Synthesis and conformational analysis of 1′- and 3′-substituted 2-deoxy-2-fluoro-D-ribofuranosyl nucleosides

Article abstract of DOI:10.1002/hlca.200790191

Convergent syntheses of the 9-(3-X-2,3-dideoxy-2-fluoro-β-D- ribofuranosyl)adenines 5 (X = N₃) and 7 (X = NH₂), as well as of their respective α-anomers 6 and 8, are described, using methyl 2-azido-5-O-benzoyl-2,3-dideoxy-2-fluoro-β-D-ribofuranoside (4) as glycosylating agent. Methyl 5-O-benzoyl-2,3-dideoxy-2,3-difluoro-β-D- ribofuranoside (12) was prepared starting from two precursors, and coupled with silylated N⁶-benzoyladenine to afford, after deprotection, 2′,3′-dideoxy-2′,3′-difluoroadenosine (13). Condensation of 1-O-acetyl-3,5-di-O-benzoyl-2-deoxy-2-fluoro-β-D-ribofuranose (14) with silylated N²-palmitoylguanine gave, after chromatographic separation and deacylation, the N⁷-β-anomer 17 as the main product, along with 2′-deoxy-2′-fluoroguanosine (15) and its N⁹-α- anomer 16 in a ratio of ca. 42:24:10. An in-depth conformational analysis of a number of 2,3-dideoxy-2-fluoro-3-X-D-ribofuranosides (X = F, N₃, NH₂, H) as well as of purine and pyrimidine 2-deoxy-2-fluoro-D- ribofuranosyl nucleosides was performed using the PSEUROT (version 6.3) software in combination with NMR studies.

Full text of DOI:10.1002/hlca.200790191

1818
Helvetica Chimica Acta – Vol. 90 (2007)
Synthesis and Conformational Analysis of 1’- and 3’-Substituted 2-Deoxy-2-
fluoro-d-ribofuranosyl Nucleosides
by Grigorii G. Sivets, Elena N. Kalinichenko, and Igor A. Mikhailopulo*
Institute of Bioorganic Chemistry, National Academy of Sciences, Acad. Kuprevicha 5, 220141 Minsk,
Belarus (phone/fax: þ 375-172-67-81-48; e-mail: igor_mikhailo@yahoo.de)
This paper is dedicated to the fond memory of our colleague Dr. Galina V. Zaitseva. She will be sorely
missed by her colleagues and friends at the Institute of Bioorganic Chemistry.
Convergent syntheses of the 9-(3-X-2,3-dideoxy-2-fluoro-b-d-ribofuranosyl)adenines 5 (X ¼ N3)
and 7 (X ¼ NH2), as well as of their respective a-anomers 6 and 8, are described, using methyl 2-azido-5-
O-benzoyl-2,3-dideoxy-2-fluoro-b-d-ribofuranoside (4) as glycosylating agent. Methyl 5-O-benzoyl-2,3-
dideoxy-2,3-difluoro-b-d-ribofuranoside (12) was prepared starting from two precursors, and coupled
with silylated N6-benzoyladenine to afford, after deprotection, 2’,3’-dideoxy-2’,3’-difluoroadenosine (13).
Condensation of 1-O-acetyl-3,5-di-O-benzoyl-2-deoxy-2-fluoro-b-d-ribofuranose (14) with silylated N2-
palmitoylguanine gave, after chromatographic separation and deacylation, the N7-b-anomer 17 as the
main product, along with 2’-deoxy-2’-fluoroguanosine (15) and its N9-a-anomer 16 in a ratio of ca.
42 :24 :10. An in-depth conformational analysis of a number of 2,3-dideoxy-2-fluoro-3-X-d-ribofurano-
sides (X ¼ F, N3, NH2, H) as well as of purine and pyrimidine 2-deoxy-2-fluoro-d-ribofuranosyl
nucleosides was performed using the PSEUROT (version 6.3) software in combination with NMR
studies.
Introduction. – 2’-Deoxy-2-fluoro ribo-nucleosides have great potential for the
investigation of chemical and/or biochemical problems in which the ribofuranosyl
moiety is involved (see, e.g., [1 – 4]). They constitute a very important family of
analogues of natural nucleosides, displaying a wide variety of biological activities (see,
e.g., [5 – 8]), and are constituents of the nucleic-acid-based gene-silencing molecules
such as antisense oligodeoxyribonucleic acids (ODNs), ribozymes, and small interfer-
ing RNAs (siRNAs) (see, e.g., [9 – 13]). Two main peculiarities of 2’-deoxy-2-fluoro
ribo-nucleosides are responsible for the continuous interest in these compounds.
Replacement of the 2’-a-OH group (or 2’-a-H-atom of 2’-deoxy nucleosides) by an a-
F-atom leads to i) significant strengthening of the glycosyl bond [14][15] and ii)
remarkable conformational changes in the pentofuranose ring [1]. Further, this type of
modification of pyrimidine and purine nucleosides was found to result in a remarkable
stabilization of the glycosyl bond towards both chemical [14] and enzymatic [15][16]
degradation.
A number of synthetic pathways to this family of biologically important nucleosides
have been elaborated, following basically three main approaches:
ꢀ Convergent synthesis employing suitable derivatives of 2-deoxy-2-fluoro-d-ribofur-
anose (for a review, see [17]) as glycosylating agents (see, e.g., [18 – 21] and refs. cit.
therein)
ꢀ 2007 Verlag Helvetica Chimica Acta AG, Zürich

Helvetica Chimica Acta – Vol. 90 (2007)
1819
ꢀ Introduction of a 2’-a-F-atom by fluorination of pyrimidine O2,2’-anhydro-nucleo-
sides [6][14] or by nucleophilic replacement of the 2’-b-OH group of pyrimidine and
purine nucleosides with F3SNEt2 (DAST) or other fluorinating agents [5][6] [22 – 26]
(for a review, see [27])
ꢀ Enzymatic transglycosylation of purines using 2’-deoxy-2’-fluoro-uridine as a donor
of the sugar moiety [15][16].
During the course of our studies, we have synthesized a number of 3’-deoxy-3-
fluoro a,b-d-ribo-nucleosides containing diverse (deoxy, azido, amino, and chloro)
substituents at the 2’-C-atom [19 – 21][28 – 32]. The conformational analysis of this
genus of ribo-nucleosides was performed with the aid of the PSEUROT (version 6.3)
program [33], using the reparameterized [34] Karplus-type relation and empirical
bond-correction terms [35]. An astonishing conformational uniformity of the
pentofuranose rings was revealed, with a dominating population of the S-conformation,
which was found to be only marginally dependent on both the character of the C(2’)-a-
substituent and the anomeric configuration [35]. This behavior could be satisfactorily
rationalized by means of the Brunck – Weinhold model [36], which invokes maximum
antiperiplanar s – s* stabilization when the donating bond is the least polar one, and
the acceptor orbital is at the most polarized bond.
These data prompted us to investigate the conformation of 2’-deoxy-2’-fluoro a,b-
d-ribo-nucleosides having variable substituents at C(3’) and C(1’). Furthermore, 3’-
amino-2’,3’-dideoxy-2’-fluoro-b-d-ribofuranosyl purine nucleosides are of significant
interest as constituents of the N(3’) ! P(5’) phosphoramidate oligonucleotides [26];
such oligonucleotides, which are complementary to the RNA template region of human
telomerase (hTR), have been shown to be effective inhibitors of the enzyme and,
consequently, inhibited cancer cell growth in vitro [37].
In the present communication, we report the convergent synthesis of 1) b- and a-
configured 9-(3-azido-2,3-dideoxy-2-fluoro-d-ribofuranosyl)adenine, 5 and 6, respec-
tively, which were transformed into the respective 3’-amino-3’-deoxy nucleosides 7 and
8; 2) of 2’,3’-dideoxy-2’,3’-difluoroadenosine (13); and 3) of 2’-deoxy-2’-fluoroguano-
sine (15), its N9-a-anomer 16, and of 7-(2-deoxy-2-fluoro-b-d-ribofuranosyl)guanine
(17) (see Schemes 1 and 2 below). Further, we performed a detailed conformational
analysis of a number of these 2,3-dideoxy-2-fluoro-3-X-d-ribofuranosides (X ¼ H, F,
NH2, N3) as well as of purine and pyrimidine 2-deoxy-2-fluoro-d-ribofuranosyl
nucleosides, the goal being to gain further insight into the role of different stereo-
electronic factors on the pseudorotational equilibrium in nucleosides and pentofur-
anose sugars.
Results and Discussion. – 1. Synthesis. The synthesis of methyl 3-azido-5-O-benzoyl-
2,3-dideoxy-2-fluoro-b-d-ribofuranoside (4), the precursor of the target compounds 5 –
8, was realized in three steps starting from the b-d-xylofuranoside derivative 1 [30], via
2 and 3, in ca. 40% combined yield, using conventional methods (Scheme 1).
Condensation of 4 with persilylated N6-benzoyladenine in the presence of an excess
SnCl4 in anhydrous MeCN under reflux, followed by chromatographic purification,
gave a mixture of the anomeric benzoylated nucleosides, along with the starting
compound 4 (20%). Debenzoylation of the nucleoside mixture, followed by chromato-
graphic purification, then afforded the pure nucleosides 5 and 6 in yields of 34 and 8%,

1820
Helvetica Chimica Acta – Vol. 90 (2007)
Scheme 1
a) TsCl, pyridine, 208, 72 h; 87%. b) [H], 10% Pd/C, EtOH, 208, 18 h; or 20% Pd(OH)2/C, EtOH,
cyclohexene, reflux, 2 h (95% of 11). c) BzCl, pyridine, 208, 24 h; 85% of 3 (two steps) or 89% of 12 (two
steps). d) NaN3, DMSO, 190– 195 8, 30min; 51%. e) 4/persilylated N6-benzoyladenine/SnCl4 10:19 :48,
MeCN, reflux, 80min. f) Ammonia-sat. MeOH, 08, 208, 18 h; 34% of 5, 8% of 6 (two steps each). g)
Ph3P, pyridine; aq. NH3; 82% of 7; 85% of 8. h) CsF, DMSO, 190– 195 8, 130min; 32% (based on
consumed 9). i) 1/DAST 10:22, toluene, 20 8, 20h; 30% (based on consumed 1). j) 12/persilylated N6-
benzoyladenine/SnCl4 100 :175 :50, MeCN, reflux, 105 min; 45% (two steps).

