To rectify the significant deficit in our understanding of the interactions of the centrosome, we have generated a detailed centrosome interactome by using a directed yeast two-hybrid (Y2H) screen to identify the PPIs among 21 centrosome proteins. We use the information from our screen to gain an in vivo understanding of protein organization within the centrosome. We also demonstrate how the interactome can lead to the discovery of novel kinase substrates. Specifically, we have uncovered a centriole duplication independent role for the kinase Plk4, showing that it phosphorylates the centriole protein Cep135 to regulate its interaction with Asterless (Asl) and influence the radial positioning of Asl on the centriole. As a non-membrane bound organelle, the assembly of centrosomes must be driven by PPIs. These interactions are likely modified by highly regulated changes in protein binding affinity in a cell-type and cell-cycle dependent manner, which can in turn modulate centrosome behavior and function. We have focused on identifying PPIs among a core set of conserved centrosome proteins, classified as proteins of the centriole, the PCM and regulatory kinases. Previous high-throughput and small-scale interactions studies have suggested that there might be a limited number of PPIs among centrosome proteins and have raised the possibility that few interactions are used to construct a simple, ultimately static structure. Our data does not support this simplistic view of the centrosome. We have uncovered a large number of direct PPIs, dramatically expanding our understanding of the centrosome interaction landscape. This highly interconnected landscape suggests a much more complex centrosome assembly and regulatory process, which we suggest can be leveraged to perform a variety of specialized tasks dictated by a broad spectrum of cellular requirements. Some, if not most, of the interactions we identified must serve purposes beyond solidifying the final structure of the centrosome, since we observe interactions between proteins that appear to be quite distant from each other as determined by light microscopy. For example, we find an interaction between the C-terminus of Cep135, a core centriole protein, and the C-terminus of Cnn, a PCM protein. This interaction is not supported by the average position of these proteins in the final mother centrosome structure as determined by SIM. An alternative model raised by this interaction is that these proteins transiently bind during the daughter centriole assembly process. Furthermore, the daughter centriole forms adjacent to the proximal end of the mother centriole where the PCM is located. Therefore, the Cep135C-term-CnnC-term interaction might be required for concentrating Cep135 during centriole duplication. To further illustrate that many interactions might be used transiently, we highlight the localization diversity of Sas-4 within the centrosome. Many of the Sas-4 structures we observed differ from the previously reported average position as a dot or ring. Furthermore, Sas-4 localization in S2 cells and the developing Drosophila wing disc, although similar, are not identical, illustrates a cell-type specific diversity in its localization. The variety of localizations we observe are consistent with Sas-4s many PPIs with centriole and PCM components, and its roles in centriole duplication and PCM recruitment. Taken together, we propose that interaction diversity begets functional diversity, not only for Sas-4, but for all centrosome proteins. An unexpected finding from our screen was the low frequency with which FL proteins interacted in comparison to protein fragments. This discrepancy might help explain why so few interactions among centrosome proteins were identified in high-throughput Y2H screens. A number of studies have suggested that using protein fragments is advantageous for Y2H screens. Therefore the difference in the interactions made by full-length protein and protein fragments may be illustrative of centrosome protein regulation and not simply a technical limitation of Y2H. An attractive model is that FL centrosome proteins are unable to interact with the full complement of their interaction partners, perhaps due to an autoinhibited confirmation, which might have been relieved by using protein fragments. One such example is our study on Cep135 and Asl, which show exactly this type of regulation. In this case, the N- and C-terminal fragments of Cep135, but not Cep135FL, could interact with AslC-term. We go on to show that phosphorylation of Cep135 by Plk4 relieves the autoinhibition present in FL Cep135 allowing it to interact with the AslC-term. We then show that in the absence of this regulated interaction in the fly, the positioning of Asl on the centriole is perturbed. Furthermore, we show that Plk4 is critical for this positioning. This is the first example of Plk4 playing a role in constructing or maintaining the organization of centriole proteins outside of its well-established role in initiating centriole duplication. While we show this role is partially via phosphorylation of Cep135, our data suggests additional unknown Plk4 substrates are important for regulating Asl positioning. By combining our interactome with in vivo experimental evidence in Drosophila, we demonstrate how large interaction information can lead to in depth mechanistic insight into macromolecular assemblies. Integrating protein localization, dynamics and functional data with direct PPI information and mutant analysis, we show that interactions can predict inter- and intra-molecular architecture, identify kinase substrates and uncover regulated interactions within the centrosome. Understanding the interaction landscape of the centrosome is a critical foundation needed to gain an understanding of the molecular basis for human diseases caused by dysfunction of centriole, centrosome and cilia proteins. The diversity of centrosome-related diseases, such as microcephally, dwarfism, polycystic kidney disease and many others, can be best explained by loss of specific PPIs, rather than a simple complete loss of protein function stemming from a null mutation. We believe that our study will also guide new avenues of centrosome research that focus on the in-depth understanding of the diverse functions of these proteins, and could serve as a framework to explore other complex cellular processes.
Showing the most recent 10 out of 13 publications
