Metabolic engineering has been very successful in addressing pathway optimization for localized branches, but multigenic or poorly understood traits have remained a challenge. In particular, enhancing the robustness of biocatalysts has relied almost exclusively in evolutionary approaches that result in untransferable phenotypes. We have successfully delivered a method that introduces global transcriptomic changes, thereby generating diversity at the phenotypic level. We have applied it to Lactobacillus plantarum for overcoming its poor resistance to high lactic acid and low pH. Using mutagenesis of the principal sigma factor, we have produced strains with increased tolerance to both stresses. In addition, the diversity of the resulting libraries was quantified and used as a measure of their usefulness and versatility. Because of its easy implementation and the tractability and transferability of the genetic modifications, this method holds great potential for becoming a routine tool for metabolic engineering.