Enantioselective Synthesis of β-l-5-[(E)-2-Bromovinyl)-1-((2S,4S)-2-(hydroxymethyl)-1,3-(dioxolane-4-yl) Uracil)] (l-BHDU) via Chiral Pure l-Dioxolane

β-l-5-((E)-2-Bromovinyl)-1-((2S,4S)-2-(hydroxymethyl)-1,3-(dioxolane-4-yl) uracil (l-BHDU, 17) is a potent and selective inhibitor of the varicella-zoster virus (VZV). l-BHDU (17) has demonstrated excellent anti-VZV activity and is a preclinical candidate to treat chickenpox, shingles (herpes zoster), and herpes simplex virus 1 (HSV-1) infections. Its monophosphate prodrug (POM-l-BHDU-MP, 24) demonstrated an enhanced pharmacokinetic and antiviral profile. POM-l-BHDU-MP (24), in vivo, effectively reduced the VZV viral load and was effective for the topical treatment of VZV and HSV-1 infections. Therefore, a viable synthetic procedure for developing POM-l-BHDU-MP (24) is needed. In this article, an efficient approach for the synthesis of l-BHDU (17) from a readily available starting material is described in 7 steps. An efficient and practical methodology for both chiral pure l- & d-dioxolane 11 and 13 were developed via diastereomeric chiral amine salt formation. Neutralization of the amine carboxylate salt of l-dioxolane 10 provides enantiomerically pure l-dioxane 11 (ee ≥ 99%). Optically pure 11 was utilized to construct the final nucleoside l-BHDU (17) and its monophosphate ester prodrug (POM-l-BHDU-MP, 24). Notably, the reported process eliminates expensive chiral chromatography for the synthesis of chiral pure l- & d-dioxolane, which offers avenues for the development and structure–activity relationship studies of l- & d-dioxolane-derived nucleosides.


■ INTRODUCTION
Varicella-zoster virus (VZV) is a highly contagious alpha herpesvirus that causes chickenpox (varicella) and shingles (herpes zoster). 1 The Centers for Disease Control and Prevention (CDC) has estimated that in the U.S., there are approximately one million cases of zoster annually. 2People over the age of 50, immunocompromised persons, HIVinfected people, and organ transplant patients are at a higher risk of VZV infections. 3,4The VZV infection induces major complications; it causes a painful skin vesicular rash.A significant problem of shingles is postherpetic neuralgia, which is long lasting pain that persist for months and years. 5herefore, there is a need for new antivirals that are more effective and safer for the treatment and management of VZV infection than the currently available regiments.However, vaccination is available for both stages of VZV infections, but these can only be utilized in healthy persons. 6cyclovir (ACV), valacyclovir (VACV), and famciclovir (FCV, Figure 1) are approved drugs for the treatment of VZV infection. 7However, these drugs are not highly effective against VZV, and large doses are required, which usually promotes drug resistance. 8Cidofovir (CDV), a broadspectrum antiviral, is active against VZV, but it is associated with nephrotoxicity and lack of oral administration. 7Another antiviral drug, brivudine (BVDU), is approved to treat VZV infection in Europe. 9,10A major drawback associated with BVDU is that it metabolizes in the liver into bromovinyl uracil (BVU). 9BVU impedes the activity of dihydropyridine dehydrogenase (DPD), which catabolizes thymidine and uracil.Thus, cancer patients undergoing cotreatment with 5fluorouracil (5-FU) and BVDU will have life threatening side effects. 11,12Due to the significant problems of currently approved drugs, there is an urgent need for new safe and effective antivirals to treat VZV infection and viral resistant strains.
