ABSTRACT
Homozygous diploid embryos were obtained by using Subtelny’s method of nuclear transplantation of haploid nuclei and selection of the eggs which were delayed in first cleavage.
Genetically uniform clones of homozygous diploids were produced by serial transplantation and their uniformity assessed as tadpoles.
Homozygous diploid nuclear transplant embryos are quite variable within clones, about 20% normal and 80% with the ‘haploid syndrome’. Hetero-zygous diploid transplant embryos are uniform and normal.
It is concluded that lack of heterozygosity leads to developmental insta-bility of which the ‘haploid syndrome’ is an expression.
INTRODUCTION
The problem of the factors involved in the development of the haploid syndrome in anuran embryos is as yet unsolved. It is known that about 90 % of all haploid frog embryos develop the haploid syndrome, which is characterized by the presence of oedema and sluggishness, by reduction in pigmentation and in the efficiency of the heart and circulation, and by a partial failure of the gut to coil and of muscle to differentiate. The two most favoured explanations of the development of the haploid syndrome have been nucleocytoplasmic imbalance, since a haploid nucleus is only half the size of a diploid nucleus in the same-sized egg, and unmasked recessive lethal genes.
There is good evidence that an abnormal nucleocytoplasmic ratio is an important contributory factor in the development of the haploid syndrome. Briggs (1949) compared populations of haploids developing from large and small eggs and Subtelny (1958) compared the development of haploids and homozygous diploids which possessed a reduplicated set of haploid chromosomes. Both these workers found that the proportion of embryos developing the haploid syndrome was smaller when the nucleocytoplasmic ratio approached normal (i.e. in the small-egged haploids and the homozygous diploids) but was not reduced to such an extent that a genetic influence could be ruled out.
Other evidence from grafting experiments (Hadorn, 1936; Baltzer, 1941) has shown that although lethal genes may be unmasked in haploids, they are not necessarily cell-lethal, and Hamilton (1963) also observed that haploid organs may become viable and functional in a diploid environment. The demonstration that the haploid syndrome can be mimicked by treatment of heterozygous diploids with lithium chloride and sodium thiocyanate (Hamilton, 1965), and that similar abnormalities are found in magnesium-starved and nucleolarless Xenopus (Brown & Gurdon, 1964), led to the conclusion that developmental stresses could give rise to the haploid syndrome in both haploid and diploid embryos but that the threshold was very much higher in heterozygous animals.
This phenomenon of heterosis or the stability of heterozygotes has been studied extensively and mention of some of the work is pertinent here. Robertson & Reeve (1952) found that the variance in wing length of hybrid Drosophila melanogaster was less than that of inbred individuals, and they suggested that the greater number of alleles possessed by heterozygotes would allow a greater selection of responses to environmental changes and thus minimize the effect of such changes. Inbred mice are also more variable than heterozygotes in their development (Grüneberg, 1954) and in their response to narcotics (McLaren & Michie, 1954).
The problem of the haploid syndrome can be attacked by producing clones of haploids or homozygous diploids by the method of nuclear transplantation. If recessive lethal alleles are responsible for the haploid syndrome, isogenic haploids or homozygous diploids should be no more variable than isogenic heterozygous diploids, whereas if these two classes of embryos differ appreciably in the variability of their development, then this is best attributed to the degree of heterozygosity, the only respect in which they are thought to differ.
METHOD
The experiments were performed with eggs and embryos of Rana pipiens and Xenopus laevis.
Rana pipiens females were induced to ovulate by injection of fresh pituitaries. Eggs were artificially fertilized and enucleated in order to make androgenetic haploids by the method of Porter (1939). Moore’s modification (1958) of the nuclear transplantation procedure of Briggs & King (1953) was followed and the injected cells were always taken from the animal pole of blastulae. Donor embryos were one of five types: diploid or haploid developing from fertilized eggs, diploid transfer, haploid transfer or homozygous diploid. The latter arose when the recipient of a haploid transfer nucleus was delayed in first cleavage. Since reduplication of any injected nucleus is fairly common it was important to notice whether the time of first cleavage was delayed or not (Subtelny, 1958).
Similar nuclear transfer experiments were performed with Xenopus laevis but in this case the females were induced to spawn by injection of gonadotrophin (Pregnyl, Organon), and fertilization of the eggs, when required, was natural since the male Xenopus could be induced to clasp by a similar injection. The female pronucleus was inactivated in eggs by UV-irradiation (Gurdon, 1960). Both methods of enucleation were 100 % successful.
When normal heterozygous controls reached early feeding stage all animals which had appeared normal at the tail-bud stage were classified into those which were normal and those which were not (‘haploid syndrome’). Many embryos survived to this age but were not scored because they had been abnormal before reaching tail bud.
