Synthesis and Anti-Hepatitis B Activities of 3′-Fluoro-2′-Substituted Apionucleosides

Nucleoside analogues have excellent records as anti-HBV drugs. Chronic infections require long-term administration ultimately leading to drug resistance. Therefore, the search for nucleosides with novel scaffolds is of high importance. Here we report the synthesis of novel 2′-hydroxy- and 2′-hydroxymethyl-apionucleosides, 4 and 5, corresponding triphosphates and phosphoramidate prodrugs. Triphosphate 38 of 2′-hydroxymethyl-apionucleoside 5 exhibited potent inhibition of HBV polymerase with an IC50 value of 120 nM. In an HBV cell-based assay, the phosphoramidate prodrug 39 demonstrated potent activity with an EC50 value of 7.8 nM.


Introduction
Hepatitis B is an infectious liver disease caused by the hepatitis B virus (HBV). An effective and safe vaccination has been available since 1982. However, in 2018, a total of 3322 cases of acute (short-term) hepatitis B were reported to CDC in the USA [1]. Since many people may not have symptoms or don't know that they are infected, their illness is often not diagnosed so it remains unreported or uncounted. The CDC estimates that the actual number of acute hepatitis B cases was closer to 21,600 in 2018. Many more people (about 296 million) are estimated to be living with chronic, long-term hepatitis B worldwide [2]. HBV is an enveloped DNA virus that can cause both acute and chronic forms of the disease. Acute HBV infection is cleared out of the body usually in a few weeks, while chronic HBV infection can lead to permanent liver damage such as cirrhosis and liver cancer. Many nucleoside analogues are effective antiviral agents, with most of them targeting the reverse transcriptase activity of viral polymerases [3]. Six nucleoside derivatives (adefovir dipivoxyl, entecavir 1, lamivudine, telbivudine, tenofovir alafenamide, and tenofovir disoproxil) and two interferons (alfa-2a/b and PEGylated alfa-2a/b) are currently approved for the treatment of hepatitis B in the United States [4]. Long-term administration of reverse transcriptase inhibitors is necessary, ultimately followed by the development of resistance [5]. The development of drug-resistant mutation to Lamivudine within five years is as high as 70%, while entecavir 1 has only a 1.2% of incidence [6]. Drug-resistant mutations to tenofovir were not reported initially, but mutations have been observed more recently [7]. Therefore, it remains important to search for new antiviral nucleoside analogues.
A series of racemic 3 -fluoro-2 ,3 -dideoxyapionucleosides (±)2 and (±)3 were found to inhibit HBV in vitro with EC 50 values in the range of 0.011 µM for 2c and 0.8 µM for both 2d and 3b (Figure 1) [8][9][10]. Asymmetric synthesis of stereoisomers was developed for 2a,c, and 3a,c [11][12][13]. In vitro anti HBV activity of separate enantiomers is reported only for 3a (EC 50 = 0.01 µM and 5.6 µM) [14]. We decided to synthesize and evaluate close The apionucleoside 4 was synthesized as presented in Scheme 2. Secondary alcohol in 11 was benzoylated and the tertiary alcohol of 12 was transformed by nucleophilic fluorination with diethylaminosulfur trifluoride (DAST) to compound 13 in 43% yield [19]. The reversal of stereochemistry at C2 of 14 was achieved by a sequence of Dess-Martin oxidation to ketone 15 and following reduction with sodium borohydride to alcohol 16 [20]. Transformation of methoxy glycoside 17 to corresponding acetoxy glycoside under acidic conditions was plagued by concomitant deprotection of the benzyl group. Therefore, the benzyl group of 17 was exchanged by benzoyl in compound 19. Acetolysis of 19 provided desired acetate 20 in high yield as an inconsequential mixture of two diastereomers. Small amounts of acyclic byproduct 21 were also isolated. Vorbrüggen condensation [21,22] of 20 with N 6 -benzoyladenine provided 22 which after debenzoylation afforded apionucleoside 4. The stereochemistry of the glycosidic bond of 4 was confirmed by the 2D-NOESY experiment ( Figure 2). The observed NOE contacts between H-8 of the base and H-2′ of the sugar in addition to NOE interactions between H-1′ and H-4′ confirmed β-orientation of the base.

