Discovery of a Series of 2′-α-Fluoro,2′-β-bromo-ribonucleosides and Their Phosphoramidate Prodrugs as Potent Pan-Genotypic Inhibitors of Hepatitis C Virus

Abstract

Hepatitis C virus (HCV) nucleoside inhibitors display pan-genotypic activity, a high barrier to the selection of resistant virus, and are some of the most potent direct-acting agents with durable sustained virologic response in humans. Herein, we report, the discovery of β-d-2′-Br,2′-F-uridine phosphoramidate diastereomers 27 and 28, as nontoxic pangenotypic anti-HCV agents. Extensive profiling of these two phosphorous diastereomers was performed to select one for in-depth preclinical profiling. The 5′-triphosphate formed from these phosphoramidates selectively inhibited HCV NS5B polymerase with no inhibition of human polymerases and cellular mitochondrial RNA polymerase up to 100 μM. Both are nontoxic by a variety of measures and display good stability in human blood and favorable metabolism in human intestinal microsomes and liver microsomes. Ultimately, a preliminary oral pharmacokinetics study in male beagles showed that 28 is superior to 27 and is an attractive candidate for further studies to establish its potential value as a new clinical anti-HCV agent.

Graphical Abstract

INTRODUCTION

An estimated 180 million people worldwide are infected with hepatitis C virus (HCV), and about 71 million people have chronic HCV infection. Among those with chronic infection, about 15–30% will develop cirrhosis within 20 years and have a significant increased risk for developing an end-stage chronic liver disease, such as hepatocellular carcinoma (HCC).¹ About 400 000+ people die each year as a result of HCV infection, mostly from cirrhosis and hepatocellular carcinoma. Just a few years ago, the combination therapy with pegylated interferon-α plus ribavirin was used to treat chronically infected patients.² However, the combination of a long treatment course, variable efficacy across genotypes (GTs), a poor safely profile, and low patient tolerability has resulted in very limited use of this therapy.

Advances in understanding HCV viral replication cycle revealed multiple proteins or steps in the viral life cycle that could be pursued as potential targets for antiviral therapy.³ Among them, NS3/4A protease, NS5A and NS5B RNA-dependent RNA polymerase (RdRp) were extensively studied and as a result, direct-acting antivirals (DAA) targeting these viral proteins now form the mainstay of current HCV treatment.⁴ Early NS3/4A protease inhibitors (telaprevir and boceprevir) and NS5A inhibitor (daclatasvir) exhibited high potency but generally showed a low barrier to resistance. Moreover, their activity profile was limited to specific genotypes and associated with significantly high rate of side effects.⁵ On the other hand, sofosbuvir (SOF), a nucleoside inhibitor of NS5B RdRp, quickly became the backbone of interferon free HCV treatment⁶ due to its high potency, pan-genotypic activity, high barrier to resistance selection, and low incidence of side effects.⁷ Thus, SOF in combination with velpatasvir or voxilaprevir (QD) is now part of the most prescribed single tablet regimens for the treatment of HCV genotypes 1–6.⁸ Another pan-genotypic combination regimen comprised of glecaprevir and pibrentasvir with a 8 week (BID) treatment course has also been approved for adults who have been previously treated with an NS5A or NS3/4A protease inhibitor and also for new subjects without cirrhosis and preferably low viral load. Although DAAs have improved the HCV treatment prospects significantly, the treatment course remains long at anywhere from 8 to 12 weeks. Furthermore, treatment has been dampened due to the high cost of currently available DAA. One company, Gilead, markets the only nucleoside (SOF) and does not yet permit other competing companies to use SOF as part of their fixed dose combination with their DAAs. Therefore, there remains a need for additional safe, pan-genotypic nucleoside analogs with a high barrier to resistance that could be formulated with other non-Gilead DAAs. These should be cost effective and when combined with potent pan-genotypic DAAs could lead to a shorter duration of treatment while maintaining very high cure rates.

NS5B RdRp is responsible for replication of HCV RNA and is conserved across all of the genotypes and, as such, is considered a target of choice for HCV treatment.³ Nucleoside inhibitors, such as SOF, are anabolized intracellularly to their 5′-triphosphate form and compete as alternative substrates of the NS5B polymerase. Once incorporated into the growing viral RNA chain, elongation is blocked, and RNA replication is stopped.⁹ During the last decade, various structural modifications have been made to natural nucleosides with the aim of developing potent and selective anti-HCV nucleoside analogs.^6,10,11 Among those structural modifications, 2′-variations showed higher selectivity rates toward the HCV polymerase versus human polymerases, thus crossing the first big hurdle for nucleoside analogs. But so far, SOF (1) is the only nucleoside inhibitor approved by the Food and Drug Administration despite a large number of nucleoside analogs that reached human clinical trials (Figure 1). Many of these analogs along with other known preclinical HCV nucleoside inhibitors possess a 2′-α-OH or isosteric group such as a fluorine atom, and many also contain a 2′-β-C-methyl.^11-13 Molecular modeling studies suggest that the 2′-C-methyl group of the incorporated nucleoside analog might exert steric clash to the next incoming nucleotide substrate.¹⁴ We theorized that replacement of the methyl group with other groups of similar size, such as halogens, may exert similar steric effects in the 2′-β-position resulting in new HCV selective inhibitors.

Figure 1.

Selected potent HCV antiviral agents and targeted β-d-2′-deoxy-2′-β-bromo-2′-α-fluoro nucleoside prodrugs.

Recently, we reported β-d-2′-deoxy-2′-β-chloro-2′-α-fluoro uridine prodrug, 2, as a nontoxic, potent, pan-genotypic anti-HCV NS5B inhibitor (Figure 1).¹⁵ Pinho et al. reported that the monophosphate prodrug of β-d-2′-deoxy-2′-dichloro-uridine showed activity against HCV NS5B (G1b), whereas in our hands β-d-2′-deoxy-2′-dibromo uridine nucleotide prodrug showed micromolar potency.¹⁶ Herein, we describe the synthesis of β-d-2′-deoxy-2′-β-bromo-2′-α-fluoro nucleosides and their phosphoramidate prodrugs, 3, as potent, pangenotypic anti-HCV agents and select the uridine analog as a potential clinical development candidate (Figure 1).¹⁷

RESULTS AND DISCUSSION

Chemistry.

For the preparation of the 2′-bromo-2′-fluoro nucleoside analogs, we chose to capitalize on our group’s recently developed protocol¹⁵ for the unsymmetrical synthesis of 2-dihalogenated compounds. Deoxyribonolactone 4 served as the starting material for synthesis of purine and pyrimidine nucleosides and was readily obtained by oxidation of commercially available 2-deoxy-d-ribose (Scheme 1).¹⁸ Protection of 4 with tert-butyldiphenylsilyl (TBDPS) groups provided lactone 5 in 85% yield. Fluorination of lactone 5 with N-fluorodibenzenesulfonimide (NFSI) in presence of lithium bis(trimethylsilyl)amide (LiHMDS) furnished 2-deoxy-2-fluoroarabinolactone 6 in 29% yield.¹⁸ Lower reaction yields were obtained in the fluorination reaction due to a competing β-elimination of tert-butyldiphenylsilyl alkoxide from the enolate along with the formation of the 2-gem-difluoro (~5%) compound and a 25% recovery of starting material, 5. Bromination of compound 6 using N-bromosuccinimide (NBS) in the presence of LiHMDS afforded the diastereomeric mixture of 2-bromo-2-fluoro lactones 7 and 8 in approximately 1:1 ratio which were separated by flash chromatography on silica gel. The stereochemistry of these 2-dihalogenated compounds was determined at a later stage in the synthesis. Subsequent reduction of 7 with lithium-tert-butoxyaluminium hydride (LiAl(t-OBu)₃H) gave lactol 9 as a mixture of anomers (α/β ratio 1.2/1) in a 90% yield. The lactol 9 was converted to 1-mesylate 10 as an anomeric mixture by treatment with methanesulfonyl chloride and triethylamine in a quantitative yield.

Scheme 1. Synthesis of 2-Br,2-F Lactone and Mesylate 10^a.

^aReagents and conditions: (a) TBDPSCl, imidazole, dimethylformamide, room temperature (rt), 24 h, 85%; (b) NFSI, LiHMDS, tetrahydrofuran (THF), −78 °C, 1 h, 29%; (c) NBS, LiHMDS, THF, −78 °C, 40 min, β-anomer 29% and α-anomer 31%; (d) Li(t-BuO)₃AlH, THF, 2 h, 0 °C to rt, 90%; (e) trimethylsulfonyl chloride (MsCl), Et₃N, dichloromethane (DCM), 0 °C to rt, 1 h, 99%.

Coupling of 1-mesylate 10 with silylated uracil in the presence of trimethysilyl trifluoromethanesulfonate (TMSOTf) afforded the N-glycosylated product 11 as an inseparable mixture of α/β (ratio 2/1) (Scheme 2). Removal of the silyl groups and flash silica gel chromatographic separation gave β-anomer 12 in 21% and α-anomer in 46% yield. Phosphorylation of nucleoside 12 using phenyl-l-isopropylalaninyl phosphorochloridate 13¹⁹ gave phosphoramidate prodrug 14 as a mixture of R_p/S_p isomers (7:3 ratio, assignment unknown). Following a similar protocol for the coupling and deprotection, we separated both α/β isomer of N-4-benzoyl-protected cytosine nucleoside 16.

Scheme 2. Syntheses of 2′-α-F,2′-β-Br Pyrimidine Nucleosides and Their Corresponding Prodrugs^a.

^aReagents and conditions: (a) (i) Uracil or N⁴-benzoyl cytosine, N,O-bis(trimethylsilyl)acetamide (BSA), 1,2-dichloroethane (DCE), 60 °C, 30 min; (ii) TMSOTf, DCE, 80 °C, 5 h, 55–57%; (b) 1 M tetra-N-butylammonium fluoride (TBAF) in THF, THF, 0 °C to rt, 1 h, U analog: β-anomer 21%, α-anomer 46%, C analog: β-anomer 24%, α-anomer: 46%; (c) NMI, THF, rt, 4 h, 45%; (d) NH₃, MeOH, overnight, 93%; (e) 13, t-BuMgCl, THF, 0 °C to rt, 3 h, 21%.

The lower yields and unfavorable anomer ratios we observed in these coupling reactions deserve some further comments. We explored the sugar protecting groups, solvents, and Lewis acids quite extensively^20,21 and found the above conditions, while not fully optimized, to provide the best yields and anomer ratios. It is worth noting that our initial attempts to use Mitsunobu conditions using key lactol 9 and N³-benzyluracil yielded a mixture of major O-glycosylated and minor N-glycosylated α and β products.

Deprotection of the N-4-amino group of β-16 with methanolic ammonia furnished fully deprotected nucleoside 17 in 93% yield. Phosphoramidate prodrug 18 was obtained as a mixture of R_p/S_p isomers (ratio 6:4, assignment unknown) by the reaction of nucleoside 17 with phenyl-l-isopropylalaninyl phosphorochloridate, 13 in presence of t-BuMgCl (Scheme 2).

Mitsunobu conditions were found to provide the best yields for the coupling of purine bases to the 2-Br, 2-F sugar. Hence, coupling of lactol 9 with bis-Boc-adenine gave a mixture of α/β isomers 19 (α/β ratio 3/7) in 39% isolated yield (Scheme 3). Removal of silyl groups and chromatographic separation of α/β isomers gave β-anomer 20 in 44% yield. Further treatment of 20 with BCl₃ afforded adenine nucleoside 21 in 69% yield. The reaction of 20 with phenyl-l-isopropylalaninyl phosphorochloridate 13 in presence of N-methylimidazole (NMI) followed by removal of the Boc groups using 50% trifluoroacetic acid (TFA) in DCM afforded prodrug 22 as a mixture of R_p/S_p isomers (ratio 7:3, assignment unknown). Likewise, the coupling of lactol 9 with bis-Boc-2-amino-6-benzyloxypurine afforded a mixture of α/β isomers 23 (α/β ratio 3/7) in 40% yield (Scheme 4). Again, removal of silyl groups and chromatographic separation of α/β isomers gave β isomer 24 in 44% yield. The benzyl and Boc protection were removed in one step using BCl₃ to furnish guanine nucleoside 25 in 65% yield. The reaction of compound 24 with phenyl-l-isopropylalaninyl phosphorochloridate 13 in presence of NMI followed by Boc group removal using BCl₃ afforded prodrug 26 as a mixture of R_p/S_p (ratio ~1:1) isomers in 20% over two steps.

Scheme 3. Syntheses of 2′-α-F,2′-β-Br Adenosine Nucleoside 21 and Its Prodrug 22^a.

^aReagents and conditions: (a) bis-N-Boc adenine, diisopropyl azodicarboxylate (DIAD), triphenylphosphine (PPh₃), THF, rt, 24 h, 39%; (b) 1 M TBAF in THF, THF, 0 °C to rt, 1 h, β-anomer 44%, α-anomer: 25%; (c) 1 M BCl₃, DCM, −78 °C to rt, 1 h, 69%; (d) (i) 13, NMI, THF, 0 °C to rt, 3 h; (ii) 50% TFA-H₂O, 0 °C to rt, overnight, 40% over two steps.

