The second step of ATP binding to DnaK induces peptide release

The interaction of the nucleotide-free molecular chaperone DnaK (Hsp70) from Escherichia coli with nucleotides was studied under equilibrium and transient kinetic conditions. These studies used the intrinsic fluorescence signal of the single tryptophan residue (Trp102) of DnaK, or of novel fluorescent nucleotide analogs of ADP and ATP, N-8-(4-N'-methylanthraniloylaminobutyl)-8-aminoadenosine 5'-di- or triphosphate (MABA-ADP and MABA-ATP) as spectroscopic probes. Titration of MABA-ADP with DnaK resulted in a 2.3-fold increase of the fluorescence signal, from which a binding stoichiometry of 1:1, and a dissociation constant (K-d) of 0.09 mu M were derived. The intrinsic rate constant of hydrolysis of ATP or MABA-ATP in single turnover experiments was found to be 1.5 x 10(-3) s(-1) and 1.6 x 10(-3) s(-1), identical with the catalytic rate constant of 1.5(+/-0.17) x 10(-3) s(-1) obtained under steady-state conditions. The dissociation rate constant of ADP was measured to be 35(+/-7) x 10(-3) s(-1) in the absence or 15(+/-5) x 10(-3) s(-1) in the presence of 2 mM inorganic phosphate (P-i) and is therefore 10 to 20 times faster than the rate of hydrolysis. These results demonstrated that processes governing ATP hydrolysis are rate-limiting in the DnaK ATPase reaction cycle. The three observed different fluorescent states of the single tryptophan residue were investigated. The binding of ATP gave a decrease of 15% in fluorescence intensity compared with the nucleotide-free state. Subsequent ATP hydrolysis, or the simultaneous addition of ADP and P-i, increased the fluorescence 7% above the fluorescence intensity of the nucleotide-free protein. Changes in the tryptophan fluorescence could not be detected when ADP, P-i or the non-hydrolyzable nucleotide analogs AMPPNP (K-d = 1.62(+/-0.1) mu M) or ATP gamma S (K-d = 0.044(+/-0.003) mu M) were added. These data suggested that DnaK exists in at least three different conformational states, depending on nucleotide site occupancy. The fluorescence increase of DnaK upon ATP binding was resolved into two steps; a rapid first step (K(d)1 = 7.3 mu M) is followed by a second slow step (k(+2) = 1.5 s(-1) and k(-2) less than or equal to 1.5 x 10(-3) s(-1)) that causes the decrease in the tryptophan fluorescence signal. The addition of ATP also resulted in the release of DnaK-bound peptide substrate with k(off) = 3.8 s(-1), comparable with the rate of the second step of nucleotide binding. AMPPNP or ATP gamma S were not able to change the fluorescence signal nor to release the peptide. We therefore conclude that the second step of ATP binding, and not the 1000-fold slower ATP hydrolysis is coupled to peptide release. (C) 1996 Academic Press Limited