Baculovirus chitinases and other family 18 glycohydrolases have been shown to possess both exo- and endochitinase activities when assayed against fluorescent chito-oligosaccharides. Homology modelling of the chitinase of Epiphyas postvittana nucleopolyhedrovirus (EppoNPV) against Serratia marcescens chitinase A indicated that the enzyme possesses an N-terminal polycystic kidney 1 (PKD1) domain for chitin-substrate feeding and an alpha/beta TIM barrel catalytic domain characteristic of a family 18 glycohydrolase. EppoNPV chitinase has many features in common with other baculovirus chitinases, including high amino acid identity, an N-terminal secretion signal and a functional C-terminal endoplasmic reticulum-retention sequence. EppoNPV chitinase displayed exo- and endochitinolytic activity against fluorescent chito-oligosaccharides, with K(m) values of 270+/-60 and 240+/-40 microM against 4MU-(GlcNAc)2 and 20+/-6 and 14+/-7 microM against 4MU-(GlcNAc)3 for native and recombinant versions of the enzyme, respectively. In contrast, digestion and thin-layer chromatography analysis of short-chain (GlcNAc)(2-6) chito-oligosaccharides without the fluorescent 4-methylumbelliferone (4MU) moiety produced predominantly (GlcNAc)2, indicating an exochitinase, although low-level endochitinase activity was detected. Digestion of long-chain colloidal beta-chitin and analysis by mass spectrometry identified a single 447 Da peak, representing a singly charged (GlcNAc)2 complexed with a sodium adduct ion, confirming the enzyme as an exochitinase with no detectable endochitinolytic activity. Furthermore, (GlcNAc)(3-6) substrates, but not (GlcNAc)2, acted as inhibitors of EppoNPV chitinase. Short-chain substrates are unlikely to interact with the aromatic residues of the PKD1 substrate-feeding mechanism and hence may not accurately reflect the activity of these enzymes against native substrates. Based upon these results, the chitinase of the baculovirus EppoNPV is an exochitinase.