Experimental Cell Research, Suppl. 6, 236-251 (1959)
The Genetical Interactions Between Nucleus and Cytoplasm in Epilobium
P. Michaelis
Max-Planck-Institut für Züchtungsfarschung, Abteilung für Plasmavererbung, Koln-Vogelsang, Germany

The cell is the unit of life, in which nucleus and cytoplasm form a self-perpetuating reaction system. This reaction system can be considered not only from the physiological viewpoint, but from a genetical aspect as well. All life reactions are specific for different species and they are inherited from the parents to the progeny, because all living units of the cell system possess the potentiality of identical reproduction.

The first task for the genetic investigator is to analyse the living substance and to differentiate between the various self-perpetuating particles of the cell. The methodical way is the comparison of genetical behaviour of inherited properties with the behaviour of cell particles during numerical increase of cells and organisms. During the last sixty years geneticists have analysed nuclear genes by comparing the Mendelian behaviour of certain characteristics with the behaviour of nucleus and chromosomes. The field of cytoplasmic inheritance is far less fully explored. In the last few decades it has been possible to demonstrate an independent cytoplasmic inheritance by investigating non-Mendelian characteristics and reciprocally different hybrids; but it has not hitherto been possible to differentiate between the different components of the cytoplasm, for example between plastids, chondriosomes, microsomes, and cytoplasm itself [6-8]. If we now consider the interactions between nucleus and cytoplasm, we can only contrast the nuclear genes with the sum of all cytoplasmic constituents. This sum is designated: plasmon. In this stage of knowledge we will only consider the effects of nucleus and cytoplasm on the phenotypic properties of organisms. The more exact investigation of the chains of physiological reactions within the cell may be postponed until we are able to investigate the interactions between nuclear genes and cytoplasmic units themselves.


If a geneticist intends to investigate the interaction between nuclear genes and the plasmon, he must first look for genetically-different nuclei and different plasmons, and he must be able to unite these nuclei and cytoplasms in an organism without disturbing its viability. In this regard, the plant Epilobium from the family Oenotheraceae, the evening primrose, is a favourable object, for two main reasons. (1) On fertilization the cytoplasm of the Epilobium zygote originates only from the mother. Paternal nuclei canbe crossed into the maternal cytoplasm without mixing the cytoplasms of the two parents. By reciprocal crosses, constant reciprocal differences may be produced. (2) Quite different cytoplasms and nuclei can be united in Epilobium and striking reciprocal differences can be produced without disturbing the viability. Epilobium is very little specialised, so that the physiological reactions proceed with a wide range of tolerance.

In order to find different nuclei and different cytoplasms representative of the genus Epilobium and suitable for experiments, a systematic search has been made among a large collection of mutants and natural strains and species.

The Search for Suitable Genes and Genomes

At first the nuclei of the largest part of this collection were investigated by introducing them into a cytoplasm which tends to produce hybrids different from their reciprocals. The Epilobium hirsutum strain from Jena possesses such a cytoplasm and was reciprocally crossed with the different species and strains of the collection (Chart 1).

CHART 1. First series of crosses.    
  x E. species 1  
  x E. species 2  
  x E. species 3  
  x E. species 4  
  x E. Species 111  
Epilobium hirsutum Jena ()      
  x E. hirsutum strain 1 ()
  x E. hirsutum strain 2  
  x E. hirsutum strain 3  
  x E. hirsutum strain 4  
  x E. hirsutum strain 390  

Crosses with E. hirsutum Jena as father are without any interest, all the hybrids being normal. All the hybrids with E. hirsutum Jena as mother have the same cytoplasm but different combinations of nuclei. The reciprocal species crosses behave similarly to the strain crosses. In 13.6 per cent of the 44 crosses with E. parviflorum a plasmon segregation has been found in the Jena cytoplasm which will be subsequently discussed. In all other crosses the cytoplasmic inheritance remained absolutely constant. The reciprocal differences were very heterogeneous. 42 per cent of the 30 strain crosses were reciprocally alike and 30 per cent showed very small differences. 25 per cent of the hybrids were distinctly different through stunting or dwarfing of the plants with Jena cytoplasm. Finally 3 per cent of the crosses were lethal with Jena cytoplasm [2, 4].

From these results it follows that the degree of reciprocal difference does not depend on the cytoplasm alone. The cytoplasm of E. hirsutum Jena is the primary condition for reciprocal differences, but the magnitude of reciprocal differences is determined by the incrossed nuclei. Some E. hirsutum strains possess strong effective nuclei, which release great differences, while other strains contain weak nuclei, which cannot release differences.

