The racemic gluco-configured norbornanes 4 and 16 were prepared and tested as inhibitors of b-glucosidases. The known alcohol 5 was deprotected to provide the triol 6. Silylation (! 7), monobenzoylation (! 8/9), and oxidation provided the regioisomeric ketones 10 and 11. Reduction of 10 gave the desired endo-alcohol 13, albeit in low yield, while reduction of the isomeric ketone 11 provided mostly the altro-configured endo-alcohol 12. The alcohol 13 was desilylated to 14. Debenzoylation to 15 followed by hydrogenolytic deprotection gave the amino triol 4 that was reductively aminated to the benzylamine 16. The amino triols 4 and 16 proved weak inhibitors of the b-glucosidase from Caldocellum saccharolyticum (4: IC 50 = 5.6 mM; 16: IC 50 = 3.3 mM) and from sweet almonds (16: IC 50 = 5.5 mM). A comparison of 4 with the manno-configured norbornane 3 shows that 3 is a better inhibitor of snail b-mannosidase than 4 is of b-glucosidases, in keeping with earlier results suggesting that these b-glycosidases enforce a different conformational itinerary.Introduction. -b-Glycosidases impose a conformational change on their substrates to comply with the stereoelectronic control of glycoside cleavage [1 -4]. Several conformations allow the required coplanar arrangement of a non-bonding, doubly occupied orbital of the ring O-atom and the s* orbital of the scissile bond [5], and it was considered likely that all of them are enforced by different glycosidases [6]. Stable mimics of the individual acceptable conformers may act as selective inhibitors and contribute in assessing the conformational itinerary enforced by a specific glycosidase. It was claimed that inhibitors mimicking such distorted conformers should be inherently more selective than mimics of an intermediate oxocarbenium cation and be more sensitive to small geometric changes [7] [8]. In keeping with these considerations, the manno-configured isoquinuclidine 1 [9] mimicking a 1,4 B/