Direct SNAr amination of fluorinated imidazo[4,5-c]pyridine nucleosides: Efficient syntheses of 3-fluoro-3-deazaadenosine analogs

Article abstract of DOI:10.1016/j.tetlet.2005.03.190

This paper describes the ready preparation of 3,6-difluoro-3-deazapurine (4,7-difluoroimidazo[4,5-c]pyridine). This novel base was glycosylated under mild conditions using three different ribose sugar analogs. 3,6-Difluoro-3- deazapurine ribonucleoside an

Full text of DOI:10.1016/j.tetlet.2005.03.190

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 3883–3887
Direct SNAr amination of fluorinated
imidazo[4,5-c]-pyridine nucleosides: efficient syntheses
of 3-fluoro-3-deazaadenosine analogs
Kandasamy Sakthivel* and P. Dan Cook
Biota Inc, 2232 Rutherford Rd, Carlsbad, CA 92008, USA
Received 27 January 2005; revised 23 March 2005; accepted 24 March 2005
Available online 11 April 2005
Abstract—This paper describes the ready preparation of 3,6-difluoro-3-deazapurine (4,7-difluoroimidazo[4,5-c]pyridine). This novel
base was glycosylated under mild conditions using three different ribose sugar analogs. 3,6-Difluoro-3-deazapurine ribonucleoside
analogs underwent direct SNAr amination reactions with liquid ammonia to give 3-fluoro-3-deazaadenosine analogs in excellent
yield; in contrast, 6-chloro-3-fluoro-3-deazapurine nucleosides were inert under similar reaction conditions.
Ó 2005 Elsevier Ltd. All rights reserved.
Fluorinated organic molecules perform a wide range of
biological functions.1 In many cases, the introduction of
one or more fluorine atoms into biologically important
molecules enhances their activity, liphophilicity, bio-
availability, and metabolic stability.2 Accordingly, an
enormous amount of work has been performed to pre-
pare fluorinated nucleoside analogs; several of these
analogs have been studied in clinical trials and a few
have been approved as drugs, including the pancreatic
cancer drug Gemcytabine (Gemzar) and the HIV drug
5F-3TC. It was reported3 quite recently that 7-fluoro-
7-deaza-20-C-methyladenosine 1, a nucleoside derivative
in which a C–F bond appears at the 7-position of
7-deaza-20-C-methyladenosine 2a or 20-C-methyladenosine
2b, inhibits the replication of hepatitis C virus (HCV)
with higher inhibitory potency (EC50 = 0.07 lM) than
its parent compounds 2a (EC50 = 0.25 lM) and 2b
(EC50 = 0.26 lM). 3-Deazaadenosine 3 analogs4 display
broad-spectrum antiviral and anticancer activities, and
30-deoxyribonucleosides exhibit interesting antiviral,
antifungal, antibacterial, antiparasitic, and anticancer
properties.5 Furthermore, ribonucleosides having 20-b-
C-methyl substituents display anti-HCV properties.6
These findings prompted us to synthesize a series of
3-fluoro-3-deazaadenosine analogs, 4a–c, that combine
the 3-fluoro-3-deazapurine heterocycle with the active
compoundsÕ ribosugar moieties. The introduction of a
C–F bond at the 3-position of purine could not only
isosterically mimic the regular adenine and guanine
bases but also the 3-deazaadenine and 3-deazaguanine
bases (Fig. 1).
Although syntheses of 3-fluoro-3-deazaadenosine (3F-3-
deaza-Ad) have been reported,7 there are difficulties
encountered when applying these procedures to the
synthesis of modified-sugar analogs of 3F-3-deaza-Ad
and, therefore, an efficient and general methodology is
desirable. 6-Chloro-3-deazapurine (4-chloroimidazo[4,5-
c]pyridine) nucleosides are highly inert toward
nucleophiles (e.g., ammonia) in SNAr reactions, which is
the main drawback for their use in the synthesis of 3-dea-
za-Ad analogs.8 Thus, an alternative two-step synthesis
(displacement of the 6-chloro-substituent with hydrazine
under reflux conditions and subsequent cleavage of
the N–N bond using Raney Ni) for the amination of
6-chloro-3-deazapurine has been developed.9,10 This
approach has its limitations, however, including poor-
to-moderate yields and the generation of side products.
