An investigation on stereospecific fluorination at the 2′-arabino-position of a pyrimidine nucleoside: radiosynthesis of 2′-deoxy-2′-[¹⁸F]fluoro-5-methyl-1-β-D-arabinofuranosyluracil

Abstract

Direct fluorination at the 2′-arabino-position of a pyrimidine nucleoside has been a long-standing challenge, yet we recently reported such a stereospecific fluorination for the first time in the synthesis of [¹⁸F]FMAU, albeit in low yields. Herein we report the results of an investigation on stereospecific fluorination on a variety of precursors for synthesis of [¹⁸F]FMAU. Several precursors were synthesized in multiple steps and fluorination was performed at the 2′-arabino position using K[¹⁸F]/kryptofix 2.2.2. All precursors produced [¹⁸F]FMAU in low yields.

1. Introduction

2′-Deoxy-2′-fluoro-5-substituted-1-β-D-arabinofuranosyluracils are biologically important analogues of thymidine, because of their anticancer ¹ and antiviral properties.^2–4 Therefore, many attempts have been made to develop facile methods for single-step fluorination of various protected pyrimidine nucleoside precursors.^5–7 In an attempt to prepare 2′-iodo-arabinothymidine from 5′-trityl-2′-tosyl-5-methyluridine by reaction with sodium iodide, an intermediate 2,2′-anhydronucleoside product was formed.^5,6 Upon further heating of the crude reaction mixture from this preparation the 2′-iodo-ribofuranose was produced instead of an arabino-derivative. Fox and Miller⁶ and Codington et al.⁷ explained that the conversion of the 2′-tosyloxy derivative to its 2′-iodo-ribo analogue via the 2,2′-anhydro intermediate was catalyzed by the presence of a small amount of p-toluenesulfonic acid (TSOH) liberated during the formation of the 2,2′-anhydro intermediate. Thus, direct nucleophilic substitution (S_N2-type) reactions with inversion of configuration at the 2′-position of a pyrimidine were not possible due to the formation of an intermediate 2,2′-anhydronucleoside.^5,6–9 It was reported that the direct introduction of a fluoro-group in the 2′-up (arabino) position from a preformed nucleoside would be “difficult, if not impossible” because of neighboring-group participation of the carbonyl oxygen at C₂-position of the pyrimidine moiety.¹⁰ Therefore, the synthesis of these 2′-arabino-fluoro-pyrimidine nucleoside analogues was developed using a multistep methodology, such as stereospecific fluorination of 1,3,5-tri-O-benzoyl-α-D-ribofuranose-2-sulfonate ester to produce 1,3,5-tri-O-benzoyl-β-D-2-fluoro-arabinofuranose, bromination of the 1,3,5-tri-O-benzoyl-β-D-2-fluoro-arabinosugar at the C₁-postion, and then coupling of the 1,3,5-tri-O-benzoyl-β-D-2-fluoro-1-bromo-arabinofuranose with pyrimidine-bis-trimethylsilyl ether to produce the protected pyrimidine nucleoside analogue. Finally hydrolysis of the protecting groups with a strong base and purification produced the desired 2′-fluoro-1-β-D-arabino-pyrimidine nucleoside.^2,4,11–13

Because of the anticancer and antiviral properties, some of these fluorinated pyrimidine nucleoside analogues have been radiolabeled with ¹⁸F for positron emission tomography (PET) of tumor proliferation^14–16 and herpes simplex virus type 1-thymidine kinase (HSV1-tk) reporter gene expression.^17–22 Alauddin et al. developed ¹⁸F-labeled 2′-fluoro-arabinofuranosyluracil derivatives, such as 2′-deoxy-2′-[¹⁸F]fluoro-5-methyl-1-β-D-arabinofuranosyluracil ([¹⁸F]FMAU) for the first time²³ and some other 2′-deoxy-2′-[¹⁸F]fluoro-5-substituted-1-β-D-arabinofuranosyluracil derivatives,²⁴ by modification of the multistep process, such as radiofluorination of 1,3,5-tri-O-benzoyl-α-D-ribofuranose-2-trifluoromethylsulfonate ester, followed by bromination of the radiolabeled sugar and coupling of the radiolabeled 1-bromosugar with pyrimidine-bis-trimethylsilyl ether. The coupled product was hydrolyzed and purified by high-performance liquid chromatography (HPLC). Other researchers have also reported the radiosynthesis of these compounds using the same methodology.²⁵ This multi-step synthetic method is currently used for synthesizing [¹⁸F]FMAU and other 5-substituted analogues for pre-clinical and clinical applications;^{14–17, 20–22, 26, 27} however, this method requires multiple steps after radiofluorination of the sugar and cannot be applied to a commercial automated synthesis module for routine production. Therefore, there is a need for one-step stereospecific fluorination at the 2′-arabino position of the intact nucleoside, especially for radiosynthesis of ¹⁸F-labeled compounds using automated synthesis module.