Scheme 4. Syntheses of 2′-α-F,2′-β-Br Guanine Nucleoside 25 and Its Prodrug 26^a.

^aReagents and conditions: (a) 6-O-Bn-bis-N-Boc guanine, DIAD, PPh₃, THF, 0 °C to rt, 24 h, 40%; (b) 1 M TBAF in THF, THF, 0 °C to rt, 1 h, β-anomer 60%, α-anomer: 19%; (c) 1 M BCl₃, DCM, −20 °C to rt, 2 h, 65%; (d) (i) 13, NMI, THF, 0 °C to rt, 3 h; (ii) 1 M BCl₃, DCM, −78 °C to rt, 3 h, 20% over two steps.

Identification of both α and β anomers was determined by two-dimensional nuclear Overhauser effect (NOE) experiments (Figure 2). In all cases, NOEs between H1′ and H4′ in the β-nucleoside and H1′ with H3′/H5′ in α-nucleoside were observed.

Figure 2.

α/β anomer assignment for nucleoside 12.

Single crystal of 12 was grown from ethanol by slow evaporation of the solvent. The crystal structure showed a β orientation of both the uracil group and bromine atom and an α orientation of fluorine atom and thus confirming S-configuration at 2′-position and correct anomer selection from the glycosylation reaction (Figure 3).

Figure 3.

ORTEP drawing of nucleoside 12 from X-ray crystal analysis.

Antiviral Profile.

Parent nucleosides 12, 17, 21, and 25 and their corresponding phosphoramidate prodrugs 14, 18, 22, and 26 were evaluated for inhibition of HCV genotype 1b RNA replication in Huh-7 cells using a subgenomic HCV replicon system.²² Cytotoxicity in Huh-7 cells was determined simultaneously by extraction and amplification of both HCV RNA and cellular ribosomal RNA.²³ In addition, cytotoxicity was determined in primary human peripheral blood mononuclear (PBM) cells, human lymphoblastoid cells (CEM), and African Green monkey Vero cells.^24,25 The results are summarized in Table 1. The uracil, adenine, and guanine nucleoside analogs 12, 21, and 25 were devoid of anti-HCV activity up to 10 μM, whereas the corresponding monophosphate prodrugs of the uridine and guanine analogs 14 and 26 were active in the submicromolar range (EC_50’s of 0.4 and 0.6 μM, respectively). Only the cytosine nucleoside analog 17 displayed an activity (EC₅₀ = 3.9 μM) without the aid of a phosphoramidate prodrug. However, the formation of its monophosphate prodrug, 18 allowed for a 5-fold increase of potency. This pattern of activity among the parent nucleoside analogs and their corresponding prodrugs is quite similar to what was observed with our recently reported 2′-Cl,2-F series¹⁵ and the 2′Me,2′-F series^10,21,26 to which SOF belongs. In addition, none of the compounds tested, except for reference compound 2′-C-Me-cytidine (NM-107), displayed any cytotoxicity versus our panel of cell lines.

Table 1.

HCV Genotype 1b Replicon Activity and Cytotoxicity of Synthesized Nucleosides and Their Phosphoramidate Prodrugs

	anti-HCV EC50 μM (SDa)		cytotoxicity, CC50 μM (SDa)
compound	EC50	EC90	Huh-7	PBM	CEM	Vero
12	>10	>10	>10	>100	>100	>100
14	0.5 (0.05)	0.9 (0.01)	>10	>100	>100	>100
27	0.7 (0.1)	2.4 (0.1)	>10	>100	>100	>100
28	0.16 (0.002)	0.28 (0.003)	>10	>100	>100	>100
17	4.0 (1.6)	9.5 (0.3)	>10	>100	>100	>100
18	0.6 (0.1)	1.8 (0.2)	>10	>100	>100	>100
21	>10	>10	>10	>100	>100	>100
22	7.1 (0.2)	>10	>10	>100	>100	>100
25	>10	>10	>10	>100	>100	>100
26	0.6 (0.1)	1.8 (0.3)	>10	>100	>100	>100
2′-C-Me-C	2.8 (0.2)	9.9 (0.5)	>100	60 (20)	22 (15)	>100
SOF	0.5 (0.1)	0.9 (0.03)	>100	> 100	> 100	>100

^aThe standard deviation (SD) were determined in Excel using the STDEV function (calculated using the “n – 1” method) based on the average of three independent EC₅₀ and EC₉₀ values.

Uridine is the only RNA nucleoside in which the de novo synthesis of its monophosphate form does not proceed through the parent nucleoside. Indeed, the biosynthesis of uridine monophosphate in humans occurs via decarboxylation of 6-carboxy uridine monophosphate (orotidylate) which is catalyzed by orotidylate decarboxylase. This intricate pathway explains why uridine nucleoside analogs are generally not converted to their triphosphate forms and therefore need to be administered as monophosphate prodrugs to display activity. This also gives uridine-derived nucleoside monophosphate prodrugs an inherent safety advantage in fighting liver diseases since it is well established that phosphoramidate prodrugs are largely absorbed and metabolized to their monophosphate forms in the liver. The polarity of the mono-, di-, and trinucleotide phosphates traps them in liver tissue until they are ultimately released to systemic circulation as the parent nucleoside analog and eliminated from the body.²⁷ These properties, unique to uridine analogs, can effectively reduce systemic exposure to uridine analog phosphates and potential unwanted off-target effects. With these considerations combined with the superior potency and lack of cellular toxicity in our panel of cell lines, the uridine phosphoramidate, 14, was chosen for further study.

As phosphoramidate 14 is a diastereomeric mixture at the phosphorous center, each stereoisomer was studied independently to determine which may be the most promising to select for human clinical development. The S_p and R_p phosphorous diastereomers were initially separated by column chromatography providing compounds 27 and 28 (Scheme 5). When larger amounts of phosphorous single diastereomers were needed, the S_p and R_p diastereomers were synthesized by the known literature protocol utilizing isopropyl ((R or S)-(perfluorophenoxy)-(phenoxy)phosphoryl)-l-alaninate (Scheme 5).²⁸ Use of this synthetic approach also allowed us to assign the absolute stereochemistry to 27 and 28.

Scheme 5. Syntheses of Diastereomers 27 and 28^a.

^aReagents and conditions: (a) t-BuMgCl, THF, −5 to 4 °C, 16 h, 68%.

The single diastereomers 27 and 28 were first evaluated for inhibition of HCV genotype 1b RNA replication in Huh-7 cells using a subgenomic HCV replicon system. As noted with SOF, the two diastereomers have different potencies, but again no cytotoxicity was seen in any of the four cell lines tested (Table 1). It is interesting to note that in the case of SOF, clone A cells were used, and an 18-fold difference was seen for the diastereomers,⁹ whereas in our case the difference was only 2.7-fold and only 1.7-fold at the EC₉₀.

Next 27, 28, and SOF were tested versus chimeric replicons GT3a, GT4a, GT5a, and polymerase mutation S282T associated with drug resistance (built on a GT1b backbone), GT1a and GT1b stably transfected replicons systems in Huh-7 cells and finally in a GT2a infectious assay (Table 2). Although both diastereomers were pan-genotypic versus all genotypes tested, 28 was more potent than 27 against all genotypes tested. In this assay, a bigger difference with 27 and 28 was noted versus GT1b with a 3.8-fold difference at the EC₅₀ and a 7.2-fold at the EC₉₀. Similar to SOF, compound 28 displayed an 8- to 9-fold decrease in potency versus the S282T mutant, whereas isomer 27 was 38 times less potent. On a highly positive note, we found that in these assays, phosphoramidate 28 was equipotent or more potent versus SOF at both the EC₅₀ and EC₉₀ levels.

Table 2.

Potency of 27, 28, and SOF in Huh-7 Cells against Various HCV Genotypes and the Mutant S282T Virus^a

		EC50 and EC90 (μM) and (fold increase vs GT1b-WT)a (chimeric replicons, 1a/1b replicons and 2a infectious assay)
cmpd		GT1a	GT2a	GT1b-WT	GT1b/3a	GT1b/4a	GT1b/5a	GTlb/S282T
27	EC50	0.12 ± 0.07 (1.3)	0.16 ± 0.0 (1.7)	0.092 ± 0.0 (1)	0.23 ± 0.01 (2.5)	0.17 ± 0.04 (1.8)	0.21 ± 0.09 (2.3)	3.6 ± 0.12 (38)
	EC90	0.48 ± 0.004 (0.9)	0.83 ± 0.01 (1.6)	0.53 ± 0.7 (1)	1.6 ± 0.02 (3.1)	0.87 ± 0.23 (1.6)	0.80 ± 1.5 (1.5)	21 ± 0.3 (40)
28	EC50	0.055 ± 0.0 (2.3)	0.046 ± 0.15 (1.9)	0.024 ± 0.0 (1)	0.038 ± 0.0 (1.5)	0.021 ± 0.0 (0.9)	0.030 ± 0.01 (1.3)	0.18 ± 0.01 (8)
	EC50	0.19 ± 0.017 (2.6)	0.077 ± 0.0 (1.0)	0.074 ± 0.002 (1)	0.20 ± 0.001 (2.7)	0.13 ± 0.003 (1.7)	0.083 ± 0.02 (1.1)	1.2 ± 0.1 (16)
SOF	EC50	0.061 ± 0.0 (1.8)	0.043 ± 0.01 (1.3)	0.034 ± 0.0 (1)	0.064 ± 0.0 (1.8)	0.058 ± 0.0 (1.7)	0.044 ± 0.01 (1.3)	0.30 ± 0.02 (9)
	EC90	0.32 ±0.06 (2.1)	0.13 ± 0.02 (0.9)	0.15 ± 0.03 (1)	0.37 ± 0.01 (2.5)	0.15 ± 0.02 (1.0)	0.32 ± 0.15 (2.2)	4.4 ± 0.62 (29)

^aThe SD were determined in Excel using the STDEV function (calculated using the “n – 1” method) based on the average of three independent EC₅₀ and EC₉₀ values.

Next, we looked at the functional strength of inhibitors 27 and 28 by evaluating the corresponding 5′-triphosphate, 12-TP, versus a panel of HCV NS5B polymerases (Table 3). Although the three inhibitors tested were all pan-genotypic versus GT1-6, the IC_50’s were somewhat surprising in that NM-107-TP was by far the most potent when compared to 12-TP and 2′-Me,2′-F UTP; which is not the case in cell culture nor in humans (in the case of NM-107-TP). These observations underscore the inherent obstacles in predicting clinical efficacy based on in vitro data. Also, 12-TP was noticeably less potent versus 2′-Me,2′-F UTP which one would not predict based on the cell culture data presented above. All three nucleoside 5′-triphosphate analogs had a loss of activity versus S282T with NM-107-TP having a 570-fold loss in activity versus this mutation. Interestingly, although 12-TP had a 70-fold loss in activity versus the S282T mutation, we have been unable to select this mutation in cell culture after multiple attempts.

Table 3.

12-TP IC₅₀ (μM) vs Various Genotypes of the NS5B Mutant S282T

IC50, μM (SD)
cmpd	GT1b	GTla	GT2a	GT3a	GT4a	GT5a	GT6a	GT1b S282T
12-TP	0.64 (0.018)	0.38 (0.0065)	1.2 (0.17)	1.8 (0.078)	1.8 (0.048)	1.2 (0.061)	0.59 (0.025)	45 (2.3)
IC50 fold change	1	0.59	1.9	2.8	2.8	1.9	0.92	70
2′-Me,2′-F-UTP	0.18 (0.0012)	0.096 (0.0015)	0.29 (0.035)	0.38 (0.013)	0.57 (0.016)	0.43 (0.012)	0.13 (0.0022)	7.3 (0.41)
IC50 fold change	1	0.53	1.6	2.1	3.2	2.4	0.72	41
2′-Me-UTP (NM-107-TP)	0.063 (0.0015)	0.025 (0.00018)	0.19 (0.020)	0.047 (0.0021)	0.07 (0.0013)	0.097 (0.0046)	0.047 (0.00089)	36 (1.7)
IC50 fold change	1	0.40	3.0	0.75	1.1	1.5	0.75	570

Cellular Pharmacology.

To better understand the difference in cell culture potency of the two diastereomers 27 and 28, and also to explain the discrepancy between the cell culture data for 28 and SOF versus the enzymology data for their TPs, we undertook a study on the cellular uptake and egress profiles of 27, 28, and SOF in Huh-7 cells, and determined the T_max and half-lives of nucleoside 5′-triphosphate metabolites. Uptake incubations were performed at 10 μM for 2, 4, 8, 12, 24, 36, and 48 h (Figure 4). Compound 28 provided higher levels of nucleoside triphosphate (NTP) versus 27 at every time point studied except 48 h and also more rapidly produces NTP (T_max = 12 and 24 h, respectively), thus explaining the difference in HCV replicon potency for the two diastereomers. Also, at every time point (except 48 h), 28 delivers substantially more NTP intracellularly versus SOF, again explaining the discrepancy observed above between the cell culture and enzymology data.

