A Chemical Strategy for Intracellular Arming of an Endogenous Broad-Spectrum Antiviral Nucleotide

Abstract

The naturally occurring nucleotide, 3’-deoxy-3’,4’-didehydro-cytidine-5’-triphosphate (ddhCTP), was recently found to exert potent and broad-spectrum antiviral activity. However, nucleoside 5’-triphosphates in general are not cell permeable, which precludes the direct use of ddhCTP as a therapeutic. To harness the therapeutic potential of this endogenous antiviral nucleotide, we synthesized phosphoramidate prodrug HLB-0532247 (1), and found it to result in dramatically elevated levels of ddhCTP in cells. We compared 1 and 3’-deoxy-3’,4’-didehydro-cytidine (ddhC) and found that 1 more effectively reduces titers of Zika and West Nile viruses in cell culture with minimal non-specific toxicity to host cells. We conclude that 1 is a promising antiviral agent based on a novel strategy of facilitating elevated levels of the endogenous ddhCTP antiviral nucleotide.

Graphical Abstract

Introduction:

Nucleoside analogues targeting viral polymerases are critically important therapeutics with proven efficacy against a spectrum of viruses.¹ Emerging viral threats such as SARS-CoV-2, West Nile (WNV), and Zika (ZIKV) are reminders that new therapeutics are needed. 3’-Deoxy-3’,4’-didehydro-cytidine-5’-triphosphate (ddhCTP, Fig. 1) is a recently discovered endogenous antiviral nucleotide.² ddhCTP inhibits the polymerases of dengue (DENV), WNV, ZIKV, hepatitis C (HCV), and polio (PV, weak inhibition in competition with ribonucleoside 5’-triphosphates),² as well as SARS-CoV-2,³ via chain termination due to the lack of a 3’-alcohol. However, cellular studies with 3’-deoxy-3’,4’-didehydro-cytidine (ddhC, Fig. 1) revealed only weak ZIKV inhibitory activity.² Harnessing the antiviral activity of ddhCTP through the use of a cell-permeable nucleoside 5’-monophosphate prodrug that can be metabolically activated to ddhCTP may confer a new broad-spectrum antiviral agent.

Figure 1.

ddhCTP, ddhC, and HLB-0532247 (1).

Phosphoramidate prodrugs of nucleotides, including the ProTide class of modifications,^4–7 bypass the often rate-limiting 5’-monophosphorylation that occurs during metabolism of nucleoside drugs into their bioactive 5’-triphosphates.^8–10 By bypassing this process, ProTides increase the efficacy of antiviral nucleosides^11–15 and has resulted in the discoveries of sofosbuvir¹⁰ and tenofovir alafenamide.^16–17 In order to arm the cell with its endogenous antiviral defense molecule, ddhCTP, we designed and synthesized HLB-0532247 (1), which is ddhC bearing the 5’-phosphoramidate moiety found in the blockbuster hepatitis C drug sofosbuvir (Fig. 1). ¹⁰ We show that 1 potently inhibits WNV and ZIKV in cell culture through intracellular production of ddhCTP to levels that surpass what cells or exogenous ddhC dosing can produce. Moreover, we show that 1 is non-toxic to host cells and is not used as a substrate by mitochondrial DNA polymerase γ, which is a known off-target of antiviral nucleosides/nucleotides.^18–21

Results and Discussion:

The synthesis of HLB-0532247 (1) was completed using established methods for the preparation of nucleosides containing dehydrated ribose rings.^22–25 The synthesis of 1 commences with a two-step protection sequence to yield bis-protected cytidine 3,^26–27 which can undergo an oxidative esterification yielding 4 in 78% yield (Scheme 1).^28–29 Base-mediated tandem elimination and deprotection of 4, followed by re-protection of the 2’-alcohol with TBSCl yields 5 in 77% yield in a one-pot procedure.^22–23 Reduction of the carboxylic ester to the primary alcohol yields 7. Nucleoside 7 was then converted to phosphoramidate 8 in 93% yield as a single P-diastereomer, which was assigned as S based on analogy to literature precedence in which in-line displacement of the pentafluorophenol by allylic alcohol 7 inverts the phosphorous stereochemistry, yet retains the S-configuration due to a change in priority.³⁰ Deprotection of the silyl group required acidic conditions to yield 9, as use of traditional fluoride deprotection strategies led to removal of the phosphoramidate yielding the undesired product N4-benzoyl-ddhC. N4-benzoyl amide 9 was deprotected using buffered hydrazine³¹ to produce 1 in high yield. ddhC was synthesized as previously described.³²

Scheme 1.

Synthesis of HLB-0532247 (1) from cytidine. PDC (pyridinium dichromate), DMP (2,2-dimethoxypropane).

The activity of 1 against virus-infected cells was investigated to determine whether the phosphoramidate prodrug offered significant potency improvements over ddhC. It has been shown previously that ZIKV levels in infected Vero cells were reduced by 50-200x upon exposure to 1 mM ddhC over 24 h.² To evaluate 1 similarly, Vero, HUH7 and HUH7.5 cell lines were treated with 1 and then infected with ZIKV (strain PRABC59) at a multiplicity of infection (MOI) of 0.1. Virus levels were quantified by plaque titration 4 days post-infection. 1 proved to be significantly more potent than ddhC against ZIKV, resulting in >1 log₁₀ pfu/mL reduction in virus levels at 0.1 mM doses in both HUH7 and HUH7.5 cell lines, with weaker activity observed in Vero cells (Fig. 2A). To directly compare the antiviral activity of 1 to ddhC against a different virus, Vero, HUH7, and HUH7.5 cells were treated with 1 or ddhC and then infecte with WNV at a MOI of 0.1. At 3 days post-infection we found ddhC to be ineffective against WNV across this panel of cell lines, in stark contrast to its published efficacy against ZIKV.² Excitingly, 1 reduced WNV levels in HUH7 and HUH7.5 cells by > 4 1og₁₀ pfu/mL at 1.0 mM, but was inactive in Vero cells. These data demonstrate that introduction of the phosphoramidate moiety significantly promotes antiviral effects. Of additional note, we synthesized 1 as a mixture of R and S phosphorous diastereomers and observed no difference in antiviral activity in comparison to 1 (single diastereomer) in preliminary studies (data not shown).

Figure 2.

A. Effect of 1 on ZIKV virus titer in a panel of cell lines. B. Effect of 1 and ddhC on WNV virus titer in a panel of cell lines. ddhC results shown in red; 1 results shown in blue. The cell line tested is indicated in each column. * p<0.05, **p<0.01, *** p<0.001, **** p<0.0001 (see Experimental Section for additional details).

We next evaluated whether 1 or ddhC is cytotoxic, which could obfuscate the observed antiviral activity. Many clinically approved viral polymerase and reverse transcriptase inhibitors are toxic with chronic dosing due to off-target effects, often the result of human polymerase inhibition.^20–21 We found that neither ddhC nor 1 were cytotoxic to Vero or HUH7 cells after 2 days in culture (Fig. S1). Extended assays to 5 days in culture revealed only modest toxicity (>60% cell viability at a 1.0 mM dose of 1; Fig. S1). In addition, we found that ddhCTP is not a substrate for DNA polymerase γ (Fig. 3A), which is the enzyme often associated with chronic toxicity due to incorporation of nucleotide analogues into mitochondrial DNA.¹⁸–²¹ These results suggest that 1 is devoid of common toxicity liabilities known for this class of compounds. The 50% cytotoxic concentration (CC₅₀) of 1 is greater than 1 mM, surpassing that of many other potent antiviral agents including sofosbuvir and ribavirin, giving it an advantage in future development efforts.^33–36

Figure 3.

DNA polymerase γ (catalytic domain) was allowed to react with a FAM-labeled DNA primer (A) and a cytidine triphosphate analogue (CTP, 2’-dCTP, ddhCTP). Reactions were conducted for 0, 10, or 20 min. The presence of starting DNA (lower band) and addition products (n+1 nt) were determined by gel electrophoresis analysis. B. Metabolism of 1 and ddhC to ddhCTP in cells. Levels of analytes (ddhCTP, black; 1, blue; ddhC, red) following treatment of HUH7 cells with 1 or ddhC. Background ddhCTP levels were < 0.01 nmols per 5x10⁶ HUH7 cells or below the limit of quantitation in each DMSO-only treatment replicate (data not shown). Values are mean ± S.D (n = 3).

