Synthesis of New C- and N-β-d-Glucopyranosyl Derivatives of Imidazole, 1,2,3-Triazole and Tetrazole, and Their Evaluation as Inhibitors of Glycogen Phosphorylase

The aim of the present study was to broaden the structure-activity relationships of C- and N-β-d-glucopyranosyl azole type inhibitors of glycogen phosphorylase. 1-Aryl-4-β-d-gluco-pyranosyl-1,2,3-triazoles were prepared by copper catalyzed azide-alkyne cycloadditions between O-perbenzylated or O-peracetylated β-d-glucopyranosyl ethynes and aryl azides. 1-β-d-Gluco-pyranosyl-4-phenyl imidazole was obtained in a glycosylation of 4(5)-phenylimidazole with O-peracetylated α-d-glucopyranosyl bromide. C-β-d-Glucopyranosyl-N-substituted-tetrazoles were synthesized by alkylation/arylation of O-perbenzoylated 5-β-d-glucopyranosyl-tetrazole or from a 2,6-anhydroheptose tosylhydrazone and arenediazonium salts. 5-Substituted tetrazoles were glycosylated by O-peracetylated α-d-glucopyranosyl bromide to give N-β-d-glucopyranosyl-C-substituted-tetrazoles. Standard deprotections gave test compounds which were assayed against rabbit muscle glycogen phosphorylase b. Most of the compounds proved inactive, the best inhibitor was 2-β-d-glucopyranosyl-5-phenyltetrazole (IC50 600 μM). These studies extended the structure-activity relationships of β-d-glucopyranosyl azole type inhibitors and revealed the extreme sensitivity of such type of inhibitors towards the structure of the azole moiety.


Results and Discussion
For the syntheses of C-glucosyl 1,2,3-triazoles of type IX several methods were published and this chemistry was reviewed last year [44]. Our work, summarized in Table 2, started with O-perbenzylated C-glucosyl acetylene 1 described in the literature [45]. Copper catalyzed azide-alkyne cycloaddition (CuAAC) [46] was effected from 1 either by pre-formed aryl azides with CuO(CO)C 3 H 7 (PPh 3 ) 2 as the catalyst [47] (method a) or azides obtained in situ from areneboronic acids [48,49] (method b) to give 1,2,3-triazoles 2a-c in very good yields. Removal of the O-benzyl protecting groups from 2a by usual catalytic hydrogenation (method c) gave excellent yield of 5a, however, under the same conditions 2c gave an inseparable product mixture. After O-peracetylation (method e) of the mixture the products could be separated and identified as 4c and a partially saturated derivative 4d. Since the formation of a tetrahydronaphthyl by-product under catalytic hydrogenation was observed previously with a 2-naphthyl substituted C-glucopyranosyl 1,2,4-triazole [50] hydrogenolytic deprotection of 2b was not attempted. Instead, the protecting groups were exchanged to acetate esters as reported to get O-peracetylated acetylene 3 [51]. CuAAC from 3 produced triazoles 4b and 4c in very good yields. Formation of 4b was also effected from 2b by a direct exchange of protective groups by method d [51]. Removal of the O-acetyl groups from 4b,c under Zemplén conditions (method f ) gave the targeted 5b,c in excellent yields.
For the preparation of an imidazole of type X a literature method [52] was adapted. Thus, acetobromoglucose 6 was reacted with 4-phenyl-imidazole in the presence of Hg(CN) 2 in acetonitrile to give 1-glucopyranosyl-4-phenyl-imidazole 7 (Scheme 1). Due to the tautomerism of imidazoles the formation of the isomeric 1-glucopyranosyl-5-phenyl-imidazole would also be possible, however, this was excluded on the basis of a HMBC measurement. Specifically, the observation of cross peaks between H-1'-C-2, H-1'-C-5, C-1'-H-2, and C-1'-H-5 clearly indicated the formation of 7. O-Deacetylation of 7 by the Zemplén method gave 8 in good yield. Actually, these compounds are the best glucose analogue inhibitors of GP known to date. Their efficiency, among other factors, is due to the formation of a H-bridge between the heterocycle and the His-377 main chain carbonyl group in the active site of the enzyme. 5-Glucosyltetrazole VIII [35], although capable of H-bond formation, proved inactive. It is remarkable that, besides the properties of the heterocycle, also the nature and the size of the substituent of the azole moiety had a very significant influence on the activity of the inhibitors. While methyl substituted derivatives in column a were practically inactive, a phenyl appendage (column b) made much more efficient compounds and the 2-naphthyl derivatives (column c) proved to be the strongest inhibitors. With these preliminaries in mind, in order to make the structure-activity relationship of N-and C-β-Dglucopyranosyl azole type compounds more complete, we envisaged to synthesize the "missing" counterparts of the above glucose derivatives. In this paper the syntheses and enzymatic evaluation of 4-β-D-glucopyranosyl-1-substituted-1,2,3-triazoles IX, 1-β-D-glucopyranosyl-4-substituted imidazoles X, 5-β-D-glucopyranosyl-2-substituted tetrazoles XI, and 2-β-D-glucopyranosyl-5substituted tetrazoles XII are presented.

