Nondividing Cells Within Colonies of Escherichia coli

A genetic strategy was designed to examine the occurrence of mutations in stationary-phase populations. In this strategy, a parental population of cells is able to survive under both permissive and restrictive conditions whereas mutants at a particular target locus exhibit a conditional-lethal phenotype. Thus, by growing the population to stationary phase under restrictive conditions and then shifting it to permissive conditions, mutations that had arisen in stationary phase can be studied without confounding effects caused by the occurrence of similar mutations during growth of the population. In two different applications of this strategy, we have studied the reversion to Lac’ in stationary phase of several Lacmutations in Escherichia coli. Our results indicate that a variety of spontaneous point mutations and deletions, particularly those that are sensitive to the mechanisms of replication slippage (for their generation) and methyldirected mismatch repair (for their correction), can arise in nondividing populations of cells within a colony. The frequency of their occurrence was also elevated in m u 6 strains, which are defective in such mismatch repair. These data have relevance to the ongoing debate on adaptive or directed mutations in bacteria. S PONTANEOUS mutations are thought to arise as errors in replication of the genetic material when cells (or organelles or viruses) increase their numbers, with the errors that either are blind to or escape the correction mechanisms becoming then fixed in the genome. Can spontaneous mutations also arise in nondividing cells? An early study by RYAN (1955) suggested that His+ revertants could arise in populations of nondividing Hiscells (of Escherichia colq starved for the amino acid. Recent interest in this question was revived by claims, in both bacteria and yeast, that directed mutations (also variously referred to as adaptive or Cairnsian or selection-induced or postplating mutations) can occur in starving cell populations that are subjected to nonlethal selections (reviewed in FOSTER 1993). It has been speculated that endogenous DNA lesions caused by base alkylation (MACKAY et al. 1994) or oxidative stress (BRIDGES 1996; BRIDGES et al. 1996), as also by stationary-phase induction of the SOS response (TADDEI et al. 1995), may contribute to the occurrence of these mutations. However, the claims for directed mutation have also generated much contention and controversy. The criticism has focused on the validity of the assumptions that the mutants were not preexisting in the population and that the population itself was truly nondividing, as also on the evidence that the mutations were “directed” by the environmental selection condiCurresponding author: J. Gowrishankar, Centre for Cellular and M e lecular Biology, Hyderabad 500 007, India. E-mail: [email protected] Genetics 147: 991-1001 (November, 1997) tions (reviewed in SMITH 1992; LENSKI and MITTLER 1993; SNIEGOWSKI and LENSKI 1995; MACPHEE and A” One example of directed mutation that has been intensively studied is the reversion to Lac+ of a frameshift mutation lucI33::lacZpresent on F’lacpro in E. colistrain FC40 (reviewed in ROSENBERG et al. 1995). The mutational spectrum of “adaptive” reversion events of this allele is different from that of the “random” reversion events that occur during exponential growth, and it is only the former that are also R e d and RecBCDdependent. It has been suggested that stationary-phase cells are physiologically deficient in DNA mismatch repair and that this deficiency accounts for the distinct spectrum of adaptive reversions in strain FC40 (FOSTER et al. 1995; LONGERICH et aZ. 1995; FENG et al. 1996). On the other hand, it is now known that adaptive reversions of lacI33::lacZ occur only when the mutation is carried on a conjugation-proficient F’ (reviewed in ROSENBERG et al. 1995); redundant homosexual transfer of the F‘ between individual cells in the population has been demonstrated under these conditions (RADICELLA et al. 1995; PETERS et al. 1996). Whether the findings with FC40 bear relevance to the occurrence of other categories of adaptive mutation or at other (nonepisomal) locations therefore remains uncertain (e.&, see GALITSKI and ROTH 1996; PRIVAL and CEBULA 1996; HALL 1997). The principal difficulty in studying, and perhaps even defining the existence of, stationary-phase mutations (including directed mutations) is that the same mutaBROSE 1996). 992 M. Reddy and J. Gowrishankar