Moreover, use of hydrazine under reflux is not a condi-
tion tolerated by various sugar or base modifications;
indeed, Matsuda and co-workers7a reported that
nucleoside 5a decomposed when attempting to perform
the displacement reaction using hydrazine under
reflux (Scheme 1).11
Keywords: Nucleosides; 3-Deazaadenosine; SNAr reactions, Fluori-
nated nucleosides.
*
Corresponding author. Tel.: +1 760 496 0432; fax: +1 858 625
2433; e-mail: sakvel@hotmail.com
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.03.190

3884
K. Sakthivel, P. D. Cook / Tetrahedron Letters 46 (2005) 3883–3887
Figure 1. Rationale for the synthesis of target compounds 4a–c.
Scheme 1. Reagents and conditions: (i) (a) hydrazine, reflux; (b) Raney nickel; (ii) liquid NH3 or MeOH saturated with NH3 at 0 °C, (iii) NaSMe,
DMF.
Likewise, our initial attempts to synthesize 3F-3-deaza-
20-C-Me-Ad 4b based on performing SNAr reactions
with 5b were also unsuccessful, but we did gather some
useful clues about what would be necessary to design a
successful synthetic scheme for the target compounds
4a–c: (i) The starting material was completely recovered
when compound 5b was reacted directly with ammonia
nucleophiles, that is, NH4OH, saturated methanolic
NH3 (MeOH saturated with NH3 at 0 °C), and liquid
NH3. (ii) In an attempt to realize the classical two-step
amination, we heated nucleoside 5b under reflux with
anhydrous hydrazine, but observed only decomposition
of the starting material.7a (iii) To proceed with SeelaÕs
methodology,10b we attempted the synthesis of 3-fluoro-
6-S-methyl-3-deazapurine-20-C-methylriboside 6c.12 When
we reacted nucleoside 5b with sodium thiomethoxide
in DMF at room temperature, the products we
isolated were, surprisingly, 6a (52%) and 6b (17%); their
formation reveals that the SNAr reaction of the 3-fluoro
substituent is more facile than that of the 6-chloro sub-
stituent.13 It is well established in the literature14 that a
halogen substitutent adjacent to the N-1 position (the 6-
position of purine or 3-deazapurine) is most reactive to-
ward nucleophiles, but it is also well known that fluoride
is the best leaving group among the halogens in most
SNAr reactions.15 In addition, the electron-withdrawing
6-chloro substituent in 5b may enhance the reactivity of
the 3-fluoro substituent by reducing the electron density
around the six-membered ring.16 As a result, the 3-flu-
oro substituent is easily displaced. By exploiting these
observations, facile SNAr reactivity at the 6-position
can be tuned by introducing fluorine substituents
at both the 3- and 6-positions. On the basis of these
concepts, we investigated the direct SNAr aminations
of 4,7-difluoro-imidazo[4,5-c]pyridine nucleosides with
NH3 nucleophiles and, consequently, in this paper we
present the efficient and general syntheses of 3F-3-dea-
za-Ad analogs using direct SNAr amination.