The Search for Plasnon Differences

From the series of crosses described some strains containing test genomes of different strength were selected, and in a second group of crosses these test genomes were placed in the different cytoplasms of the collection (Chart 2).

CHART 2. Second series of crosses.
Epilobium species or strain 1

4 x test genomes of different hirsutum strains
  with 610 crosses

Through these crosses strains have been found, especially in central Germany, containing cytoplasms similar to the cytoplasm of the Jena strain, which with strong test genomes produced stunted and dwarfed plants. Other cytoplasms have been found in southern Germany producing disturbances of flower development, and in the Mediterranean areas and other territories cytoplasms occur which produce pollen sterility. But the cytoplasms within each one of these three groups were similar to one another, but not identical, for varying strengths of the test genomes were required to release a certain degree of reciprocal difference. No two strains could be found in the large collection possessing completely identical cytoplasms, not even in strains which had come from the same locality within a distance of only a few metres. It is a very interesting fact that differences among cytoplasms occur very frequently even if reciprocally-different hybrids are rare. Reciprocally-different hybrids arise only if a test genome possessing a special strength is placed in a cytoplasm with a certain degree of action.

Now, the intensity of reaction of cytoplasm has been compared with the strength of the nucleus in natural strains. The triple-cross-procedure used is complicated, and an explanation would take too much time [1]. I will present only the results. Strains with strong effective nuclei contain weak cytoplasms and strains with weak nuclei contain strong cytoplasms. These contrasting effects of nucleus and cytoplasm on the degree of reciprocal difference demonstrate an equilibrium between nucleus and cytoplasm. The disturbance of this balance produces the disturbances of development in reciprocally-different hybrids.

To summarize very briefly the results of the first part of the investigations presented in this paper: In Epilobium reciprocal differences by cytoplasmic inheritance are not caused by cytoplasm alone, but by a disturbed balance between nucleus and cytoplasm.

Through such a systematic search within the Epilobium collection a large number of different cytoplasms and different nuclei has been found. The most interesting nucleus-cytoplasm combinations have been selected and investigations made of the constancy of cytoplasm and the influence of cytoplasm on nuclear genes. First of all we shall consider the interactions between cytoplasm and nuclear genes.


Change of Gene-Manifestation

The simplest and clearest results are obtained if one single gene is influenced by the cytoplasm. These cases occur rarely on hybridisation of natural strains, so I shall discuss an example from my mutation experiments [11]. Following treatment of Epilobium parviflorum with radioactive isotopes, 67 mutations of nuclear genes arose. Three of these mutations were sensitive to the cytoplasm of E. hirsutum. The E. hirsutum-parviflorum hybrid showed very great reciprocal differences (Fig. 1). The hybrids with E. parviflorum cytoplasm were normal and about 50 cm high, while the reciprocal hybrids with E. hirsutum as the mother remained dwarfs only a few centimetres tall. One of the plasma-sensitive mutants was "resistens", a recessive mutant showing resistance against the mildew fungus Erysiphe. The growth of the plant has not been altered by the mutation. In the hirsutum-parviflorum hybrid with the normal parviflorum cytoplasm the mutation has no effect either. In the reciprocal hybrid, however, the dwarf-effect of the control disappeared (Fig. 2). In the foreign hirsutum cytoplasm the recessive parviflorum gene "resistens" was predominant and caused a new effect on reciprocal differences.

This experiment demonstrates a change of gene-dominance and gene-manifestation under the influence of a foreign cytoplasm.

As a contrast to this case, where only one single plasma-sensitive gene is involved, we shall now consider the gene behaviour in the hybrid with the greatest reciprocal differences. This is a hybrid between the hirsutum strain Jena and the South African hirsutum strain Parys from Natal. The hybrid with Parys cytoplasm is a normal fertile plant, while the reciprocal hybrid with Jena as the mother dies as an embryo of a few cells. Since an investigation of the lethal hybrid by crossing is not possible, we can only investigate the gene-segregation in the normal cytoplasm and then cross the segregants into Jena cytoplasm (see Table I).

