The role of particular residues of the PvuII endonuclease in DNA binding and cleavage was studied by mutational analysis using a number of in vivo and in vitro approaches. While confirming the importance of residues predicted to be involved directly in function by the crystal structure, the analysis led to several striking results. Aspartate 34, which contacts the central base pair of the PvuII site (5 -CAGCTG-3 ) through the minor groove, plays a critical role in binding specificity. A D34G mutant binds with high affinity to any of the sequences in the set CANNTG, although its low level of cleavage activity acts only on the wild-type site. In addition, a His to Ala mutation at the residue that contacts the central G and is predicted to be blocked by PvuII methylation still requires the PvuII methylase to be maintained in vivo, arguing against this hypothesis as the only mechanism for methylation protection. Finally, four of the five mutations that reduce cleavage activity while still exhibiting binding in the gel shift assay are at residues that form DNA-or subunit-subunit contacts rather than in the catalytic center. This provides further evidence for a strong linkage between specific binding and catalysis.The structures of the five type II endonucleases determined by x-ray diffraction, EcoRI (1), BamHI (2, 3), EcoRV (4, 5), PvuII (6, 7) and Cfr10I (8), share substantial elements of similarity. Analysis of the enzymes complexed to DNA, where known, suggests a preliminary classification into two different groups (9). The endonucleases that produce 5Ј-overhanging ends, EcoRI and BamHI, approach the DNA from the major groove, where most of their base-specific contacts are made. The endonucleases that produce blunt-ended DNA products, EcoRV and PvuII, approach the DNA from the minor groove side and establish contacts in the major groove by wrapping around the DNA. Structure-function studies have been limited to a small number of type II enzymes. Years of studies have produced an extensive amount of information about EcoRI (10 -18), EcoRV (18 -24), and NaeI (25) and, more recently, BamHI (26,27). These studies have focused on the identification and analysis of residues involved in catalysis, testing alternative mechanisms of catalysis, and understanding how a specific DNA sequence is recognized. Attempts to alter the sequence specificity have proven to be difficult, with successful examples limited to mutants showing relaxed specificity (10,11,28) or displaying activity toward DNA substituted with unnatural bases (15,29). One explanation for this difficulty is that a change in specificity not only requires recognition of a different DNA sequence, but also requires that the linkage between recognition and catalysis be retained. A class of mutations that might illuminate this linkage is the catalytic mutations, where specific DNA binding is retained but catalysis is reduced (26, 30). In particular, mutants in this class that are outside the catalytic center might be deficient in the linkage between recognition and ca...