We present a new version of P systems called Client-Server P Systems (CSPS). The client membranes are characterized by their states; the server membrane stores the states of the clients andtriggers the corresponding interaction rules. We show that CSPS have the same expressive power as Turing machines. CSPS is used to model various molecular processes in which interaction and state transitions are causally linked. Signaling pathways and T cell activation are described by using a CSPS software environment called MOlNET (MOlecular NETworks). MOlNET can describe the dynamics of molecular interactions, including both qualitative and quantitative aspects and simulating the signaling pathways that tune the activation thresholds for T cells.