Fig. 1.—Reciprocal differences of the hybrid: Epilobium hirsutum Essen x E. parviflorum Tübingen. On the left: the hybrid hirsutum x parviflorum with hirsutum cytoplasm; on the right: the hybrid parviflorum x hirsutum with parviflorum cytoplasm. Fig. 2.—The influence of nuclear genes on reciprocal differences of the Epilobium hirsutum x parviflorum hybrid (see Fig. I). On the left: Control; on the right: Nuclear gene resistens.


F2 and F3 segregants with increasing similarity to E. hirs. Jena

Results: Crosses
Dwarfs Stunted
Normal Plants
(only from
plants like Jena)
  36.2% 55.4% 1.1% 5.3% 1.9% 0.1%

The normal hirsutum Parys x Jena d' hybrid was selfed and back-crossed over two generations. In these crosses there was observed a segregation of all the genes in which the parents differ—genes causing differences in the height of plants, in ramification, in leaf shape and hairiness, in flower size and colour, and so on. In these crosses, however, no embryo lethality, dwarfing, stunting, or pollen-sterility has been found. These plants were then arranged according to their similarity to the two parents and in a fourth generation were crossed into the cytoplasm of hirsutum Jena. In these crosses with cytoplasm from Jena abnormal plants reappeared. After backcrossing with the plasma-foreign Parys-genome and parallel with the similarity to E. hirsutum Parys, the disturbances of development increased. Only segregates from selfings and from backcrosses with the plasma-own hirsutum Jena genome were viable. Only 7 plants which could not be differentiated from the hirsutum Jena parent were completely normal.

The results of these crosses demonstrate that most, perhaps all, nuclear genes of the Parys strain were influenced by the cytoplasm, no matter what their effect in the normal plant was. We must assume that plasma-sensitivity is not a characteristic of special genes. Most nuclear genes interact with the cytoplasm in an unknown manner and can be effective only if the cytoplasmic reaction partner of the cell system is well matched with the nucleus.

I am not going to discuss further details of gene-segregations in these and other crosses. They are similar to what we already know of gene-behaviour, and they offer nothing new.

The Significance of Cytoplasm for the Stability of the Idiotype

We shall now look at the other partner of the genetic cell system, the cytoplasm. When suitable test genomes were crossed into strains of the collection, it was found that the cytoplasms differed greatly. In a search for reciprocal differences depending upon the cytoplasm alone, not only have morphological properties of the plants been investigated, but also physiological reactions such as transpiration, assimilation, respiration, the role of auxin in the development of dwarfs, the redox-potential in the cells, the permeability, the viscosity, and the isoelectric point of the cytoplasm. Great differences have been found which at first seemed to be dependent on the cytoplasm alone. But in all cases in which the nuclear component of the cell has been varied by introducing different nuclei, these nuclei also changed the extent of reciprocal differences. It therefore appears that all the genetic components of the cell induce independent primary reactions, but that most essential reaction-chains of the cell arise through interaction of nucleus and cytoplasm. We cannot expect to find these primary reaction-links earlier in the cytoplasm, as we are only investigating the sum of all cytoplasmic components, the plasmon—and we do not analyse and differentiate the genetic units of the cytoplasm. Since an analysis of the cytoplasm is not possible by exchanging nuclei and cytoplasm as in reciprocal crosses, we must find other methods which correspond to the method of analysing the nucleus by gene-segregation.

Segregation of plasmagenes

It was a favourable coincidence that during the systematic search in the collection several species-hybrids were found with a remarkable instability of the plasmon [11]. Cytoplasms which had been absolutely constant in other crosses for more than 20 years and 20 backcross-generations produced in these species crosses numerous eytoplasmic alterations by plasmon-segregation. The extensive investigation of these plasmon alterations led to the assumption that plasmagenes do not segregate only at the end of the generations, as the nucleus does, but already intra-individually during the cell divisions. Such a behaviour had also been supposed for the segregation of plastids. After the intra-individual segregation had been recognised, mutations of plastids and of plasmon could also be found in mutation experiments with radioactive isotopes. The investigation of these plastid mutations [8, 12] had a great advantage. We could here find morphologically different plastids, which could be followed in their behaviour, and thus the laws of plastid segregation could be derived. If we count the normal and mutated plastids and if we take into consideration the development of the tissue, the relations between normal and mutated plastids correspond to the expectation which can be derived from calculation and model experiments [6, 7, 9]. Through these plastid segregations characteristic patterns arise on the leaves (Fig. 3). If the development of the leaves is exactly known, the shape of these patterns may be determined beforehand through model experiments. Vice-versa, the behaviour and the number of segregating units may be recognised by an analysis of the patterns. These patterns are not limited to the leaves. In many cases lateral branches arise within the spots. Within one plant several very different branches can be produced by cytoplasmic segregation (Fig. 4). These cases are very important, because these branches can be separated and multiplied by cuttings. The branches of one individual can be crossed reciprocally and thus a cytoplasmic inheritance may be proved with certainty. These and other crosses show that the different branches and the different spots of one plant possess the same nuclei and differ only in their plasmon.