Figure 4.

Cellular uptakes of 27, 28, and SOF in Huh-7 cells and primary human hepatocytes (pmol/10⁶ cells).

Next, to understand if the difference on potency and NTP formation for 27 and 28 was an artifact of the Huh-7 replicon system or if it might behave this way in other cell systems and ultimately in humans, we repeated the above uptake study but this time with primary human hepatocytes (Figure 4). The observed NTP levels for all phosphoramidates derivatives were substantially higher than that observed in Huh-7 cells and, more importantly, 27 and 28 gave virtually identical NTP levels and T_max’s (8 h). On the other hand, SOF was slower to deliver its NTP with a T_max of 12 h, but it ultimately delivered NTP levels that were higher versus 27 and 28.

The egress profiles of 27, 28, and SOF were next studied in Huh-7 cells to determine their T_1/2. Egress incubation was done in triplicate at 10 μM with pretreatment for 24 h based on the above uptake studies, then a change to new media followed by harvesting cells at 0, 2, 4, 8, 12, 24, 36, and 48 h (Figure 5). As one might predict 27 and 28 did not show much difference in egress profile (T_1/2 = 15.4 and 15.7 h, respectively), but both had shorter half-lives versus SOF (T_1/2 = 28.2 h). We also did a similar egress study in primary human hepatocytes in triplicate at 10 μM with pretreatment for 12 h, and 27, 28, and SOF show similar egress profiles with T_1/2 of 15.3, 15.0, and 15.3 h, respectively. Armed with this information, selection of one diastereomer to move forward could not be achieved based solely on cell culture potency and enzymology data as was done with SOF.⁹

Figure 5.

Cellular egress profiles of prodrug 27, 28, and SOF in Huh-7 cells and primary human hepatocytes (pmol/10⁶ cells).

Cytotoxicity Profile.

Nucleoside analogs are known to potentially exhibit mitochondrial (Mt) toxicity.²⁹ Mitochondrial dysfunction commonly involves the peripheral nervous system and brain, the retina, cardiac and other muscles, and the endocrine, renal, gastrointestinal (GI), hematologic, and hepatic systems which can ultimately result in stroke, pancreatitis, lactic acidosis with liver failure, hepatic steatosis, and myopathy.³⁰ Therefore, the potential liabilities for the phosphoramidates 14, 27, 28 and the parent nucleoside 12 were evaluated and compared to SOF for their effects on mitochondrial DNA levels. HepG2 cells were propagated in the presence of nucleotide analogs (up to 50 μM) for 14 days prior to quantification of mitochondrial COXII DNA (mtDNA) and ribosomal DNA using real-time polymerase chain reaction. Lamivudine (3TC) and β-d-2′,3′-dideoxycytidine (ddC) (at 10 μM) were used as negative and positive controls, respectively (Table 4). At the end of the 14 day assay, neither the parent nucleoside, 12 nor the prodrug diastereomer mixture, 14 showed measurable mitochondrial toxicity up to 50 μM in a HepG2 cell line (Table 4), whereas ddC, as anticipated, was highly toxic at 10 μM. Similar suppression of nuclear DNA for 28 and SOF at 50 μM was noted, with 28 having more impact on mitochondrial DNA levels versus SOF. Lactic acid levels were also measured in the culture supernatant after 14 days of incubation with each drug. Increased production of lactic acid (generally above 100% when normalized to ribosomal DNA control) is a marker for the HepG2 cells being under stress and associated with mitochondrial toxicity.³¹ In this study, we did not observe increased lactic acid production with the parent nucleoside 12 nor its phosphoramidate diastereomer mixture 14 up to 50 μM. Conversely, at 50 μM, an increased lactic acid production was observed for 28, SOF, and the positive control, ddC (at 10 μM) (Table 4). However, no mitochondrial toxicity was noted with 28 when evaluated at 10 μM.

Table 4.

Effects of Compounds 12, 27, 28, and SOF on Mitochondrial (Mt), Nuclear DNA Levels, and Lactic Acid Production in HepG2 Cells (14 day Assay)^a

cmpd	conc., μM	% inhibition MtDNA/nDNA	IC50, μM MtDNA/nDNA	MtDNA content % of control (range)	lactic acid production (% of control)
12	50	14 ± 3.7/<1 ± 0.004	>50/>50	72 (63–83)	69 ± 14
14	50	39 ± 5.8/40 ± 3.9	>50/>50	100 (99–104)	140 ± 17
27	50	31 ± 2.8/<1 ± 0.002	>50/>50	65 (65–65)	100 ± 6.2
28	12.5	23 ± 7.9/11 ± 6.5	46 ± 8.0/32 ± 6.0	86 (79–93)	90 ± 14.6
	25	31 ± 5.6/39 ± 2.9		110 (110–110)	120 ± 1.1
	50	58 ± 7.7/72 ± 9.0		150 (150–150)	220 ± 6.5
SOF	12.5	6.7 ± 5.1/3.1 ± 6.4	>50/45 ± 2.5	96 (90–100)	79 ± 17
	50	44 ± 2.8/58 ± 3.0		130 (120–150)	180 ± 12.5
3TC	10	<10/<10	>10/>10	140	83 ± 35
ddC	10	84 ± 9.0/50 ± 4.5	<10/<10	7.3	200 ± 11
untreated control	10	0/0	N/A	100 (72–140)	100 ± 5.0

^aCmpd, compound; conc., concentration; MtDNA, mitochondrial DNA; nDNA, nuclear DNA; N/A, not applicable.

HepG2 cells are highly proliferative immortalized cells that derive a significant proportion of their energy from glycolysis rather than mitochondrial oxidative phosphorylation. The use of glucose containing media with this cell line can mask the effects of potential mitochondrial toxicants. This is referred to as the Crabtree effect and can be evaluated by comparing the cytotoxic effect of drug candidates in glucose versus galactose-supplemented media. In galactose containing media, the HepG2 cells are forced to generate energy by mitochondrial oxidative phosphorylation so toxic effects of a drug on mitochondrial function are more pronounced. Therefore, 14 was also tested in HepG2 cells for 3 days in glucose- and galactose-supplemented media up to 100 μM, and no toxicity signal was observed in either media. Extending this assay to 14 days for 27 and 28 showed no significant difference between using glucose- or galactose-supplemented assay media, and again both 27 and 28 were not toxic up to 100 μM.

Pluripotent hematopoietic stem cells are found in the primary bone marrow of healthy adults. These cells proliferate and differentiate into all mature hematopoietic cells. When cultured in methylcellulose, individual progenitors called colony-forming cells proliferate and differentiate to form colonies of identifiable progeny. Inhibition of the burst forming unit-erythroid (BFU-E) lineage will present in vivo as red blood cell loss and anemia. Inhibition of colony-forming unit-granulocyte, macrophage (CFU-GM) lineage will present in vivo as neutropenia.³²

Although 12, 14, and 27 were devoid of BFU-E and CFU-GM bone marrow toxicity up to 100 μM versus three separate donors, when 28 was tested we found an IC₅₀ = 72 μM for BFU-E and an IC₅₀ = 76 μM for CFU-GM. However, no marked cytotoxicity was noted at more physiologic concentrations of 10 μM. In all of these bone marrow toxicity assays, we found that the positive control AZT exhibited the expected high toxicity in both erythroid and myelomonocytic progenitors at the concentration tested, and 3TC exhibited no toxicity with both progenitors with an IC_50’s of >100 μM.

Compounds 27 and 28 were also evaluated for cytotoxicity in HepaRG cells in a 14 day assay, and the CC₅₀ values were 7.2 and 35 μM, respectively (Table 5). Although there is a mild toxicity signal in this HepaRG cell line, we found SOF to have a CC₅₀ of 44 μM, very similar to what we found with 28. The cytotoxicity potential of 27 and 28 was also evaluated in primary human hepatocytes up to 200 μM and found IC_50’s of 140 and 46 μM, respectively. We tested in vitro for nephrotoxicity in HK-2 cells in a 3 day assay and found that 28, 27, and SOF all had CC_50’s > 300 μM. We evaluated the potential for cardiac toxicity by incubating 28, 27, and SOF with human embryonic stem cell-induced ventricular cardiomyocytes, and all three compounds had CC_50’s > 100 μM (Table 5).

Table 5.

Cytotoxicity (IC₅₀; μM) of Compounds 27 and 28 in Various Cell Lines

cmpd	HepG2 3 day glucose/galactose	HepG2 14 day glucose/galactose	bone marrow BFU-E	bone marrow CFU-GM	HepRG 14 day glucose vs galactose	primary human hepatocytes	HK-2 cells 3 day	cardiomyocytes	hERG
27	>100	>100	>100	>100	7.2	140	>300	>100	>30
28	>100	>100	72a	76b	35	46	>300	>100	>30

^aDonor specific data 15.8, >100, >100 μM.

^bDonor specific data 26.5, >100, >100 μM.

Next, 27 and 28 were evaluated for potential effects on human ether-à-go-go-related gene (hERG) potassium channels using CHO cells stably expressing hERG potassium channels at room temperature utilizing the whole-cell patch clamp technique. Both compounds were determined to have IC₅₀ values greater than 30 μM, whereas the positive control, amitriptyline had an IC₅₀ of 3 μM (Table 5). For comparison, SOF has been reported to be free from hERG channel inhibition,^33,34 although symptomatic bradycardia has been reported when co-administered with amiodarone.³⁵ However, these effects were found to be non-hERG related.³⁶

A mini Ames study was conducted to evaluate 27 and 28 abilities to induce reverse mutations both in the presence and absence of S9 mix at the histidine locus in the genome of four strains of Salmonella typhimurium (TA98, TA100, TA1535, and TA1537) and at the tryptophan locus in the genome of Escherichia coli WP2 uvrA (pKM101). For 27 and 28, cytotoxicity or precipitate was not observed under any of the test conditions. Both test articles did not induce more than 2-fold increase in strain TA98, TA100, and WP2 uvrA (pKM101) nor 3-fold increase in strain TA1535 and TA1537 in the mean number of revertant colonies at any dose level relative to the concurrent negative/solvent control, either in the presence or absence of the S9 mix, and no dose response was observed with any strain.

In contrast, for all of the bacterial tested strains used in this study, the mean number of his⁺ and trp⁺ revertant colonies observed for the negative/solvent control was comparable to the laboratory historical negative control data. All positive controls induced the expected increase more than 3-fold in the mean number of revertant colonies, in the presence and absence of S9 mix, when compared to the concurrent negative/solvent control. This study concluded that, like SOF,³³ both 27 and 28 were negative for mutagenicity under the conditions of this study.

Selectivity for Cellular RNA Polymerase.

We also tested whether nucleoside 5′-triphosphate analog of 14 was a substrate for the human mitochondrial RNA polymerase (POLRMT). Each nucleoside triphosphate was incubated at 100 μM with a POLRMT enzyme, and the appropriate DNA/RNA primer/template hybrid and incorporation was evaluated at 2 h. NTP analog incorporation was normalized to that of natural rNTP substrates. As shown in Table 6, 12-TP was incorporated 7.4% as compared to natural UTP. This value was comparable to the active metabolite of SOF (3.1%) and 2′-C-Me-UTP (9.8%). Similar to 2′-C-methyl-2′-F-UTP, the active metabolite of SOF, 12-TP was an exceedingly poor substrate for POLRMT.

Table 6.

Incorporation of 12-TP into the Mitochondrial RNA Polymerase (POLRMT)

inhibitor	POLRMT % incorporation
12-TP	7.4 ± 1.3
2′-F,2′-C-Me-UTP	3.1 ± 1.4
2′-C-Me-UTP	9.8 ± 4.7
UTP	100

^aNA: not available.

The effect of 12-TP with human DNA polymerases was evaluated (Table 7). Inhibition of cellular DNA polymerases α, β, and γ was determined in vitro using commercially available enzymes and appropriate DNA primers and templates and incubated with increasing concentrations of compound from 0 to 100 μM in a tris-buffered reaction at 37 °C. Aphidicolin was used as a positive control for DNA polymerase α and ddTTP for DNA polymerases β and γ. Both 12-TP and 2′-F,2′-C-Me-UTP were found to have no inhibitory potency against DNA polymerases α, β, and γ (IC₅₀ > 100 μM). The positive controls, aphidicolin and ddTTP, inhibited as expected verifying the validity of the test system.

Table 7.

Impact of 12-TP on Human DNA Polymerase Activity^a

	IC50 (μM)
compound	DNA pol α	DNA pol β	DNA pol γ
12-TP	>100	>100	>100
2′-F,2′-C-Me-UTP	>100	>100	>100
aphidicolin	4.3 ± 0.006	NA	NA
ddTTP	NA	8.4 ± 0.11	0.10 ± 0.07

^aNA: not available.

Stability in Human Liver Microsomes and Gastric Fluid.