Helvetica Chimica Acta – Vol. 90 (2007)
1821
respectively. It is noteworthy that no anomerization of 4 was observed, in accord with
its low reactivity (see, e.g., [29][30]).
Transformation of the individual azido nucleosides 5 and 6 into the respective 3’-
amino-3’-deoxy nucleosides 7 and 8, respectively, was effected using the Staudinger
reaction, and the desired compounds were purified by flash chromatography
(Scheme 1).
Recently, we briefly reported on the synthesis of 13 (see supporting information in
[35]), and now provide the full experimental details. Two different routes were tested
for the synthesis of the starting difluoride 10. First, treatment of the xyloside 1 [30] with
an excess of DAST in toluene at room temperature proceeded smoothly and furnished,
after 20h, the desired difluoride 10 in 30% yield, based on consumed 1, 43% of which
was recovered. The second route consisted in the nucleophilic displacement of the 2-
TsO group of the arabinoside 9 [29][30] upon treatment with CsF in DMSO at 190–
1958, which gave, after chromatography, 32% of 10 (based on consumed starting
material), along with the initial arabinoside 9 (25%). Attempts to improve the yield of
10, making use of KF or NaF, failed.
Two-step replacement of the 5-BnO group of 10 by a BzO group afforded 12 in high
yield. The latter was allowed to react with persilylated N6-benzoyladenine, as described
above, to afford, after purification, the benzoylated derivative of 13, together with the
starting material 12 (27%) and some impurities. Once again, very low reactivity of the
difluorinated methyl glycoside 12 was observed, similar to that of the analogue 4.
Standard deprotection of this fraction, followed by chromatographic purification,
finally afforded the desired nucleoside 13 in 45% overall yield.
The synthesis of the target compounds 15 – 17 (Scheme 2) was based on
condensation of the acetate 14 [19] with silylated N2-palmitoylguanine in MeCN in
the presence of trimethylsilyl triflate (TMS-Tf) under reflux, which gave a complex
mixture from which three individual acylated nucleosides were isolated by column
chromatography, in 27, 12, and 5% yield, respectively. Standard deacylation of these
nucleosides then afforded 17 as the main product, together with the b- and a-anomeric
compounds 15 and 16 in good yields as well. The same reaction, but in the presence of
an increased amount of TMS-Tf in CH2Cl2 at reflux, afforded, after chromatography,
the same three nucleosides, in 27%, 19%, and 8% yield, respectively.
Scheme 2
a) 14/persilylated N2-palmitoylguanine/TMS-Tf 100:120:155, MeCN, reflux, 3 h [or 100:120:155,
CH2Cl2, reflux, 5 h]. b) Column chromatography (SiO2). c) Ammonia-sat. MeOH, 08, 208, 24 h; 10 –
13% of 15, 2 – 5% of 16, and 22% of 17.

1822
Helvetica Chimica Acta – Vol. 90 (2007)
It is noteworthy that the predominant formation of the N7-b-nucleoside 17 vs. the
N9-b-anomer 15 (ratio ca. 2 :1) is an exact antithesis to the predominant formation of
the N9-b-anomer in the condensation of silylated N2-acylguanines with 1-O-acetyl-2,5-
di-O-benzoyl-3-deoxy-3-fluoro-b-d-ribofuranose under similar reaction conditions, as
reported before [30]. We have previously shown that the use of an excess of a Friedel –
Crafts catalyst in the glycosylation reaction allows one to minimize the formation of the
N7-purine isomer (see, e.g., [30][35]). One can hypothesize that the N7-b-nucleoside 17
is formed under kinetic control, but cannot be isomerized into the thermodynamically
more stable N9-b-isomer 16 owing to i) the remarkable stabilization of the glycosyl
bond (vide supra), or ii) the absence of the neighboring 2’-O-acyl group, which would
facilitate breakdown of the glycosyl bond by intramolecular attack of the C¼O
function onto the C(1’)-atom.
2. Spectroscopic Analyses. The structures of all synthesized compounds were
corroborated by in-depth spectroscopic analyses. The assignments of the configurations
1
at the anomeric centers were based upon H- and 13C-NMR data (Tables 1– 3 ). Thus,
the b ! a change of configuration at the anomeric center of each pair of related
nucleosides is accompanied by displacement of the HÀC(1’) and HÀC(4’) resonances
to lower field, and a high-field shift of the HÀC(2’) resonance (Table 1; cf. also [30]).
Diagnostic for the a-anomeric configurations of the nucleosides 6, 8, 16, and 19 is a
long-range coupling of HÀC(8) to the 2’-a-F-atom (Table 2, last column: J(H,F) ¼ 2.3,
2.3, 3.0, and 2.0 Hz, resp.), and of C(8) to the 2’-a-F-atom (Table 3, last column:
J(C,F) ¼ 4.9, 3.3, 6.5, and 4.2 Hz, resp.) in their 1H- and 13C-NMR spectra, respectively.
These couplings are generally indicative of the physical proximity of the nuclei involved
(see, e.g., [19][35] and refs. cit. therein). These couplings were not observed in the
corresponding b-anomers 5, 7, 13, 15, 17, 18, and 20.
A common feature of the above compounds was the very low absolute values of
3J(C(4’),F), which points to a high preference for the N-type pentofuranose-ring
conformation. In this conformation, the FÀC(2’)ÀC(3’)ÀC(4’) fragment is in a clinal
(ca. 908) arrangement, consistent with a coupling constant smaller than ca. 1.5 Hz.
The CD spectrum of the b-anomer 5 displayed, like that of adenosine [38], a
negative long-wavelength Cotton effect; in contrast, the 220-nm envelope(s) were
missing, and a positive molar ellipticity in the short-wavelength region was observed
(Fig. 1). As expected, the long-wavelength Cotton effect in the CD spectrum of the a-
anomer 6 was positive, its maximum being shifted to the far-UV compared to 5, with an
additional shoulder at ca. 230nm and clear trough in the short-wavelength region. The
shoulder in the region at 220– 240nm was positive, similar to the 220-nm envelope of a-
d-adenosine [38]. Two curves bear a mirror-image relationship. The differences in the
shape of the a-anomer 6 vs. the b-anomer 5 was attributed to the electronic interaction
between the 2’-a-F-atom and the imidazole ring of the purine base in the former.
The CD spectra of the anomeric pair 7 and 8 were found to be approximately mirror
images, with a weaker rotational strength in the long-wavelength region compared to 5
and 6. In contrast to adenosine, the positive 225-nm envelope in the CD spectrum of the
b-anomer 7 was absent. However, as in the case of adenosine, the CD curve of the a-
anomer 8 showed a clearly resolved, positive Cotton effect at 228 nm, which was
apparent as a shoulder in the CD spectrum of the a-azide 6. It is remarkable that both
anomers showed clearly resolved negative Cotton effects in the short-wavelength

