Regio- and Stereoselective Synthesis of 2′-Deoxy-4′-thioguanine Nucleosides: Evaluation of Anti-Hepatitis B Virus Activity and ­Cytotoxicity Leading to Improved Selectivity Index by 4′-C-Cyanation

Abstract

An N ⁹-regio- and β-anomer-selective 4′-thioglycosidation of purine bases has been developed. The reaction between a 2-deoxy-2-iodo-4-thioribofuranosyl glycosyl donor and N-(6-chloro-9H-purin-2-yl)-2-methylpropanamide gave the corresponding 2′-deoxy-4′-thiopurine nucleoside in 87% yield along with its N ⁷-regioisomer in 6% yield, without the formation of the α-anomer. By using a derivative obtained from 17, a practical chemical synthesis of 2′-deoxy-4′-thioguanosine was developed. 4′-α-C-Cyano-2′-deoxy-4′-thioguanosine was synthesized, starting from a 4-(acetoxymethyl)-2-deoxy-2-iodo-4-thioribofuranose derivative as a glycosyl donor. An evaluation of the anti-hepatitis B virus (HBV) activity and the cytotoxicity toward the host cell revealed that 4'-C-cyano-2'-deoxy-4'-thioguanosine exhibited about 100 times more potent anti-HBV activity than 2′-deoxy-4′-thioguanosine with a comparative cytotoxicity, resulting in the identification of a novel molecule having better selectivity index value than that of 2′-deoxy-4′-thioguanosine. This finding might provide a guideline for the development of the next generation of anti-HBV agents.

Keywords

nucleosides - glycals - deoxythioguanosine - medicinal chemistry - cytotoxicity - hepatitis B virus

Hepatitis B is one of the most prevalent viral diseases, with an estimated 400 million people affected globally, and is known to be a major cause of chronic disease, leading to cirrhosis and/or hepatocellular carcinoma.[1] Five nucleoside/nucleotide analogues have been approved for the treatment of chronic hepatitis B; these include three nucleoside derivatives (entecavir, telbivudine, and lamivudine) and two nucleotide analogues (adefovir and tenofovir) (Figure [1]). The target enzyme of these molecules is hepatitis B virus (HBV) polymerase, a multifunctional protein with RNA- and DNA-dependent DNA polymerase functions that are essential for viral replication. Despite this successful development of anti-HBV agents, the synthesis of novel nucleoside derivatives with different core structures is essential to combat the emergence of resistant strains of the virus.

Figure 1 Structure of anti-HBV nucleoside/nucleotide derivatives

Figure 2 Structure of compounds 1–3

4′-Thionucleosides, in which the oxygen atom in the furanose ring is replaced by a sulfur atom, have attracted much attention since the discovery of the antiviral and antitumor activities of 4′-thiothymidine (1) and 2′-deoxy-4′-thiocytidine (2) (Figure [2]).[2] [3] [4] [5] [6] The 4′-thionucleoside 2′-deoxy-4′-thioguanosine (3) has been reported to exhibit potent inhibitory activity toward HBV and has been recognized as a lead compound for the next generation of anti-HBV agents.[7]

In addition, recent reports have revealed that 4′-C-Cyano-2′-deoxyguanosine (4) shows highly potent anti-HBV activity (Figure [3]).[8] Stimulated by these facts, we became interested in the synthesis and the evaluation of the anti-HBV activity of the novel 2′-deoxy-4′-thioguanosine 5, substituted with a cyano group at the 4′-position. This letter describes the development of a method for the synthesis of 5, and the evaluation of its anti-HBV activity and cytotoxicity toward host cells.

Figure 3 Structure of 4′-C-cyano-2′-deoxyguanosine (4) and the target molecule 5

We have previously described an N-iodosuccinimide (NIS)-mediated electrophilic glycosidation by the 4-C-ethynyl-4-thiofuranoid glycal 6 of the trimethylsilylated derivative of compound 7 (TMS-7), leading to the formation of three isomeric glycosides: the N ⁹-glycoside 8 (25% yield), the N ⁷-glycoside 9 (12% yield), and the N ¹-glycoside 10 (29% yield) (Scheme [1]).[9] Although the glycosidation gave the corresponding β-anomer as the sole stereoisomer, three regioisomers were formed.