Results and Discussion
For the syntheses of C-glucosyl 1,2,3-triazoles of type IX several methods were published and this chemistry was reviewed last year [44]. Our work, summarized in Table 2, started with Operbenzylated C-glucosyl acetylene 1 described in the literature [45]. Copper catalyzed azide-alkyne cycloaddition (CuAAC) [46] was effected from 1 either by pre-formed aryl azides with CuO(CO)C3H7(PPh3)2 as the catalyst [47] (method a) or azides obtained in situ from areneboronic acids [48,49] (method b) to give 1,2,3-triazoles 2a-c in very good yields. Removal of the O-benzyl protecting groups from 2a by usual catalytic hydrogenation (method c) gave excellent yield of 5a, however, under the same conditions 2c gave an inseparable product mixture. After O-peracetylation (method e) of the mixture the products could be separated and identified as 4c and a partially saturated derivative 4d. Since the formation of a tetrahydronaphthyl by-product under catalytic hydrogenation was observed previously with a 2-naphthyl substituted C-glucopyranosyl 1,2,4triazole [50] hydrogenolytic deprotection of 2b was not attempted. Instead, the protecting groups were exchanged to acetate esters as reported to get O-peracetylated acetylene 3 [51]. CuAAC from 3 produced triazoles 4b and 4c in very good yields. Formation of 4b was also effected from 2b by a direct exchange of protective groups by method d [51]. Removal of the O-acetyl groups from 4b,c under Zemplén conditions (method f) gave the targeted 5b,c in excellent yields.
For the preparation of an imidazole of type X a literature method [52] was adapted. Thus, acetobromoglucose 6 was reacted with 4-phenyl-imidazole in the presence of Hg(CN)2 in acetonitrile to give 1-glucopyranosyl-4-phenyl-imidazole 7 (Scheme 1). Due to the tautomerism of imidazoles the formation of the isomeric 1-glucopyranosyl-5-phenyl-imidazole would also be possible, however, this was excluded on the basis of a HMBC measurement. Specifically, the observation of cross peaks between H-1'-C-2, H-1'-C-5, C-1'-H-2, and C-1'-H-5 clearly indicated the formation of 7. O-Deacetylation of 7 by the Zemplén method gave 8 in good yield.    (f) ~1M NaOMe in MeOH, r.t.