Results and Discussion
The synthesis of apionucleosides 4 and 5 started with compound 6, which was prepared in four steps from D-ribose according to known procedures [15][16][17][18]. Formation of methyl glycoside under acidic conditions using p-toluenesulfonic acid provided methoxymethylene protected compound 7 and acetonide 8 in high overall yield (Scheme 1). The anomeric centers in both 7 and 8 formed with exclusive β-stereoselectivity. A mixture of compounds 7 and 8 was benzylated to provide 9 and 10, respectively. Acidic treatment of the mixture resulted in diol 11.
The apionucleoside 4 was synthesized as presented in Scheme 2. Secondary alcohol in 11 was benzoylated and the tertiary alcohol of 12 was transformed by nucleophilic fluorination with diethylaminosulfur trifluoride (DAST) to compound 13 in 43% yield [19]. The reversal of stereochemistry at C2 of 14 was achieved by a sequence of Dess-Martin oxidation to ketone 15 and following reduction with sodium borohydride to alcohol 16 [20]. Transformation of methoxy glycoside 17 to corresponding acetoxy glycoside under acidic conditions was plagued by concomitant deprotection of the benzyl group. Therefore, the benzyl group of 17 was exchanged by benzoyl in compound 19. Acetolysis of 19 provided desired acetate 20 in high yield as an inconsequential mixture of two diastereomers. Small amounts of acyclic byproduct 21 were also isolated. Vorbrüggen condensation [21,22] of 20 with N 6 -benzoyladenine provided 22 which after debenzoylation afforded apionucleoside 4. The stereochemistry of the glycosidic bond of 4 was confirmed by the 2D-NOESY experiment ( Figure 2). The observed NOE contacts between H-8 of the base and H-2′ of the sugar in addition to NOE interactions between H-1′ and H-4′ confirmed β-orientation of the base. Scheme 1. Reagents and conditions: (a) pTsOH·H 2 O, HC(OMe) 3 /MeOH, reflux, 6 h, 83% combined yield of 7 and 8; (b) BnBr, NaH, TBAI, THF, rt, 24 h, 90% combined yield of 9 and 10; (c) i. 80% AcOH, rt, 1 h; ii. NH 4 OH, rt, 30 min, 98%.
The apionucleoside 4 was synthesized as presented in Scheme 2. Secondary alcohol in 11 was benzoylated and the tertiary alcohol of 12 was transformed by nucleophilic fluorination with diethylaminosulfur trifluoride (DAST) to compound 13 in 43% yield [19]. The reversal of stereochemistry at C2 of 14 was achieved by a sequence of Dess-Martin oxidation to ketone 15 and following reduction with sodium borohydride to alcohol 16 [20]. Transformation of methoxy glycoside 17 to corresponding acetoxy glycoside under acidic conditions was plagued by concomitant deprotection of the benzyl group. Therefore, the benzyl group of 17 was exchanged by benzoyl in compound 19. Acetolysis of 19 provided desired acetate 20 in high yield as an inconsequential mixture of two diastereomers. Small amounts of acyclic byproduct 21 were also isolated. Vorbrüggen condensation [21,22] of 20 with N 6 -benzoyladenine provided 22 which after debenzoylation afforded apionucleoside 4. The stereochemistry of the glycosidic bond of 4 was confirmed by the 2D-NOESY experiment ( Figure 2). The observed NOE contacts between H-8 of the base and H-2 of the sugar in addition to NOE interactions between H-1 and H-4 confirmed β-orientation of the base.  