A yeast-based assay for protein tyrosine kinase substrate specificity and inhibitor resistance

Date

2017-12

Authors

Taft, Joseph Michael

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Phosphorylation of tyrosines by protein kinases is a fundamental mode of signal transduction in all eukaryotic cells, leading to a wide variety of cellular outcomes, including proliferation, differentiation, transcriptional activation, and programmed cell death. Perturbations to tyrosine kinase signaling networks by activation, overexpression, or mutation is the driving factor in many diseases, most notably cancers. The development of tyrosine kinase inhibitors, 37 of which are currently FDA-approved, has led to a revolution in cancer treatment. Imatinib, the first FDA-approved kinase inhibitor, has drastically improved prognosis for patients with Bcr-abl-positive leukemias. Despite this unprecedented success, however, up to one-third of patients lose response to imatinib due to mutations within the tyrosine kinase domain of Bcr-abl. Subsequent generations of Bcr-abl inhibitors, including dasatinib and ponatinib, have been developed to overcome these resistance mutations, but in each case, novel resistance mutations have arisen. We present a high-throughput yeast-based assay for the prediction of dasatinib- and ponatinib-resistant mutations in the ABL1 kinase domain. Our results not only recapitulate all known dasatinib-resistant mutations, but confirm recent patient data emphasizing the importance of compound mutations in ponatinib resistance. Furthermore, with hundreds of kinase inhibitors in development for the treatment of a wide range of diseases, understanding the cellular pathway of each kinase is critically important to the selection of ideal drug targets and avoiding potentially toxic side effects. Discovery of novel tyrosine kinase substrates is hindered by the presence of 90 human tyrosine kinases, which are often active in the same pathways. Phosphoproteomics, chemical genetics, and in vitro assays have been used to great success, yet only 30% of phosphorylated tyrosines in the human proteome have been assigned to a specific kinase. Recent advances in predicting tyrosine kinase substrates have been made by combining large data sets on kinase domain specificity, cellular localization, and protein-protein interactions in probabilistic algorithms. However, the high-quality data sets required for accurate predictions are often lacking. In chapter 2, we present a high-throughput yeastbased assay for screening millions of putative kinase substrates, which we then use to build a probabilistic model to accurately predict the in vitro phosphorylation of candidate substrates

Department

Description

LCSH Subject Headings

Citation