Direct S_N2-type fluorination reactions with inversion of configuration at the 2′-position of the sugar moiety in a purine nucleoside have been successful;²⁸ however, for pyrimidine nucleoside, these S_N2 type reactions have seldom been performed or attempted because of neighboring-group participation of the carbonyl oxygen at C₂-position of the pyrimidine moiety. The assumption that substitution of the N³-position with a suitable protecting group would prevent the formation of a 2,2′-anhydro compound and therefore increase the likelihood of an S_N2 reaction at the 2′-position is reasonable. Surprisingly, this strategy has rarely been used, and only a few attempts have been made to stereospecifically substitute at the 2′-arabino position on a pyrimidine nucleoside.^29–31 In some instances, the N³-position was protected by substituting the hydrogen with an electron withdrawing group, such as benzoate;³⁰ however, subsequent substitution of the 2′-hydroxyl group with a triflate was not successful.³¹ In the other instance the N³-proton was substituted with a nitro group and successfully prepared a 2′-triflate using a onestep reaction.²⁹ Thus, an N³-nitro-3′,5′-O-1,1,3,3-tetraisopropyl-1,3-disiloxane-2′-triflate was prepared and the precursor was reacted with tetrabutylammonium halides (Bu₄NX); the halogens were iodide, chloride, and bromide. The corresponding 2′-halo-arabinopyrimidine nucleosides were obtained in reasonably good yields. This was the first demonstration of an S_N2 substitution at the 2′-position of an intact pyrimidine nucleoside with the arabino configuration; however, in this methodology no 2′-arabino-fluorinated compound has been reported.

The 4′-thioanalogue of the pyrimidine nucleoside underwent fluorinated at the 2′-position with an arabino configuration using diethylaminosulfur trifluoride, whereby the 4′-sulfur behaved differently than the 4′-oxygen. In this reaction, sulfur took an active part during fluorination and assisted to produce the desired 2′-arabino fluoro-compound.³² The synthesis of a 4′-seleno-2′-fluoro-arabinofuranosyluracil has also been reported,³³ in which selenium played a role similar to that of the sulfur in 4′-thio-pyrimidine. Most recently, we reported a synthesis of 2′-arabino-fluoro-derivative ([¹⁸F]FMAU) for the 4′-oxo-nucleoside by stereospecific fluorination.³⁴ In this method the N³-proton was substituted with t-butyloxycarbonyl (Boc) and a precursor compound, 2′-methylsulfonyl-3′,5′-O-tetrahydropyranyl-N³-Boc-5-methyl-1-β-D-ribofuranosil-uracil, was synthesized in multiple steps. Radiofluorination of this precursor was performed with K[¹⁸F]/kryptofix 2.2.2. to produce 2′-deoxy-2′-[¹⁸F]fluoro-3′,5′-O-tetrahydropyranyl-N³-Boc-5-methyl-1-β-D-arabinofuranosyluracil. Acid hydrolysis of the intermediate compound produced the desired product [¹⁸F]FMAU, which was purification by HPLC. The average radiochemical yield was 2.0% (decay corrected, d. c.) from the end of bombardment. To improve the yield of this fluorination reaction, we have synthesized several new precursor compounds with two different leaving groups and various protecting groups at the 3′- and 5′-postions, and tested them for direct fluorination at the 2′-position to synthesize [¹⁸F]FMAU. In this article, we report syntheses of those new precursors and results of direct fluorination of the intact pyrimidine nucleoside precursors at the 2′-carbon with an arabino-configuration under various reaction conditions.

2. Results

Figure 1 shows the synthetic scheme for preparation of the precursor compounds and radiosynthesis of [¹⁸F]FMAU 9. Chemical yields for compounds 2, 3, 4, 5, 6a, 7a and 8a were comparable as reported previously.³⁴ Compound 6b was obtained in 70% yield, which was lower than that of 6a (90%). Hydrolysis of compounds 6a and 6b produced 7a and 7b in 97% and 70% yields, respectively. The yield of compound 7b was also lower than that of 7a. The 3′- and 5′-hydroxyl groups of compounds 7a and 7b were protected using the appropriate protecting groups, such as acetate, benzoate, methoxy methyl (Mom) and tetrahydropyranyl (THP) groups, and the products 8b–d, 8e, and 8f–g were obtained in 60%, 57% and 50% yields, respectively. Radiofluorination of compounds 8a–g using K¹⁸F/kryptofix 2.2.2. produced the desired product [¹⁸F]FMAU in very low yields (Table 1).

Figure 1.

Scheme for preparation of the precursors 8a–g and radiosynthesis of [¹⁸F]FMAU.

Table 1.

Results of radiofluorination reactions on various precursor compounds 8a–g under various conditions,.

