Search

Polo-like kinases (Plks) are key cell cycle regulators. They contain a kinase domain followed by a polo-box domain that recognizes phosphorylated substrates and enhances their phosphorylation. The regulatory subunit of the Dbf4-dependent kinase complex interacts with the polo-box domain of Cdc5 (the sole Plk in Saccharomyces cerevisiae) in a phosphorylation-independent manner. We have solved the crystal structures of the polo-box domain of Cdc5 on its own and in the presence of peptides derived from Dbf4 and a canonical phosphorylated substrate. The structure bound to the Dbf4-peptide reveals an additional density on the surface opposite to the phospho-peptide binding site that allowed us to propose a model for the interaction. We found that the two peptides can bind simultaneously and non-competitively to the polo-box domain in solution. Furthermore, point mutations on the surface opposite to the phosphopeptide binding site of the polo-box domain disrupt the interaction with the Dbf4 peptide in solution and cause an early anaphase arrest phenotype distinct from the mitotic exit defect typically observed in cdc5 mutants. Collectively, our data illustrates the importance of non-canonical interactions mediated by the polo-box domain and provide key mechanistic insights into the combinatorial recognition of substrates by Polo-like kinases.

Forkhead-associated (FHA) domains are phosphopeptide recognition modules found in many signaling proteins. The Saccharomyces cerevisiae protein kinase Rad53 is a key regulator of the DNA damage checkpoint and uses its two FHA domains to interact with multiple binding partners during the checkpoint response. One of these binding partners is the Dbf4-dependent kinase (DDK), a heterodimer composed of the Cdc7 kinase and its regulatory subunit Dbf4. Binding of Rad53 to DDK, through its N-terminal FHA (FHA1) domain, ultimately inhibits DDK kinase activity, thereby preventing firing of late origins. We have previously found that the FHA1 domain of Rad53 binds simultaneously to Dbf4 and a phosphoepitope, suggesting that this domain functions as an 'AND' logic gate. Here, we present the crystal structures of the FHA1 domain of Rad53 bound to Dbf4, in the presence and absence of a Cdc7 phosphorylated peptide. Our results reveal how the FHA1 uses a canonical binding interface to recognize the Cdc7 phosphopeptide and a non-canonical interface to bind Dbf4. Based on these data we propose a mechanism to explain how Rad53 enhances the specificity of FHA1-mediated transient interactions.

The cAMP signaling cascade is one of the most frequently targeted pathways for the development of pharmaceutics. A plethora of recent genetic and pharmacological studies suggest that exchange proteins directly activated by cAMP (EPACs) are implicated in multiple pathologies. Selective EPAC inhibitors have been recently developed. One specific inhibitor, ESI-09, has been shown to block EPAC activity and functions, as well as to recapitulate genetic phenotypes of EPAC knockout mice when applied in vivo. However, a recent study raised concern that ESI-09 might act as a non-specific protein denaturant. Herein, we present a detailed biochemical and pharmacological characterization, as well as a structure-activity relationship (SAR) analysis of ESI-09. Our studies show that ESI-09 dose-dependently inhibits activity of both EPAC1 and EPAC2 with apparent IC50 values well below the concentrations shown to induce "protein denaturation". Moreover, the ESI-09's action towards EPAC proteins is highly sensitive to minor modifications of the 3-chlorophenyl moiety. Taken together, these results demonstrate that ESI-09 indeed acts as an EPAC specific antagonist and does not significantly destabilize/denature proteins at pharmacological effective concentrations. This conclusion is further supported by NMR data showing that ESI-09 induces residue-dependent chemical shift changes at low concentrations, while preserving well dispersed peaks.

Allostery is an essential regulatory mechanism of biological function. Allosteric sites are also pharmacologically relevant as they are often targeted with higher selectivity than orthosteric sites. However, a comprehensive map of allosteric sites poses experimental challenges because allostery is driven not only by structural changes, but also by modulations in dynamics that typically remain elusive to classical structure determination methods. An avenue to overcome these challenges is provided by the NMR chemical shift covariance analysis (CHESCA), as chemical shifts are exquisitely sensitive to redistributions in dynamic conformational ensembles. Here, we propose a set of complementary CHESCA algorithms designed to reliably detect allosteric networks with minimal occurrences of false positives or negatives. The proposed CHESCA toolset was tested for two allosteric proteins (PKA and EPAC) and is expected to complement traditional comparative structural analyses in the comprehensive identification of functionally relevant allosteric sites, including those in otherwise elusive partially unstructured regions.