The Shokat lab uses chemical genetics to study and target signaling proteins, with the aim of finding new ways to treat human diseases. Many projects in the lab have focused on kinases, GTPases and RNA helicases, including some of the most important proteins in cancer, such as mTOR, Ras and CDK4. We are also interested in other signaling proteins, such as the dual kinase/GTPase LRRK2, which is relevant in Parkinson’s disease, and the translational regulator helicase DDX3.

Chemical genetics combines the timing and dose control of small molecules with the absolute specificity of genetics. Small molecules are invaluable tools for studying proteins, but it can be hard to find one that targets a single member of a family of proteins. We take advantage of the chemical difference that results from a mutation to allow us to target a protein specifically, leaving other similar proteins untouched.

Here are a few examples of the lab’s projects to illustrate our research strategy:

Drugging the most common oncogene, K-Ras: The small GTPase K-Ras is the most commonly mutated oncogene in human cancer, and cancers with mutated K-Ras are hard to treat with standard therapies. Developing small-molecule drugs to directly target K-Ras is difficult because K-Ras has extremely high affinity for GTP/GDP (the small molecules that it naturally binds) and no known allosteric regulatory sites. We pursued a chemical-genetic strategy and were able to develop small molecules to target a common oncogenic mutant, K-Ras G12C (PMID: 24256730). These compounds rely on the mutant cysteine for binding and so do not affect the wild-type protein. This approach has been recently validated in clinical studies, culminating in the FDA approval of sotorasib, the first approved K-Ras-targeting drug. We have also uncovered additional druggable vulnerabilities in K-RAS and are continuing to explore new therapeutic approaches in ongoing projects (PMID: 29033317)(PMID: 28621541)(PMID: 31138768).

Developing new mTOR inhibitors: Phosphatidylinositol 3-kinase α (PIK3CA) is the second-most-frequently mutated oncogene across all cancers. Downstream of PIK3CA in the growth factor pathway lies the mechanistic target of rapamycin (mTOR) kinase, which integrates nutrient and growth hormone availability. The natural product rapamycin inhibits mTOR by acting as a molecular glue, recruiting the chaperone protein FKBP12 to mTOR’s FRB domain to block substrate binding to mTOR. In 2008 we published PP42, a second-generation ATP-competitive inhibitor of mTOR (PMID: 19209957), which was later optimized for human clinical studies in collaboration with Intellikine and named INK128/sapanisertib (PMID: 22367541). Sapanisertib is undergoing phase II clinical studies, but we are also continuing to pursue new ways of targeting mTOR. In 2016 we reported a third-generation bitopic inhibitor of mTOR called RapaLink, which is made of rapamycin linked to INK128 by a flexible linker. RapaLink has exceptionally high affinity even for mutants of mTOR which are resistant to first- and second-generation inhibitors (PMID: 27279227). The RapaLink molecule has been modified for human clinical investigation (DOI: 10.1038/s41589-021-00813-7) and is now undergoing phase I trials (ClinicalTrials.gov identifier NCT04774952). We continue to explore chemical modifications to RapaLink and ways to exploit the molecular glue mechanism of rapamycin (bioRxiv 619551).

Identifying the substrates of every protein kinase: Matching protein kinases to their substrates is a major goal of many laboratories, but thousands of proteins are phosphorylated by hundreds of human kinases, so the problem is challenging. To address this problem we have devised a chemical-genetic method, often called the “bump–hole” approach, for identifying the direct substrates of any protein kinase (PMID: 9108016) using specialized analogs of ATP. (PMID: 17486086)(PMID: 18234856). These analogs are based on ATP-γ-S, so they transfer a thiophosphate group to kinase substrates, and they have bulky functional groups (the “bump”) at the N6 position that make them poor substrates for wild-type protein kinases. But a kinase of interest can be easily engineered to accept the bulky analogs without changing substrate specificity by mutating a residue in the ATP-binding site to create extra space (the “hole”). Thiophosphate-labeled substrate proteins can then be chemically captured and analyzed. This approach has been used to identify hundreds of novel substrates of over 50 widely divergent kinases, such as v-Src, CDK2, JNK, Cdc28, Erk2, Srb10, and Kin28 (PMID: 23836541). It is our long-term goal to identify all the direct substrates of each kinase in the human genome, and projects based on this technique are ongoing in the lab.

Protein Kinase Inhibitors: We have also applied the “bump-hole” approach to inhibit kinases that have never been targeted with specific inhibitors. The “hole” in the active site of a kinase of interest sensitizes it to inhibition by certain potent, cell-permeable small molecule inhibitors. This allows studying the effect of inhibiting the kinase in lysate, whole cells, tissues, and even whole organisms. Tyrosine and serine/threonine kinases are equally amenable to this approach, which has been used to selectively inhibit kinases in the Src, Abl, cyclin-dependent kinase (CDK), mitogen-activated kinase (MAPK), p21-activated kinase (PAK), and Ca2+/calmodulin-dependent kinase (CAMK) families, as well as over 50 other protein kinases. The ability to generate the very first inhibitors of many diverse protein kinases has enabled the discovery of fundamentally new roles of kinases in transcription, the cell cycle, cell-fate determination, the unfolded protein response, oncogenic transformation, and more.