Synthesis and anticancer evaluation of 4'-C-methyl-2'-fluoro arabino nucleosides

Article abstract of DOI:10.1080/15257770903091946

As part of an ongoing program to develop novel antitumor agents over the years, we have synthesized and evaluated a number of 4'-C-substituted nucleosides. A few years ago, we reported the first synthesis of 4'-C-hydroxymethyl-2'-fluoro arabino nucleoside

Full text of DOI:10.1080/15257770903091946

This article was downloaded by: [Florida State University]
On: 06 May 2013, At: 01:14
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Nucleosides, Nucleotides and Nucleic
Acids
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/lncn20
Synthesis and Anticancer Evaluation
of 4′-C-Methyl-2′-Fluoro Arabino
Nucleosides
Kamal N. Tiwari a , Anita T. Shortnacy-Fowler a , William B. Parker a
, William R. Waud a & John A. Secrist III a
a Southern Research Institute, Drug Discovery Division, Birmingham,
Alabama, USA
Published online: 29 Jul 2009.
To cite this article: Kamal N. Tiwari , Anita T. Shortnacy-Fowler , William B. Parker , William R. Waud
& John A. Secrist III (2009): Synthesis and Anticancer Evaluation of 4′-C-Methyl-2′-Fluoro Arabino
Nucleosides, Nucleosides, Nucleotides and Nucleic Acids, 28:5-7, 657-677
To link to this article: http://dx.doi.org/10.1080/15257770903091946
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation
that the contents will be complete or accurate or up to date. The accuracy of any
instructions, formulae, and drug doses should be independently verified with primary
sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand, or costs or damages whatsoever or howsoever caused arising directly or
indirectly in connection with or arising out of the use of this material.

Nucleosides, Nucleotides and Nucleic Acids, 28:657–677, 2009
Copyright ꢀC Taylor & Francis Group, LLC
ISSN: 1525-7770 print / 1532-2335 online
DOI: 10.1080/15257770903091946
SYNTHESIS AND ANTICANCER EVALUATION OF
4ꢀ-C-METHYL-2ꢀ-FLUORO ARABINO NUCLEOSIDES
Kamal N. Tiwari, Anita T. Shortnacy-Fowler, William B. Parker,
William R. Waud, and John A. Secrist III
Southern Research Institute, Drug Discovery Division, Birmingham, Alabama, USA
As part of an ongoing program to develop novel antitumor agents over the years, we have
2
synthesized and evaluated a number of 4ꢁ-C-substituted nucleosides. A few years ago, we reported
the first synthesis of 4ꢁ-C-hydroxymethyl-2ꢁ-fluoro arabino nucleosides, which did not exhibit any
cytotoxicity. In our exploration of related compounds, we synthesized and evaluated the 4ꢁ-C-methyl-
2ꢁ-fluoro arabino nucleosides in both the purine and pyrimidine series. In the pyrimidine series,
1-(4-C-methyl-2-fluoro-β-D-arabinofuranosyl) cytosine (13) was found to be highly cytotoxic and
had significant antitumor activity in mice implanted with human tumor xenografts. The synthesis
and anticancer activity of this series of nucleosides are reported.
Keywords 4ꢁ-c-methylnucleosides; anticancer activity; 2ꢁ-fluoroarabino purine and
pyrimidine; nucleosides
INTRODUCTION
In our program to develop novel nucleosides as potential antitumor
agents, we have been successful in obtaining useful clinical agents from ara-
binofuranosyl nucleosides.[1–4] This success has been demonstrated in our
program by the approval for human use of clofarabine and the 5ꢁ-phosphate
of fludarabine, both arabino nucleosides. Recently, several groups have
reported the synthesis and biological activities of 4ꢁ-C-methylnucleosides
in the 2ꢁ-deoxy, ribo and arabino series.[5–10] The reported in vivo
Received 19 March 2009; accepted 29 May 2009.
In honor and celebration of the 70th birthday of Professor Morris J. Robins.
This investigation was supported by NIH Grant No. CA34200. We sincerely acknowledge the data
received from the following laboratories at Southern Research Institute: Jim Riordan and Jackie Truss
(NMR), Mark Richardson (Mass), Sheila Campbell (HPLC), Joan Bearden (CHN analysis), and Jerry
Frye and Ron Carter (large scale synthesis). The authors appreciate the help of Stephanie Brannon with
preparing the manuscript.
Address correspondence to John A. Secrist III, Southern Research Institute, P.O. Box 55305,
Birmingham, AL 35255-5305. E-mail: Secrist@sri.org
657

658
K. N. Tiwari et al.
anticancer activity of 4ꢁ-C-methyl-araC suggests that incorporation of a
4ꢁ-C-methyl group probably allows the retention of substrate activity for
deoxycytidine kinase, the enzyme by which the arabinocytosine nucleoside
is converted to its monophosphate analog.[8] Recently, we incorporated a
4ꢁ-C-hydroxymethyl into both the 2ꢁ-hydroxy and 2ꢁ-fluoro arabino series and
have reported these results.[11, 12] The marked activities from arabinonu-
cleosides having either the 2ꢁ-hydroxyl or 2ꢁ-fluoro group in the arabino
configuration as in the case of fludarabine and clofarabine, respectively,
prompted us to explore the corresponding 4ꢁ-C-methyl modification in
2ꢁ-fluoroarabinonucleosides. Very recently, there have been two reports of
compounds from this series of nucleosides that include selected antiviral
evaluations.[13, 14] In the present report, we describe the details of our
synthetic methods and the anticancer activity evaluations of several purine
and pyrimidine nucleosides in this series.
DISCUSSION
Chemistry
Our routes to the target compounds are presented in Scheme 1.
We have already reported the synthesis of 4ꢁ-C-hydroxymethyl-2ꢁ-fluoro-
arabinofuranoside 1 and the corresponding nucleosides.[12] Selective pro-
tection of 1 with the monomethoxytrityl (MMT) group was carried
out using MMT chloride in pyridine in 30% yield.[15] The undesired
isomer 2b and the unreacted 1 were recycled to increase the yield.
The selectively blocked intermediate 2a was benzoylated to give 3 in
92% yield, which was then detritylated to afford the sugar intermedi-
ate 4 in 89% yield. This 4ꢁ-C-hydroxymethyl analog 4 was converted
into 4ꢁ-C-phenoxythiocarbonyloxymethyl derivative 5 in 90% yield using
phenyl chlorothionoformate. Compound 5 was deoxygenated using 1,1ꢁ-
azobis(cyclohexane-carbonitrile) (ACCN) and tris(trimethyl)silane to pro-
vide 4ꢁ-C-methyl analog 6 in 84% yield.[16] Acetolysis of compound 6 using
traditional methods[17] failed to give 1-O-acetyl sugar 8, resulting in either

4ꢁ-C-Methyl-2ꢁ-fluoro Arabino Nucleosides
659
SCHEME 1 Conditions: (a) MMTr-Cl, pyridine, room temperature, overnight; (b) BzCl, pyridine, room
temperature, overnight; (c) 80% AcOH, room temperature, overnight; (d) PhOC( =S)Cl, DMAP, MeCN,
room temperature, 3 hours; (e) (TMS)3SiH, ACCN, toluene, 100◦C, 5 hours; (f) TFA/H2O, 65◦C,
24 hours; (g) Ac2O/pyridine, room temperature, overnight; (h) HBr/AcOH, 5◦C, overnight; (i) Bases,
BSA, MeCN, room temperature, 1–2 hours; (j) persilylated bases, compound 9, ClCH2CH2Cl, 100◦C,
4 hours; (k) 0.5 N NaOCH3, MeOH, room temperature, 2–7 hours.; then for 13 only, HCl; (l) NaH,
MeCN, room temperature, 6 hours; (m) EtOH, NH3, 80◦C, 16 hours; (n) NaN3, EtOH, reflux, 1/ hour;
2
(o) 10% Pd/C, H2, 1 atm, EtOH/DMAC, 18 hours; (p) NaOCH3/MeOH, room temperature, 3 hours;
(q) Adenosine deaminase.

