Recombination-Dependent Mutation in Escherichia coli Occurs in Stationary Phase

“ADAPTIVE” or “stationary-phase” mutation occurs in apparently stationary-phase populations of cells, over time, after exposure to a nonlethal selection. In one assay system, Escherichia coli with an amber mutation in lacZ forms Lac+ colonies over time when plated on medium with lactose as the sole carbon source (Cairns et al. 1988). These Lac+ mutations had been postulated to form after exposure to, and in response to, the selective medium (Cairns et al. 1988), but in an article in Genetics last year, most of these were shown to be slow-growing mutants that formed during preselective growth of the culture (Prival and Cebula 1996). For the lac amber reversion assay system, the hypothesis of mutations occurring by a different mechanism after plating is not supported. Is this generally the case? In a well-characterized system for studying presumed stationary-phase mutation, E. coli with a lacI-lacZ fusion gene with a +1 frameshift mutation appears to revert to Lac+ over time after exposure to minimal medium with lactose as the sole carbon source (Cairns and Foster 1991). In this system, the late arising Lac+ mutants form via a different molecular mechanism from the early mutants (reviewed by Bridges 1997; Rosenberg 1994, 1997; Rosenberg et al. 1995, 1996, 1998). The late appearing mutations (1) require proteins of the RecBCD recombination system to form (Harris et al. 1994, 1996; Foster et al. 1996); (2) possess a unique mutation spectrum mostly of −1 deletions at mononucleotide repeats (Foster and Trimarchi 1994; Rosenberg et al. 1994); (3) arise in a hypermutable subpopulation of cells whose whole genome is susceptible to mutation (Torkelson et al. 1997); and (4) occur in cells in which postsynthesis mismatch repair activity is diminished transiently at the level of limiting MutL protein (Longerich et al. 1995; Harris et al. 1997; Rosenberg et al. 1998). Early arising mutants have different sequences, are rec and ruv gene-independent, and form under conditions in which mismatch repair protein MutL is not limiting. The late mutants arise by a process mechanistically distinct from that giving rise to early Lac+ revertants. However, this does not establish that this novel mechanism is specific to stationary phase or occurs while cells are under selective conditions. The finding of Prival and Cebula (1996) that slow-growing mutants make up the late-arising lac amber revertants raises the possibility that the recombination-dependent lac frameshift revertants are also slowly growing mutants that arose during growth (i.e., a subset of all growth-dependent mutants). We tested this possibility using the lac frameshift reversion assay system (Cairns and Foster 1991) with a reconstruction experiment similar to that employed by Prival and Cebula (1996). Lac+ early and late mutants were obtained in a typical stationary-phase mutation assay (e.g., Cairns and Foster 1991; Harris et al. 1996) in which lac frameshift-bearing cells were plated in the presence of a 25-fold excess of scavenger cells on minimal lactose medium. Scavenger cells are E. coli cells with all the lac genes deleted and are used to remove any contaminating carbon sources in the medium. One colony that appeared on day 2 and 1 on day 5 were isolated from each of 26 independent cultures and purified by streaking on minimal lactose medium three times. These were regrown in minimal glycerol medium (Harris et al. 1996) and plated at approximately 102 Lac+ cells per plate in the presence of greater than 109 scavenger cells, mimicking the original conditions under which Lac+ mutants were isolated. If late mutants form on the plate during starvation, then the late-arising Lac+ colonies might require only 2 days to form a colony, the usual length of colony formation for Lac+ E. coli on minimal medium under these conditions. However, if the late-arising Lac+ mutants are slow-growers, then they may require 3 to 5 days to form a colony when plated with excess scavenger cells on selective medium. The majority of mutant cultures produced most of their Lac+ colonies on day 2 (Table 1). Under growth situations identical to the original experiment, all but one early and one late-arising Lac+ mutant required Length of colony formation of independent early and late Lac+ mutants in reconstruction experiments