Prodrug 1 was next subjected to stability assays to gauge its suitability for systemic use and cell permeability assays (Fig. S2). In the presence of human plasma, 1 was stable up to two hours (> 87% remaining). In simulated gastric fluid (SGF), approximately 76% of 1 remained after 6 h (Fig. S2B). Taken together, these stability properties of 1 are suitable for future in vivo studies. Caco-2 permeability studies were also conducted, revealing that 1 has similar cell permeability properties to that reported for sofosbuvir in a related PAMPA assay (mean P_app 0.3 nm/sec for 1 compared to 0.46 nm/sec for sofosbuvir; Figure S2C).¹⁰

We next investigated the conversion of 1 to ddhCTP in cells. Since we utilized the same phosphoramidate moiety (ProTide) in 1 as that found in sofosbuvir, we hypothesized that 1 likely utilizes the same enzymes for metabolic activation as sofosbuvir. Consequently, we first established that HUH7 cells express the phosphoramidase enzyme HINT1 (histidine triad nucleotide-binding protein 1), which is responsible for cleaving P-N bonds found in phosphoramidate prodrugs such as sofosbuvir.^37,11 Vero cells also express HINT1, but at lower levels (Fig. S3). Carboxyesterase 1 (CES1) also is implicated in the metabolism of sofosbuvir¹¹ and its expression levels in HUH7 cells are known to be high.³⁸ These data may suggest why 1 is more active in HUH7 versus Vero cells, as also observed for sofosbuvir.³⁴ Second, we measured ddhCTP production in cells. DMSO, ddhC, and 1 were dosed to HUH7 cells at 100 μM for 24 hours and then the cell lysates were evaluated by mass spectrometry. In DMSO-control-treated cells, basal levels of ddhCTP were found (< 0.01 nmols per 5x10⁶ cells or below the limit of quantitation). Both 1 and ddhC were detected in HUH7 lysate 24 h following dosage, indicating that both compounds are cell permeable (Fig. 3B). However, the amount of the ddhCTP metabolite detected in the treated cells were vastly different: HUH7 cells treated with 1 produced 3.8 nmol ddhCTP per 5-million cells, whereas cells treated with ddhC produced 0.13 nmol ddhCTP per 5-million cells. Consequently, exposing cells to phosphoramidate 1 results in a 29-fold increase in ddhCTP compared with treatment of ddhC. If the HUH7 cell volume is approximated, the intracellular concentration of ddhCTP produced by exposure to 1 is greater than 1 mM (see experimental section for details).³⁹ This concentration far surpasses those arising from other methods of ddhCTP production, namely, HEK293 cells exposed to ddhC produced ~100 μM ddhCTP after 48 h, and macrophages (RAW 264.7) induced with IFNα produced 350 μM ddhCTP after 19 h.² Our results with 1 are similar to the concentrations of sofosbuvir-triphosphate found in animal models and cell studies.¹⁰, ³⁴ These data support the hypothesis that 1 undergoes a sequence of metabolic activations to the 5’-triphosphate in a manner analogous to sofosbuvir,¹¹ and that ddhCTP is the source of the observed antiviral activity of 1.

ddhCTP has been shown to inhibit a variety of viral polymerases including those of DENV, WNV, ZIKV, HCV and SARS-CoV-2.^2–3 ddhCTP inhibits the polymerases by chain termination, preventing the synthesis of viral RNA beyond its own incorporation. To further study these processes, we report a chemical synthesis of ddhCTP that is free of contaminating CTP, which is a limitation of enzymatic preparations (Scheme S1).² Of note, another method for synthesizing ddhCTP was reported recently.⁴⁰ Using synthetic ddhCTP (Scheme S1) and ZIKV RdRp in a reported primer extension assay (Fig. 4A & S4),⁴¹ we confirm that ddhCTP acts as expected for an obligate chain terminator. Building on these results, we show that ddhCTP potently inhibits ZIKV RdRp in a competitive fashion against 10 μM CTP (Fig. 4). Processing of ddhCTP inhibited production of full-length RNA in a dose-dependent manner (IC₅₀ 320 ± 10 μM) (Figs. 4B,C).

Figure 4.

Synthetic ddhCTP is utilized as a substrate by ZIKV NS5 RdRp and chain terminates RNA synthesis. A. Schematic of primer extension assay for evaluating chain-terminating activity in the presence of CTP. ddhCTP effectively competes for incorporation with CTP resulting in the accumulation of the chain terminated n+1 product B. ZIKV RdRp-catalyzed nucleotide incorporation with increasing concentrations of ddhCTP (0, 1, 5, 10, 50, 100, 500, 1000, 2500, and 5000 μM) in the presence of 10 μM CTP. Lane 1 has no CTP or UTP. Lane 2 has CTP and UTP. Lanes 3-11 have CTP, UTP, and increasing amounts of competitor ddhCTP. C. Representative plot of the percentage inhibition as a function of ddhCTP concentration and mean IC₅₀ (with standard error) from the individual curve fit data. All experiments were repeated independently 3x with similar results.

Conclusion:

In conclusion, we report the development of HLB-0532247 (1), which facilitates dramatically enhanced levels of ddhCTP in cells, yielding potent antiviral effects. Prodrug 1 is non-cytotoxic to host cells and primer-extension assays with mitochondrial DNA polymerase γ, an off-target for nucleotide-based drugs, show that ddhCTP is not a substrate. Taken together, 1 is a promising antiviral agent that merits further development Moreover, our strategy of using ddhC prodrugs that are metabolically activated to ddhCTP offers opportunities to tailor physiochemical and metabolic activation properties for various therapeutic applications.

EXPERIMENTAL SECTION

General.

Reactions were performed in flame-dried glassware under inert gas (N₂ or Ar) and stirred using a Teflon-coated magnetic stir bar. Reaction solvents tetrahydrofuran (THF) and dichloromethane (DCM) were dried by passage over a column of activated alumina and dimethylformamide (DMF) over crushed molecular sieves using a solvent purification system (MBraun). Other solvents were purchased as ACS grade or higher and used as received unless otherwise noted. TMSCl was purified by short path distillation.⁴² Silica gel chromatography was accomplished using a Teledyne-Isco Combiflash Rf-200 instrument using Redisep Rf High Performance silica gel columns from Teledyne-Isco. ¹H NMR (500 MHz), ¹³C NMR (125 MHz), and ³¹P NMR (202 MHz) were collected on a Bruker Advance NMR spectrometer at room temperature. NMR chemical shifts (δ) are recorded relative to solvent signal for ¹H NMR (δ = 7.26 for CDCl₃, δ = 4.79 for D₂O, δ = 3.31 for MeOD, and δ = 2.50 for (CD₃)₂SO) and the solvent signal for ¹³C NMR (δ = 77.0 for CDCl₃, δ = 49.0 for MeOD, and δ = 39.7 for (CD₃)₂SO). High resolution mass spectra were obtained at the Analytical Biochemistry Core Facility at the University of Minnesota Masonic Cancer Center using an LTP Orbitrap Velos Mass Spectrometer (Thermo Fisher). Compound purities of synthesized molecules were determined by analytical HPLC analysis using an Agilent 1200 series instrument with a diode array detector monitoring 215 or 254 nm. A Zorbax SB-C18 column (4.6 mm x 150 mm x 5.0 μm, Agilent Technologies) was used. A two solvent system was used as the eluents: solvent A = distilled and deionized H₂O (containing 0.1% TFA) and solvent B: MeCN (containing 0.1% TFA). The gradient (1 mL / min flow rate) consisted of: A:B 90:10 from 0-2 minutes, followed by a linear gradient to A:B 20:80 from 2-24 minutes, a second gradient to A:B 5:95 from 24-26 minutes, and an isocratic A:B 5:95 from 26-30 minutes. Purity of ddhCTP was determined by analytical HPLC analysis using an Agilent 1200 series instrument with a diode array detector monitoring 215 nm, or 260 nm. A Zorbax SB-C18 column (4.6 mm x 150 mm x 5.0 μm) was used. A two solvent system was used as the eluents: solvent A = 100 mM aqueous KH₂PO₄ (pH 6) and solvent B = MeCN. The gradient (1.0 mL / min flow rate) consisted of: A:B 99:1 from 0-5 minutes, followed by a linear gradient to 20:80 from 5 to 20 minutes.

Nucleosides and nucleoside prodrugs tested in biological assays were >95% (254 nm) and >90% (215 nm) pure by HPLC. Prodrug 1 (3 batches) was 95-99% (254 nm) and 96-98% (215 nm) pure; ddhC³² (3 batches) was 95-99% (254 nm) and 90-99% (215 nm) pure; and synthesized ddhCTP was >99% (270 nm) and 94% (215 nm) pure by HPLC (see Supporting Information).

