Synthesis of Chiral 1,4-Disubstituted-1,2,3-Triazole Derivatives from Amino Acids

A versatile method for the synthesis of chiral 1,4-disubstituted-1,2,3-triazole derivatives starting from easily accessible naturally occurring D-or L-amino acids as chiral synthons is described. The amino acids were converted into azido alcohols, followed by copper catalyzed [3+2] cycloaddition reactions between the azido alcohols and methyl propiolate and subsequent ester aminolysis with primary and secondary amines furnished the target compounds, which were obtained in excellent yields with no racemization. Docking of selected target compounds shows that the chiral 1,4-disubstituted-1,2,3-triazoles derivatives has the potential of mimicking the binding mode of known purine analogues.

In our search for synthetic non-natural frameworks that could mimic nucleosides, we wanted to investigate the possibility of using 1,2,3-triazole moiety as a mimic of the imidazole part of the adenine system (Figure 1). Imidazole and triazole based nucleoside derivatives have been reported as potent enzyme inhibitors. Ribavarin [15], a 1,2,4-triazole based purine analogue, has high potency as an antiviral agent and is currently used in the clinic to inhibit RNA virus infections. Furthermore, imidazole based derivatives have been reported as potent adenosine deaminase (ADA) inhibitors ( Figure 1) [16][17][18][19]. Therefore, we wanted to investigate the possibility of the 1,2,3-triazole scaffold to be docked into proteins that are inhibited of these nucleoside analogues and to develop efficient synthesis protocols for the preparation of 1,4-disubstituted-1,2,3-triazoles.

Results and Discussion
In order to explore the potential of the 1,4-disubstituted-1,2,3-triazoles as a possible scaffold for nucleoside inhibitors, compound (R)-3 derived from D-phenylalanine, was docked into the nucleoside binding pocket of a viral enzyme of the Dengue virus (NS5MTase DV ) [20] and adenosine deaminase [16][17][18][19]. Initially, the triazole scaffold was compared with the ribavirin nucleotide that inhibits the NS5MTase DV enzyme. Upon entry into the cell, ribavirin gets converted to the 5'-triphosphate derivative before it binds to the enzyme ( Figure 1A). However, due to poor electron density in the X-ray structure (1R6A) only the monophosphate has been modelled. Using Schrödinger's Glide (XP docking mode), compound (R)-3 docks into the same pose (yellow ligand) as the ribavirin 5'-monophosphate in the NS5MTase DV enzyme suggesting that compound (R)-3 mimics the binding mode of ribavirin (dark blue) ( Figure 1B). The second example comes from the inhibition of the ADA enzyme. In this case, several crystal structures have been reported with imidazole-based nano-molar inhibitors [16][17][18][19].
Therefore, (R)-3 was also docked into the ADA system in order to investigate the similarity of the 1,4-disubstituted-1,2,3-triazoles to the imidazole-based nucleoside analogue 2 ( Figure 1C). The docking results from Glide (XP docking mode) show that our model compounds (yellow) superimposes the imidazole of the inhibitor in the binding pocket of the X-ray structure ( Figure 1D).
The docking results from NS5MTase DV and adenosine deaminase (ADA) suggest that chiral 1,4disubstituted-1,2,3-triazoles derivatives has the potential of mimicking the binding mode of known purine analogues.
The 1 and 4 position of the triazole scaffold allow diversification and the 1,4-disubstituted-1,2,3triazoles should be easy accessible from [3+2] cycloaddition reactions between azides and alkynes using the copper catalyzed procedure reported by Meldal [21] and Sharpless [22] (Figure 2). We anticipated that by using azido alcohol derivatives, prepared from amino acids, a variety of different Rgroups could be introduced resulting in chiral 1,4-disubstituted-1,2,3-triazoles. Furthermore, by using propiolate in the [3+2] cycloaddition reactions an ester functionality would be present in 4 position of the disubstituted-1,2,3-triazoles and allow further derivatisation, for example by reactions with primary or secondary amines.
Here, we report an efficient synthesis of compounds represented by structure 4 ( Figure 2) by performing [3+2] cycloaddition reactions followed by ester aminolysis. According to the general strategy outlined in Figure 2, our synthesis of the 1,4-disubstituted-1,2,3-triazoles commenced with the preparation of amino alcohols 10-14 by reduction of the corresponding amino acids 5-9 (Table 1).

