucleoplasm with the chromonema. These polypeptides come from the cytoplasm, where they have been elaborated by the cell ferments. Only small parts of the nemameres, those which have distended polypeptide chains, serve as templets for the further elaboration of the aminoacids or simple polypeptides received from the cytoplasm, which are elaborated to greater units with a certain pattern and certain active groups, functioning like haptenes.
8. The amount of apoferment formed with basis on the haptogene produced by the gene is a.hq, directly proportional to the quantity of haptogene. The action of the ferments in phenogenesis is primarily exponential but when a sufficient dosage is attained the final action follows a sigmoid curve c(1-e-ahq)k. In these expressions, a, b, c, k are constants characteristic for each gene and e is the basis of natural logarithms.
9. Gene reduplication is different from haptogene production. Very probably the chromonema behaves in this respect as a whole, growing by apposition growth. Polypeptide chains already folded in specific ways, due to cytoplasmic eiabora tion serving the haptogenes as templets, are received from the cytoplasm and further elaborated by nucleoplasmic-chromonema interaction to give the nemameres' chains. When a sufficient thickness is attained, the chromonema is divided into two halves, individualized by thymonucleoproteins. Phosphorus compounds must play a role in supplying energy-rich bonds to the gene sytheses, not only of the nemameres' polypeptide chains as also of the haptogenes (further detail under the following heading). Probably the small amount of permanent ribonucleotides of the chromonema play a structural role and are not utilized to give the anabolic energy while the thyrnonucleoproteins must also play chiefly a structural and protective role in the chromosome as a whole, although they may also furnish in their catabolism energy and building blocks for the chromonema.
10. Inactivation of the active zones is ordinary, hypomorphic, mutation and will be further discussed under the following heading.
As much of the relevant data on gene mutation and their discussion has already been dealt with above, the question of the nature of gene mutation in comparison with chromonema structural changes may now be treated rather dogmatically; the treatment will be apologetic only when the subject was not sufficiently discussed in previous chapters.
The two chief alternatives in what concerns gene mutation are: 1. Gene mutation takes place by a kind of error during gene reduplication, the mutated gene being a faulty copy of the original gene. 2. Gene mutation is an alteration in the composition, or in the structure, or both, of the gene, not necessarily linked with gene reduplication. To the second alternative belong the structural theories of gene mutation, while the various hypotheses of an alteration in the composition of one side group of the gene polypeptide chains, or in the polypeptide chains themselves, may belong to one or the other alternative. As these theories have repeatedly been criticized in former chapters, we will refer here only to the points essential to our main question.
Metabolism of chromosome compounds
|* The literature on this subject cited in the present work without date is referred to in Hevesy's book.|
Desoxyribonucleic acit presents a low turnover rate, of the order of 1-2% per day for brain, kidney and liver and only in organs which are in active division is the turnover rate greater, of the order of 6-16%, for spleen and intestinal mucosa (Hevesy and Ottesen, Halm and Hevesy). Ribonucleic acid shows a much greater turnover than desoxyribonucleic acid in liver and kidney and only a slightly greater one in intestine and regenerating liver (2 hours after injection of labelled phosphate, in rats). These results accord well with the known cytological fact that desoxyribonucleic acid is active during cell growth and synthetic activities.
With respect to the role of ribonucleic acid during cell syntheses, it has been concluded, and this is now a much accepted idea, that nucleic acid takes part in the synthesis of proteins (Brachet 1933, Caspersson 1936, Hyden 1943, see also Spiegelman and Kamen 1947). The evidence in favour of this assumption is only indirect and to judge from tracer results it seems that in reality nucleic acids simply respond to the rise in the level of the general metabolism during cell syntheses, more or less as do other cell constituents; the nucleic acid fraction is not especially active, the phosphate and phosphatide fractions of P compounds being more active than the nucleotide fraction. The energy-rich phosphorus compounds very probably serve as energy donators in protein syntheses (Wiame 1947, Spiegelman and Kamen 1947). It is possible that this is in part effected by nucleotide phosphorus (Muller, Spiegelman and Kamen) but in view of the more rapid turnover of phosphate compounds it seems more probable that this is realized by means of protein bound phosphate. This mechanism must also be operative in the synthesis of the chromonema proteins during the resting stage. During cell division the role of the nucleoproteins is a structural one, "protecting" the chromonema and furnishing a means to chromosome individualization.
Of the proteins it is known that they may have from a very rapid to a very slow turnover rate, according to the physiological role they play. Ordinarily, the half life of the proteins may go from some hours to some days, with an average of about one or two days (Rittenberg 1941). From the very low frequency of mutation, it is to be expected that the chromonemata proteins shall have a relatively low turnover rate, due to their almost fibrous structure and to probably taking part only in synthetic activities and not, or only at a very low rate, in disintegrative, catabolic, metabolism (Serra 1947b). It would be very important to submit this hypothesis to experimental test.
