Controlled release dosage forms of proteins and other biomolecules can be prepared by microencapsulating them in polymeric microspheres. Proteins are subjected to potentially damaging effects of sonication and exposure to organic solvents during the microencapsulation process. The relatively stable enzyme lysozyme was dissolved in aqueous buffer and sonicated in the presence of methylene chloride to mimic the initial step of the microencapsulation process. The stability of lysozyme was evaluated by determining the enzyme activity before and after sonication, size-exclusion chromatography, native polyacrylamide gel electrophoresis, and by measuring the amount of precipitates formed. Following sonication, the total protein introduced was distributed between a soluble and an insoluble fraction. Sonication of lysozyme solutions in the presence of methylene chloride led to an increase in precipitates. The precipitates were enzymatically inactive, did not dissolve easily, and were held by non-covalent interactions. No fragments or aggregates of lysozyme were detectable in the soluble fraction. Sonicating aqueous lysozyme solutions with and without methylene chloride decreased the specific activity of the enzyme in the soluble fraction. Excipients such as dimethyl sulfoxide (DMSO), mannitol, sucrose, and tween 80 were included in the sonication mixtures containing lysozyme. With the exception of tween 80, the addition of the excipients to aqueous solutions of lysozyme led to a greater decrease in the specific activity of lysozyme when sonicated in the presence of methylene chloride. DMSO caused the greatest loss of enzyme activity following sonication. Sonication of lysozyme with water, methylene chloride, and DMSO yielded methyl radicals, which were trapped with alpha-phenyl N-tert-butylnitrone and detected by ESR spectroscopy.