Ar
Conditions and Yields (%) Next we turned to the synthesis of C-glucopyranosyl tetrazoles of type XI. While 5-(β-D-glucopyranosyl)tetrazoles (e.g., 9) are long known compounds ( [35,49] and references cited therein), no Nsubstituted derivatives could be located in the literature. For the preparation of the phenyl substituted derivatives a copper catalyzed reaction [53] of 9 and benzeneboronic acid was applied (Table 3, conditions a). Although the tautomerism of the tetrazole moiety could have facilitated the formation of regioisomers, only 10a was obtained in excellent yield as it was claimed in the cited paper. By modifying a literature procedure [54], compound 10a was also prepared, albeit in lower yield, from tosylhydrazone 12 [55,56] and benzenediazonium tetrafluoroborate [57,58] (conditions c). For the methylation of 9 a method [59] applied for the synthesis of C-glycofuranosyl tetrazoles was adapted. Thus, 9 was reacted with diazomethane to give a 1:1 mixture of the regioisomeric tetrazoles 10e and 11e in very good overall yield (conditions b). Removal of the ester protecting groups by the Zemplén protocol gave the test compounds 13 and 14 in very good yields (conditions d). The regioisomers of the formed C,N-disubstituted tetrazoles could easily be identified by the 13 C-NMR signal of the C-5 carbons. It is well known that the tetrazole carbon of 2,5-disubstituted derivatives (162-167 ppm) is shifted downfield by ~10 ppm in comparison to that of the 1,5-disubstituted counterparts (152-156 ppm) [60], and this is clearly visible in the obtained data shown in Table 3. In addition, for 11e 1 H-1 H NOEs were observed between the CH3 protons and the pyranose H-1' and H-2', while for 10e the NOE spectrum did not indicate proximity between the substituents of the tetrazole.
Next we turned to the synthesis of C-glucopyranosyl tetrazoles of type XI. While 5-(β-D-glucopyranosyl)tetrazoles (e.g., 9) are long known compounds ( [35,49] and references cited therein), no Nsubstituted derivatives could be located in the literature. For the preparation of the phenyl substituted derivatives a copper catalyzed reaction [53] of 9 and benzeneboronic acid was applied (Table 3, conditions a). Although the tautomerism of the tetrazole moiety could have facilitated the formation of regioisomers, only 10a was obtained in excellent yield as it was claimed in the cited paper. By modifying a literature procedure [54], compound 10a was also prepared, albeit in lower yield, from tosylhydrazone 12 [55,56] and benzenediazonium tetrafluoroborate [57,58] (conditions c). For the methylation of 9 a method [59] applied for the synthesis of C-glycofuranosyl tetrazoles was adapted. Thus, 9 was reacted with diazomethane to give a 1:1 mixture of the regioisomeric tetrazoles 10e and 11e in very good overall yield (conditions b). Removal of the ester protecting groups by the Zemplén protocol gave the test compounds 13 and 14 in very good yields (conditions d). The regioisomers of the formed C,N-disubstituted tetrazoles could easily be identified by the 13 C-NMR signal of the C-5 carbons. It is well known that the tetrazole carbon of 2,5-disubstituted derivatives (162-167 ppm) is shifted downfield by ~10 ppm in comparison to that of the 1,5-disubstituted counterparts (152-156 ppm) [60], and this is clearly visible in the obtained data shown in Table 3. In addition, for 11e 1 H-1 H NOEs were observed between the CH3 protons and the pyranose H-1' and H-2', while for 10e the NOE spectrum did not indicate proximity between the substituents of the tetrazole.
Next we turned to the synthesis of C-glucopyranosyl tetrazoles of type XI. While 5-(β-D-glucopyranosyl)tetrazoles (e.g., 9) are long known compounds ( [35,49] and references cited therein), no Nsubstituted derivatives could be located in the literature. For the preparation of the phenyl substituted derivatives a copper catalyzed reaction [53] of 9 and benzeneboronic acid was applied (Table 3, conditions a). Although the tautomerism of the tetrazole moiety could have facilitated the formation of regioisomers, only 10a was obtained in excellent yield as it was claimed in the cited paper. By modifying a literature procedure [54], compound 10a was also prepared, albeit in lower yield, from tosylhydrazone 12 [55,56] and benzenediazonium tetrafluoroborate [57,58] (conditions c). For the methylation of 9 a method [59] applied for the synthesis of C-glycofuranosyl tetrazoles was adapted. Thus, 9 was reacted with diazomethane to give a 1:1 mixture of the regioisomeric tetrazoles 10e and 11e in very good overall yield (conditions b). Removal of the ester protecting groups by the Zemplén protocol gave the test compounds 13 and 14 in very good yields (conditions d). The regioisomers of the formed C,N-disubstituted tetrazoles could easily be identified by the 13 C-NMR signal of the C-5 carbons. It is well known that the tetrazole carbon of 2,5-disubstituted derivatives (162-167 ppm) is shifted downfield by ~10 ppm in comparison to that of the 1,5-disubstituted counterparts (152-156 ppm) [60], and this is clearly visible in the obtained data shown in Table 3. In addition, for 11e 1 H-1 H NOEs were observed between the CH3 protons and the pyranose H-1' and H-2', while for 10e the NOE spectrum did not indicate proximity between the substituents of the tetrazole.