660
K. N. Tiwari et al.
no reaction or gradual decomposition. The methyl glycoside 6 was instead
hydrolyzed using 9:1 trifluoroacetic acid/water to provide the hydroxy sugar
7 in 83% yield, which was acetylated to produce compound 8 in 91%
yield. This sugar intermediate was converted cleanly into glycosyl bromide
9 using 33% HBr in acetic acid. Attempted conversion of 7 directly to 9
resulted in a complex mixture and a very low yield of 9. The bromosugar 9
was highly reactive and was used directly without purification for coupling
reactions.
Coupling of bromosugar 9 with silylated N4-benzoylcytosine in situ with
BSA gave cytosine nucleosides 10/10α in 54% yield.[18, 19] Separation of α,
β anomers afforded pure β anomer 10 as the major product (48%) and α
anomer 10α as the minor product (6%). Similarly, uracil and thymine were
coupled with bromo sugar 9 to obtain corresponding nucleosides 11/11α
and 12/12α in 71% and 68% yields, respectively, with the β anomer as
the predominant product. Both anomers of cytosine nucleoside 10 were
deblocked using sodium methoxide to get the target compound 13 as well
as 13α. After purification the β anomer 13 was isolated as a hydrochloride
salt in 89% yield, and the α anomer 13α was isolated as the free base
in 77% yield. In the case of nucleosides 11 and 12, only the β anomers
were deblocked using the same procedure to obtain compounds 14 (77%)
and 15 (92%), respectively. The α anomers of compounds 11 and 12 were
not further utilized and were isolated only for characterization purposes to
compare with the β anomers.
A series of purine nucleoside analogs were prepared through the
coupling of bromosugar 9 with 6-chloropurine and 2, 6-dichloropurine.
Sodium salt coupling of 9 with 6-chloropurine gave the desired β nucle-
oside 16 (36%) and α anomer 16α (14%).[20] Separate treatment with
ethanolic ammonia gave the target compound 17 (74%) and α anomer
17α (49%), respectively. Similarly 2, 6-dichloropurine was coupled with
bromo sugar 9 to obtain the corresponding nucleoside as an anomeric
mixture (2:1, β:α ratio) in 64% yield. Both anomers were separated by
preparative TLC to provide 18 and 18α as white foams. Separate treat-
ment of 18 and 18α with sodium azide in aqueous ethanol at reflux
produced the corresponding 2,6-diazido intermediates 19 and 19α, which
were subjected to reduction with Pd/C to afford blocked diaminopurine
nucleosides 20 and 20α, respectively. Deblocking of 20 and 20α with
NaOMe produced the target 2-aminoadenine nucleosides 21 and 21α. Con-
version of 21 to the guanine nucleoside 22 was accomplished by treatment
with adenosine deaminase.[21] Though the deamination was slow, it went
to completion at room temperature in 68 hours. The 2-chloroadenine
nucleosides 24 (84%) and 24α (75%) were prepared by first converting
the dichloropurine nucleosides 18 and 18α to their 6-methoxy interme-
diate 23 with sodium methoxide followed by treatment with ethanolic
ammonia.[21]

4ꢁ-C-Methyl-2ꢁ-fluoro Arabino Nucleosides
661
BIOLOGICAL RESULTS
In Vitro Cytotoxicity
The concentration of compound required to inhibit cell growth by 50%
(IC50) after 72 hours of incubation was determined for each unprotected
analog with eight human tumor cell lines (SNB-7 CNS, DLD-1 colon, CCRF-
CEM leukemia, NCI-H23 NSCL, ZR-75–1 breast, LOX melanoma, PC-3
prostate, and CAKI-1 renal). The most active compound in this series was
methyl-F-araC (13, Table 1), which was found to have significant cytotoxicity
against four of the cell lines in the panel. The purine analogs demonstrated
modest cytotoxicity against the solid tumor cell lines (IC50’s between 5 and
80 µM), while the uracil and thymine analogs were not active against any
cell line (IC50’s greater than 200 µM). The α-anomers 13α, 17α, 21α,
and 24α were also screened but were not found to be cytotoxic (data not
shown).
CCRF-CEM cells are a T-cell leukemia cell line that is known to be very
sensitive to nucleoside analogs. Methyl-F-araC was a very potent inhibitor
of this cell line with an IC50 of 0.012 0.003 µM. CCRF-CEM cell growth
was also inhibited by the 2-Cl-adenine (24), 2,6-diaminopurine (21), and
guanine (22) analogs with IC50’s of approximately 0.5 µM. The inhibition of
CCRF-CEM cell growth caused by either methyl-F-araC or the 2-Cl-adenine
analog (4ꢁ-C-methyl-clofarabine, 24) was prevented by adding dCyd to the
culture medium and neither compound was active in cells that lacked dCyd
kinase. These results indicated that dCyd kinase was the primary enzyme
responsible for the initial activation step of these two agents in CCRF-CEM
cells. The cytotoxicity of the diaminopurine analog 21 was prevented by the
addition of deoxycoformycin, a potent inhibitor of adenosine deaminase,
TABLE 1 Effect of 4ꢁ-methyl-2ꢁ-F-2ꢁ-deoxy analogs on cell growth
Cell line
13
14
15
17
21
22
24
IC50 (µM)
>200
>200
3.9
SNB-7
DLD-1
CCRF-CEM
NCI-H23
ZR-75–1
LOX
>200
>200
0.012
0.19
>200
0.24
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
0.18
39
>200
37
>200
>200
0.45
52
>200
43
>200
>200
0.64
79
>200
32
17
>200
>200
>200
5.8
PC-3
CAKI-1
>200
0.54
29
16
22
51
140
17
The cell lines were incubated with various concentrations of the above compounds for 72 hours, and
the concentration of drug that resulted in 50% inhibition of cell growth (IC50) was determined. The
numbers shown are the result of one experiment except for the cytosine analog, which is the average of
three separate determinations. The α-anomers 13α, 17α, 21α, and 24α were evaluated in all cell lines
and were not found to be cytotoxic at the highest concentrations utilized (>200 µM).

662
K. N. Tiwari et al.
which indicated that 21 was deaminated to the dGuo analog before conver-
sion to cytotoxic nucleotides.
In Vitro Metabolic Studies in CCRF-CEM Cells
CCRF-CEM cells were incubated with methyl-F-araC (13), araC, and
gemcitabine, and the amount of intracellular triphosphate (TP) of each
compound was determined. There was significant metabolism of each of
these compounds, and their triphosphates did not co-elute with any of
the natural nucleotides (ATP, GTP, CTP, or UTP). Incubation of CCRF-
CEM cells with 100 nM of each compound for two hours resulted in
an intracellular concentration of methyl-F-araC-TP (16
cells) that was similar to those of araC-TP (12
2 pmoles/106
1 pmoles/106 cells) and
gemcitabine-TP (17 3 pmoles/106 cells; mean SD, N = 3). These results
indicated that methyl-F-araC was a good substrate for deoxycytidine kinase.
The intracellular half-life of each triphosphate was similar: methyl-F-araC-TP
(7.1 hours, N = 2); araC-TP (5.6 hours, N = 2); gemcitabine-TP (5.0 hours,
N = 2).
In Vivo Activity
Because of its potent in vitro activity, methyl-F-araC (13) was evaluated
for in vivo activity against three solid tumor xenografts (CAKI-1 renal,
NCI-H23 NSCL, and LOX melanoma). Prior to these studies the maximally
tolerated dose of methyl-F-araC was determined to be 3 mg/kg given
once per day for 9 consecutive days. Methyl-F-araC demonstrated excellent
activity against the CAKI-1 tumors (Figure 1). In each treatment group there
were 3 of 6 tumor-free survivors 62 days post implant. Good results were
also seen against NCI-H23 and LOX human tumor xenografts (Table 2).
Therefore, methyl-F-araC demonstrated good to excellent in vivo antitumor
activity in the three solid tumor xenografts that have been tested to date.
SUMMARY
We have synthesized a series of 4ꢁ-C-methyl-2ꢁ-fluoro arabino nucleo-
sides in both the purine and pyrimidine series. 1-(4-C-Methyl-2-fluoro-β-
D-arabinofuranosyl) cytosine (methyl-F-araC, 13) was found to be highly
cytotoxic and had significant antitumor activity in mice implanted with
human tumor xenografts. This compound is a substrate for deoxycytidine
kinase and shows significant levels of its 5ꢁ-triphosphate accumulated in
CCRF-CEM cells. The excellent in vivo anticancer activity of 13 warrants
continued investigation and suggests that evaluation of nucleosides with

4ꢁ-C-Methyl-2ꢁ-fluoro Arabino Nucleosides
663
TABLE 2 Response of s.c. implanted human tumor xenografts to methyl-F-araC (13)
Optimal i.p. dosage
(mg/kg/dose)
Tumor size range
(mm3)
Tumor-free
survivors
Tumor
T-C (days)
CAKI-1 renal
NCI-H23 NSCL
LOX melanoma
3
4
3
100–245
100–270
100–221
>34.6a
18.4b
3/6
0/5
0/6
15.0c
Xenografts were implanted sc on the flanks of female nude mice. When tumors were approximately
100–250 mg, they were treated ip with 3 or 4 mg/kg/dose of methyl-F-araC (q1d × 9) and tumor size
was measured twice weekly thereafter. Tumor-free survivors are the number of mice that were tumor-free
at the end of the experiment/total number of mice in the treatment group.
aThe difference in the median of times poststaging for tumors of the treated (T) and control (C)
groups to double in mass three times.
bThe difference in the median of times poststaging for tumors of the treated (T) and control (C)
groups to double in mass two times.
cThe difference in the median of times poststaging for tumors of the treated (T) and control (C)
groups to double in mass four times.
FIGURE 1 Effect of methyl-F-araC on CAKI-1 tumor growth. Female NCr-nu athymic mice were
implanted subcutaneously with CAKI-1 tumor fragments. When tumors were approximately 100–
250 mg, mice were treated ip with methyl-F-araC at 1, 2, or 3 mg/kg/dose (1 treatment each day
for 9 consecutive days starting on day 14). Each treatment group contained 6 mice. The tumors were
measured with calipers twice each week, and the weight (mg) was calculated. In this experiment there
were 3 of 6 tumor-free survivors at the end of the experiment (62 days post implant) in each treatment
group.