The rough picture of the metabolism of chromosome compounds which the actually existing data allow is the following: The proteins of the chromonema and the genes have an intense integrative metabolism when reduplication and haptogene formation occur, during the cell resting stage and during growth, but their catabolism appears to be very low. Energy-rich phosphorus compounds probably furnish energy for these syntheses, the chief compounds involved being phosphate bound to proteins, and probably also phosphatides. Nucleic acids, while contributing also to furnish energy-rich phosphate bonds during chromonema syntheses, probably play their essential role chiefly in serving as structural compounds which form part of the structures necessary for cell activities: cytoplasmic granula, secretion droplets, distension of certain chromonema parts etc., for ribonucleic acid; matrix formation for thymonucleic acid.
Mutation and faulty reduplication
The hypothesis that mutation may result from a faulty reduplication of the gene is widely accepted; it seems, however, that generally without sufficient criticism. This hypothesis assumes that a change in one group of atoms, or even in one atom, of the compound which makes the gene may bring about a mutation, that is in such a view it is implicitly admitted that each gene is composed of only one backbone structure, of one polypeptide chainFig. 6. We have seen above, however, that the chromonema has about 104-106 polypeptide chains and therefore such a view is highly improbable. For a change in a certain active or prosthetic group of a whole gene to take place, it would be necessary that 104-106 groups are changed at once, by a faulty reduplication, which seems unlikely. The other alternatives are: 1. that a change in only one group of one of the 104 chains is capable of changing at once the properties of the whole gene; and 2. that the change is a statistical one, taking place gradually but altering the properties of the gene only after many groups have been substituted. Alternative 1. seems highly improbable and we will not consider it further. Alternative 2., on the contrary, is interesting and merits being considered in some detail.
|FIG. 6. Two alternaive explanations of mutation. At the left, a change in an aminoacid residue (change of a phenylalanil to a tyrosil group) results in mutation, M. At the right, mutation Mut is due to inactivation of two nemameres due to the break at B. This break restituted but left the contact between the nemameres changesd, which caused the folding of the distended chains. The haptogene h of these two nemameres is not formed in the mutated gene Mut. The active nemameres pass from 4 to 2. See also the text.|
Supposing that by faulty copy a specific group of one of the 104 polypeptide chains was changed and that this occurred repeatedly, it could happen that when a chromonema divides could go to one side all, or at least the great majority, of the changed chains and this should result in a mutated gene. This would explain in part the rarity of mutations, which would be effective only when by chance the chains with the altered group were included in one half-chromonema. On the other hand, it could be supposed that instead of a total change, gradual changes were possible, each time the chromonema enriched in the altered group a step mutation taking place. It could even be supposed that these gradual alterations and not the number of active zones corresponded to the quanta numbers of the gene. However, this assumption of a statistical mutation seems improbable; in fact, if a change in a single group, or even a change in some groups at once, corresponded to mutation, owing to the existence of 104 groups the number of alleles in multiple allelic series would be expected to be much greater than usually it is. It should also result that the gene was always in a state of change, because fluctuations in the average composition would be frequent. There are sufficient reasons to reject such an assumption.
The conclusion, therefore, is that the gene is a relatively stable composite structure in which only with difficulty statistical fluctuations in its composition may bring about genetical consequences. The hypothesis that gene mutation appears by faulty reduplication seem really little probable, or better, to be operative only in very rare cases of mutation. These rare cases may, however, be very important in evolution, because a new gene composition could perhaps be the starting point of a new evolutionary line, for instance if it corresponded to the appearing of new growth or cell multiplication substances operative in opening the way to the appearance of new, revolutionary, phenotypes.
By the same reason that a change in chemical composition during
reduplication does not seem to furnish a valid explanation of gene
mutation of the ordinary type, a change due to a catabolic error or
otherwise induced only in one or some of the groups of the aggregate which
forms the gene, also does not furnish a basis of explanation for this kind
of mutation. Besides this negative argument, there are other reasons to
admit that structural alterations, that is alterations in the structure of
the sub-units which form the gene, are the key to gene mutations — FIG. 6.
First, it is verified that all agents which induce gene mutations, also
cause breakage of the chromosomes; second, chromosome breakage and gene
mutations have the same genetical results. This similarity in cause and
effect would be surprising if the two kinds of mutations, genie and
chromosomal, had not a great similarity in nature. This is one of the
facts which more strongly advocate in favour of the architectural
theories. The weakness of these theories lies not in the postulate of a
change in the architecture of the gene bringing about a genetic change,
but in they not explaining how a genetical alteration is correlated with
the structural change.
