Controlling oligomerization through protein engineering: in vivo analysis of Hsp90

Abstract

Many homo-oligomeric proteins are vital for biology including ion-channels, the p53 tumor suppressor, and the essential kinase-associated chaperone Hsp90. Mutational analyses of these homo-oligomeric systems in vivo is complicated by cross-oligomerization between wild-type and mutant subunits. We have devised a generalizable thermodynamic strategy to prevent cross-dimerization. Appending an oligomerization domain to the mutant subunits reduces the free energy of homocomplexes relative to wild-type/mutant heterocomplexes. We have used this strategy to engineer super-stabilized Hsp90 dimers that do not cross-oligomerize with wild-type Hsp90. Super-stabilized Hsp90 supports yeast viability and is fully active in the maturation of v-src kinase. Thus, our stabilization strategy does not disturb the biochemical function of Hsp90.We have used superstabilized Hsp90 to address a fundamental and long-unanswered question regarding Hsp90: what clients or substrate proteins depend on Hsp90 ATPase activity in vivo. The identification of ATP dependent Hsp90 substrates has been a major challenge both in vitro and in vivo. In vitro studies are complicated by the large number of co-chaperones required for Hsp90 to function efficiently. In vivo studies are complicated both because ATPase deficient Hsp90 mutants do not support viability and because when different Hsp90 variants are co-expressed they form a mixture of different dimer species. We have used our engineered super-stabilized Hsp90 to developed a yeast system to identify clients that rely on Hsp90 ATPase activity. Using this approach, we find that Hsp90 mutants deficient for ATP binding or hydrolysis have differential impacts on the activation of kinase and hormone receptor clients in vivo. These results provide a rationale for understanding anti-cancer drugs that competitively bind to the ATPase site of Hsp90.