Figure 2.
Retrosynthetic strategy for the generation of chiral 1,4-disubstituted-1,2,3triazole derivatives. obtain 1,4-disubstituted-1,2,3-triazoles with the phenyl group further away from the triazole ring. These aromatic groups could potentially make interactions with hydrophobic pockets adjacent to the nucleoside binding site in target proteins [23]. We also wanted to explore the use of amino acids with alkyl side chains and selected L-Val and L-Ser. Although the copper(I)-catalyzed 1,3-dipolar cycloaddition is insensitive to functional group interference, a benzyl ether derivative of L-Ser was used in order to facilitate column chromatography of the target compound. The amino acids were reduced using LiAlH 4 and the resulting amino alcohols converted to the corresponding azides via treatment with TfN 3 . In this way, a variety of azides derived from various amino alcohols were obtained in good yields (70-76% yields) [24][25][26][27][28][29]. In general, the copper(I)-catalyzed 1,3-dipolar cycloaddition is known as a mild reaction proceeding well in aqueous solutions without protection from oxygen. A range of different reaction conditions have been reported, particularly with respect to generation of the active Cu I species. Sources of Cu I include Cu I salts, most commonly copper iodide [21], in situ reduction of Cu II salts, particularly Cu II sulfate [22], and comproportionation of Cu 0 and Cu II [30,31]. Recent reports suggest that nitrogenbased ligands can stabilize the Cu I oxidation state under aerobic, aqueous conditions and promote the desired transformation [32].
We got very good results using 5 mol % of sodium ascorbate and 1 mol % of copper(II) sulphate in a 1:1 mixture of water and tert-butyl alcohol ( Table 2). The reaction between methylene propiolate and 15 or 18 using these reaction conditions furnished the two 1,4-disubstituted triazole products 24 and 25 in 71% and 74% isolated yields respectively, after stirring at RT in a capped vial. The reactions were monitored by TLC and showed complete conversion after 15 h. The product yields in the reactions therefore reflect the efficiency of product purification rather than the coupling efficiency. In order to improve the efficiency of purification of the products by column chromatography we decided to block the hydroxyl function of the azido alcohols either as benzyl ethers (as in compounds 20-22) or as TBDMS ethers (as in compound 23). In addition, the benzyl ethers could make potential interactions in the phosphate-and sugar binding area of the ATP binding site of target kinases [23].
The corresponding 1,4-disubstituted triazole products 26-28 were obtained high yields (83-97% yields). However, the reaction between methylene propiolate and the TBDMS-protected azido alcohol did not proceed to completion, and the corresponding 1,4-disubstituted triazole 30 was only obtained in 66% yield. Prolonged reaction times (30 h) or running the reaction at elevated temperature (50 o C for 15 h) did not improve the yield (data not shown). Steric interference of the bulky TBDMS ether in the cycloaddition reaction is a possible explanation for this effect.
As well as high yields and prevention of by-product formation, it is essential to avoid racemization when functionalizing enantiomerically pure compounds. Hence, a chiral HPLC system was used to analyze the [3+2] cycloaddition products 27 and 28. The two enantiomers separated well on the chiral HPLC column with retention times of 16.6 min (27) and 13.4 min (28) respectively. No racemization could be detected. The ester functionality of the disubstituted-1,2,3-triazoles was further derivatised by reaction with primary and secondary amines. Compound 24 and 25 were reacted with 7N NH 3 in MeOH for 15 h at RT and the corresponding target compounds were obtained in 97% and 93% isolated yields respectively. However, reacting 24 with other amines (3 eq) such as butylamine, benzylamine, diethylamine, morpholine or ethanolamine in MeOH, resulted in sluggish reactions. Prolonged reaction times (30 h), larger excess of amines (10 eq) or elevated temperature, either by classical thermal heating (50 o C, 10 h) or by microwave assisted heating (100 o C, 10-60 min in a sealed vessel) had only a marginal effect on the yield.
The reactions were repeated in the presence of 2 eq of sodium methoxide and TLC showed complete conversion after 15 h at RT. Nevertheless, several products were formed and the target compounds 31-35 were only isolated in low to moderate yields (4-58%). In general, the primary amines gave the best result while the secondary amines proceeded more slowly (Table 3). On the other hand, when the 4-carboxy methyl ester-triazole 27 was reacted with 7N NH 3 in MeOH or amines in the presence of sodium methoxide the 4-carbamoyltriazoles 37-42 were obtained in high yields. Ammonia and primary amines gave excellent yields (81-92%) while diethyl amine and morpholine did not go to completion and gave the corresponding 4-carbamoyltriazoles in 8 and 68% yields respectively. The 4-carbamoyltriazoles 43-46 were also obtained in excellent yields (88-98%). In all cases TLC showed clean reactions with only formation of target compounds. The results clearly show that it is essential to block the primary alcohol in order to prevent side-reaction and to secure high yields of the 4-carbamoyltriazoles.
The two enantiomers 38 and 43 were analyzed by chiral HPLC to determine if any racemization occurred in the amination step. The two enantiomers separated well on the chiral HPLC column with retention times of 21.2 min (38) and 14.9 min (43) respectively. No racemization could be detected.