The synthesis of 3′-fluoro-2′-hydroxymethyl apionucleoside 5 started with unsuccessful attempts to oxidize the secondary alcohol moiety of 11. Therefore, a different synthetic approach was designed (Scheme 3). Tetrahydropyranyl protection of the secondary alcohol of 11 gave compound 25 in high yield as a mixture of two diastereomers. Protection of the tertiary alcohol with a 4-methoxybenzyl group (26) followed by THP deprotection afforded alcohol 27. The Dess-Martin oxidation of 27 gave ketone 28 in excellent yield. The Wittig olefination of 28 provided alkene 29 in a 69% yield [23]. Removal of the PMB protecting group provided tertiary vinyl alcohol 30 in high yield [24,25]. Subsequent hydroboration of 30 with 9-BBN proceeded in very high yield and with exclusive stereoselectivity providing a single diastereoisomer of 31 [26][27][28]. Primary alcohol in 31 was protected with the 4-methoxybenzoyl group to give 32. Electron-rich 4-methoxybenzoyl group was chosen because, as compared to the unsubstituted benzoyl group, it stabilizes the oxocarbenium ion by a more effective neighboring group participation during the glycosylation step [29]. In the next step, the tertiary alcohol of 32 was converted to fluoride  The synthesis of 3′-fluoro-2′-hydroxymethyl apionucleoside 5 started with unsuccessful attempts to oxidize the secondary alcohol moiety of 11. Therefore, a different synthetic approach was designed (Scheme 3). Tetrahydropyranyl protection of the secondary alcohol of 11 gave compound 25 in high yield as a mixture of two diastereomers. Protection of the tertiary alcohol with a 4-methoxybenzyl group (26) followed by THP deprotection afforded alcohol 27. The Dess-Martin oxidation of 27 gave ketone 28 in excellent yield. The Wittig olefination of 28 provided alkene 29 in a 69% yield [23]. Removal of the PMB protecting group provided tertiary vinyl alcohol 30 in high yield [24,25]. Subsequent hydroboration of 30 with 9-BBN proceeded in very high yield and with exclusive stereoselectivity providing a single diastereoisomer of 31 [26][27][28]. Primary alcohol in 31 was protected with the 4-methoxybenzoyl group to give 32. Electron-rich 4-methoxybenzoyl group was chosen because, as compared to the unsubstituted benzoyl group, it stabilizes the oxocarbenium ion by a more effective neighboring group participation during the glycosylation step [29]. In the next step, the tertiary alcohol of 32 was converted to fluoride The synthesis of 3 -fluoro-2 -hydroxymethyl apionucleoside 5 started with unsuccessful attempts to oxidize the secondary alcohol moiety of 11. Therefore, a different synthetic approach was designed (Scheme 3). Tetrahydropyranyl protection of the secondary alcohol of 11 gave compound 25 in high yield as a mixture of two diastereomers. Protection of the tertiary alcohol with a 4-methoxybenzyl group (26) followed by THP deprotection afforded alcohol 27. The Dess-Martin oxidation of 27 gave ketone 28 in excellent yield. The Wittig olefination of 28 provided alkene 29 in a 69% yield [23]. Removal of the PMB protecting group provided tertiary vinyl alcohol 30 in high yield [24,25]. Subsequent hydroboration of 30 with 9-BBN proceeded in very high yield and with exclusive stereoselectivity providing a single diastereoisomer of 31 [26][27][28]. Primary alcohol in 31 was protected with the 4-methoxybenzoyl group to give 32. Electron-rich 4-methoxybenzoyl group was chosen because, as compared to the unsubstituted benzoyl group, it stabilizes the oxocarbenium ion by a more effective neighboring group participation during the glycosylation step [29].
