Mismatch repair proteins correct errors in DNA via an ATP-driven process. In eukaryotes, the Msh2-Msh6 complex recognizes base pair mismatches and small insertion/deletions in DNA and initiates repair. Both Msh2 and Msh6 proteins contain Walker ATP-binding motifs that are necessary for repair activity. To understand how these proteins couple ATP binding and hydrolysis to DNA binding/mismatch recognition, the ATPase activity of Saccharomyces cerevisiae Msh2-Msh6 was examined under presteady-state conditions. Acid-quench experiments revealed that in the absence of DNA, Msh2-Msh6 hydrolyzes ATP rapidly (burst rate = 3 s −1 at 20 °C) and then undergoes a slow step in the pathway that limits catalytic turnover (k cat = 0.1 s −1 ). ATP is hydrolyzed similarly in the presence of fully matched duplex DNA; however, in the presence of a G:T mismatch or +T insertioncontaining DNA, ATP hydrolysis is severely suppressed (rate = 0.1 s −1 ). Pulse-chase experiments revealed that Msh2-Msh6 binds ATP rapidly in the absence or in the presence of DNA (rate = 0.1 μM −1 s −1 ), indicating that for the Msh2-Msh6·mismatched DNA complex, a step after ATP binding but before or at ATP hydrolysis is the rate-limiting step in the pathway. Thus, mismatch recognition is coupled to a dramatic increase in the residence time of ATP on Msh2-Msh6. This mismatchinduced, stable ATP-bound state of Msh2-Msh6 likely signals downstream events in the repair pathway.Replicative DNA polymerases are responsible for accurately reproducing the genetic code of organisms; however, even the most accurate polymerases make errors that result in approximately one mismatched base pair per 10 7 nucleotides as well as insertion/deletion loops (1). These defects must be corrected prior to subsequent rounds of replication, to minimize accumulation of potentially deleterious mutations and genome instability. A multi-protein mismatch repair system is responsible for this task, which involves recognition and removal of the defects followed by replacement with correct DNA. This repair system was initially identified and studied extensively in bacteria, and homologous systems were discovered in a variety of other organisms, including humans (2). The critical role of DNA mismatch repair in maintaining genome and cellular integrity is highlighted by the links between defective repair protein function and predisposition to cancer (e.g., hereditary nonpolyposis colorectal cancer (3)).The repair process begins with DNA binding and mismatch/defect recognition by prokaryotic MutS or eukaryotic MutS homologue proteins. MutS/Msh 1 proteins then signal other proteins downstream in the pathway to initiate DNA excision. In bacteria, MutH endonuclease nicks the unmethylated new DNA strand at a GATC site, which is followed by UvrD/Helicase II and † This work was supported by a grant from the N.I.H. .