Abbreviations:

acetic anhydride (Ac₂O), acetonitrile (MeCN), dichloromethane (DCM), dimethyl formamide (DMF), ethyl acetate (EtOAc), methanol (MeOH), potassium tert-butoxide (t-BuOK), pyridinium dichromate (PDC), tert-butyl alcohol (t-BuOH), tert-butyldimethylsilyl chloride (TBSCl), tetrahydrofuran (THF), triethylamine (TEA or Et₃N), triethylammonium acetate (TEAA), trifluoroacetic acid (TFA or CF₃CO₂H), trimethylsilyl chloride (TMSCl), Tri(tetrabutylammonium)pyrophosphate (TBAPP).

4-N-Benzoyl-2’,3’-O-isopropylidenecytidine (3).

3 was synthesized as previously described (see Scheme 1).²⁶,²⁷

4-N-Benzoyl-2’,3’-O-isopropylidenecytidine-5’-tert-butylcarboxylate (4).

4 was synthesized in a manner similar to a previous report.²⁸ 3 (1.92 g, 4.96 mmol, 1.0 eq) was suspended in DCM (20 mL). To this solution was added pyridinium dichromate (PDC, 3.73 g, 9.91 mmol, 2.0 eq), acetic anhydride (Ac₂O, 4.69 mL, 4.96 mmol, 10.0 eq), and tert-butyl alcohol (t-BuOH, 9.48 mL, 99.1 mmol, 20.0 eq) sequentially and stirred for 3 hours. The reaction was diluted in an excess of EtOAc (400 mL), filtered through a pad of silica gel, rinsed with excess EtOAc, and the solution concentrated in vacuo. The residue was purified by chromatography on silica gel using a gradient of 0-100% EtOAc in hexanes to yield 4 as a light-yellow foam (1.77 g, 3.87 mmol, 78% yield). ¹H NMR (500 MHz, CDCl₃): δ 9.08 (s, 1H), 8.01 (d, J = 7.4 Hz, 1H), 7.86 (d, J = 7.3 Hz, 2H), 7.56 (app t, J = 7.5 Hz, 1H), 7.51 – 7.37 (m, 3H), 5.71 (s, 1H), 5.28 (dd, J = 6.1, 2.1 Hz, 1H), 5.16 (d, J = 6.1 Hz, 1H), 4.61 (d, J = 2.0 Hz, 1H), 1.53 (s, 3H), 1.46 (s, 9H), 1.35 (s, 3H) ppm. ¹³C NMR (126 MHz, CDCl₃): δ 168.3, 166.6, 162.8, 154.6, 148.0, 133.0, 128.8, 127.5, 113.2, 98.4, 96.6, 88.3, 84.6, 84.4, 82.4, 27.9, 26.5, 25.0 ppm. One carbon signal is merged into another; possible merged peak is visible in the ¹³C NMR of 3, 5, 7, and 9 at approximately 132.9 ppm. HRMS-ESI⁺ (m/z) calc’d [M + H]⁺ for C₂₃H₂₈N₃O₇: 458.1922, found: 458.1903.

4-N-Benzoyl-3’-deoxy-3’,4’-didehydro-2’-O-tert-butyldimethylsilylcytidine-5’-tert-butylcarboxylate (5).

5 was synthesized in a manner similar to a previous report.²³ 4 (6.55 g, 14.3 mmol, 1.0 eq) was dissolved in t-BuOH (90 mL) and to this solution was added t-BuOK (3.37 g, 30.1 mmol, 2.0 eq). The reaction was stirred until the starting material was observed to be consumed by TLC (approximately 5 minutes). The reaction was neutralized to approximately pH 7 with glacial acetic acid and then concentrated in vacuo. The reaction was dried on a high vacuum for at least 1 hour before proceeding. The residue was then dissolved in DMF (50 mL) and to this solution was added tert-butyldimethylsilyl chloride (TBSCl, 8.62 g, 57.2 mmol, 2.0 eq) and imidazole (3.89 g, 57.2 mmol, 2.0 eq) and the reaction stirred for 30 minutes. The reaction was monitored by TLC to ensure reaction completion (if the reaction was incomplete this indicates residual t-BuOH is still present and excess TBSCl should be added until all starting material is converted). The reaction was then directly poured into a separatory funnel and diluted in a 1:1 mixture of EtOAc and hexanes (800 mL), and washed with 1M aqueous HCl (300 mL), saturated aqueous NaHCO₃ (300 mL), and brine (300 mL). The organic layer was dried over Na₂SO₄, filtered, and concentrated in vacuo. The residue was purified by chromatography on silica gel using a gradient of 0-60% EtOAc in hexanes to yield 5 as a white foam (5.68 g, 11.1 mmol, 77% yield). ¹H NMR (500 MHz, CDCl₃): δ 8.83 (br s, 1H), 7.89 (d, J = 7.7 Hz, 2H), 7.59 (app t, J = 7.4 Hz, 1H), 7.56 – 7.46 (m, 4H), 6.28 (d, J = 2.2 Hz, 1H), 5.94 (d, J = 2.6 Hz, 1H), 5.03 (app t, J = 2.4 Hz, 1H), 1.55 (s, 9H), 0.89 (s, 9H), 0.16 (s, 3H), 0.10 (s, 3H) ppm. ¹³C NMR (126 MHz, CDCl₃): δ 166.8, 162.6, 158.1, 154.0, 150.6, 143.6, 133.0, 132.8, 128.7, 127.6, 111.7, 97.0, 95.5, 83.2, 79.8, 27.8, 25.5, 17.9, −4.8, −5.0 ppm. HRMS-ESI⁺ (m/z) calc’d [M + H]⁺ for C₂₆H₃₆N₃O₆Si: 514.2368, found: 514.2341.

4-N-Benzoyl-3’-deoxy-3’,4’-didehydro-2’-O-tert-butyldimethylsilylcytidine (7).

5 (700.0 mg, 1.36 mmol, 1.0 eq) was dissolved in DCM (9 mL). To this was added NaOH (3M solution in MeOH, 908 μL; the final solvent ratio is approximately 9:1 DCM:MeOH) and the solution stirred until the mixture solidifies into a gel-like solid (< 1 minute to 30 minutes). The crude mixture was transferred to a separatory funnel, washed with 1M aqueous HCl (300 mL), extracted with EtOAc (600 mL; mix in separatory funnel until the solid solubilizes), and the organic layer was dried over Na₂SO₄ and concentrated in vacuo. Crude 4-N-benzoyl-3’-deoxy-3’,4’-didehydro-2’-O-tert-butyldimethylsilylcytidine-5’-carboxylate (6, 537 mg, 1.21 mmol, 79% yield) is used without further purification. ¹H NMR (500 MHz, MeOD): δ 7.98 (d, J = 7.2 Hz, 2H), 7.87 (d, J = 7.5 Hz, 1H), 7.64 (app t, J = 7.4 Hz, 1H), 7.61 (d, J = 7.4 Hz, 1H), 7.54 (app t, J = 7.8 Hz, 2H), 6.32 (d, J = 2.4 Hz, 1H), 6.11 (d, J = 2.6 Hz, 1H), 5.20 (app t, J = 2.6 Hz, 1H), 0.94 (s, 9H), 0.18 (s, 3H), 0.15 (s, 3H) ppm. ¹³C NMR (126 MHz, MeOD): δ 129.2, 113.7, 99.0, 97.1, 81.5, 26.2, 18.9, −4.6, −4.7 ppm. HRMS-ESI⁺ (m/z) calc’d [M + H]⁺ for C₂₂H₂₈N₃O₆Si: 458.1742, found: 458.1726.

Crude 6 was dissolved in THF (30 mL), cooled to 0°C, and to this solution was added ethyl chloroformate (393 μL, 4.11 mmol, 3.6 eq), and triethylamine (261 μL, 1.87 mmol, eq). After reagent addition the reaction was allowed to warm back to room temperature and stirred for 2 hours. After 2 hours, the mixture was cooled back to 0°C. To the solution was added NaBH₄ (88.5 mg, 2.34 mmol, 2 eq) and then water (15 mL) was added dropwise over 30 minutes to fully dissolve all NaBH₄ and then the reaction was stirred for an addition 1 hour at 0°C. The solution was then concentrated in vacuo, redissolved in EtOAc (300 mL), washed with 1M aqueous HCl (200 mL), saturated aqueous NaHCO₃ (200 mL), and brine (200 mL). The organic layer was then dried over Na₂SO₄ and concentrated in vacuo. The residue was purified by chromatography on silica gel using a gradient of 0-100% EtOAc in DCM to yield 7 as a white foam (226 mg, 0.509 mmol, 44% yield). ¹H NMR (500 MHz, CDCl₃): δ 9.18 (s, 1H), 7.81 (d, J = 7.2 Hz, 2H), 7.66 (d, J = 7.5 Hz, 1H), 7.52 (app t, J = 7.4 Hz, 1H), 7.46 – 7.33 (m, 3H), 6.34 (d, J = 1.5 Hz, 1H), 5.17 (d, J = 2.5 Hz, 1H), 4.81 (app t, J = 2.0 Hz, 1H), 4.34 (d, J = 14.2 Hz, 1H), 4.29 (d, J = 14.3 Hz, 1H), 0.85 (s, 9H), 0.10 (s, 3H), 0.05 (s, 3H) ppm. ¹³C NMR (126 MHz, CDCl₃): δ 166.7, 162.6, 161.2, 154.7, 143.7, 133.0, 132.7, 128.8, 127.6, 101.3, 97.1, 94.0, 80.6, 57.2, 25.6, 18.0, −4.7, −4.9 ppm. HRMS-ESI⁺ (m/z) calc’d [M + H]⁺ for C₂₂H₃₀N₃O₅Si: 444.1949, found: 444.1933.

