from the atomic bombs (Radiation Effects Research Foundation [RERF] 1987).
The classical experimental studies on the genetic effects of radiation were conducted in the era before biochemical indicators were conveniently available. When the present study was conceived, in 1970, there were no published data on the production by radiation of germinal mutations resulting in transmitted electrophoretic variants in the mouse, the usual human surrogate in these matters. The present study was thus launched because, on theoretical grounds, it was difficult to believe that some fraction of the mutations resulting from radiation of germ-line cells would not be characterized by nucleotide substitutions. The probability of spontaneous loss of the purine moiety of a nucleotide in a mammalian cell has been estimated at 3×10-11/s (Lindahl 1977). This low probability suggests a remarkable stability for any single nucleotide, but, because there are so many nucleotides in a mammalian nucleus, by this same calculation a mammalian cell should lose by spontaneous hydrolysis approximately 10,000 purines (and 500 pyrimidines) from its DNA during a 20-h generation period. Failure to repair each of these losses exactly would result in a mutation. Radiation might be expected not only to accelerate this spontaneous disintegration of nucleotides, with additional mutations as a result of misrepair, but to impair the repair process as well.
Furthermore, the policy of the genetic studies in Japan has been to examine as many valid indicators of the potential genetic effects of the bombs as possible. “Dogma” has had it that, in contrast to the results with Neurospora (Malling and de Serres 1973), the germinal mutations produced by the radiation of mammals are very predominantly deletions and other major chromosomal events. These should manifest themselves as children with congenital defects and/or decreased life spans. A search for such defects and for impaired survival was a major aspect of the early studies of the genetic effects of the bombs (Neel and Schull 1956). This dogma has been based primarily on the well-documented clastogenic effects of radiation and on the high frequency with which radiation-induced mutations are lethal when homozygous. A further, more recent argument has been that, in a variety of mammalian cell systems, ionizing radiation does not induce mutations to ouabain resistance, a type of mutation believed, on theoretical grounds, to arise from base-pair substitutions (see references in Liber et al. 1987). In our opinion, even if this view-point were essentially correct, it would have been a mistake, in any effort at a comprehensive study of the genetic effects of atomic bombs, to have concentrated only on phenotypes reflecting endpoints known to be especially sensitive to radiation.
There was a further point of principle for undertaking a study directed toward indicators reflecting mutations in single nucleotides. Studies of chemical mutagenesis—conducted some decades after the bulk of the studies on radiation mutagenesis—do demonstrate that the production of mutations resulting in electromorphs (as well as nulls) is an important component of the spectrum of mutations induced in mammalian systems by a variety of chemical mutagens. Lewis and Johnson (1986), in a summary of germinal mutations induced in the mouse by such agents as ethylene oxide, procarbazine, ethylnitrosourea, and triethylene-melamine, find that, in a system in which both loss of protein and/or enzyme activity and a change in electrophoretic mobility could be studied, 17 of the induced mutations were of the former type and 12 were of the latter type. Thus far, only a single spontaneous mutation has been observed in the controls of these experiments; it was an electromorph. In the ultimate, our ability to evaluate the relative genetic risks of the exposure of human populations to radiation and to chemical agents requires comparable data for the two types of mutagens.
Since the present study was initiated, two kinds of data directly relevant to the potential role of electrophoresis in evaluating X-ray-induced mutagenesis in humans have begun to become available. The first entails the molecular basis of the chromosomal lesions that occur spontaneously or are induced by X-rays; the second entails some empirical observations on the offspring of irradiated mice. With respect to the first development, we note that some 20%–40% of the spontaneous germinal and somatic cell mutations at the hypoxanthine phosphoribosyltransferase, adenine phosphoribosyltransferase, thymidine kinase, factor VIII, and Duchenne muscular dystrophy loci can be shown with the Southern blot technique to involve readily detectable genomic alterations, whereas the proportion of such readily detected lesions among X-ray-induced mutations at these loci in somatic cells is more like 50%–60%, the results dif-