Experiment #	Compound #	Temperature (20 min)	Yield	Average yield
1	8a	90 °C	2.0%
2	8a	85 °C	1.2%	1.56%
3	8a	100 °C	1.5%
4	8b	100 °C	0.9%
5	8b	90 °C	0.5%	0.60%
6	8b	90 °C	0.4%
7	8c	90 °C	0.1
8	8c	90 °C	0.11%a	0.08%
9	8c	90 °C	0.08%a
10	8d	90 °C	0.06%a
11	8d	90 °C	0.2%a	0.22%
12	8d	90 °C	0.4%a
13	8e	90 °C	0.4%a
14	8e	90 °C	0.9%a	0.63%
15	8e	95 °C	0.6%a
16	8f	90 °C	1.00%
17	8f	90 °C	0.92%	1.00%
18	8f	100 °C	1.08%
19	8g	90 °C	0.9%a
20	8g	90 °C	1.8%	1.10%
21	8g	90 °C	0.8%a

^aThe yield of [¹⁸F]FMAU was calculated from the analytical HPLC chromatograms of the crude reaction mixtures.

Table 1 represents the results of radiofluorination reactions with radiochemical yields of [¹⁸F]FMAU (9) from all precursors 8a–g at various temperatures. All precursors produced the desired product [¹⁸F]FMAU in variable yields with an average ranging from 0.08%–2.00%. At 90°C, 8a produced the highest yield of 2.0% (d. c.) with radiochemical purity > 99% and specific activity >1500 mCi/μmol. The synthesis time was 95–100 min from the end of bombardment.

Figure 2 represents a quality-control analysis of [¹⁸F]FMAU using HPLC, which shows that the product (radioactive trace) was co-eluted with standard FMAU (UV trace).

Figure 2.

HPLC chromatogram of [¹⁸F]FMAU co-injected with standard FMAU; analytical C₁₈ column, 9% MeCN/water, flow 1mL/min.

3. Discussion

Stereospecific fluorination at the 2′-arabino-position of a pyrimidine nucleoside has been a longstanding challenge. We recently reported such a stereospecific fluorination for the first time in the synthesis of [¹⁸F]FMAU in low yields. To improve the yield of the fluorination reaction we have investigated several precursor compounds through their synthesis and application in radiofluorination reactions. Syntheses of the precursor compounds 8a–g (Fig. 1) involved a sequence of protection, deprotection and derivatization. Protection of 4′- and 5′-hydroxyl groups of 5-methyluridine 1 produced compound 2; then protection of 2′-hydroxyl group with trimethylsilane (TMS) produced compound 3, and finally protection of the NH proton with Boc produced compound 4. Deprotection of the 2′-hydroxyl group of 4 produced compound 5, which by derivatization with mesylate and nosylate produced compounds 6a and 6b, respectively. Deprotection of 4′- and 5′-hydroxyl groups of 6a and 6b produced compounds 7a and 7b, which were then derivatized with a variety of protecting groups to obtain compounds 8a–g. The purpose of protecting the NH proton was to prevent its participation to form the 2,2′-anhydro-compound during fluorination as previously reported.⁵

Previously, we reported protection of the 2′-hydroxyl group by TMS followed by protection of the NH proton with Boc,³⁴ because our attempt to protect the NH proton after protection of the 2′-hydroxyl group with mesylate was not successful. However, now we have successfully protected the NH proton after protecting the 2′-OH with an electronegative group such as mesylate. Figure 3 shows the scheme for preparation of compound 6a from compound 2 via 10. This is a significant improvement over the previously reported synthesis of 6a, because this method involves only two-steps, which is three-steps shorter than the previous one, and both steps produced very high yields, 93% and 90% for compounds 10 and 6a, respectively. However, an attempt to adapt this synthetic methodology to prepare the 2′-nosylate precursor 6b from compound 2 was not successful.

Figure 3.

Scheme for alternative synthesis of 6a

The yield of compound 6b from 5 (Fig. 1) was lower (70%) than that of 6a (90%). ¹H NMR spectrum of 6b showed characteristic peaks in the aromatic region due to the nosylate group in addition to the other peaks, which were consistent with the identity and structure of the compound 6b, and the compound was further characterized by high-resolution mass spectrometry (HRMS). Hydrolysis of the siloxane group from 6b also produced lower yield (70%) of 7b compared with that of 7a (97%). This lower yield of 7b may be due to higher reactivity of nosylate compared with that of mesylate. During the preparation and work-up of 7b, an unidentified by-product was obtained in variable yields (15%–20%) from run to run. The unknown compound was less polar than 7b as observed by thin-layer chromatography (TLC). The ¹H NMR spectrum of the unknown compound clearly showed the absence of the peak at 8.35 ppm for the 3,5-protons next to the nitro-group of the 4-nitro-phenylsulphonyl moiety; and reappearance of a new peak in the higher field at 8.10 ppm, which clearly indicates the absence of the nitro-group. Unfortunately, the mass spectrum of the unknown compound could not yield the molecular ion or any reasonable ion to interpret; therefore, the compound could not be completely identified. Due to the formation of this unknown by-product, the yield of the desired product 7b was much lower than that of the mesylate 7a.

