Genetics 18(2): 223-232 (1927)
THE HYPOTHESIS OF CHROMOSOME AFFINITY AND THE PHENOMENON OF SUPPRESSION OF CHARACTERS ON CROSSING
SEVERAL papers have appeared recently which try to prove that some linkage phenomena may be caused by an association between particular chromosomes.
In. the cytological literature we frequently find data concerning linkage of chromosomes. These linkage phenomena may be conditioned either by material connections between particular chromosomes, or by an unknown affinity which causes the chromosomes to show a tendency to go in groups to each pole of the dividing cell. S. Nawashin(16), R. R. Gates(5), Ö. Winge(23,24), Z. Wóycicki(25), R. E. Cleland(2) and other authors point out the existence of material connections between the chromosomes of certain plants. F. Schrader(19) shows that in some Pseudococcus species there are two groups of chromosomes, and that the perfectly definite distinguishing characteristics of each of the two chromosome groups make it possible to follow them through the entire division, and it is plain that the clumped group acts as a unit and that all its members go to one and the same pole." Schrader does not, however, figure any material connections between the chromosomes of those groups.
Cleland(2) points out another kind of affinity which leads to the same results. According to him "every chromosome seems to have some sort of affinity for a particular pole, otherwise the picking out of alternate chromosomes to be sent to the same pole is inexplicable. Each chromosome must go to one pole rather than the other, and as this applies to every chromosome in the cell, it naturally follows that in every cell the same chromosomes will go together to the poles. The effect will be the same as though the chromosomes were actually bound together structurally in such a way as to form one pair of large compound chromosomes."
A. B. Stout(20) describes in Carex an interesting case of arrangement of chromosomes in series. This serial arrangement is not lost in the division stages. The daughter chromosomes are arranged in a series which remains in evidence during the resting stages.
R. R. Gates(5) believes that the coalescence of chromosomes, observed by him in Lactuca, may furnish a basis for the phenomena of linkage as distinct from those based upon relations between the two members of a given pair of chromosomes.
The idea as to the possibility of existence of chromosome coupling as a basis for linkage phenomena was suggested also by Rennet in 1917 ((17) footnote, p. 248).
In 1925 several papers were published in which we find genetical arguments for the possibility of the existence of such a higher order of linkage. I myself published four papers during that year(10-13) concerning this phenomenon in wheat, and the arguments given in those papers were repeated by me in vol. XVII of this Journal(14).
In the same year, 1925, O. Renner published a paper(18) in which he discusses the possibility of the existence of "Chromosomenkoppelung" in Oenothera, and brings forward the following arguments in favour of this supposition:
(1) If the linked factors (for instance the factors M, N, P, S, R) were located in one chromosome "das fragliche Chromosom müsste also sehr lang sein, wenn es drei oder noch mehr Genen die Möglichkeit geben sollte sich als selbstandig zu gebarden."
(2) If the linked factors were located in one chromosome the mechanism of crossing-over would be different from that in Drosophila because in the cases observed by Renner "die Koppelung bald mehrfach, bald einmal, bald gar nicht gebrochen wird, je nach dem Grad der Aehnlichkeit des Partnerchromosoms."
Recently Hurst(8) has put forward some evidence for the existence of superlinkage in Rosa. The fundamental number of chromosomes in Rosa is 7. Somatic numbers are diploid 14, triploid 21, tetraploid 28, pentaploid 35, hexaploid 42, or octoploid 56. Hurst believes that "the significance of the septuple numbers of chromosomes in Rosa is apparent in various stages of gametogenesis and also in some somatic divisions in diploids and polyploids, in which it is evident that the chromosomes are working in sets of seven or septets." When comparing the taxonomic characters of the species in the living collection at Kew Hurst was struck by the fact that the tetraploid species showed the combined characters of two distinct diploid species, while the hexaploid species showed the combined characters of three distinct diploid species, and the octoploid species showed the combined characters of four distinct diploid. species.