Isopropyl ((S)-(((4R,5R)-5-(4-benzamido-2-oxopyrimidin-1(2H]-yl]-4-((tert-butyldimethylsilyl]oxy]-4,5-dihydrofuran-2-yl]methoxy](phenoxy]phosphoryl]-L-alaninate (8).

8 was synthesized utilizing a reported methodology.³⁰ 7 (153 mg, 0.345 mmol, 1.0 eq) and N-[(S)-(2,3,4,5,6-pentafluorophenoxy)phenoxyphosphinyl]-L-alanine-1-isopropyl ester (234 mg, 0.517 mmol, 1.5 eq) were mixed as dry powders in a flame dried 10 mL round bottom flask and vacuum dried overnight before use. The reaction was backfilled with N₂, then dissolved in anhydrous pyridine (1.5 mL) and cooled to 0°C. To this solution was added AlMe₂Cl (1.0 M solution in hexanes, 173 μL, 0.173 mmol, 0.5 eq) and stirred at 0°C for 10 minutes before warming to room temperature and stirring for 48 hours at room temperature under N₂. The reaction was quenched by the addition of L-tartaric acid (30% w/v solution, 1 mL), the mixture transferred to a separatory funnel, extracted with EtOAc (100 mL), washed with brine (100 mL, 2x), the organic layer dried over Na₂SO₄ and concentrated in vacuo. The residue was purified by chromatography on silica gel using a gradient of 0-60% EtOAc in hexanes to yield 8 as a light-yellow oil (203 mg, 0.319 mmol, 93% yield). ¹H NMR (500 MHz, CDCl₃): δ 9.28 (br s, 1H), 7.93 (d, J = 7.5 Hz, 2H), 7.56 (app t, J = 7.4 Hz, 1H), 7.51 (d, J = 7.5 Hz, 1H), 7.46 (app t, J = 7.7 Hz, 2H), 7.43-7.33 (m, 1H), 7.30 (app t, J = 8.1 Hz, 2H), 7.22 (d, J = 8.5 Hz, 2H), 7.14 (app t, J = 7.3 Hz, 1H), 6.30 (d, J = 1.5 Hz, 1H), 5.19 (d, J = 2.4 Hz, 1H), 5.00 (hept, J = 6.2 Hz, 1H), 4.78 (s, 1H), 4.74 (s, 1H), 4.72 (s, 1H), 4.30 (q, J = 9.6 Hz, 1H), 4.01 – 3.91 (m, 1H), 1.41 (d, J = 7.1 Hz, 3H), 1.21 (d, J = 4.4 Hz, 3H), 1.20 (d, J = 4.3 Hz, 3H), 0.86 (s, 9H), 0.12 (s, 3H), 0.06 (s, 3H) ppm. ¹³C NMR (126 MHz, CDCl₃): δ 172.8 (d J_C-P = 6.9 Hz, 1C), 166.5, 162.6, 156.3 (d, J_C-P = 8.4 Hz, 1C), 154.2, 150.6 (d, J_C-P = 6.8 Hz, 1C), 143.1, 133.0, 129.7, 128.7, 125.0, 120.0 (d, J_C-P = 4.9 Hz), 104.0, 97.0, 94.3, 80.4, 69.1, 60.5 (d, J_C-P = 4.3 Hz, 1C), 50.3, 25.6, 21.6, 21.6, 20.6 (d, J_C-P = 5.3 Hz, 1C), 18.0, −4.7, −5.0 ppm. One carbon signal is merged into another; possible merged peak is visible in the ¹³C NMR of 3, 5, 7, and 9 at approximately 132.9 ppm. ³¹P NMR (202 MHz, CDCl₃): δ 2.69 ppm. HRMS-ESI⁺ (m/z) calc’d [M + H]⁺ for C₃₄H₄₆N₄O₉PSi: 713.2766, found: 713.2739.

Isopropyl ((S)-(((4R,5R)-5-(4-benzamido-2-oxopyrimidin-1(2H)-yl)-4-hydroxy-4,5-dihydrofuran-2-yl)methoxy)(phenoxy)phosphoryl)-L-alaninate (9).

8 (48.1 mg, 0.0755 mmol, 1.0 eq) was dissolved in THF (1 mL) and aqueous CF₃CO₂H (80% CF₃CO₂H, v/v, 1 mL) and stirred for 3 hours. The reaction was concentrated in vacuo and purified by chromatography on silica gel using a gradient of 0-10% MeOH in DCM to yield 9 as a colorless oil (23.5 mg, 0.0393 mmol, 52% yield). ¹H NMR (500 MHz, CDCl₃): δ 9.38 (br s, 1H), 7.91 (d, J = 7.3 Hz, 2H), 7.67 (d, J = 7.5 Hz, 1H), 7.56 (app t, J = 7.4 Hz, 1H), 7.54 – 7.49 (m, 1H), 7.45 (app t, J = 7.7 Hz, 2H), 7.31 (app t, J = 7.8 Hz, 2H), 7.21 (d, J = 8.1 Hz, 2H), 7.14 (app t, J = 7.4 Hz, 1H), 6.25 (d, J = 2.1 Hz, 1H), 5.32 (d, J = 2.4 Hz, 1H), 4.99 (hept, J = 6.3 Hz, 1H), 4.89 (s, 1H), 4.75 (s, 1H), 4.74 (s, 1H), 4.61 (br s, 1H), 4.19 (q, J = 9.7 Hz, 1H), 4.03 – 3.91 (m, 1H), 1.39 (d, J = 7.1 Hz, 3H), 1.21 (d, J = 1.7 Hz, 3H), 1.20 (d, J = 1.8 Hz, 3H) ppm. ¹³C NMR (126 MHz, CDCl₃): δ 172.9 (d, J_C-P = 7.1 Hz, 1C), 166.7, 162.8, 156.1 (d, J_C-P 8.2 Hz, 1C), 155.2, 150.5 (d, J_C-P = 6.8 Hz, 1C), 143.1, 133.0, 132.9, 129.7, 128.8, 127.8, 125.0, 120.0 (d, J_C-P = 4.9 Hz, 1C), 103.6, 97.2, 95.3, 80.0, 69.2, 60.5 (d, J_C-P = 4.4 Hz, 1C), 50.3, 21.6, 21.6, 20.7 (d, J_C-P = 5.3 Hz, 1C) ppm. ³¹P NMR (202 MHz, CDCl₃): δ 2.69 ppm. HRMS-ESI⁺ (m/z) calc’d [M + H]⁺ for C₂₈H₃₂N₄O₉P: 599.1901, found: 599.1878.

Isopropyl ((S)-(((4R,5R)-5-(4-amino-2-oxopyrimidin-1(2H)-yl)-4-hydroxy-4,5-dihydrofuran-2-yl)methoxy)(phenoxy)phosphoryl)-L-alaninate (1).

9 (79.1 mg, 0.132 mmol, 1.0 eq) was dissolved in acetic acid and pyridine (AcOH:pyridine 1:4, 500 μL) and to this flask was added hydrazine hydrate (64% NH₂NH₂, 30.0 μL, 0.396 mmol, 3.0 eq) and the reaction was stirred for 24 hours. The reaction was concentrated in vacuo and purified by chromatography on silica gel using a gradient of 0-15% MeOH in DCM to yield 1 as a pale-yellow foam (55.8 mg, 0.113 mmol, 85% yield). The Harki laboratory identifier for 1 is HLB-0532247. ¹H NMR (500 MHz, CDCl₃): δ 7.35 – 7.25 (m, 3H), 7.21 (d, J = 8.1 Hz, 2H), 7.15 (app t, J = 7.4 Hz, 1H), 6.22 (s, 1H), 5.76 (d, J = 7.4 Hz, 1H), 5.28 (s, 1H), 5.00 (hept, J = 6.2 Hz, 1H), 4.90 (s, 1H), 4.70 (s, 1H), 4.69 (s, 1H), 4.13 (app t, J = 10.7 Hz, 1H), 4.01 – 3.90 (m, 1H), 1.37 (d, J = 7.0 Hz, 3H), 1.23 (d, J = 2.4 Hz, 3H), 1.21 (d, J = 2.4 Hz, 3H) ppm. ¹³C NMR (126 MHz, CDCl₃): δ 173.2 (d, J_C-P = 6.4 Hz, 1C), 166.0, 156.1, 156.0, 150.6 (d, J_C-P = 6.7 Hz, 1C), 139.8, 129.7, 125.1, 120.1 (d, J_C-P = 4.8 Hz, 1C), 103.5, 95.8, 94.8, 79.6, 69.3, 60.9, 50.4, 21.7, 21.6, 20.7 (d, J_C-P = 5.6 Hz, 1C) ppm. ³¹P NMR (202 MHz, CDCl₃): δ 2.64 ppm. HRMS-ESI⁺ (m/z) calc’d [M + H]⁺ for C₂₁H₂₈N₄O₈P: 495.1639, found: 495.1617.