-phenylbutyric acid and (S)-2-amino-3-benzyloxypropionic acid (40)
were obtained from Sigma Aldrich and Bachem. Trifluoromethanesulfonic anhydride was obtained from Acros. All other chemicals were reagent grade or better obtained from reputable suppliers. 1 Hand 13 C-NMR spectra were recorded on a JEOL JNM-EX 400 spectrometer at 400 and 100 MHz, respectively. Optical rotation was measured with a Perkin Elmer Polarimeter 341 LC. The reactions were monitored by TLC, on silica plated aluminium sheets (Silica gel 60F254, E. Merck) in case of flash chromatography, Merck silica gel CC (230-400 mesh) was used. Elemental analyses were performed at H Kolbe Mikroanalytisches Laborium, Mühlheim an der Ruhr, Germany.

General procedure for preparation of amino alcohols
A solution of lithium aluminium hydride (1.4 eq) in dry THF (2 mL/mmol amino acid) was stirred and cooled in an ice bath. The amino acid (1 eq) was added and the mixture was stirred at 0 o C for 1 h, warmed to room temperature, and stirred for an additional 1 h. The reaction mixture was then heated at reflux for 16 h. The solution was cooled in an ice bath and diluted slowly with ether (30 mL). Water (7.5 mL) and 15% NaOH (2.5 mL) were added dropwise to the solution in succession. The mixture was filtered through Celite to remove the aluminium salts. The Celite was washed with ether (2 × 15 mL). The filtrate was collected and the solvent was removed under reduced pressure affording the amino alcohol. The amino alcohols were used in the following steps without further purification. (10): The target compound was prepared from 5 (24.2 mmol, 4 g) using the general procedure described above and obtained as a yellow solid (2.9 g, 79% yield). Analytical data was consistent with the literature [33]. (11): The target compound was prepared from 6 (18.2 mmol, 3.00 g) using the general procedure described above and obtained as a yellow oil (2.49 g, 91% yield). The amino alcohol was used for the following steps without further purification [34]. (12): The target compound was prepared from 7 (5.5 mmol, 1.00 g) using the general procedure described above and obtained as a yellow oil (50 mg, 92% yield). Analytical data was consistent with the literature [34]. (13): The target compound was prepared from 8 (8.23 mmol, 965 mg) using the general procedure described above and obtained as a yellow oil (805 mg, 96% yield). Analytical data was consistent with the literature [35]. (14): The target compound was prepared from 9 (10.24 mmol, 2.00 g) using the general procedure described above and obtained as an oil (1.80 g, 97% yield). The amino alcohol was used for the following steps without further purification. 1