In the next step, the tertiary alcohol of 32 was converted to fluoride 33 by using DAST. The yield of nucleophilic fluorination was low due to the formation of various elimination byproducts. Following hydrogenolysis of benzyl ether provided primary alcohol 34, which was protected as a benzoyl ester in a straightforward way to give compound 35. Acetolysis of a glycosidic bond in 35 gave acetate 36 as a mixture of two diastereomers together with some amounts of open chain byproduct; this mixture was used without separation in Vorbrüggen condensation with N 6 -benzoyladenine. After chromatography and deprotection, nucleoside 5 was obtained in good yield [30][31][32]. Stereochemistry was confirmed by 2D-NOESY NMR spectroscopy ( Figure 2  33 by using DAST. The yield of nucleophilic fluorination was low due to the formation of various elimination byproducts. Following hydrogenolysis of benzyl ether provided primary alcohol 34, which was protected as a benzoyl ester in a straightforward way to give compound 35. Acetolysis of a glycosidic bond in 35 gave acetate 36 as a mixture of two diastereomers together with some amounts of open chain byproduct; this mixture was used without separation in Vorbrüggen condensation with N 6 -benzoyladenine. After chromatography and deprotection, nucleoside 5 was obtained in good yield [30][31][32]. Stereochemistry was confirmed by 2D-NOESY NMR spectroscopy ( Figure 2) and elemental composition by HRMS. The observed NOE contacts between H-8 of the base and H-4′ of the sugar confirmed the β orientation of the base. A prerequisite for a nucleoside antiviral drug is the cellular activation of its triphosphate. Direct administration of the triphosphate is not possible due to the highly charged nature of phosphate moiety. Cellular bioactivation of nucleosides to triphosphates is often limited by the first phosphorylation step. To bypass this rate-limiting step, monophosphate prodrug technology has been applied [33][34][35]. We synthesized phosphoramidate prodrugs 24 and 39 as well as triphosphate 23 and 38 ( Figure 3).
3′-Fluoroapionucleosides 4 and 5, and their phosphoramidate prodrugs 24 and 39 were tested in cell-based assays against HBV activity. The data are summarized in Table  1. Neither 4 nor its prodrug 24 showed any anti-HBV activity. While 2′-hydroxymethyl substituted nucleoside 5 itself did not show any anti-HBV activity, its prodrug 39 exhibited strong inhibition of HBV in this cell-based assay with EC50 of 7.8 nM. This may be explained by the kinetically difficult first phosphorylation step of 5 in cells [35][36][37]. Prodrug 39 can be considered a masked monophosphate, thus overcoming the rate-limiting first phosphorylation step [21,[33][34][35]. Moreover, prodrug 39 showed anti-HIV activity in a cell-based assay. The corresponding entecavir prodrug (1-PD) inhibited HBV replication with EC50 of 31 nM. None of the tested apionucleosides 4 and 5 or apionucleoside prodrugs 24 and 39 showed any cytotoxicity at their highest tested concentrations (Table 1). A prerequisite for a nucleoside antiviral drug is the cellular activation of its triphosphate. Direct administration of the triphosphate is not possible due to the highly charged nature of phosphate moiety. Cellular bioactivation of nucleosides to triphosphates is often limited by the first phosphorylation step. To bypass this rate-limiting step, monophosphate prodrug technology has been applied [33][34][35]. We synthesized phosphoramidate prodrugs 24 and 39 as well as triphosphate 23 and 38 ( Figure 3).