664
K. N. Tiwari et al.
modifications at the 4ꢁ position could lead to the identification of a new
analog with superior antitumor activity.
EXPERIMENTAL
TLC analysis was performed on Analtech (Newark, DE, USA) precoated
(250 µm) silica gel GF plates. Melting points were determined on a
Mel-Temp apparatus (Laboratory Devices, Cambridge, MR, USA) and are
uncorrected. Purifications by flash chromatography were carried out on
Merck silica gel (230–400 mesh). Evaporations were performed with a rotary
evaporator, higher boiling solvents (DMF, pyridine) were removed in vacuo
(<1 mm, bath to 35◦C). Products were dried in vacuo (<1 mm) at 22–25◦C
over P2O5. The mass spectral data were obtained with a Varian-MAT 311A
mass spectrometer (Palo Alto, CA, USA) in the fast atom bombardment
(FAB) mode or with a Bruker BIOTOF II (Billerica, MA, USA) by electro-
spray ionization (ESI). 1H NMR spectra were recorded on a Nicolet NT-300
NB spectrometer (Madison, WI, USA) operating at 300.635 MHz. Chemical
shifts in CDCl3 and Me2SO-d6 are expressed in parts per million downfield
from tetramethylsilane (TMS), and in D2O chemical shifts are expressed
in parts per million downfield from sodium 3-(trimethylsilyl)propionate-
2,2,3.3-d4 (TMSP). Chemical shifts (δ) listed for multiplets were measured
from the approximate centers, and relative integrals of peak areas agreed
with those expected for the assigned structures. UV absorption spectra were
determined on a Perkin-Elmer lambda 9 spectrophotometer (Norwalk, CT,
UK) by dissolving each compound in MeOH or EtOH and diluting 10-fold
with 0.1 N HCl, pH 7 buffer, or 0.1 N NaOH. Numbers in parentheses
are extinction coefficients (ε × 10−3). Microanalyses were performed
by Atlantic Microlab, Inc. (Atlanta, GA, USA) or the Spectroscopic and
Analytical Department of Southern Research Institute (Birmingham, AL,
USA). Analytical results indicated by element symbols were within 0.4% of
the theoretical values, and where solvents are indicated in the formula, their
presence was confirmed by 1H NMR.
Cell Culture Cytotoxicity
All cell lines were grown in RPMI 1640 medium containing 10% fetal
bovine serum, sodium bicarbonate, and 2 mmol glutamine. For in vitro
evaluation of the sensitivity of these cell lines to compounds, cells were
plated in 96-well microtiter plates and then were exposed continuously
to various concentrations of the compounds for 72 hours at 37◦C. Cell
viability was measured using the MTS assay [3-(4,5-dimethylthiazol-2-yl)-5-(3-
carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt; MTS
and an electron coupling reagent (phenazine ethosulfate; PES)]. Ab-
sorbance was read at 490 nm. The background absorbance mean was

4ꢁ-C-Methyl-2ꢁ-fluoro Arabino Nucleosides
665
subtracted from the data followed by conversion to percent of control. The
drug concentrations producing survival just above and below 50% level were
used in a linear regression analysis to calculate the IC50.
Measurements of Intracellular Triphosphates
CCRF-CEM cell extracts were collected by centrifugation and resus-
pended in ice-cold 0.5 M perchloric acid. The samples were centrifuged
at 12,000 × g, and the supernatant fluid was neutralized and buffered
by adding 4 M KOH and 1 M potassium phosphate, pH 7.4. KClO4 was
removed by centrifugation, and a portion of the supernatant fluid was
injected onto a strong anion exchange HPLC (Bio Basic anion exchange
column, Thermo Electron Corp., Bellefonte, PA, USA). Nucleotides were
eluted with a 30-minutes linear salt and pH gradient from 6 mM ammonium
phosphate (pH 2.8) to 900 mM ammonium phosphate (pH 6). Peaks were
detected as they eluted from the column by their absorbance at 254 nm.
Experimental Chemotherapy
Mice, obtained from various commercial suppliers, were housed in
microisolator cages and were allowed commercial mouse food and water ad
libitum. The three human tumors were obtained from the Developmental
Therapeutics Program Tumor Repository (Frederick, MD, USA) and were
maintained in in vivo passage. Only tumor lines that tested negative for
selected viruses were used. For the in vivo evaluation of the sensitivity
of human tumors to the compounds, female NCr-nu athymic mice were
implanted subcutaneously (sc) with 30–40 mg tumor fragments. In each ex-
periment, methyl-F-araC (13) was tested at three dosage levels. Procedures
were approved by the Southern Research Institutional Animal Care and Use
Committee, which conforms to the current Public Health Service Policy on
Humane Care and Use of Laboratory Animals and the Guide for the Care and Use
of Laboratory Animals.
Antitumor activity was assessed on the basis of delay in tumor growth
(T-C). The delay in tumor growth is the difference in the median of times
poststaging for tumors of the treated and control groups to double in mass
two, three, or four times. Drug deaths and any other animal whose tumor
failed to attain the evaluation size were excluded. Tumors were measured
in two dimensions (length and width) twice weekly, and the tumor weight
was calculated using the formula (length × width2)/2 and assuming unit
density. The mice were also weighed twice weekly.
Methyl
4-C-(p-Anisyldiphenylmethoxymethyl)-2-deoxy-2-fluoro-β-D-
arabinofuranoside (2a) and Methyl 5-O-(p-Anisyldiphenylmethyl)-4-C-
hydroxymethyl-2-deoxy-2-fluoro-β-D-arabinofuranoside (2b). To a solution
of 1 (342 mg, 1.75 mmol) in dry pyridine (15 mL) was added in one

666
K. N. Tiwari et al.
portion solid 97% p-anisylchlorodiphenylmethane (816 mg, 2.56 mmol).
The reaction mixture was stirred at room temperature for 20 hours then
evaporated. The resulting residue was coevaporated with two portions
of toluene before being purified by flash chromatography on silica
gel (45 g) using a gradient from CHCl3 to 97: 3 CHCl3/MeOH. The
first eluted fraction gave 2a ((246 mg, 30%) as a white foam: TLC 97:
3 CHCl3/MeOH, Rf 0.45; MS m/z 491 (m+Na)+; 1H NMR (CDCl3)
7.20–7.45 (m, 12H, aromatic H’s), 6.86–6.88 (m, 2H, para- H’s), 5.10–5.33
(m, 2H, H-1 and H-2, J2,F = 54 Hz), 4.44–4.52 (m, 1H, H-3), 3.81 (s, 3H,
OCH3 of 4-methoxyphenyl), 3.56 (s, 3H, 1-OCH3), 3.46–3.58 (m, 3H, two
4-C-hydroxymethyl and one 5-CH2 hydrogens), 3.04 (d, 1H, 5-CH2, J5a,5b
= 10.3 Hz), 2.95(d, 1H, 3-OH, J3,3-OH = 10 Hz), 2.18(t, 1H, 5-OH, J5a,5-OH
= 6.3 Hz). The second fraction provided 2b (137 mg, 17%): TLC 97: 3
CHCl3/MeOH, Rf 0.39; MS m/z 491 (m+Na)+; 1H NMR (CDCl3) 7.20–7.48
(m, 12H, aromatic H’s), 6.82–6.86 (m, 2H, para- H’s), 4.83–5.06 (m, 1H,
H-2, J2,F = 55 Hz), 4.92 (dd, 1H, H-1, J1,2 = 2 Hz and J1,F = 6 Hz), 4.3–4.40
(m, 1H, H-3), 3.94–4.02 (m, 1H, 5-CH2), 3.82–4.02 (m, 1H, 5-CH2), 3.80 (s,
3H, OCH3 of 4-methoxyphenyl), 3.25 (s, 3H, 1-OCH3), 3.26–3.28 (m, 1H,
4-C-hydroxymethyl), 3.18–3.20 (m, 1H, 4-C-hydroxymethyl), 2.80 (d, 1H,
3-OH, J = 12 Hz), 2.12 (t, 1H, 5-OH, J = 8 Hz).
Methyl 4-C-(p-Anisyldiphenylmethoxymethyl)-3, 5-di-O-benzoyl-2-deoxy-
2-fluoro-β-D-arabinofuranoside (3). To a solution of 2a (52 mg, 0.11 mmol)
in dry pyridine (5 mL) at 0◦C was added benzoyl chloride (91 µl,
0.77 mmol) dropwise. After 5 minutes, the cooling bath was removed, and
stirring was continued for 18 hours. The solution was evaporated to a solid
that was coevaporated once with toluene. The solid was purified by silica gel
preparative TLC (Analtech GF, 10 × 20 cm, 1000 µ) with 3:1 hexane/EtOAc
as solvent to give 3 (69 mg, 92%) as a white foam: TLC 3:1 hexane/EtOAc,
Rf 0.44; MS m/z 699 (m+Na)+; 1H NMR (CDCl3) 7.82–7.92 (m, 4H, ortho
H’s of benzoyl), 7.50–7.62 (m, 2H, para- H’s of benzoyl), 7.12–7.92 (m, 16H,
aromatic H’s), 6.63–6.68 (m, 2H, p- H’s), 6.02 (dd,1H, H-3, J2,3 = 8 Hz
and J3,F = 18 Hz), 5.46–5.70 (m,1H, H-2), 5.18 (bd, 1H, H-1, J1,2 = 6 Hz),
4.50–4.60 (m, 2H, 5CH2), 3.70 (s, OCH3 of 4-methoxyphenyl), 3.48 (s, 3H,
1-OCH3), 3.36–3.42 (m, 1H, 4-CH2), 3.14–3.18 (m,1H, 4-CH2).
Methyl
3,
5-Di-O-benzoyl-2-deoxy-2-fluoro-4-C-hydroxymethyl-β-D-
arabinofuranoside (4). A solution of 3 (576 mg, 0.85 mmol) in 4:1 acetic
acid/water (20 mL) was stirred at room temperature for 19 hours and
then evaporated. The residue obtained was partitioned between EtOAc
and ice-cold saturated NaHCO3. The aqueous layer was extracted twice
with EtOAc, and the combined organic layers were washed with saturated
NaCl, dried (MgSO4), and evaporated. The residue was purified by flash
chromatography on silica gel (20 g) using a gradient from hexane to
2:1 hexane/EtOAc to give 4 (306 mg, 89%) as a clear syrup: TLC 2:1
hexane/EtOAc, Rf 0.29; MS m/z 405 (m+H)+; 1H NMR (CDCl3) 8.0–8.10

