, 1993) (data not shown). However, when a blast search was performed, several proteins and cDNA sequences encoding for proteins having a significant sequence similarity were found (Fig. 3). A rooted phylogenetic tree of the Endo T sequence and 17 close matches are represented in Fig. 4. Three fungal proteins

from N. crassa, M. grisea and Podospora anserina are grouped together with Endo T (cluster A in Fig. 4). One can observe that these three species possess a common gene, probably originating ATM/ATR cancer from an ancient gene duplication (cluster B in Fig. 4). This duplicated gene also seems to occur in two other species: Botryotinia fuckeliana and Sclerotinia sclerotiorum. The latter species appear to have a gene, different from the Endo T gene, that is also originating from an ancient gene duplication (cluster C in Fig. 4). The presently purified enzyme was shown to be a Sotrastaurin research buy true ENGase: it released Man5–9GlcNAc structures from the glycoprotein RNAse B (Fig. 5a). The preparation is devoid of cellobiohydrolase I/endoglucanase I (CNP-Lac), α-mannosidase

(PNP-Man and Man9GlcNAc2), β-N-acetylglucosaminidase (PNP-GlcNAc) and chitinase (4MU-chitotriose and powdered chitin) activities (data not shown). The specific activity (220 mU mg−1) of Endo T contrasted to that found with Endo H from S. plicatus (5200 mU mg−1). Substrate specificity was examined with several glycoproteins. Band shifting on SDS-PAGE (not shown) and FACE analysis of the released N-glycans was performed for a qualitative comparison (Fig. 5). The release of high-mannose, hyperglycosylated and phosphorylated-type N-glycans from, respectively, RNAse B, Saccharomyces cerevisiae invertase BCKDHA and T. reesei Cel7A (Stals et al., 2004a) was readily observed (Fig. 5, gels A, B and D, lanes 3). Similarly to Endo H, Endo T does not catalyse the hydrolysis of any of the sialylated complex-type oligosaccharides, present

in fetuin (Fig. 5 gel C, lanes 1 and 3). The presence of single N-acetylglucosamine residues on N-glycosylation sites of T. reesei proteins (Klarskov et al., 1997; Bower et al., 1998; Hui et al., 2001, 2002; Stals et al., 2004b; Selinheimo et al., 2006) has been attributed to the action of intra- or extracellular ENGase-type activity in the fungus. Efforts to identify the Endo T gene/protein (Nevalainen et al., 1998) did not lead to clear-cut results probably due to the low sequence homology with other ENGases. From our work, it becomes evident that it cannot unambiguously be traced in the T. reesei genome (Martinez et al., 2008) without adequate sequence information. The purified enzyme is the first fungal representative in family GH18 with ENGase activity. Apart from the family gh18 motif, the homology of Endo T with the bacterial ENGases and the fungal chitinases from this family is very low. Database searches have identified several cDNA sequences encoding proteins and predicted proteins with high homology.