Regarding selection induced mutation
The paper in Science is Harris et al. Science 264: 258-260
Also there is a commentary in the same issue by Thaler (p224-225)
In addition there is a recent paper in Mol. Biol. Evol. (vol 11 p159-168) by
Barry Hall on some aspects of selection-induced mutation in the bgl operon.
Warren Gallin suggests that the RecA dependence suggested by Harris et al.
might be due to recombination occurring more actively in transcribed
regions of the genome. This seems like a reasonable possibility since
transcription is known to affect recombination and repair (I am currently
working on how some aspects of transcription-coupled repair might affect
selection-induced mutations). However, Cairns and Foster suggest that
transcription is not one of the causes of selection-induced mutations because
1) tRNA suppressors account for a large % of selection induced
2) IPTG does not increase the number of revertants in lacZ
3) lac mutants that are constitutively active are not reverted
unless lactose is around.
Nevertheless, the only experiment that has attempted to address whether the
seleciton induced mutation rate varies with transcription is that with the IPTG
induction of the lacZ gene. Unfortunately, they do not provide any information
about how much the gene is induced in their conditions. Thus I think it is
still as open question as to whether the transcription rate affects the rate of
Finally, a personal opinion. I believe that the points raised by Lenski and
Mittler (e.g. Science 259:188-194) about some of the problems of the
experiments have not been well addressed. Their criticisms have some
experimental support. For example, Macphee showed that the control used
by Cairns and Foster for whether the mutations were specifically occurring in
the lacZ gene was faulty. The control was to overlay plates at different times
during selection for lac reversion with valine and to see how many colonies
would grow thus counting the number of valine resistant mutations. Macphee
showed that the conditions used to detect valine resistance (glucose as a
carbon source) were not optimal and that when glycerol was used as a carbon
source he detected essentially the same number of valine resistant colonies as
lac revertants. This suggests that at least in regard to valine resistance
versus lac reversion the selection induced mutations are not directed.
At least Harris et al. acknowledge that their experiments on recABCD
do not address whether the phenomena is directed but instead what the
mechanisms of the mutation processes in stationary phase cells under selective
pressure for genetic changes.
Jonathan A. Eisen
Department of Biological Sciences
Stanford, CA 94305