4-N-Benzoyl-3’-deoxy-3’,4’-didehydro-2’-O-tert-butyldimethylsilylcytidine-5’-H-phosphonate (10).

10 was synthesized using a reported methodology.⁴³ 7 (106 mg, 0.239 mmol, 1.0 eq) was vacuum dried overnight before use. 2-Chloro-1,3,2-benzodioxaphosphorin-4-one (58.1 mg, 0.287 mmol, 1.2 eq) was vacuum dried overnight before use. 2-Chloro-1,3,2-benzodioxaphosphorin-4-one was dissolved in anhydrous dioxane and pyridine (1:1 v/v solution, 4 mL). To this solution was added a solution of 7 dissolved in anhydrous dioxane and pyridine (1:1, 1 mL) and the mixture was stirred at room temperature for 2 hours. After 2 hours, the reaction was quenched with H₂O (1 mL) and the mixture evaporated in vacuo. The residue was purified by chromatography on Et₃N-deactivated silica gel using a gradient of 0-20% MeOH in DCM to yield 10 (triethylammonium salt) as a glassy solid (105 mg, 0.172 mmol, 72% yield). Loss of the TBS group can occur on acidic silica gel, which in this case also yielded 4-N-benzoyl-3’-deoxy-3’,4’-didehydrocytidine-5’-H-phosphonate (11, 26.4 mg, 0.0534 mmol, 22%yield). Spectral data for 10: ¹H NMR (500 MHz, MeOD): δ 7.98 (d, J = 7.0 Hz, 2H), 7.93 (d, J = 7.6 Hz, 1H), 7.64 (m, J = 8.2 Hz, 2H), 7.54 (app t, J = 7.7 Hz, 2H), 6.86 (d, J_H-P = 625 Hz, 1H), 6.30 (d, J = 1.4 Hz, 1H), 5.40 (d, J = 2.6 Hz, 1H), 4.92 (s, 1H), 4.60 (s, 1H), 4.58 (s, 1H), 3.21 (q, J = 7.3 Hz, 8H), 1.31 (t, J = 7.3 Hz, 12H), 0.92 (s, 9H), 0.18 (s, 3H), 0.13 (s, 3H) ppm. (Quartet at 3.47 and 3.20, and triplets at 1.35 and 1.31 ppm correspond to the triethylamine salt form of the product, as well as excess triethylamine-HCl in the product that was not removed in this preparation). ¹³C NMR (126 MHz, MeOD): δ 169.0, 165.1, 160.1 (d, J_C-P = 7.7 Hz, 1C), 157.6, 145.2, 134.7, 134.1, 129.8, 129.2, 103.9, 98.9, 95.8, 82.0, 63.5, 59.2 (d,J_C-P = 3.3 Hz, 1C), 53.6, 47.8, 26.2, 18.9, 9.2, 7.6, −4.5, −4.6 ppm. ³¹P NMR (202 MHz, MeOD): δ 4.30 ppm. HRMS-ESI⁺ (m/z) calc’d [M + Et₃N + H]⁺ for C₂₈H₄₆N₄O₇PSi: 609.2868, found: 609.2844.

4-N-Benzoyl-3’-deoxy-3’,4’-didehydrocytidine-5’-H-phosphonate (11).

10 (870. mg, 1.43 mmol, 1.0 eq) was dissolved in DMF (10 mL) and to this solution was added dropwise triethylamine trihydrofluoride (Et₃N·3HF, 583 μL, 3.58 mmol, 2.5 eq). The reaction was monitored by TLC until complete (approximately 3 hours). The reaction was directly evaporated in vacuo and the residue purified by chromatography on silica gel using a gradient of 0-20% MeOH in DCM to yield 11 (triethylammonium salt) as a glassy colorless solid (436 mg, 0.882 mmol, 62% yield). ¹H NMR (500 MHz, MeOD): δ 7.97 (d, J = 7.5 Hz, 2H), 7.92 (d, J = 7.5 Hz, 1H), 7.67 – 7.60 (m, 2H), 7.54 (app t, J = 7.7 Hz, 2H), 6.85, (d, J = 620 Hz, 1H), 6.33 (d, J = 1.5 Hz, 1H), 5.42 (d, J = 2.4 Hz, 1H), 4.80 (s, 1H), 4.58 (s, 1H), 4.57 (s, 1H), 3.47 (q, J = 7.3 Hz, 6H), 1.35 (t, J = 7.3 Hz, 9H).¹³C NMR (126 MHz, MeOD): δ 169.0, 165.1, 160.3 (d, J_C-P = 7.4 Hz, 1C), 157.6, 145.2, 134.7, 134.1, 129.8, 129.2, 103.2, 98.9, 95.8, 80.5, 63.5, 59.1 (d, J_C-P 3.6 Hz, 1C), 53.6, 7.63 ppm. HRMS-ESI⁺ (m/z) calc’d [M + H; no salt form]⁺ for C₁₆H₁₇N₃O₇P: 394.0799, found: 394.0786.

3’-Deoxy-3’,4’-didehydrocytidine-5’-triphosphate (ddhCTP).

ddhCTP was synthesized using a reported methodology.⁴³ 11 (78.2 mg, 0.158 mmol, 1.0 eq) and a stir bar were vacuum dried overnight in a flame dried flask before use. Tri(tetrabutylammonium)pyrophosphate (TBAPP, 285 mg, 0.316 mmol, 2.0 eq) was similarly vacuum dried overnight in a flame dried flask before use. 11 was then purged and flushed with nitrogen gas before dissolving in anhydrous DMF (1.5 mL) and anhydrous pyridine (318 μL, 3.95 mmol, 25 eq). To this solution was added fresh distilled trimethylsilyl chloride (TMSC1, 100. μL, 0.790 mmol, 5 eq; distilled as described in the General section) and the solution was stirred for 10 minutes. Solid I₂ (60.2 mg, 0.237 mmol, 1.5 eq) was dissolved in anhydrous DMF (1 mL) and this solution was added dropwise to the reaction mixture until the brown color persists and then the reaction was allowed to stir for an additional 5 minutes. TBAPP was dissolved in anhydrous DMF (1 mL) and was added all at once to the reaction mixture and then stirred for 1 hour. Next, the crude mixture was transferred to a sealed tube and MeNH₂ in EtOH (30% w/v, 3 mL) was added, the flask was sealed, and the mixture stirred at 50°C for 18 hours. The crude reaction mixture was cooled and concentrated in vacuo. A Sephadex LH-20 column (2.5 x 40 cm) was equilibrated with aqueous triethylammonium acetate (10 mM TEAA, pH 8) buffer. The crude reaction mixture was dissolved in the same buffer, transferred to the column and the molecule eluted at 2 mL/min with the same buffer (10 mM TEAA, pH 8). Fractions were frozen, lyophilized, and analyzed by ¹H and ³¹P NMR. Fractions containing product were further purified by preparative HPLC (Column: Agilent Technologies Zorbax SB-C18, 21.2 x 250 mm size, 7 μm pore; Method: Solvent A: 10 mM TEAA pH 7.5; Solvent B: 1:1 solution of 10 mM TEAA pH 7.5 and MeCN; Gradient: A:B 98:2 from 0-5 minutes, followed by a linear gradient to A:B 60:40 from 5-30 minutes, and from 30-35 minutes an isocratic A:B 60:40) and the pure fractions confirmed by NMR were collected and lyophilized. ddhCTP (tetratriethylammonium salt) was isolated as a glassy colorless solid (7.67 μmol, 5% yield, which was determined using UV-Vis spectrometry and Beer’s Law using the extinction coefficient for cytidine). ¹H NMR (500 MHz, D₂O): δ 7.49 (d, J = 7.5 Hz, 1H), 6.33 (d, J = 1.8 Hz, 1H), 6.06 (d, J = 7.6 Hz, 1H), 5.53 (d, J = 2.7 Hz, 1H), 4.90 (s, 1H), 4.74 – 4.65 (m, 2H), 3.21 (q, J = 7.3 Hz, 14H), 1.29 (t, J = 7.3 Hz, 20H) ppm. ³¹P NMR (202 MHz, D₂O): δ −10.4 (d, J = 19.9 Hz), −11.4 (d, J = 19.8 Hz), −23.2 (t, J = 19.9 Hz) ppm. LRMS-ESI⁺ (m/z) calc’d [M + H] ⁺ for C₉H₁₅N₃O₁₃P₃: 466.0, found 465.9. LRMS-ESI⁻ (m/z) calc’d [M - H] ⁻ for C₉H₁₃N₃O₁₃P₃: 464.0, found 463.8.