RESULTS
1. Rana pipiens
The overall results of transplanting the five types of nuclei are summarized in Table 1, and so far as haploid and diploid nuclei are concerned agree very closely with the results of Subtelny’s (1958) experiments on haploid and diploid transplantability. Unfortunately the poor transplantability of haploid nuclei, together with the fact that not every embryo was suitable for scoring, made the production of sufficiently large haploid clones a more difficult task than initially envisaged. The only embryos that could usefully be scored for degree of normal development were those that had appeared to be perfect at the tail-bud stage, for abnormalities at this stage can lead to the development of the ‘haploid syndrome’ in heterozygous nuclear-transplant diploids and it was essential to compare uniformly normal heterozygous embryos with the homozygous clones.
Development of embryos obtained by nuclear transfer in Rana pipiens: a comparison of the transplantability of different types of nuclei

On the basis of Subtelny’s (1958) experiments showing good survival of homozygous diploids, it was hoped that homozygous diploid clones might be easier to produce than haploid clones and that the results might be easier to assess since there was no nucleocytoplasmic imbalance to consider. Table 1 also includes the results of transplanting homozygous diploid nuclei and shows that they survive up to hatching about as well as heterozygous diploids; yet, despite this, only one clone with more than twenty scorable members was produced, and this at the expense of nuclear transfer controls. Unfortunately the produc-tion of a large clone of homozygous diploid embryos by nuclear transfer together with the production of isogenic heterozygous diploids in the same experiment proved impossible. The single large homozygous diploid clone will now be described in more detail. The experiment lasted 7 days and was planned as follows (Table 2). On the first day haploid embryos were made by mechanical enucleation of fertilized eggs when extrusion of the second polar body was visible (Porter, 1939). On the second day one of these haploids became the donor of nuclei for transplantation and those embryos in which first cleavage was delayed were set aside as possible donors of homozygous diploid nuclei for the following day. Only two sibs of this first homozygous diploid donor in the series were haploid at scoring, and a third was diploid (homozygous). On all subsequent days the donor embryos were selected from those that cleaved on time so that further generations were also homozygous diploid.
Out of the 169 eggs injected with homozygous diploid nuclei in this experi-ment, 47 (27 %) reached the age at which they should be scored. Unfortunately many transfer embryos had tended to exogastrulate and because a persistent yolk plug was present at the neurula stage spina bifida developed. This type of defect even developed in heterozygous diploid transfers, no doubt from faulty transfer technique, and in order to avoid complications in scoring all such abnormal embryos were discounted. This meant that a further ten embryos had to be eliminated from the experiment under discussion so that eventually only 37 were scored. Of these 37, 30 had developed the ‘haploid syndrome’ and 7 were indistinguishable from normal heterozygous diploids. The original 169 eggs injected may be subdivided into 99 of one host female on days 2-5 and 70 eggs of a second female on days 6 and 7. The results in this case were 23 oedematous, 4 normal from the first host eggs; and 7 oedematous, 3 normal from the second. There were 5 spinae bifidae or otherwise seriously malformed embryos in each group that were not scored. All the embryos that were scored were genetically identical and yet were not phenotypically identical. Heterozygous diploid controls for this experiment had not been obtained by nuclear transfer, but in all other experiments where heterozygous nuclear transfer embryos were normal at tail bud their subsequent development was normal.
Tail-tip preparations of the embryos were made which confirmed that all supposedly diploid animals were in fact diploid. Figs. 1 and 2 show the experimental animals from days 4 and 5 of the experiment which has been described, together with control haploids and heterozygous diploids.
Camera lucida drawings of Rana pipiens larvae from day 4 of the experiment at the time of scoring. Haploids and heterozygous diploids have developed from fertilized eggs, homozygous diploids by nuclear transfer. Notice that five of the homozygous diploids are oedematous and two display spina bifida.
Camera lucida drawings of Rana pipiens larvae from day 4 of the experiment at the time of scoring. Haploids and heterozygous diploids have developed from fertilized eggs, homozygous diploids by nuclear transfer. Notice that five of the homozygous diploids are oedematous and two display spina bifida.
Camera lucida drawings of Rana pipiens larvae from the 5th day of the experi-ment at the time of scoring. The homozygous diploids produced by nuclear transfer were isogenic with those shown in Fig. 1. In this case two were indistinguishable from heterozygous diploids, four were oedematous, and three display spina bifida.
Camera lucida drawings of Rana pipiens larvae from the 5th day of the experi-ment at the time of scoring. The homozygous diploids produced by nuclear transfer were isogenic with those shown in Fig. 1. In this case two were indistinguishable from heterozygous diploids, four were oedematous, and three display spina bifida.
Other clones of homozygous diploids were produced in the same way and showed similar variation but were composed of fewer individuals; their heterozygous nuclear transfer controls were uniformly normal.
Doubling of the transferred nucleus is a fairly common occurrence (Briggs & King, 1957) and many of the homozygous diploid transfers were delayed in entering first cleavage. However, none of the homozygous tetrapioids so produced survived to scoring.
