Reverse engineering GTPase programming languages with reconstituted signaling networks

Abstract

The Ras superfamily GTPases represent one of the most prolific signaling currencies used in Eukaryotes. With these remarkable molecules, evolution has built GTPase networks that control diverse cellular processes such as growth, morphology, motility and trafficking. (1-4) Our knowledge of the individual players that underlie the function of these networks is deep; decades of biochemical and structural data has provided a mechanistic understanding of the molecules that turn GTPases ON and OFF, as well as how those GTPase states signal by controlling the assembly of downstream effectors. However, we know less about how these different activities work together as a system to specify complex dynamic signaling outcomes. Decoding this molecular "programming language" would help us understand how different species and cell types have used the same GTPase machinery in different ways to accomplish different tasks, and would also provide new insights as to how mutations to these networks can cause disease. We recently developed a bead-based microscopy assay to watch reconstituted H-Ras signaling systems at work under arbitrary configurations of regulators and effectors. (5) Here we highlight key observations and insights from this study and propose extensions to our method to further study this and other GTPase signaling systems.

Keywords: Ras; biochemistry; dynamics; in vitro reconstitution; signaling.

Figure 1.

A bead-based microscopy assay for watching Ras GTPase networks signal under arbitrary system configurations. Depiction of the assay developed in ref. . Beads loaded with the H-Ras GTPase are placed in a solution containing fluorescent effector molecules and user-defined concentrations of network components like GAP and GEF. When an INPUT that changes the activity of one of these components is applied to the system (such as a change in [GEF]), activation of the GTPase will lead to recruitment of fluorescent effectors to the bead surface, defining the OUTPUT of the system. This recruitment can be monitored in real time by microscopy for hundreds of unique beads in a multiplexed fashion to see how different network configurations result in different dynamic signaling OUTPUTs.

Figure 2.

Each Ras network component programs Ras signaling output dynamics in a unique way. When an input is applied to a Ras signaling system, the underlying network configuration will shape the resulting output dynamics. We found that each signaling component in the network impacted the timing, duration, and amplitude of signaling output in a unique way. The way each network component—GEF, GAP, Ras, or effector—programs the system output is summarized in this figure.

Figure 3.

The extent to which oncogenic alleles distort signal processing depends on the underlying network configuration. The difference between signaling outputs of networks harboring either wildtype Ras or oncogenic Ras networks was compared using our assay. When an input was applied to a low-GAP network context, the outputs from wildtype and oncogenic Ras networks were largely indistinguishable. In contrast, when that same input was applied to a high-GAP network context, the outputs were completely different: wildtype Ras produced a low-amplitude transient pulse of output, while oncogenic Ras produced a high-amplitude sustained signal.