Inferences concerning the ATPase properties of DnaK and other HSP70s are affected by the ADP kinase activity of copurifying nucleoside-diphosphate kinase

Preparations of Escherichia coli DnaK from our lab as well as preparations of DnaK and other HSP70 proteins from several major labs in the field produce a stoichiometric initial burst of [alpha-P-32]ADP when incubated with [alpha-P-32]ATP and contain an ADP kinase activity. We determined that the initial burst activity results from the transfer of gamma-phosphate from the radiolabeled substrate [alpha-P-32]ATP to unlabeled ADP bound by the DnaK and is the same activity that results in ADP phosphorylation. The purification of DnaK from E. coli cells that carry a disrupted ndk gene, ndk::km, results in preparations with greatly reduced ADP kinase activities compared with preparations of DnaK purified from ndk(+) cells. The reduction in the amount of ADP kinase activity in preparations of DnaK purified from ndk::lm cells shows that nucleoside-diphosphate kinase (NDP kinase) is responsible for most of the ADP kinase activity present in DnaK preparations isolated from ndk(+) cells. The remaining ADP kinase activity in preparations hom ndk::km cells, which varies between preparations, is also a property of NDP kinase, which is most likely expressed because of a low frequency reversion of the disrupted ndk gene. A weak, but measurable physical interaction exists between DnaK. and NDP kinase and may be at least partially responsible for the co-purification of NDP kinase with DnaK The presence of contaminating NDP kinase can explain the range of k(cat) values reported for the ATPase activity of DnaK as well as recent reports of initial burst kinetics by DnaK (Banecki, B., and Zylicz, M. (1996) J. Biol. Chem. 271, 6137-6143) and an ADP-ATP exchange activity of DnaK (Hiromura, M., Yano, M., Mori, H., Inoue, M., and Kido, H. (1998) J. Biol. Chem. 273, 5435-5438).