We synthesized (Scheme 2) 3,6-difluoro-3-deazapurine
15 from commercially available 3-chloro-2,4,5,6-tetra-
fluoropyridine 7. The synthesis of N-(2-amino-
5-chloro-3,6-difluoropyridine-4-yl)phthalimide
8
was
realized in three steps from 7 by following a reported
procedure.17 Diazotization of compound 8 using tert-
butylnitrite in THF produced the deaminated product
9. Treatment of compound 9 with 28% aqueous ammo-
nium hydroxide at room temperature, followed by cata-
lytic dehalogenation under H2 (50 psi) using 10% Pd on
activated carbon, afforded 4-amino-2,5-difluoro-pyri-
dine 11 (55% overall yield from compound 7). Gratify-
ingly, the synthesis of compound 11 from compound 7
required only a single purification by column chroma-
tography after the last step, because all of the synthetic
steps proceeded with clean conversions and excellent
yields. Compound 11 was nitrated using potassium

K. Sakthivel, P. D. Cook / Tetrahedron Letters 46 (2005) 3883–3887
3885
Scheme 2. Reagents and conditions: NPhthol = Phthalimido; (i) See Ref. 17 or Supplementary data; (ii) tert-butylnitrite, THF, 60 °C, 1 h; (iii) 28%
aqueous NH4OH, rt, 1.5 h; (iv) 10% Pd/C, 50 psi [H] Et3N, 24 h; (v) concd H2SO4, KNO3 4 °C to rt, 1.5 h; (vi) concd H2SO4, rt, 24 h; (vii) Raney
nickel, anhydrous EtOH, 36 psi [H]; (viii) diethoxymethyl acetate, 100 °C, 2 h.
nitrate in concentrated H2SO4 to give the nitrated
amino compound 12 in 73% yield, which on further
reaction with concentrated H2SO4 produced 4-amino-
2,5-difluoro-3-nitropyridine 13 in 87% yield. Catalytic
reduction of 13 under H2 (36 psi) using Raney Ni pro-
duced the diamino compound 14 in 95% yield, which
we reacted with diethoxymethyl acetate18 to give the
ring-closed product, 3,6-difluloro-3-deazapurine 15 in
78% yield.
In contrast, the coupling reaction between 30-deoxy su-
gar analog 19 and base 15 afforded the anomeric mixture
20ab (1:2) in 70% yield; we also observed the anomeric
1
mixture of the N-7 positional isomer (ca. 7%) by H
NMR spectroscopic analysis of the crude reaction mix-
ture. We isolated the major product (20ab) as a pure
mixture (a/b = 1:2) by flash column chromatography
(SiO2; 3% EtOAc in CH2Cl2) (Scheme 3).
To realize our synthetic goal, we first investigated the di-
rect SNAr amination of compound 17a, which we re-
acted with liquid NH3 at 80 °C in a steel bomb.23
After 48 h, we observed complete conversion of 17a
and the only product isolated was indeed 3F-3-deaza-
adenosine 4a (81% yield). Neither a prolonged reaction
time (96 h) nor a higher temperature (110 °C) resulted
in the formation of the diamino compound. We
Next we investigated the coupling reactions between
base 15 and three different ribosugar analogs: 1,2,3,5-
tetra-O-benzoyl-b-D-ribofuranose 16a, 1,2,3,5-tetra-O-
benzoyl-2-C-methyl-b-D-ribofuranose 16b,19 and 1,2-
di-O-acetyl-5-O-(4-methylbenzoyl)-3-deoxy-b-D-ribofu-
ranose 19.20 The Vorbruggen-type glycosylation reac-
¨
tion21 between the silylated base 15 and sugar 16a in
the presence of trimethylsilyl trifluoromethanesulfonate
(TMSOTf) proceeded smoothly to give 17a in 70% yield.
In contrast, glycosylation reactions between 15 and the
1
confirmed the structure of 4a by comparing its H and
13C NMR and UV spectra with those of a compound
synthesized by Matsuda et al.7a This comparison
revealed that the slower-moving (by TLC), major
glycosylation product was the N-9 positional isomer.