General Cell Culture.

All cell lines were maintained in a humidified 5% CO2 environment at 37 °C. HUH7 and HUH7.5 cells (Human hepatocyte derived cells from carcinoma, Apath LLC, new York, NY, USA) were cultured in DMEM (4500 mg/L “high glucose,” + L-glutamine, + 25 mM HEPES, - Sodium Pyruvate; Life Technologies 12430-062) supplemented with fetal bovine serum (FBS, Gibco; 10% FBS was used for cytotoxicity assays and 2% was used in antiviral plaque assays), penicillin (100 I.U./mL), and streptomycin (100 μg/mL, ATCC) and 1% non-essential amino acids (NEAA; Sigma Aldrich M7145-100ML) at a density of 2×10⁵ – 2×10⁶ cells/mL. Vero cells (ATCC, CCL-81) were cultured in EMEM media (ATCC) supplemented with FBS (10% FBS was used for cytotoxicity assays and 2% was used in antiviral plaque assays), penicillin (100 I.U./mL), and streptomycin (100 μg/mL) at a density of 2×10⁵ – 2×10⁶ cells/mL. HUH7 Cells were verified by short tandem repeat (STR) profiling by Creative Bioarray.

Antiviral Plaque Assay.

ddhC and 1 DMSO stock solutions were diluted into the appropriate cell media (0 mM, 0.1 mM and 1 mM). Vero cells, HUH7, and HUH7.5 cells grown in 24 well plates were used to characterize differential antiviral response by cell type. Twenty-four hours prior to confluency cells were pretreated with drug-containing media (250 μL). After pretreatment, cells were checked for confluency, drug containing media was removed and viral dilutions of either West Nile virus (WNV, originally isolated from American Crow from Albany Co, NY in 2003; Vero pass X2 + C6/36 pass x 2. Accession #DQ164189)⁴⁴ or Zika virus (ZIKV) PRABC59 were prepared. Plates were infected at an MOI of 0.1 with diluted virus (100 μL) for one hour and rocked every 30 minutes to ensure even distribution. After inoculation, the inoculum was removed, and wells were washed three times with Hank’s balanced salt solution (BA-1, 1 mL each time). Plates were then overlaid with the appropriate drug containing media (1 mL) and were incubated at 37°C with 5% CO₂ for either 3 days for WNV or 4 days for ZIKV. After incubation, cells were checked for cytopathic effect to ensure infection and contamination to ensure purity. Media aliquots were harvested (800 μL) from each well and stored in FBS (200 μL) and placed in a −80°C freezer. Each concentration of each drug in all cell lines was tested in triplicate. Infectious virus was enumerated from stored samples using plaque titration as in reference ⁴⁵. Titers were compared across drug concentrations using an ordinary one-way ANOVA with Dunnett’s multiple comparisons test Samples which failed the Brown-Forsythe test were re-analyzed using a Welch ANOVA test and Dunnett’s T3 multiple comparisons test using GraphPad Prism.

Western Blot of HINT1 in Vero and HUH7 Cells.

HUH7 or Vero cells were grown to confluency in a T75 flask, detached using a cell dissociation buffer (Gibco Cat# 13151-014), collected by pelleting in a centrifuge at 1000 rpm, and rinsed three times with cold PBS (10 mL) and pelleted using a centrifuge after each wash and decanting the PBS solution. After decanting the last PBS wash, the cells were suspended in approximately RIPA buffer (500 μL, Pierce Cat# 89901) containing protease and nuclease inhibitors (tablets, Pierce Cat# 32965). Cell lysates were stored at −80°C until analysis. Cell protein concentration was normalized to the lowest concentration sample using a Thermo Scientific Pierce BSA Protein Assay Standards calibration curve, which was constructed according to the manufacturer protocol. Normalized cell lysis samples were run on a 4-12% Bis-Tris gradient gel at 184V, 0.38A, 70W for approximately 50 minutes. Samples were transferred using a Bio-Rad Trans-Blot Turbo system per manufacturer’s instructions. Blocking performed using a blocking solution (5%, Bio-Rad Cat# 1706404) for 4 hours at room temperature (volume used was enough to cover the whole blot in a plastic container; approximately 10 mL was prepared). Blotting performed using HINT1 polyclonal antibody (Proteintech, Cat# 10717-1-AP; 1:500 dilution in 5% blocking solution with 1% sodium azide) for 18 hours. The membrane was then washed with TBST (10 mL) for 5 minutes five times. Goat anti-rabbit 2° antibody (HRP conjugation, Bio-Rad, Cat#170-6515; 1:1000 dilution in 5% blocking solution with 1% sodium azide) was incubated for 2 hours before HRP imaging. The membrane was washed with TBST (10 mL) for 5 minutes five times before re-blotting for beta-actin (Sigma Aldrich, Cat# A1978, 1:2000 dilution in 5% blocking solution with 1% sodium azide). A second series of washing with TBST (10 mL) for 5 minutes five times was followed by treatment with goat anti-mouse Alexa Fluor 680 antibody for imaging of beta-actin as described above (Alexa Fluor Plus 680, Invitrogen, Cat# A32729, 1:1000 dilution in 5% blocking solution with 1% sodium azide). Imaging performed on a Lycor Odyssey FC as per manufacturer’s instructions.

In Cellulo Metabolism of 1 and ddhC to ddhCTP.

Note: Mass spectrometry quantitation of prodrug metabolites were contracted to WuXi AppTec using cell samples and standards prepared by KTP. Cell culture collection procedure was based on a previously report.⁴⁶ HUH7 cells were seeded at 5x10⁶ each into two T75 flasks in appropriate cell media (10 mL) and incubated for 24 hours. A DMSO solution of 1 was diluted to yield a stock solution in media (1000 μM of 1 in 5% DMSO in media). An aliquot of media (1 mL) was removed from each of the T75 flasks, and then 1 (1 mL of the 1000 μM stock solution in 5% DMSO in media) was dosed to one flask (final concentration of 1 was 100 μM; final DMSO concentration was 0.5%). The second T75 flask received a negative, no-compound control (1 mL of 5% DMSO in media; results in a 0.5% final DMSO concentration) and both flasks were incubated for 24 hours. After 24 hours, the media was decanted from the flasks, the cells trypsinized (2 mL, Thermo Fisher), counted with an automated cell counter and Trypan blue staining (Thermo Fisher), and then centrifuged at 1000 rpm for 5 minutes. The pellets were resuspended in cold aqueous MeOH (60% MeOH, 1 mL), incubated at −20°C overnight, and then underwent three freeze-thaw cycles in liquid nitrogen. The cell mixture was centrifuged at 14,000 rpm for 5 minutes, the supernatant collected, lyophilized, and stored at −20°C until use.

To analyze levels of 1 and ddhCTP found in lysates, the following procedure was employed. Lysate samples were reconstituted in a 1:1 solution of MeOH:H₂O (100 μL). An aliquot (4 μL) was treated with an internal standard solution (200 μL of a 200 nM adenosine-¹³C₁₀ 5’-triphosphate in MeOH; referred to as the internal standard (IS)), the mixture vortexed, and then centrifuged at 3900 rpm for 10 minutes at 4°C. An aliquot of the supernatant (100 μL) was further diluted with water (100 μL) in a sample plate, the plate shaken at 800 rpm for 10 minutes. An aliquot of this mixture (20 μL) was injected for LC-MS analysis.