4ꢁ-C-Methyl-2ꢁ-fluoro Arabino Nucleosides
667
(m, 4H, ortho H’s of benzoyl), 7.34–7.64 (m, 6H, para and meta H’s of
benzoyl), 6.08 (dd, 1H, H-3, J2,3 = 8 Hz and J3,F = 18 Hz), 5.27–5.50 (m,
1H, H-2), 5.12 (dd, 1H, H-1, J1,2 = 2 Hz and J1,F = 8 Hz), 4.52–4.65 (m, 2H,
5-CH2), 3.76 (s, 2H, 4-CH2), 3.50 (s, 3H, OCH3), 2.12 (broad hump, 1H,
5-OH).
Methyl 3, 5-Di-O-benzoyl-2-deoxy-2-fluoro-4-C-phenoxythiocarbonyloxy-
methyl-β-D-arabinofuranoside (5). To a solution of 4 (75 mg, 0.19 mmol)
and 4-(dimethylamino) pyridine (93 mg, 0.75 mmol) in dry MeCN (7
mL) at room temperature was added dropwise phenyl chlorothionoformate
(39 µl, 0.28 mmol). The resulting yellow solution was stirred at room
temperature for 3 hours and then evaporated. The residue obtained was
partitioned between ice-cold 5% citric acid and EtOAc. The aqueous
layer was extracted twice with EtOAc, and the combined organic layers
were washed with water, dried (MgSO4), and evaporated to a gum. This
crude 5 was purified by silica gel preparative TLC (Analtech GF, 10 ×
20 cm, 1000 µ) with 3:1 hexane/EtOAc as solvent to obtain pure 5 (92
mg, 90%) as a white foam: TLC 3:1 hexane/EtOAc, Rf 0.55; MS m/z 563
1
(m+Na)+; H NMR (CDCl3) 8.00–8.10 (m, 4H, ortho H’s of benzoyl),
7.26–7.68 (m, 9H, aromatic H’s), 6.90–6.94 (m, 2H, ortho H’s of phenyl),
6.10 (dd, 1H, H-1, J1,2 = 8 Hz and J1,F = 18 Hz), 5.16–5.42 (m, 1H, H-2),
5.10 (dd, 1H, H-3, J2,3 = 1.5 Hz and J3,F = 7 Hz), 4.62–4.74 (m, 4H, 4 and
5-CH2), 3.52 (s, 3H, OCH3).
Methyl 3,5-Di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-β-D-arabinofurano-
side (6). A solution of 5 (4.0 g, 7.4 mmol) in anhydrous toluene
(125 mL) was purged with argon before solid 98% 1,1ꢁ-azobis
(cyclohexanecarbonitrile) (657 mg, 2.6 mmol) was added in one portion.
The argon purge was repeated followed by a syringe addition of 97% tris
(trimethylsilyl) silane (10 mL, 31 mmol) over 5 minutes. The reaction
solution was warmed over 0.5 hours to 100◦C, maintained at 100◦C for 5
hours, cooled to room temperature and reduced under vacuum to an oil.
The crude product was purified by column chromatography on silica gel
with 5:1 cyclohexane/EtOAc as solvent to provide 6 (2.4 g, 84%) as a clear
1
oil: TLC 85:15 cyclohexane/EtOAc, Rf 0.38; MS m/z 389 (m+H)+; H
NMR (CDCl3) 8.08–8.12 (m, 4H, ortho H’s of benzoyl), 7.38–7.66 (m, 6H,
para and meta H’s of benzoyl), 6.28 (dd, 1H, H-3, J2,3 = 8 Hz and J3,F
=
18 Hz), 5.15–5.37 (m, 1H, H-2), 5.04 (dd, 1H, H-1, J1,2 = 2 Hz and J1,F = 8
Hz), 4.46–4.62 (m, 2H, 5-CH2), 3.44 (s, 3H, OCH3), 1.34 (s, 3H, CH3).
3,5-Di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-α,β-D-arabinofuranoside
(7). A solution of 6 (1.3 g, 3.35 mmol) in 9:1 trifluoroacetic acid/water
(23 mL) was maintained at 60–65◦C for 24 hours, cooled to room
temperature, and diluted with CH2Cl2 (100 mL). The solution was added
dropwise to a stirred mixture of ice (300 g) and saturated NaHCO3 (300
mL). Solid NaHCO3 was added during the addition to maintain a pH
of 7. The mixture was extracted with CH2Cl2 (3 × 100 mL), and the