Compounds 27 and 28 were both evaluated for their stability to human liver microsomes. Both compounds were rapidly metabolized with 27 having 27% remaining after 60 min, whereas 28 was even more extensively metabolized with only 3% remaining after 60 min (Table 8). Incubations of both 27 and 28 with human liver microsomes lacking the NADPD regeneration system demonstrated that these compounds are metabolized by a non-NADPH-dependent metabolism path. Furthermore, these two diastereomers were tested for stability to human intestinal microsomes and found to be quite stable with 27 having a T_1/2 = 132 min, whereas 28 had a T_1/2 > 145 min (Table 9). Both 27 and 28 were stable in a simulated intestinal fluid at 24 h and had about a 2% loss at 24 h in simulated gastric fluid, both at 37 °C. These data combined with good stability in human plasma (27 72%; 28 90% remaining at 120 min @ 37 °C) indicated that upon absorption in the GI tract, the compounds could pass to the liver via the portal vein largely intact and rapidly convert to the monophosphate form once in liver tissue and thus minimizing systemic exposure to the prodrug form.^27,37

Table 8.

Stability of Compounds 27 and 28 in Human Liver Microsomes

	human liver microsomes (0.5 mg protein/mL)
cmpd	R2 b	T1/2 (min)c	Clint(mic)(μL/min mg))d	Clint(liver) (mL/(min kg))e	remaining, % (T =60 min)	remaining, % (NCF = 60 min)a
27	0.9827	18.4	75.3	67.8	8.9	26.6
28	0.8767	19.7	70.2	63.2	8.0	3.7
testosterone	0.9921	15.5	89.4	80.4	6.5	86.1
diclofenac	0.9935	10.7	129.3	116.3	2.0	88.1
propafenone	0.9638	6.1	226.4	203.8	0.1	95.6

^aNCF: no co-factor. No NADPH regenerating system is added into NCF sample (replaced by buffer) during the 60 min incubation, if the NCF remaining is less than 60%, then non-NADPH-dependent metabolism occurred.

^bR²: correlation coefficient of the linear regression for the determination of kinetic constant.

^cT_1/2: half-life.

^dCl_int(mic): the intrinsic clearance; Cl_int(mic) = 0.693/half-life/mg microsome protein per mL.

^eCl_int(liver) = Cl_int(mic) × mg microsomal protein/g liver weight × g liver weight/kg body weight.

Table 9.

Stability of Compounds 27 and 28 in Biorelevant Fluids at 37 °C

cmpd	human intestinal microsomes, T1/2 (min)	simulated intestinal fluid (24 h)	simulated gastric fluid (24 h)	human plasma (remaining at 120 min)
27	132	stable	~2% loss	72%
28	>145	stable	~2% loss	90%

The presence of an aliphatic bromide in this series of nucleoside analogs warrants some discussions of its stability and potential as an alkylation agent in biological systems. Typically, primary and secondary aliphatic bromides are suitable targets for alkylation by proteins containing nucleophilic amino acids such as arginine, lysine, or cysteine via substitution reactions, whereas tertiary bromides are more prone to elimination reactions and carbocation or free radical formation and subsequent reaction. It has long been understood that electron withdrawing groups attached to tertiary bromides reduce their reactivity.³⁸ In the series presented herein, the powerful electron withdrawing effects of the 2′-fluorine atoms suppress both carbocation and free radical formation and, as such, provide a suitably stable tertiary alkyl bromide for in vivo utilization.

Compound 14 was found to be 100% stable in male beagle plasma up to 2 h but highly unstable to rat plasma (0.2% at 2 h). Thus, we choose a preliminary animal study in male beagles to determine if there was a significant difference in the liver pharmacokinetics of 27 versus 28. We utilized a partially optimized formulation that was administered by oral gavage tube. After a single dose at 10 mg/kg in portal vein cannulated fasted male beagles (2/group; >6 months old), portal vein and peripheral vein plasma were collected at 0.5, 1, 2, and 4 h; in addition, livers were collected at 4 h. We analyzed plasma and liver tissue for prodrugs 27 and 28, parent nucleoside 12 and, in addition, we looked for 12-TP in liver tissue only. Not surprisingly, the level of 27 and 28 observed in liver tissue was quite low to below the level of detection, and the level of 12 in the liver was 10–20 times lower than the level of 12-TP. Portal vein levels observed for prodrug 28 were higher than that observed for 27 at every time point except 4 h. For example, at 30 min, the level of 28 (average 3635 ng/mL) was 2.7 times higher than that observed for 27. There was no marked difference in systemic exposure with 27 and 28 based on low levels of both parent nucleoside and prodrug levels observed in peripheral blood. Despite the small sample size and some variabilities among the phosphoramidates, the group of dogs that received 28 had 12-TP levels (average 20 130 ng/mL) in the liver that were 2 times higher than that seen with 27; clearly indicating the superior absorption and liver uptake of phosphoramidate 28.

CONCLUSIONS

Herein, we disclose the synthesis and biological evaluation of a unique series of 2′-bromo,2′-fluoro nucleosides. Among the synthesized compounds, 27 and 28 were potent and specific inhibitors of HCV in culture, but based on all of the data presented and weighted on preliminary liver pharmacokinetic data in dogs, which showed 28 to be better adsorbed and produce 2 times the level of 12-TP in liver tissues, compound 28 was chosen to further evaluate as an HCV clinical candidate. Compound 28 had excellent pan-genotypic anti-HCV replicon activity similar to that of SOF. Its NTP (12-TP) was a specific inhibitor of HCV NS5B polymerase GT1-6 with no inhibition of human α, β, γ DNA polymerase and showed low incorporation by human mitochondrial RNA polymerase (POLRMT). No marked mitochondrial (MtDNA, nuclear DNA), including lactic acid and bone marrow, toxicities were observed up to 10 μM, and only mild toxicities were observed for both 28 and SOF at 50 μM. At physiological concentrations, no increase in lactic acid was noted for 28 and SOF. No toxicities were observed in a large number of cell lines, a mini Ames was negative versus five strains, and there was no in vitro hERG liability. Compound 28 was highly stable in human blood for up to 2 h, was rapidly metabolized in human hepatocytes, and showed low metabolism in human intestinal microsomes. The novel nucleotide analog 28 has an excellent preclinical profile, suggesting further development to establish its potential value as a clinical anti-HCV nucleoside analog.

EXPERIMENTAL SECTION

General Procedures.

Anhydrous solvents were purchased from (Milwaukee, WI). All commercially available reagents were used without further purifications. Reagents were purchased from commercial sources. All of the reactions were carried out under nitrogen in oven-dried glassware unless otherwise noted. Thin layer chromatography was performed on Analtech GHLF silica gel plates. Column chromatography was accomplished on Combiflash Rf200 or via reverse-phase high-performance liquid chromatography. ¹H, ¹³C, ¹⁹F, and ³¹P NMR spectra were recorded on a Bruker Ascend 400 spectrometer at 25 °C (400, 101, 377, and 162 MHz) as noted, and residual proton solvent signals were used as internal standards. Deuterium exchange and decoupling experiments were utilized to confirm proton assignments. NMR processing was performed with MestReNova version 10.0.2-15465. Signal multiplicities are represented by s (singlet), d (doublet), dd (doublet of doublets), t (triplet), q (quadruplet), br (broad), bs (broad singlet), m (multiplet). Coupling constants (J) are in hertz (Hz). Mass spectra were determined on a Micromass Platform LC spectrometer using electrospray ionization. The purity of final compounds was determined to be >95%, using ultraperformance liquid chromatography (UPLC) analyses performed on a Waters Acquity UPLC System with a Kinetex LC column (2.1 mm A, 50 mm, 1.7 μm, C18, 100 Å) and further supported by clean NMR spectra. Mobile phase flow was 0.4 mL/min with a 1.20 min gradient from 95% aqueous media (0.05% formic acid) to 95% CH₃CN (0.05% formic acid) and a 4.5 min total acquisition time. Photodiode array detection was from 190 to 360 nm.

2-Deoxy-2-bromo-2-fluoro-3,5-di-O-(tert-butyldiphenylsilyl)-d-ribonolactone (7, 8).

To a solution of 6 (5.6 g, 8.94 mmol) and NBS (3.18 g, 17.9 mmol) in THF (45 mL) was added LiHMDS in THF (1 M in THF, 14.31 mL, 14.31 mmol), at −78 °C under N₂, dropwise over a period of 15 min. The suspension was stirred at −78 °C for 40 min and then quenched with a saturated aqueous solution of NH₄Cl (30 mL). The mixture was allowed to warm to rt and extracted with hexanes (3 × 50 mL). The combined organic layers were washed with a saturated aqueous solution of NaHCO₃ (30 mL), water (30 mL), and brine (30 mL). The organic layer was dried over anhydrous Na₂SO₄, filtered, and concentrated. The residue was purified two times by flash column chromatography using 0–1% EtOAC/hexane gradient to give 7 (1.82 g, 29%) and 8 (1.94 g, 31%) as a colorless liquid.

Compound 7: ¹H NMR (400 MHz, CDCl₃ δ 7.72–7.66 (m, 4H), 7.54–7.35 (m, 16H), 4.68–4.65 (m, 2H), 3.68–3.66 (m, 2H), 1.15 (s, 9H), 0.97 (s, 9H). ¹³C NMR (101 MHz, CDCl₃) δ 166.1 (d, J = 26.1 Hz), 136.1, 135.8, 135.6, 135.5, 132.4, 132.3, 132.1, 131.4, 130.5, 130.5, 130.0, 128.1, 128.0, 127.9, 127.8, 93.0 (d, J = 276.1 Hz), 85.4, 75.7 (d, J = 15.0 Hz), 61.7, 26.8, 26.7, 19.4, 19.1. ¹⁹F NMR (377 MHz, CDCl₃) δ −136.0. High-resolution mass spectrometry (HRMS) (electrospray ionization (ESI)): m/z [M + Na]⁺ calcd for C₃₇H₄₂BrFNaO₄Si₂: 727.1687, found: 727.1672.

Compound 8: ¹H NMR (400 MHz, CDCl₃) δ 7.63–7.68 (m, 4H), 7.29–7.49 (m, J = 78.6 Hz, 16H), 4.56 (dd, J = 14.6, 7.9 Hz, 1H), 4.26–4.22 (m, 1H), 3.70–3.72 (dd, J = 12.5 Hz, 3.4 Hz, 1H), 3.46 (dd, J = 12.5, 3.5 Hz, 1H), 1.13 (s, 9H), 0.86 (s, 9H). ¹³C NMR (101 MHz, CDCl₃) δ 165.4 (d, J = 27.7 Hz), 136.1, 135.8, 135.7, 135.4, 132.7, 132.2, 132.0, 130.7, 130.6, 130.4, 129.9, 129.8, 128.1, 128.0, 127.8, 127.7 99.3 (d, J = 279.1 Hz), 81.2, 81.1, 73.2 (d, J = 20.9 Hz), 60.2, 26.7, 26.6, 19.6, 19.1. ¹⁹F NMR (377 MHz, CDCl₃) δ −128.0 (d, J = 14.9 Hz). HRMS (ESI): m/z [M + Na]⁺ calcd for C₃₇H₄₂BrFNaO₄Si₂: 727.1687, found: 727.1667.

(3S,4R,5R)-3-Bromo-4-((tert-butyldiphenylsilyl)oxy)-5-(((tert-butyldiphenylsilyl)oxy)methyl)-3-fluorotetrahydrofuran-2-ol (9).

To a suspension of 7 (1.81 g, 2.57 mmol) in anhydrous THF (15 mL) was added 1 M solution of LiAl(O^tBu)₃H in THF (5.14 mL, 5.14 mmol) at 0 °C. After being stirred at rt for 2 h, the reaction was quenched with saturated NH₄Cl (25 mL) at 0 °C. The mixture was allowed to warm slowly to rt and stirred for 2 h. The reaction mixture was filtered through a pad of celite and washed with ethyl acetate (EA, 30 mL). The aqueous layer was extracted with ethyl acetate (30 mL), and the combined organic layer was washed with saturated NaHCO₃ (25 mL), water (20 mL), and brine (20 mL). The solution was dried over Na₂SO₄, filtered, and concentrated in vacuo to give crude product 9 (1.63 g, 90%) as a mixture of anomers (α/β ratio 1.2:1). The crude product was used as such in the next step.

