Final Report Summary - NETWORK MODULATION (Protein interaction interference: linking chemical biology to short linear motifs)
Cellular signaling is a complex process that requires the concerted regulation of many signaling proteins in response to an environment or internal stimulus. Protein kinases, phosphatases and their regulatory subunits, or kinome, are key molecules in this process of decision making by the cell. How the kinome is organized, how it interacts with other proteins within the cell, and how signals are captured, integrated and transduced into appropriate responses by the global kinome network were open questions at the beginning of this project.
In collaboration with the Gingras and Nesvishkii groups, the Tyers group set out to systematically study all the protein kinases and phosphatases, their regulatory subunits and substrates in budding yeast to deduce their interactions and phosphorylation patterns. This approach entailed the isolation of kinase and phosphatase complexes from yeast lysates and the interrogation of these complexes by mass spectrometry. Dr Neduva's analysis of the overall structure of the kinome revealed unexpected properties, including very dense connectivity between different, previously functionally separated, components of the yeast proteome. The most connected interaction sub-network was amongst the kinases themselves, with more than 30% more interactions than expected by chance. This finding implies that traditional view of linear kinase pathways is a highly unlikely scenario in cellular signaling. These results also correlate well with the discovery that protein kinases are highly phosphorylated compared to other proteins, suggesting extensive cross-regulation of kinase activities.
The highly connected kinase backbone that runs through the yeast protein interaction network has the potential to coordinate virtually all aspects of cellular behavior; indeed, recent phosphoproteomic studies suggest substantial cross-pathway flux can arise in response to discrete stimuli. Dense signaling networks have the potential to form attractors that can define stable cellular states or function as multiplexed input-output devices. In particular, we suggest that the numerous metabolic connections revealed in our kinome network may serve to coordinate nutrient signaling and cell cycle control with the recently described global metabolic cycle. The integration of our interaction-based approach with other genetic and biochemical means to map kinase substrates and phosphorylation sites should extend these predictions and form the basis for comprehensive models of cellular responses.
This systematic interrogation of the budding yeast kinome has uncovered numerous novel kinase associations, many of which have prompted new avenues of investigation in our lab and by other groups. For example, the cell cycle phosphatase Cdc14 association with multiple kinases revealed roles for Cdc14 in mitogen-activated protein kinase signaling, the DNA damage response, and metabolism. Another example is the interaction profile of the target of rapamycin complex 1 (TORC1) that revealed new effector kinases in nitrogen and carbon metabolism. This data has been pursed by the Tyers group to reveal new connections by TORC1 and metabolic enzyme regulation, including the regulation of a central enzyme in carbon and nitrogen metabolism by the TORC1-assocaited kinase Nnk1. We have deposited all our kinome data in a publically available resource (http://www.kinome.org). The data should continue to prompt further new avenues of investigation in other research groups.
The modulation of protein interactions is paramount to gain further insight into proteome organization and function. The ubiquitin-proteasome system (UPS) mediates selective protein degradation across most if not all intracellular processes and misregulation of the UPS is heavily implicated in cancer and many other diseases. In collaboration with the Sicheri group in Toronto, the Tyers laboratory has pursued the interactions of the SCF ubiquitin ligase family with its targets. Dr Neduva participated in a study to identify and characterize chemical modulators of SCF-substrate interactions. In this screen, we identified a biplanar dicarboxylic acid compound, called SCF-I2, as an inhibitor of substrate recognition by yeast SCF-Cdc4. To identify inhibitors of Cdc4, we used a fluorescence polarization assay that monitors the displacement of a fluorescein-labeled phosphodegron peptide from Cdc4, one of the F-box adapter proteins responsible for substrate recognition by the core SCF complex. SCF-I2 inhibits the binding and ubiquitination of full-length phosphorylated substrates by Cdc4. A co-crystal structure of the SCF-I2-Cdc4 complex reveals that the inhibitor inserts itself between two blades of the WD40 domain of Cdc4, to which the phosphodegron normally binds. Interestingly, the site of compound interaction lies far away from the substrate binding site. SCF-I2 perturbs the binding pocket in an allosteric manner via series of transmitted structural changes that lead to occlusion of the pocket by a normally buried tyrosine residue. Structural and sequence comparison of large number of WD40 domains suggests that this surprising mechanism of action might reflect a general mechanism of allosteric modulation of WD40 function that is conserved from yeast to mammals. WD40 domains represent one of the most abundant domain classes in the human proteome, and are present in a host of proteins involved numerous biological process, including signaling, transcription, chromatin remodeling, mRNA splicing, DNA replication and repair, protein synthesis, the ubiquitin system, autophagy, vesicle trafficking, the cytoskeleton and organelle biogenesis. As many of these proteins and processes are implicated in human disease the development allosteric inhibitors of WD40 domains could lead to completely novel small molecule leads for many therapeutically important targets.