3 -Fluoroapionucleosides 4 and 5, and their phosphoramidate prodrugs 24 and 39 were tested in cell-based assays against HBV activity. The data are summarized in Table 1. Neither 4 nor its prodrug 24 showed any anti-HBV activity. While 2 -hydroxymethyl substituted nucleoside 5 itself did not show any anti-HBV activity, its prodrug 39 exhibited strong inhibition of HBV in this cell-based assay with EC 50 of 7.8 nM. This may be explained by the kinetically difficult first phosphorylation step of 5 in cells [35][36][37]. Prodrug 39 can be considered a masked monophosphate, thus overcoming the rate-limiting first phosphorylation step [21,[33][34][35]. Moreover, prodrug 39 showed anti-HIV activity in a cell-based assay. The corresponding entecavir prodrug (1-PD) inhibited HBV replication with EC 50 of 31 nM. None of the tested apionucleosides 4 and 5 or apionucleoside prodrugs 24 and 39 showed any cytotoxicity at their highest tested concentrations (Table 1). 3 -Fluoroapionucleoside triphosphates 23 and 38 were tested in HBV polymerase and HIV-1 reverse transcriptase (HIV-1 RT) enzymatic assays. The data in Table 1 shows that 2 -hydroxymethyl substituted triphosphate 38 had an IC 50 of 0.12 µM with HBV polymerase, and it was a much more potent inhibitor of HIV-1 RT than 2 -hydroxy-substituted triphosphate 23 (IC 50 of 0.32 µM vs. 4.2 µM). The triphosphate compound 38 was also tested with human DNA polymerases α, β, and γ, and it had an IC 50 of 80 µM, >100 µM, and >100 µM, respectively, suggesting the 3 -fluoroapionucleoside triphosphate 38 has low or no inhibition activity against human DNA polymerases.  Table 1 shows that 2′-hydroxymethyl substituted triphosphate 38 had an IC50 of 0.12 μM with HBV polymerase, and it was a much more potent inhibitor of HIV-1 RT than 2′-hydroxy-substituted triphosphate 23 (IC50 of 0.32 μM vs 4.2 μM). The triphosphate compound 38 was also tested with human DNA polymerases α, β, and γ, and it had an IC50 of 80 μM, >100 μM, and >100 μM, respectively, suggesting the 3′-fluoroapionucleoside triphosphate 38 has low or no inhibition activity against human DNA polymerases.
Values of EC 50 , CC 50 , and IC 50 are expressed as means ± SDs, followed by numbers of independent experiments in parentheses.

Materials and Methods
All commercially obtained solvents and reagents were used as received. All solvents used for chemical reactions were anhydrous grade unless specifically indicated. Structures of the target compounds in this work were assigned by use of NMR spectroscopy and MS spectrometry. The purities of all non-salt compounds were >95% as determined on an Agilent 1200 HPLC, XTerra 3.5 µm 4.6 × 150 mm MS C18 column, using 0.04% (v/v) TFA in water and 0.02% (v/v) TFA in acetonitrile as mobile phase. The purities of all nucleotides were >95%, determined on an Agilent 1100 HPLC, 50 mM TEAA in water, and 50 mM TEAA in acetonitrile as mobile phase. 1 H-, 19 F-, and 13 C-NMR spectra were recorded on a Bruker Avance III (400 MHz) or a Varian 400MR (400 MHz) NMR spectrometer. Chemical shifts are reported in parts per million (ppm, δ) using the residual solvent line as an internal reference. Splitting patterns are reported as s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), or br s (broad singlet). Coupling constants (J) are reported in hertz (Hz). Mass spectrometric analyses for nucleosides were performed on an Agilent 1200 HPLC with Agilent 6110/6140/1956C MSD mass spectrometer using ESI as ionization, Phenomenex Luna C18 5 µm 5.0 × 20 mm column; mobile phase: 0.1% (v/v) TFA in water and 0.1% (v/v) TFA in acetonitrile, 40 • C, flow rate 0.4 mL/min. Mass spectrometric analyses for nucleotides were performed on an Agilent 1100 HPLC with API 2000 LC-MS/MS System using ESI as ionization, Synergi 75 × 2.0 mm, 4 µm Hydro-RP80Å column, 50 mM TEAA in water, and 50 mM in acetonitrile, flow rate 0.4 mL/min. Work-up procedures for most of the chemical reactions are the same or similar, therefore, unless specifically indicated the work-up refers to the following procedure: the