668
K. N. Tiwari et al.
organic extract was washed with water (2 × 50 mL), dried (MgSO4), and
concentrated to a syrup (1.3 g). This material was flash chromatographed
on silica gel (100 g) with 3:1 hexane/EtOAc as solvent to yield pure 7
(1.05 g, 83%) as a white solid: TLC 3:1 hexane/EtOAc, Rf 0.40; MS m/z
1
375 (m+H)+; H NMR (CDCl3) 8.04–8.12 (m, 4H, ortho H’s of benzoyl),
7.40–7.64 (m, 6H, para and meta H’s of benzoyl), 5.70–5.88 (m, 1H, H-3
α,β), 5.65 (dd, 0.65H, H-1 α, J1,2 = 4 Hz and J1,F = 12 Hz), 5.52–5.57 (m,
0.35H, H-1 β), 5.07–5.28 (m, 1H, H-2 α,β), 4.34–4.78 (m, 2H, 5-CH2), 3.52
(d, 0.35H, 1-OH β), 2.79 (dd, 0.65H, 1-OH α, J = 1 and 4 Hz), 1.51 (s, 2H,
CH3 α), 1.36 (s, 1H, CH3 β).
1-O-Acetyl-3,5-di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-α,β-D-arabino-
furanose (8). To a solution of 7 (1.04 g, 2.78 mmol) in dry pyridine
(25 mL) at 5◦C was added dropwise acetic anhydride (0.79 mL, 8.37
mmol) over 5 minutes. After 15 minutes, the solution was allowed to
warm to room temperature where it was held for 18 hours. The reaction
solution was concentrated in vacuum and coevaporated with toluene (3 ×
2 mL). The crude product was purified by flash chromatography on silica
gel (70 g) using 3:1 hexane/EtOAc as solvent to provide pure 8 (1.06
g, 91%) as a white solid: TLC 3:1 hexane/EtOAc, Rf 0.50; MS m/z 439
1
(m+Na)+; H NMR (CDCl3) 8.06–8.12 (m, 4H, ortho H’s of benzoyl),
7.40–7.68 (m, 6H, para and meta H’s of benzoyl), 6.41–6.46 (m, 1H, H-1
α,β), 5.16 (m, 1H, H-3 α,β), 5.16–5.48 (m, 1H, H-2 α,β), 4.40–4.64 (m,
2H, 5-CH2), 2.16 (s, 2.25H, CH3 of 1-O-acetyl, α), 2.0 (s, 0.75H, CH3 of
1-O-acetyl, β), 1.50 (s, 2.25H, CH3 α), 1.40 (s, 0.75H,CH3 β).
3,5-Di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-α,β-D-arabinofuranosyl
Bromide (9). A solution of 8 (507 mg, 1.22 mmol) in CH2Cl2 (35 mL) was
stirred with MgSO4 for 1.5 hours, filtered, and evaporated to a stiff syrup.
After being dried under vacuum for 2 hours, the residue was dissolved
in anhydrous CH2Cl2 (20 mL), chilled to 5◦C, and treated dropwise with
33% HBr in acetic acid (5.5 mL). The clear yellow solution in a tightly
sealed flask was placed in a nitrogen filled bag, refrigerated for 20 hours,
and evaporated in vacuum. The dark orange highly reactive residue was
coevaporated with toluene (2 × 3 mL) and then used directly in the various
pyrimidine and purine couplings: TLC 3:1 hexane/EtOAc, Rf 0.85.
N4-Benzoyl-1-(3,5-di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-β-D-arabino-
furanosyl)cytosine (10). A suspension of 98% N4-benzoyl cytosine (863
mg, 3.93 mmol) in dry MeCN (20 mL) at room temperature was
treated dropwise with 95% N, O-bis (trimethylsilyl) acetamide (BSA)
(3.6 mL) and stirred for 2 hours. The clear solution obtained was
evaporated in vacuum to a fluid oil that was dried under vacuum for an
additional 2 hours before being dissolved in ClCH2CH2Cl (20 mL). To
this solution was added in one portion a solution of 9 [prepared from 8
(507 mg, 1.22 mmol)] in ClCH2CH2Cl (10 mL). The reaction solution
was heated at 100◦C for 4 hours, cooled, and quenched with MeOH

4ꢁ-C-Methyl-2ꢁ-fluoro Arabino Nucleosides
669
(15 mL) at 5◦C. The reaction was stirred at room temperature for 0.5 hours
before being filtered through a celite pad to remove excess pyrimidine.
The solid was washed with CHCl3 and MeCN until free of 10, and the
combined filtrate and washings were evaporated to a yellow solid. This
crude product was purified by flash chromatography on silica gel (45 g)
with 1:1 hexane/EtOAc as solvent to give pure 10 (336 mg, 48%) as a white
1
solid: TLC 1:1 hexane/EtOAc, Rf 0.28; MS m/z 572 (m+H)+; H NMR
(CDCl3) 8.70 (broad hump, 1H, NH), 7.86–8.14 (m, 7H, H-6 and ortho H’s
of benzoyl), 7.46–7.70 (m, 10H, H-5 and para and meta H’s of benzoyl),
6.47 (dd, 1H, H-1ꢁ, J1 ,2 = 4 Hz and J1 ,F = 20 Hz), 5.88 (dd, 1H, H-3ꢁ, J3 ,F
ꢁ
ꢁ
ꢁ
ꢁ
= 18 Hz), 5.52 (dd, 1H, H-2ꢁ, J2 ,3 = 4 Hz and J2 ,F = 50 Hz), 4.58–4.68 (m,
ꢁ
ꢁ
ꢁ
2H, 5ꢁ-CH2), 1.50 (s, 3H, 4ꢁ-CH3).
From impure fractions, the α-anomer 10α (41 mg, 6%) was recovered
as a white solid by silica gel preparative TLC (Analtech GF, 10 × 20 cm,
500 µ) with 99:1 CHCl3/MeOH as solvent: TLC 1:1 hexane/EtOAc, Rf 0.45;
1
MS m/z 572 m+H)+; H NMR (CDCl3) 8.80 (bh, 1H, NH), 7.40–8.16 (m,
17H, H-5, H-6 and aromatic H’s), 6.28 (dd, 1H, H-1ꢁ, J1 ,2 = 1 Hz and J1 ,F
ꢁ
ꢁ
ꢁ
= 18 Hz), 5.90 (dd, 1H, H-3ꢁ, J3 ,F = 14 Hz), 5.46 (dd, 1H, H-2ꢁ, J2 ,3 = 1 Hz
ꢁ
ꢁ
ꢁ
and J2 ,F = 48 Hz), 4.44–4.50 (m, 2H, 5ꢁ-CH2), 1.64 (s, 3H, 4ꢁ-CH3).
ꢁ
1-(3,5-Di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-β-D-arabinofuranosyl)
uracil (11). A suspension of uracil (55 mg, 0.49 mmol) in dry MeCN
(3 mL) at room temperature was treated dropwise with 95% N, O-bis
(trimethylsilyl) acetamide (BSA) (0.44 mL) and stirred for 1 hour. The
clear solution obtained was evaporated under vacuum to a syrup that was
dried for an additional 1 hour before being dissolved in ClCH2CH2Cl (3
mL). To this solution was added in one portion a solution of 9 [prepared
from 8 (57 mg, 0.14 mmol)] in ClCH2CH2Cl (2 mL), and the mixture was
heated at 100◦C for 4 hours, cooled, and quenched with MeOH (1 mL) at
5◦C. After being stirred at room temperature for 1.5 hours, the mixture
was filtered through a Celite pad to remove excess uracil, and the filtrate
was concentrated to a yellow solid. This anomeric mixture was resolved
by preparative TLC (Analtech GF, 10 × 20 cm, 1000 µ) with multiple
development in 97:3 CHCl3/MeOH to provide 11 (42 mg, 65%) and 11α
(4 mg, 6%) as white solids. 11:TLC 97:3 CHCl3/MeOH, Rf 0.50; MS m/z
1
469 (m+H)+; H NMR (CDCl3). 8.32 (bs, 1H, H-3), 8.06–8.14 (m, 4H,
ortho H’s of benzoyl), 7.44–7.70 (m, 7H, H-6 and para and meta H’s of
benzoyl), 6.37 (dd, 1H, H-1ꢁ, J1 ,2 = 4 Hz and J1 ,F = 20 Hz), 5.86 (dd,1H,
ꢁ
ꢁ
ꢁ
H-3ꢁ, J2 ,3 = 2 Hz and J3 ,F = 18 Hz), 5.66 (bd, 1H, H-5, J5,6 = 8 Hz), 5.32
ꢁ
ꢁ
ꢁ
(ddd, 1H, H-2ꢁ, J1 ,2 = 4 Hz and J2 ,F = 50 Hz), 4.56–4.64 (m, 2H, 5ꢁ-CH2),
ꢁ
ꢁ
ꢁ
1.46 (s, 3H, 4ꢁ-CH3). 11α:TLC 97:3 CHCl3/MeOH, Rf 0.46; MS m/z 469
1
(m+H)+; H NMR (CDCl3). 8.32 (bs, 1H, H-3), 7.98–8.14 (m, 4H, ortho
H’s of benzoyl), 7.44–7.68 (m, 7H, H-6 and para and meta H’s of benzoyl),
6.24 (dd, 1H, H-1ꢁ, J1 ,2 = 3 Hz and J1 ,F = 16 Hz), 5.80–5.92 (m, 2H, H-3ꢁ
ꢁ
ꢁ
ꢁ

