Synthesis and Evaluation of Bicyclo[3.1.0]hexane-Based UDP-Galf Analogues as Inhibitors of the Mycobacterial Galactofuranosyltransferase GlfT2

UDP-galactofuranose (UDP-Galf) is the donor substrate for both bifunctional galactofuranosyltransferases, GlfT1 and GlfT2, which are involved in the biosynthesis of mycobacterial galactan. In this paper, a group of UDP-Galf mimics were synthesized via reductive amination of a bicyclo[3.1.0]hexane-based amine by reacting with aromatic, linear, or uridine-containing aldehydes. These compounds were evaluated against GlfT2 using a coupled spectrophotometric assay, and were shown to be weak inhibitors of the enzyme.


Introduction
Mycobacterium tuberculosis and other mycobacterial species have a unique lipidated polysaccharide structure in their cell wall, the mycolyl-arabinogalactan (mAG) complex, which provides the organism with significant protection from the environment [1][2][3][4]. The polysaccharide portion of the mAG complex contains a galactan domain with approximately 30 galactofuranose (Galf ) residues attached via alternating β-(1→5) and β-(1→6) linkages. All of the galactose residues in mycobacterial galactan are in the furanose form, an isomer of this monosaccharide that is absent in humans [5]. Thus, the glycosyltransferases that are involved in the biosynthesis of mycobacterial galactan are viewed as potential targets for development of new antibacterial agents [6][7][8].

Design Considerations
As targets, we chose molecules containing different groups that could fill the binding pocket of GlfT2 that would normally be occupied by the uridine diphosphate moiety of 1. In total, eight compounds (3-10, Figure 2) were targeted for synthesis. The key step was to use the amino group of 2 in a reductive amination strategy to form the corresponding N-alkylated derivatives. The amino group in the compounds was expected to play an important role in the inhibition of the enzyme. Under physiological conditions, this group would be protonated and would thus provide a positive charge close to the sugar ring, which could mimic the anticipated electrophilic transition state of the GlfT2-catalyzed glycosylation. In the search for inhibitors of mycobacterial galactosyltransferases and other galactofuranose-recognizing proteins, UDP-Galf analogues have drawn significant attention. Amongst those synthesized, a common structural modification is the decoration of hydroxyl groups on the galactose [11,12] and the replacement of the ring oxygen by other atoms [13][14][15]. In a previous paper, we reported the synthesis of a potential mimetic of 1, the bicyclo[3.1.0]hexane-based derivative 2 ( Figure 1) [16]. Based on previous computational investigations, we anticipated that the five-membered ring in 1 would adopt an envelope conformation in which C-2 was above the plane formed by C-1, O-4, C-4, and C-3 [17,18]. In 2, the five-membered ring is locked into an envelope in which the cyclopropane methylene group is on the same side of the ring as the flap formed by the cyclopentane carbon [19]. Thus, we hypothesized that compound 2 functionalized on the nitrogen with different groups could mimic 1 and serve as GlfT2 inhibitors. In this paper, we describe an exploration of this hypothesis. adopt an envelope conformation in which C-2 was above the plane formed by C-1, O-4, C-4, and C-3 [17,18]. In 2, the five-membered ring is locked into an envelope in which the cyclopropane methylene group is on the same side of the ring as the flap formed by the cyclopentane carbon [19]. Thus, we hypothesized that compound 2 functionalized on the nitrogen with different groups could mimic 1 and serve as GlfT2 inhibitors. In this paper, we describe an exploration of this hypothesis.

Design Considerations
As targets, we chose molecules containing different groups that could fill the binding pocket of GlfT2 that would normally be occupied by the uridine diphosphate moiety of 1. In total, eight compounds (3-10, Figure 2) were targeted for synthesis. The key step was to use the amino group of 2 in a reductive amination strategy to form the corresponding N-alkylated derivatives. The amino group in the compounds was expected to play an important role in the inhibition of the enzyme. Under physiological conditions, this group would be protonated and would thus provide a positive charge close to the sugar ring, which could mimic the anticipated electrophilic transition state of the GlfT2-catalyzed glycosylation.