Helvetica Chimica Acta – Vol. 90 (2007)
1823

1824
Helvetica Chimica Acta – Vol. 90 (2007)
3
Table 2. Selected J(H,H) and J(H,F) Coupling Constants (in Hz). Recorded at 500 MHz in CD3OD, unless
noted otherwise.
Cmpd. J(1,2) or J(2,3) or J(3,4) or J(4,5) or J(4,5’) or J(1,F) or J(3,F) or 2J(2,F) or Others
J(1’,2’) J(2’,3’) J(3’,4’) J(4’,5’) J(4’,5’’) J(1’,F)
J(3’,F)
2J(2’,F)
4
< 1.03.7
< 1.03.7
3.1
8.3
8.4
3.7
n.d.
4.4 .3 10
27.6
10.1
16.5
52.8
27.8
52.2
53.9
J(5’,5’’) ¼ 12.0
b
4a)
5
)
n.d.
16.5
3.018.0 21.7
53.4
4.5
3.5
6.7
8.1
3.03.1
2.4
J(5’,5’’) ¼ 12.8
J(5’,5’’) ¼ 12.5
J(HÀC(8),F) ¼ 2.3
J(5’,5’’) ¼ 12.2
J(5’,5’’) ¼ 12.3
J(8,F) ¼ 2.3
6
3.5
7
8
< 1.04.0
3.5
9.4
3.9
2.2
8.2
2.9
2.3
18.0 26.0 52.5
3.5
19.6
24.054.2
12a)
< 1.03.6
2.8
6.8
4.2
4.2
[
J(1,F(2) [J(3,F(2) 52.8
J(5,5’) ¼ 11.7
¼ 9.6]
¼ 21.0
51.6
J(1,F(3)) ¼ 1.8
J(4,F(2)) < 1.0
[J(2,F(3) [J(4,F(3)
< 2.0]
4.016.0 18.2
¼ 16.2]
15
3.6
6.2
2.8
53.0
J(3’,OH) ¼ 5.8
J(5’,OH)
¼ J(5’’,OH) ¼ 5.5
J(5’,5’’) ¼ 13.0
J(3’,OH) ¼ 6.2
J(5’,OH)
16
3.9
4.07.02.5 .504.5148.0.200
¼ J(5’’,OH) ¼ 5.0
J(5’,5’’) ¼ 12.5
J(8,F) ¼ 3.0
17
2.5
4.2
7.02.5
3.5 5.03.0162.00
J(3’,OH) ¼ 6.0
J(5’,OH)
¼ J(5’’,OH) ¼ 5.5
J(5’,5’’) ¼ 12.8
J(5’,5’’) ¼ 12.6
J(5’,5’’) ¼ 12.5
J(8,F) ¼ 2.0
18
19
3.4
3.9
4.7
4.06.9
5.7
2.2
2.2
3.5
3.1
16.2
15.7
19.2
14.8
54.1
53.3
51.9
20
< 1.04.3
10.5
[J(2’,3’’) [J(3’’,4’)
< 1.0]
3.3
4.2
18.3
42.1
J(5’,5’’) ¼ 12.1
J(3’,3’’) ¼ 14.8
[J(3’’,F)
¼ 20.8]
22.5
¼ 5.6]
21
0.9
0.9
4.0
8.0
2.0
2.5
3.5
17.0
18.0
53.0
52.8
J(C(5),C(6)) ¼ 7.5
J(5’,5’’) ¼ 12.5
J(C(5),C(6)) ¼ 7.5
J(5’,OH)
¼ J(5’’,OH) ¼ 5.0
J(5’,5’’) ¼ 12.5
J(3’,OH) ¼ 6.3
J(5’,OH)
21c)
4.2
3.4
7.8
n.d.
22.8
22c)
1.2
8.7
< 1.5
< 1.9
17.4
20.2
53.2
¼ J(5’’,OH) ¼ 4.6
J(5’,5’’) ¼ 12.4
J(5’,5’’) ¼ 12.4
J(5’,5’’) ¼ 12.2
23
< 1.04.4
3.1
8.4
2.5
4.9
21.0 23.6
4.4.020 17.3
53.4
53.6
24c)
4.9
6.6
2.1
a) In CDCl3 solution. b) Not determined. c) In (D6)DMSO solution.

Helvetica Chimica Acta – Vol. 90 (2007)
1825
Table 3. Selected 13C-NMR Chemical Shifts (in ppm) and J(C,H) and J(C,F) Coupling Constants (in Hz). Recorded at
125 MHz in CD3OD, unless noted otherwise.
Cmpd. C(1) or
C(2) or
C(2’)
C(3) or
C(3’)
C(4) or C(5)
C(2)
[J(C(2), {J(C(4),
C(4)
C(5)
C(6)
C(8)
C(1’)
C(4’)a) or
[J(C(6), [J(C(8),F)]
H(1’)]b) {J(C(8),
H(1’))}c)
[2J(C(1’),F)] [2J(C(2’),F)] [2J(C(3’),F)]
C(5’) H(1’))]b) H(1’))}c)
4d)
4e)
5
106.9 (d)
[29.5]
105.3 (d)
[29.2]
88.7 (d)
[33.5]
83.7 (d)
[15.6]
87.5 (d)
[23.0]
84.5 (d)
[15.4]
84.9 (d)
[32.5]
81.4 (d)
[15.6]
87.4 (d)
[32.7]
85.7 (d)
[32.7]
83.3 (d)
[15.8]
88.4 (d)
[35.8]
90.7 (d)
[34.1]
88.0( d)
[33.4]
87.2 (d)
[34.2]
89.7 (d)
[35.5]
95.6 (d)
[184.1]
95.6 (d)
[186.1]
95.2 (d)
[190.9]
91.4 (d)
[194.8]
96.3 (d)
[193.1]
92.1 (d)
[191.6]
93.6 (d)
[186.9]
90.5 (d)
[193.6]
94.1 (d)
[186.8]
93.3 (d)
[186.9]
90.81 (d)
[194.1]
96.8 (d)
[177.6]
95.4 (d)
[186.4]
93.9 (d)
[185.2]
93.5 (d)
[185.7]
95.4 (d)
[182.5]
92.5 (d)
[184.6]
62.0( d)
[15.7]
60.7 (d)
[15.8]
60.5 (d)
[15.9]
60.4 (d)
[15.1]
51.5 (d)
[17.9]
53.5 (d)
[17.1]
67.9 (d)
[16.5]
69.2 (d)
[16.0]
67.1 (d)
[16.0]
68.2 (br. d) 84.1
[18.9]
70.1 (d)
[16.2]
31.9 (d)
[20.3]
69.1 (d)
[16.9]
67.1 (d)
[15.0]
67.4 (d)
[16.2]
79.5
65.2 (s) –
s) –
–
–
–
–
–
78.06(4.0
–
–
–
84.061.8 (
82.5.06(0
s) 153.8
s) 152.8
150.0
149.1
148.7
148.8
121.0 157.4
118.0156.0 1.140
119.2 156.5
118.5 156.1
141.3
6
[4.9]
139.8
7
84.3
83.8
83.6
83.5
83.5
60.1 (s) 152.7
60.9 (s) 152.6
8
140.4
[3.3]
135.1
{4.0}
136.2
[6.5] {5.0}
15f)
16f)
17f)
18f)
19
60.0
60.4
59.6
153.7
153.7
154.3
150.6g) 116.6g) 156.6
{2.0}
150.8g) 115.6g) 156.5
{ < 1.5}
160.4h) 107.2h) 153.0141.3
{2.0}
60.3 (s) 152.4
148.7h) 119.0g) 155.8
{3.8}
139.4
{3.8}
140.7
[4.2]
84.3
81.3
84.3
82.6
83.3
84.3
82.8
60.8
61.5
60.9
59.2
59.4
62.5
60.8
152.5
152.6
149.0
148.6
167.7
165.6
163.1
149.5
170.8
118.3 155.9
20f)
21
119.0156.0 138.6
157.9
[ < 1.5]
154.8
[2.0]
95.9i) 142.6i)
[4.8]
–
–
–
–
–
21f)
22f)
23
93.8
141.0
[2.8]
150.3
101.6 140.4
70.8 (d)
[16.6]
69.0( d)
[16.0]
158.2
153.3
137.7
36.8
–
24f)
86.9 (d)
[33.6]
31.2
a) In all compounds, the 3J(C(4’),2(F)) or 3J(C(4’),2’(F)) values were < 1.5 Hz. b)c) Coupling constants of
pyrimidine and purine nucleosides, resp. d) MeO Resonance at d(C) 55.2 (CD3OD) and 55.5 (CDCl3); benzoyl
(Bz) group: 167.61 (C¼O); 134.49 (p-C); 131.06 (ipso-C); 130.70, 129.62 (o- and m-C). e) In CDCl3. f) In
g
1
(D6)DMSO. ) J(C(5),H(8)) (11.2 and 11.0Hz) and J(C(4),H(8)) (4.5 and 3.8 Hz) were observed in the H-
coupled 13C-NMR spectra of 15 and 18, resp. h) J(C(4),H(8)) (13.0Hz) and the br. s of C(5) were observed in
i
the H-coupled 13C-NMR spectrum of the N7-b-isomer 17. ) J(C(5),H(6)) (3.0Hz), J(C(4),H(5)) (2.0Hz),
1
and J(C(4),H(6)) (9.0Hz) were observed in the 1H-coupled 13C-NMR spectrum in CD3OD.
region (Fig. 1). Another unusual example of anomeric pairs having a CD transition of
the same sign had been reported earlier by Ingwall [38] for the a- and b-pairs of the N9-
d-xylo- and N9-d-arabinofuranosyl nucleosides of adenine.

