Characterization of the p23 interaction network

Molecular chaperones have evolved to support the general stability and maintenance of the cellular proteome. Through their broad binding capacity, transient interactions, and high cellular abundance, the Hsp70 and Hsp90 chaperone families promote folding of a diverse clientele in numerous cellular compartments and organelles. An often-underappreciated aspect of this activity is the ability to modulate conformations of native substrate proteins. By guiding substrates over potentially unstable intermediates, chaperones are able to shuttle clients through multiple functional forms to alter the composition and activity of complex protein assemblies. Thus, molecular chaperones contribute significantly to cellular homeostasis by supporting all steps of protein maturation. Hsp90 is one of the most abundant chaperones in the cell capable of physically associating with partially folded (e.g., steroid hormone receptors) or native clients (e.g., telomerase). In yeast, the Hsp90 interaction network encompasses nearly one-sixth of the genome, operating in numerous pathways and cellular processes. Given the ubiquitous nature of chaperone association, mechanisms to efficiently direct and modulate the activity of the chaperone must exist. This is accomplished largely through the Hsp90-associated cofactors, termed cochaperones, which modulate the activity of the central chaperone in many instances. Furthermore, numerous Hsp90 cochaperones are able to independently suppress aggregation of denatured substrates, opening the possibility of direct modulation of client proteins. Despite the well-characterized association with Hsp90, little is known about how the cochaperones function within the cell or connect to the Hsp90 network. For this reason, we used high-throughput genetic and proteomic approaches to identify the interaction network of the Hsp90 cochaperone, p23. Our study uncovered a large network of interactions that, like Hsp90, contains genes that operate in numerous cellular processes and locations. Furthermore, the p23 network revealed novel functions in protein transport and DNA repair that were functionally validated and conserved from yeast to mammalian cells. Comparison of the p23 and Hsp90 networks resulted in the surprising observation of limited overlap (~25%). This suggested to us that p23 may operate outside the sphere of Hsp90, a view that contradicts the typical model of cochaperone activity. Despite the seemingly high level of independence, both networks displayed similar patterns of process enrichment and were connected to many common complexes. Therefore, our data supports a model in which p23 operates along similar pathways and with common complexes but in a manner that is not strictly dependent on Hsp90 presence. Further analysis of the p23 network indicated nearly 30% of the interactions participate in nuclear processes. This supported previously identified nuclear functions of p23 and predicted numerous novel p23 nuclear functions. Hidden within the sizeable nuclear network were six subunits of chromatin remodelers, accounting for five of the eight complexes present in Saccharomyces cerevisiae. Given the recent finding that p23 deletion alters global chromatin states (Zelin et al. 2012), we sought to study p23 functions with chromatin remodelers. Therefore, we began by following up on the genetic interaction with RSC7, a subunit of the Remodels the Structure of Chromatin (RSC) complex. A common characteristic of all chromatin remodeling complexes is the utilization of ATP-dependent mechanisms to alter the local chromatin states during transcription, replication, and DNA repair. Despite the apparent critical role remodelers play within the cell, only RSC is essential for viability in S. cerevisiae. Our study identified the ability of p23 to dissociate RSC: DNA complexes in vitro. Furthermore, screening all 17 RSC subunits by yeast 2-hybrid analysis indicated that p23 interacts with Rsc30, which heterodimerizes with Rsc3 to form a RSC DNA binding module. Cloning of Rsc3 and Sfh1 subunits next to the Gal4-AD resulted in an autoactivation phenotype that was suppressed by expression of Gal4-DBD-p23. Collectively, our data support a model that p23 modulates RSC-DNA interaction through modulation of the Rsc3/30 component. We next compared the ChIP genome binding profiles of p23 to that of a previously published high-confidence RSC profile and found an approximately 25% overlap. Finally, operating under the assumption that p23 is acting as a general dynamic-promoting factor, we hypothesized that we might see altered kinetics of RSC binding to inducible loci. We measured RSC association with ethanol-stress-induced gene promoters and observed a significant reduction in the rate of localization in the absence of p23. Collectively, our data suggest that p23 is an important factor that promotes mobilization of RSC, probably in a general capacity, but especially under conditions that result in a significant relocalization of RSC along the genome. Our work has dramatically expanded the known p23 interaction network, improving our understanding of how the Hsp90 cochaperone functions within the cell. Furthermore, this network led to the discovery of novel p23 activities in protein transport, cell motility, and chromatin remodeling. Together, my project has challenged the existing model of cochaperones functioning exclusively through modulation of the central chaperone; rather, our study of p23 suggests cochaperones may have a much more expansive and independent interaction network than previously realized