¹H NMR (400 MHz, CDCl₃) δ 7.74 (m, 4H), 7.67–7.61 (m, 4H), 7.55–7.29 (m, 37H), 5.39 (dd, J = 11.5, 5.3 Hz, 1H), 5.23 (dd, J = 9.1, 6.0 Hz, 1.2H), 4.76 (dd, J = 16.4, 6.0 Hz, 1.2H), 4.60 (dd, J = 10.8, 5.3 Hz, 1H), 4.33 (m, 1H), 4.18 (s, 0H), 3.69 (d, J = 9.1 Hz, 1H), 3.60 (td, J = 12.0, 11.5, 3.0 Hz, 2H), 3.51–3.41 (m, 2H), 3.25 (dd, J =11.5, 2.5 Hz, 1H), 1.10 (d, J = 10.1 Hz, 20H), 0.92 (d, J = 2.3 Hz, 20H). ¹³C NMR (101 MHz, CDCl₃) δ 136.2, 135.9, 135.9, 135.7, 135.6, 135.6, 135.5, 133.0, 132.9, 132.6, 132.5, 132.1, 132.0, 131.9, 131.8, 130.3, 130.2, 130.2, 130.2, 130.1, 129.0, 129.7, 127.9, 127.8, 127.8, 127.8, 127.7, 127.7, 108.4 (d, J = 265.5 Hz), 103.2 (d, J = 273.9 Hz), 101.4 (d, J = 19.2 Hz), 99.3 (d, J = 31.5 Hz), 83.8, 83.4, 77.6 (d, J = 15.7 Hz), 76.7 (d, J = 15.1 Hz), 63.1, 62.5, 26.9, 26.9, 26.8, 26.8, 19.44, 19.11, 19.01. ¹⁹F NMR (377 MHz, CDCl₃) δ −131.45 (s), −139.68 (d, J = 10.9 Hz). HRMS (ESI): m/z [M + Na]⁺ calcd for C₃₇H₄₄BrFNaO₄Si₂: 729.1843, found: 729.1831.

(3S,4R,5R)-3-Bromo-4-((tert-butyldiphenylsilyl)oxy)-5-(((tert-butyldiphenylsilyl)oxy)methyl)-3-fluorotetrahydrofuran-2-yl Methanesulfonate (10).

To a solution of 9 (1.6 g, 2.26 mmol) in CH₂Cl₂ (15 mL) were added Et₃N (0.62 mL, 4.5 mmol) and MsCl (0.26 mL, 3.4 mmol) 0 °C. After stirring 1 h at 0 °C, the mixture was allowed to warm up to rt and stirred for 1 h. The reaction mixture was then diluted with CH₂Cl₂ (100 mL), washed with 1 N HCl (25 mL) followed by 5% NaHCO₃ (25 mL) and brine (25 mL). The organic layer was dried over Na₂SO₄, filtered, and concentrated in vacuo to give crude product 10 (1.72 g, quantitative) as a mixture of anomers (α/β ratio 2.5:1). This anomeric mixture was dried under high vacuum and used as described in the next step.

¹H NMR (400 MHz, CDCl₃) δ 7.75–7.34 (m, 54H), 6.25 (s, 1.7H), 6.06 (d, J = 7.4 Hz, 1H), 4.61–4.59 (m, 1.7H), 4.53–4.51 (m, 2.7H), 4.32 (dd, J = 19.5, 7.7 Hz, 1H), 3.72–3.54 (m, 3.7H), 3.41 (dd, J = 11.9, 6.5 Hz, 1H), 3.17 (s, 5.4H), 2.78 (s, 3H), 1.15 (2s merged, 24.3H), 1.01 (2s merged, 24.3H). ¹³C NMR (101 MHz, CDCl₃) δ 136.2, 136.0, 135.8, 135.8, 135.6, 135.6, 135.5, 135.5, 132.9, 132.7, 132.7, 132.6, 132.5, 132.2, 131.9, 131.3, 130.4, 130.3, 130.3, 130.3, 129.9, 129.8, 128.0, 127.9, 127.8, 127.8, 127.7, 127.7, 104.9 (d, J = 19.6 Hz), 104.7 (d, J = 260.1 Hz), 102.8 (d, J = 36.5 Hz), 101.7 (d, J = 288.7 Hz), 87.9, 84.9, 77.0 (d, J = 15.8 Hz), 76.7 (d, J = 15.4 Hz), 63.6, 62.5, 40.2, 39.9, 26.8, 26.8, 26.7, 19.47, 19.42, 19.17, 19.13. ¹⁹F NMR (377 MHz, CDCl₃) δ −131.1, −133.4 (dd, J = 20.4, 6.7 Hz).

1-((3S,4R,5R)-3-Bromo-4-((tert-butyldiphenylsilyl)oxy)-5-(((tert-butyldiphenylsilyl)oxy)methyl)-3-fluorotetrahydrofuran-2-yl)-pyrimidine-2,4(1H,3H)-dione (11).

To a suspension of uracil (0.187 g, 1.67 mmol) in DCE (1 mL) was added BSA (0.817 mL, 3.34 mmol). The reaction mixture was stirred at 60 °C for 30 min and then allowed to cool to rt. To the resulting homogeneous solution, 10 (0.655 g, 0.835 mmol) in DCE (2 mL) and TMSOTf (0.604 mL, 3.34 mmol) were added. The reaction mixture was then stirred at 80 °C for 5 h. The reaction was quenched by addition of 5% aqueous solution of NaHCO₃ (15 mL) at 0 °C, filtered through celite, and washed with ethyl acetate (25 mL). The aqueous layer was extracted with ethyl acetate (25 mL), and the combined organic layers were washed with a saturated solution of NaHCO₃ (10 mL), water (10 mL), and brine (10 mL). The solution was dried over Na₂SO₄ and concentrated in vacuo. The residue was purified by flash chromatography using EtOAc/hexane gradient to afford 11 (0.385 g, 57%) as a 2:1 (α/β) mixture.

¹H NMR (400 MHz, CDCl₃) δ 9.36 (2s merged, 1.5H), 7.79–7.12 (m, 31.5 H), 6.46–6.38 (2d merged, 1.5H), 5.80–5.77 (m, 1.5H), 4.80–4.72 (m, 1H), 4.57 (dd, J = 15.1, 8.2 Hz, 0.5H), 4.33–4.29 (m, 1H), 4.17–4.07 (m, 0.5H), 3.98 (dd, J = 12.1, 2.4 Hz, 0.5H), 3.82 (dd, J = 12.1, 2.4 Hz, 0.5H), 3.69 (dd, J = 11.8, 2.7 Hz, 1H), 3.45 (dd, J = 11.8, 3.6 Hz, 1H), 1.10–1.08 (2s merged, 13.5H), 13.5 (s, 13.5H). ¹³C NMR (101 MHz, CDCl₃) δ 163.0, 162.7, 150.2, 150.1, 140.7, 140.6, 138.9, 136.3, 136.1, 135.9, 135.6, 135.5, 135.5, 135.2, 133.1, 132.7, 132.5, 132.4, 132.1, 131.8, 131.6, 131.5, 130.3, 130.1, 129.9, 129.8, 129.8, 128.0, 127.9, 127.9, 127.8, 127.8, 127.7, 108.3 (d, J = 268.1 Hz), 102.6, 102.5 (d, J = 273.2 Hz), 102.5, 88.1 (d, J = 39.3 Hz), 87.7 (d, J = 15.4 Hz), 84.1, 82.5, 77.7 (d, J = 15.4 Hz), 76.8 (d, J = 16.2 Hz), 62.7, 61.0, 26.9, 26.8, 26.7, 19.4, 19.4, 19.3, 19.0. HRMS (ESI): m/z [M + H]⁺ calcd for C₄₁H₄₇BrFN₂O₅Si₂: 801.2191, found: 801.2186.

1-((2R, 3S, 4R, 5R)-3-Bromo-3-fluoro-4-hydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)pyrimidine-2,4(1H,3H)-dione (β-12) and 1-((2S,3S,4R,5R)-3-Bromo-3-fluoro-4-hydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)pyrimidine-2,4(1H,3H)-dione (α-12).

To a stirred solution of 11 (0.385 g, 0.48 mmol) in anhydrous THF (2.5 mL), was added 1 M solution of TBAF in THF (0.962 mL, 0.962 mmol) at 0 °C. The reaction mixture was stirred for 1 h at rt. The solvent was evaporated under reduced pressure, and the residue was purified by flash chromatography using 0–6% MeOH/CH₂Cl₂ gradient to afford 12 (β-isomer, 34 mg, 21%) and α-isomer (72 mg, 46%).

β-Anomer (12): ¹H NMR (400 MHz, MeOD-d₄) δ 7.96 (d, J = 8.1 Hz, 1H), 6.36 (d, J = 16.5 Hz, 1H), 5.76 (d, J = 8.2 Hz, 1H), 4.44 (dd, J = 19.8, 9.2 Hz, 1H), 4.01 (dd, J = 12.7, 2.2 Hz, 1H), 3.98–3.92 (m, 1H), 3.80 (dd, J = 12.7, 2.7 Hz, 1H). ¹³C NMR (101 MHz, MeOD-d₄) δ 164.2, 150.6, 139.9, 109.3 (d, J = 261.5 Hz), 101.8, 88.4 (d, J = 39.7 Hz), 81.4, 75.2 (d, J = 17.2 Hz), 58.5. ¹⁹F NMR (400 MHz, MeOD-d₄) δ −122.60 (s). HRMS (ESI): m/z [M + H]⁺ calcd for C₉H₁₁BrFN₂O: 324.9835, found: 324.9833. α-Anomer: ¹H NMR (400 MHz, MeOD-d₄) δ 7.62 (dd, J = 8.2, 3.4 Hz, 1H), 6.59 (d, J = 18.1 Hz, 1H), 5.75 (d, J = 8.2 Hz, 1H), 4.62 (dd, J = 20.6, 8.8 Hz, 1H), 4.22–4.17 (m, 1H), 3.89 (dd, J = 12.7, 2.5 Hz, 1H), 3.70 (dd, J = 12.7, 3.4 Hz, 1H). ¹³C NMR (101 MHz, MeOD-d₄) δ 164.3, 150.7, 141.4 (d, J = 6.1 Hz), 103.5 (d, J = 268.2 Hz), 101.5, 87.3 (d, J = 14.8 Hz), 82.6, 76.0 (d, J = 16.5 Hz), 59.8 ¹⁹F NMR (377 MHz, MeOD-d₄) δ −139.05 (s). HRMS (ESi): m/z [M + H]⁺ calcd for C₉H₁₁BrFN₂O: 324.9835, found: 324.9834.

(2S)-Isopropyl ((((2R,3R,4S,5R)-4-Bromo-5-(2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)-4-fluoro-3-hydroxytetrahydrofuran-2-yl)-methoxy)(phenoxy)phosphoryl)-propanoate (14).

To a stirred solution of high-vacuum-dried 12 (21 mg, 0.064 mmol) and (2S)-isopropyl 2-((chloro(phenoxy)phosphoryl)amino)propanoate, 13 (39 mg, 0.13 mmol) in 1 mL of anhydrous THF under nitrogen atmosphere was added NMI (10 μL, 0.13 mmol) slowly. After stirring for 2 h at 0 °C, the reaction was warmed slowly to rt and stirred for 2 h. The reaction was quenched with isopropyl alcohol (0.2 mL). The solvent was removed under reduced pressure, and the residue was purified by flash chromatography using 0–6% MeOH/CH₂Cl₂ to afford 14 (17 mg, 45%) as a diastereomeric (R_p/S_p ~ ratio 7:3) mixture.

¹H NMR (400 MHz, MeOD-d₄) δ 7.57 (2d merged, each J = 8.1 Hz, 1H), 7.42–7.38 (m, 2H), 7.32–7.15 (m, 3H), 6.39–6.32 (m, merged, 1H), 5.70 (2d, each J = 8.1 Hz, 1H), 5.04–4.96 (m, 1H), 4.65–4.33 (m, 4H), 4.23–4.08 (m, 1H), 3.95–3.90 (m, 1H), 1.38–1.32 (m, 3H), 1.36–1.23 (m, 6H). ¹⁹F NMR (377 MHz, MeOD-d₄) δ −121.77, −122.10. ¹³C NMR (101 MHz, MeOD-d₄) δ 173.2 (d, J = 4.5 Hz), 172.9 (d, J = 5.4 Hz), 164.1, 150.7, 150.4, 139.9, 139.6, 129.5, 124.9, 124.9, 120.0, 119.9, 119.9, 109.9, 107.3, 102.2, 102.1, 79.2, 79.2, 76.1 (d, J = 17.2 Hz), 75.8 (d, J = 17.3 Hz), 68.8, 68.8, 64.1, 63.7, 50.4, 50.3, 20.6, 20.5, 19.1 (d, J = 6.4 Hz), 18.9 (d, J = 7.4 Hz). ³¹P NMR (162 MHz, MeOD-d₄) δ 3.64, 3.54. HRMS (ESI): m/z [M + H]⁺ calcd for C₂₁H₂₇BrFN₃O₉P: 594.0652, found: 594.0642.

N-(1-((3S,4R,5R)-3-Bromo-4-((tert-butyldiphenylsilyl)oxy)-5-(((tert-butyldiphenylsilyl)oxy)methyl)-3-fluorotetrahydrofuran-2-yl)-2-oxo-1,2-dihydropyrimidin-4-yl)benzamide (15).