Fig. 3.—Intra-individual plasmon-segregation. (a) Segregation of "white" and "green" plastids. (b-c) Leaves of the plasmon-alteration irregulare. During the leaf-development new plasmotypes arose. Fig. 4.—Plasmon alterations of two plants or the Epilobium hirsutum x parviflorum (see Fig. 1, left). (a) Plant with three different rosettes. Plasmotypes stenophyllum, rhytidiophyllum, adaequatum. (b) Plant with three different branches. Plasmotype: longifolium-rhytidiophyllum combination, macrodentatum, rhytidiophyllum.

The frequency of plasmon alteration under the influence of nuclear genes

Table II demonstrates that the plasmon of the E. hirsutum strain Essen is constant in strain crosses. Many species-crosses behave simiIarly. The plasmotype irregulare arises spontaneously with a frequency of 0.04 per cent. By treatment with radioactive isotopes this frequency was increased to 7.65 per cent in the treated plants and decreased in the next progeny to 0.09 per cent [12]. In the presence of the E. parviflorum genome the E. hirsutum Essen cytoplasm becomes unstable and plasmon alterations arise with a frequency of 0-100 per cent depending upon the peculiarities of the introduced genomes or gene-mutations. After the incrossing of different nuclear genemutations, not only the frequency of plasmon-segregation, but also the specificity of the plasmotypes occurring was influenced. Finally it must be mentioned that plasmon segregation only occurs under a certain manner of plant cultivation. At certain temperatures the segregation happens very frequently, while other conditions suppress the segregation completely [5].

TABLE II. Constancy of the plasmon of the Epilobium hirsutum strain "Essen''.

Hitherto it has not been known, in what way nucleus and environment induce plasmon-segregation. As a working hypothesis I assume that in normal genome-plasmon combinations (1) the multiplying of plasmagenes and plasma organelles is correlated with the multiplying of nuclear genes and cell division and that (2) the distribution of cytoplasmic units during cell division proceeds in an orderly fashion. In other genome-plasmon combinations this coordination is apparently broken and a segregation of plasma units is possible by an irregular un-coordinated multiplication and by an irregular distribution. This hypothesis has still to be confirmed by further investigations on Epilobium. From the experiments of Sonneborn on Paramecium, interrelations between the nuclear genes and the kappa particles are well known.

The mutabiity of plastids

Now the interactions between nucleus and cytoplasm must be discussed in regard, to the stability and mutability of the genes themselves. Little is known about the mutability of cvtopiasmuie units (see Table Il). More results exist having to do with plastid mutation. The work of Rhoades on Zea Mays and of Wood du Buy on Mentha reveal that the mutability of plastids can be increased through influence of nuclear genes. In my experiments an increase of plastid-mutations through influence of an altered plasmon was found (Table III). The spontaneous rate for albomaculate plants amounts to 0.08 per cent.

TABLE III. Frequency of albomaculate plants.


Number of
cultivated plants


Spontaneous mutation rate 68,699 0.08
Treatment with 35S    
    < 0.06 mc. 4,626 0.13
    0.12-0.15 mc. 2,958 0.27
    0.23-0.47 mc. 3,252 0.77
Secondary plastid mutations in albomaculate plants 83 albomaculate plants 13.25

This frequency can be increased by treatment with radioactive sulphur to 0.77 per cent. In these experimentally-induced plasmon and plastid mutants the spontaneous mutation rate of secondary plastid alterations is increased to 13.25 per cent [13].

The increase of nuclear mutations in cyloplasmic alterations

Conversely an increase of nuclear mutations seems to be possible through the influence of the cytoplasm. During hybridisation experiments with the cytoplasmic alteration vernicosum it has been found that among only 20 plants with this cytoplasm the labile nuclear gene pallidovariabile arose independently eight times in six plants [3].