Scheme 1 Previous report on the β-face-selective electrophilic glycosidation of 6 with 7

By using this protocol, we synthesized the β-anomer of the 2′-deoxy-4′-thioguanosine derivative 12, a key intermediate for the synthesis of the parent 3 (Scheme [2]). Thus, the glycal 11 [6] reacted with the TMS-7 in the presence of NIS to give the desired N ⁹-glycoside 12 in 11% isolated yield, along with the N ⁷-glycoside 13 (48% yield) and the N ¹-glycoside 14 (22% yield). The structures of products 12–14 were determined by comparison of their UV spectral data (λ_max and λ_min) with those of compounds 8–10 (Scheme [1] and 2), and on the basis of NOE experiments (Figure [4]).

Scheme 2 Electrophilic glycosidation of 11 with 14 to give the regioisomeric glycosides 12–14

Figure 4 NOE correlations of 12–14

In the Lewis acid-mediated glycosidation between a 4-thioribofuranose derivative and a purine base, it has been reported that N ⁷-4′-thiopurine ribonucleosides isomerize to the corresponding N ⁹-glycosides under heated reaction conditions.[6] [10] [11] To improve the low yield of the desired 12, a Lewis acid-promoted glycosidation using 15 [12] as a glycosyl donor was examined (Scheme [3]). Thus, when 15 reacted with the TMS-7 in the presence of TMSOTf, the target product 12 was obtained as the major regioisomer in 38% isolated yield, along with the N ⁹-isomer 13 (21% yield). In this glycosidation reaction, the formation of the N ¹-glycoside 14 was suppressed to a trace amount. Furthermore, by using N-(6-chloro-9H-purin-2-yl)-2-methylpropanamide (16)[13] as a basis, the target N ⁹-glycoside 17 was obtained in 87% isolated yield, along with its N ⁷-isomer 18 in 6% isolated yield. The depicted structures of 17 and 18 were determined on the basis of heteronuclear multiple bond correlation (HMBC) and NOE experiments (Figure [5]).[14]

Scheme 3 Lewis acid-mediated glycosidations of 15 with the purine bases 7 and 16

Figure 5 NOE correlations of 17 and 18

Scheme 4 Equilibrium experiment with the N ⁷-glycosides 13 and 18 and the N ⁹-glycosides 12 and 17

To clarify the reaction pathway of the Lewis acid-mediated glycosidation shown in Scheme [3], the N ⁷-glycosides 13 and 18 and the N ⁹-glycosides 12 and 17 were subjected to the above reaction conditions (Scheme [4]). The N ⁷-glycoside 13 gave a mixture of the isomerized N ⁹-glycoside 12 and recovered 13 in isolated yields of 30 and 33%, respectively, whereas N ⁷-glycoside 18 gave the N ⁹-glycoside 17 in a 75% yield along with recovered 18 in a 14% yield. The N ⁹-glycoside 12 gave a mixture of recovered 12 and the isomerized N ⁷-isomer 13 in yields of 45 and 24%, respectively. A similar result was obtained in the case of 17, which gave recovered 17 (85% yield) and the isomerized 18 (9% yield).

These results suggested that the kinetically controlled N ⁷-glycoside isomerizes to the corresponding thermodynamically controlled target N ⁹-2′-deoxy-4′-thiopurine ribonucleoside at ambient temperature due to the weaker glycosidic bond of the 2′-deoxynucleoside compared with that of the 4′-thioribonucleoside (Scheme [5]). At present, there is no clear explanation why the yield of 17 was higher than that of the corresponding N ⁹-glycoside 12.

Scheme 5 Proposed pathway for the Lewis acid mediated glycosidation between 15 and 7/16

This highly regio- and stereoselective glycosidation enabled us to develop a practical chemical synthesis leading to the β-anomer of the parent 3. Thus, 17 was subjected to Et₃B-initiated radical reduction to give 19 in 99% yield (Scheme [6]). Next, removal of the silyl-protecting group of 19, followed by the transformation of 19 into the corresponding guanine nucleoside with HOCH₂CH₂SH and NaOMe at 80 °C gave the β-anomer of 3 in 91% isolated yield over two steps. The ¹H NMR spectral data for the sample of 3 synthesized in this study were consistent with those reported in the literature.[7] This is the first example of a regio- and stereoselective chemical synthesis of the β-anomer of 3.

