Roy Kishony and Stanislas Leibler*
Corresponding author: Stanislas Leibler
Journal of Biology 2003, 2:14 doi:10.1186/1475-4924-2-14
(2003-07-07 19:59) Michigan State University
Fascinating work! I'm reminded of results from Mikkola & Kurland FEMS Microbiology
Letters 56:265-269...they found that the growth rate depression relative to wild type
of E. coli ribosome mutants was reduced in poor media compared to rich media. The
work is based on earlier modeling work by Ehrenberg and Kurland reported in 1984,
Quarterly Review of Biophysics 17:45-82.
The logic is as follows: Suppose the mutation is such that for a fixed investment
of mass in the protein synthesis system, the mutant only gets 2/3 the protein output
of wild type. Consider that in a poor medium, growth is slow and the protein synthesis
system is normally a relatively small fraction of cell mass, say 20%. The mutant
can compensate for reduced protein output by increasing the size of its protein synthesis
system. It could match the protein output of the wild type by investing 1.5 times
the biomass in the system, an amount of biomass equal to 10% of the total wild type
cell mass. Of course, this does result in a growth rate penalty, since the cell has
to replicate this extra biomass in each generation.
But compare that to the situation in rich medium where growth is rapid and the protein
synthesis system may normally comprise 50% of cell mass. Now the mutant has to make
a proportionately much larger increase in cell mass to compensate for the relative
inefficiency of its protein synthesis system, equal to 25% of the wild type cell mass
if it is to match total wild type protein output. Hence the relative fitness penalty
is greater in rich medium.
Mikkola and Kurland had a specific mutation in the protein synthesis system, but the
logic of the Ehrenberg-Kurland model they invoked to explain their results would apply
to any comparison of conditions where the mass normally invested in the protein synthesis
system differs (i.e., conditions where growth rate normally differs). The condition
under which the protein synthesis system comprises a smaller fraction of cell mass
will show a smaller relative effect of a mutation reducing the mass-normalized efficiency
of protein synthesis; the condition under which the protein synthesis system is normally
a larger fraction of cell mass will have a larger relative effect from the same mutation.
Insofar as each of the random mutants generated by Kishony and Leibler can be considered
to contribute to the efficiency of some functional subsystem of the cell, as with
the protein synthesis system above, the logic should hold. Clearly, this is not the
only way that a mutation by environment interaction can be generated, but it may be
a fairly general phenomenon that acts in addition to any specific results (such as
a mutation in a gene that is critical in a particular stress response). I wonder
whether the fact that most of the stress environments tested seemed to ameliorate
mutational effects is explained by something akin to the phenomenon reported by Mikkola
BioMed Central Ltd unless otherwise stated. Part of Springer Science+Business Media.