The LD(50) for soman is 10-20-fold higher for a mouse than a human. The difference in susceptibility is attributed to the presence of carboxylesterase in mouse but not in human plasma. Our goal was to make a mouse lacking plasma carboxylesterase. We used homologous recombination to inactivate the carboxylesterase ES1 gene on mouse chromosome 8 by deleting exon 5 and by introducing a frame shift for amino acids translated from exons 6 to 13. ES1-/- mice have no detectable carboxylesterase activity in plasma but have normal carboxylesterase activity in tissues. Homozygous ES1-/- mice and wild-type littermates were tested for response to a nerve agent model compound (soman coumarin) at 3 mg/kg sc. This dose intoxicated both genotypes but was lethal only to ES1-/- mice. This demonstrated that plasma carboxylesterase protects against a relatively high toxicity organophosphorus compound. The ES1-/- mouse should be an appropriate model for testing highly toxic nerve agents and for evaluating protection strategies against the toxicity of nerve agents.