To a suspension of N-benzoyl cytosine (0.502 g, 2.33 mmol) in DCE (1 mL) was added BSA (1.142 mL, 4.67 mmol). The reaction mixture was stirred at 60 °C for 30 min and then allowed to cool to rt. To the resulting homogeneous solution, 10 (0.916 g, 1.16 mmol) in DCE (2 mL) and TMSOTf (0.846 mL, 4.67 mmol) was added. The reaction mixture was then stirred at 80 °C for 5 h. The reaction was quenched by addition of 5% aqueous solution of NaHCO₃ (15 mL) at 0 °C. The aqueous layer was extracted with ethyl acetate (50 mL), and the combined organic layers were washed with a saturated solution of NaHCO₃ (25 mL), water (25 mL), and brine (25 mL). The solution was dried over Na₂SO₄ and concentrated in vacuo. The residue was purified by flash chromatography using EtOAc/hexane gradient to afford 15 (0.576 g, 55%) as a 2:1 (α/β) mixture.

¹H NMR (400 MHz, CDCl₃) δ 8.05–7.25 (m, 38H), 6.68 (d, J = 15.0 Hz, 1H), 6.59 (d, J = 14.2 Hz, 0.4H), 4.77 (dd, J = 15.3, 6.7 Hz, 1H), 4.62 (dd, J = 12.5, 7.7 Hz, 0.4H), 4.49–4.39 (m, 1H), 4.24–4.16 (m, 0.4H), 3.95 (dd, J = 12.1, 2.4 Hz, 0.4H), 3.80–3.71 (m, 1.4H),3.54 (dd, J = 11.8, 4.0 Hz, 1H), 1.13 (s, 3.6H), 1.11 (s, 9H), 0.99 (s, 3.6H), 0.96 (s, 9H). ¹³C NMR (101 MHz, CDCl₃) δ 166.9, 162.7, 162.4, 154.5, 145.4, 144.0, 136.8, 136.3, 136.2, 136.1, 136.0, 136.0, 135.9, 135.7, 135.7, 135.7, 135.6, 135.6, 135.4, 135.3, 133.7, 133.5, 133.2, 133.1, 132.8, 132.7, 132.5, 132.1, 132.0, 131.6, 131.5, 130.3, 130.3, 130.3, 130.1, 130.0, 129.8, 129.7, 129.5, 129.4, 129.3, 128.9, 128.9, 127.9, 127.8, 127.8, 127.8, 127.7, 127.7, 127.6, 127.5, 127.5, 127.4, 127.3, 127.2, 108.3 (d, J = 271.8 Hz), 102.0 (d, J = 273.9 Hz), 88.8 (d, J = 38.0 Hz), 88.6 (d, J = 15.6 Hz), 84.4, 82.6, 78.2 (d, J = 15.5 Hz), 77.3 (d, J = 16.2 Hz), 62.7, 61.3, 26.9, 26.8, 26.8, 26.7, 19.4, 19.0. ¹⁹F NMR (377 MHz, CDCl₃) δ −118.98 (t, J = 13.3 Hz), −136.16. HRMS (ESI): m/z [M + H]⁺ calcd for C₄₈H₅₂BrFN₃O₅Si₂: 904.2613, found: 904.2605.

N-(1- ((2R, 3S, 4R, 5R) -3-Bromo-3-fluoro-4-hydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)-2-oxo-1,2-dihydropyrimidin-4-yl)benzamide (16).

To a stirred solution of 15 (0.560 g, 0.62 mmol) in anhydrous THF (2.5 mL), was added 1 M solution of TBAF in THF (0.930 mL, 0.930 mmol) at 0 °C. The reaction mixture was stirred for 1 h at rt. The solvent was evaporated under reduced pressure, and the residue was purified by flash chromatography using 0–5% MeOH/CH₂Cl₂ gradient to afford β-16 (β-isomer, 62 mg, 24%) and α-isomer (122 mg in 46%).

β-Anomer: ¹H NMR (400 MHz, MeOD-d₄) δ 8.47 (d, J = 7.6 Hz, 1H), 8.01–7.99 (m, 2H), 7.69–7.64 (m, 2H), 7.58–7.54 (m, 2H), 6.52 (d, J = 16.0 Hz, 1H), 4.51 (dd, J = 19.2, 9.0 Hz, 1H), 4.13–3.97 (m, 1H), 3.85 (dd, J = 12.5, 2.5 Hz, 1H). ¹³C NMR (101 MHz, MeOD-d₄) δ 163.9, 156.4, 144.4, 133.2, 132.8, 128.4, 127.8, 108.9 (d, J = 262.9 Hz), 97.5, 89.4 (d, J = 40.5 Hz), 81.7, 75.3 (d, J = 17.1 Hz) 58.5. ¹⁹F NMR (377 MHz, MeOD-d₄) δ −123.31 (s). HRMS (ESI): m/z [M + H]⁺ calcd for C₁₆H₁₆BrFN₃O₅: 428.0257, found: 428.0251.

α-Anomer: ¹H NMR (400 MHz, DMSO-d₆) δ 11.39 (s, 1H), 8.16 (d, J = 5.9 Hz, 1H), 8.02 (d, J = 7.3 Hz, 2H), 7.65 (t, J = 7.4 Hz, 1H), 7.53 (t, J = 7.7 Hz, 2H), 7.42 (d, J = 7.6 Hz, 1H), 6.69 (d, J = 17.6 Hz, 1H), 6.60 (d, J = 6.9 Hz, 1H), 5.13 (dd, J = 6.5, 4.9 Hz, 1H), 4.57 (dt, J = 21.2, 7.8 Hz, 1H), 4.22 (d, J = 8.4 Hz, 1H), 3.77–3.71 (m, 1H), 3.65–3.52 (m, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 167.9, 164.2, 154.9, 146.9 (d, J = 5.0 Hz), 133.5, 133.3, 129.0, 128.9, 104.2 (d, J = 269.5 Hz), 97.2, 87.9 (d, J = 14.8 Hz), 83.3, 76.4 (d, J = 16.2 Hz), 60.3. ¹⁹F NMR (377 MHz, DMSO-d₆) δ −135.76 (t, J = 19.3 Hz). HRMS (ESI): m/z [M + H]⁺ calcd for C₁₆H₁₆BrFN₃O₅: 428.0257, found: 428.0250.

4-Amino-1-((2R,3S,4R,5R)-3-bromo-3-fluoro-4-hydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)pyrimidin-2(1H)-one (17).

Compound 16 (60 mg, 0.14 mmol) was dissolved in 20% NH₃/MeOH (5 mL), and the reaction mixture was stirred overnight at rt. After the solvent was removed under reduced pressure, the crude product was purified by flash chromatography using 0–10% MeOH/CH₂Cl₂ gradient to afford nucleoside 17 (42 mg, 93%).

¹H NMR (400 MHz, MeOD-d₄) δ 7.93 (d, J = 7.5 Hz, 1H), 6.44 (d, J = 16.8 Hz, 1H), 5.95 (d, J = 7.6 Hz, 1H), 4.42 (dd, J = 19.5, 9.3 Hz, 1H), 4.01 (dd, J = 12.7, 2.2 Hz, 1H), 3.96–3.90 (m, 1H), 3.81 (dd, J = 12.7, 2.7 Hz, 1H). ¹³C NMR (101 MHz, MeOD-d₄) δ 166.2, 156.6, 140.5, 109.5 (d, J = 261.9 Hz), 95.2, 89.1 (d, J = 38.1 Hz), 81.1, 75.4 (d, J = 17.2 Hz), 58.6. ¹⁹F NMR (377 MHz, MeOD-d₄) δ −122.31. HRMS (ESI): m/z [M + H]⁺ calcd for C₉H₁₂BrFN₃O₄: 323.9995, found: 323.9992.

Isopropyl ((((2R,3R,4S,5R)-5-(4-Amino-2-oxopyrimidin-1(2H)-yl)-4-bromo-4-fluoro-3-hydroxytetrahydrofuran-2-yl)methoxy)-(phenoxy)phosphoryl)-l-alaninate (18).

To vacuum-dried 17 (20 mg, 0.062 mmol) in 1 mL anhydrous THF under nitrogen atmosphere was added t-BuMgCl (0.093 mL, 0.093 mmol, 1.5 equiv) at 0 °C. After stirring at 0 °C for 30 min, (2S)-isopropyl 2-((chloro(phenoxy)-phosphoryl)amino)propanoate, 13 (18 mg, 0.062 mmol) in 1 mL of anhydrous THF was added. The reaction mixture was allowed to attain rt and stirred for 3 h. The reaction was quenched with isopropyl alcohol (0.2 mL). The solvent was removed under reduced pressure, and the residue was purified by flash chromatography using 0–6% MeOH/CH₂Cl₂ to afford 18 (7.7 mg, 21%) as a diastereomeric (R_p/S_p ~ ratio 6:4) mixture.

¹H NMR (400 MHz, MeOD-d₄) δ 7.55 (2d merged, 1H), 7.42–7.37 (m, 2H), 7.29–7.20 (m, 3H), 6.48–6.40 (m, 1H), 5.92–5.88 (2d merged, 1H), 5.04–4.97 (m, 1H), 4.61–4.34 (m, 4H), 4.14–4.09 (m, 1H), 3.95–3.90 (m, 1H), 1.37–1.31 (m, 3H), 1.26–1.23 (m, 6H). ¹³C NMR (101 MHz, MeOD-d₄) δ 173.2 (d, J = 4.5 Hz), 172.9 (d, J = 5.3 Hz), 166.2, 166.1, 156.4, 156.4, 150.7, 150.7, 150.7, 150.6, 140.4, 140.2, 129.5, 124.9, 120.0, 120.0, 119.9, 119.9, 108.9 (d, J = 262.8 Hz), 95.5, 95.5 79.0, 78.9, 76.2 (d, J = 17.4 Hz), 75.9 (d, J = 17.4 Hz), 68.8, 68.8, 64.1, 63.8, 50.4, 50.3, 20.6, 20.5, 19.1 (d, J = 6.5 Hz), 18.9 (d, J = 7.5 Hz). ¹⁹F NMR (377 MHz, MeOD-d₄) δ −121.75, −122.03. ³¹P NMR (162 MHz, MeOD-d₄) δ 3.59, 3.49. HRMS (ESI): m/z [M + H]⁺ calcd for C₂₁H₂₈BrFN₄O₈P: 593.0812, found: 593.0813.

3',5'-Bis-O-tert-butyldiphenylsilyl-2'-β-bromo,2'-α-fluoro-N⁶-bis-tert-butylcarbonate-2'-deoxyadenosine (19).

To a solution of lactol 9 (0.42 g. 0.6 mmol), triphenyl phosphine (0.283 g, 1.08 mmol), and N-Boc₂ adenine (0.3 g, 1.08 mmol) in THF (5 mL) under nitrogen atmosphere was added DIAD (0.212 mL, 1.08) dropwise at 0 °C. The resulting solution was stirred for 24 h at rt and then evaporated under reduced pressure to give a yellow syrup. This syrup was purified by silica gel column chromatography with a mixture of hexane and ethyl acetate (15% EA) to give mixture of α/β product 19 (1:0.4) as a white solid (0.242 g, 39%).

¹H NMR (400 MHz, CDCl₃) δ 8.87 (s, 0.4H), 8.67 (s, 1H), 8.40 (d, J = 3.5 Hz, 0.4H), 8.12 (s, 1H), 7.76–7.66 (m, 28H), 6.67 (d, J = 16.9 Hz, 0.4H), 6.51 (d, J = 13.1 Hz, 1H), 5.00 (dd, J = 17.1, 7.0 Hz, 0.4H), 4.81 (dd, J = 12.6, 7.2 Hz, 1H), 4.49–4.30 (m, 1.4 H), 3.79–3.68 (m, 2.4H), 3.50 (dd, J = 11.9, 2.9 Hz, 0.4H), 1.49 (s, 7H), 1.42 (s, 18H), 1.12–0.97 (m, 25H). ¹⁹F NMR (377 MHz, CDCl₃) δ −122.77, −136.83 (t, J = 17.2 Hz). ¹³C NMR (101 MHz, CDCl₃) δ 153.5, 152.5, 152.5, 152.2, 150.5, 150.4, 150.3, 150.1, 143.6, 143.5, 142.5, 136.1, 135.9, 135.9, 135.5, 135.5, 135.5, 135.4, 132.7, 132.7, 132.4, 132.3, 132.2, 131.5, 130.2, 130.3, 130.3, 130.3, 129.8, 129.8, 129.7, 128.8, 128.3, 127.9, 127.9, 127.8, 127.7, 127.7, 127.6, 127.6, 127.6, 108.2 (d, J = 269.6 Hz), 102.3 (d, J = 272.4 Hz), 88.4 (d, J = 37.2 Hz), 87.6 (d, J = 16.4 Hz), 83.9, 83.8, 83.7, 77.9 (d, J = 15.3 Hz), 77.6 (d, J = 15.5 Hz). 62.6, 27.8, 27.8, 27.7, 26.8, 26.8, 19.4, 19.4, 19.1, 19.0. ¹⁹F NMR (377 MHz, CDCl₃) δ −122.77, −136.83 (t, J = 17.2 Hz). HRMS (ESI): m/z [M + H]⁺ calcd for C₅₂H₆₄BrFN₅O₇Si₂: 1024.3512, found: 1024.3505.