Independent early Lac+ revertants of the lacI33 frameshift-bearing strain (rec+; Cairns and Foster 1991) and its Δ recA derivative SMR624 (Harris et al. 1994) were isolated on day 2, and late revertants were isolated on day 5. One early and one late mutant was isolated from 26 independent cultures from an adaptive mutation assay (methods of Harris et al. 1996). Each independent mutant was purified and regrown in M9 glycerol medium to saturation, washed three times in M9 medium with no carbon source, and plated with greater than 109 scavenger cells (Cairns and Foster 1991) to give approximately 102 Lac+ colonies/plate on M9 lactose plates. Colonies were counted on days 2–8. The percentage of colonies arising on days 2, 3, and 5 are shown, the total number of colonies per culture being approximately 102 but never less than 23. These isolates produced 57 and 67% of colonies by day 2. This isolate produced 19% of colonies by day 2. This isolate was unusual in displaying only 28% viability on minimal lactose medium as compared with complex medium, and of those colonies that arose on minimal lactose medium, 13, 18, 47, and 82% arose by days 2–5, respectively. 92% of the colonies first appeared on day 5. Length of colony formation of independent early and late Lac+ mutants in reconstruction experiments Independent early Lac+ revertants of the lacI33 frameshift-bearing strain (rec+; Cairns and Foster 1991) and its Δ recA derivative SMR624 (Harris et al. 1994) were isolated on day 2, and late revertants were isolated on day 5. One early and one late mutant was isolated from 26 independent cultures from an adaptive mutation assay (methods of Harris et al. 1996). Each independent mutant was purified and regrown in M9 glycerol medium to saturation, washed three times in M9 medium with no carbon source, and plated with greater than 109 scavenger cells (Cairns and Foster 1991) to give approximately 102 Lac+ colonies/plate on M9 lactose plates. Colonies were counted on days 2–8. The percentage of colonies arising on days 2, 3, and 5 are shown, the total number of colonies per culture being approximately 102 but never less than 23. These isolates produced 57 and 67% of colonies by day 2. This isolate produced 19% of colonies by day 2. This isolate was unusual in displaying only 28% viability on minimal lactose medium as compared with complex medium, and of those colonies that arose on minimal lactose medium, 13, 18, 47, and 82% arose by days 2–5, respectively. 92% of the colonies first appeared on day 5. only 2 days to form a colony. This supports the hypothesis that the late-arising, recombination-dependent Lac+ mutants form on the plate in response to starvation. Although the vast majority of late Lac+ reversion events are dependent on recombination functions, including recA+, a small number of late Lac+ colonies can be seen in a stationary-phase mutation assay with recA-deleted cells (Harris et al. 1994; Rosenberg et al. 1995; Harris et al. 1996). These could be either novel RecA-independent stationary-phase mutants, or slow-growing growth-dependent mutants. RecA-independent stationary-phase mutation has been reported in other systems (Bridges 1993; Hall 1995; Galitski and Roth 1996; and reviewed by Foster 1993). We found that 14 out of 15 late-arising colonies of recA cells are not slow-growers (Table 1). The single slow-growing mutant had a majority of colonies appearing on day five. The data show that in the best characterized system for study of stationary-phase mutation, late-arising, recombination-dependent Lac+ mutants are not mostly slow-growing mutants that form during growth preselection. This supports the idea that recombination-dependent hypermutation (Torkelson et al. 1997) occurs while cells are starving, in stationary phase (Cairns and Foster 1991; Rosenberg 1997). The minor RecA-independent mutation route also appears to occur after exposure to selective medium. We thank P. J. Hastings for comments on the manuscript. This work was supported by a grant from the National Cancer Institute of Canada with funds provided by the Canadian Cancer Society, and grants R01GM53158 from the U.S. National Institutes of Health. M.-J.L. held an Alberta Heritage Foundation for Medical Research postdoctoral fellowship. S.M.R. was supported in part as a Medical Research Council of Canada Scientist and as an Alberta Heritage Senior Medical Scholar.