As was repeatedly said above, ordinary gene mutation of the hypomorphic-hypermorphic type corresponds to the inactivation or reactivation of the nemameres which compose the gene. The q nemameres which remain active give the amount h.q of haptogene. It is assumed that inactivation and reactivation in each nemamere is a quantum process, that is an all-or-none happening. If it is accepted that the action of mutagenic agents upon the genes is the same as that which provokes chromonema rearrangements, then inactivation (or reactivation) must be initiated by a break within the gene, which restitutes but leaves the gene changed. In a former work (SERRA 1944) we have exposed the hypothesis of the gene mutations consisting in intra-genic rearrangements, for instance the inversion of one of the sub-units of the gene. This would require at least two breaks within a gene in order to cause a gene mutation. Although it is not impossible that the energy of one ionization is capable of being propagated at a small distance along the chromonema (see SERRA 1944, 1945a) it seems that the necessity of the occurrence of two simultaneous breaks is a too narrow condition. It is preferable to postulate that a single intra-genic break will be sufficient to cause a gene mutation, if restitution does not give rise exactly to the original links.
Now it is necessary to explain why a break which has not restituted in exactly the original manner may cause an inactivation of the neighboring nemameres. Two hypotheses offer themselves as possible: It may be that during the time that the intragenic bonds were open a small amount of non-chromonematic material, probably matricial nucleoproteins or nucleoplasmic compounds, was trapped between the extremes and this would result in a modification of the region of contact between the two nemameres; or else, it is conceivable that the new links formed were different in their chemical bonds or in the strength of the Van der Waals forces involved, due to the end groups of the proteins having been folded in a different way during the period in which the links remained open. This second hypothesis seems the most probable, since if the first was true it would be expected that the pure chromonema should present extraneous materials included within the genes, probably thymonucleoproteins. We will, therefore, adopt the second alternative see Fig. 7.
Providing that the contacts between two adjacent nemameres have been modified, how is the production of haptogene hindered in one or the two nemameres? The most obvious hypothesis is that the different folding of the polypeptide chains at the contacts extends its effect to the whole nemamere, for instance by modifying the state of folding and distension of the rest of the polypeptide chains in the whole active zone. 1f modification at the contacts caused that the relatively small region of the polypeptide chains which serve as templet for haptogene formation, and which we postulated to be formed of distended chains, was now folded, then haptogene formation would be difficult or impossible.
FIG. 7. Cross-sectional scheme of the changes in the folding of the polypeptide chains of a nemamere, following a change in the bonds of the contact (C) with an adjacent nemamere. In B the region with distended chains has now folded chains, which hinders haptogene production and results in inactivation. R, aminoacid residues with opposite charges; H, hydrogen bonds.
An intimate knowledge of the nature of the gene and gene mutations would eventually allow a control of the direction of mutation. As yet, although it is already possible to increase considerably the rate of mutation, it is not possible to choose the loci which will be affected; at most, there are some indications how to attain this important aim of the qualitative control of mutation.
It has been claimed that the transformation of bacteria types by certain extracts is in fact directed mutation. Pneumococci forming smooth colonies in agar may be induced to lose its virulence and to give rough colonies. This transformation is accompanied by the loss of the specific polysacharide envelope which is characteristic of the smooth forms (results Grifith, Dawson, Avery, Heidelberger, and others, reviews in Dobzhansky 1941, Boivin 1947). It is possible to induce the reversion of the rough or degenerated form to the smooth form by means of extracts of dead smooth forms and later on it was demonstrated that the principle desoxyribonucleic acid of the smooth form, or at least a contaminant which in the extracts is not possible to separate from the thymonucleic fractions. To induce type reversion, the desoxyribonucleic acid must be in its polymerized form and blood serum or ascitic fluids must be present; depolymerization of the nucleic acid inactivates the power to induce type reversion.
A similar phenomenon happens to colon bacilli (Escherichia coli), which have a multiplicity of antigenic types and, concomittantly with a polysacharide loss, may suffer degradation from the smooth to the rough form. Reversion of the rough to the smooth form is induced by highly polymerised thymonucleic acid or a small amount of a contaminant, very probably protein, which induces type transformation, although Boivin is in favour of the first alternative, which really seems more probable (see discussion by Mirsky to the article of Boivin 1947). If thymonucleic acid is the inducing principle, this would implicate that this acid has specific types of polymerization in long chains; this is not yet proved from the biochemical side. However important it is to know the exact compound which induces type transformation, for the problem of directed mutation this is of no immediate significance.
