Technical Comments

Comment on "High Deleterious Genomic Mutation Rate in Stationary Phase of Escherichia coli"

See allHide authors and affiliations

Science  23 Apr 2004:
Vol. 304, Issue 5670, pp. 518
DOI: 10.1126/science.1094911

Loewe et al. (1) presented an estimate of the deleterious mutation rate for Escherichia coli in stationary phase that is two orders of magnitude higher than an earlier estimate using the same strain of E. coli (2). From this striking discrepancy in mutation rate estimates, they concluded that cells held in stationary phase become transient mutators and rapidly accumulate deleterious mutations. However, we believe that problems with experimental design and data interpretation render these conclusions erroneous.

The conventional method to study deleterious mutations is to conduct a mutation accumulation experiment, in which replicate populations started from a single ancestor are serially passaged through bottlenecks (3). In such small populations, where natural selection is largely absent, deleterious mutations can accumulate relatively freely by genetic drift. As a result of mutation accumulation, the mean fitness of replicate populations is expected to decrease over time while the variance across populations increases. Estimates of deleterious mutation rate and mean effect of each mutation can be inferred from these data using standard Bateman-Mukai and maximum-likelihood analyses (3).

Loewe et al. (1) conducted such an experiment, but rather than passaging cells through bottlenecks, the cells were maintained in stationary phase at high densities [on the order of 107 cells after an initial decline from about 109 cells (4)] for about 100 days. Maximum growth rates were then determined in fresh medium and the anticipated response was observed: Mean growth rate declined while the variance across populations increased. These changes in fitness were then used to derive estimates of rate and mean effect of deleterious mutations. Loewe et al. compared their high estimate of deleterious mutation rate (translating per-generation into per-day estimates by assuming a certain number of cell generations per day) with estimates from studies in which stationary phase conditions were less prominent, and concluded that the mutation rate must have been transiently increased during stationary phase conditions (1).

We believe that an alternative explanation is more likely. The probability for mutation fixation by genetic drift within the ∼100 days of the experiment is extremely small for populations of 107 cells. Instead, the long-term maintenance of large populations of E. coli in stationary phase, as in (1), allows for the frequent and rapid selection of spontaneous mutants with growth advantages during stationary phase [so-called GASP mutants (4)]. Furthermore, GASP mutants are known to have reduced fitness during rapid growth conditions due to antagonistic pleiotropy (AP) (5). It is thus likely that Loewe et al., rather than measuring the effects of accumulated deleterious mutations, were instead measuring the negative pleiotropic effects under rapid-growth conditions of mutations that reached fixation by natural selection under conditions of restricted growth. This would also cause a mean decline in maximum growth rate, as well as an increase in across-population variance if these nonselected pleiotropic effects are diverse. To distinguish between these possibilities, Loewe et al. would need to assay fitness of their populations under the same conditions in which the mutations arose (i.e., during stationary phase). Instead, their assays were conducted during rapid growth conditions, which would not measure the effects of mutations accumulated by genetic drift— however unlikely those would be. These two problems—large populations experiencing strong natural selection, and fitness assays performed under conditions different from those that prevailed in stationary phase—make the inferences drawn in (1) doubtful.

In a subsequent analysis, Loewe et al. interpreted the loss of niche breath observed in a long-term evolution experiment with E. coli (6) as resulting from mutation accumulation, while the conclusion of this study was instead that antagonistic pleiotropy caused most of the observed loss. It may be true that E. coli held in stationary phase have elevated mutation rates (7), and it remains an important and worthwhile problem to determine the extent of these rate increases and their associated effects on fitness. However, it will be a challenge to develop protocols that allow the free accumulation in stationary phase of deleterious mutations by genetic drift, while simultaneously avoiding selection of GASP mutants.


Navigate This Article