Design Considerations
As targets, we chose molecules containing different groups that could fill the binding pocket of GlfT2 that would normally be occupied by the uridine diphosphate moiety of 1. In total, eight compounds (3-10, Figure 2) were targeted for synthesis. The key step was to use the amino group of 2 in a reductive amination strategy to form the corresponding N-alkylated derivatives. The amino group in the compounds was expected to play an important role in the inhibition of the enzyme. Under physiological conditions, this group would be protonated and would thus provide a positive charge close to the sugar ring, which could mimic the anticipated electrophilic transition state of the GlfT2-catalyzed glycosylation. compounds (3-10, Figure 2) were targeted for synthesis. The key step was to use the amino group of 2 in a reductive amination strategy to form the corresponding N-alkylated derivatives. The amino group in the compounds was expected to play an important role in the inhibition of the enzyme. Under physiological conditions, this group would be protonated and would thus provide a positive charge close to the sugar ring, which could mimic the anticipated electrophilic transition state of the GlfT2-catalyzed glycosylation.

Synthesis of Target Molecules
Three analogues (3)(4)(5), containing an aromatic domain, could interact with amino acids in the active site either through cation-π or π-π stacking interactions [20]. To access these molecules (Scheme 2) commercially-available aldehydes 11, 12, or 13 were treated with 2 in freshly distilled methanol to form the imines, which were then reduced with either NaBH 4 or borane-pyridine (BH 3 ·Py) complex leading to 3, 4, and 5, respectively. The yields of these reactions were moderate, ranging from 53% to 77%. Normally, NaCNBH 3 is used in reductive amination reactions [21]; however, NaBH 4 was used here given its more potent reducing ability of both the imine and the unreacted aldehyde, which minimized the formation of dialkylated compounds. Reductive amination of 13 using BH 3 ·Py, gave a better yield than when NaBH 4 was used as the reducing agent. However, a similar effect was not seen for 11 or 12; indeed, in the case of 11, partial reduction of the double bond was observed, as was an increase in the amount of dialkylated byproducts.

