Protein interactions are often mediated through binding of protein interaction domains to short peptide motifs. However, how the specificity of interactions is regulated within domain families is not well understood. In this thesis, I investigated this issue through the study of two cross-reactive SH3 domains from the yeast adaptor proteins Nbp2p and Bem1p. Despite low sequence similarity and different cellular roles, the Nbp2p SH3 and Bem1p SH3b domains recognize targets containing the same consensus sequences, raising the question of how the specificity of these domains is controlled. The Bem1p SH3b domain is also unusual in containing a 40 amino acid long C-terminal extension required for folding and interaction with the Cdc42p GTPase, which I also investigated in this thesis. I demonstrated that although Nbp2p SH3 and Bem1p SH3b domains recognize the same consensus sequence, each domain possesses its own unique specificity within the context of eight yeast peptides. Through structural and mutagenesis studies, I demonstrated that the Nbp2p SH3 and Bem1p SH3b domains employ different mechanism for binding, explaining how each domain can still possess its own unique specificity, while recognizing the same consensus sequence. I found that the specificity of Bem1 SH3b domain is largely regulated through inhibitory interactions, where a conserved residue in the binding interface serves to decrease binding to non-specific targets. Most significantly I observed that reducing the specificity of Bem1 SH3b domain, without reducing its binding affinity decreases its ability to function in vivo. My structural and mutagenesis studies with the Bem1p SH3b domain also demonstrated that its C-terminal extension is involved in peptide binding surface and creates a separate surface required for interaction with Cdc42p. I also showed that the Bem1p SH3b domain is capable of simultaneously interacting with peptide targets and Cdc42p, suggesting a role for this domain in Cdc42p mediated activation. Overall, my findings demonstrated the ability of SH3 domains to perform complex functions and bind their targets with finely tuned specificity. Most importantly, I demonstrated the significance of negative interactions for specificity determination and the importance of precise specificity for in vivo function.

Protein interactions are often mediated through binding of protein interaction domains to short peptide motifs. However, how the specificity of interactions is regulated within domain families is not well understood. In this thesis, I investigated this issue through the study of two cross-reactive SH3 domains from the yeast adaptor proteins Nbp2p and Bem1p. Despite low sequence similarity and different cellular roles, the Nbp2p SH3 and Bem1p SH3b domains recognize targets containing the same consensus sequences, raising the question of how the specificity of these domains is controlled. The Bem1p SH3b domain is also unusual in containing a 40 amino acid long C-terminal extension required for folding and interaction with the Cdc42p GTPase, which I also investigated in this thesis. I demonstrated that although Nbp2p SH3 and Bem1p SH3b domains recognize the same consensus sequence, each domain possesses its own unique specificity within the context of eight yeast peptides. Through structural and mutagenesis studies, I demonstrated that the Nbp2p SH3 and Bem1p SH3b domains employ different mechanism for binding, explaining how each domain can still possess its own unique specificity, while recognizing the same consensus sequence. I found that the specificity of Bem1 SH3b domain is largely regulated through inhibitory interactions, where a conserved residue in the binding interface serves to decrease binding to non-specific targets. Most significantly I observed that reducing the specificity of Bem1 SH3b domain, without reducing its binding affinity decreases its ability to function in vivo. My structural and mutagenesis studies with the Bem1p SH3b domain also demonstrated that its C-terminal extension is involved in peptide binding surface and creates a separate surface required for interaction with Cdc42p. I also showed that the Bem1p SH3b domain is capable of simultaneously interacting with peptide targets and Cdc42p, suggesting a role for this domain in Cdc42p mediated activation. Overall, my findings demonstrated the ability of SH3 domains to perform complex functions and bind their targets with finely tuned specificity. Most importantly, I demonstrated the significance of negative interactions for specificity determination and the importance of precise specificity for in vivo function.