Latent mutants in chemostats.

HE correlation of spontaneous mutation with active growth in cell culture Thas led in the past to a common tacit assumption that division is required for mutation, an assumption exposed by the expression of mutation rates as some function of division. A ready explanation was provided by the error (copy-error) hypothesis which views mutation as an error in gene duplication, and which predicts therefore, that mutation rate is proportional to the rate of cell division. However, in measurements of the spontaneous mutation rates of bacterial cultures in chemostats in which growth was limited with tryptophan, NOVICK and SZILARD (1950) found that spontaneous mutation to resistance to bacteriophage T5 appeared independent of growth rate. Attempts were made to reconcile this result with the error hypothesis by assuming compensatory processes ( NOVICK 1955), but the absence of experimental confirmation of the latter has caused their finding to be regarded as paradoxical, rather than fundamental, in most reviews on mutation. On the other hand, when glucose is used to limit growth rate the accumulation of caffeine-induced mutants is proportional to growth rate (KUBITSCHEK 1960), seemingly providing prima facie evidence for the error hypothesis. Nevertheless, the error hypothesis is shown in this report to fail to account satisfactorily for newly-forming mutants when glucose limits growth. Furthermore, the constancy of mutation rate when tryptophan limits growth is not an isolated case peculiar to spontaneous mutation but occurs also for caff eine-induced mutations. An examination of the nature of the difference between results from tryptophan-limited and glucose-limited cultures has been made possible by the development of a method of delayed selection that permits, for the first time, measurement of concentrations of cells which are latent mutants for T5 resistance in the chemostat in the absence of accumulated mutations arising spontaneously after plating (plate mutants). Since plate mutant concentrations are insignificant, this method allows simultaneous measurements of concentration of expressed and of latent mutants (those cells in which the mutational process has been initiated, but which have not yet expressed the mutant phenotype), making possible a study of the kinetics of mutation. Previously, with the time-course measurement of the appearance of mutant colonies introduced by DEMEREC (1946) it had been necessary to assume latent mutant concentrations as a difference between the maximum yield and those mutations already expressed. Correspondingly, there has been no clear determination of latent mutant concentrations in later modifica-