During the past two decades, in search of effective antivirals against VZV, herpes simplex viruses 1 and 2 (HSV 1 & 2), and Epstein−Barr virus (EBV), our group has been involved in the discovery of modified L-and D-dioxolane-nucleos(t)ides.During these efforts, we have discovered 2-(hydroxymethyl)-1,3-dioxolan-4-yl]5-vinyl uracil (L-HDVD, Figure 2) 13 and 2-bromovinyl-2-(hydroxymethyl)-1,3 dioxolan uracil (L-BHDU) 14,15 with potent antiviral activity against EBV, VZV & HSV-1.L-HDVD is highly active against EBV (EC 50 value of 0.01 μM) and Kaposi's sarcoma-associated herpesvirus (KSHV) (EC 50 = 0.09 μM). 13-BHDU has demonstrated an EC 50 value of 0.25 μM without cytotoxicity in human foreskin fibroblasts up to 200 μM, with a selectivity index (SI) of >909. 14In in vivo studies, L-BHDU significantly reduces the viral load in comparison to the ACV and VACV.It is noteworthy that in the metabolic studies, it was found that L- BHDU does not inhibit dihydropyridine dehydrogenase (DPD), and it provided a better safety and antiviral profile than BVDU. 14ecently, to increase the bioavailability and uptake of L-BHDU with enhanced pharmacokinetic properties, our group has developed the L-BHDU monophosphate pivaloyloxymethyl ester (POM-L-BHDU-MP) prodrug (Figure 2). 16,17POM-L-BHDU-MP has expressed antiviral activity similar to that of L- BHDU in infected cells.In contrast, POM-L-BHDU-MP was superior to L-BHDU in a VZV mouse model. 18The pharmacokinetic studies revealed that POM-L-BHDU-MP had a better oral absorption profile compared to L-BHDU. 16epeated assay in humanized mice showed that POM-L-BHDU-MP was effective when administered orally once per day at 11.3 mg/kg or higher, which indicates its potency and bioavailability in vivo. 18POM-L-BHDU-MP was evaluated as a topical treatment for VZV and HSV in a human skin explant model, which demonstrated highly effective antiviral activities against both viruses at 0.2% formulated in cocoa butter. 18ased on these findings, POM-L-BHDU-MP was selected as a preclinical candidate against VZV infection (shingles).
To develop POM-L-BHDU-MP as a preclinical candidate, extended biological, pharmacokinetic, and toxicological studies are required, which demand a significant amount of POM-L-BHDU-MP.Consequently, the development of a robust, practical, and cost-effective synthesis of POM-L-BHDU-MP  The Journal of Organic Chemistry was needed.Earlier synthesis of L-BHDU was carried out by the L-dioxolane key intermediate 5 (Scheme 1).We and several other research groups followed the procedure reported by Sznaidman et al. 19,20 for the synthesis of L-dioxane 5, which is a critical intermediate for the synthesis of L-modified nucleoside analogues.
However, this reported process involved the chiral chromatographic separation of racemic dioxolane 4, which limits its scalable and commercial utility.Furthermore, during the large-scale synthesis, in the presence of excess boron trifluoride diethyl etherate (BF 3 •Et 2 O), the decomposition of intermediate 4 was observed.Chiral pure L-& D-dioxolane (5 and 6) also revealed stability issues during the chiral chromatographic separation, and undesired impurities were observed during the concentration or lyophilization of collected chiral pure fractions.These downsides (Scheme 1) restrict the process for the large-scale synthesis of chiral L-& Ddioxolane (5 and 6).To overcome these challenges of Scheme 1, herein we report a scalable, practical synthesis of L-BHDU (17) in 7 steps from the methyl (R)-2,2-dimethyl-1,3dioxolane-4-carboxylate 7. To accomplish the scalable synthesis of L-BHDU (17), we revisited synthesis of chiral L-& Ddioxolane, which was previously reported by Bera et al. 21Here, we report an improved process for synthesis of both chiral pure L-& D-dioxolane (11 and 13, enantiomeric excess ≥99%), which is devoid of chromatography separations and feasible for large-scale synthesis.All previously reported methods described chromatographic and chiral chromatographic separation for the synthesis of the chiral pure L-& D-dioxolane, 19,20 which limits construction of diverse L-& D-dioxolane-derived nucleosides/nucleotides of pharmaceutical interest.Nevertheless, the present communication discloses a straightforward synthesis and separation of the chiral pure L-& D-dioxolane (11 and 13) via the formation of a diastereomeric chiral amine salt.The reported procedure is devoid of a chiral separation and reduces column chromatographic steps, which makes the process feasible for the scalable synthesis of the chiral L- dioxolane key intermediate (11) and POM-L-BHDU-MP (24).