reaction mixture at 0 • C is quenched with water, diluted with EtOAc or dichloromethane, washed with 5% sodium bicarbonate and then with brine, dried over anhydrous sodium sulfate, filtered, and concentrated to dryness. Purification on silica gel refers to flash chromatography on a silica gel column. . After refluxing for 6 h the reaction mixture was stirred at rt overnight. Solid NaHCO 3 (700 mg, 8.3 mmol) was added in portions until the mixture had neutral pH. The mixture was filtered, the filtrate was evaporated, and the residue was purified by column chromatography, 0-20% EtOAc in hexanes, to give colorless oil of 7 and 8 as a 2.5:1 mixture (13.8 g, 83%). TBAI (284 mg, 0.77 mmol) was added to a solution of 7 and 8 (15.8 g, 77 mmol) in anhydrous THF (160 mL). NaH (4.6 g, 115 mmol, 60% dispersion in mineral oil) was added in small portions and stirred at rt for 30 min, followed by BnBr (9 mL, 92 mmol). The reaction mixture was stirred at rt for 24 h. Florisil (6 g) was added, and the solvent was evaporated. The residue was dispersed in hexanes and insoluble material was filtered off and washed with additional quantities of hexanes. The combined filtrate was purified by column chromatography (hexane/EtOAc 1:3) to give 20.6 g (90% combined) of a mixture of 9 and 10 (46 and 22 mmol, respectively) as a colorless oil. The obtained mixture was dissolved in 80% aqueous acetic acid (50 mL), stirred for 1 h, and then neutralized with ammonium hydroxide (28-30%). After 30 min the solvents were evaporated. The residue was suspended in CH 2 Cl 2 , filtered, and the filtrate evaporated. The crude residue was purified by column chromatography with 25-80% EtOAc in hexanes to give 5.7 g of recovered 10 and 11.5 g of 11 (98%) as a colorless oil.  (21). Concentrated H 2 SO 4 (44 µL) was added to a solution of 19 (614 mg, 1.6 mmol) in acetic acid (40 mL) and Ac 2 O (20 mL). After stirring for 3 h at 0 • C, the reaction mixture was poured into saturated NaHCO 3 solution and stirred for 15 min. Usual work-up and purification on silica gel with 10-75% EtOAc in hexanes afforded 578 mg (87%) of 20 as an inseparable mixture of two diastereomers in a 1.5:1 ratio as colorless oil, followed by 93 mg (12%) of 21 as an inseparable mixture of two diastereomers in a 1:1 ratio as colorless oils. Complete stereochemistry of individual diastereomers of 20 and 21 was not assigned.

2S,4S)-4-((Benzyloxy)methyl)-2-methoxy-4-((4-methoxybenzyl)oxy)-3-methylenetetrahydrofuran (29)
. KHMDS (87 mL, 44 mmol, 0.5 M in toluene) was added to a solution of PPh 3 CH 3 I (21 g, 52 mmol) in toluene (50 mL). The mixture was stirred at 80 • C for 30 min, during which yellow colored solution ensued. It was transferred to a solution of 28 (6.4 g, 17 mmol) in toluene (50 mL). The reaction mixture was stirred at rt for 2 h, quenched with water and diluted with diethyl ether. The aqueous layer was separated and washed with diethyl ether (3×), combined organic fractions were washed with brine and dried with anhydrous Na 2 SO 4 . The solvents were evaporated, and the crude product was purified by column chromatography with 2-50% of EtOAc in hexanes to give 4. in THF (10 mL) at rt under argon. After stirring for 24 h at 40 • C in a sealed vessel, NaBO 3 ·4H 2 O (5.2 g, 34 mmol), EtOH (50 mL), and water (50 mL) were added, and the reaction mixture was stirred at 50 • C for 1.5 h. Another portion of NaBO 3 ·4H 2 O (5 g, 32 mmol) was added and the reaction mixture was stirred at 50 • C for an additional 1.5 h, then acidified with acetic acid to pH = 6 and evaporated. The crude residue was dissolved in CH 2 Cl 2 , water was added, solids were filtered off, and the layers were separated. The aqueous layer was subsequently extracted with CH 2 Cl 2 (2×). The combined organic extract was washed with brine and dried with anhydrous Na 2 SO 4 . Evaporated residue was purified on silica gel with 2-8% MeOH in CH 2 Cl 2 to afford 1.9 g (96%) of 31 as a colorless oil. 1  ((3S,4S,5R)-5-(6-Amino-9H-purin-9-yl)-3-fluoro-4-(hydroxymethyl)tetrahydrofuran-3yl)methyl triphosphate (38). 