Synthesis of Target Molecules
Three analogues (3)(4)(5), containing an aromatic domain, could interact with amino acids in the active site either through cation-π or π-π stacking interactions [20]. To access these molecules (Scheme 2) commercially-available aldehydes 11, 12, or 13 were treated with 2 in freshly distilled methanol to form the imines, which were then reduced with either NaBH4 or borane-pyridine (BH3·Py) complex leading to 3, 4, and 5, respectively. The yields of these reactions were moderate, ranging from 53% to 77%. Normally, NaCNBH3 is used in reductive amination reactions [21]; however, NaBH4 was used here given its more potent reducing ability of both the imine and the unreacted aldehyde, which minimized the formation of dialkylated compounds. Reductive amination of 13 using BH3·Py, gave a better yield than when NaBH4 was used as the reducing agent. However, a similar effect was not seen for 11 or 12; indeed, in the case of 11, partial reduction of the double bond was observed, as was an increase in the amount of dialkylated byproducts. In previous molecular modeling studies by van Boom and coworkers [22], a five-atom linker between the uridine and the sugar moiety was shown to provide the required distance to span a pyrophosphate moiety. Hence a group of analogues containing five-or six-member chains attached to the nitrogen were selected for synthesis (6)(7)(8)(9)(10). We chose as targets compound 6, which has five atoms between the nitrogen and oxygen, and 7, which has a six-atom linker, but with more hydroxyl groups that might act as the chelating sites to metal ions involved in the transferase reaction [23]. Compounds 8-10 contain the uridine moiety, and have five or six atoms between the bicyclohexane moiety and the uridine. In previous molecular modeling studies by van Boom and coworkers [22], a five-atom linker between the uridine and the sugar moiety was shown to provide the required distance to span a pyrophosphate moiety. Hence a group of analogues containing five-or six-member chains attached to the nitrogen were selected for synthesis (6)(7)(8)(9)(10). We chose as targets compound 6, which has five atoms between the nitrogen and oxygen, and 7, which has a six-atom linker, but with more hydroxyl groups that might act as the chelating sites to metal ions involved in the transferase reaction [23]. Compounds 8-10 contain the uridine moiety, and have five or six atoms between the bicyclohexane moiety and the uridine.
The synthesis of 6 is shown in Scheme 3. Aldehyde 14 [24] and 2 were mixed in freshly distilled methanol and deoxygenated phosphate buffer (pH 6.8) and then reacted with BH 3 ·Py to afford 15 in 69% yield. The phosphate buffer was added to increase the rate of imine reduction [25]. Solvent deoxygenation was important to prevent N-methylation through aerobic oxidation of methanol to formaldehyde, imine formation, and reduction. Hydrogenolysis of 15 in H 2 O and THF afforded the target 6 in quantitative yield. In previous molecular modeling studies by van Boom and coworkers [22], a five-atom linker between the uridine and the sugar moiety was shown to provide the required distance to span a pyrophosphate moiety. Hence a group of analogues containing five-or six-member chains attached to the nitrogen were selected for synthesis (6)(7)(8)(9)(10). We chose as targets compound 6, which has five atoms between the nitrogen and oxygen, and 7, which has a six-atom linker, but with more hydroxyl groups that might act as the chelating sites to metal ions involved in the transferase reaction [23]. Compounds 8-10 contain the uridine moiety, and have five or six atoms between the bicyclohexane moiety and the uridine.
The synthesis of 6 is shown in Scheme 3. Aldehyde 14 [24] and 2 were mixed in freshly distilled methanol and deoxygenated phosphate buffer (pH 6.8) and then reacted with BH3·Py to afford 15 in 69% yield. The phosphate buffer was added to increase the rate of imine reduction [25]. Solvent deoxygenation was important to prevent N-methylation through aerobic oxidation of methanol to formaldehyde, imine formation, and reduction. Hydrogenolysis of 15 in H2O and THF afforded the target 6 in quantitative yield.  To access 7 (Scheme 4) commercially-available 1,4-dimethyl-L-tartrate (16) was treated with benzyl bromide and freshly prepared silver oxide to give the expected dibenzyl ether, which was reduced to the corresponding diol with LiBH4 in ether; subsequent monobenzylation of the product To access 7 (Scheme 4) commercially-available 1,4-dimethyl-L-tartrate (16) was treated with benzyl bromide and freshly prepared silver oxide to give the expected dibenzyl ether, which was reduced to the corresponding diol with LiBH 4 in ether; subsequent monobenzylation of the product with sodium hydride and benzyl bromide gave 17 in 51% over the three step sequence [26]. The primary alcohol was oxidized by Dess-Martin periodinane reagent to afford, in 76% yield, aldehyde 18. The compound was subsequently treated with 2 and BH 3 ·Py in methanol and phosphate buffer (pH 6.8) to give a 49% yield of 19. Finally, target 7 was obtained in quantitative yield by hydrogenolysis over Pd-C in H 2 O and THF. with sodium hydride and benzyl bromide gave 17 in 51% over the three step sequence [26]. The primary alcohol was oxidized by Dess-Martin periodinane reagent to afford, in 76% yield, aldehyde 18. The compound was subsequently treated with 2 and BH3·Py in methanol and phosphate buffer (pH 6.8) to give a 49% yield of 19. Finally, target 7 was obtained in quantitative yield by hydrogenolysis over Pd-C in H2O and THF. The first step required for the preparation of compounds 8-10 was to generate an activated uridine derivative 20 (Scheme 5) [27], which could then be attached to a linker and finally coupled with the bicyclo[3.1.0]hexane amine 2. Reaction of 20 with 10 equivalents of diol 21 or 22 and 1.2 equivalents of NaH in dimethylformamide (DMF) afforded compounds 23 and 24 in moderate yield (36% and 52%, respectively). Attempts to improve the yield of this transformation by changing the ratio of the starting materials and the sequence of adding the reagents were unsuccessful. These primary alcohols were then oxidized, in ~80% yield, into aldehydes 25 and 26, which were treated with 2 and BH3·Py in methanol and phosphate buffer (pH 6.8) to give the ammonium salts 27 and 28, respectively. The acetate anion was introduced during work up, which involves acidification with acetic acid. The yield of the reductive amination reaction was low and both the amine and the aldehyde were found unconsumed at the end of the reaction. However, elongation of the reaction time led to the formation of an N-methylated byproduct, which was inseparable from the target compound. The first step required for the preparation of compounds 8-10 was to generate an activated uridine derivative 20 (Scheme 5) [27], which could then be attached to a linker and finally coupled with the bicyclo[3.1.0]hexane amine 2. Reaction of 20 with 10 equivalents of diol 21 or 22 and 1.2 equivalents of NaH in dimethylformamide (DMF) afforded compounds 23 and 24 in moderate yield (36% and 52%, respectively). Attempts to improve the yield of this transformation by changing the ratio of the starting materials and the sequence of adding the reagents were unsuccessful. These primary alcohols were then oxidized, in~80% yield, into aldehydes 25 and 26, which were treated with 2 and BH 3 ·Py in methanol and phosphate buffer (pH 6.8) to give the ammonium salts 27 and 28, respectively. The acetate anion was introduced during work up, which involves acidification with acetic acid. The yield of the reductive amination reaction was low and both the amine and the aldehyde were found unconsumed at the end of the reaction. However, elongation of the reaction time led to the formation of an N-methylated byproduct, which was inseparable from the target compound. primary alcohols were then oxidized, in ~80% yield, into aldehydes 25 and 26, which were treated with 2 and BH3·Py in methanol and phosphate buffer (pH 6.8) to give the ammonium salts 27 and 28, respectively. The acetate anion was introduced during work up, which involves acidification with acetic acid. The yield of the reductive amination reaction was low and both the amine and the aldehyde were found unconsumed at the end of the reaction. However, elongation of the reaction time led to the formation of an N-methylated byproduct, which was inseparable from the target compound. With 27 and 28 in hand, the remaining step was to remove both the isopropylidene ketal and the benzyloxymethyl (BOM) aminal. We initially explored hydrogenation to remove the BOM aminal, but treatment of H2 in the presence of Pd/C or Pd(OH)2 only resulted in starting material being returned. We then explored the use of Lewis acids to cleave both protecting groups. However, when treated with boron trichloride or boron tribromide in a number of solvent systems, none of the desired product With 27 and 28 in hand, the remaining step was to remove both the isopropylidene ketal and the benzyloxymethyl (BOM) aminal. We initially explored hydrogenation to remove the BOM aminal, but treatment of H 2 in the presence of Pd/C or Pd(OH) 2 only resulted in starting material being returned. We then explored the use of Lewis acids to cleave both protecting groups. However, when treated with boron trichloride or boron tribromide in a number of solvent systems, none of the desired product was formed. The starting compound was decomposed in the reaction and thin-layer chromatography (TLC) revealed a number of products. However, when 27 and 28 were treated with trifluoroacetic acid (TFA), only the isopropylidene acetal was cleaved and compounds 29 and 30 were obtained in good yield. Given the problems we had in removing the BOM acetal, we chose instead to test compounds 29 and 30 against GlfT2.
We then turned our attention to the synthesis of the final target, 10. Mindful of the low yields obtained in the alkylation of 20 with diols (synthesis of 23 and 24), and our inability to remove the BOM aminal from compounds 27 and 28, we chose to modify the target, by replacing O-5 in the nucleoside domain with a sulfur (derivative 31, Scheme 6). This approach necessitated the preparation of thioacetate 33 from 32 (Scheme 6). Upon treatment with a weak base, 33 would generate a thiolate that could react with an electrophile. We anticipated that the enhanced nucleophilicity of this thiolate compared to an alkoxide would facilitate the alkylation reaction. Moreover, this modification would require less basic conditions, which would obviate the need for the use of an N-protecting group (i.e., BOM).
In executing this approach, compound 32 underwent monosilylation, tosylation, and displacement with potassium thioacetate to form 33. This compound was then deacylated and then treated with 34 to produce 35 in 48% over five steps. After desilylation of 35 with tetra-n-butylammonium fluoride (TBAF), alcohol 36 was obtained in 98% yield [28]. Compound 36 was then oxidized with Dess-Martin periodinane in CH 2 Cl 2 to give the aldehyde, which was coupled, without purification, with 2 and BH 3 ·Py to afford a 34% yield 38. The target 31 was obtained successfully after the cleavage of isopropylidene acetal upon treatment with trifluoroacetic acid in 33% yield. TLC showed complete conversion of the substrates. The poor yield is because of losses during chromatography.
nucleoside domain with a sulfur (derivative 31, Scheme 6). This approach necessitated the preparation of thioacetate 33 from 32 (Scheme 6). Upon treatment with a weak base, 33 would generate a thiolate that could react with an electrophile. We anticipated that the enhanced nucleophilicity of this thiolate compared to an alkoxide would facilitate the alkylation reaction. Moreover, this modification would require less basic conditions, which would obviate the need for the use of an N-protecting group (i.e., BOM). In executing this approach, compound 32 underwent monosilylation, tosylation, and displacement with potassium thioacetate to form 33. This compound was then deacylated and then treated with 34 to produce 35 in 48% over five steps. After desilylation of 35 with tetra-n-butylammonium fluoride (TBAF), alcohol 36 was obtained in 98% yield [28]. Compound 36 was then oxidized with Dess-Martin periodinane in CH2Cl2 to give the aldehyde, which was coupled, without purification, with 2 and BH3·Py to afford a 34% yield 38. The target 31 was obtained successfully after the cleavage of isopropylidene acetal upon treatment with trifluoroacetic acid in 33% yield. TLC showed complete conversion of the substrates. The poor yield is because of losses during chromatography.