The 4′- and 5′-hydroxyl groups of 7a and 7b were protected with various protecting groups to produce compounds 8a–g. The THP group was inserted into 7a and 7b to produce 8a and 8e according to a previously reported method³⁴ using a catalytic amount of TsOH. Compound 8e was obtained in slightly higher yield (57%) compared with that of 8a (48%). However, these yields were low in general for both compounds compared with the other steps. One of the reasons for the lower yields of 8a and 8e was incomplete formation of the disubstituted THP-ethers; a small amount of the monosubstituted THP-ether was also isolated during this preparation, as a result 8a and 8e were produced in lower yields. Furthermore, low yield may be due to partial hydrolysis of the N-Boc by TsOH used during the preparation of THP-ethers. We have synthesized precursors 8b and 8f with another protecting group, Mom, at the 4′- and 5′-positions. In these preparations, yield of 8f was lower (50%) than that of 8b (60%), probably due to the more reactive nosylate group at the 2′-position of 8f compared with the mesylate in 8b. The benzoate protected precursors 8c and 8g were obtained in reasonably good yields, 60% and 50%, respectively. Similarly, compound 8d with the mesylate group was obtained in higher yield (60%) than that of the nosylate. In general, compounds containing the nosylate group had lower yields in the subsequent steps compared with the mesylate group, probably due to the difference in reactivity between these two leaving groups.

In this study, we investigated stereospecific fluorination of intact pyrimidine nucleoside towards the radiosynthesis of [¹⁸F]FMAU using two different leaving groups, mesylate and nosylate, and various other protecting groups at the 4′- and 5′-positions. We anticipated that nosylate would be more reactive than mesylate and would therefore provide better yields of [¹⁸F]FMAU. However, no significant improvement in fluorination was achieved using these nosylate-containing precursors. In some instances (expt. # 16–21), nosylate-containing precursors produced relatively better yields, but lower than 8a. Moreover, the precursors containing nosylate produced several radioactive by-products after fluorination as observed by HPLC, and these were not related to the nucleoside; rather, one of this product was p-nitrophenylsulfonyl-¹⁸F-fluoride, which was verified by co-injection with the standard compound in HPLC. The precursors containing mesylate produced a clean desired product and appeared to be better than those with nosylate in stereospecific fluorination of pyrimidine nucleoside. The most reactive leaving group known, triflate, could not be prepared for this study.

With respect to the other protecting groups at the 4′- and 5′-position, we anticipated that the THP group might have steric hindrance because of its ring size when folded over the sugar ring and prevents the incoming fluoride ion from the same side of the sugar; therefore, selected a smaller and linear protecting group Mom. However, Mom did not show any significant improvement of the yield and had a similar effect in terms of product formation. Changing the leaving group from mesylate to nosylate in these Mom-containing precursors led to a similar yield. We also used another type of protecting groups, esters such as acetate and benzoate, at the 4′- and 5′-position. In both protecting groups, neither acetate nor benzoate with either mesylate or nosylate improved the yield.

For hydrolysis of the protecting groups from the crude products, two methods were used depending on the protecting groups. For precursors 8a, 8b, 8e, and 8f, hydrolysis was performed using HCl in MeOH and all protecting groups were hydrolyzed in a single step. For precursors 8c, 8d, and 8g, hydrolysis required two steps: acid hydrolysis of the N-Boc using trifluoroacetic acid followed by basic hydrolysis of the esters using Na-OMe. Thus, precursors 8a, 8b, and 8e have the benefit of one-step hydrolysis that saves time in radiosynthesis. Compound 8a appeared to be the best precursor in the synthesis of [¹⁸F]FMAU as reported earlier.³⁴ This precursor produced higher yields and requires one-step hydrolysis of the protecting groups; therefore, this precursor may be used for routine production of the compound in an automated synthesis module. However, further studies, especially those using a triflate, as leaving group, are needed to improve the yields.

All precursors produced the desired product in variable low yields. The reactions that produced yields > 1% were purified by HPLC using a semipreparative column and reported as isolated yields, and the reactions with yield < 1% were calculated from the analytical HPLC chromatograms of the crude reaction mixtures. Reactions in MeCN at 90°C for 20 min appeared to be optimal. This synthetic method with precursor 8a using an automated synthesis module may be suitable in production of [¹⁸F]FMAU for clinical application.

4. Summary

Several precursor compounds of intact pyrimidine nucleoside have been investigated for stereospecific fluorination at the 2′-arabino position. Out of seven precursors, all produced the desired product in variable low yields. Compound 8a appeared to be the best precursor in terms of product yield and simplicity of its synthesis. Precursors like 8a and this methodology should be applicable for radiosynthesis of the other 2′-deoxy-2′-fluoro-5-substituted-1-β-D-arabinofuranosyluracil analogues, including [¹⁸F]FEAU, [¹⁸F]FIAU, [¹⁸F]FFAU, [¹⁸F]FCAU, and [¹⁸F]FBAU for PET imaging.

5. Experimental

5.1. Reagents and Instrumentation

All reagents and solvents were purchased from Aldrich Chemical Co. (Milwaukee, WI), and used without further purification. Solid-phase extraction cartridges (Sep-Pak, silica gel, 900 mg) were purchased from Alltech Associates (Deerfield, IL). FMAU was prepared in-house for the HPLC standard.