In collaboration with the Gingras and Nesvishkii groups, the Tyers group set out to systematically study all the protein kinases and phosphatases, their regulatory subunits and substrates in budding yeast to deduce their interactions and phosphorylation patterns. This approach entailed the isolation of kinase and phosphatase complexes from yeast lysates and the interrogation of these complexes by mass spectrometry. Dr Neduva's analysis of the overall structure of the kinome revealed unexpected properties, including very dense connectivity between different, previously functionally separated, components of the yeast proteome. The most connected interaction sub-network was amongst the kinases themselves, with more than 30% more interactions than expected by chance. This finding implies that traditional view of linear kinase pathways is a highly unlikely scenario in cellular signaling. These results also correlate well with the discovery that protein kinases are highly phosphorylated compared to other proteins, suggesting extensive cross-regulation of kinase activities.
The highly connected kinase backbone that runs through the yeast protein interaction network has the potential to coordinate virtually all aspects of cellular behavior; indeed, recent phosphoproteomic studies suggest substantial cross-pathway flux can arise in response to discrete stimuli. Dense signaling networks have the potential to form attractors that can define stable cellular states or function as multiplexed input-output devices. In particular, we suggest that the numerous metabolic connections revealed in our kinome network may serve to coordinate nutrient signaling and cell cycle control with the recently described global metabolic cycle. The integration of our interaction-based approach with other genetic and biochemical means to map kinase substrates and phosphorylation sites should extend these predictions and form the basis for comprehensive models of cellular responses.
This systematic interrogation of the budding yeast kinome has uncovered numerous novel kinase associations, many of which have prompted new avenues of investigation in our lab and by other groups. For example, the cell cycle phosphatase Cdc14 association with multiple kinases revealed roles for Cdc14 in mitogen-activated protein kinase signaling, the DNA damage response, and metabolism. Another example is the interaction profile of the target of rapamycin complex 1 (TORC1) that revealed new effector kinases in nitrogen and carbon metabolism. This data has been pursed by the Tyers group to reveal new connections by TORC1 and metabolic enzyme regulation, including the regulation of a central enzyme in carbon and nitrogen metabolism by the TORC1-assocaited kinase Nnk1. We have deposited all our kinome data in a publically available resource (http://www.kinome.org). The data should continue to prompt further new avenues of investigation in other research groups.
The modulation of protein interactions is paramount to gain further insight into proteome organization and function. The ubiquitin-proteasome system (UPS) mediates selective protein degradation across most if not all intracellular processes and misregulation of the UPS is heavily implicated in cancer and many other diseases. In collaboration with the Sicheri group in Toronto, the Tyers laboratory has pursued the interactions of the SCF ubiquitin ligase family with its targets. Dr Neduva participated in a study to identify and characterize chemical modulators of SCF-substrate interactions. In this screen, we identified a biplanar dicarboxylic acid compound, called SCF-I2, as an inhibitor of substrate recognition by yeast SCF-Cdc4. To identify inhibitors of Cdc4, we used a fluorescence polarization assay that monitors the displacement of a fluorescein-labeled phosphodegron peptide from Cdc4, one of the F-box adapter proteins responsible for substrate recognition by the core SCF complex. SCF-I2 inhibits the binding and ubiquitination of full-length phosphorylated substrates by Cdc4. A co-crystal structure of the SCF-I2-Cdc4 complex reveals that the inhibitor inserts itself between two blades of the WD40 domain of Cdc4, to which the phosphodegron normally binds. Interestingly, the site of compound interaction lies far away from the substrate binding site. SCF-I2 perturbs the binding pocket in an allosteric manner via series of transmitted structural changes that lead to occlusion of the pocket by a normally buried tyrosine residue. Structural and sequence comparison of large number of WD40 domains suggests that this surprising mechanism of action might reflect a general mechanism of allosteric modulation of WD40 function that is conserved from yeast to mammals. WD40 domains represent one of the most abundant domain classes in the human proteome, and are present in a host of proteins involved numerous biological process, including signaling, transcription, chromatin remodeling, mRNA splicing, DNA replication and repair, protein synthesis, the ubiquitin system, autophagy, vesicle trafficking, the cytoskeleton and organelle biogenesis. As many of these proteins and processes are implicated in human disease the development allosteric inhibitors of WD40 domains could lead to completely novel small molecule leads for many therapeutically important targets.