In mice W. H. Gates(6) found some characters which show an association apart from linkage phenomena in Morgan's sense; for there was a tendency for the entire group of characters derived from each parent to associate together. This association appears only in specific crosses, and is believed by the author to be due to a grouping of the chromosomes.
Phenomena of this nature appear to be not uncommon, and the hypothesis of "superlinkage" has been supported by means of more or less numerous genetic arguments by several authors almost simultaneously.
It is probable that the affinity between certain chromosomes may be greater than between others. The existence of different degrees of affinity between particular chromosomes may furnish the basis for the phenomenon of exchange of elements between two sets of chromosomes of different origin. Cleland(2) believes that such a phenomenon exists in Oenothera and he calls it "chromosome exchange." If this really happens we should expect that in the F2 generation from certain crosses one or both parental types would not appear at all. Possibly in this way we might explain some cases of suppression of characters on crossing, and so throw light upon the so-called "Wichura" type of segregation.
Suppose for instance that we have two types differing from one another in three cumulative factors, and that these factors are each located in a different chromosome. We have thus three chromosomes A, B, C in one parent and three respectively homologous chromosomes a, b, c in the other one. The genetic constitution of the first parent will be AABBCC and of the second one aabbcc. Let us suppose further that two of these chromosomes, A and B, are linked with, one another, and that the chromosomes a and b are also linked. The precise nature of the linkage does not matter. It may be structural, or there may exist a sort of chemical affinity between the chromosomes. If this affinity be greater between chromosomes A and b than between A and B, then in F1 the normal reassortment of A and a with B and b will be upset and we shall obtain the new permanent combinations Ab and aB. If chromosomes C and c are independent of the others the F1 hybrid will produce the following four kinds of gametes: AbC, aBC, Abc, aBc. In F2 we shall obtain various intermediate types, as shown on. Diagram 1, but the parental types, AABBCC and aabbcc, will not reappear. In the F2 generation we shall obtain three kinds of individuals, according as they contain 2, 3, or 4 factors; but none will possess 0 or 6 factors, i.e. there will be no parental combinations in F2. This is what Vavilov(21) speaks of as the Wichura type of segregation.
Wichura(21) worked with the hybrids of Salix species, and obtained in his crosses an intermediate F1 generation without apparent segregation in F2. From the photographs reproduced in Heribert-Nilsson's paper(7) it would seem that the parental types did not reappear in the F2 generation from Salix viminalis x S. caprea. Individuals resembling S. viminalis did not appear at all. One single F2 individual noted by Heribert-Nilsson as possessing "ganz caprea-ähnliche Blätter" cannot be identified with S. caprea in respect of the width of the leaves. Heribert-Nilsson made a back cross (viminalis x caprea) x viminalis and obtained in the subsequent generation "ganz viminalis ähnliche Blatter." But these "viminalis ahnliche Blätter" were not identical with the leaves of the viminalis parent in respect of width.
F. L. Engledow(4) has described the phenomenon of "shift" for the shape of the glumes in a Triticum polonicum x Tr. durum cross. F1 was intermediate. "For F2 the frequency-distribution of glume-length was very clearly trimodal, and its form suggested simple segregation on the 1:2:1 basis." Engledow examined a large material, making over 10,000 measurements of glume-length and following his investigations up to the F3 generation. He found that the mean glume-length for polonicum is in F2 "shifted" down by more than 20 per cent., that the shifted value breeds true, and is again exhibited by F3 polonicum ex intermediate F2 forms. For durum "a corresponding small upward shift seems to have occurred." Engledow writes ((4), p. 93) that "the parental polonicum does not reappear in F2. In its place are found polonicum plants which closely resemble parental polonicum in general appearance, but whose mean glume-length is more than 20 per cent. lower than that of P (parental polonicum). This 'shifted' form, when selfed, breeds true."