4-Methoxytriphenylmethyl chloride (47 mg, 0.15 mmol) was added to a solution of nucleoside 5 (39 mg, 0.14 mmol) in pyridine (3 mL) and the reaction mixture was stirred overnight at rt, an additional amount (5 mg, 0.02 mmol) of MMTrCl was added, and the reaction mixture was kept at rt for 6 more hours. The reaction was quenched with methanol, solvents were evaporated, and the crude mixture was purified by column chromatography with 0-12% in CH 2 Cl 2 to give 66 mg (85%) of protected intermediate which was treated as described for the synthesis of 23. The fractions containing triphosphate were concentrated to a volume of~1 mL and treated with 80% formic acid to remove the 4-methoxytriphenylmethyl protecting group. Triphosphate 38 was isolated as
The antiviral activity was measured using a Real-Time quantitative polymerase chain reaction (RT qPCR) assay directly measuring the HBV viral DNA copy numbers from the supernatant of HepG2.117 cells. The HBV Core primers and probes used in qPCR: the core forward primer was 5 -CTGTGCCTTGGGTGGCTTT-3 ; the core reverse primer was 5 -AAG-GAAAGAAGTC AGAAGGCAAAA-3 ; the core probe was 5 /FAM/AGCTCCAAA/ZEN/ TCCTTTATAAGGGTCGA TGTCCATG/3IBFQ/-3 . The RT qPCR was run for 20 min at 95 • C and 20 min at 60 • C for each cycle, 40 cycles in total. HBV viral DNA copy numbers were normalized to the level observed in the absence of an inhibitor, which was defined as 100%. EC 50 was defined as the concentration of compound at which the HBV viral DNA copy numbers from the HepG2.117 cells were reduced by 50% relative to its level in the absence of the compound.
In parallel, cell cytotoxicity (CC 50 ) against HepG2 cells was measured using a luminescent cell viability assay to determine the number of viable cells in the culture based on quantitation of the adenosine triphosphate (ATP) present after a 4-day incubation period. On the first day, HepG2 cells were seeded at 15,000 cells per 100 µL per well with assay media containing DEME (Corning, cat#10-013-CV), 3% FBS (Coning, cat#35-011-CV), 1× penicillin/streptomycin and 1× Non-essential Amino Acid in Biocoat collagen-coated 96-well flat bottom plates. Cells were incubated in a 37 • C, 5% CO 2 incubator for 4 h before compound dosing. The compound dilution and dosing procedure were identical to that for the determination of anti-HBV activity. After 96 h incubation, CellTiter-Glo ® reagent (Promega, Madison, WI, USA) was added to each well and incubated for 10 min at RT. Luminescence was measured on a Victor X3 multi-label plate reader. Cell viability is normalized to the level observed in the absence of an inhibitor, which was defined as 100%.

HIV Single Cycle Infection Assay
Twenty-four hours prior to infection, CEM human T lymphoblast cells (ATCC, Manassas, VA, USA) were plated in assay media (MEM supplemented with 10% FBS, 1% penicillin/streptomycin (all Mediatech, Manassas, VA, USA) and 1% DMSO (Sigma-Aldrich, St Louis, MO, USA)) at a density of 5 × 10 5 cells/mL (5 × 10 4 cells/well) in white 96-well plates. Serially diluted compounds were added to cells and incubated overnight in a 37 • C, 5% CO 2 incubator. The following day, cells were infected with VSV-G pseudotyped HIV NL4-3, in which parts of the env and nef genes were replaced with Renilla-luciferase. Infected cells were incubated for 72 h in a 37 • C, 5% CO 2 incubator. The viral inoculum was titrated to achieve a Renilla-luciferase signal of approximately 100-fold over the background. Antiviral activity of compounds was measured by the addition of 100 µL of Renilla-Glo ® reagent (Promega, Madison, WI, USA) to infected cells. After a 10-min incubation at room temperature, luminescence was measured on a Victor X3 multi-label plate reader (Perkin Elmer, Waltham, MA, USA). Cytotoxicity of compounds to uninfected parallel cell cultures was determined by the addition of 100 µL CellTiter-Glo ® reagent (Promega, Madison, WI, USA), and incubation for 10 min at room temperature. Luminescence was measured on a Victor X3 multi-label plate reader.

