The contribution of bacterial hypermutators to mutation in stationary phase.

IN their discussion of mutation in stationary populaTheir data, and that of Rosche and Foster (1999), show that any given class of mutation is z10 times less tions of Escherichia coli, Bull et al. (2000) conclude that all the mutations are arising in a small fraction of common in single (Lac) mutants than in double mutants (Lac plus a second mutation). For example, Torthe population. They give as their reason the fact that they could fit the data of Torkelson et al. (1997) on kelson et al. found that the frequency of Mal mutations was 31 per 42,617 Lac mutants and 6 per 394 double multiple mutations to a Poisson distribution simply by assuming that all mutations are confined to a subpopulamutants (Lac plus one or more of the other five classes of mutation); similarly, Rosche and Foster found that tion comprising z1 in 10,000–100,000 cells. Their argument is unsound and, furthermore, makes a prediction the frequency of mutations causing a defect in motility (Mot*) was 210 per 3168 Lac mutants and 8 per 13 that is contradicted by their own results. Lac doubles. These results imply that most single They monitored seven classes of mutation: reversion (Lac) mutants are arising in cells that have a lower of a frameshift in the lac operon (Lac), and forward mutation rate than the cells that produce double mumutations leading to 5-fluorocytosine resistance (5-FCR), tants. As Rosche and Foster pointed out, the most rea5-fluorouracil resistance (5-FUR), inability to ferment sonable hypothesis is that all cells are undergoing mutaxylose (Xyl), maltose (Mal), and fructose (Fruc), tion, but a small minority have a much higher mutation and inability to grow on minimal medium at high temrate. In other words, when there is selection for just a peratures [temperature sensitive (ts)]. Unfortunately, single novel trait, most of the survivors will not be burthe target sizes for the seven classes of mutation are dened with multiple changes in the rest of their genome. clearly not the same. In the wild-type strain, the frequency of Lac revertants in the total population (the hypermutators and the rest) is z1026 (Bull et al. 2000, LITERATURE CITED Figure 2A), which is an order of magnitude less than