Next we turned to the synthesis of C-glucopyranosyl tetrazoles of type XI. While 5-(β-D-glucopyranosyl)tetrazoles (e.g., 9) are long known compounds ( [35,49] and references cited therein), no Nsubstituted derivatives could be located in the literature. For the preparation of the phenyl substituted derivatives a copper catalyzed reaction [53] of 9 and benzeneboronic acid was applied (Table 3, conditions a). Although the tautomerism of the tetrazole moiety could have facilitated the formation of regioisomers, only 10a was obtained in excellent yield as it was claimed in the cited paper. By modifying a literature procedure [54], compound 10a was also prepared, albeit in lower yield, from tosylhydrazone 12 [55,56] and benzenediazonium tetrafluoroborate [57,58] (conditions c). For the methylation of 9 a method [59] applied for the synthesis of C-glycofuranosyl tetrazoles was adapted. Thus, 9 was reacted with diazomethane to give a 1:1 mixture of the regioisomeric tetrazoles 10e and 11e in very good overall yield (conditions b). Removal of the ester protecting groups by the Zemplén protocol gave the test compounds 13 and 14 in very good yields (conditions d). The regioisomers of the formed C,N-disubstituted tetrazoles could easily be identified by the 13 C-NMR signal of the C-5 carbons. It is well known that the tetrazole carbon of 2,5-disubstituted derivatives (162-167 ppm) is shifted downfield by ~10 ppm in comparison to that of the 1,5-disubstituted counterparts (152-156 ppm) [60], and this is clearly visible in the obtained data shown in Table 3. In addition, for 11e 1 H-1 H NOEs were observed between the CH3 protons and the pyranose H-1' and H-2', while for 10e the NOE spectrum did not indicate proximity between the substituents of the tetrazole.   Next we turned to the synthesis of C-glucopyranosyl tetrazoles of type XI. While 5-(β-D-glucopyranosyl)tetrazoles (e.g., 9) are long known compounds ( [35,49] and references cited therein), no Nsubstituted derivatives could be located in the literature. For the preparation of the phenyl substituted derivatives a copper catalyzed reaction [53] of 9 and benzeneboronic acid was applied (Table 3, conditions a). Although the tautomerism of the tetrazole moiety could have facilitated the formation of regioisomers, only 10a was obtained in excellent yield as it was claimed in the cited paper. By modifying a literature procedure [54], compound 10a was also prepared, albeit in lower yield, from tosylhydrazone 12 [55,56] and benzenediazonium tetrafluoroborate [57,58] (conditions c). For the methylation of 9 a method [59] applied for the synthesis of C-glycofuranosyl tetrazoles was adapted. Thus, 9 was reacted with diazomethane to give a 1:1 mixture of the regioisomeric tetrazoles 10e and 11e in very good overall yield (conditions b). Removal of the ester protecting groups by the Zemplén protocol gave the test compounds 13 and 14 in very good yields (conditions d). The regioisomers of the formed C,N-disubstituted tetrazoles could easily be identified by the 13 C-NMR signal of the C-5 carbons. It is well known that the tetrazole carbon of 2,5-disubstituted derivatives (162-167 ppm) is shifted downfield by ~10 ppm in comparison to that of the 1,5-disubstituted counterparts (152-156 ppm) [60], and this is clearly visible in the obtained data shown in Table 3. In addition, for 11e 1 H-1 H NOEs were observed between the CH3 protons and the pyranose H-1' and H-2', while for 10e the NOE spectrum did not indicate proximity between the substituents of the tetrazole.
--e 3 (from 2c) --Next we turned to the synthesis of C-glucopyranosyl tetrazoles of type XI. While 5-(β-D-glucopyranosyl)tetrazoles (e.g., 9) are long known compounds ( [35,49] and references cited therein), no N-substituted derivatives could be located in the literature. For the preparation of the phenyl substituted derivatives a copper catalyzed reaction [53] of 9 and benzeneboronic acid was applied (Table 3, conditions a). Although the tautomerism of the tetrazole moiety could have facilitated the formation of regioisomers, only 10a was obtained in excellent yield as it was claimed in the cited paper. By modifying a literature procedure [54], compound 10a was also prepared, albeit in lower yield, from tosylhydrazone 12 [55,56] and benzenediazonium tetrafluoroborate [57,58] (conditions c). For the methylation of 9 a method [59] applied for the synthesis of C-glycofuranosyl tetrazoles was adapted. Thus, 9 was reacted with diazomethane to give a 1:1 mixture