670
K. N. Tiwari et al.
and H-5), 5.41 (dt, 1H, H-2ꢁ, J2 ,3 = 2 Hz and J2 ,F = 50 Hz), 4.64–4.68 (m,
ꢁ
ꢁ
ꢁ
1H, 5ꢁ-CH2), 4.40–4.46 (m, 1H, 5ꢁ-CH2), 1.52 (s, 3H, 4ꢁ-CH3).
1-(3,5-Di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-β-D-arabinofuranosyl)
thymine (12). Compound 9 [prepared from 8 (60 mg, 0.14 mmol)] was
treated with 99% thymine (62 mg, 0.48 mmol) as described for 11 to
give 12 (43 mg, 62%) and 12α (4 mg, 6%) as white solids. 12:TLC 98:2
1
CHCl3/MeOH, Rf 0.43; MS m/z 483 (m+H)+; H NMR (CDCl3) 8.24 (bs,
1H, H-3), 7.44–7.70 (m, 10H, aromatic H’s), 7.34 (s,1H, H-6), 6.38 (dd,
1H, H-1ꢁ, J1 ,2 = 4 Hz and J1 ,F = 20 Hz), 5.88 (dd,1H, H-3ꢁ, J3 ,F = 18 Hz),
ꢁ
ꢁ
ꢁ
ꢁ
5.31 (ddd, 1H, H-2ꢁ, J2 ,3 = 2 Hz, J2 ,1 = 4 Hz and J2 ,F = 50 Hz), 4.62
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
(s, 2H, 5ꢁ-CH2), 1.74 (s, 3H, 5-CH3), 1.46 (s, 3H, 4ꢁ-CH3). 12α:TLC 98:2
1
CHCl3/MeOH, Rf 0.39; MS m/z 483(m+H)+; H NMR (CDCl3) 8.36 (bs,
1H, H-3), 7.46–7.68 (m, 10H, aromatic H’s), 7.30 (s, 1H, H-6), 6.28 (dd,
1H, H-1ꢁ, J1 ,2 = 3 Hz and J1 ,F = 16 Hz), 5.87 (dd,1H, H-3ꢁ, J2 ,3 = 2 Hz and
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
J3 ,F = 16 Hz), 5.40 (dt, 1H, H-2ꢁ, J2 ,F = 50 Hz), 4.40–4.66 (m, 2H, 5ꢁ-CH2),
ꢁ
ꢁ
1.98 (s, 3H, 5-CH3), 1.52 (s, 3H, 4ꢁ-CH3).
1-(2-Deoxy-2-fluoro-4-C-methyl-β-D-arabinofuranosyl)cytosine (13).[13]
A
suspension of 10 (334 mg, 0.58 mmol) in MeOH (30 mL) at room
temperature was treated dropwise with 0.5 M sodium methoxide in MeOH
(0.58 mL). The solid dissolved after 15 minutes, and the solution was stirred
for 2 hours, neutralized to pH 7 with glacial acetic acid, and evaporated to
an oil. Preparative TLC on silica gel (Analtech GF, 20 × 20 cm, 2000 µ)
with 3:1:0.10 CHCl3/MeOH/concentrated NH4OH as eluant provided the
nucleoside as a hygroscopic white foam. This residue was dissolved in 2-
PrOH (10 mL), 1.0 M HCl in ether (1.17 mL) was added and the mixture
was evaporated. From a suspension of this material in acetone was recovered
hydrochloride 13 (157 mg, 89%) as a white solid: m.p. 230–231◦C; TLC
3:1:0.1 CHCl3/MeOH/NH4OH, Rf 0.40; HPLC 99%, tR = 8.6 minutes, 9:1
NH4H2PO4 (0.01 M, pH 5.1)/MeOH; MS m/z 260 (M+H)+; UV λmax pH
1
1, 279 (13.8), pH 7, 269 (9.4), pH 13, 271 (9.6); H NMR (DMSO-d6) 9.70
(s, 1H, 4-NH2), 8.62 (s, 1H, 4-NH2), 8.24 (d, 1H, H-6, J = 8 Hz), 6.10–6.17
(m, 2H, H-1ꢁ and H-5 overlapped), 5.95 (broad hump, 1H, OH), 5.23 (dt,
1H, H-2ꢁ, J2 ,3 = 4 Hz and J2 ,F = 52 Hz), 4.26 (dd,1H, H-3ꢁ, J3 ,F = 20
Hz), 3.40–3.50 (m, 3H, 5ꢁ-CH2 and OH), 1.25 (s, 3H, 4ꢁ-CH3). Anal. Calcd.
For C10H14FN3O4·HCl 0.10 C3H8O, 0.20 H2O: C, 40.52; H, 5.35; N, 13.76.
Found: C, 40.82; H, 5.18; N, 13.58.
ꢁ
ꢁ
ꢁ
ꢁ
The α-anomer was prepared from 10α as described for 13 except the
HCl salt formation was omitted. Pure 13α (77%) was obtained from acetone
as a white solid: m.p. 222–223◦C; TLC 3:1:0.1 CHCl3/MeOH/NH4OH,
Rf 0.40; HPLC 100%, tR = 7.7 minutes, 9:1 NH4H2PO4 (0.01 M, pH
5.1)/MeOH; MS m/z 260 (M+H)+; UV λmax pH 1, 278 (13.3), pH 7,
1
270 (9.3), pH 13, 271 (9.3); H NMR (DMSO-d6) 7.59 (d, 1H, J = 3.5,
H-6), 7.28 (bs, 1H, 4-NH2), 7.20 (bs,1H, 4-NH2), 6.0 (dd, 1H, H-1ꢁ J1 ,2
=
ꢁ
ꢁ
4 Hz and J1 ,F = 18 Hz), 5.74–5.78 (m, 2H, H-5 and 3ꢁ-OH overlapped),
ꢁ

4ꢁ-C-Methyl-2ꢁ-fluoro Arabino Nucleosides
671
4.9–5.10 m, 2H, H-2ꢁ and 5ꢁ-OH overlapped), 4.24 (dt, 1H, H-3ꢁ, J3 ,F
=
ꢁ
18 Hz), 3.30–3.40 (m, 2H, 5ꢁ-CH2), 1.24 (s, 3H, 4ꢁ-CH3). Anal. Calcd. For
C10H14FN3O4: C, 46.33; H, 5.44; N, 16.21. Found: C, 46.19; H, 5.23; N,
16.09.
1-(2-Deoxy-2-fluoro-4-C-methyl-β-D-arabinofuranosyl)uracil
(14).[13]
Compound 14 was prepared from 11 using the method described for
13α. Pure 14 (77%) was obtained as a clear glass from acetone and
was subsequently ground to a white powder: m.p. 70–75◦C; TLC 3:1:0.1
CHCl3/MeOH/NH4OH, Rf 0.62; HPLC 99%, tR = 6.4 minutes, 85:15
NH4H2PO4 (0.01 M, pH 5.1)/MeOH; MS m/z 261 (M+H)+; UV λmax
pH 1, 261 (10.5), pH 7, 261 (10.3), pH 13, 261 (8.0); 1H NMR (DMSO-d6)
11.42 (broad hump,1H, H-3), 7.86 (d, 1H, H-6, J = 8 Hz), 6.20 (dd, 1H,
H-1ꢁ, J1 ,2 = 4 Hz and J1 ,F = 12 Hz), 5.92 (bs, 1H, 3ꢁ-OH), 5.72 (d, 1H,
ꢁ
ꢁ
ꢁ
H-5, J = 8 Hz), 5.23 (dt, 2H, H-2ꢁ, J2 ,3 = 4 Hz and J2 ,F = 52 Hz and 5ꢁ-OH
ꢁ
ꢁ
ꢁ
overlapped and exchanged with D2O), 4.28 (dd, 1H, H-3ꢁ, J3 ,3 -OH = 4 Hz
ꢁ
ꢁ
and J3 ,F = 22 Hz), 3.40–3.44 (m, 2H, 5ꢁ-CH2), 1.10 (s, 3H, 4ꢁ-CH3). Anal.
ꢁ
Calcd. For C10H13 F N2 O5 ·0.25 H2O: C, 45.37; H, 5.14; N, 10.58. Found: C,
45.32; H, 4.89; N, 10.42.
1-(2-Deoxy-2-fluoro-4-C-methyl-β-D-arabinofuranosyl)thymine
(15).
Compound 15 was synthesized from 12 using the conditions described for
13α but with a 7 hour reaction time and with 5:1 CHCl3/MeOH + 1%
concentrated NH4OH as eluent for preparative TLC. Pure 15 (92%) was
recovered from acetone (as for 14) as a white powder: m.p. 80◦C; TLC 5:1
CHCl3/MeOH + 1% NH4OH, Rf 0.52; HPLC 100%, tR = 5.2 minutes, 3:1
NH4H2PO4 (0.01 M, pH 5.1)/MeOH; MS m/z 275 (M+H)+; UV λmax pH
1
1, 267 (9.9), pH 7, 267 (9.8), pH 13, 266 (7.8); H NMR (DMSO-d6) 11.40
(s, 1H, H-3), 7.74 (s, 1H, H-6), 6.14 (dd, 1H, H-1ꢁ, J1 ,2 = 6 Hz and J1 ,F
=
ꢁ
ꢁ
ꢁ
12 Hz), 5.88 (bd, 1H, 3ꢁ-OH, J = 6 Hz), 5.3–5.40 (m, 1H, 5ꢁ-OH), 5.18 (dt,
1H, H-2ꢁ, J2 ,3 = 4 Hz and J2 ,F = 52 Hz), 4.31 (dt, 1H, H-3ꢁ, J3 ,F = 22 Hz),
3.40–3.48 (m, 2H, 5ꢁ-CH2), 1.80 (s, 3H, 5-CH3), 1.08 (s, 3H, 4ꢁ-CH3). Anal.
Calcd. For C11H15 F N2 O5·0.50 H2O: C, 46.64; H, 5.69; N, 9.89. Found: C,
46.54; H, 5.33; N, 9.68.
ꢁ
ꢁ
ꢁ
ꢁ
6-Chloro-9-(3,5-di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-β-D-arabinofu-
ranosyl)purine (16). A suspension of 98% 6-chloropurine (102 mg, 0.65
mmol) in dry MeCN (5 mL) at room temperature was treated in one
portion with 60% NaH (35 mg, 0.88 mmol). The mixture was stirred for
40 minutes before the immediate addition of 9 [prepared from 8 (177
mg, 0.43 mmol)] dissolved in MeCN (2 mL). After 6 hours, the stirred
mixture was adjusted to about pH 6 with glacial acetic acid, and stirring was
continued 15 minutes before the solids present were collected, washed with
MeCN, and discarded. The combined filtrate and washings were evaporated
to a yellow residue that was purified by silica gel preparative TLC (Analtech
GF, 20 × 20 cm, 2000 µ) with 99:1 CHCl3/MeOH as eluent. The recovered
anomeric product was resolved by preparative TLC developed twice in 2:1

