Microsomal epoxide hydrolase (MEH) catalyzes the addition of water to epoxides in a two-step reaction involving initial attack of an active site carboxylate on the oxirane to give an ester intermediate followed by hydrolysis of the ester. An efficient bacterial expression system for the enzyme from rat that facilitates the production of native and mutant enzymes for mechanistic analysis is described. Pre-steady-state kinetics of the native enzyme toward glycidyl-4-nitrobenzoates, 1, indicate the rate-limiting step in the reaction is hydrolysis of the alkyl-enzyme intermediate. The enzyme is enantioselective, turning over (2R)-1 about 10-fold more efficiently than (2S)-1, and regiospecific toward both substrates with exclusive attack at the least hindered oxirane carbon. Facile isomerization of the monoglyceride product is observed and complicates the regiochemical analysis. The D226E and D226N mutants of the protein are catalytically inactive, behavior that is consistent with the role of D226 as the active-site nucleophile as suggested by sequence alignments with other alpha/beta-hydrolase fold enzymes. The D226N mutant undergoes hydrolytic autoactivation with a half-life of 9.3 days at 37 degreesC, suggesting that the mutant is still capable of catalyzing the hydrolytic half-reaction (in this instance an amidase reaction) and confirming that D226 is in the active site. The indoylyl side chain of W227, which is in or near the active site, is not required for efficient alkylation of the enzyme or for hydrolysis of the intermediate. However, the W227F mutant does exhibit altered stereoselectivity toward (2R)-1, (2S)-1, and phenanthrene-9,10-oxide, suggesting that modifications at this position might be used to manipulate the stereo- and regioselectivity of the enzyme.