This article presents evidence that starvation for leucine in an Escherichia coli auxotroph triggers metabolic activities that specifically target the leu operon for derepression, increased rates of transcription, and mutation.
—Write et al
That seems a monumentally unsupported claim. It seems to me that their results are better explained in two ways, one of which is discussed in their own discussion section and one of which isn't.
1. As discussed in their discussion and supported by other papers: transcription increases the chances of mutation. This should surprise no-one: DNA is usually packaged in a manner to ensure its protection and only unravelled when copied or expressed.
2. As not discussed in the paper. A mutant that is already
derepressed will experience a significant selective advantage when it is put into the starvation conditions as it will already have the required enzymes in place; this means a sample taken post-starvation will have a vastly increased chance of finding a bacterium that is descended from a pre-starvation mutant. I see nothing in their method that would distinguish between a pre-starvation mutant and one that occurred during starvation.