Scheme 6 Practical chemical synthetic pathway of the β-anomer of 3

With this successful synthesis of 3, a synthesis of the target molecule 4′-C-cyano 5 was performed (Scheme [7]). Glycosidation by the glycosyl donor 20 [12] of TMS-16 proceeded in an N ⁹-regioselective and β-stereoselective manner to furnish the desired glycoside 21 in 72% isolated yield, along with the regioisomer 22 (7% isolated yield).[15] The structures of 21 and 22 were confirmed by comparison of their UV spectra with those of 17 and 18, and by the NOE experiments shown in Figure [6]. Radical reduction of 21 gave 23 in 97% yield. When 23 was treated with NaOMe in MeOH at –20 °C, the 4′-C-(hydroxymethyl)-6-methoxypurine derivative 24 was obtained in 52% yield, with 42% recovery of 23. Oxidation of 24 with Dess–Martin periodinane, followed by the reaction of the resultant aldehyde 25 with HONH₂ gave oxime 26. Oxime 26 was treated with MsCl/Et₃N in CH₂Cl₂ to give the cyano derivative 27 in 70% yield from 24. Compound 27 was then converted into 28 in 53% in two steps: (1) TMSCl/NaI and (2) Bu₄NF/Ac₂O. Global deprotection of 28 gave the target molecule 5 in 88% yield.

Scheme 7 Synthesis of the target molecule 5

Figure 6 Representative NOE correlations of 21 and 22

An evaluation of the anti-HBV activity of the parent compound 3 and the 4′-C-cyano derivative 5 synthesized in this study was conducted with HepG2 2.2.15 cells transfected with the HBV genome.[16] As shown in Table [1, 5] exhibited about 100 times more potent anti-HBV activity than 3, with a comparative cytotoxicity to that of 3. Thus, the novel 4′-C-cyano nucleoside 5 was found to have a significantly better selectivity index (SI) than that of the parent nucleoside 3.

Table 1 Inhibitory Activities of 3 and 5 toward HBV and Their Cytotoxicity to Host Cells

Compound	EC50 a (μM)	CC50 HepG2.2.15b (μM)	SI
3 (R = H)	1.6	8.7	5.4
5 (R = C≡N)	0.015	6.7	446.7

^a Half-maximal effective concentration.

^b Half-maximal comparative cytotoxicity.

In conclusion, an N ⁹-regioselective and β-anomer-selective 4′-thioglycosidation of purine bases has been developed. The reaction of the 2-deoxy-2-iodo-4-thioribofuranosyl glycosyl donor 15 with the N-(6-chloro-9H-purin-2-yl)-2-methylpropanamide (16) gave the corresponding 2′-deoxy-4′-thiopurine nucleoside 17 in 87% yield, along with its N ⁷-regioisomer 18 (6% yield), without the formation of the α-anomers. By using 17, a practical chemical synthesis of 2′-deoxy-4′-thioguanosine (3) was developed for the first time. The target 4′-α-C-Cyano-2′-deoxy-4′-thioguanosine (5) was synthesized from 4-(acetoxymethyl)-2-deoxy-2-iodo-4-thioribofuranose (20) as a glycosyl donor.

Evaluation of the anti-HBV activity and the cytotoxicity toward the host cells revealed that the target molecule 5 exhibited about 100 times more potent anti-HBV activity than 3, with a comparative cytotoxicity to that of 3. The novel 4′-thioguanine nucleoside 5 was therefore found to demonstrate a better SI value than that of the parent 3. This finding could provide a guideline for the development of the next generation of anti-HBV agents.

Conflict of Interest

The authors declare no conflict of interest.