■ RESULTS AND DISCUSSION
Numerous efforts by various groups to construct modified nucleosides of five-membered dioxolane often lead to a racemic mixture of L-& D-dioxolane and demand chiral SFC purification to obtain the chiral pure L-and D-dioxolane.To avoid the chiral SFC separation, Bera et al. 21reported the synthesis of chiral dioxolane via methyl-(R)-2,2-dimethyl-1,3dioxalane-4-carboxylate.However, the reported procedure was ambiguous with no analytical data of chiral dioxolanes and required a close analytical technique to confirm the L-& Denantiomers, which may result in undesired conformational final products.With the L-BHDU (17) authentic sample in hand from a previously synthesized method, 22 we tried to compare the optical value of the newly synthesized L-BHDU (17) by the method of Bera et al. with that of our authentic sample without success.Thus, to accomplish the scalable synthesis of L-BHDU (17), a new protocol for the synthesis of the chiral pure L-dioxolane was required.The reported method in this communication is entirely focused on the synthesis of the chiral L-& D-dioxolane with an excellent enantiomeric excess (ee ≥ 99%).Notably, this procedure may also be suitable for the large-scale synthesis of the other dioxolane nucleosides such as troxacitabine 23 and L-HDVD. 13It may also expedite the extended structure−activity relationship studies of D-& L-dioxane-derived molecules of therapeutic interest, which have yet to be explored.
The synthesis of chiral pure L-dioxolane key intermediate 11 was initiated by condensation of commercially available methyl (R)-2,2-dimethyl-1,3-dioxolane-4-carboxylate 7 with 2-benzyloxyacetaldehyde via transketalization in the presence of Dowex 50W X8 to give the racemic 1,3-dioxolane ester 8 in a 1:1 diastereomeric ratio (Scheme 2) in 76% yield.In the earlier reported process, p-TSA was utilized to convert 7 to 8, which reveals hurdles in scale-up and requires a chromatography purification of the product. 21Next, hydrolysis of methyl ester 8 was carried out by a 1 M aqueous solution of LiOH in water/THF to afford a diastereomeric mixture of acids 9 in 92% yield.
Next, our goal was to obtain pure L-dioxolane acid 11 with a high enantiomeric purity (ee ≥ 99%) by avoiding the chromatographic separation of racemic acid 9.However, the selective separation of L-dioxolane was also challenging.One approach was to insert a chiral auxiliary via esterification of carboxylic acid 9, followed by fractional crystallization to give enantiomerically pure L-dioxolane 11.However, this strategy unnecessarily increases the two-step process of esterification and hydrolysis, which is not appropriate for the large-scale synthesis of L-dioxolane.To reduce the additional steps, it was thought to perform a resolution of racemic 9 via a diastereomeric salt formation with a chiral amine.It is noteworthy that compound 9 is oily in nature, and other fractional crystallization techniques may not be applicable for the resolution of chiral pure L-dioxolane 11 as well as Ddioxolane 13.Singh et al. reported chiral resolution of racemic amines via chiral pure L-(+)-tartaric acid. 24Taking the lead The Journal of Organic Chemistry from the reported procedure by Singh et al., we thought to adopt a reverse strategy, which was chiral resolution of racemic acid 9 via salt formation with a chiral pure amine.
To achieve a diastereomerically pure salt formation of racemic 9, several (R) or (S) chiral amines were utilized (Table 1) in various solvent systems.In the investigational resolution of racemic acid 9, diastereomeric salt formation of 9 with (S)phenylethylamine produced pure chiral dioxolane salt 10 with a diastereomeric excess (de) of 99.37% (entry 4 in Table 1).Racemic acid 9 was taken in acetonitrile and treated with 0.8 eq. of (S)-phenylethylamine, which gives a white salt of the acid amine after precipitation as L-dioxolane salt 10.Our next goal was to determine the chiral purity of the precipitated salt as well as the stereo conformation of the diastereomeric salt.The precipitated diastereomeric salt 10 was treated with the 1 N aq.HCl to afford free diastereomeric L-dioxolane 11.The chiral purity of the obtained free L-dioxolane 11 was determined by chiral HPLC, which indicates that the optical purity of the precipitated isomers is 85% with a 15% presence of another isomer.The obtained results of diastereomeric salt formation via resolution of racemic 9 were not encouraging because, for the construction of optically pure L-BHDU (17), there was a need for L-dioxolane 11 with an optical purity of more than 99% (ee ≥ 99%).