An identical procedure was performed for the culturing of cells and dosing of ddhC. To analyze levels of ddhCTP found in ddhC-dosed lysates, the previous procedure was employed. To analyze levels of ddhC found in lysates, the following procedure was employed. Cell lysates were reconstituted with cold aqueous MeOH (100 μL, 1:1 MeOH:H₂O). An aliquot (30 μL) was treated with a second internal standard solution (200 μL, of a 100 ng/mL Adenosine ¹³C₅ in MeOH; referred to as internal standard 2, IS2), the mixture vortexed, and then centrifuged at 12000 rpm for 5 minutes at 4°C. An aliquot of the of the resulting supernatant (180 μL) was blow-dried in a sample plate using nitrogen gas flow under 35°C. The resulting residue was resuspended in water (150 μL), shaken for 5 minutes, and aliquots of this solution (1 μL) were injected for LC-MS analysis. All cell lysate treatments were done in triplicate on different days paired with a DMSO control.

Sample concentrations of 1, ddhC, and ddhCTP analytes were calculated by measuring the concentration of the analytes in three biological replicates for each compound dosed (DMSO or 1 or ddhC) relative to a calibration curve of synthetic standards (see Figure S5, Figure S6, Figure S7; see below for technical details). Concentrations of analytes measured are reported as nanomoles (nmol) detected per 5x10⁶ cells in the collected lysates (cell counts and viabilities were measured during cell lysate collection using Trypan Blue staining).

Mass spectrometry data for the detection of 1 and ddhCTP was collected using a Sciex Triple Quad 6500 (ESI: negative; multiple reaction monitoring (MRM) detection of ddhCTP (464.2), 1 (493.1), IS (516.3)) equipped with a Kentex EVO C18 column (2.6 ×m, 20 × 2.1 mm). A two solvent system was used as the eluents: A = 5 mM N,N-dimethylhexylamine (DMHA) in water, pH 7 and solvent B = 5 mM N,N-dimethylhexylamine(DMHA) in MeCN. The gradient (0.5000 mL / min flow rate) consisted of: A:B 95:5 from 0-0.1 minutes, followed by a linear gradient to A:B 70:30 from 0.1-2 minutes, an isocratic A:B of 70:30 from 2-2.70 minutes, followed by a return to the starting condition of A:B 95:5 from 2.70-2.71 minutes followed by a second isocratic A:B 95:5 from 2.71-3.50 minutes. Observed retention times of the standards were as follows: ddhCTP: 1.69 minutes; 1: 2.31 minutes; IS: 1.75 minutes. A calibration curve of both 1 and ddhCTP was calculated by preparing a stock solution and diluting to form an 8-point dilution series from 1-3000 ng/mL (in PBS buffer) of the analyte standard along with 1 ng/mL of IS. A plot of A_s/A_IS versus C_s/C_IS was made and linear regression with 1/x² weighing was performed to calculate a calibration curve where A_s is the analyte peak area, A_IS is the IS peak area, C_s is the concentration of the analyte, and C_IS is the concentration of the IS. Data analysis was performed using integrated Instrument Control and Data Processing Software Analyst 1.6.3 software. The experimental sample concentrations of 1 and ddhCTP were calculated based on these curves (see Figure S5 and Figure S6). Final analyte concentrations are reported as the average of three replicates ± the standard deviation.

Mass spectrometry data for the detection of ddhC was collected using a Waters Zevo TQs LC-MS/MS system (ESI: positive, MRM detection of ddhC (226.1) and IS2 (273.1)) equipped with an Acquity UPLC BEH C18 column (1.7 μm, 50 × 2.1 mm). A two solvent system was used as the eluents: A = 1% acetic acid in water pH 8.5 and solvent B = MeCN. The gradient (0.5000 mL / min flow rate) consisted of: A:B 99:1 from 0-0.90 minutes, followed by a linear gradient to A:B 70:30 from 0.90-1.30 minutes, followed by a return to the starting condition of A:B 95:1 from 1.30-1.31minutes followed by a isocratic A:B 99:1 from 1.31-3.00 minutes. Observed retention times of the standards were as follows: ddhC: 0.75 minutes; IS2: 1.65 minutes. A calibration curve of ddhC and IS2 was calculated by preparing a stock solution and diluting to form an 8-point dilution series from 5-2500 ng/mL of the ddhC standard (2:8 MeOH:H₂O, v/v) along with 1 ng/mL of IS2. Curves using the analyte and IS2 signal were prepared as above and used to calculate experimental sample concentrations of ddhCTP and ddhC (see Figure S5 and Figure S7). Final analyte concentrations are reported as the average of three replicates ± the standard deviation.

In the text, we made a note of approximating the molarity of cellular ddhCTP produced in the course of this dosing study. Approximate cell volumes and ddhCTP concentrations were calculated using previously published images of HUH7 cells to estimate cell diameter,³⁹ and calculating cell volume using the formula 4/3*π*r³ where r is an estimated 5 μm giving an approximate cell volume of an HUH7 cell as 532 fL. Concentrations were then calculated after taking the measured amount of ddhCTP per sample and dividing it among the total number of cells collected at the end of compound treatment.

Protocol for Alamar Blue Cell Viability Assay.

Cytotoxicity assays were done similar to those previously described.⁴⁷ HUH7 or Vero cells were seeded at a density of 2,000 cells/well or 1,500 cells/well, respectively, in the appropriate cell culture media (50 μL) in 96-well clear plates (Costar) and incubated for 24 hours. A DMSO solution of 1 or ddhC was serially diluted in the appropriate cell line media and each solution (50 μL) was dosed to cells (well volume total of 100 μL, final DMSO concentration 0.5%). DMSO-only and no-cell controls wells were dosed with a DMSO-media solution (50 μL of 1% DMSO in media) rather than 1 or ddhC. The cells were then incubated for 2 days or 5 days. 2 hours before the designated time point, Alamar blue (10 μL, Invitrogen) was added to each well. At the time point, fluorescence data was measured using a BioTek Synergy H1 microplate reader. Cell viability was calculated by subtracting background fluorescence (no cell control) from the measured fluorescence of each well. Wells with compound were normalized to the DMSO only control (100% viability) and no-cell controls. Each individual experiment was performed in groups of three technical replicates and then repeated at least in biological triplicate. The final cell viability percentage was calculated by averaging % viability from each of the biological replicates and the uncertainty was calculated by propagating the standard deviations from each biological replicate (taking the square root of the sum of the squares of individual standard deviations). Cytotoxicity values were compared across drug concentrations using an ordinary one-way ANOVA with Dunnett’s multiple comparisons test using GraphPad Prism and Microsoft Excel.

Caco-2 Cell Permeability Assays.

Note: Caco-2 permeability assays were contracted to WuXi AppTec using compounds synthesized by KTP. Caco-2 cells (ATCC HTB-37) cells were seeded at 1 x 10⁵ cells/cm² onto polyethylene membranes (PET) in 96 well plates and the medium refreshed every 4-5 days until confluent monolayers formed (21-28 days). Transport solutions were prepared as follows: HBSS (Hank’s balanced salt solution), 10 mM HEPES pH 7.4 and test compound (2 μM or 10 μM of either ddhC or 1) where the final DMSO concentration was < 1%. Digoxin was used as a bidirectional assay control (2 μM); nadolol and metoprolol were tested at 2 μM in the apical (A) to basolateral (B) direction only. Cell monolayers were incubated with the corresponding transport solutions for 2 hours in a humidified incubator set to 37°C at 5% CO₂. After 2 hours, samples of starting solution, donor solution, and receiver solution were quantified for each tested and control compound using LC-MS by analyzing peak area ratios of the different analytes relative to an internal standard that was added before analysis. At the end of each assay, Lucifer Yellow rejection assays were performed to certify Caco-2 cell monolayer integrity. The apparent permeability coefficient Papp (cm/s) was then calculated using Equation 1:

$${\text{P}}_{\text{app}}=\frac{\left(\frac{{\text{dC}}_{\text{r}}}{\text{dt}}\right)\times {\text{V}}_{\text{r}}}{\text{A}\times {\text{C}}_{\text{0}}}$$

where dC_r/dt is the cumulative concentration of compound in the chosen receiving chamber as a function of time (μM/s); V_r is the volume of the chosen chamber solution (0.075 mL apical side; 0.25 mL basolateral side); A is the surface area of the cell monolayer (0.0804 cm²); C₀ is the initial concentration of compound in the donor chamber in μM. The efflux ratio was calculated by Equation 2:

$$\text{Efflux Ratio}=\frac{{\text{P}}_{\text{app}}\left(\text{BA}\right)}{{\text{P}}_{\text{app}}\left(\text{BA}\right)}$$

where P_app (AB) is the permeability coefficient for the direction A to B and P_app (BA) is the permeability coefficient for the direction B to A.

Human and Mouse Plasma Stability Assay.