2. Xenopus laevis
Essentially the same results were obtained in the nuclear transfer experiments using Xenopus. The overall results are set out in Table 3. The greatest handicap in clone production in Xenopus is the need to use a different host every day and the rarity of Xenopus which lay good recipient eggs. No clones of sufficient size were produced but even so there was variation within the small clones. Fig. 3 shows the development of the ‘haploid’ syndrome in some members of a homozygous diploid clone while one remains free from it; the heterozygous clone is uniformly normal. In Xenopus again, none of the eggs with delayed first cleavage which should be homozygous tetrapioids survived to scoring.
Development of embryos obtained by nuclear transfer in Xenopus laevis: a comparison of the transplantability of different types of nuclei

Camera lucida drawings of one group of nuclear transplant homozygous and heterozygous diploid Xenopus to show their development. The ‘haploid syndrome’ develops in two out of the three scorable homozygous diploids.
DISCUSSION
Subtelny (1958) pointed out in his discussion of the development of homozygous diploids that although they possessed a duplicate set of haploid chromosomes and displayed normal early development they all had deficiencies in later embryonic or larval development. Two-thirds of them developed the ‘haploid syndrome but if recessive lethals were its cause the proportion should be the same as in haploids (90-95 %). Subtelny was thus reluctant to support the idea of ‘genetic influence on the expression of these deficiencies’ but at the same time he said that the differences between homozygous and heterozygous diploid development ‘must be due somehow to the nuclear condition of the homozygous diploids’. This should also be a nuclear condition shared by haploids and I suggest it is lack of heterozygosity.
The work of Robertson & Reeve (1952), McLaren & Michie (1954) and Griineberg (1954) already referred to presents general evidence not only for the phenomenon of heterosis but also for greater variation within populations of homozygotes than heterozygotes. These may be two sides of the same coin, for if we assume that a species is fitted to its environment its individuals must respond to changes in the environment by deviating as little as possible for the norm. In any case highly inbred mice and fruit-flies and homozygous diploid amphibians cannot be considered to be genetically normal members of their species, so there is no reason for expecting them to be phenotypically normal.
My suggestion is that the haploid syndrome is not specific but may develop in any embryo in response to conditions to which it cannot adapt. For instance, heterozygous diploid Xenopus develop the ‘haploid syndrome’ if they are treated with sodium thiocyanate or lithium chloride (Hamilton, 1965), are reared in magnesium-free medium, or are nucleolarless (Brown & Gurdon, 1964); under ordinary conditions, however, the tadpoles are uniformly normal because of the developmental stability afforded by heterosis. Homozygous diploids, on the other hand, are developmentally less stable and their responses to stress less predictable, so that under normal laboratory conditions some may develop the ‘haploid syndrome ‘. One also finds, as one might expect, that haploids are more sensitive to lithium chloride and similar treatments than heterozygous diploids (Hamilton, 1965).
The extreme sensitivity of homozygous embryos can explain why no homo-zygous tetrapioids and so few haploids survived nuclear transplantation in the present experiments. The possibility that chromosome damage or loss during manipulation of nuclei causes the development of the haploid syndrome is unlikely for two reasons. First, members of heterozygous diploid clones do not develop the haploid syndrome despite the fact that they are produced by the same technique that is used for producing homozygous diploid clones; and, secondly, some normal embryos did develop from each day of serial trans-plantation and the variability of transfer clones was as great at the end of an experiment as at the beginning.
The real crux of the matter is that clones of genetically identical homozygous diploids do not develop uniformly, and those individuals that are abnormal develop the ‘haploid syndrome’. Lack of heterozygosity is the nuclear condition that haploids and homozygous diploids have in common. By rendering them less stable genetically this makes them more susceptible to stresses, such as un-masked deleterious genes and nucleocytoplasmic imbalance, and thus reduces their chances of developing normally.
ZUS AMMENFASSUNG
Die Rolle des Genoms bei der Entwicklung des Haploid-Syndroms in Anura
Man erhielt homozygote diploide Embryonen durch Kerntransplantation auf haploide Nuclei und durch Auslese der Eier, welche in der ersten Teilungs-phase zuriickgeblieben waren.
Genetisch einheitliche Klone der homozygoten Diploiden wurden an hand von Serientransplantation erzielt, und ihre Identitat an Kaulquappen gemessen.
Homozygote Diploide sind recht variabel innerhalb der Klone : circa 20 % sind normal und 80% zeigen das ‘Haploid-Syndrom’. Heterozygote Diploide sind einheitlich und normal.
Es wird gefolgert, dass ein Mangel an Heterozygozität zu Entwicklungs-labilitat fiihrt, die seinen Ausdruck im ‘Haploid-Syndrom’ findet.
ACKNOWLEDGEMENTS
I wish to thank Dr J. A. Moore for guidance and hospitality at Columbia University, New York, during the work on Rana pipiens’, and the National Science Foundation for grant G 9001. I am also grateful to the Wellcome Trust for a travel grant and the Nuffield Foundation for support in London.