2'-β-Bromo-2'-α-fluoro-N⁶-bis-tert-butylcarbonate-2'-deoxyadenosine (20).

To a stirred solution of 19 (0.242 g, 0.24 mmol) in THF (3 mL), was added 1 M solution of TBAF in THF (0.519 mL, 0.516 mmol) at 0 °C. The reaction mixture was allowed to stir for 1 h at the same temperature. The solvent was evaporated under reduced pressure, and the residue was purified by flash chromatography using 0–6% MeOH/CH₂Cl₂ gradient to afford 20 (β-isomer, 58 mg, 44%) and α-isomer (25 mg, 19%).

α-Anomer: ¹H NMR (400 MHz, MeOD-d₄) δ 8.90 (s, 1H), 8.71 (s, 1H), 6.91 (d, J = 17.2 Hz, 2H), 4.83–4.75 (m, 1H), 4.43 (d, J = 10.1 Hz, 1H), 3.96 (dd, J = 12.7, 2.4 Hz, 1H), 3.77 (dd, J = 12.7, 3.3 Hz, 1H), 1.39 (s, 18H). ¹³C NMR (101 MHz, MeOD-d₄) δ 154.9, 153.6, 151.4, 151.3, 146.6, 146.5, 129.8, 104.5 (d, J = 268.0 Hz), 89.5 (d, J = 15.9 Hz), 85.4, 84.6, 77.6 (d, J = 16.6 Hz), 61.2, 27.9. ¹⁹F NMR (377 MHz, MeOD-d₄) δ −136.09. HRMS (ESI): m/z [M + H]⁺ calcd for C₂₀H₂₈BrFN₅O: 548.1156, found: 548.1146.

β-Anomer: ¹H NMR (400 MHz, MeOD-d₄) δ 8.96 (s, 1H), 8.91 (s, 1H), 6.71 (d, J = 14.7 Hz, 1H), 4.92–4.82 (m, 1H), 4.14 (d, J = 10.4 Hz, 1H), 4.07 (dd, J = 12.7, 2.2 Hz, 1H), 3.93 (dd, J = 12.7, 3.3 Hz, 1H), 1.37 (s, 18H). ¹³C NMR (101 MHz, MeOD-d₄) δ 154.4, 153.5, 151.4, 151.2, 145.7, 130.3, 110.5 (d, J = 261.3 Hz), 89.9 (d, J = 38.8 Hz), 85.4, 83.6, 76.3 (d, J = 16.9 Hz), 60.6, 27.9. ¹⁹F NMR (377 MHz, MeOD-d₄) δ −123.63. HRMS (ESI): m/z [M + H]⁺ calcd for C₂₀H₂₈BrFN₅O: 548.1156, found: 548.1147.

(2R,3R,4S,5R)-5-(6-Amino-9H-purin-9-yl)-4-bromo-4-fluoro-2-(hydroxymethyl)tetrahydrofuran-3-ol (21).

To a solution of β product 20 (0.040 g, 0.072 mmol) in dry DCM (1 mL) at −78 °C was added dropwise a 1 M solution BCl₃ in DCM (0.145 mL, 0.145 mmol), and the reaction mixture was allowed to warm to rt and stirred for 1 h. The reaction was quenched by careful addition of MeOH, and volatiles were evaporated under reduced pressure. The residue was purified by a silica gel column chromatography using 0–10% MeOH/CH₂Cl₂ gradient to afford the nucleoside derivative 21 (17.3 mg, 69%) as a white solid.

¹H NMR (400 MHz, DMSO-d₆) δ 8.42 (s, 1H), 8.18 (s, 1H), 7.43 (bs, 2H, NH₂), 6.54 (d, J = 7.2 Hz, 1H), 6.47 (d, J = 15.8 Hz, 1H), 5.37 (t, J = 5.1 Hz, 1H, OH), 4.79 (dt, J = 22.1, 8.3 Hz, 1H), 3.99–3.97 (m, 1H), 3.89–3.84 (m, 1H), 3.79–3.73 (m, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 156.6, 153.4, 149.5, 139.0, 119.2, 110.4 (d, J = 258.9 Hz), 88.1 (d, J = 38.1 Hz), 82.2, 75.0 (d, J = 16.7 Hz), 59.9. ¹⁹F NMR (377 MHz, DMSO-d₆) δ −120.95. HRMS (ESI): m/z [M + H]⁺ calcd for C₁₀H₁₂BrFN₅O₃: 348.0108, found: 348.0105.

Isopropyl ((((2R,3R,4S,5R)-5-(2-Amino-6-oxo-1,6-dihydro-9H-purin-9-yl)-4-bromo-4-fluoro-3-hydroxytetrahydrofuran-2-yl)-methoxy)(phenoxy)phosphoryl)-L-alaninate (22).

To a stirred solution of 20 (60 mg, 0.11 mmol) in anhydrous THF (1 mL) were added (2S)-isopropyl 2-((chloro(phenoxy)phosphoryl)amino)-propanoate (66 mg, 0.21 mmol) in 1 mL of anhydrous THF and NMI (34 μL, 0.44 mmol) slowly at 0 °C under nitrogen atmosphere. The reaction mixture allowed to warm slowly to rt and stirred for 3 h. The reaction was quenched with isopropyl alcohol (0.5 mL). The solvents were evaporated under reduced pressure. A 50% TFA in DCM (3 mL) solution was added to the crude product at 0 °C and then stirred at rt overnight and concentrated in vacuo. A 5% aqueous solution of NaHCO₃ (2 mL) was slowly added, and the water was evaporated. The residue was purified by flash chromatography using 0–10% MeOH/CH₂Cl₂ to afford 22 (26 mg, 40%) as a diastereomeric (R_p/S_p ~ ratio 7:3) mixture over two steps.

¹H NMR (400 MHz, MeOD-d₄) δ 8.26–8.24 (m, 2H), 7.37–7.33 (m, 2H), 7.26–7.18 (m, 3H), 6.56–6.50 (m, 1H), 5.07–4.96 (m, 2H), 4.62–4.57 (m, 2H), 4.31–4.26 (m, 1H), 3.93–3.86 (m, 1H), 1.32–1.29 (m, 3H), 1.22–1.13 (m, 6H). ⁹F NMR (377 MHz, MeOD-d₄) δ −124.58 (dd, J = 19.4, 16.1 Hz), −124.71 (dd, J = 19.4, 16.2 Hz). ¹³C NMR (101 MHz, MeOD-d₄) δ 173.1 (d, J = 4.6 Hz), 172.9 (d, J = 5.6 Hz), 156.1, 156.1, 152.8, 152.8, 150.7, 150.6, 149.1, 149.0, 139.3, 139.0, 129.4, 129.4, 124.8, 120.0, 119.9, 119.0, 118.9, 108.4 (d, J = 260.2 Hz), 108.4 (d, J = 260.3 Hz), 88.9 (d, J = 39.1 Hz), 88.7 (d, J = 39.2 Hz), 80.0 (d, J = 8.3 Hz), 79.9 (d, J = 8.7 Hz), 75.9 (d, J = 17.1 Hz), 75.6 (d, J = 17.2 Hz), 68.8, 68.7, 65.2, 64.6, 50.4, 50.2, 20.5, 20.5, 20.5, 20.4, 19.1 (d, J = 6.5 Hz), 18.9 (d, J = 7.1 Hz). ³¹P NMR (162 MHz, MeOD-d₄) δ 3.63, 3.51. HRMS (ESI): m/z [M + H]⁺ calcd for C₂₂H₂₈BrFN₆O₇P: 617.0924, found: 617.0925.

3',5'-Bis-O-tert-butyldiphenylsilyl-2'-deoxy-2'-β-bromo-2'-α-fluoro-6-benzyloxy-N²-bis-tert-butylcarbonatepurine Ribonucleoside (23).

To a solution of lactol 9 (0.5 g. 0.71 mmol), triphenyl phosphine (0.427 g, 1.27 mmol), and O⁶-benzyl-N-Boc₂ guanine (0.57 g, 1.27 mmol) in THF (5 mL) under nitrogen atmosphere was added DIAD (0.25 mL, 1.27 mmol) dropwise at 0 °C. The resulting solution was stirred for 24 h at rt and then evaporated under reduced pressure to give a yellow syrup. This syrup was purified by a silica gel column chromatography with a mixture of hexane and ethyl acetate (15%) to give a mixture of α/β product (ratio ~1:0.4) 23 as a white solid (0.320 g, 40%).

¹H NMR (400 MHz, CDCl₃) δ 8.25 (d, J = 3.6 Hz, 0.3H), 7.95 (s, 1H), 7.69–7.29 (m, 32.5 H), 6.56–6.47 (m, 1.3 H), 5.66–5.62 (m, 2.6H), 4.91 (dd, J = 16.5, 6.8 Hz, 0.3H), 4.68–4.64 (m, 1H), 4.42–4.40 (m, 0.3H), 4.26–4.24 (m, 0.3H), 3.90–3.58 (m, 2.6H), 3.47–3.42 (m, 0.3H), 1.40–1.28 (m, 23.5H), 1.09–0.96 (m, 23.5H). ¹³C NMR (101 MHz, CDCl₃) δ 161.1, 160.9, 153.3, 152.7, 152.2, 152.1, 150.7, 150.7, 150.4, 142.1, 142.1, 140.7, 140.5, 136.1, 135.9, 135.8, 135.8, 135.8, 135.6, 135.5, 135.5, 135.4, 135.4, 132.7, 132.7, 132.6, 132.5, 132.4, 132.4, 132.1, 131.6, 131.6, 130.4, 130.3, 130.3, 130.3, 130.1, 130.1, 129.9, 129.9, 129.8, 129.8, 128.6, 128.5, 128.5, 128.4, 128.3, 128.3, 128.3, 128.2, 127.9, 127.8, 127.8, 127.7, 127.6, 119.7, 119.6, 119.3, 108.4 (d, J = 272.8 Hz), 102.4 (d, J = 273.7 Hz), 87.7 (d, J = 16.4 Hz), 87.5 (d, J = 36.4 Hz) 84.0, 83.1, 83.0, 82.8, 78.0 (d, J = 15.4 Hz), 77.6 (d, J = 15.5 Hz), 72.2, 68.9, 68.8, 62.7, 62.3, 27.8, 27.8, 26.8, 26.6, 21.6, 19.4, 19.4, 19.1, 19.00. ¹⁹F NMR (377 MHz, CDCl₃) δ −122.03 (t, 9.8 Hz), −134.67 (td, J = 16.5, 3.9 Hz). HRMS (ESI): m/z [M + H]⁺ calcd for C₅₉H₇₀BrFN₅O₈Si₂: 1130.3930, found: 1130.3928.

3',5'-Bis-O-tert-butyldiphenylsilyl-2'-deoxy-2'-β-bromo-2'-α-fluoro-6-benzyloxy-N²-bis-tert-butylcarbonatepurine Ribonucleoside (24).

To a stirred solution of 23 (0.300 g, 0.24 mmol) in THF (5 mL), was added 1 M solution of TBAF in THF (0.58 mL, 0.58 mmol) at 0 °C. The reaction mixture was allowed to stir for 1 h at the same temperature. The solvent was evaporated under reduced pressure, and the residue was purified by flash chromatography using a 0–5% MeOH/CH₂Cl₂ gradient to afford 24 (β-isomer, 104 mg, 60%).

¹H NMR (400 MHz, CDCl₃) δ 8.54 (s, 1H), 7.51 (d, J = 7.8 Hz, 1H), 7.40–7.34 (m, 3H), 6.42 (d, J = 14.8 Hz, 1H), 5.61 (s, 2H), 4.90–4.93 (m, 1H), 4.12–4.07 (m, 2H), 3.86 (d, J = 4.4 Hz, 1H), 1.39 (s, 18H). ¹³C NMR (101 MHz, CDCl₃) δ 161.0, 152.1, 152.1, 150.6, 141.6, 135.5, 128.6, 128.4, 128.4, 119.8, 108.8 (d, J = 261.2 Hz), 89.1 (d, J = 38.5 Hz), 83.6, 82.0, 77.3, 74.3 (d, J = 16.7 Hz), 69.1, 59.1, 27.9. ¹⁹F NMR (377 MHz, CDCl₃) δ −122.41. HRMS (ESI): m/z [M + H]⁺ calcd for C₂₇H₃₄BrFN₅O₈: 654.1575, found: 654.1566.

2-Amino-9-((2R,3S,4R,5R)-3-bromo-3-fluoro-4-hydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)-1,9-dihydro-6H-purin-6-one (25).

To a solution of β product 24 (0.058 g, 0.088 mmol) in dry DCM (1 mL) at −78 °C was added 1 M solution BCl₃ in DCM (0.177 mL, 0.177 mmol) dropwise, and the reaction mixture was allowed to warm to −20 °C and stirred for 1 h. The reaction was next stirred at rt for 2 h. The reaction was quenched by careful addition of MeOH (0.3 mL), and volatiles were evaporated under reduced pressure. The residue was purified by a silica gel column chromatography using a 0–10% MeOH/CH₂Cl₂ gradient to afford the nucleoside derivative 25 (21 mg, 65%) as a white solid.