1826
Helvetica Chimica Acta – Vol. 90 (2007)
Fig. 1. Circular dichroism spectra of compounds 5 – 8 in EtOH
The shape of the CD curve of the b-anomer 18 was anomalous, being unexpectedly
similar to that of 9-(b-d-xylofuranosyl)adenine [38], with an additional broad positive
Cotton effect centered at 292 nm (Fig. 2). The dependence of the CD spectrum of the
nucleosides on the syn vs. anti orientation of the heterocyclic base about the glycosyl
bond was repeatedly delineated (see, e.g., [38], and refs. cit. therein). To assess the base
orientation, we measured the 3J(C(8),HÀC(1’)) and 3J(C(4),HÀC(1’)) coupling
constants in the 13C-NMR spectrum of the b-anomer 18 by means of broad-band
decoupling of the 1H-NMR resonances (Table 3). The value of 3.8 Hz for both
coupling constants clearly point to the high rotational mobility of the adenine base
about the glycosyl bond [39][40], which is somewhat unexpected, because there are
numerous indications for predominant population of the anti orientation of purine
bases about this bond (compare with the data for some other nucleosides in Table 3). It
is also noteworthy that the ribofuranose ring of the b-anomer 18 exists in an N/S 68 :32
pseudorotational equilibrium (Table 4), as discussed below. One can suppose that these
two factors, i.e., high mobility of the base about the glycosyl bond and the ribofuranose
ring, are responsible for the observed anomalous CD curve of 18.
The CD spectrum of the a-anomer 19 was very similar to that of the a-3’-amino-
2’,3’-dideoxy-2’-fluoro nucleoside 8, the former displaying positive Cotton effects more
shifted to the far UVand a very stout Cotton effect at 215 nm (Fig. 2). Further, the CD
spectrum of the difluoride 13 was diffuse from 320to 220nm, and showed a poorly
resolved, negative Cotton effect at 268 nm, and another negative CD band at 215 nm
(Fig. 2).
It was previously discussed that the broad negative and positive Cotton effects of
the respective b-d- and a-d-pentofuranosyl adenine nucleosides at ca. 260nm contain
both the B2u and B1u transitions, together with an n ! p* transition [38]. The adequate
resolution of the B2u and B1u bands in the CD spectra of adenine nucleosides was not
observed. However, theoretical considerations and analysis of the shapes of the UVand

Helvetica Chimica Acta – Vol. 90 (2007)
1827
Fig. 2. Circular dichroism spectra of compounds 13, 18, and 19 in H2O
CD envelopes revealed that the Cotton effects are controlled by many factors,
especially the base orientation about the glycosyl bond and the puckered conforma-
tions of the ribofuranose ring. The results of the present work, thus, point to an essential
contribution of the electronic character of substituents at C(2’) and C(3’) to the sign
and rotational strength of positive or negative Cotton effects, especially in the case of
the a-anomers.
3. Conformational Analysis. The pseudorotational parameters of the compounds
investigated are presented in Table 4. The conformational analysis of all compounds
was performed employing the PSEUROT (version 6.3) program [33], with simulta-
neous analysis of vicinal H,H and H,F couplings, as described previously [35]. The
pseudorotational parameters of compounds 5 – 8 and 15 – 17 were compared with those
for the closely related b-anomeric 2’-deoxy-2’-fluoroadenosine (18) and its a-anomer
19 [19], as well as the 2’,3’-dideoxy-2’-fluoro derivative 20 [41]. The data for the b-
nucleoside 5 were compared with the pseudorotational parameters for the methyl
glycoside 4. Moreover, to gain further insight into the role of the heterocyclic base in
the N/S equilibrium of the 2’-deoxy-2’-fluoro-b-d-pentofuranose ring, conformational
analysis of the 2’-deoxy-2’-fluoro pyrimidine nucleosides 21 – 24 was performed
(Table 4).
All calculations were performed such as to achieving the minimal r.m.s. deviations
and j DJ j values. Scale factors for 3J(H,H) and 3J(H,F) were 1.0and 0.2, respectively.
Moreover, we calculated the geometry-minimized structures for all the compounds
studied, using the MM þ and AMBER force fields, and the results were compared with

1828
Helvetica Chimica Acta – Vol. 90 (2007)
Table 4. Pseudorotational Parameters of the Compounds Investigated. Recorded in CD3OD, unless noted
1
1
otherwise. All the PSEUROT data include the H,1H and H,F couplings. The P and n values given in
underlined italics were kept fixed during the final minimization.
Cmpd.
PN
[8]
nmaxN
[8]
PS
[8]
nmaxS
[8]
r.m.s.
[Hz]
j DJ j
S
[%]
[Hz]
4
5.5 (32 T)
41.3
(42.8
37
44
43.4
44
41(.0216.0
39.9
44.7
37.9
40.7
37
44
44
40
36.7
38.1
37.2
46.7
34.9
33.1
183.5 (32 T)
216 (43 T)
126.6 (1E)
133.2 (1T2)
108 (01 T)
107.7 (01 T)
43 T)
216 (43 T)
180.6 (23 T)
198.9 (3E)
126.4 (1E)
28
40
44.6
44
44
44
44
44
36
44
36
43.6
44
44
44
44
44
44
44
40
36
0.05
0.02
0.01
0.02
0.02
0.01
0.19
0.16
0.03
0.04
0.03
0.02
0.03
0.01
0.06
0.05
0.04
0.04
0.02
0.09
0.01
0.10
0.03
0.02
0.07
0.04
0.03
0.40
0.29
0.08
0.09
0.09
0.06
0.04
0.02
0.09
0.09
0.08
0.08
0.07
0.15
0.03
10
10
29
16
10
20
9
4a)
5
11.0 3T2)
À 8.0( T3)
2
6
7
8
1.6 (32 T)
25.5 (3T4)
À 5.8 (3 T)
2
12a)
[H,F]
15b)
16b)
17b)
18
À 11.8 (2T3)
À 10.3 (2T3)
3.8 (32 T)
7
31
14
24
32
39
32
38
5
5
5
7
1
4.3 (32 T)
1.5 (32 T)
À 18.6 (2E)
22.2 (3E)
À 1.5 (32 T)
À 26.8 (2T1)
4.6 (32 T)
16.5 (3E)
5.9 (32 T)
16.9 (3E)
17.6 (3E)
27.4 (3T4)
125.0( E)
1
18c)
19
210.6 (43 T)
127.0( E)
1
19c)
20
197.5 (3E)
90 (0E)
21
192.8 (3E)
201.3 (3E)
136.3 (1T2)
198 (3E)
21b)
22b)
23
24b)
149.1 (21 T)
24
a) In CDCl3. b) In (D6)DMSO. c) The relevant H,H and H,F coupling constants were taken from the
NMR spectra recorded in (D6)DMSO [19].
those resulting from the PSEUROT calculations. Both force fields afford essentially the
same spatial arrangement of the pentofuranose ring for each individual compound, and
the outcome of the PSEUROT and molecular-mechanics calculations were mutually
consistent as far as the geometry of the most-populated conformer is concerned. Thus,
the C(4’)ÀC(3’)ÀC(2’)ÀF torsion angles for the most-populated conformers included
in Table 4 were within 77 – 838, which is in fair agreement with the aforementioned
3J(C(4’),F) value of < 1.5 Hz.
In terms of the antiperiplanar effect (APE) proposed by Brunck and Weinhold
[36], and as discussed in detail earlier [35], predominant population of the N-type 2T3
conformer (71%) of the ribofuranose ring of the azide 5 results mainly from the strong
APE for FÀC(2’)ÀC(3’)ÀH (torsion angle V ꢁ 1708) and the N-driving anomeric
effect (ANE) of the base. The APE for N3ÀC(3’)ÀC(2’)ÀC(1’) (V ꢁ 1658) and the
APE for OÀC(1’)ÀC(2’)ÀH (V ꢁ 1508) also contribute to the N-conformer popula-
tion. The APEs for N3ÀC(3’)ÀC(2’)ÀH and FÀC(2’)ÀC(3’)ÀC(4’) (V ꢁ 1508 each),
favoring the S-form (1E), are weaker in comparison with the strong N-driving APEs in
the 2T3 conformation.

Helvetica Chimica Acta – Vol. 90 (2007)
1829
It is remarkable that the conformational equilibrium of the pentofuranose rings of
the azide 5 and the fluorides 15 and 12 are very similar1), implying a similar influence of
the 3’-OH and 3’-N3 functions on the S/N population. On the contrary, the S-type
conformation (2E) prevails in 2’,3’-dideoxy-2’,3’-difluoroadenosine (13) [35]; interest-
ingly, the most APEs mutually cancel in the S- vs. N-form, except for a strongly S-
driving APE for FÀC(3’)ÀC(4’)ÀC(5’), counteracting the N-driving ANE. It should
be stressed that the anomeric effect of the heterocyclic base diminishes with increasing
electronegativity of the 2’-substituent, and in the b-d-ribonucleosides it is lower than
the ANE in its 2’-deoxy congeners [42] (for a review, see [43]).
The population of the prevailing N-type conformers as well as the geometries of the
a-anomers 6, 8, 16, and 19 are similar. There is a close resemblance of the N-type
conformers of 6 and 8, and, as a consequence, of their population. On the contrary, the
reason for decreased population of the N-type conformer of the adenine a-anomer 19
vs. the related guanine nucleoside 16 most likely is due to variable strength of numerous
related APEs, depending on the geometry of the 3E vs. 1E segments of the
pseudorotational wheel. On the whole, the S/N equilibrium of the b- and a-anomers
18 and 19 appears to be similar, with predominant population of the N-type rotamers,
i.e., 68% in CD3OD, and ca. 62% in (D6)DMSO. However, some peculiarities of this
equilibrium are notable. The S-type conformers of the 18/19 pair in each individual
solvent are analogous, but the occupied segments of the pseudorotational wheel, 1E in
CD3OD, and E in (D6)DMSO, are rather different. The N-type conformers of both
3
anomers in the studied solvents demonstrate remarkable diversity, which, however, is
not reflected on the population level.
In the case of the a-anomer 6, the N-type (32 T) conformation issued from a
calculation was found to be more populated (84%) compared to that of the b-
counterpart 5. By going from the b- to the a-anomer, the removal of ANE in the N-
conformer is compensated by the emerging strong N-driving APE for
FÀC(2’)ÀC(1’)ÀH (V ꢁ 1708), far overweighing the S-driving ANE of the S (1T2)
conformation. In addition, a slight 1E (5) ! 1T2 (6) change of the S-type sugar
conformer resulted in the clinal arrangement of the 2’-F-atom and the 3’-H-atom, which
cannot be compensated by the emerging APE for N3ÀC(3’)ÀC(2’)ÀH (V ꢁ 1608),
leading to a reduced percentage of the latter conformation.
The close similarity of the minor S-type conformers of the anomeric pairs 5/6, 7/8,
15/16, and 18/19 is remarkable. On the other hand, the b ! a transfer of anomeric
configuration results in the decrease of the S-type conformation in the case of the 5/6
pair (in CD3OD) and the 15/16 pair (in (D6)DMSO), remains unchanged in the 18/19
pair (in both solvents), and is displaced in opposite direction in the 7/8 pair (in
CD3OD). The most plausible explanation of such diversity of related anomeric pairs
consists in the interplay of subtle differences of APEs and ANEs of the individual
conformers. At the same time, differential stabilization by the solvent cannot be
excluded.
The MM þ calculation yielded for these three compounds the 3E conformation, with APEs for
FÀC(2’)ÀC(3’)ÀH (V ꢁ 1708), HOÀC(3’)ÀC(2’)ÀC(1’) (V ꢁ 1658), and OÀC(1’)ÀC(2’)ÀH
(V ꢁ 1458), resp.
1
)

