Abstract: Reverse transcriptase polymerase chain reaction (RT-PCR) was employed to amplify cDNAs constructed from Trimeresurus stejnegeri venom glands poly(A)+ RNA to facilitate the cloning and sequencing of serine protease genes.The PCR products were then subcloned into pGEM-T vector and transformed into E.coli strain JM109.Five serine protease cDNAs (named TSSP-1,-2,-3,-4 and -5,respectively) were cloned and sequenced.The protease encoded by TSSP-1 is characterized by a His41-Arg41 mutation in the catalytic triad.Combined the results of N-terminal sequence determinations of the purified proteins,it is concluded that TSSP-2,-3 and -4 encode stejnobin (thrombin-like enzyme),and stejnefibrase 1 and 2 (fibrinolytic enzymes),respectively.Mature proteases encoded by TSSP-1 and -2 are composed of 236 amino acid residues,and those proteases encoded by TSSP-3,-4 and -5 are all composed of 234 amino acid residues.The cloned 5 mature proteases contain 1 to 6 N-type glycosylation sites,indicating that the differences among calculated molecular weights and apparent molecular weights determined by SDS-PAGE are caused by carbohydrate content variations.Sequence identities among the cloned proteases are around 60%-90%.The differences in their structures resulted in their various substrate specificities.