Evaluation of Analogues as Inhibitors of GlfT2
These UDP-Galf analogues were investigated as potential inhibitors of mycobacterial GlfT2 using a reported coupled spectrophotometric assay [29]. In these assays, a potential inhibitor is added to the incubation mixture together with a known trisaccharide substrate (β-D-Galf -(1→5)-β-D-Galf -(1→6)-β-D-Galf -Octyl) [30] and UDP-Galf (1). All of the UDP-Galf analogues were screened at a concentration of 4 mM against GlfT2 at 37 • C, using 750 µM UDP-Galf. The percentage activities compared to the no inhibitor control are shown in Figure 3. Of all of the compounds, the one incorporating the furan moiety (compound 4) was the most potent, showing only 18% activity (82% inhibition). The next most potent compound was the BOM-protected derivative 29 which inhibited the enzyme by approximately 50%. All of the other compounds showed less than 40% inhibition; given the concentration at which they were tested compared to the K M of the 1 (~250 µM) [31], this is reflective of these compounds being poor inhibitors. Compounds 4 and 29 do not share obvious structural similarities, except that both possess a heterocyclic ring. It is therefore difficult to draw conclusions about the enhanced potency of these compounds compared to the others. no inhibitor control are shown in Figure 3. Of all of the compounds, the one incorporating the furan moiety (compound 4) was the most potent, showing only 18% activity (82% inhibition). The next most potent compound was the BOM-protected derivative 29 which inhibited the enzyme by approximately 50%. All of the other compounds showed less than 40% inhibition; given the concentration at which they were tested compared to the KM of the 1 (~250 µM) [31], this is reflective of these compounds being poor inhibitors. Compounds 4 and 29 do not share obvious structural similarities, except that both possess a heterocyclic ring. It is therefore difficult to draw conclusions about the enhanced potency of these compounds compared to the others.

