Proteins that interact coevolve their structures. When mutation disrupts the interaction, compensation by the partner occurs to restore interaction otherwise counterselection occurs. We show in this study how a destabilizing mutation in one protein is compensated by a stabilizing mutation in its protein partner and their coevolving path. The pathway in this case and likely a general principle of coevolution is that the compensatory change must tolerate both the original and derived structures with equivalence in function and activity. Evolution of the structure of signaling elements in a network is constrained by specific protein pair interactions, by requisite conformational changes, and by catalytic activity. The heterotrimeric G protein-coupled signaling is a paragon of this protein interaction/function complexity and our deep understanding of this pathway in diverse organisms lends itself to evolutionary study. Regulators of G protein Signaling (RGS) proteins accelerate the intrinsic GTP hydrolysis rate of the Gα subunit of the heterotrimeric G protein complex. An important RGS-contact site is a hydroxyl-bearing residue on the switch I region of Gα subunits in animals and most plants, such as Arabidopsis. The exception is the grasses (e.g., rice, maize, sugarcane, millets); these plants have Gα subunits that replaced the critical hydroxyl-bearing threonine with a destabilizing asparagine shown to disrupt interaction between Arabidopsis RGS protein (AtRGS1) and the grass Gα subunit. With one known exception (Setaria italica), grasses do not encode RGS genes. One parsimonious deduction is that the RGS gene was lost in the ancestor to the grasses and then recently acquired horizontally in the lineage S. italica from a nongrass monocot. Like all investigated grasses, S. italica has the Gα subunit with the destabilizing asparagine residue in the protein interface but, unlike other known grass genomes, still encodes an expressed RGS gene, SiRGS1. SiRGS1 accelerates GTP hydrolysis at similar concentration of both Gα subunits containing either the stabilizing (AtGPA1) or destabilizing (RGA1) interface residue. SiRGS1 does not use the hydroxyl-bearing residue on Gα to promote GAP activity and has a larger Gα-interface pocket fitting to the destabilizing Gα. These findings indicate that SiRGS1 adapted to a deleterious mutation on Gα using existing polymorphism in the RGS protein population.