General procedure for preparation of azido alcohols by diazo transfer
To a solution of amino alcohol (1 eq) and DMAP (2.5 eq) in CH 2 Cl 2 (1mL/mmol amino alcohol) triflyl azide solution (4.4 eq) was added dropwise with stirring and exclusion of moisture. The reaction mixture was stirred at rt for 15 h. The reaction mixture was concentrated under reduced pressure and purified by flash chromatography. (15): The target compound was prepared from 10 (19.27 mmol, 2.91 g) using the general procedure described above and purified by flash chromatography on silica gel using EtOAc/hexane(1:2 v/v) as eluent. The target compound was obtained as a yellow oil (2.45 g, 66% yield). Analytical data was consistent with the literature [28]. (16): The target compound was prepared from 11 (16.4 mmol, 2.49 g) using the general procedure described above and purified by flash chromatography on silica gel using EtOAc/hexane (1:2 v/v) as eluent. The target compound was obtained as a yellow oil (2.10 g, 72% yield). 1  (S)-2-Azido-3-methylbutan-1-ol (18): The target compound was prepared from 13 (7.8 mmol, 805 mg) using the general procedure described above and purified by flash chromatography on silica gel using EtOAc/hexane (1:3 v/v) as eluent, affording the target compound as a yellow oil (698 mg, 70% yield). Analytical data was consistent with the literature [35].

General procedure for benzylation of the hydroxyl functionality
The azido alcohol (1 eq) was dissolved in dry THF (10 mL), NaH (suspension in mineral oil 60%, 2 eq) was added and the suspension was stirred 5 min at rt. Benzyl bromide (2 eq) and TBAI (0.05 eq) were added and the reaction mixture stirred for 15 h. The reaction was quenched with sat. NH 4 Cl aq (15 mL) at which the reaction mixture became clear, and was then extracted with EtOAc (2 × 30 mL). The organic phases were pooled and washed once with brine, dried with MgSO 4 , filtered and concentrated under reduced pressure. The residue was purified by column chromatography on silica gel using hexane and hexane/EtOAc (95:5 v/v) as eluents.

General procedure for copper catalyzed cycloaddition reactions
The unprotected azido alcohol (1eq) and the alkyne (1.5 eq) were suspended in a 1:1 mixture of water and tert-butyl alcohol (5 mL/mmol azido alcohol). Benzyl protected azido alcohol (1 eq) and the alkyne (1.5 eq) were suspended in H 2 O/THF (1:2, v/v). Sodium ascorbate (0.1 eq as a 1 M solution in water) was added, followed by copper(II) sulfate (0.01 eq as a 0.1 M solution in water). The reaction mixture was stirred at rt for 15 h. The solvents were removed under reduced pressure and the residue adsorbed on silica gel and purified by column chromatography. (24): The target compound was prepared from 15 (14.9 mmol, 2.64 g) and methyl propiolate (22.35 mmol, 1.87 g) using the general procedure described above. The product was purified by column chromatography on silica gel (2:1 EtOAc/hexane) affording the target compound as a pale oil (3.15 g, 81% yield).
The desired product was recrystallized from EtOH affording 3 as a white solid (184 mg, 97% yield).
The target compound was obtained as a white solid (44 mg, 88% yield). 1

Docking
X-ray structures with inhibitors were used as starting point for all dockings. The enzymes were prepared according to the standard procedure in the Schrödinger package. Docking was performed by using Glide (Schrödinger) [37][38][39] with extra precision (XP) settings and standard parameters for ligand docking.

Conclusions
An efficient synthetic route was developed for the synthesis of chiral 1,4-disubstituted-1,2,3-triazole derivatives starting from D-or L-amino acids. The target compounds were obtained in excellent yields with no racemization. This straight forward and flexible strategy allow the synthesis of novel, 4disubstituted-1,2,3-triazole derivatives with desired properties and is currently being explored in our laboratory for the development of kinase inhibitors.