Evolving a Direct Inhibitor of the Ras Proteins

Abstract

In recent years, great advances have been made in understanding the molecular causes of human disease, but our ability to exploit these discoveries for therapeutic benefit is frequently limited by the inability to make drugs that target the processes responsible. Many diseases can be linked to the aberrant activity of proteins, and while the development of inhibitors for enzymes and extracellular targets is often feasible, these proteins account for only a small fraction of all the proteins in cells. The remaining proteins are, in most cases, considered therapeutically intractable and are sometimes referred to as "undruggable." Many proteins, particularly in higher organisms, carry out their activity in part through interactions with other proteins and biomolecules. The ability to specifically disrupt these interactions could have great therapeutic benefit, as it may provide a means of targeting otherwise intractable processes. The focus of this dissertation is on the development and characterization of molecules that inhibit the interactions of an “undruggable” protein target, Ras, which is linked to both the initiation and progression of a wide array of human cancers. Our approach has been to use high-throughput screening, coupled with directed evolution, to identify and improve small proteins (peptides) that bind Ras and block its ability to engage the effector proteins necessary for its oncogenic activity. We report these efforts, along with a series of biochemical experiments aimed at characterizing the properties and binding mechanism of the peptides discovered in the screen. These peptides bind the three human Ras proteins with mid-to-low nanomolar affinity, and with high specificity for Ras proteins over their close family members. The peptides directly engage the Ras effector domain, and can block Ras from binding a canonical effector protein in the context of cancer cell lysates. Based on a series of observations, we hypothesize that the peptides bind Ras as head-to-tail homodimers, and report preliminary attempts to exploit this observation and identify peptides with improved affinity to Ras. Finally, we discuss the preliminary results from a conceptually related effort to identify peptide inhibitors of the Myc transcription factor, which is another protein heavily implicated in human cancer.