672
K. N. Tiwari et al.
hexane/EtOAc to provide 16 (79 mg, 36%) and 16α (31 mg, 14%) as white
solids: 16, TLC 2:1 hexane/EtOAc, Rf 0.40; MS m/z 511 (M+H)+; 1H NMR
(CDCl3) 8.78 (s,1H, H-2), 8.42 (d, 1H, H-8, J = 4 Hz), 8.10–8.16 (m, 4H,
ortho H’s of benzoyl), 7.46–7.72 (m, 6H, para and meta H’s of benzoyl),
6.76 (dd, 1H, H-1ꢁ, J1 ,2 = 4 Hz and J1 ,F = 22 Hz), 6.04 (dd, 1H, H-3ꢁ, J3 ,F
=
ꢁ
ꢁ
ꢁ
ꢁ
16 Hz), 5.39 (ddd, 1H, H-2ꢁ, J2 ,3 = 2 Hz and J2 ,F = 50 Hz), 4.61–4.76 (m,
ꢁ
ꢁ
ꢁ
2H, 5ꢁ-CH2), 1.58 (s, 3H, 4ꢁ-CH3). 16α, TLC 2:1 hexane/EtOAc, Rf 0.52;
1
MS m/z 511 (M+H)+; H NMR (CDCl3) 8.74 (s, 1H, H-2), 8.38 (s, 1H,
H-8), 7.86–8.66 (m, 4H, ortho H’s of benzoyl), 7.42–7.66 (m, 6H, para and
meta H’s of benzoyl), 6.51 (dd, 1H, H-1ꢁ, J1 ,2 = 3.5 Hz and J1 ,F = 14 Hz),
ꢁ
ꢁ
ꢁ
6.24 (dt, 1H, H-2ꢁ, J2 ,3 = 3 Hz and J2 ,F = 50 Hz), 6.00 (dd, 1H, H-3ꢁ, J3 ,F
=
ꢁ
ꢁ
ꢁ
ꢁ
18 Hz), 4.46–4.72 (m, 2H, 5ꢁ-CH2), 1.60 (s, 3H, 4ꢁ-CH3).
9-(2-Deoxy-2-fluoro-4-C-methyl-β-D-arabinofuranosyl)adenine (17).[13]
Compound 16 (101 mg, 0.20 mmol) in a glass lined stainless steel bomb
was diluted with ethanolic ammonia (100 mL, saturated at 5◦C). The sealed
bomb was heated at 80◦C for 26 hours before the contents were evaporated.
The residue was purified by multiple development on silica gel preparative
TLC (Analtech GF, 10 × 20 cm, 1000 µ) with 5:1 CHCl3/MeOH + 1%
NH4OH as solvent. Pure 17 (45 mg, 74%) was obtained as a white powder
from acetone: m.p. 160–162◦C; TLC 5:1 CHCl3/MeOH + 1% NH4OH, Rf
0.45; MS m/z 284 (M+H)+; UV λmax pH 1, 256 (15.0), pH 7, 259 (15.7),
pH 13, 259 (16.3); 1H NMR (DMSO-d6) 8.30 (d, 1H, H-8, J = 1 Hz), 8.14 (s,
1H, H-2), 7.32 (s, 2H, 6-NH2), 6.43 (dd, 1H, H-1ꢁ, J1 ,2 = 5 Hz and J1 ,F = 10
ꢁ
ꢁ
ꢁ
Hz), 5.94 (bs, 1H, 3ꢁ-OH), 5.35 (dt, 1H, H-2ꢁ, J2 ,3 = 4 Hz and J2 ,F = 52 Hz),
ꢁ
ꢁ
ꢁ
5.22–5.30 (m, 1H, 5ꢁ-OH), 4.56 (dd, 1H, H-3ꢁ, J3 ,F = 20 Hz), 3.46–3.52 (m,
ꢁ
2H, 5ꢁ-CH2), 1.16 (s, 3H, 4ꢁ-CH3). Anal. Calcd. For C11H14 F N5 O3 ·0.40
H2O·0.30 C3H6O: C, 46.42; H, 5.43; N, 22.75. Found: C, 46.44; H, 5.17; N,
22.83. 17α was prepared from 16α as described for 17. Pure 17α (49%) as
a glass from acetone was ground to an off white powder: m.p. 185–187◦C;
TLC 5:1 CHCl3/MeOH + 1% NH4OH, Rf 0.51; MS m/z 284 (M+H)+;
1
UV λmax pH 1, 257 (15.0), pH 7, 259 (15.5), pH 13, 259 (16.0); H NMR
(DMSO-d6) 8.34 (s,1H, H-8), 8.18 (s, 1H, H-2), 7.38 (s, 2H, 6-NH2), 6.16
(dd, 1H, H-1ꢁ, J1 ,2 = 5 Hz and J1 ,F = 10 Hz), 6.10 (bs, 1H, 3ꢁ-OH), 5.84
ꢁ
ꢁ
ꢁ
(dt, 1H, H-2ꢁ, J2 ,3 = 4 Hz and J2 ,F = 52 Hz), 5.14 (t, 1H, 5ꢁ-OH, J = 4 Hz),
ꢁ
ꢁ
ꢁ
4.49 (dd, 1H, H-3ꢁ, J3 ,F = 20 Hz), 3.32–3.40 (m, 2H, 5ꢁ-CH2), 1.22 (s, 3H,
ꢁ
4ꢁ-CH3). Anal. Calcd. For C11H14FN5O3·0.50 H2O·0.10 C3H6O: C, 45.53; H,
5.28; N, 23.49. Found: C, 45.40; H, 5.02; N, 23.48.
2,6-Dichloro-9-(3,5-di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-β-D-arabin-
ofuranosyl)purine (18). Compound 18 was synthesized as described for
16 except using 2,6-dichloropurine. The anomeric product mixture
was isolated by silica gel preparative TLC (developed twice in 3:1
1
hexane/EtOAc) as a white solid (64%, as 2:1 β: α ratio by H NMR).
Pure anomers were obtained separately as white foams by preparative TLC
with multiple developments in CHCl3: 18, TLC 100:1 CHCl3/MeOH, Rf

