UC Berkeley

Ras GTPase cycles between the signaling active GTP-bound and signaling inactive GDP-bound states. GTP-bound Ras binds to the Ras-binding domains of effector proteins, such as Raf kinases and PI-3 kinase, triggering signaling cascades that result in cell proliferation. Ras is mutated frequently in cancers and some hyperproliferative developmental disorders. Data from cancer genomics show that most of these mutations occur at just three sites in Ras (Gly 12, Gly 13, and Gln 61). However, mutations at many sites in Ras lead to activation, as shown previously by site-saturation mutagenesis using a bacterial-two-hybrid assay. Here, we analyze the data from site-saturation mutagenesis of H-Ras in the mammalian Ba/F3 cell line. Comparison with the bacterial experiment indicates that in Ba/F3 cells, Ras activity is inhibited by a GTPase-activating protein (GAP). Activating mutations identified for H-Ras using the Ba/F3 experiment, and for H-Ras and K-Ras in the presence of a GAP in the bacterial assay, are present in cancer databases. However, the mutagenesis data do not explain why cancer mutations in Ras occur predominantly at three sites: Gly 12, Gly 13, and Q61. This is the question I address in this thesis.

In Chapter I, Ras constructs of different stabilities are mutagenized, and the results showcase a feature that distinguishes these three sites from the rest. While mutations at Gly 12, Gly 13, and Gln 61 activate Ras regardless of construct stability, mutations at lower-frequency sites (e.g., Val 14 or Asp 119) are activating or deleterious depending on the stability of the particular Ras construct. Mutations that decrease Ras stability can reduce nucleotide affinity, thereby permitting nucleotide exchange and reactivating Ras. However, the stability decrease may lead to insufficient levels of folded Ras. Collaborative work done in the Kuriyan lab uses hydrogen-deuterium exchange (HDX), measured by Nuclear Magnetic Resonance (NMR), to characterize the dynamics of four Ras mutants. Three of these mutants lead to activation and demonstrate global increases in HDX, consistent with Ras destabilization. In sum, these results demonstrate the importance of protein stability and GAP surveillance in determining the sensitivity of Ras to activation by mutation.

Chapter II analyzes gain-of-function Ras mutations in the context of constitutively GTP-bound Ras and shows that activation is not an additive property. Mutations that activate wild-type Ras become neutral when the activating Q61L mutation is in the background. This chapter also examines the mutational-fitness landscape of two distantly-related Ras proteins to determine if human Ras has any sequence-specific features that make it more prone to activation. The mutational sensitivity of the two proteins is strikingly similar to H-Ras and K-Ras. In the presence of a GAP, the canonical gain-of-function mutations are also activating. Thus, activation hotspots in Ras are a byproduct of the Ras switch cycle and not a characteristic specific to human Ras. Taken together, the first two chapters lay the foundation for a greater understanding of gain-of-function mutations in Ras.

Chapter III introduces mutagenesis-visualization, an open-source bioinformatics software that handles the data processing, analysis, and visualization in this dissertation. Site-saturation mutagenesis experiments have been transformative in our described study of protein function. Despite the rich data generated from such experiments, current tools for processing, analyzing, and visualizing the data offer a limited set of static visualization tools that are difficult to customize. Furthermore, usage of the tools requires extensive experience and programming knowledge, slowing the research process for those in the biological field who are unfamiliar with programming. Mutagenesis-visualization is a Python package for creating publication-quality figures for site-saturation mutagenesis datasets without the need for prior Python or statistics experience, whereby each of the graphs is generated with a one-line command. Thanks to this software, we have conducted a more precise analysis of gain-of-function mutations in Ras.