The cloning of a frog

In 1962, a paper examining the developmental capacity of nuclei taken from intestinal epithelial cells of Xenopus laevis tadpoles was published (Gurdon, 1962) (Boxes 1, 2). This paper, which used the technique of nuclear transplantation originally described by Briggs and King (Briggs and King, 1952), showed that the nuclei of some intestinal epithelial cells from feeding tadpoles, after transplantation into enucleated eggs, could develop into normal feeding tadpoles. Several years later, the tadpoles produced in this work had matured into frogs and had been tested for sexual maturity and fertility. A subsequent paper (Gurdon and Uehlinger, 1966) was based entirely on the animals that had been produced and described in the 1962 paper. Below, I highlight some interesting features and outcomes of these early experiments.

In all experiments, the transplanted nuclei carried the nucleolar genetic marker (Elsdale et al., 1960), which was crucial for these experiments to be accepted by the scientific community. Intestinal epithelial cells from feeding tadpoles (Box 3) were used as donors of nuclei for transplantation into unmarked enucleated recipient eggs, and these were interspersed with experiments using nuclei taken from early embryo (blastula or gastrula) embryos. The proportion of total transplant embryos that reached a feeding tadpole stage was 1.5% when using intestinal nuclei as donors but 36% when using early embryo nuclei (Box 4). It was therefore already evident, at this time, that, as cells differentiate, their nuclei are a great deal less able to support development within enucleated eggs than are nuclei from embryonic cells. This reflects the phenomenon of ‘resistance’ that is so evident in current induced pluripotent stem cell (iPSC) experiments.

To investigate the reason for this decline in the success of nuclear transfers, the fate of transplanted nuclei in eggs was investigated by wax embedding and serial sectioning. In some cases no transplanted nucleus was found at all and this could have been because the nucleus was broken or pulled out of the egg as the injection pipette was withdrawn. Therefore, technical difficulties accounted for some of the failures. Furthermore, previous work had shown the desirability of breaking the donor cell to the least extent possible, thereby enabling it to be protected by donor cell cytoplasm. Therefore, in these experiments, I used the least amount of donor cell distortion that gave a reasonable proportion of broken donor cells. A number of the recipient eggs may therefore have received unbroken donor cells, which cannot participate in development.

It was also known from work in Rana pipiens (Di Berardino and King, 1967) that transplanted somatic nuclei often undergo severe chromosome breakage. This has been attributed to incomplete DNA replication after nuclear transfer. DNA synthesis normally requires ~6 hours in somatic cells but has to be completed within 1 hour of nuclear transfer to eggs. This is because eggs always divide into the two-cell stage 90 minutes after activation (which results from a micropipette injection). A failure to make the transition from a slow to a fast DNA replication cycle can therefore account for some of the nuclear damage sustained after transfer to eggs and, hence, for the abortive cleavages often observed with nuclei from differentiated cells.

Another common consequence of transplanting nuclei from differentiated cells is the formation of partially cleaved embryos, which either die prior to gastrulation or form abnormal gastrulae. Typically, these have normal-looking blastomeres in about half of the embryo while about half are undivided. A favoured interpretation of this is that somatic nuclei often fail to complete their DNA replication by the time the egg divides into two. It is thought that, in such cases, the whole replicating transplanted nucleus moves into one of the first two blastomeres where it can then further complete its DNA replication, while fertilised eggs continue carrying out their second replication, from two-cell to four-cell stages. The other blastomere hence receives no nucleus and so remains undivided.

To test this idea, serial nuclear transplantation was carried out in which normal cells from a partially cleaved embryo served as donors of nuclei for transplantation to another set of recipient eggs. The results published in the 1962 paper were quite striking. The first important observation is that the serial transfer of nuclei from partially cleaved embryos very often gave completely cleaved embryos and, hence, complete blastulae. In turn, many of these developed much further, even forming, in some cases, normal feeding tadpoles (Box 5). The conclusion that I believe to be correct is that the initially transplanted nuclei were spared from chromosome damage by having a second chance to complete their DNA replication as the recipient egg divided into two cells. Only when that had happened did the originally transplanted nucleus have to start division as one of the first two blastomeres divided into two cells at the time that control embryos from fertilised eggs were dividing from the two- to four-cell stage. In this way, I believe that the true genetic potential of an originally transplanted somatic nucleus can be revealed especially well when the initial first transfer embryo divides into a partial blastula. It was noticed that subsequent serial nuclear transfers up to a third or fourth generation did not yield any further improvement in the normality of development ultimately achieved. In some cases, as many as ten serial transfer experiments were carried out, a physically demanding kind of experiment requiring re-transfer every 12 hours for 5 days. There was therefore no case for believing that the developmental capacity of a transplanted nucleus can increase with serial transfer. Rather, the full genetic potential of such a somatic nucleus is often revealed most clearly by a serial nuclear transfer experiment (Gurdon, 1960).

The major conclusion from these nuclear transplant experiments was that two-thirds of the nuclei transplanted from intestinal epithelium cells could generate embryos that reach the muscular response stage (Box 6). This stage of development indicates normal muscle and nerve function; the muscle and nerve lineages are entirely unrelated to that of the intestine, which comes from the endoderm, highlighting that a nucleus is able to promote the formation of a differentiated cell type whilst still retaining the genetic information required for the formation of other differentiated somatic cell types in a normal tadpole.

The normal tadpoles derived from these early experiments were taken to Geneva where my supervisor Michail Fischberg moved to take up the professorship of Zoology. He and his assistant Vreni Uehlinger cared for my tadpoles during my post-doctoral period in USA. They grew them up to adult frogs and returned them to me in Oxford for fertility testing. This led to the 1966 paper (Gurdon and Uehlinger, 1966) reporting ‘fertility’ of intestine nuclei and, of course, to many more studies that examine the concept of reprogramming.

I thank especially my PhD supervisor Dr Michail Fischberg, and members of Dr Fischberg's group in the 1960s.