The usual interpretation of type transformation in bacteria by desoxyribonucleic preparations is that the preparation induces a specific mutation. However, another interpretation is also possible: the smooth form has a limited capacity of synthesis of an active compound which is furnished by the preparation, or which is synthesized by the cells when the active compound is present. A minimal amount of the active compound is necessary in order that a bacterial cell may produce the characteristic polysacharide and the smooth colony form; the supply of this minimal amount restores the smooth type. That is, the preparation acts like a substract, not as a mutation agent, and the case is rather similar to that of the Killer principle in Paramoecium. Another hypothesis is that the active compound is a component of the gene which determines the smooth form (Wright, 1945); this component would be lost during bacterial reproduction and its supply by a preparation would restore the smooth form. The hypothesis of a substract or a kind of plasmagene seems more likely; at any rate, it seems not probable that this is a case of directed gene mutation.
|* Of course, this does not signify that the genes play no role in cancerization. It is well known that the susceptibility to cancerization is genetically determined. We refer only to the immediate stimulus, which causes the déclanchement of the uncontrolled multiplication of a cell.|
Another supposed case of directed mutation would be the induction of cancer by chemicals of the phenanthrene group and by certain viruses. The interpretation of cancerization as being due to mutation in a gene of a cell is, however, very doubtful or, better, improbable. It seems much more likely to attribute cancerization of a cell to the disturbance of the normal balance between its capacity for multiplication and for differentiation; the factors capable of upsetting this balance towards multiplication power will cause cancerization. It seems that a reversion to a condition normally found in embryonic cells must not be attributed simply to a mutation but to a change in the cytoplasm-nucleus systems of multiplication, growth and differentiation. These systems include the energetic and structural compounds, that is not only nucleoproteins as also simpler phosphorus compounds and several metabolites. In the actual state of knowledge it is not possible to define the limiting factors, though it seems that steroids and nucleoproteins, given a rich supply of energetic metabolites, play the decisive role, upsetting the normal balance towards multiplication of the uncontrolled type. In any case, the interpretation of cancerization as being due to mutation has no serious basis and represents that kind of pretensely resolving biological problems by postulating that an unknown, generally something which happened to the genes, is the responsible. The alterations which produce cancerization probably lie more in the cytoplasm than in the genes.*
From the fallacy of these cases it must not be concluded that the goal of directed gene mutation can not be attained. Two ways seem possible: by acting directly upon the germen, or by ambiental actions through the soma. The first way has yet three alternatives: 1) induction of breaks in a selected chromosome region; 2) an action upon haptogene formation by a competitive compound; and 3) a similar action upon the building blocks of the genes. The second procedure, of the attack through the soma, seems the most remote but may in the end turn out to be very interesting for a comprehension of evolution. By ambiental actions the formation of certain compounds or the realization of new physiological conditions may be obtained which modify the functioning of the cell by altering either the nucleus or the cytoplasm, and this may modify the realization of the whole phenogenesis. We will discuss briefly these possibilities.
Direct action upon the germen by inducing breaks in a certain chromosome region has already been applied in a practice with interesting results (McClintock 1941-1944). Chromosome aberrations which mechanically produce a great frequency of breaks in a certain region also give rise to mutations of factors located in that region. Inversions and translocation and the nucleolus-zone provide the necessary means to induce breaks. It is not impossible to induce breaks in a certain region by means other than the mechanical ones, for instance by micro-radiation experiments, irradiating certain small chromosomes or certain chromosome parts with ultra-violet light in growing pollen tubes, or by applying locally at chromosomes, with suitable apparatus, minute amounts of breaking compounds like the mustards; these possibilities are difficult to realize in practice.
The second means of directed action upon the germ, interference with the production of the normal haptogene, is as yet only a theoretical possibility but offers great perspectives for the control of mutation. If during the process of haptogene systhesis by the nemameres, instead of the right kind of components, probably simple polypeptides, somewhat different compounds, which by their similarity with the original compounds can compete with them, are furnished to the chromonemata by the nucleoplasm, an inactivation of the gene would result. This, in its turn, could subsequently induce a change in the same or other genes because gene reduplication is subsidiary to the furnishing by the cytoplasm-nucleoplasm system of the right kind of building blocks for the genes.
The third possibility of directly inducing germ modifications, by interference with the power of multiplication of the gene, could thus be a corollary of the competition for haptogene formation, although it could also be attempted, probably with much more difficulty since the necessary compounds must be more complicated, by furnishing competitive compounds somewhat different of the blocks which build the genes.
Ambiental actions capable of influencing the germen through the soma are for instance those which modify the heterochromatinization, to which we referred in chapter 6. Other cases are the modification of the composition of the cytoplasm, and especially of the enzymatic systems, which may influence the production of the building blocks for haptogene synthesis and gene reduplication furnished to the nucleus. A strong modification in dietary conditions may bring about changes in the composition of the cytoplasm which afterwards could reflect in the chromonemata. As other parts of the cell, the chromonemata and genes very probably are capable of being influenced by metabolic changes and nutrition, although the conditions for such an influence only very rarely realize in nature and are difficult to know. It is for future work to investigate the ways of attaining these changes, by means of which evolution may eventually be directed.