• References and Notes

• 1 Fung J, Lai C.-L, Seto W.-K, Yuen M.-F. J. Antimicrob. Chemother. 2011; 66: 2715
• 2 Yokoyama M. Synthesis 2000; 1637
• 3 Gunaga P, Moon H.-R, Choi W.-J, Shin D.-H, Park JG, Jeong LS. Curr. Med. Chem. 2004; 11: 2585
• 4 Mulamoottil VA, Majik MS, Chandra G, Jeong LS. In Chemical Synthesis of Nucleoside Analogues, Chap. 14. Merino P. Wiley; Hoboken: 2013. 655
• 5 Rodrigues L, Tilve SG, Majik MS. Eur. J. Med. Chem. 2021; 224: 113659
• 6 Haraguchi K, Kumamoto H, Tanaka H. Curr. Med. Chem. 2022; 29: 3684
• 7 Van Draanen NA, Freeman GA, Short SA, Harvey R, Jansen R, Szczech G, Koszalka G. J. Med. Chem. 1996; 39: 538
• 8 Takamatsu Y, Tanaka Y, Kohgo S, Murakami S, Singh K, Das D, Venzon DJ, Amano M, Higashi-Kuwata N, Aoki M, Delino NS, Hayashi S, Takahashi S, Sukenaga Y, Haraguchi K, Sarafianos SG, Maeda K, Mitsuya H. Hepatology (Baltimore, MD, U. S.) 2015; 62: 1024
• 9 Haraguchi K, Shimada H, Kimura K, Akutsu G, Tanaka H, Abe H, Hamasaki T, Baba M, Gullen EA, Dutschman GE, Cheng Y.-C, Balzarini Y. ACS Med. Chem. Lett. 2011; 2: 692
• 10 Naka T, Minakawa N, Abe H, Kaga D, Matsuda A. J. Am. Chem. Soc. 2000; 122: 7233
• 11 Haraguchi K, Matsui H, Takami S, Tanaka H. J. Org. Chem. 2009; 74: 2616
• 12 Haraguchi K, Kumamoto H, Konno K, Yagi H, Tatano Y, Odanaka Y, Shimbara Matsubayashi S, Snoeck R, Andrei G. Tetrahedron 2019; 75: 4542
• 13 Pitsch S, Wendeborn S, Krishnamurthy R, Holzner A, Minton M, Bolli M, Miculca C, Windhab N, Micura R, Stanek M, Jaun B, Eschenmoser A. Helv. Chim. Acta 2003; 86: 4270
• 14 Purines 17 and 18 N,O-Bis(trimethylsilyl)acetamide (0.22 mL, 0.90 mmol) was added to a suspension of 16 (215.7 mg, 0.90 mmol) in MeCN (5.0 mL) at rt under Ar, and the mixture was stirred for 1 h. A solution of 15 (360.4 mg, 0.60 mmol) in MeCN (7.0 mL)–CH₂Cl₂ (4.0 mL) and TMSOTf (0.22 mL, 1.20 mmol) were added sequentially at –10 °C under Ar. The resulting mixture was stirred at –10 °C for 1 h, 0 °C for 1 h, and r.t. overnight. The mixture was then partitioned between CHCl₃ and sat. aq NaHCO₃, and the organic layer was subjected to column chromatography [silica gel, hexane–EtOAc (7:1)] to give 17 as a colorless foam [yield: 387.6 mg (87%)] and 18 as a colorless syrup [yield: 21.4 mg (6%)]. 