TLC was performed on pre-coated Kieselgel 60 F254 (Merck, Darmstadt, Germany) glass plates. Proton and ¹⁹F NMR spectra were recorded on a Bruker 300 MHz spectrometer with tetramethylsilane used as an internal reference and hexafluorobenzene as an external reference, and ¹³C NMR spectra were recorded on a Bruker 600 MHz spectrometer at The University of Texas M. D. Anderson Cancer Center. HRMS were obtained on a Bruker BioTOF II mass spectrometer at the University of Minnesota using the electrospray ionization technique.

HPLC was performed with an 1100 series pump (Agilent Technologies, Stuttgart, Germany) with a built-in UV detector operated at 254 nm and a radioactivity detector with a single-channel analyzer (Bioscan, Washington, DC), using a semipreparative C₁₈ reverse-phase column (Alltech, Econosil, 10×250 mm) and an analytical C₁₈ column (Alltech, Econosil, 4.6×250 mm). An acetonitrile/water (MeCN/H₂O) solvent system (9% MeCN/H₂O) was used for purification of the radiolabeled product at a flow of 4 mL/min. Quality control analyses were performed on an analytical HPLC column with the same solvents at a flow of 1 mL/min.

5.2. Methods

5.2.1

Compounds 1–5, 6a, 7a, and 8a, were synthesized as described previously.³⁴

5.2.2. Preparation of N³-Boc-3′,5′-O-1,1,3,3-tetraisopropyl-1,3-disiloxane-2′-O-p-nitrophenylsulfonyl-5-methyluridine 6b

Compound 5 (0.50 g, 0.83 mmol) was dissolved in dichloromethane (5.0 mL) and pyridine (5.0 mL) was added, followed by the addition of 4-nitrobenzensulfonyl chloride (0.37 mL, 2 mmol). The mixture was stirred at room temperature (RT) for 12 h. The reaction mixture was filtered, and the solvent was removed under reduced pressure to render oil that was purified by flash chromatography on a silica gel column using 20% ethyl acetate (EtOAc) in hexane to give 6b as white foam in 90% yield. ¹H NMR (CDCl₃) δ: 8.35 (d, J=9.0 Hz, 2H, aromatic, C₃H & C₅H), 8.22 (d, J=9.0 Hz, 2H, aromatic, C₂H & C₆H), 7.53 (s, 1H, C₆H), 5.58 (s, 1H, 1′H), 5.15 (d, 1H, J=4.5 Hz, 2′H), 4.40 (m, 1H, 3′H), 4.20 (m, 1H, 4′H), 4.11 (m, 1H, 5′H), 4.03 (m, 1H, 5′H), 1.92 (s, 3H, CH₃), 1.63 (s, 9H, t-But), 1.06–1.09 (m, 28H, iso-Pr). ¹³C NMR (CDCl₃) δ: 161.09, 150.73, 147.78, 147.56, 142.90, 133.73, 129.40, 124.31, 110.62, 88.68, 87.21, 84.00, 81.93, 66.82, 58.87, 27.41, 17.38, 17.32, 17.21, 16.92, 16.82, 16.81, 13.55, 12.86, 12.79, 12.73, 12.61. HRMS (m/z): [M+Na] for C₃₃H₅₁N₃O₁₃SSi₂, calculated, 808.2579; found, 808.2567.

5.2.3. Preparation of N³-Boc-2′-O- p-nitrophenylsulfonyl-5-methyluridine 7b

A solution of 6b (0.27 g, 0.4 mmol) in THF (10 mL) was cooled to 0°C and n-Bu₄NF (1M in THF, 0.5 mL, 0.5 mmol) was added. The reaction mixture was stirred at RT for 15 min, when TLC showed no starting material remained. The reaction mixture was concentrated under vacuum and the residue purified by flash chromatography on a silica gel column eluted with 10% methanol in dichloromethane to give 7b as white foam in 70% yield. Based on ¹H NMR spectrum, the compound was > 98% pure. ¹H NMR (CDCl₃) δ: 8.35 (d, J=9.0 Hz, 2H, aromatic, C₃H & C₅H), 8.22 (d, J=9.0 Hz, 2H, aromatic, C₂H & C₆H), 7.49 (s, 1H, C₆H) 5.89 (d, 1H, J=4.5 Hz,1′H), 5.27 (t, 1H, J=4.5 Hz, 2′H), 4.60 (t, 1H, J=5.2 Hz, 3′H), 4.20 (m, 1H, 4′H), 4.08-3.83 (m, 2H, 5′H), 2.20–2.80 (broad s, 2H, OH), 1.95 (s, 3H, CH₃), 1.62 (s, 9H, t-But). ¹³C NMR (CDCl₃) δ: 161.09, 150.73, 147.78, 147.56, 142.90, 133.73, 129.40, 124.31, 110.62, 88.62, 87.21, 84.00, 81.93, 66.82, 58.87, 27.14, 12.61. HRMS: (m/z): [M+Na] for C₂₁H₂₅N₃O₁₂S, calculated, 566.1051; found, 566.1046.