Hence, the process of the diastereomeric salt formation of racemic compound 9 was revisited.To enrich the diastereomeric excess (de) of the precipitated salt (via acetonitrile), the obtained salt 10 was resuspended in EtOAc/isopropanol (IPA) [3:1] and stirred for 4 h at room temperature, then filtered (filtrate contains a racemic mixture of major D-and minor Ldioxolane) and neutralized by 1 N aqueous HCl.The afforded L-dioxolane 11 was re-examined for chiral purity via chiral HPLC, which demonstrated an ee of more than 99% (ee = 99.18%) in 32% yield.Therefore, it was concluded that treatment of a double solvent system, first in acetonitrile followed by EtOAc/IPA (3:1), enhances the diastereomeric excess (de) of the salt precipitate (10, de ≥ 99.3%).The subsequent objective was the stereo conformation determination of the precipitated salt, either the D-or L-form of dioxolane.The X-ray crystal of the precipitated salt was developed, and it confirmed that the formed diastereomeric salt matches with the stereo conformation of L-dioxolane 10.The X-ray structure reveals that hydrogen atoms present at carbon-2 (C-2) and C-4 are trans (Figure 3b) to each other, which confirms L-dioxolane salt precipitation.

The Journal of Organic Chemistry
However, a single solvent salt formation approach was also attempted with various chiral amines mentioned in Table 1, but in each case, the results of de were unsatisfactory.To improve the yield of intermediate 11, altered ratios of EtOAc & IPA were used, but in all efforts, either less yield or less chiral purity of intermediate 11 was achieved.
Encouraged by the above finding, it was worth examining the D-dioxolane salt 12 precipitation via racemic 9.It was predicted that by application of chiral (R)-phenylethyl amine [inverse conformation of (S)-phenylamine)] it may provide the diastereomeric salt of D-dioxolane.Similarly, racemic acid 9 was taken in acetonitrile and treated with 0.8 eq. of (R)phenylethylamine, which gives a white salt of the acid amine after precipitation as D-dioxolane salt 12. Furthermore, Ddioxolane salt 12 was resuspended in EtOAc/IPA (3:1) and stirred for 4 h at room temperature, and then it was filtered and neutralized by 1 N aqueous HCl solution to afford D-dioxolane 13 in 28% yield with ee ≥ 99.3% (entry 5, Table 1).The direct precipitation of chiral salts of L-& D-dioxolane by the racemic 9 is a crucial finding, and the described method may expedite the synthesis of numerous novel derivatives of chiral pure L- and D-dioxolane-derived small molecules.
However, the exothermic final deprotection of the benzyl group by using BCl 3 was not suitable for large-scale synthesis.To avoid the use of BCl 3 , a protection group replacement strategy was adopted, and 5-O-benzyl of intermediate 11 was The Journal of Organic Chemistry replaced with an isobutyric ester.The debenzylation of 11 was carried out with 5% Pd/C under hydrogenation conditions to afford intermediate 18 in 91% yield.Intermediate 18 was treated with isobutyric anhydride in the presence of pyridine to produce 19 in 88% yield.The acylation of compound 19 followed by the coupling with BVU afforded coupled desired (β) nucleoside 22 (54% yield) and undesired (α) nucleoside 21 (yield 22%) in an approximately 2:1 ratio (Scheme 4).
