A structure-based model describing the interaction of the two-domain PI-SceI endonuclease with its 31-base pair DNA substrate suggests that the endonuclease domain (domain II) contacts the cleavage site region of the substrate, while the protein splicing domain (domain I) interacts with a distal region that is sufficient for high affinity binding. To support this model, alanine-scanning mutagenesis was used to assemble a set of 49 PISceI mutant proteins that were purified and assayed for their DNA binding and cleavage properties. Fourteen mutant proteins were 4-to >500-fold less active than wild-type PI-SceI in cleavage assays, and one mutant (T225A) was 3-fold more active. The yeast PI-SceI endonuclease catalyzes the hydrolysis of two specific phosphodiester bonds within an asymmetrical recognition site (1). This enzyme is a homing endonuclease (for a review, see Ref. 2) that occurs as an intein situated within an H Ļ© -ATPase protein subunit. Like other homing endonucleases, PI-SceI recognizes an extremely long sequence (31 bp) 1 and cuts DNA to yield 5Š-phosphate and 3Š-hydroxyl ends (3, 4). Mutagenesis and biochemical studies indicate that the PI-SceI recognition sequence can be divided into two regions (4, 5).Region I contains the cleavage site that is cut by the enzyme to generate a 4-base pair overhang, and region II includes an adjacent 17-bp sequence (the minimal binding sequence) that is sufficient for high affinity binding. Mutagenesis of the substrate reveals that PI-SceI tolerates substitutions at numerous positions, since substitutions at only nine positions in the substrate lead to severely reduced activity (4). Like the other homing endonucleases that have been studied, PI-SceI requires Mg 2Ļ© as a cofactor. The metal ion is likely to be required for the hydrolytic reaction, since it is required for catalysis but not for specific binding. Mn 2Ļ© can substitute for Mg 2Ļ© , and it stimulates more efficient cleavage by the enzyme at cognate and noncognate sites (1, 5).The three-dimensional structure of PI-SceI has been recently determined by x-ray crystallography and reveals a bipartite domain structure (6). Domain I contains the protein splicing active site, which is composed of the N-and C-terminal amino acids and two other His residues that have been shown to be required for activity or have been implicated in the reaction (7,8). The residues that compose the putative endonucleolytic active site, a lysine (Lys 301 ) and two aspartic acid residues (Asp 218 and Asp 326 ), are present in domain II and form a catalytic triad that displays structural similarity to charged clusters found in restriction enzymes (6, 9). By using the PISceI structural information and the knowledge that the enzyme contacts two discrete regions of the recognition sequence, a model for the docking of PI-SceI with its substrate was constructed where domains I and II of the protein contact regions II and I, respectively, of the substrate (Fig. 1). In this model, both domains are proposed to contact the substrate, since the bindi...