Synthesis of Target Molecules
All reagents were purchased from commercial sources and were used without further purification. Solvents used in reactions were pre-dried by PURESOLV-400 System from Innovative Technology Inc. (Amesbury, MA, USA). All reactions were monitored by TLC on silica gel G-25 UV254 (0.25 mm,

Synthesis of Target Molecules
All reagents were purchased from commercial sources and were used without further purification. Solvents used in reactions were pre-dried by PURESOLV-400 System from Innovative Technology Inc. (Amesbury, MA, USA). All reactions were monitored by TLC on silica gel G-25 UV254 (0.25 mm, Macherey-Nagel). Spots were detected under UV light and/or by charring with acidified ethanolic anisaldehyde. Solvents were evaporated under reduced pressure and below 50 • C (water bath). Column chromatography was performed on silica gel 60 (40-60 µm). The ratio between silica gel and crude product ranged from 100:1 to 20:1 (w/w). Iatrobeads refers to a beaded silica gel 6RS-8060, which was manufactured by Iatron laboratories (Tokyo, Japan). 1 H-NMR spectra were recorded on VARIAN INOVA-NMR spectrometers (Varian, Inc., Salt Lake City, UT, USA) at 400, 500, or 600 MHz and chemical shifts are referenced to CDCl 3 (7.26, CDCl 3 ) or CD 3 OD (4.78, CD 3 OD). 13 C-NMR APT spectra were recorded at 100 or 125 MHz, and chemical shifts are referenced to CDCl 3 (77.23, CDCl 3 ) or CD 3 OD (48.9, CD 3 OD). 1 H-NMR data are reported as though they are first order, and the peak assignments are made by 2D-NMR spectroscopy ( 1 H-1 H COSY and HMQC). The numbering system used for NMR assignment is shown in Figure 4. HRMS-ESI spectra were recorded on samples suspended in THF or CH 3