¹H NMR (400 MHz, DMSO-d₆) δ 10.90 (bs, 1H),7.99 (s, 1H), 6.81 (bs, 2H), 6.54 (d, J = 7.2 Hz, 1H), 6.17 (d, J = 15.8 Hz, 1H), 5.38 (t, J = 5.2 Hz, 1H), 4.66–4.55 (m, 1H), 3.92 (d, J = 9.2 Hz, 1H), 3.85–3.78 (m, 1H), 3.73–3.68 (m, 1H). ¹³C NMR (101 MHz, DMSO-d₆) δ 157.1, 154.6, 151.5, 134.8, 116.8, 110.4 (d, J = 258.8 Hz), 87.3 (d, J = 37.7 Hz), 82.0, 74.7 (d, J = 16.5 Hz), 59.5. ¹⁹F NMR (377 MHz, DMSO-d₆) δ −122.73. HRMS (ESI): m/z [M + H]⁺ calcd for C₁₀H₁₂BrFN₅O₄: 364.0057, found: 364.0054.

Isopropyl ((((2R,3R,4S,5R)-5-(2-Amino-6-oxo-1,6-dihydro-9H-purin-9-yl)-4-bromo-4-fluoro-3-hydroxytetrahydrofuran-2-yl)-methoxy)(phenoxy)phosphoryl)-l-alaninate (26).

To a stirred solution of 24 (52 mg, 0.08 mmol) in anhydrous THF (1.5 mL) were added (2S)-isopropyl 2-((chloro(phenoxy)phosphoryl)amino)-propanoate (48 mg, 0.16 mmol) in 1 mL of anhydrous THF and NMI (25 μL, 0.32 mmol) slowly at 0 °C under a nitrogen atmosphere. The reaction mixture was allowed to warm slowly to rt and stirred for 3 h. The reaction was quenched with isopropyl alcohol (0.5 mL). The solvents were evaporated under reduced pressure. The residue was purified by column using 0–10% MeOH/CH₂Cl₂ gradient to obtain the crude 6-O-Bn-2-N-Boc₂ prodrug intermediate (25 mg, 0.027 mmol). This crude product was dissolved in anhydrous DCM (3 mL) and cooled at −78 °C. A 1 M solution of BCl₃ (32 μL, 0.032 mmol) in DCM was added dropwise, and the reaction mixture was allowed to warm to −20 °C then stirred for 1 h. The reaction was next stirred at rt for 2 h. The reaction was quenched by careful addition of MeOH (0.3 mL), and the volatiles were evaporated under reduced pressure. The residue was first purified by a silica gel column chromatography using a 0–10% MeOH/CH₂Cl₂ gradient and second by a C18 reverse-phase column chromatography using water/acetonitrile (95:5–80:20) to afford the guanine nucleoside prodrug 26 (10 mg, 20% over two steps) as a diastereomeric (R_p/S_p ~ ratio 1:1).

¹H NMR (400 MHz, DMSO-d₆) δ 7.82 and 7.79 (2S, 1H), 7.37–7.33 (m, 2H), 7.15–7.23 (m, 3H), 6.86 (bs, 2H), 6.28–6.22 (m, 1H), 6.09 (b, 1H), 4.87–4.77 (m, 2H), 4.43–4.39 (m, 2H), 4.18–4.07 (m, 1H), 3.84–3.72 (m, 1H), 1.22–1.20 (m, 3H), 1.12–1.10 (m, 1H). ¹³C NMR (101 MHz, MeOD-d₄) δ 174.6, 174.5, 174.4, 174.3, 159.3, 155.5, 155.4, 152.9, 152.7, 152.1, 152.0, 137.7, 137.3, 130.8, 130.3, 126.2, 121.4, 121.4, 121.3, 118.1, 117.9, 111.2, 108.6, 90.4, 90.1, 90.0, 89.7, 81.3, 81.2, 81.1, 81.0, 77.5, 77.3, 77.2, 77.1, 70.2, 70.1, 66.7, 66.1, 66.1, 51.8, 51.6, 21.9, 21.9, 21.8, 20.5, 20.5, 20.4, 20.3. ¹⁹F NMR (377 MHz, DMSO-d₆) δ −121.50, −121.80. ^3lP NMR (162 MHz, DMSO-d₆) δ 3.61, 3.40. HRMS (ESI): m/z [M + H]⁺ calcd for C₂₂H₂₈BrFN₆O₈P: 633.0874.1575, found: 633.0864.

Isopropyl ((R)-(((2R,3R,4S,5R)-4-Bromo-5-(2,4-dioxo-3,4-dihydro-pyrimidin-1(2H)-yl)-4-fluoro-3-hydroxytetrahydrofuran-2-yl)-methoxy)(phenoxy)phosphoryl)-l-alaninate (27).

The nucleoside 12 (600 mg, 1.8 mmol) was dried at 50 °C under high vacuum for 1 h before adding dry THF (6 mL) at 25 °C. The mixture was cooled to −5 °C, and tert-butylmagnesium chloride (3.6 mL, 3.6 mmol, 1 M in THF) was introduced. The reaction mixture was stirred at −5 °C for 15 min, then warmed to 25 °C, and stirred for an additional 15 min. A solution of isopropyl ((R)-(perfluorophenoxy)-(phenoxy)phosphoryl)-l-alaninate (800 mg, 1.8 mmol, dried 2 h under high vacuum) in THF (6 mL) was added dropwise to the white suspension at 0 °C. After 16 h stirring at 4 °C, the white suspension was treated with 1 M HCl (10 mL). The mixture was extracted with EtOAc (2 × 10 mL). The combined organic layers were washed with water (10 mL), brine (10 mL), dried over MgSO₄, and concentrated under reduced pressure. The residue was purified by flash column chromatography (CH₂Cl₂ then CH₂Cl₂/MeOH 1–5%) to afford 27 (747 mg, 68%) as a white foam along with 104 mg (17%) of the unreacted nucleoside 12.

¹H NMR (400 MHz, MeOD-d₄) δ 7.57 (d, J = 8.2, 1H), 7.42–7.38 (m, 2H), 7.27–7.20 (m, 3H), 6.37 (d, J = 16.9 Hz, 1H), 5.73 (d, J = 8.2, 1H), 5.04–4.97 (m, 1H), 4.59 (d, 1H), 4.47–4.37 (m, 2H), 4.16–4.13 (m, 1H), 4.01–3.82 (m, 1H), 1.37–1.32 (m, 3H), 1.27–1.24 (m, 6H). ¹³C NMR (101 MHz, MeOD-d₄) δ 173.2 (d, J = 4.4 Hz), 164.1, 150.7, 150.5, 139.6, 129.5, 124.9, 119.9, 119.9, 108.6 (d, J = 262.5 Hz), 102.2, 79.2 (d, J = 8.7 Hz), 75.8 (d, J = 17.2 Hz), 68.8, 63.8, 50.4, 20.6, 20.5, 18.9 (d, J = 7.4 Hz). ¹⁹F NMR (377 MHz, MeOD-d₄) δ −122.08. ^3lP NMR (162 MHz, MeOD-d₄) δ 3.55. HRMS (ESI): m/z [M + H]⁺ calcd for C₂₁H₂₇BrFN₃O₉P: 594.0652, found: 594.0657.

Isopropyl ((S)-(((2R,3R,4S,5R)-4-Bromo-5-(2,4-dioxo-3,4-dihydro-pyrimidin-1(2H)-yl)-4-fluoro-3-hydroxytetrahydrofuran-2-yl)-methoxy)(phenoxy)phosphoryl)-l-alaninate (28).

Compound 12 (600 mg, 1.8 mmol) was dried at 50 °C under high vacuum for 1 h before adding dry THF (6 mL) at 25 °C. The mixture was cooled to −5 °C, and tert-butylmagnesium chloride (3.6 mL, 3.6 mmol, 1M in THF) was introduced. The reaction mixture was stirred at −5 °C for 15 min, then warmed to 25 °C, and stirred for an additional 15 min. A solution of isopropyl ((S)-(perfluorophenoxy)-(phenoxy)phosphoryl)-l-alaninate (750 mg, 1.7 mmol, dried 2 h under high vacuum) in THF (6 mL) was added dropwise to the white suspension at 0 °C. After 16 h stirring at 4 °C, the clear solution was quenched by an addition of 1 M HCl (10 mL). The mixture was extracted with EtOAc (2 × 10 mL). The combined organic layers were washed with water (10 mL), brine (10 mL), dried over MgSO₄, and concentrated under reduced pressure. The residue was purified by a flash column chromatography (CH₂Cl₂ then CH₂Cl₂/MeOH 1–5%) to afford 28 (744 mg, 68%) as a white foam along with 92 mg (15%) of the unreacted nucleoside 12.

¹H NMR (400 MHz, MeOD-d₄) δ 7.56 (d, J = 8.2 Hz, 1H), 7.42–7.37 (m, 2H), 7.31–7.17 (m, 3H), 6.34 (d, J = 16.9 Hz, 2H), 5.68 (d, J = 8.1 Hz, 1H), 5.05–4.95 (m, 1H), 4.59–4.36 (m, 3H), 4.20–4.08 (m, 1H), 3.97–3.88 (m, 1H), 1.37 (d, J = 7.1 Hz, 3H), 1.25–1.23 (m, 6H). ¹³C NMR (101 MHz, MeOD-d₄) δ 172.9 (d, J = 5.4 Hz), 164.1, 150.7 (d, J = 6.8 Hz), 150.4, 139.9, 129.5, 124.9, 120.0, 119.9, 108.6 (d, J = 262.6, Hz), 102.1, 79.3 (d, J = 7.8 Hz), 76.1 (d, J = 17.3 Hz), 68.8, 64.1, 50.3, 20.6, 20.5, 19.1 (d, J = 6.4 Hz). ¹⁹F NMR (377 MHz, MeOD-d₄) δ −121.77. ³¹P NMR (162 MHz, MeOD-d₄) δ 3.55. HRMS (ESI): m/z [M + H]⁺ calcd for C₂₁H₂₇BrFN₃O₉P: 594.0652, found: 594.0651.

Crystallography.

A suitable crystal (0.83 × 0.57 × 0.51 mm³) was selected and mounted on a loop with paratone oil on a Bruker APEX-II CCD diffractometer. The crystal was cooled to T = 100(2) K during data collection. The structure was solved with the XT³⁹ structure solution program using combined Patterson and dual-space recycling methods, and by using Olex2⁴⁰ as the graphical interface. The crystal structure was refined with version 2014/7 of XL⁴¹ using least-squares minimization.

Results from X-ray structure determination of 12 are the following. Crystal data for C₉H₁₀BrFN₂O₈ (M = 325.10 g/mol): tetragonal, space group P4₁2₁2 (no. 92), a = 17.6568(3) Å, b = 17.6568(3) Å, c = 14.3690(3) Å, α = 90°, β = 90°, γ = 90°, V = 4479.69(18) ų, Z = 16, T = 100(2) K, μ(Mo Kα) = 3.700 mm⁻¹, D_calc = 1.928 g/cm³. Intensity data were collected on a Bruker APEX-II CCD diffractometer with a monochromated Mo Kα radiation (λ = 0.7103 Å) at 100(2) Kin the 2θ range 3.654–61.012°. The user interface Olex2 was used for the crystallographic calculations and crystal structure visualization (Dolomanov et al., 2009). The structure was solved with Superflip by charge flipping and refined by least-squares minimization using SHELXL (Sheldrick, 2008, 2015). All nonhydrogen atoms were refined anisotropically, and hydrogen atoms were refined using a “riding” model. A total of 40 560 reflections were measured (3.654 ≤ 2θ ≤ 61.012), whereas 6853 unique data (R_int = 0.0456, R_sigma = 0.0264), which were used in all calculations. The final R₁ was 0.0284 (I > 2σ(I)), and wR₂ was 0.0665 (all data). GOF = 1.046. The maximum and the minimum peaks on the final difference Fourier map corresponded to 0.59 and −0.47 e/ų, respectively. The Flack absolute structure parameter was refined to −0.003(4), thus corroborating the stereochemistry of the title compound.

ABBREVIATIONS

BSA

N,O-bis(trimethylsilyl)acetamide

DAA

direct-acting antivirals

DCE

1,2-dichloroethane

DCM

dichloromethane

DIAD

diisopropyl azodicarboxylate

GT

genotype

LiHMDS

lithium bis(trimethylsilyl)amide

MsCl

trimethylsulfonyl chloride

NBS

N-bromosuccinimide

NFSI

N-fluorobenzenesulfonimide

NMI

N-methylimidazole

NOE

nuclear Overhauser effect

rt

room temperature

SD

standard deviation

SOF

sofosbuvir

TBDPS

t-butyldiphenylsilyl

TFA

trifluoroacetic acid

THF

tetrahydrofuran

TMSOTf

trimethysilyl trifluoromethanesulfonate

TP

triphosphate

NTP

nucleoside triphosphate

RdRp

RNA-dependent RNA polymerase