HIV-1 Reverse Transcriptase and HBV Polymerase Activity Assays
HIV-1 reverse transcriptase (HIV-1 RT) was purchased from Abcam (cat#ab63979). Recombinant HBV polymerase (Hepatitis B virus genotype D subtype ayw, full length) was cloned using the baculovirus system, expressed in sf9 cells, and purified with similar strategy and methods described by Lanford et al. [39]. A DNA primer (5 -CCGAGTAGTGTTGG-3 ) was synthesized by IDTDNA and a 358-nt RNA template was synthesized in-house using Megascript T7 transcription kit (ThermoFisher, cat#AM1334). dNTPs were purchased from Thermo Fisher and 3 H-dTTP from Perkin Elmer. Filter plates were purchased from Millipore (cat#MABN0V050) and microscint-20 was purchased from Perkin Elmer (cat#6013621).
The RNA-dependent DNA polymerization (RdDp) activity of HIV-1 RT was measured by the incorporation of radioactively labeled nucleotides by HIV-1 RT into acid-insoluble DNA products from the DNA primer primed RNA template. To test compound inhibition, the reactions were performed at 30 • C for 40 min in a reaction mixture containing reaction buffer (50 mM Tris-Cl, pH 7.5, 100 mM KCl, 12.5 mM MgCl 2 , 4 mM DTT), 1 nM HIV-1 RT, 0.1 µM DNA primer, 0.02 µM RNA template, 10% DMSO, 0.1 µM dATP, 1 µM dGTP, 0.1 µM dCTP, 0.32 µM 3 H-dTTP, and compounds at various concentrations. The 50 µL reactions were performed in 96-well plates. The reactions were quenched with a 60 µL cold mixture of 20% (w/v) TCA and 0.5 mM ATP and incubated at 4 • C for 1 h. The reactions were loaded onto 96-well filter plates. The filters on plates were washed three times with 10% TCA and once with 70% ethanol on a Millipore plate wash station with vacuum applied. The filters on the plate were air-dried and 40 µL Microscint-20 was added to each well. The acid-precipitated tritiated DNA products retained on the filters were detected by a Trilux MicroBeta scintillation reader (Perkin Elmer).
All data were analyzed with GraphPad Prism. The compound concentration at which the enzyme-catalyzed rate was reduced by 50% (IC 50 ) was calculated by fitting the data to the equation Y = % Min + (% Max − % Min) / (1 + 10ˆ((logIC 50 -X)*h)), where Y corresponds to the percent inhibition to the enzyme activity, % Min is the residual inhibition activity without compound, % Max is the maximum inhibition of enzyme activity at saturating compound concentration, and X corresponds to the log of compound concentrations, and h is the hillslope.

Conclusions
We have synthesized two novel 3 -fluoro-2 -substituted apionucleoside analogues with adenine nucleobase. Phosphoramidate prodrug of 2 -hydroxymethyl analogue 39 showed promising inhibition against HBV with an EC 50 of 7.8 nM. None of the synthesized nucleoside analogues and their respective prodrugs exhibited any cytotoxicity at their highest tested concentrations. 3 -Fluoroapionucleoside triphosphate 38 had an IC 50 of 0.12 µM against HBV polymerase, and it was also a potent inhibitor of HIV-1 RT. Triphosphate 38 demonstrated low or no activity against human DNA polymerases.