5.2.4. Preparation of N³-Boc-3′,5′-O-bis-methoxymethyl-2′-O-methylsulfonyloxy-5-methyluridine 8b

To a solution of compound 7a (100.0 mg, 0.23 mmol) in dry CH₂Cl₂ (5 mL), chloromethyl methyl ether (0.5 mL, 36 mmol) and N,N-diisopropylethylamine (0.5 mL) were added. The reaction mixture was stirred at RT for 16 h, and the solvent was evaporated. The residue was purified by flash chromatography on a silica gel column using 30% EtOAc in hexane to obtain 8b as white foam in 60% yield. Based on HPLC chromatogram, the compound was > 98% pure. ¹H NMR (CDCl₃) δ: 7.49 (s, 1H, C₆H) 5.89 (d, 1H, J=4.5 Hz,1′H), 5.27 (t, 1H, J=4.5 Hz, 2′H), 4.75 (s, 4H, methylene), 4.60 (t, 1H, J=5.2 Hz, 3′H), 4.20 (m, 1H, 4′H), 4.08-3.83 (m, 2H, 5′H), 3.44 (s, 6H, OMe), 3.23, (s, 3H, Ms), 1.95 (s, 3H, CH₃), 1.62 (s, 9H, t-But). ¹³C NMR (CDCl₃) δ: 161.20, 148.10, 147.63, 133.77, 110.56, 89.03, 87.23, 84.86, 84.00, 82.86, 81.92, 66.74, 58.90, 57.56, 57.50, 39.21, 27.47, 12.64. HRMS (m/z): [M+H] for C₂₁H₃₅N₂O₁₂S, calculated, 437.1224; found, 437.1222.

5.2.5. Preparation of N³-Boc-3′,5′-O-bis-benzoyl-2′-O-methylsulfonyloxy-5-methyluridine 8c

Compound 7a (100.0 mg, 0.23 mmol) was dissolved in dry THF (5 mL), triethylamine (0.5 mL) and benzoyl chloride (0.5 mL, 46 mmol) were added. The reaction mixture was stirred at RT for 10 h. The reaction mixture was filtered and the solvent was evaporated. The residue was purified by flash chromatography on a silica gel column using 30% EtOAc in hexane to give 8c as white foam in 60% yield. Based on HPLC chromatogram, the compound was > 98% pure. ¹H NMR (CDCl₃) δ: 7.93 (m, 4H, aromatic), 7.63 (m, 2H, aromatic), 7.49 (s, 1H, C₆H), 7.45 (m, 4H, aromatic), 5.89 (d, 1H, J=2.7 Hz,1′H), 5.27 (dd, 1H, J=5.7 Hz and 7.2 Hz, 3′H), 4.60 (dd, 1H, J=2.7 Hz and 5.7 Hz, 2′H), 4.84 (dd, 1H, J=2.7 Hz, 12.7 Hz), 4.72-4.67 (m, 1H, 4′H), 4.57 (dd, 1H, J= 2.7 Hz, 12.7 Hz, 5′H), 3.10, (s, 3H, Ms), 1.66 (s, 3H, CH₃), 1.62 (s, 9H, t-But). ¹³C NMR (CDCl₃) δ: 165.91, 165.42, 161.00, 148.20, 147.37, 133.77, 133.63, 133.41, 129.93, 129.86, 129.67, 129.63, 128.63, 128.60, 128.49, 110.56, 89.03, 87.23, 82.86, 81.92, 66.74, 58.90, 39.21, 27.43, 12.64. HRMS (m/z): [M+Na] for C₃₀H₃₂N₂O₁₂S, calculated, 667.1575; found, 667.1579.

5.2.6. Preparation of N³-Boc-3′,5′-O-bis-acetyl-2′-O-methylsulfonyloxy-5-methyluridine 8d

To a solution of compound 7a (100.0 mg, 0.23 mmol) in dry pyridine (5 mL) was added acetic anhydride (0.1 mL, 1.15 mmol). The reaction mixture was stirred at RT for 8 h and the solvent was evaporated. The residue was purified by flash chromatography on a silica gel column using 30% EtOAc in hexane to give 8d as white foam in 60% yield. ¹H NMR (CDCl₃) δ: 7.49 (s, 1H, C₆H) 5.89 (d, 1H, J=4.5 Hz, 1′H), 5.27 (t, 1H, J=4.5 Hz, 2′H), 4.60 (t, 1H, J=5.2 Hz, 3′H), 4.20 (m, 1H, 4′H), 4.08-3.83 (m, 2H, 5′H), 3.23, (s, 3H, Ms), 2.21 (s, 3H, acetyl), 1.95 (s, 3H, CH₃), 1.62 (s, 9H, t-But). ¹³C NMR (CDCl₃) δ: 170.16, 169.89, 161.00, 148.20, 147.39, 135.11, 110.93, 91.22, 87.22, 79.14, 78.83, 68.40, 61.63, 38.35, 27.43, 20.72, 20.49, 12.65. HRMS (m/z): [M+H] for C₂₁H₃₅N₂O₁₂S, calculated, 437.1224; found, 437.1222.