There is a tendency among geneticists to believe that segregation phenomena like those in the Salix crosses may be explained by means of the multiple-factor hypothesis. Non-appearance of parental types in the F2 generation is usually believed to be due to a very great number of factors conditioning the differences between these types. Engledow, however, has given much evidence which favours the view that in the polonicum x durum crosses it is "impossible to devise a suitable multiple-factor explanation," or at least that any such explanation is a problematical one. He says that "'Multiplying factors' have been suggested by some in explanation of the phenomena but no feature of these results appeared to lend any encouragement to that idea" ((4), p. 81).
It is difficult also to explain a "shift" of the type of the apical tooth of durum wheat observed by VaviIov(21) in a Triticum persicum x Tr. durum cross by the hypothesis of cumulative factors. Vavilov reports that in the F2 generation no plants similar to Tr. durum were observed, nor were any found reproducing exactly Tr. persicum in glumes and teeth. A detailed investigation of over 4000 F3 plants revealed no individuals with glumes of the type of Tr. durum. "This cross," says Vavilov(20), "was perfectly fertile and there could be no missing series of forms as it happens with distant crosses." Vavilov concludes that the explanation of this case by means of the multiple-factor hypothesis "is practicable if we admit a very great number of factors conditioning the shape of the glumes." But, as he says later on, "it is possible that in this case the scheme of cumulative factors hypothesis can not be applied at all."
|1I suggested this idea in one of my former papers (15).|
If we assume that in crosses with a Wichura type of segregation the phenomena of chromosome exchange1 occur with the production of new permanent chromosome compounds, then with such restrictions the multiple factor hypothesis will be applicable to this type of segregation.
In the Triticum polonicum x Tr. durum cross it is possible that only two pairs of allelomorphs are involved.
|2A certain analogy exists between the factor B and the factor increasing the glume-length which I observed in my polonicum x dicoccum crosses (14). But the factor increasing the glume-length in my experiments exerted a much more pronounced influence upon the polonicum glumes than upon the glumes of dicoccum. This factor, however, did not show any linkage with other factors.|
Let us suppose that the glume-length of polonicum is determined by two factors A and B, one of which, A, produces the essential glume-length characteristic of Tr. polonicum, while the second, B, increases this essential length by more than 20 per cent. The allelomorph of A will be a, which determines the length of durum glumes. The heterozygote Aa is intermediate. The factor B, which increases the glume-length of polonicum by more than 20 per cent., is scarcely appreciable on the durum type2.
The factors A and B are located in different chromosomes which are linked with one another. The chromosomes a and b of Tr. durum are also linked. In F1 new permanent combinations will occur, viz. Ab and aB, and the F1 individual will produce only two kinds of gametes, viz. Ab and aB. In F2 we shall get three types of individuals, namely: AAbb, AaBb, aaBB in the proportion of 1:2:1. The type AAbb will be polonicum "shifted" clown by more than 20 per cent., the type aaBB will be durum with slightly increased glume-length and the type AaBb will be intermediate like F1. This type, according to Engledow's statement, "remains singularly constant from generation to generation."
In the case of Engledow's experiments the formation of a new permanent chromosome association is rendered more probable owing to linkage between several characters. Engledow points out that in the "shifted" polonicum "grain-length exhibited shift on lines precisely corresponding to the case of glume-length and a group of ten characters—five of the glume and five of the grain—behaved as `genetic inseparables’” ((4), p. 81). It is possible therefore that the chromosome carrying the factor B contained also some other factors influencing grain and glume characters, and that all these factors were conveyed together from one chromosome group to another.
Some cases of the Wichura type of segregation are known in which one of the parental types undoubtedly appears in the F2 generation whilst another one does not appear at all. Such a case was observed by E. M. East (3) in Nicotiana Langsdorfii x N. alata crosses in relation to flower shapes. East writes as follows: "Individuals reproducing the N. Langsdorfii were found in the F2 generation .... Certain of these F2 individuals reproduced N. Langsdorfii populations in the F3 generation….No F2 individuals reproducing N. alata were found, but F3 plants approaching such a type were produced." The F1 generation was intermediate, nearer however to N. Langsdorfii. East concludes as follows: "Though extremes like each parent were not produced, it is hardly possible to see any other cause for this great difference in variability than segregation and recombination of Mendelian factors." But East does not propose any definite Mendelian. scheme of segregation, and Engledow(4) believes that the multiple-factor action "is perhaps...a matter of uncertainty...even in such careful investigations as those of East."