Note: Plasma stability experiments were contracted to WuXi AppTec using prodrug 1 prepared by KTP. Pooled plasma samples (Mouse: 20 males, from BioreclamationIVT, Cat# MSE00PLK2P2N, batch MSE321336; Human: 3 male, 3 female from BioreclamationIVT, Cat# HUMANPLK2P2N, batch HMN51524) were thawed at 37°C, centrifuged at 4000 rpm for 5 minutes, and clots removed. pH was adjusted to 7.4 if necessary. 1 was diluted in DMSO (to 100 μM) and positive control propantheline diluted (to 100 μM) in a solution of aqueous MeOH (45% MeOH). Plasma solutions (98 μL) were dosed with the prepared solutions of test and control compounds to make a final 2 μM dose. Samples were then incubated at 37°C in a water bath and time points taken at 0, 10, 30, 60, and 120 minutes. Time point samples were mixed with quench solution containing internal standards tolbutamide (200 ng/mL) and labetalol (200 ng/mL) in a 1:1 solution of MeCN:MeOH. Quenched samples were centrifuged at 4000 rpm for 10 minutes before being subject to LC-MS. The amount of sample was quantified by peak analyte ratio (PAR) relative to internal standards according to Equation 3:

$$\text{\%Remaining=}\frac{\text{PAR}\left({\text{t}}_{\text{n}}\right)}{\text{PAR}\left({\text{t}}_{\text{0}}\right)}$$

where the PAR(t_n) was the PAR measured at time (t) = 0, 10, 30, 60, or 120 minutes and PAR(t₀) was the PAR measured at t = 0 minutes.

Simulated Gastric Fluid Stability Assay.

Note: SGF stability experiments were contracted to WuXi AppTec using prodrug 1 prepared by KTP. Simulated gastric fluid (SGF) was prepared by the dissolution of NaCl (0.08 g), pepsin (0.128 g), HCl (0.28 mL) and H₂O (40 mL) after which the pH is approximately 1.20. A 96 well plate was prepared by the addition of the chosen compound solution (2 μL; DMSO stock at 200 μM) to each of the wells corresponding to time points 0, 60, 120, 360, and 1440 minutes. SGF solution (198 μL) was added to each well except the 0 time point and the samples incubated at 37°C and shaken at 300 rpm for the appropriate time. When the time point is reached, the sample was mixed with 400 μL cold MeCN containing internal standards (200 ng/mL tolbutamide and labetalol). All samples were centrifuged at 4000 rpm at 4°C for 20 minutes and aliquots prepared for LC-MS-MS analysis. Remaining compound was measured as per Equation 3.

Enzymatic Incorporation of ddhCTP by Human DNA Polymerase Gamma.

Equimolar amounts of the template and FAM-labeled primer (see Figure 3A) were diluted in 1x annealing buffer (5 mM Mg(OAc)₂, 20 mM HEPES pH 7.5 in DNase/RNase free water) and heated to 95°C for 10 minutes. The heat was then turned off, and the DNA allowed to cool to room temperature overnight for complete annealing. Extension reaction mixtures were performed by combining DNA polymerase gamma (10 nM), primer/template DNA annealed duplex (100 nM), and 100 μM of CTP, dCTP, or ddhCTP in a reaction buffer (50 mM TrisHCl, 100 mM KCl, 10 mM MgCl₂, 0.4 mg/mL BSA, 15% glycerol). After the components were mixed, the assay was run at 37°C and time points taken at the indicated times zero (taken before addition of the nucleotide), 10, and 20 minutes. Time point samples were stored in aqueous quenching buffer (80% v/v formamide, 0.1 M EDTA, 0.01% w/v bromophenol blue, and 0.1% w/v xylene cyanol). Samples were then directly ran on a 7M urea 20% polyacrylamide gel and run for approximately 2.5 hours until substrate and product bands were separated on the single nucleotide level. Gels were imaged using a BioRad ChemiDoc. DNA sequences were purchased from Eurogentec. DNA polymerase gamma (catalytic subunit, catalog number 85) was purchased from Enzymax (Lexington, KY). Nucleotides CTP and dCTP were purchased from Thermo Fisher Scientific. ddhCTP sample was synthesized as described in this experimental section.

Construction of ZIKV RdRp Bacterial Expression Plasmid.

The ZIKV RdRp (NS5 gene) was cloned into the pET26Ub-CHIS bacterial expression plasmid using a similar procedure as described for WNV RdRp (NS5 gene).⁴¹ This system allows for the production of ubiquitin fusion proteins containing a carboxy-terminal hexahistidine tag that are then co- and/or post-translationally processed by the ubiquitin protease, co-expressed from a second plasmid, pUBPS.^48–49 Briefly, the ZIKV RdRp coding region was amplified using a synthetic NS5 gene construct as template (synthesized by GenScript) based upon the NS5 amino acid sequence of the Zika virus strain BeH815744 (AMA12087.1), oligonucleotides 5’-TGGTCCTGCGTCTCCGCGGTGGAGGTGGCGGTACCGGCGAAACCCTGGGC-3’ and 5’-GGTGACCAGAGGATCCCAGAACGCCCGGGGTGCT-3’. The PCR product ZIKV RdRp was gel purified and cloned into the pET26Ub-CHIS plasmid using SacII and BamHI sites. The final construct (pET26Ub-ZIKV-RdRp-CHIS) was confirmed by sequencing at The Pennsylvania State University’s Nucleic Acid Facility.

Expression and Purification of ZIKV RdRp.

ZIKV RdRp was expressed and purified using the same procedure reported for WNV RdRp.⁴¹ Briefly, expression was performed at 15 °C by auto-induction, cells harvested, lysed by French press, subjected to PEI precipitation followed by AMS04 precipitation, Ni-NTA chromatography, gel filtration and the protein concentrated using Vivaspin concentrators.

ZIKV RdRp-Catalyzed Nucleotide Incorporation Assays.

To assemble ZIKV RdRp elongation-competent complexes, purified ZIKV RdRp (1 μM) was mixed with pGGC RNA primer (10 μM), RNA template (5’-UUUAGCUCUUCCUCUUUGCC-3’, 1 μM), ATP (10 μM), GTP (10 μM) and [α-32P]-ATP (0.1 μCi/μL) for 30 min at 30 °C in the reaction buffer (50 mM HEPES pH 7.5,20 mM NaCl, 5 mM MgCi2 and 10 mM 2-mercaptoethanol). For single nucleotide incorporation assays, elongation complexes were assembled, reactions were initiated with either 100 μM CTP, 3’-dCTP, or ddhCTP nucleoside triphosphate substrate and allowed to proceed for 5 minutes before being quenched with EDTA (50 mM). For chain termination experiments, reactions were allowed to proceed for an additional 1 min in the presence of the next correct nucleotide substrate (10 μM UTP) before being quenched. To evaluate the ability of ddhCTP or 3’-dCTP to inhibit RdRp elongation, elongation complexes were assembled, reactions were initiated with the addition of UTP and CTP and varying concentrations of ddhCTP and allowed to proceed for 5 min before being quenched with EDTA (50 mM). Products were resolved from substrates by denaturing PAGE. ZIKV RdRp was diluted immediately prior to use in enzyme dilution buffer (50 mM HEPES, pH 7.5, 10 mM 2-mercaptoethanol, 20% glycerol). The volume of enzyme added to any reaction was always less than or equal to one-tenth the total volume.

Denaturing PAGE Analysis of Polymerase-Catalyzed Reaction Products.

Two volumes of loading buffer (10 μL of 90% formamide, 0.025% bromophenol blue and 0.025% xylene cyanol) was added to quenched reaction mixtures (5 μL) and heated to 90 °C for 2-5 min prior to loading (5 μL) on a denaturing 23% polyacrylamide gel containing 1X TBE (89 mM Tris base, 89 mM boric acid, 2 mM EDTA) and urea (7 M). Formamide gels contained an additional 20% formamide. Electrophoresis was performed in 1XTBE at 90 W. Gels were visualized by using a PhosphorImager (GE) and quantified by using ImageQuant TL software (GE).

ABBREVIATIONS

CTP

cytidine triphosphate

ddhC

3’-deoxy-3’,4’-didehydro-cytidine

ddhCTP

3’-deoxy-3’,4’-didehydro-cytidine-5’-triphosphate

DENV

dengue virus

DMF

N,N’-dimethylformamide

DMP

2,2-dimethoxypropane

FAM

carboxy fluorescein

HINT1

histidine triad nucleotide-binding protein 1

HCV

hepatitis C virus

IFNα

interferon alpha

NTP

nucleotide triphosphate

PV

polio virus

ProTide

phosphoramidate monophosphate prodrug

RdRp

RNA dependent RNA polymerase

RNA

ribonucleic acid

rPol

RNA polymerase

RT

reverse transcriptase

SARS-CoV-2

Severe Acute Respiratory Syndrome Coronavirus 2

SGF

simulated gastric fluid

WNV

West Nile virus

ZIKV

Zika virus