The GLUT5 transporter catalyses the specific uptake Of D-fructose and can accept this hexose in its furanose and pyranose ring forms. The transporter does not accept fructose epimers and has very limited tolerance of bulky groups substituted at the 2-, 3-, 4- and 5-OH positions [Tatibouet, Yang, Morin and Holman (2000) Bioorg. Med. Chem. 8, 1825-1833]. To further explore whether bulky groups can be tolerated at the primary OH positions, a D-fructose analogue with an allylamine group substitution to replace the 1-OH group was synthesized and was found to be quite well tolerated (K-i = 27.1 mM). However, this analogue occurs in multiple ring forms. By contrast, 2,5-anhydro-D-mannitol is a symmetrical molecule that occurs only in a furanose ring form in which C-1 and C-6 are equivalent. We have therefore synthesized new 2,5-anhydro-D-mannitol analogues (substituted at the equivalent of the 6-OH Of D-fructose) and from studies in Chinese hamster ovary cells expressing GLUT5 cells report that (i) the allylamine derivative of 2,5-anhydro-D-mannitol is well tolerated (K-i = 2.66 mM); (ii) introduction of a di-nitrophenyl-substituted secondary amine group enhances affinity (K-i = 0.56 mM); (iii) introduction of amide-linked biotinylated photolabel moieties is possible without loss of affinity relative to 2,5-anhydro-D-mannitol but a small secondary amine spacer between the biotinylated photolabelling moiety and the fructofuranose ring increases affinity (fructose photolabel 2; A = 1.16 mM); (iv) introduction of a hydrophilic tartarate spacer between biotin and the diazirine photoreactive groups can be accomplished without reduction in affinity and (v) photoactivation of biotinylated fructose photolabels leads to specific biotin tagging of GLUT5. These data suggest that substitution of a secondary amine group (-NH) to replace the C-6 (or C-1) -OH of 2,5-anhydro-D-mannitol results in compounds of high affinity; the affinity is enhanced over 10-fold compared with D-fructose.