To generalize this direct amination reaction, we sub-
jected compounds 17b and 18b independently to similar
reactions with liquid NH3 to give 4b (88%) and 22
(72%), respectively. The assignment of the positional
isomers 4b and 22 is based on the UV absorption spectra
of these derivatives:24 the N-9 positional isomer 4b dis-
played its kmax at 269 nm while the undesired (N-7) iso-
mer 22 exhibited kmax at 295 nm. In addition, the value
of kmax of 4b is consistent with that obtained for com-
2-C-methylribose 16b failed under both the Vorbruggen
¨
and SnCl4-mediated coupling conditions. The successful
coupling of 6-chloropurine and the 2-C-methyl sugar
16b has been reported in the literature22 when a combi-
nation of two coupling reagents, for example, TMSOTf
and 1,8-diazabicyclo[4,5,0]undec-7-ene (DBU), was
used. Using this procedure, we reacted base 15 with 2-
C-methylribose 16b in dichloroethane in the presence
of TMSOTf and DBU. After the usual work-up, we iso-
lated the corresponding glycosylated products 17b and
18b as a mixture of positional isomers (9-N/7-N = 5:1)
in a combined yield of 84%. For comparison, we also
realized the coupling reaction between base 15 and
ribose sugar 16a under similar conditions to give a
mixture of the corresponding glycosylated products 17a
and 18a in a 6:1 ratio (82% combined yield). The posi-
tional isomers 17 and 18 were readily separated by flash
column chromatography (SiO2; 3% EtOAc in CH2Cl2).
The major product, the slower-moving one by thin-layer
chromatography (TLC), was 17a, which is also the prod-
1
pound 4a. In their H NMR spectra, the 8-H protons
of the N-9 isomers are always located downfield (ca.
0.1 ppm) relative to the 8-H protons of the N-7 isomers.
These data are consistent with those reported for other
3-deazapurine nucleosides.24 We reacted the pure mix-
ture 20ab with liquid NH3 in a similar fashion to give
a mixture of 4c and 4c-a in 85% yield after separation
by reverse-phase HPLC. We confirmed the glycosylation
site and the anomeric configurations of 4c and 4c-a
by using UV and 1H NMR spectral data, which are
uct obtained under the Vorbruggen reaction conditions.
¨

3886
K. Sakthivel, P. D. Cook / Tetrahedron Letters 46 (2005) 3883–3887
Scheme 3. Reagents and conditions: (i) (a) compound 15, HMDS, cat. ammonium sulfate, reflux 3 h; (b) dichloroethane, TMSOTf, rt, 12 h; (ii)
compound 15, DBU, TMSOTf, dichloroethane, 0 °C to rt, 24 h.
Scheme 4.
comparable to those reported for 30-deoxy-3-deaza-Ad
(Scheme 4).25
Acknowledgments
We thank Drs. John Lambert and Susan M. Daluge for
their support.
In summary, we have executed an efficient synthesis of a
novel base, 3,6-difluoro-3-deazapurine. This base was
glycosylated under mild conditions using three different
ribose sugar analogs. Our studies on the direct SNAr
amination reactions of fluorinated imidazo[4,5-c]pyri-
dine nucleosides have yielded an efficient and general-
ized methodology for the synthesis of such biologically
important molecules as 3-fluoro-3-deazaadenosine
analogs. A study of the biological properties of these
nucleosides will be published in due course.
Supplementary data
The supplementary data available online with the paper
in ScienceDirect. Supplementary data contains experi-
mental procedures and characterization data of all the
compounds. Supplementary data associated with this
article can be found, in the online version at
doi:10.1016/j.tetlet.2005.03.190.

K. Sakthivel, P. D. Cook / Tetrahedron Letters 46 (2005) 3883–3887
3887
11. Sartorelli and co-workers reported (Ref. 7b), however,
that compound 5a produced 4a upon reaction with
hydrazine under reflux followed by treatment with Raney
Ni.
12. On the basis of a report by Seela and co-workers (Ref.
10b), we anticipated that compound 6c, after oxidation to
6-methanesulfonyl-3-fluoro-3-deazapurine-20-C-methylribo-
side, would react readily with NH3 to give the target
compound 4b.
13. A similar observation has been reported in the literature.
When 2-chloro-5-fluoropyridine was reacted with aqueous
NH4OH at 180 °C, the 5-fluoro substituent was most
readily displaced. See: Hand, E. S.; Baker, D. C. Synthesis
1989, 905–908.
14. (a) When the 2,6-difluoro-3-deazapurine or 2,6-dichloro-3-
deazapurine ribonucleosides were reacted with NH3, only
the 6-halogen substituents were displaced to give 2-fluoro-
3-deazaadenosine and 2-chloro-3-deazaadenosine, respec-
tively (see Refs. 7b and 10); (b) May, J. A.; Townsend, L.