Deprotection of the isobutyl ester of 22 was executed by a 7 N NH 3 solution in MeOH to give L-BHDU (17) in 88% yield.Optical rotation and other analytical data of 17 were consistent with previously synthesized authentic L-BHDU. 22o synthesize the POM-L-BHDU-MP prodrug, first synthesis of bis(POM)phosphorochloridate ( 23) was carried out according to the reported protocol by Hawang and Cole. 26ext, coupling of L-BHDU (17) was performed with POM chloride (23) in the presence of N-methyl imidazole in THF to furnish POM-L-BHDU-MP (24) in 55% yield (Scheme 5). 16

■ CONCLUSIONS
To determine the full biological profile of POM-L-BHDU-MP (24), an efficient and scalable synthetic method of L-BHDU (17) has been developed via commercially available methyl (R)-2,2-dimethyl-1,3-dioxolane-4-carboxylate, 7. The selective diastereomeric salt formation of racemic 9 via a chiral (S)phenyl ethylamine followed by neutralization of 10 gave optically pure L-dioxolane, 11 (ee ≥ 99%).Compound 11 was converted to acetate intermediate 14, which was utilized for Vorbruggen coupling with BVU followed by the benzyl deprotection of coupled nucleoside to afford the target compound 17 (L-BHDU) in 7 steps with approximately 4.9% overall yield.This process removes the expensive chiral separation of racemic L-& D-dioxolane and is more efficient than the previously reported method for the synthesis of L-B H D U ( 1 7 ) .F u r t h e r c o u p l i n g o f b i s ( P O M )phosphorochloridate 23 with L-BHDU (17) produced L-BHDU-monophosphate ester prodrug 24 (POM-L-BHDU-MP).The removal of chiral separation and reduction of column chromatography in synthetic steps and the use of economical reagents make the reported methodology amenable for large-scale preparation of POM-L-BHDU-MP (24).Additionally, the reported synthesis of chiral pure L-& Ddioxolane (11 and 13) opens new paths for the synthesis of diverse nucleoside/nucleotide analogues of pharmaceutical interest.

■ EXPERIMENTAL SECTION
General Analytical Methods.Reagents and anhydrous solvents were purchased from commercial sources and used without further purification.Moisture-sensitive reactions were performed by using oven-dried glassware under a nitrogen or argon atmosphere.
Reactions were monitored by thin-layer chromatography plates (TLC silica gel GF 250 μm) that were visualized using a Spectroline UV lamp (254 nm) and developed with 15% solution of sulfuric acid in methanol.Column chromatography was performed on silica gel 60 Å, 40−63 μM (230 × 400 mesh, Sorbent Technologies).Preparative normal phase chromatography was performed on a CombiFlash Rf 150 (Teledyne Isco) with prepacked RediSep Rf silica gel cartridges or on RediSep gold C18 reverse phase columns.Melting points were recorded on a Mel-temp II laboratory device and are uncorrected.Nuclear magnetic spectra were recorded on Varian Inova 500 spectrometer at 500 MHz for 1 H NMR, 202 MHz for 31 P NMR, and 125 MHz for 13 C NMR with tetramethylsilane as an internal standard.Chemical shifts (δ) are quoted as s (singlet), bs (broad singlet), d (doublet), t (triplet), q (quartet), m (multiplet), dd (double doublet), and dt (double triplet).Optical rotations were measured on a JASCO DIP-370 digital polarimeter.Structural assignments were determined using additional information from gCOSY, gHSQC, and gHMBC experiments.High-resolution mass spectroscopy (HRMS) spectra were recorded on a Bruker Ultrahigh resolution QTOF MS Impact II spectrometer.Samples were infused at 3 μL/min, and spectra were obtained in the positive or negative ionization mode with a typical resolution of 20,000 or greater.Optical purity of chiral intermediates and final chiral compounds were determined by the chiral HPLC.Chiral HPLC/UV were performed with a Waters HPLC coupled to a photodiode array.Ten microliters of samples (0.5 mg/mL in methanol) were injected using a CHIRALCEL OX-H, 5 μmm (4.6 × 250 mm) column at 30 °C with a flow rate of 3.0 mL/min.

Figure 1 .
Figure 1.Structures of antiviral drugs for the treatment of VZV.