Experimental evidence for an alternative to directed mutation in thebgl operon

THE directed mutation hypothesis1–6 suggests that some mutations occur more often when selectively advantageous than when neutral or disadvantageous, challenging the principle that the selective value of a mutation does not affect the rate of its occurrence7–11. Mutations in the bgl operon ofEscherichia coli have been reported to be a case of directed mutation2. E. coli K12 strain Ξ342LD cannot grow on salicin but derivatives with two mutations in the bgl operon, an excision of IS 150 (formally called 1S103; ref. 12) 500 T from bglF and a point mutation or insertion in bglR 13–15 ,grow rapidly on this sugar. When x342LD is grown on a medium containing salicin, bglF excision mutants accumulate to a frequency of >1%, even though these mutants are reportedly2 unable to grow on salicin, and Sal+ double mutants subsequently attain a high frequency. Comparable accumulations of excision mutants and Sal+ double mutants are not observed in the absence of salicin. As salicin is not mutagenic, it has been suggested that excision mutations in bglFmight serve only to create the potential for a secondary selectively advantageous mutation2. We show here, however, that these double mutants can be accounted for by spontaneous mutation to intermediate genotypes in non-growing populations, coupled with slow growth of some of these intermediates on salicin, which enables their populations to reach a size where secondary mutations allowing rapid growth on salicin become common.