B. Nucl. Acid Chem. 1978, 2, 693–699; (c) Robins, M. J.;
Basom, G. L. Can. J. Chem. 1973, 51, 3161–3169; (d)
Lakshman, M. K.; Hilmer, J. H.; Martin, J. Q.; Keeler, J.
C.; Dinh, Y. Q. V.; Ngassa, F. N.; Russon, L. M. J. Am.
Chem. Soc. 2001, 123, 7779–7787; (e) Ding, S.; Gray, N.
S.; Ding, Q.; Schultz, P. G. Tetrahedron Lett. 2001, 42,
8751–8755.
References and notes
1. (a) Welch, T.; Eswarakrishnan, S. Fluorine in Bioorganic
Chemistry; Wiley-Interscience: New York, 1991, and
references cited therein; (b) Mikami, K.; Itoh, Y.; Yama-
naka, M. Chem. Rev. 2004, 104, 1–16.
2. (a) Park, B. K.; Kitteringham, N. R.; OÕNeil, P. M. Annu.
Rev. Pharmacol. Toxicol. 2001, 41, 443–470; (b) Man, H.;
Corral, L. G.; Stirrling, D. I.; Muller, G. W. Bioorg. Med.
Chem. Lett. 2003, 13, 3415–3417; (c) Kool, E. T. Acc.
Chem. Res. 2002, 35, 936–943, and references cited therein;
(d) Dai, Q.; Piccirilli, J. A. Org. Lett. 2003, 5, 807–810,
and reference cited therein; (e) Lai, J. S.; Kool, E. T. J.
Am. Chem. Soc. 2004, 126, 3040–3041; (f) Robins, M. J.;
MacCoss, M.; Naik, S. R.; Ramani, G. J. Am. Chem. Soc.
1976, 98, 7381–7389; (g) Blackburn, G. M.; Brown, D.;
Martin, S. J.; Parratt, M. J. J. Chem. Soc., Perkin Trans. 1
1987, 181–186.
3. Eldrup, A. B.; Prhavc, M.; Brooks, J.; Bhat, B.; Prakash,
T. P.; Song, Q.; Bera, S.; Bhat, N.; Dande, P.; Cook, P. D.;
Cook, P. D.; Bennet, C. F.; Carrol, S. S.; Ball, R. G.;
Bosserman, M.; Burlein, C.; Colwell, L. F.; Fay, J. F.;
Flores, O. A.; Getty, K.; LaFamina, R. L.; Leone, J.;
MacCoss, M.; McMasters, D. R.; Tomassini, J. E.;
Langen, D. V.; Wolanski, B.; Olseon, D. B. J. Med.
Chem. 2004, 47, 5284–5297.
15. (a) The direct SNAr amination of 2-halopyridine with
lithium aminoborohydride proceeds in the following order
of reactivity: 2-F-pyridine > 2-Br-pyridine > 2-Cl-pyri-
dine. See: Thomas, S.; Roberts, S.; Pasumansky, L.;
Gamsey, S.; Singaram, B. Org. Lett. 2003, 5, 3867–3870;
(b) Smith, M. B.; March, J. In Advanced Organic
Chemistry, 5th ed.; Wiley Inter-Science: New York,
2001; Chapter 13, p 851.
16. Deazapurines are more electron rich than the correspond-
ing purines. See Refs. 7a and 9a.
17. Yazaki, A.; Niino, Y.; Ohshita, Y.; Hirao, Y.; Amano, H.;
Hayashi, N.; Kuramoto, Y. Intl. Patent Appl. WO 97/
11068, 1997.
18. Kroon, C.; van den Brink, A. M.; Vlietstra, E. J.;
Salemink, C. A. Recl. Trav. Chim Pays-Bas 1976, 95, 127.
19. (a) Wolfe, M. S.; Harry-Okuru, R. E. Tetrahedron Lett.