5.2.6. Preparation of N³-Boc-3′,5′-O-bis-tetrahydropyranyl-2′-O-p-nitrophenylsulfonyl-5-methyluridine 8e

Compound 7b (50.0mg, 0.09 mmol) was dissolved in dry THF (5 mL), a catalytic amount of TsOH (10.0 mg) and 3,4-dihydro-2H-pyrane (DHP, 0.20 mL, 2.2 mmol) were added. The reaction mixture was stirred at RT for 16 h, and then neutralized by the addition of triethylamine (30.0 μL), and the solvent was evaporated. The residue was purified by flash chromatography on a silica gel column using 30% EtOAc in hexane to give 8e (a mixture of diasteriomers) as a colorless oil in 57% yield. ¹H NMR (CDCl₃) δ: 8.35 (d, J=9.0 Hz, 2H, aromatic, C₃H & C₅H), 8.22 (d, J=9.0 Hz, 2H, aromatic, C₂H & C₆H), 7.79, 7.78, 7.68, 7.67 (4s, 1H, C₆H), 6.06, 5.98 (2d, J=3.3Hz, 2.4Hz, 1H, 1′H), 4.55-4.47 (m, 1H, 2′-H), 4.72-3.54 (m, 9H, 3′-5′H and THP), 1.96, 1.92 (2s, 3H, CH₃), 1.83-1.70 (m, 4H, THP), 1.62 (m, 9H, t-But), 1.63-1.58 (m, 8H, THP). ¹³C NMR (CDCl₃) δ: 161.09, 150.73, 147.78, 147.56, 142.90, 133.73, 129.40, 124.31, 110.62, 99.93, 98.80, 97.98, 97.37, 88.68, 87.21, 85.31, 85.65, 84.65, 84.09, 83.69, 81.93, 76.93, 66.82, 62.42, 58.87, 39.24, 38.00, 30.70, 27.41, 25.26, 25.23, 19.31, 19.09, 12.39. HRMS (m/z): [M+Na] for C₃₁H₄₁N₃O₁₄S, calculated, 734.2207; found, 734.2233.

5.2.7. Preparation of N³-Boc-3′,5′-O-bis-methoxymethyl-2′-O-p-nitrophenylsulfonyl-5-methyluridine 8f

To a solution of compound 7b (100.0mg, 0.18 mmol) in dry THF (5 mL) was added chloromethyl methyl ether (0.5 mL, 36 mmol). The reaction mixture was stirred at RT for 16 h, and the solvent was evaporated. The residue was purified by flash chromatography on a silica gel column using 30% EtOAc in hexane to give 8f as colorless oil in 50% yield. ¹H NMR (CDCl₃) δ: 8.35 (d, J=9.0 Hz, 2H, aromatic, C₃H & C₅H), 8.22 (d, J=9.0 Hz, 2H, aromatic, C₂H & C₆H), 7.49 (s, 1H, C₆H) 5.89 (d, 1H, J=4.5 Hz,1′H), 5.27 (t, 1H, J=4.5 Hz, 2′H), 4.75 (s, 4H, methylene), 4.60 (t, 1H, J=5.2 Hz, 3′H), 4.20 (m, 1H, 4′H), 4.08-3.83 (m, 2H, 5′H), 3.44 (s, 6H, OMe), 1.95 (s, 3H, CH₃), 1.62 (s, 9H, t-But). ¹³C NMR (CDCl₃) δ: 161.09, 150.73, 147.78, 147.56, 142.90, 133.73, 129.40, 124.31, 110.62, 88.68, 87.21, 85.10, 84.61, 84.00, 81.93, 66.82, 58.87, 57.56, 57.00, 27.41, 12.61. HRMS (m/z): [M+Na] for C₂₅H₃₃N₃O₁₄S, calculated, 654.1581; found, 654.1714.