1830
Helvetica Chimica Acta – Vol. 90 (2007)
The b-nucleoside 7 exists in the N (3T4) > S (10 T) equilibrium, with a strong
preference for the former. The strong N drive is due to an APE for FÀC(2’)ÀC(3’)ÀH
(V ꢁ 1658) and an ANE, whereas in the S-type conformation there are no trans
arrangements between an F-atom and HÀC(3’) and C(4’). By going from the b-anomer
7 to its a-counterpart 8, the geometry of the prevailing N-conformation slightly changes
3
(3T4 ! 2T) and, contrary to the b,a-pair 5/6, gets less populated. The difference in the S-
type population between the a-anomers 6 and 8 may be connected with the more
stabilizing APE for FÀC(2’)ÀC(3’)ÀH in the 01T conformation (V ꢁ 1608) vs. that of
the T2 conformation (a clinal arrangement of 2’-F and 3’-H) as well as by the
1
contribution of an APE for H2NÀC(3’)ÀC(4’)ÀC(5’) (V ꢁ 1608).
As mentioned above, 2’,3’-dideoxy-2’,3’-difluoroadenosine (13) exists in the N
(3E) > S (2E) equilibrium, with a strong preference (85%) for the latter [35].
Assuming similar strengths of the APE for FÀC(2’)ÀC(3’)ÀH and APE for
FÀC(3’)ÀC(2’)ÀC(1’) in the N-form on one hand, and of the APE for
FÀC(3’)ÀC(2’)ÀH and APE for FÀC(2’)ÀC(3’)ÀC(4’) in the S-form on the other
hand, we hypothesized that the prevailing S-pseudorotamer results from the interplay
between the N-driving ANE and the S-driving APE for FÀC(3’)ÀC(4’)ÀC(5’). Taking
into account that the ANE of the heterocyclic base diminishes with increasing
electronegativity of the 2’-substituent [42], one can expect its remarkable weakening in
going from b-d-ribonucleosides to their 2’-deoxyfluoro derivatives. The results of the
present study, thus, further support this suggestion. Indeed, replacement of the 3’-ribo
F-atom with OH, N3, or NH2 groups resulted in a consecutive decrease of the S-rotamer
population, originating from the dramatic drop of electronegativity and, as a
consequence, by weakening of the APEs for XÀC(3’)ÀC(2’)ÀH and
XÀC(3’)ÀC(4’)ÀC(5’) (X ¼ OH, N3, NH2). As might be expected, the N-type
conformation becomes dominating in the case of 9-(2,3-dideoxy-2-fluoro-b-d-erythro-
pentofuranosyl)adenine (20) (Table 4).
The strong preference for the N-type rotamer (90%) of the methyl glycoside 4
compared to the nucleoside 5 is in accord with the stronger anomeric effect of the MeO
group vs. the adenine base [44]. It is noteworthy that the N (32 T) > S (32 T) equilibrium
of 4 was found to be analogous in CDCl3 and in MeOH. Moreover, there is a close
resemblance of the pseudorotational parameters of the N/S equilibrium of the methyl
b-d-glycosides 4 and 12 in CDCl3 solution (Table 4). The dominating population of the
N-type rotamers appears to be opposite to predictions based on the conventional
gauche effect (GE). Indeed, the GE for FÀC(2’)ÀC(1’)ÀO(4’) in the N-form should
balance the GE for FÀC(3’)ÀC(4’)ÀO in the S-form, but there remains the strongly S-
driving GE for FÀC(2’)ÀC(1’)ÀOMe, which counteracts the N-driving ANE of the
MeO group. We have earlier shown that the GE for FÀC(3’)ÀC(4’)ÀO is much larger
than the ANE for the MeO group [35]; and one can expect a similar relationship
between the aforementioned competing GE for FÀC(2’)ÀC(1’)ÀOMe and the ANE
for MeO, displacing the N/S equilibrium to the S-type rotamer.
The dominating population of the N-type conformers of glycosides 4 and 12 is in
harmony with the APE proposed by Brunck and Weinhold [36]. In the case of the
difluoride 12, the APEs for FÀC(2’)ÀC(3’)ÀH, for HÀC(2’)ÀC(1’)ÀO, and for
FÀC(3’)ÀC(2’)ÀC(1’) in the N-form balance the APEs for FÀC(3’)ÀC(2’)ÀH, for
HÀC(3’)ÀC(4’)ÀO, and for FÀC(3’)ÀC(4’)ÀC(5’) in the S-form, respectively. As a

Helvetica Chimica Acta – Vol. 90 (2007)
1831
result, only one weak APE, that for C(3’)ÀC(2’)ÀC(1’)ÀOMe, remains in the S-form,
and the ANE for the MeO group assumes the predominant role, so that the N-form is
strongly favored. Similar cancellations are operating in the case of 4, and the N/S
equilibrium is dominated by the strong N-driving ANE for the MeO group.
Consideration of the conformational behavior of a number of compounds studied
using the viewpoint of Brunck and Weinhold on the gauche effect (GE) [36] appears to
be satisfactory. We have, however, earlier noted that, in some cases, the calculated
pseudorotational equilibrium can be rationalized by the standard model of the GE [44].
For example, the conformational behavior of 13 results from the balancing of all GEs,
except for a strongly S-driving GE [for FÀC(2’)ÀC(1’)ÀN(9)]2), counteracting the
weak N-driving ANE [36].
Recently, the conformations of the erythro- and threo-configured diastereoisomers
of 1,2-difluorodiphenylethanes and 2,3-difluorosuccinic acids and their derivatives,
with the general formula RCHF – CFHR, were investigated and discussed in detail
[45]. The predominant population of these molecules with a gauche arrangement of F-
atoms (torsion angle ca. 608) was rationalized on the basis of the so-called ꢁfluorine
gauche effectꢂ. It is interesting to note that, in most cases, the gauche arrangement of F-
atoms can be rationalized using the viewpoint of Brunck and Weinhold [36]. There, the
relative stabilization of the gauche arrangement of vicinal F-atoms originates from the
approximately antiperiplanar orientation of the most electronegative (F, COOÀ, etc.)
and the least electronegative ligand (H, Ar).
The very important role of the electronic nature of the heterocyclic base in terms of
the N/S equilibrium is well-documented (see, e.g., [46][47]; for a review, see [43]). We
have also noticed in our previous studies essential differences in the ratio of the N/S
population, depending on the type of substituent at the anomeric C-atom, e.g., MeO vs.
heterocyclic base [20][35] (see above), or pyrimidine vs. purine base [35], as well as on
the interplay of stereoelectronic effects of the 2’-deoxy-2’-fluoro-b-d-arabinofuranosyl
moiety of nucleosides and heterocyclic base [48][49]. In extension of this work, we
studied the role of the electronic structure of the heterocyclic base of 2’-deoxy-2’-
fluoro-b-d-ribofuranosyl nucleosides in the N/S equilibrium.
The conformational behavior of the pentofuranose rings of the adenine nucleosides
5 and 18 (in CD3OD) and of the regioisomeric guanine nucleosides 15 and 17 (in
(D6)DMSO) is rather similar. The prevailing N-type conformation of this group of b-d-
nucleosides occupies a rather narrow segment (2T3 > 2E) of the pseudorotational
wheel, and the S-type E conformation is populated in compounds 5, 17, and 18.
1
3
2
Interestingly, the T conformation is the minor and dominating one in the case of 13
and 20, respectively. On the other hand, the pentofuranose ring of the pyrimidine
nucleosides 21 – 23 displays a strong preference for the N-conformation, characterized
3
by a narrow E > 32T segment of the pseudorotational wheel, quite similar to the
aforementioned segment of the N-type conformations of purine nucleosides. Note that
2’-deoxy-2’-fluorocytidine (21) manifests the same N/S equilibrium in CD3OD and
(D6)DMSO solutions.
Reduction of the C(5)¼C(6) bond of 2’-deoxy-2’-fluorouridine (22) gave rise to
nucleoside 24, with increased population of the S-conformation, but retention of its
2
2
)
MM þ Calculations yielded for the E conformation a V value of ca. 788.