of the regioisomeric tetrazoles 10e and 11e in very good overall yield (conditions b). Removal of the ester protecting groups by the Zemplén protocol gave the test compounds 13 and 14 in very good yields (conditions d). The regioisomers of the formed C,N-disubstituted tetrazoles could easily be identified by the 13 C-NMR signal of the C-5 carbons. It is well known that the tetrazole carbon of 2,5-disubstituted derivatives (162-167 ppm) is shifted downfield by~10 ppm in comparison to that of the 1,5-disubstituted counterparts (152-156 ppm) [60], and this is clearly visible in the obtained data shown in Table 3. In addition, for 11e 1 H-1 H NOEs were observed between the CH 3 protons and the pyranose H-1' and H-2', while for 10e the NOE spectrum did not indicate proximity between the substituents of the tetrazole.  For the synthesis of N-(β-D-glucopyranosyl)-5-substituted-tetrazoles a literature protocol was applied to give 15a and 16a [61] in the reaction of acetobromoglucose 6 and 5-phenyltetrazole [62] ( Table 4). From a similar transformation of 6 with 5-methyltetrazole [62] only the 2,5-disubstituted 15e could be isolated in moderate yield and the formation of the HBr elimination product 2-acetoxy-D-glucal 17 was observed in a significant amount. Protecting group removal was effected by the Zemplén method to furnish the test compounds 18 and 19 in very good yields. The regioisomeric tetrazoles 15, 18 vs. 16, 19 were identified on the basis of the C-5 chemical shifts as described above (see respective data in Table 4).
The new compounds were assayed against rabbit muscle glycogen phosphorylase b enzyme (RMGPb) as described earlier [27] and the results are collected in Table 5. The inefficiency of Cglucopyranosyl 1,2,3-triazoles 5 (entries 4-6) as compared to the micromolar inhibition of the Nglucopyranosyl counterparts II in Table 2 came as a surprise, since the size of the heterocycle and the position of the H-bond donor and acceptor sites of the ring must not have been altered by the interchange of the substituents. A comparison of the inhibitory efficiency of N-benzoyl-β-D-glucopyranosylamine Ib (entry 1) with its "reversed" counterpart N-phenyl-2,6-anhydro-D-glycero-D-guloheptonamide 20 (shown in entry 3) results in a ratio of 38-67 (~53 as an average). Multiplication of the inhibition constants of IIb (entry 2) with this average factor to predict the efficiency of 5a (entry 4) gives values of ~7900-8500 μM, a range being well beyond the concentrations investigated in this study (max 625 μM). Nevertheless, our observations may refer to a strong directionality in the amide-1,2,3-triazole bioisosterism (mostly ignored in related studies [63,64]) indicating that the proper replacement must correspond to the pairs Ib-IIb and 20-5a in entries 1-2 and 3-4, respectively.  For the synthesis of N-(β-D-glucopyranosyl)-5-substituted-tetrazoles a literature protocol was applied to give 15a and 16a [61] in the reaction of acetobromoglucose 6 and 5-phenyltetrazole [62] ( Table 4). From a similar transformation of 6 with 5-methyltetrazole [62] only the 2,5-disubstituted 15e could be isolated in moderate yield and the formation of the HBr elimination product 2-acetoxy-D-glucal 17 was observed in a significant amount. Protecting group removal was effected by the Zemplén method to furnish the test compounds 18 and 19 in very good yields. The regioisomeric tetrazoles 15, 18 vs. 16, 19 were identified on the basis of the C-5 chemical shifts as described above (see respective data in Table 4).    The new compounds were assayed against rabbit muscle glycogen phosphorylase b enzyme (RMGPb) as described earlier [27] and the results are collected in Table 5. The inefficiency of C-glucopyranosyl 1,2,3-triazoles 5 (entries 4-6) as compared to the micromolar inhibition of the N-glucopyranosyl counterparts II in Table 2 came as a surprise, since the size of the heterocycle and the position of the H-bond donor and acceptor sites of the ring must not have been altered by the interchange of the substituents. A comparison of the inhibitory efficiency of N-benzoyl-β-D-gluco-pyranosylamine Ib (entry 1) with its "reversed" counterpart N-phenyl-2,6-anhydro-D-glycero-D-gulo-heptonamide 20 (shown in entry 3) results in a ratio of 38-67 (~53 as an average). Multiplication of the inhibition constants of IIb (entry 2) with this average factor to predict the efficiency of 5a (entry 4) gives values of~7900-8500 µM, a range being well beyond the concentrations investigated in this study (max 625 µM). Nevertheless, our observations may refer to a strong directionality in the amide-1,2,3-triazole bioisosterism (mostly ignored in related studies [63,64]) indicating that the proper replacement must correspond to the pairs Ib-IIb and 20-5a in entries 1-2 and 3-4, respectively. Table 5. Inhibitory effect of the new and some earlier compounds against rabbit muscle glycogen phosphorylase b (RMGPb).

