The relationship between dipeptidase activity variation and larval viability in Drosophila melanogaster.

The enzyme dipeptidase-A (DIP-A) in Drosophila melanogaster is coded by a second chromosome locus that is polymorphic for three allozymes in natural populations. DIP-A appears to be the only enzyme in D. melanogaster capable of hydrolyzing the dipeptide glycyl-L-isoleucine, since flies homozygous for null alleles at this locus have no detectable glycyl-L-isoleucine-ase activity. DIP-A activity occurs in many tissues and throughout development, but is particularly high in the larval midgut, suggesting an important role in protein digestion. These observations suggested an experimental design for investigating the adaptive significance of genetic variation in DIP-A activity. Fitness components of DIP-A variants could be estimated and compared under two environmental conditions (defined diets under axenic conditions). In the restrictive environment, the essential amino acid L-isoleucine is provided only in the form of glycyl-L-isoleucine, whereas in the permissive environment, L-isoleucine is provided in free form. We predicted that DIP-A activity would be essential in the restrictive, but not in the permissive environment. The results reported here clearly contradict this prediction. Two stocks homozygous for DIP-A null alleles from different geographic locations are each viable on the restrictive diet. Furthermore, relative viability experiments in which null allele larvae compete with larvae having DIP-A activity provide no evidence for even a partial reduction in egg to adult survival on the restrictive diet. Apparently, the null allele larvae have some alternative mechanism for obtaining L-isoleucine from the dipeptide, even though no glycyl-L-isoleucine-ase activity can be detected in vitro. These results, along with the viability of null alleles for many other enzymes, support the idea that eukaryotes have an intricate network of alternative biochemical pathways through which the same necessary function may be achieved. Such "buffering capacity" makes it very difficult to analyze the effects of enzyme variants on fitness components.