Sydney Peng

Protein phosphatases play an integral role in cellular senescence and have been demonstrated to facilitate malignant transformation of normal healthy cells. The protein phosphatase PP2A is a known tumor suppressor in its wild-type form. PP2A is composed of three protein chains (A, B and C). However, PP2A with missense mutations at residues 179, 183, 217, and 256 on chain A no-longer remain a tumor suppressor, and have been identified in uterine cancer. Here, we utilize dynamic network analysis to computationally investigate interactions between the three chains of PP2A at the atomic level to probe why these mutations correlate with changing PP2A from its tumor suppressor role towards that of a molecule correlated with malignancy. An all-atom model of PP2A was created using previously published x-ray coordinates and molecular dynamics simulations were performed. Communities were formed from the trajectory of the molecular dynamics simulations using the Girvan–Newman algorithm. A correlation map was generated to determine the level of dynamic correlation between every C-alpha atom, which illustrated a set of communities comprising the entire biomolecule. Nodes are links through which dynamic correlation occurs between atoms. Critical nodes are nodes connecting two communities where there are no other nodes connecting the same communities. Critical nodes from our analysis were compared against known mutations associated with uterine cancer. The dynamic network path analysis uses the correlation map to trace a link of nodes for which each node is dynamically correlated. An optimal path bridges the shortest connection between the source atoms of one chain and the sink atoms of another chain on the protein, and a suboptimal path analysis was run to identify the optimal paths between chain A and B, as well as between regulatory chain B and catalytic chain C. Our computation model identified critical nodes to occur at residues 180, 217, and 257, highly correlating with known missense mutation sites of 179, 183, 217, 256. We additionally recreated each reported missense mutation and reran the simulation. Under these simulation conditions, the critical nodes disappeared, implying a loss of the dynamic link between chains A and B and potentially a loss of the protein’s function. Our suboptimal path analysis also revealed a significant increase in the number of connections in the optimal paths between chains A and B and a significant decrease in the number of connections in the optimal paths between chains B and C, suggesting a probable short-circuit in the communication between these chains. Thus, the changes found in the dynamic network analysis between the wild type PP2A and PP2A mutants show potentially why these mutations alter the function of the enzyme. Such analysis may further facilitate the design of targeted agents for potential therapeutic intervention.