Possibly such cases may be explained by means of cumulative factors which in the heterozygous state produce almost the same effect as in the homozygous one.
I found some evidence for the existence of such factors in my crosses between varieties of Phaseolus vulgaris(9). In the crosses Hinrich's Riesen x Bagnolet and Flageolet rouge x Bagnolet I found that the seeds of F1 are as long as the seeds of Bagnolet, which was the largest parent in regard to the shapes of seeds. In F2 a segregation occurred; both parental types appeared but the Mode of the polygon of variation of seed-length was the same in F2 as in Bagnolet. I suggested that there are several cumulative factors increasing the basal seed-length. I assumed this basal length to be approximately 10 mm., and that each factor increased this basal length by approximately 2 mm. I assumed also that each of those factors in the heterozygous state increased the length of the seeds by the same or almost the same amount.
The case observed by East in Nicotiana might possibly be explained in the same way, with, however, one essential difference, viz. the occurrence of a "chromosome exchange."
Let us suppose that in a case of seed-length we have three cumulative factors A, B, C, each located in a different chromosome, and that A and B are linked. Suppose the basal length to be 10 units, and that each factor increases it by 2 units, the seed-length being 16 units in one of the parental types and 10 in the other. The genetical constitution of one parent will be AABBCC (chromosomes A and B are linked) and of the second one—aabbcc (chromosomes a and b are linked). If the supposed affinity between the chromosomes A and B is less than between A and b, then in F1 we shall obtain new permanent combinations Ab and aB instead of AB and ab. The third chromosome being independent we shall get the following kinds of gametes: AbC, aBC, Abc, aBc. In F2 we shall obtain three kinds of individuals, some of which will contain two factors, and others three or four factors. But the parental combinations will not occur (Diagram 2).
Since we are assuming that each factor here produces as great an effect in the heterozygous state as in the homozygous one we should in such a case obtain only three classes of individuals as judged by their external appearance, viz. showing 16 units of length (10 + 6), or 14 units (10 + 4) or 12 units (10 + 2). The larger parental type would appear in F2 but not the smaller.
This scheme may possibly be applied to the case of Nicotiana. We must assume, however, that instead of dominant factors increasing the shapes there are dominant factors decreasing the shapes; and we must further assume a larger number of factors.
Perhaps, too, the hypothesis of chromosome affinity will help to elucidate some aberrant cases of Mendelian ratios to which Bateson called attention in his paper on Segregation(1); and we may now examine two theoretical possibilities resulting from "chromosome exchange," without, however, definitely relating them to concrete cases of "aberrant inheritance,"
Let us suppose that two factors A and B determine red colour. These two factors are located in different chromosomes which are linked with one another. After crossing a red (AB) individual with a white one (ab), we shall obtain an F1 of the genetic constitution AaBb. The F1 individuals will be red because A and B are dominant factors. If the affinity between the chromosomes A and b is greater than between A and B, then we shall obtain in F1 two kinds of gametes, namely Ab and aB and in F2—the combinations BBaa (white): BbAa (red): bbAA (white) in the proportion of 1:2:1. Thus the number of white individuals will be the same as that of red individuals and, in such a way, instead of expected ratio 3 red : 1 white; we shall obtain the ratio 1 red : 1 white.
Let us suppose now that A and B are not dominant factors and that the F1 generation is intermediate in respect of colour. Then in the F2 generation instead of red and white individuals in the ratio of 1:1 we shall obtain intermediate and white individuals in the ratio of 1:1. In the last case red individuals will not appear at all in the F2 generation.