If we consider these differing results of the experiments with Epilobium we shall not only find complicated interactions between the nucleus and the cytoplasm in regard to the origin of reciprocal differences of hybrids, but also mutual influences on cell physiology, distribution and augmentation of cell particles and even on stability and mutability of the different genetic units. These different units form a genetically determined reaction system in the cell, each member of which independently propagates its specific properties by identical reproduction, but which react with one another and influence one another. The significance of the nucleus and the chromosomal genes is relatively well known. As to the cytoplasm, investigations have only been started.

The Significance of the Cytoplasm for Differentiation, Determination and Environment

The significance of the cytoplasm has been neglected, because the cytoplasm has always been considered from the standpoint of nuclear inheritance and because the differences of behaviour between nucleus and cytoplasm have not been taken sufficiently into account.

Fig. 5.—Rosette of the plasmon alteration rhytidiophyllum with folded leaf sectors.

To end this report a cytoplasmic alteration will be described which may demonstrate in what fields cytoplasmic inheritance can play an important part. In mutation experiments and during spontaneous plasmon-segregation of the E. hirsutum x parviflorum hybrid, the alteration irrequlare arises frequently. Spots and defects on the leaves are characteristic for irregulare (Fig. 3 b, c). Several types of irregulare have been found which differ in the frequency and in the size of the spots and defects. These differences were inherited by vegetative propagation and in hybridisation experiments by the mother over many generations. This demonstrates a constancy of the cytoplasmic constitution in the meristematic cells. The spots and defects, however, are not physiological patterns. They arise by a new plasmon-segregation during somatic development. From these spots 4-5 new plasmotypes can be isolated by regeneration. One of these new plasmon alterations is the plasmotype rhytidiophyllum (Fig. 5). Reciprocal crosses between irregulare and rhytidiophyllum reveal that the new cytoplasmic differences are inherited by the mother over many generations. If within the leaves spots of rhytidiophyllum arise, these spots are wrinkled and folded. An anatomical investigation of these wrinkled leaves demonstrates that on these wrinkles a new epidermis arises inside the leaf (Fig. 6). This new epidermis possesses stomata and hair-mother-cells, which naturally could not function within the leaf and which only disturbed the growth of the leaves. In this case a newly differentiated cell-type arose by a segregation of plasmagenes [10].

When considering these experiences, it seems possible that the intra-individual segregation of cytoplasmic units in connection with interactions between nucleus arid cytoplasm and with intercellular correlations plays a significant part also in other cases of differentiation and determination, even in these cases, where we have no opportunity of carrying out cross-experiments with the different plasmotypes.

Fig. 6.—Cross-sections through the folded sectors of rhytidiophyllum leaves. During the first cell-divisions of the leaf-development a new epidermis arose inside of the leaf.


Through a systematic search within large collections of Epilobium species, strains, and mutants a number of different nuclei and different cytoplasms has been found, the combination of which demonstrates that reciprocal differences by cytoplasmic inheritance are not caused by cytoplasm alone but by a disturbed balance between nucleus and cytoplasm. A change of gene dominance and gene manifestation under the influence of a foreign cytoplasm could be demonstrated, furthermore mutual influences on the stability and mutability of the different genetic units. The influence of nuclear genes and of intercellular correlations on the intra-individual segregation of cytoplasmic particles may play an important part in determination and differentiation. The different genetic units of the cell form a genetically-determined reaction-system, each member of which independently propagates its specific properties by identical reproduction, but which react with one another and influence one another.

I wish to acknowledge the valuable help of Prof. and Mrs. Sears for their kind correction of the style of the report.


For literature before 1953 see 2 and 4.

  1. MICHAELIS, P., Z. Vererbungsl. 82, 3-13 (1948).
  2. ————— Cold Spring Harbor Symposia Quant. Biol. 16, 121 (1951).
  3. ————— Z. Vererbungsl. 85. 282 (1953).
  4. ————— Advances in Genetics 6, 287 (1954).
  5. ————— Z. Vererbungsl. 86, 101 (1954).
  6. ————— Cytologia 20, 315 (1955).
  7. ————— Züchter 25, 209 (1955).
  8. ————— Protoplasma 48, 403 (1957).
  9. ————— Planta 50, 60 (1957).
  10. ————— Science 126, 261 (1957).
  11. ————— Biol. Z. 77, 165 (1958).
  12. ————— Planta 51 600, 722 (1958).
  13. ————— not published.

Darlington: Cytoplasmic particles in Epilobium (1958)

Sager: Cytoplasmic inheritance in Epilobium (1972)
Cytoplasmic Inheritance