A molecular investigation of genotype by environment interactions

The fitnesses conferred by seven lactose operons, which had been transduced into a common genetic background from natural isolates of Escherichia coli, were determined during competition for growth rate-limiting quantities of galactosyl-glycerol, a naturally occurring galactoside. The fitnesses of these same operons have been previously determined on lactose and three artificial galactosides, lactulose, methyl-galactoside and galactosyl-arabinose. Analysis suggests that although marked genotype by environment interactions occur, changes in the fitness rankings are rare. The relative activities of the beta-galactosidases and the permeases were determined on galactosyl-glycerol, lactose, lactulose and methyl-galactoside. Both enzymes display considerable kinetic variation. The beta-galactosidase alleles provide no evidence for genotype by environment interactions at the level of enzyme activity. The permease alleles display genotype by environment interactions with a few causing changes in activity rankings. The contributions to fitness made by the permeases and the beta-galactosidases were partitioned using metabolic control analysis. Most of the genotype by environment interaction at the level of fitness is generated by changes in the distribution of control among steps in the pathway, particularly at the permease where large control coefficients ensure that its kinetic variation has marked fitness effects. Indeed, changes in activity rankings at the permease account for the few changes in fitness rankings. In contrast, the control coefficients of the beta-galactosidase are sufficiently small that its kinetic variation is in, or close to, the neutral limit. The selection coefficients are larger on the artificial galactosides because the control coefficients of the permease and beta-galactosidase are larger. The flux summation theorem requires that control coefficients associated with other steps in the pathway must be reduced, implying that the selection at these steps will be less intense on the artificial galactosides. This suggests that selection intensities need not be greater in novel environments.