1832
Helvetica Chimica Acta – Vol. 90 (2007)
geometry. It appears to be reasonable that the alteration of the pseudorotational
equilibrium by going from nucleoside 22 to its hydrogenated derivative 24 results from
the change in electronic structure of the base.
Conclusions. – We have described the synthesis of a number of purine 2’-deoxy-2’-
fluoro-d-ribofuranosyl nucleosides with variable substituents (N3, NH2, F, OH, H) at
the 3’-position in the ribo [a-C(3’)] configuration. The methyl glycosides 4 and 12,
containing a 2’-a-F-atom, showed very low reactivity, remarkably lower compared to
previously studied methyl glycosides with an azido or amino function at this position
[29][30].
In order to gain further insight into the role of different stereoelectronic factors in
the N/S equilibrium of the pentofuranose rings, we analyzed the conformational
behavior of a number of 2,3-dideoxy-2-fluoro-3-X-d-ribofuranosides (X ¼ F, N3, NH2,
H), as well as of some purine and pyrimidine 2-deoxy-2-fluoro-d-ribofuranosyl
nucleosides, using the PSEUROT (version 6.3) software. It was found that the
pentofuranose rings of the purine nucleosides are predominantly in the N-type
conformations, the degree of population of which is decreased with increasing
electronegativity of the 3’-a-substituent (H, 95%; NH2, 90 %; N, 71%; OH, 68%; F,
3
15%). The N-type E > 32T conformations are dominating (> 93%) in the case of
3
pyrimidine 2’-deoxy-2’-fluoro-b-d-ribofuranosyl nucleosides. The observed pseudoro-
tational behavior can be rationalized in terms of the antiperiplanar effect proposed by
Brunck and Weinhold [36].
We are grateful to Prof. Cornelis Altona (Leiden University, The Netherlands) for numerous fruitful
suggestions and discussions of conformational peculiarities of nucleosides. We are thankful to Dr. Natalja
B. Khripach (Institute of Bioorganic Chemistry) and Jarkko Roivainen (University of Kuopio) for
recording NMR spectra, and to Dr. Nicolai I. Garbuz (Institute of Bioorganic Chemistry) for recording
CD spectra. The authors are indebted to Prof. Frank Seela (Universität Osnabrück, Germany) for a
sample of 2’-deoxy-2’-fluoro-6-azauridine (23). I. A. M. thanks the Technology Development Center
(TEKES) of Finland for partial financial support of this study.
Experimental Part
1. General. 2’-Deoxy-2’-fluoroadenosine (18) and its a-anomer 19 were prepared as described earlier
[19]. 9-(2,3-Dideoxy-2-fluoro-b-d-erythro-pentofuranosyl)adenine (20) was prepared essentially as
described in [41]. 2’-Deoxy-2’-fluorocytidine (21) was prepared according to [50], and its deamination
[27] gave 2’-deoxy-2’-fluorouridine (22). Catalytic hydrogenation of 22 (0.25 g, 1.0 mmol) was carried out
in water (40ml) at r.t. and atmospheric pressure for 55 min in the presence of 5% Rh on activated
alumina as catalyst (Fluka), as reported in [51], to afford 2’-deoxy-2’-fluoro-5,6-dihydrouridine (24) as a
viscous oil in quant. yield. 6-Aza-2’-deoxy-2’-fluorouridine (23) was kindly supplied by Prof. Frank Seela,
Universität Osnabrück, Germany.
Unless noted otherwise, all reactions were carried out at r.t. (208). For workup, org. solns. were dried
over anh. Na2SO4 for 4 h. Column chromatography (CC): silica gel 60 H (70– 230 mesh; Merck,
Darmstadt, Germany), unless otherwise indicated. TLC: aluminum-backed silica gel 60 F254 sheets
(Merck, Germany); eluents: hexane/AcOEt 3 :1 (A), CHCl3/MeOH 10:1 ( B), CHCl3/MeOH 4 :1 (C),
or CHCl3/MeOH/H2O 40 :10 :1 D( ). M.p.: Boetius apparatus; uncorrected. UV Spectra: Carl-Zeiss
Second M-400 apparatus; lmax (e) in nm. [a]2D0 Values: Jusco J-20 spectropolarimeter; concentration in g/l.
CD Spectra: Jasco J-20 spectropolarimeter; l in nm, [V]l in deg cm2 dmolÀ1. IR Spectra: Carl-Zeiss UR-
20 apparatus; in cmÀ1. 1H- and 13C-NMR Spectra: Bruker Avance-500-DRX spectrometer, at 500.13 and

Helvetica Chimica Acta – Vol. 90 (2007)
1833
125.76 MHz, resp., at 300 K; d in ppm rel. to Me4Si, J in Hz. All NMR assignments were confirmed by 2D
1
(1H,1H and H,13C) correlation spectroscopy. Solvents employed for NMR spectra were CD3OD and
(D6)DMSO, unless stated otherwise.
2. Methyl 3-Azido-5-O-benzoyl-2,3-dideoxy-2-fluoro-b-d-ribofuranoside (4). 2.1. Methyl 5-O-
Benzyl-2,3-dideoxy-2-fluoro-3-O-[(4-methylphenyl)sulfonyl]-b-d-ribofuranoside (2). Reaction of xylo-
side 1 (3.5 g, 13.65 mmol) [30] with TsCl (5.0 g, 27.0 mmol) in anh. pyridine (20 ml) for 72 h, followed by
purification by CC (SiO2; 0 ! 20% AcOEt in hexane) afforded 2. Yield: 4.9 g (87%). M.p. 50– 52 8
(Et2O/hexane). TLC (A): Rf 0.42. [a]D20 ¼ À33.0( c ¼ 1.0, CHCl3). Anal. calc. for C20H23FO6S (410.46): C
58.52, H 5.65; found: C 58.41, H 5.69.
2.2. Methyl 5-O-Benzoyl-2,3-dideoxy-2-fluoro-3-O-[(4-methylphenyl)sulfonyl]-b-d-ribofuranoside
(3). Compound 2 (4.9 g, 11.94 mmol) was catalytically hydrogenated in the presence of 10% Pd/C
(4.9 g) in EtOH (250ml), followed by standard benzoylation. Yield: 4.3 g (85%). M.p. 80– 82 8 (Et2O/
hexane). TLC (A): Rf 0.36. [a]2D0 ¼ À41.5 (c ¼ 1.0, CHCl3). Anal. calc. for C20H21FO7S (424.44): C 56.60,
H 4.99; found: C 56.70, H 4.87.
2.3. Conversion to 4. To a soln. of 3 (0.81 g, 1.91 mmol) in anh. DMSO (10 ml), NaN3 (0.6 g,
9.23 mmol) was added, and the mixture was heated at 190– 195 8 (bath temp.) for 30min, cooled, poured
into H2O (10ml), and extracted with AcOEt (4 ꢂ 40ml). The combined org. extracts were dried and
concentrated, and the residue was subjected to CC (SiO2; AcOEt/hexane 1:15). Yield: 0.29 g (51%).
Viscous oil that slowly crystallized on cooling. M.p. 31 – 328. TLC (A): Rf 0.64. [a]2D0 ¼ þ2.0( c ¼ 1.0,
CHCl3). IR (KBr): 1735 (C ¼ O), 2120(N ). Anal. calc. for C13H14FN3O4 (295.27): C 52.88, H 4.78, N
3
14.23; found: C 52.91, H 4.69, N 14.29.
3. 3’-Azido-2’,3’-dideoxy-2’-fluoroadenosine (5) and Its a-Anomer 6. A mixture of 4 (0.26 g,
0.88 mmol), SnCl4 (0.49 ml, 4.19 mmol), and the bis(trimethylsilyl) derivative of N6-benzoyladenine3) in
anh. MeCN (7.0ml) was heated at reflux for 80min. After standard workup, the residue was purified by
CC (SiO2; CHCl3) to afford initial 4 (52 mg, 20%) and the benzoylated derivatives of 5/6 as a viscous oil
(0.2 g), the latter being directly used in the next step. Thus, debenzoylation in ammonia-sat. MeOH
(30ml) at 0 8 for 18 h and purification by CC (SiO2; 0 ! 9% MeOH in CHCl3) gave 70mg (34%) of
5
and 17 mg (8%) of 6.
Data of 5. M.p. 184 – 1868 (Et2O/EtOH). TLC (B): Rf 0.37. UV (EtOH): 259 (15,360). [a]2D0 ¼ þ2.5
(c ¼ 1.0, EtOH). CD (EtOH): 267 (À6,500), 235 (0), 290 (0). Anal. calc. for C10H11FN8O2 (294.25): C
40.82, H 3.77, N 38.08 , found: C 40.79, H 3.80, N 37.99.
Data of 6. M.p. 87 – 898 (Et2O/EtOH). TLC (B): Rf 0.25. UV (EtOH): 259 (14,300). [a]2D0 ¼ þ31.0
(c ¼ 1.0, EtOH). CD (EtOH): 213 (À 8,700), 230 (sh), 255 (þ 8,400), 220(0), 280(0). Anal. calc. for
C10H11FN8O2 (294.25): C 40.82, H 3.77, N 38.08; found: C 40.80, H 3.82, N 38.19.
4. 3’-Amino-2’,3’-dideoxy-2’-fluoroadenosine (7) and Its a-Anomer 8. Ph3P (30mg, 0.11 mmol) was
added to a soln. of 5 (20 mg, 0.068 mmol) in pyridine (0.13 ml), and the mixture was stirred for 4 h. Then,
conc. aq. ammonia (1.1 ml) was added, and stirring was continued for 40h. The mixture was
concentrated, H2O (3 ml) was added to the residue, the mixture was extracted with benzene (2 ꢂ 20ml),
and then concentrated in vacuo. The residue was purified by CC (SiO2 (Woelm, Germany); CHCl3,
CHCl3/MeOH 10:1, and CHCl 3/MeOH 4 :1) to afford 15 mg (82%) of 7. In a similar way, starting from
the a-anomer 6 (0.07 g, 0.238 mmol), 54 mg (85%) of 8 was obtained.
Data of 7. Oil. TLC (C): Rf 0.13. UV (EtOH): 260 (15,100). CD (EtOH): 220 (À 7,000), 260
(À4000), 214 (0), 284 (0). Anal. calc. for C10H13FN6O2 (268.25): C 44.77, H 4.88, N 31.33; found: C 44.59,
H 4.70, N 31.40.
Data of 8. Oil. TLC (C): Rf 0.11. UV (EtOH): 260 (14,900). CD (EtOH): 214 (À 7,600), 228
(þ1,600), 255 (þ 3,300), 210(0), 220(0), 285 (0). Anal. calc. for C 10H13FN6O2 (268.25): C 44.77, H 4.88, N
31.33; found: C 44.70, H 4.75, N 31.45.
5. Methyl 5-O-Benzyl-2,3-dideoxy-2,3-difluoro-b-d-ribofuranoside (10). Method A. To a soln. of 9
(1.2 g, 2.92 mmol) [30] in anh. DMSO (35 ml) was added freshly dried CsF (2.9 g, 19.09 mmol), and the
reaction mixture was stirred at 195 – 2008 (bath temp.) for 130min. After cooling to r.t., the mixture was
Obtained from 0.4 g (1.67 mmol) of N6-benzoyladenine and an excess of hexamethyldisilazane
3
)
(HMDS) in the presence of (NH4)2SO4 under reflux for 3 h.