17 ¹H NMR (500 MHz, CDCl₃): δ = 0.86–0.90 and 1.07–1.14 (both m, 28 H, Si-i-Pr), 1.32 and 1.33 [both d, J _CH,_CH3 = 3.5 Hz, 6 H, CH(CH₃)₂], 3.41–3.42 [br s, 1 H, CH(CH₃)₂], 3.61 (dd, J _2′,_3′ = 4.6 and J _3′,4′ = 8.6 Hz, 1 H, H-3′), 3.76 (ddd, J _3′,_4′ = 8.6, J _4′,_5′a = 1.7 and J _4′,_5′b = 2.9 Hz, 1 H, H-4′), 4.08 (dd, J _4′,_5′a = 1.7 and J _5′a,_5′b = 12.6 Hz, 1 H, CH_2a-5′), 4.18 (dd, J _4′,_5′b = 2.9 and J _5′a,_5′b = 12.6 Hz, 1 H, CH_2b-5′), 4.73 (d, J _2′,_3′ = 4.6 Hz, 1 H, H-2′), 6.16 (s, 1 H, H-1′), 8.24 (br s, 1 H, NH), 8.66 (s, 1 H, H-8). NOE experiment: H-8/H-2′ and H-1′/H-4′. ¹³C NMR (125 MHz, CDCl₃): δ = 12.6, 13.1, 13.2, 13.3, 16.88, 16.93, 17.0, 17.31, 17.33, 17.5, 19.1, 19.2, 35.2, 37.7, 53.6, 58.2, 64.6, 71.9, 128.5, 143.3, 151.8, 152.0, 152.2, 176.8. HMBC experiment: C-4/H-1′. ESI-MS: m/z = 762 [M + Na]⁺. ESI-HRMS: m/z [M + Na]⁺ calcd for C₂₆H₄₃ClIN₅NaO₄SSi₂: 762.11997; found: 762.11896. UV (MeOH): λ_max 290, 263, 230 nm; λ_min 272 and 248 nm. 18 ¹H NMR (500 MHz, CDCl₃): δ = 0.79–0.85 and 1.07–1.20 (both m, 28 H, Si-i-Pr), 1.29 [d, J _CH,_CH3 = 6.9 Hz, 6 H, CH(CH₃)₂]. 3.12–3.25 [m, 1 H, CH(CH₃)₂], 3.52 (dd, J _2′,_3′ = 4.0 and J _3′,_4′ = 9.2 Hz, 1 H, H-3′), 3.74 (dd, J _3′,_4′ = 9.2 and J _4′,_5′b = 2.3 Hz, 1 H, H-4′), 4.13 (d, J _5′a,_5′b = 13.2 Hz, 1 H, CH_2a-5′), 4.21 (dd, J _4′,_5′b = 2.3 and J _5′a,_5′b = 13.2 Hz, 1 H, CH_2b-5′), 4.53 (d, J _2′,_3′ = 4.0 Hz, 1 H, H-2′), 6.40 (s, 1 H, H-1′), 8.13 (br s, 1 H, NH), 9.25 (s, 1 H, H-8). NOE experiment: H-8/ H-2′ and H-1′/H-4′. ¹³C NMR (125 MHz, CDCl₃): δ = 12.6, 13.1, 13.2, 13.3, 16.8, 16.88, 16.93, 17.27, 17.35, 17.5, 19.19, 19.22, 35.6, 39.4, 53.5, 57.9, 66.5, 71.0, 118.8, 143.2, 149.0, 152.7, 163.9, 176.2. HMBC experiment: C-5/H-1′. ESI-MS: m/z = 762 [M + Na]⁺. ESI-HRMS: m/z [M + Na]⁺ calcd for C₂₆H₄₃ClIN₅NaO₄SSi₂: 762.11997; found: 762.11788. UV (MeOH): λ_max 298 and 236 nm, λ_min 275 nm.
• 15 Purines 21 and 22 N,O-Bis(trimethylsilyl)acetamide (0.24 mL, 0.99 mmol) was added to a suspension of 16 (237.3 mg, 0.99 mmol) in MeCN (6.0 mL) at r.t. under Ar, and the mixture was stirred for 1 h. A solution of 20 (448.3 mg, 0.66 mmol) in MeCN (5.0 mL)–CH₂Cl₂ (3.0 mL) and TMSOTf (0.36 mL, 1.98 mmol) were added sequentially at –10 °C under Ar. The resulting mixture was stirred at –10 °C for 1 h, 0 °C for 1 h, and r.t. overnight. The mixture was partitioned between CHCl₃ and sat. aq NaHCO₃, and the organic layer was subjected to column chromatography [silica gel, hexane–EtOAc (4:1 to 2:1)] to give 21 as a colorless foam [yield: 387.1 mg, (72%)] and 22 as a colorless syrup [yield: 37.2 mg (7%)]. 