1995, 36, 7611–7614; (b) Harry-OÕkuru, R. E.; Smith, J.
M.; Wolfe, M. S. J. Org. Chem. 1997, 62, 1754–
1759.
4. (a) Ueland, P. M. Pharmacol. Rev. 1982, 34, 223–253; (b)
Prus, K. L.; Wolberg, G.; Keller, P. M.; Fyfe, J. A.;
Stopford, C. R.; Zimmerman, T. P. Biochem. Pharmacol.
1989, 38, 509–517.
5. (a) Shim, J.; Larson, G.; Lai, V.; Naim, S.; Wu, J. Z.
Antiviral Res. 2003, 58, 243–251; (b) Ismaili, H.; Moulay,
A.; Cheng, Y.; Lavalle, J.; Siddiqui, A.; Storrer, R. Intl.
Patent Appl. WO 01/60315, 2001; (c) Kumar, A.; Khan, S.
I.; Manglani, A.; Khan, Z. K.; Katti, S. B. Nucleos.
Nucleot. 1994, 13, 1049–1058, and references cited therein.
6. (a) Eldrup, A. B.; Allerson, C. R.; Bennet, C. F.; Bera, S.;
Bhat, B.; Bosserman, M.; Brooks, J.; Burlein, C.; Carrol,
S. S.; Cook, P. D.; Getty, K. L.; MacCoss, M.; McMas-
ters, D. R.; Olseon, D. B.; Prakash, T. P.; Prhavc, M.;
Song, Q.; Tomassini, J. E.; Xia, J. J. Med. Chem. 2004, 47,
2283–2295; (b) Sommadossi, J. P.; Lacolla, P. Intl. Patent
Appl. WO 01/92282, 2001.
7. (a) Minakawa, N.; Kojima, N.; Matsuda, A. J. Org.
Chem. 1999, 64, 7158–7172; (b) Liu, M. C.; Luo, M. Z.;
Mozdziesz, D. E.; Lin, T. S.; Dutschman, G. E.; Gullen, E.
A.; Cheng, Y. C.; Sartorelli, A. C. Nucleos. Nucleot.
Nucleic Acids 2001, 20, 1975–2000.
8. For clarity, we use regular purine base numberings
throughout this manuscript. Imidazo[4,5-c]pyridine num-
berings are provided in the Supplementary data.
9. (a) Rousseau, R. J.; Townsend, L. B.; Robins, R. K.
Biochemistry 1966, 5, 756–760; (b) Srivastava, P. C.;
Robins, R. K.; Meyer, R. B., Jr. In Chemistry of
Nucleosides and Nucleotides; Townsend, L. B., Ed.;
Plenum: New York, 1988; Vol. 1, p 113.
10. Two other methodologies have been reported: (a) May, J.
A.; Townsend, L. B. J. Chem. Soc., Chem. Commun. 1973,
64–65; (b) Seela, F.; Grein, T.; Samnick, S. Helv. Chim.
Acta 1992, 75, 1639–1650.
20. Rhie, S. Y.; Pfleiderer, W. Nucleos. Nucleot. 1994, 13,
1425–1452.
21. Niedballa, U.; Vorbruggen, H. J. J. Org. Chem. 1974, 39,
3654–3660.
¨
22. Franchetti, P. F.; Cappellacci, S. M.; Trincavelli, L.;
Martini, C.; Mazzoni, M. R.; Lucacchini, A.; Grifantini,
M. J. Med. Chem. 1998, 41, 1708–1715.
23. To avoid the possibility of attack by a methoxy nucleo-
phile, we did not use saturated methanolic ammonia
solution.
24. (a) Montgomery, J. A.; Shortnacy, A. T.; Clayton, S. D. J.
Heterocycl. Chem. 1977, 14, 195–197; (b) Poonian, M. S.;
McComas, W. W. J. Med. Chem. 1979, 22, 958–962.
25. Volpini, R.; Camaioni, E.; Costanzi, S.; Vittori, S.;
Cristalli, G. Helv. Chim. Acta 1998, 81, 2326–2331.