Global visualization of the proteome perturbed upon treatment with GA and its protein-protein interactions. (A) Proteins showing a significant change in their relative abundance (stSILAC) at 6hs and/or 20hs after treatment with GA were organized in a network-based map using the software Cytoscape (Cline MS, 2007). Nodes (proteins) are connected by edges (lines) representing their known physical interactions (source Echeverria et al, Plos One, 2011). Nodes are colored with a red-to-blue gradient (heat map thermogram) according to their log2 ratios/fold change values, where red and blue represent enrichment and depletion, respectively (see inset for color gradient). Visualization of the colored graph allowed the detection of different patterns along the GA treatment. Subsequently, supervised clustering methods permitted to establish the global patterns, which are indicated in the Figure. Enriched proteins are localized in the “yang” region of the graph while depleted ones era in the “ying” part. (B) The GO terms for "Cellular processes"  for the two main groups (604 depleted proteins and 580 enriched proteins) in the graph were retrieved and compared using the Cytoscape plugin ClueGO. Statistically highly enriched terms indicate the cellular functions that are potentially augmented or impaired upon treatment with GA. Functional terms shared by both groups are located in the center of the scheme
Dynamic changes of the Hsp90 chaperone machinery after GA treatment. Components of the Hsp90 molecular chaperone machine (Hsp90Int [Echeverria et al, Plos One 2011]) showing significant changes in the stSILAC data are schematized in a graph. Edges (lines) represent protein-protein interactions among members of the machinery. stSILAC data is integrated in the graph and represented as a color gradient (see inset).
Dynamic changes of the proteasomal/ubiquitination machinery and its connected Hsp90 clients after GA treatment. Members of the proteasomal complexes, ubiquitination machinery, molecular chaperones and known or potential Hsp90 “clients” proteins interconnected by protein-protein interactions were extracted from the network represented in Fig1_PE. Relative levels of proteins at 6hs and at 20hs after GA treatment are integrated in the graph and represented as a color gradient (see inset).
Both beneficial and detrimental effects of Hsp90 inhibitors on cancer-related proteins.  Cancer proteins categorized by Higgins et al (Higgins ME, NAR 2006) as oncogenes and tumor suppressors were retrieved and identified in the stSILAC data. These results were further refined and confirmed by literature mining, and organized in a network. Relative levels of proteins at 20hs after GA treatment are integrated in the graph and represented as a color gradient (see inset).
Validation of new Hsp90 clients. (A) Network analysis of a selected set of potentially new Hsp90 client proteins. stSILAC data and the Hsp90 interaction network Hsp90Int (Echeverria et al, PlosOne 2011) were combined to identify interesting candidates (OGT, ITK and BRAT1) with no reported interactions with Hsp90 at the time of the analysis. Edges connecting candidate proteins with known Hsp90 interacting proteins are highlighted in red. (B) Co-immunoprecipitation (co-IP) experiment demonstrating interactions between BRAT1, OGT and ITK with Hsp90b in Jurkat cells. Equal concentrations of specific antibodies against BRAT1 (rabbit), OGT (rabbit), ITK (mouse), Hsp90b (mouse) and the corresponding non-immune control antibodies from rabbit and mouse were used in co-IP experiments, and then analyzed by immunoblotting (WB). (C) GA-induced degradation of BRAT1, ITK and OGT in Jurkat cells. Lysates from cells treated with GA or with the equivalent volume of the solvent DMSO (control) for 6 and 20 hs were analyzed by WB for these three mentioned proteins and also for Hsp70 and CDK6 as positive controls of GA action.
published here
Back to Top