Fig. 2

Constructive neutral evolution as illustrated by a simple toxin–antitoxin system. Shown are g (the antitoxin gene), h (the toxin gene) and p (a gene producing beneficial phenotype P). When both g and h products are present the phenotype P is expressed, since the g product suppresses (inhibits) the toxic effects of the h product. When g is deleted, h-product toxically inhibits expression of P, leading to the inference that g is essential for the function of P and may have undergone positive selection for the performance of P. But if g in fact were present before the h-product acquired its potential toxicity against P, and “pre-suppressed” such toxicity, no positive selection need be invoked. The suppression by chaperones of the effects of mutations in target proteins is an example of such CNE. In many other systems a simpler evolved dependence of one component “A” on another “B” permitted by the “pre-suppressive” activity of the second on possible, otherwise deleterious, mutations in the first can be imagined [38]. Again the conclusion that B evolved positively to support the activity of A would be unwarranted