The native form of serpins (serine protease inhibitors) is a metastable conformation, which converts into a more stable form upon complex formation with a target protease. It has been suggested that movement of helix-F (hF) and the following loop connecting to strand 3 of beta-sheet A (thFs3A) is critical for such conformational change. Despite many speculations inferred from analysis of the serpin structure itself, direct experimental evidence for the mobilization of hF/thFs3A during the inhibition process is lacking. To probe the mechanistic role of hF and thFs3A during protease inhibition, a disulfide bond was engineered in alpha(1)-antitrypsin, which would lock the displacement of thFs3A from beta-sheet A. We measured the inhibitory activity of each disulfide-locked mutant and its heat stability against loop-sheet polymerization. Presence of a disulfide between thFs3A and s5A but not between thFs3A and s3A caused loss of the inhibitory activity, suggesting that displacement of hF/thFs3A from strand 5A but not from strand 3A is required during the inhibition process. While showing little influence on the inhibitory activity, the disulfide between thFs3A and s3A retarded loop-sheet polymerization significantly. This successful protein engineering of alpha(1)-antitrypsin is expected to be of value in clinical applications. Based on our current studies, we propose that the reactive-site loop of a serpin glides through between s5A and thFs3A for the full insertion into beta-sheet A while a substantial portion of the interactions between hF and s3A is kept intact.