5.2.8. Preparation of N³-Boc-3′,5′-O-bis-benzoyl-2′-O-p-nitrophenylsulfonyl-5-methyluridine 8g

Compound 7b (100.0 mg, 0.18 mmol) was dissolved in dry THF (5 mL), triethylamine (0.1 mL), 4-dimethylaminopyridine (45 mg, 0.37 mmol) and benzoyl chloride (0.2 mL) were added. The reaction mixture was stirred at RT for 1 h, and the solvent was evaporated. The residue was purified by flash chromatography on a silica gel column using 30% EtOAc in hexane to give 8g as colorless oil in 50% yield. ¹H NMR (CDCl₃) δ: 8.18 (d, J=9.0 Hz, 2H, aromatic, C₃H & C₅H), 8.10 (d, J=9.0 Hz, 2H, aromatic, C₂H & C₆H), 7.93 (m, 4H, benzoyl), 7.63 (m, 2H, benzoyl), 7.45 (m, 4H, benzoyl), 7.07 (s, 1H, C₆H), 5.93 (d, 1H, J=3.6 Hz,1′H), 5.71 (t, 1H, J=5.7 Hz, 3′H), 5.50 (dd, 1H, J=3.6 Hz, 5.4 Hz, 2′H), 4.80 (dd, 1H, J=12.3 Hz, 2.7 Hz, 5′H), 4.65 (m, 1H, 4′H), 4.59 (dd, 1H, J=12.3 Hz, J=3.9 Hz, 1H, 5′H), 1.7 (s, 3H, CH₃), 1.62 (s, 9H, t-But). ¹³C NMR (CDCl₃) δ: 165.91, 165.42, 161.20, 148.10, 147.63, 142.90, 133.77, 133.63, 133.41, 129.93, 129.86, 129.67, 129.63, 129.22, 128.67, 128.60, 124.31, 110.62, 88.63, 87.21, 84.00, 81.92, 66.82, 58.87, 27.47, 12.64. HRMS (m/z): [M+Na] for C₃₅H₃₃N₃O₁₄S, calculated, 774.1575; found, 774.1587.

5.2.9. Preparation of 3′,5′-O-1,1,3,3-tetraisopropyl-1,3-disiloxane-2′-O-methylsulfonyloxy-5-methyluridine 10

Compound 2 (0.50 g, 1.0 mmol) was dissolved in THF (5 mL) and cooled to 0°C, then triethylamine (0.56 mL, 4.1 mmol) was added, followed by the addition of methanesulfonyl chloride (0.15 mL, 1.9 mmol). The mixture was stirred at 0°C for 10 min, warmed to room temperature and stirred for an additional 1.5 h. The reaction mixture was filtered and THF was removed under reduced pressure to give colorless oil, which was purified by flash chromatography on a silica gel column eluted with 20% ethyl acetate in hexane to give 10 as white foam in 90% yield. ¹H NMR (CDCl₃) δ: 8.83 (s, 1H, NH), 7.54 (d, 1H, C₆H), 5.80 (s, 1H, 1′H), 5.04 (d, 1H, J=4.5 Hz, 2′H), 4.40 (m, 1H, 3′H), 4.20 (m, 1H, 4′H), 4.11 (m, 1H, 5′H), 4.03 (m, 1H, 5′H), 3.25 (s, 3H, Ms), 1.94 (s, 3H, CH₃), 1.06–1.09 (m, 28H, iso-prop). ¹³C NMR (CDCl₃) δ: 161.20, 148.10, 147.63, 133.77, 110.56, 89.03, 87.23, 82.86, 81.92, 58.87, 39.21, 17.39, 17.33, 17.22, 17.20, 16.95, 16.82, 16.78, 13.54, 12.86, 12.75, 12.64. HRMS (m/z): [M+Na] for C₂₃H₄₂N₂O₉SSi₂Na, calculated, 601.2047; found, 601.2060.

5.2.10. Fluorination of the precursors: preparation of [¹⁸F]FMAU 9

All precursors were reacted with K¹⁸F/kryptofix 2.2.2. under similar conditions with a slight variation in temperature. The aqueous solution of [¹⁸F]fluoride/kryptofix 2.2.2. was purchased from Cyclotope Inc. (Houston, TX). (CAUTION: A hot cell or highly shielded area should be used for using radioactive material). Water was removed by an azeotropic evaporation at 90°C with acetonitrile (1.0 mL) under a stream of argon. A solution of the precursors 8a–g (3–5 mg) in dry acetonitrile (0.3 mL) was added to the dried K¹⁸F/kryptofix 2.2.2. The reaction mixture was heated at 90°C for 20 min. The crude reaction mixture was passed through a silica Sep-Pak cartridge followed by elution with two portions of ethyl acetate (2.5 mL, total), which was evaporated at 90°C under a stream of argon. The residue from reactions of 8a, 8b, 8e and 8f was dissolved in methanol (0.3 mL), 1M methanol/HCl solution (0.1 mL) was added, and the mixture was heated at 80°C for 10 min. The solvent was evaporated, and the residue was dissolved in HPLC solvent (9% acetonitrile/water, 1.0 mL) and purified by HPLC using a semipreparative column. The residue from reactions of 8c, 8d, and 8g was dissolved in trifluoracetic acid (0.3 mL), heated for 5 min at 80°C then trifluoracetic acid was evaporated completely. The residue was dissolved in MeOH (0.3 mL) and NaOMe solution (0.5M, 0.2 mL) was added. The reaction mixture was heated for 7 min at 80°C. Solvent was evaporated and the crude product was neutralized with 1M HCl (0.2 mL) and purified by HPLC as describe above. The product was eluted with 9% acetonitrile/water at a flow of 4 mL/min. The appropriate fraction (radioactive) was collected between 11.5 and 12.5 min, and the solvent was partially evaporated under reduced pressure. An aliquot of the product [¹⁸F]FMAU 9 was analyzed on an analytical HPLC column to verify its purity and identity by co-injection with the nonradioactive authentic sample of FMAU. For those reactions with very low yield, the crude products were analyzed on an analytical column and the % yield was calculated from the radiochromatograms.