From purulent cystic fibrosis (CF) sputum, previous investigators partially purified a trypsinlike protease. A similar purified enzyme is available commercially as "human sputum trypsin." To explore the nature and origin of this preparation, we purified and NH2 terminally sequenced its major protein component. The resulting sequence, Ile-Val-Gly-Gly-Tyr-Thr-(Cys)-Ala-Ala-Asn-Ser-Val/Ile-Pro-Tyr-Gln-Val -Ser-Leu-Asn-Ser, differs from known human proteins but is identical to porcine trypsin, including the Val/Ile polymorphism at residue 12. Specific activity and electrophoretic and inhibition profiles and immunoreactivity of sputum and porcine pancreatic trypsin are nearly identical. Because porcine trypsin is a major ingredient of digestive enzyme supplements taken by CF patients with pancreatic dysfunction, we propose that one or more lots of human sputum trypsin derive from enzyme supplements and are of porcine origin. The path by which trypsin ends up in sputum is unknown. Because sputum trypsin is active but susceptible to inactivation by plasma alpha1-proteinase inhibitor, it is unlikely to derive from trypsin absorbed into the bloodstream. However, it may originate from tracheally aspirated stomach contents or from digestive supplement-contaminated saliva mixed with expectorated sputum. The imbalance between proteases and antiproteases in CF bronchial secretions allows trypsin to remain active despite sensitivity to serpins and secretory leukocyte proteinase inhibitor. Furthermore, because sputum trypsin activates human progelatinase B, it may be responsible in part for the reported presence of activated matrix metalloproteinases in CF sputum.