Synthesis of Target Molecules
All reagents were purchased from commercial sources and were used without further purification. Solvents used in reactions were pre-dried by PURESOLV-400 System from Innovative Technology Inc. (Amesbury, MA, USA). All reactions were monitored by TLC on silica gel G-25 UV254 (0.25 mm, Macherey-Nagel). Spots were detected under UV light and/or by charring with acidified ethanolic anisaldehyde. Solvents were evaporated under reduced pressure and below 50 °C (water bath). Column chromatography was performed on silica gel 60 (40-60 µm). The ratio between silica gel and crude product ranged from 100:1 to 20:1 (w/w). Iatrobeads refers to a beaded silica gel 6RS-8060, which was manufactured by Iatron laboratories (Tokyo, Japan). 1 H-NMR spectra were recorded on VARIAN INOVA-NMR spectrometers (Varian, Inc., Salt Lake City, UT, USA) at 400, 500, or 600 MHz and chemical shifts are referenced to CDCl3 (7.26, CDCl3) or CD3OD (4.78, CD3OD). 13 C-NMR APT spectra were recorded at 100 or 125 MHz, and chemical shifts are referenced to CDCl3 (77.23, CDCl3) or CD3OD (48.9, CD3OD). 1 H-NMR data are reported as though they are first order, and the peak assignments are made by 2D-NMR spectroscopy ( 1 H-1 H COSY and HMQC). The numbering system used for NMR assignment is shown in Figure 4. HRMS-ESI spectra were recorded on samples suspended in THF or CH3OH and added NaCl. Optical rotations were measured on Perkin-Elmer 241 Polarimeter    (11,7.5 mg, 0.046 mmol) in freshly distilled CH 3 OH (2 mL) was stirred at rt for 1 h, before being cooled to −30 • C. NaBH 4 (5.3 mg, 0.14 mmol) was added, and the solution was stirred for 5 min before being warmed to rt followed by stirring for an additional 10 min. The solution was then acidified with HOAc to pH 5 and concentrated. The resulting residue was purified by chromatography on Iatrobeads (CH 2 Cl 2 -CH 3 OH 10:1→1:10) to give 3 (14 mg, 77%) as a white foam.  (4). A solution of 2 (6.7 mg, 0.035 mmol) and furan-2-carbaldehyde (12, 3.4 mg, 0.035 mmol) in freshly distilled CH 3 OH (2 mL) was stirred at rt for 1 h, before being cooled to −30 • C. NaBH 4 (5.0 mg, 0.13 mmol) was added, and the solution was stirred for 5 min before being warmed to rt followed by  (15). A solution of 2 (9.0 mg, 0.047 mmol) and 14 (7.8 mg, 0.047 mmol) in fresh CH 3 OH (2 mL) was stirred at rt for 1 h. To this mixture was added phosphate buffer (0.1 M, pH 6.8, 0.2 mL) and the solution was cooled to 0 • C before BH 3 ·pyridine (20 mL, 0.16 mmol) was added. The reaction mixture was then warmed to rt and stirred overnight before being acidified with HOAc to pH 5 and concentrated. The resulting residue was purified by chromatography on Iatrobeads (CH 2 Cl 2 -CH 3 OH from 10:1 to 1:10) to give 15 (13.4 (19). A solution of 2 (11.6 mg, 0.05 mmol) and 18 (21 mg, 0.05 mmol) in freshly distilled CH 3 OH (2 mL) was stirred at rt for 1 h. To this mixture was added phosphate buffer (0.1 M, pH 6.8, 0.2 mL) and the solution was cooled to 0 • C, before and BH 3 ·pyridine (20 mL, 0.16 mmol) was added. The reaction mixture was then warmed to rt and stirred overnight before being acidified with HOAc to pH 5 and then concentrated. The resulting residue was purified by chromatography on Iatrobeads (CH 2 Cl 2 -CH 3 OH 10:1 to 1:1) to give 19 ( (20). To a solution of 2 ,3 -O-isopropylidene-3-(benzyloxylmethyl)uridine (4.0 g, 10 mmol) in pyridine (25 mL) was added p-toluenesulfonyl chloride (2.26 g, 12 mmol) at rt. The reaction was stirred overnight and then the excess reagents were quenched by the addition of CH 3 OH (3 mL). The solution was then concentrated and the residue was purified by chromatography (EtOAc-Hexane 1:2) to give 20 (3.

Conclusions
In this paper we have described the preparation of a small library of UDP-Galf mimetics based upon the bicyclo[3.1.0]hexane derivative 2. The key step in generating the products was the reductive amination between the amino group in 2 and aldehydes. In some cases, protecting group issues led to some of the desired targets not being accessible. Evaluation of these compounds analogues against the mycobacterial galactosyltransferase GlfT2 revealed that most of the compounds were poor inhibitors of the enzyme and, moreover, inclusion of the uridine moiety did not enhance inhibitory binding.