1834
Helvetica Chimica Acta – Vol. 90 (2007)
poured into H2O (30ml) and extracted with AcOEt (4 ꢂ 100 ml). The combined org. extracts were dried
and concentrated, and the oily residue was subjected to CC (SiO2; 0 ! 33% AcOEt in hexane). Yield:
0.18 g (32%, based on consumed (75%) starting material). Oil. TLC (A): Rf 0.74.
Method B. To a soln. of 1 (0.6 g, 2.34 mmol) [30] in anh. toluene (8 ml) was added DAST (0.68 ml,
5.19 mmol), and the mixture was stirred for 20h. Then, the mixture was poured into 20% aq. NH 4HCO3
soln. and extracted with CH2Cl2 (3 ꢂ 100 ml). The combined org. extracts were dried and concentrated.
The residue was subjected to CC (SiO2; 0 ! 25% AcOEt in hexane). Yield: 0.103 g (30%, based on
consumed (57%) starting material).
6. Methyl 2,3-Dideoxy-2,3-difluoro-b-d-ribofuranoside (11). To a soln. of 10 (0.18 g, 0.7 mmol) in
anh. EtOH (4 ml) was added 20% Pd(OH)2/C (0.39 g) and freshly dist. cyclohexene (5 ml), and the
mixture was heated at reflux for 2 h. After cooling to r.t., the catalyst was filtered off and washed with
EtOH (50ml), and the combined filtrates were concentrated. The residue was purified by CC (SiO 2; 0 !
33% AcOEt in hexane). Yield: 0.11 g (95%). Oil. TLC (A): Rf 0.18.
7. Methyl 5-O-Benzoyl-2,3-dideoxy-2,3-difluoro-b-d-ribofuranoside (12). Prepared by standard
benzoylation of 11 (0.11 g, 0.66 mmol), followed by CC (SiO2; 0 ! 16%AcOEt in hexane). Yield: 0.16 g
(89%). Oily substance. TLC (A): Rf 0.71.
8. 2’,3’-Dideoxy-2’,3’-difluoroadenosine (13). A mixture of 12 (0.15 g, 0.55 mmol), SnCl4 (0.32 ml,
2.74 mmol), and the bis(trimethylsilyl) derivative of N6-benzoyladenine (obtained from 0.23 g
(0.96 mmol) of N6-benzoyladenine)3) in anh. MeCN (7.5 ml) was heated at reflux for 105 min. After
standard workup, the residue was purified by CC (SiO2; 0 ! 5%MeOH in CHCl3) to afford unreacted 12
(40mg, 27%) and the benzoylated derivative of 13 (0.15 g, with some impurities) as a viscous oil, which
was used in the next step without further purification. Thus, standard debenzoylation of the latter product
(25 ml ammonia-sat. MeOH, 08, 18 h), followed by CC (SiO2; 0 ! 6 % MeOH in CHCl3) gave 50mg
(45%) of 13. Amorphous powder. TLC (B): Rf 0.38. UV (H2O): 261 (14,700). CD (H2O): 215 (À 9,600),
250( À 3,400), 265 (À 5,000), 300 (À1,700), 210 (0). Anal. calc. for C10H11F2N5O2 (271.22): C 44.28, H
4.09, N 25.82; found: C 44.20, H 4.12, N 25.84.
9. Condensation of Trimethylsilylated N2-Palmitoylguanine and 1-O-Acetyl-3,5-di-O-benzoyl-2-
deoxy-2-fluoro-b-d-ribofuranose (14). A soln. of the acetate 14 (0.8 g, 2.0 mmol) [19], persilylated N2-
palmitoylguanine [obtained from N2-palmytoylguanine (0.78 g, 2.4 mmol) in analogy to the method
described above3)], and trimethylsilyl triflate (TMS-Tf; 0.69 mg, 0.56 ml, 3.1 mmol) in anh. MeCN
(45 ml) was heated at reflux for 3 h. After standard workup, the residue was purified by CC (SiO2; 0 !
5% MeOH in CHCl3) to afford the following pure acylated nucleosides (in order of elution): the N7-b-
isomer as the main product (0.39 g, 27%), the N9-b-isomer (0.18 g, 12%), and the N9-a-isomer (75 mg,
5%). [Condensation of the same reagents, but in a molar ratio of 1:1.7:2.4, in CH2Cl2 under reflux for 5 h
gave, after standard workup and CC, the above acylated nucleosides in yields of 27, 19, and 8%, resp.].
Finally, standard deacylation of the individual acylated isomers (25 ml ammonia-sat. MeOH, 08, 24 h)
gave the pure, deprotected nucleosides, i.e., 7-(2-deoxy-2-fluoro-b-d-ribofuranosyl)guanine (17), and 2’-
deoxy-2’-fluoroguanosine (15) and its a-anomer 16, in yields of 80– 85%.
Data of 15. M.p. 262 – 2648 (H2O) (lit. m.p. 262 – 2648 (H2O) [18]). TLC (D): Rf 0.38. UV (H2O):
252 (13,100), 276 (sh, 8,000). Anal. calc. for C10H12FN5O4 · H2O (303.24): C 39.61, H 4.65, N 23.09; found:
C 39.71, H 4.60, N 22.97. The 1H- and 13C-NMR data were in fair agreement with those previously
published [18][24].
Data of 16. M.p. 283 – 2858 (MeOH). TLC (D): Rf 0.24. UV (H2O): 253 (14,200), 276 (sh, 8,100).
Anal. calc. for C10H12FN5O4 (285.23): C 42.11, H 4.24, N 24.55; found: C 42.29, H 4.30, N 24.00.
Data of 17. M.p. > 2458 (H2O; dec.). TLC (D): Rf 0.3. UV (H2O): 217 (17,900), 240 (sh, 5,000), 287
(6,200). Anal. calc. for C10H12FN5O4 · 0.5 H2O (294.24): C 40.82, H 4.45, N 23.80; found: C 40.90, H 4.50,
N 23.70.