Ib
Reagents and conditions: (a) K2CO3, 4 Å molecular sieves, dry acetone, reflux; (b) ~1 M NaOMe in MeOH, r.t. showed very weak inhibition (entry 11). This study has corroborated that the inhibition of glycogen phosphorylase by N-and C-glucopyranosyl azole type compounds is extremely sensitive to the properties of the heterocycle. Table 5. Inhibitory effect of the new and some earlier compounds against rabbit muscle glycogen phosphorylase b (RMGPb).

IIb
Reagents and conditions: (a) K2CO3, 4 Å molecular sieves, dry acetone, reflux; (b) ~1 M NaOMe in MeOH, r.t. showed very weak inhibition (entry 11). This study has corroborated that the inhibition of glycogen phosphorylase by N-and C-glucopyranosyl azole type compounds is extremely sensitive to the properties of the heterocycle. Table 5. Inhibitory effect of the new and some earlier compounds against rabbit muscle glycogen phosphorylase b (RMGPb).

20
Reagents and conditions: (a) K2CO3, 4 Å molecular sieves, dry acetone, reflux; (b) ~1 M NaOMe in MeOH, r.t. showed very weak inhibition (entry 11). This study has corroborated that the inhibition of glycogen phosphorylase by N-and C-glucopyranosyl azole type compounds is extremely sensitive to the properties of the heterocycle.

5a
Reagents and conditions: (a) K2CO3, 4 Å molecular sieves, dry acetone, reflux; (b) ~1 M NaOMe in MeOH, r.t. showed very weak inhibition (entry 11). This study has corroborated that the inhibition of glycogen phosphorylase by N-and C-glucopyranosyl azole type compounds is extremely sensitive to the properties of the heterocycle. Table 5. Inhibitory effect of the new and some earlier compounds against rabbit muscle glycogen phosphorylase b (RMGPb).

5b
Reagents and conditions: (a) K2CO3, 4 Å molecular sieves, dry acetone, reflux; (b) ~1 M NaOMe in MeOH, r.t. showed very weak inhibition (entry 11). This study has corroborated that the inhibition of glycogen phosphorylase by N-and C-glucopyranosyl azole type compounds is extremely sensitive to the properties of the heterocycle.

5c
Reagents and conditions: (a) K2CO3, 4 Å molecular sieves, dry acetone, reflux; (b) ~1 M NaOMe in MeOH, r.t. showed very weak inhibition (entry 11). This study has corroborated that the inhibition of glycogen phosphorylase by N-and C-glucopyranosyl azole type compounds is extremely sensitive to the properties of the heterocycle.  showed very weak inhibition (entry 11). This study has corroborated that the inhibition of glycogen phosphorylase by N-and C-glucopyranosyl azole type compounds is extremely sensitive to the properties of the heterocycle.

General Procedure 3 for Removal of the O-Acetyl Protecting Groups
An O-acyl protected compound (100 mg) was dissolved in anhydr. MeOH (5 mL), a few drops of 1 M solution of NaOMe/MeOH was added and the mixture was left to stand at r.t. After complete conversion (TLC monitoring, CHCl 3 -MeOH 7:3) the reaction mixture was neutralized with Amberlyst 15 (hydrogen form). After removal of the resin by filtration, the solvent was evaporated in vacuo and the crude product was purified by column chromatography (CHCl 3 -MeOH 9:1).