21 ¹H NMR (500 MHz, CDCl₃): δ = 0.90–0.97 and 1.03–1.12 (both m, 28 H, Si-i-Pr), 1.29 and 1.30 [both d, J _CH,_Me = 3.5 Hz, 6 H, CH(CH₃)₂], 2.16 (s, 3 H, Ac), 3.26–3.27 [br s, 1 H, CH(CH₃)₂], 4.03 (d, J _5′a,_5′b = 12.0 Hz, 1 H, CH_2a-5′), 4.10 (d, J _5′a,_5′b = 12.0 Hz, 1 H, CH_2b-5′), 4.30 (d, J _2′,_3′ = 6.5 Hz, 1 H, H-3′), 4.57 (d, J _CH2a,_CH2b = 9.8 Hz, 1 H, CH_2aOAc), 5.15 (d, J _CH2a,_CH2b = 9.8 Hz, 1 H, CH_2bOAc), 5.34 (dd, J _1′,_2′ = 1.7 and J _2′,_3′ = 6.5 Hz, 1 H, H-2′), 6.24 (d, J _{1′ 2′} = 1.7 Hz, 1 H, H-1′), 8.12 (br s, 1 H, NH), 8.41 (s, 1 H, H-8). NOE experiment: H-8/H-2′, H-8/H-3′, H-8/CH_2a-5′, H-8/CH_2b-5′, H-1′/CH_2b-OAc, H-2′/CH_2a-5′, H-2′/CH_2b-5′. ¹³C NMR (125 MHz, CDCl₃): δ = 12.7, 12.9, 13.0, 13.1, 17.0, 17.1, 17.2, 17.3, 17.4, 19.1, 19.2, 21.0, 35.5, 36.0, 63.18, 65.0, 66.0, 67.0, 76.2, 128.7, 143.0, 151.91, 151.98, 152.2, 170.6, 176.1. ESI-MS: m/z = 834 [M + Na]⁺. FAB-HRMS: m/z [M + Na]⁺ calcd for C₂₉H₄₇ClIN₅NaO₆SSi₂: 834.14110 found: 834.13980. UV(MeOH) λ_max 290, 263 and 230 nm, λ_min 272 and 247 nm. 22 ¹H NMR (500 MHz, CDCl₃): δ = 0.83–0.90 and 1.01–1.13 (both m, 28 H, Si-i-Pr), 1.29 [d, J _CH,_CH3 = 6.9 Hz, 6 H, CH(CH₃)₂], 2.14 (s, 3 H, Ac), 3.09–3.16. [br s, 1 H, CH(CH₃)₂], 4.09 (d, J _5′a,_5′b = 12.1 Hz, 1 H, CH_2a-5′), 4.13 (d, J _5′a,_5′b = 12.1 Hz, 1 H, CH_2b-5′), 4.16 (d, J _2′,_3′ = 6.3 Hz, 1 H, H-3′), 4.53 (d, J _CH2a,_CH2b = 11.5 Hz, 1 H, CH_2aOAc), 4.75 (d, J _2′,_3′ = 6.3 Hz, 1 H, H-2’), 5.28 (dd, J _CH2a,_CH2b = 11.5 Hz, 1 H, CH_2bOAc), 6.47 (s, 1 H, H-1′), 8.14 (br s, 1 H, NH), 9.04 (s, 1 H, H-8). NOE experiment: H-8/H-2′, H-8/H-3′, H-8/CH_2b-5′, H-1′/CH_2b-OAc. ¹³C NMR (125 MHz, CDCl₃): δ = 12.86, 12.90, 13.0, 13.2, 17.0, 17.06, 17.09, 17.2, 17.3, 17.4, 19.2, 20.9, 35.7, 37.2, 62.5, 66.0, 74.9, 118.8, 143.4, 148.1, 152.7, 163.9, 170.4, 175.9. ESI-MS: m/z = 834 [M⁺ + Na]. FAB-HRMS: m/z [M + Na]⁺ calcd for C₂₉H₄₇ClIN₅NaO₆SSi₂: 834.14110; found: 834.13938. UV(MeOH) λ_max 299 and 237 nm, λ_min 275 nm.
• 16 Sells MA, Chen ML, Acs G. Proc. Natl. Acad. Sci. U.S.A. 1987; 84: 1005

Publikationsverlauf

Eingereicht: 29. Juli 2023

Angenommen nach Revision: 30. Oktober 2023

Artikel online veröffentlicht:
05. Dezember 2023