Protein kinases have proved to be largely resistant to the design of highly
specific inhibitors, even with the aid of combinatorial chemistry(1). The
lack of these reagents has complicated efforts to assign specific signallin
g roles to individual kinases. Here we describe a chemical genetic strategy
for sensitizing protein kinases to cell-permeable molecules that do not in
hibit wildtype kinases(2). From two inhibitor scaffolds, we have identified
potent and selective inhibitors for sensitized kinases from five distinct
subfamilies. Tyrosine and serine/threonine kinases are equally amenable to
this approach. We have analysed a budding yeast strain carrying an inhibito
r-sensitive form of the cyclin-dependent kinase Cdc28 (CDK1) in place of th
e wild-type protein. Specific inhibition of Cdc28 in vivo caused a pre-mito
tic cell-cycle arrest that is distinct from the G1 arrest typically observe
d in temperature-sensitive cdc28 mutants(3). The mutation that confers inhi
bitor-sensitivity is easily identifiable from primary sequence alignments.
Thus, this approach can be used to systematically generate conditional alle
les of protein kinases, allowing for rapid functional characterization of m
embers of this important gene family.