Helvetica Chimica Acta – Vol. 90 (2007)
1835
REFERENCES
[1] W. Guschlbauer, Acta Biochim. Pol. 1996, 43, 77.
[2] K. Krawiec, B. Kierdaszuk, E. N. Kalinichenko, E. B. Rubinova, I. A. Mikhailopulo, S. Eriksson, B.
Munch-Petersen, D. Shugar, Nucleosides Nucleotides Nucl. Acids 2003, 22, 153.
[3] H. Ford Jr., F. Dai, L. Mu, M. A. Siddiqui, M. C. Nicklaus, L. Anderson, V. E. Marquez, J. J. Barchi
Jr., Biochemistry 2000, 39, 2581.
[4] L. Mu, S. G. Sarafianos, M. C. Nicklaus, P. Russ, M. A. Siddiqui, H. Ford Jr., H. Mitsuya, R. Le, E.
Kodama, C. Meier, T. Knispel, L. Anderson, J. J. Barchi Jr., V. E. Marquez, Biochemistry 2000, 39,
11205, and refs. cit. therein.
[5] J. L. Clark, J. C. Mason, L. Hollecker, L. J. Stuyver, P. M. Tharnish, T. R. McBrayer, M. J. Otto, P. A.
Furman, R. F. Schinazi, K. A. Watanabe, Bioorg. Med. Chem. Lett. 2006, 16, 1712.
[6] J. Shi, J. Du, T. Ma, K. W. Pankiewicz, S. E. Patterson, P. M. Tharnish, P. R. McBrayer, L. J. Stuyver,
M. J. Otto, C. K. Chu, R. F. Schinazi, K. A. Watanabe, Bioorg. Med. Chem. 2005, 13, 1641, and refs.
cit. therein.
[7] M. Tisdale, M. Ellis, K. Klumpp, S. Court, M. Ford, Antimicrob. Agents Chemother. 1995, 39, 2454.
[8] E. I. Boreko, N. I. Pavlova, G. V. Zaitseva, I. A. Mikhailopulo, Voprosy Virusologii (Moscow) 2001,
46, 40 (Chem. Abstr. 2002, 137, 57010).
[9] F. Eckstein, Biochimie 2002, 84, 841.
[10] Y. Dorsett, T. Tuschl, Nat. Rev. Drug Discov. 2004, 3, 318.
[11] D. M. Dykxhoorn, J. Liberman, Annu. Rev. Med. 2005, 56, 40 1.
[12] N. S. Que-Gewirth, B. A. Sullinger, Gene Ther. 2007, 14, 283.
[13] J. M. Layzer, A. P. McCaffrey, A. K. Tanner, Z. Huang, M. A. Kay, B. A. Sullenger, RNA 2004, 10,
766.
[14] J. F. Codington, I. Doerr, D. Van Praag, A. Bendich, J. J. Fox, J. Am. Chem. Soc. 1961, 83, 50 30 ; J. F.
Codington, I. Doerr, J. J. Fox, J. Org. Chem. 1964, 29, 558.
[15] J. V. Tuttle, M. Tisdale, T. A. Krenitsky, J. Med. Chem. 1993, 36, 119.
[16] G. V. Zaitseva, A. I. Zinchenko, V. N. Barai, N. I. Pavlova, E. I. Boreko, I. A. Mikhailopulo,
Nucleosides Nucleotides 1999, 18, 687.
[17] K. Dax, M. Albert, J. Ortner, B. J. Paul, Carbohydr. Res. 2000, 327, 47.
[18] H. J. Thomas, K. N. Tiwari, S. J. Clayton, J. A. Secrist III, J. A. Montgomery, Nucleosides Nucleotides
1994, 13, 309.
[19] I. A. Mikhailopulo, G. G. Sivets, N. E. Poopeiko, N. B. Khripach, Carbohydr. Res. 1995, 278, 71.
[20] I. A. Mikhailopulo, G. G. Sivets, Helv. Chim. Acta 1999, 82, 2052.
[21] G. V. Zaitseva, G. G. Sivets, Z. Kazimierczuk, J. A. Vilpo, I. A. Mikhailopulo, Bioorg. Med. Chem.
Lett. 1995, 5, 2999.
[22] R. Ranganathan, Tetrahedron Lett. 1977, 18, 1291.
[23] M. Ikehara, H. Miki, Chem. Pharm. Bull. 1978, 26, 2449; M. Ikehara, J. Imura, Chem. Pharm. Bull.
1981, 29, 3281.
[24] A. M. Kawasaki, M. D. Casper, S. M. Freier, E. A. Lesnik, M. C. Zounes, L. L. Cummins, C.
Gonzalez, P. D. Cook, J. Med. Chem. 1993, 36, 831.
[25] B. S. Ross, R. H. Springer, K. G. Sprankle, G. Vazquez, Nucleosides Nucleotides 1997, 16, 1645.
[26] R. G. Schultz, S. M. Gryaznov, Nucl. Acids Res. 1996, 24, 2966.
[27] K. W. Pankiewicz, Carbohydr. Res. 2000, 327, 87.
[28] G. Kowollik, P. Langen, E. I. Kvasjuk, I. A. Mikhailopulo, DD 158903, Publ. 1983 – 02 – 09; G. W.
Saizewa, G. Kowollik, P. Langen, I. A. Mikhailopulo, E. I.Kvasjuk, DD 209197 (Akad. Wissen-
schaften DDR [DD], Publ. 1984 – 04 – 25); E. I. Kvasyuk, G. V. Zaitsewa, L. P. Savochkina, Z. G.
Chidgeavadze, R. S. Beabealashvilli, P. Langen, I. A. Mikhailopulo, Bioorg. Chem. (Moscow) 1989,
15, 781 (Chem. Abstr. 1990, 112, 3274); G. V. Zaitsewa, E. I. Kvasyuk, N. E. Poopeiko, T. I. Kulak,
V. E. Pashinnik, V. I. Tovstenko, L. N. Markovskii, I. A. Mikhailopulo, Bioorg. Chem. (Moscow)
1988, 14, 1275 (Chem. Abstr. 1989, 111, 7722).
[29] I. A. Mikhailopulo, G. G. Sivets, T. I. Pricota, N. E. Poopeiko, J. Balzarini, E. De Clercq, Nucleosides
Nucleotides 1991, 10, 1743.

1836
Helvetica Chimica Acta – Vol. 90 (2007)
[30] I. A. Mikhailopulo, N. E. Poopeiko, T. I. Pricota, G. G. Sivets, E. I. Kvasyuk, J. Balzarini, E. De
Clercq, J. Med. Chem. 1991, 34, 3195.
[31] I. A. Mikhailopulo, T. I. Pricota, N. E. Poopeiko, T. V. Klenitskaya, N. B. Khripach, Synthesis 1993,
700.
[32] J. E. Van der Boogaart, E. N. Kalinichenko, T. L. Podkopaeva, I. A. Mikhailopulo, C. Altona, Eur. J.
Biochem. 1994, 221, 759; E. N. Kalinichenko, T. L. Podkopaeva, N. E. Poopeiko, M. Kelve, M.
Saarma, I. A. Mikhailopulo, J. E. Van der Boogaart, C. Altona, Recl. Trav. Chim. Pays-Bas 1995, 114,
43.
[33] J. van Wijk, C. A. G. Haasnot, F. A. A. M. de Leeuw, B. D. Huckriede, A. J. A. Westra Hoekzema, C.
Altona, ꢁPSEUROT 6.3ꢂ, Leiden Institute of Chemistry, Leiden, The Netherlands, 1999.
[34] C. Thibaudeau, J. Plavec, J. Chattopadhyaya, J. Org. Chem. 1998, 63, 4967.
[35] I. A. Mikhailopulo, T. I. Pricota, G. G. Sivets, C. Altona, J. Org. Chem. 2003, 68, 5897.
[36] T. K. Brunck, F. J. Weinhold, J. Am. Chem. Soc. 1979, 101, 1700.
[37] M. Egli, S. M. Gryaznov, Cell. Mol. Life Sci. 2000, 57, 1440.
[38] J. S. Ingwall, J. Am. Chem. Soc. 1972, 94, 5487.
[39] D. B. Davies, Prog. Nucl. Magn. Reson. Spectrosc. 1978, 12, 135.
[40] A. A. Akhrem, I. A. Mikhailopulo, A. F. Abramov, Org. Magn. Reson. 1979, 77, 7.
[41] P. Herdewijn, R. Pauwels, M. Baba, J. Balzarini, E. De Clercq, J. Med. Chem. 1987, 30, 2131.
[42] C. Thibaudeau, J. Plavec, J. Chattopadhyaya, J. Org. Chem. 1996, 61, 266.
[43] C. Thiebaudeau, J. Chattopadhyaya, ꢁStereoelectronic Effects in Nucleosides and Nucleotides and
Their Structural Implicationsꢂ, Uppsala University Press, Uppsala, Sweden, 1999.
[44] S. Wolfe. Acc. Chem. Res. 1972, 5, 102.
[45] D. Oꢂ Hagan, H. S. Rzepa, M. Schuler, A. M. Z. Slawin, Beilstein J. Org. Chem. 2006, 2, 19.
[46] H. Rosemeyer, F. Seela, J. Chem. Soc., Perkin Trans. 2 1997, 2341.
[47] F. Seela, G. Becher, H. Rosemeyer, H. Reuter, G. Kastner, I. A. Mikhailopulo, Helv. Chim. Acta
1999, 82, 105.
[48] I. A. Mikhailopulo, Y. A. Sokolov, J. He, P. Chittepu, H. Rosemeyer, F. Seela, Nucleosides
Nucleotides Nucl. Acids 2005, 24, 70 1.
[49] F. Seela, X. Peng, H. Li, P. Chittepu, K. I. Shaikh, J. He, Y. He, I. A. Mikhailopulo, Collect. Czech.
Chem. Commun. Symp. Ser. 2005, 7, 1.
[50] R. Mengel, W. Guschlbauer, Angew. Chem., Int. Ed. 1978, 29, 525.
[51] W. E. Cohn, D. G. Doherty, J. Am. Chem. Soc. 1956, 78, 2863.
Received April 12, 2007