4ꢁ-C-Methyl-2ꢁ-fluoro Arabino Nucleosides
673
0.48; MS m/z 545 (M+H)+; 1H NMR (CDCl3) 8.40 (d, 1H, H-8, J = 4 Hz),
8.10–8.16 (m, 4H, ortho H’s of benzoyl), 7.46–7.72 (m, 6H, para and meta
H’s of benzoyl), 6.68 (dd, 1H, H-1ꢁ, J1 ,2 = 3 Hz and J1 ,F = 20 Hz), 6.01
ꢁ
ꢁ
ꢁ
(dd, 1H, H-3ꢁ, J3 ,4 = 1.5 and 16 Hz), 5.38 (ddd, 1H, H-2ꢁ, J2 ,3 = 1.5 Hz and
ꢁ
ꢁ
ꢁ
ꢁ
J2 ,F = 50 Hz), 4.60–4.74 (m, 2H, 5ꢁ-CH2), 1.56 (s, 3H, 4ꢁ-CH3). 18α, TLC
ꢁ
100:1 CHCl3/MeOH, Rf 0.42; MS m/z 545 (M+H)+; H NMR (CDCl3)
1
8.34 (s,1H, H-8), 7.92–8.16 (m, 4H, ortho H’s of benzoyl), 7.45–7.68 (m,
6H, para and meta H’s of benzoyl), 6.48 (dd, 1H, H-1ꢁ, J1 ,2 = 4 Hz and J1 ,F
ꢁ
ꢁ
ꢁ
= 16 Hz), 6.12 (dt, 1H, H-2ꢁ, J2 ,3 = 1.5 Hz and J2 ,F = 50 Hz), 6.01 (dd, 1H,
ꢁ
ꢁ
ꢁ
H-3ꢁ, J3 ,F = 16 Hz), 4.48–4.72 (m, 2H, 5ꢁ-CH2), 1.60 (s, 3H, 4ꢁ-CH3).
ꢁ
2,6-Diazido-9-(3, 5-di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-β-D-arabin-
ofuranosyl)purine (19). To a solution of 18 (108 mg, 0.20 mmol) in EtOH
(5 mL) was added solid NaN3 (30 mg, 0.46 mmol) and H2O (0.5 mL).
The mixture was placed in a 100◦C bath, refluxed 30 minutes, cooled, and
evaporated. The residue was partitioned between CHCl3 and H2O. The
aqueous layer was extracted twice with CHCl3, and the combined organic
layers were washed with H2O, dried (MgSO4), and evaporated to a syrup.
Purification by silica gel preparative TLC (Analtech GF, 10 × 20 cm, 1000 µ)
in 2:1 hexane/EtOAc afforded 19 (105 mg, 95%) as a white solid that was
used directly in the next step. TLC 99:1 CHCl3/MeOH, Rf 0.65; MS m/z
1
559 (M+H)+; H NMR (CDCl3) 8.18 (d, 1H, H-8, J = 4 Hz), 8.10–8.16
(m, 4H, ortho H’s of benzoyl), 7.46–7.72 (m, 6H, para and meta H’s of
benzoyl), 6.62 (dd, 1H, H-1ꢁ, J1 ,2 = 4 Hz and J1 ,F = 22 Hz), 6.00 (dd, 1H,
ꢁ
ꢁ
ꢁ
H-3ꢁ, J2 ,3 = 1.5 Hz and J3 ,F = 16 Hz), 5.34 (ddd, 1H, H-2ꢁ, J2 ,F = 50 Hz),
4.90–4.70 (m, 2H, 5ꢁ-CH2), 1.54 (s, 3H, 4ꢁ-CH3). 19α was prepared from 18α
as described for 19: TLC 99:1 CHCl3/MeOH, Rf 0.58; MS m/z 559 (M+H)+;
1H NMR (CDCl3) 8.14 (s,1H, H-8), 7.92–8.14 (m, 4H, ortho H’s of benzoyl),
ꢁ
ꢁ
ꢁ
ꢁ
7.44–7.66 (m, 6H, para and meta H’s of benzoyl), 6.41 (dd, 1H, H-1ꢁ, J1 ,2
=
ꢁ
ꢁ
4 Hz and J1 ,F = 12 Hz), 6.16 (dt, 1H, H-2ꢁ, J2 ,3 = 1.5 Hz and J2 ,F = 50 Hz),
ꢁ
ꢁ
ꢁ
ꢁ
6.01 (dd, 1H, H-3ꢁ, J3 ,F = 14 Hz), 4.48–4.68 (m, 2H, 5ꢁ-CH2), 1.58 (s, 3H,
ꢁ
4ꢁ-CH3).
2,6-Diamino-9-(3,
5-di-O-benzoyl-2-deoxy-2-fluoro-4-C-methyl-β-D-
arabinofuranosyl)purine (20). A solution of 19 (105 mg, 0.19 mmol) in
2:1 EtOH/DMAc (15 mL) was treated with 10% palladium on carbon
(17 mg) and hydrogenated for 18 hours at room temperature and
atmospheric pressure. The catalyst was removed by filtration and washed
thoroughly with CHCl3. The combined filtrate and washings were
evaporated under vacuum to a syrup. Purification by silica gel preparative
TLC (Analtech GF, 10 × 20 cm, 1000 µ) with multiple development in 95:5
CHCl3/MeOH + 1% NH4OH gave 20 (84 mg, 87%) as a white residue that
was used directly in the next step. TLC 95:5 CHCl3/MeOH + 1% NH4OH,
1
Rf 0.43; MS m/z 507 (M+H)+; H NMR (CDCl3) 8.08–8.14 (m, 4H, ortho
H’s of benzoyl), 7.80 (d, 1H, H-8, J = 3 Hz), 7.42–7.70 (m, 6H, para and
meta H’s of benzoyl), 6.48 (dd, 1H, H-1ꢁ, J1 ,2 = 4 Hz and J1 ,F = 22 Hz),
ꢁ
ꢁ
ꢁ

674
K. N. Tiwari et al.
6.01 (dd, 1H, H-3ꢁ, J3 ,F = 16 Hz), 5.30 (ddd, 1H, H-2ꢁ, J2 ,3 = 1.5 Hz, J2 ,F
= 50 Hz), 5.38 (bs, 2H, 2-NH2), 4.74 (bs, 2H, 6-NH2), 4.60–4.68 (m, 2H,
5ꢁ-CH2), 1.52 (s, 3H, 4ꢁ-CH3). 20α was prepared from 19α as described for
20: TLC 95:5 CHCl3/MeOH + 1% NH4OH, Rf 0.48; MS m/z 507 (M+H)+;
1H NMR (CDCl3) 8.12–8.16 (m, 4H, ortho H’s of benzoyl), 7.78 (s, 1H,
H-8), 7.44–7.96 (m, 6H, para and meta H’s of benzoyl), 6.30 (dd, 1H, H-1ꢁ,
ꢁ
ꢁ
ꢁ
ꢁ
J1 ,2 = 3 Hz and J2 ,F = 16 Hz), 5.95 (dd, 1H, H-3ꢁ, J2 ,3 = 2.8 Hz and J3 ,F
= 14 Hz), 5.32 (bs, 2H, 2-NH2), 4.66 (bs, 2H, 6-NH2), 4.46–4.62 (m, 2H,
5ꢁ-CH2), 1.56 (s, 3H, 4ꢁ-CH3).
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
2,6-Diamino-9-(2-deoxy-2-fluoro-4-C-methyl-β-D-arabinofuranosyl) pur-
ine (21).[13] To a solution of 20 (84 mg, 0.17 mmol) in MeOH (5 mL)
at room temperature was added 0.5 N NaOCH3 in MeOH (0.17 mL).
The solution was stirred 3 hours, neutralized to pH 6 with glacial acetic
acid, and evaporated. Purification by development on silica gel preparative
TLC (Analtech GF, 10 × 20 cm, 500 µ) using 5:1 CHCl3/MeOH + 1%
NH4OH as solvent gave 21 (43 mg, 85%) as a white solid from acetone:
m.p. 210–215◦C; TLC 5:1 CHCl3/MeOH + 1% NH4OH, Rf 0.39; HPLC
99%, tR = 9.6 minutes, 20 minute linear gradient from 10–90% MeOH in
0.01 M NH4H2PO4 (pH 5.1); MS m/z 299 (M+H)+; HRMS m/z 299.12634
(M+H)+, Calcd. 299.12624 (M+H)+; UV λmax pH 1, 252 (11.5), 290
1
(9.9), pH 7, 255 (9.6), 280 (10.2), pH 13, 255 (9.8), 280 (10.5); H NMR
(DMSO-d6) 7.84 (d, 1H, H-8, J = 1 Hz), 6.74 (s, 2H, 2-NH2), 6.20 (dd, 1H,
H-1ꢁ, J1 ,2 = 5 Hz and J1 ,F = 13 Hz), 5.82–5.92 (m, 3H, 6-NH2 and 3ꢁ-OH),
ꢁ
ꢁ
ꢁ
5.24 (t,1H, 5ꢁ-OH, J = 4 Hz), 5.22 (dt, 1H, H-2ꢁ, J2 ,3 = 4 Hz and J2 ,F
=
ꢁ
ꢁ
ꢁ
52 Hz), 4.51 (dd, 1H, H-3ꢁ, J3 ,F = 20 Hz), 3.40–3.48 (m, 2H, 5ꢁ-CH2), 1.14
(s, 3H, 4ꢁ-CH3). Anal. Calcd. For C11H15 F N6 O3·0.40 H2O: C, 43.25; H,
5.21; N, 27.51. Found: C, 43.59; H, 5.09; N, 27.21. 21α was prepared from
20α as described for 21 except white solid was obtained from MeOH: m.p.
130–135◦C; TLC 5:1 CHCl3/MeOH + 1% NH4OH, Rf 0.45; HPLC 99%,
tR = 9.9 minutes, 20 minute linear gradient from 10–90% MeOH in 0.01
M NH4H2PO4 (pH 5.1); MS m/z 299 (M+H)+; HRMS m/z 299.12583
(M+H)+, Calcd. 299.12624 (M+H)+; UV λmax pH 1, 252 (12.4), 292
(10.5), pH 7, 255 (10.0), 280 (10.5), pH 13, 256 (9.8), 280 (10.5); 1H NMR
(DMSO-d6) 7.92 (s, 1H, H-8), 6.76 (s, 2H, 2-NH2), 6.78 (s, 1H, 3ꢁ-OH), 6.00
ꢁ
(dd, 1H, H-1ꢁ, J1 ,2 = 4 Hz and J1 ,F = 16 Hz), 5.88 (s, 2H, 6-NH2), 5.68
ꢁ
ꢁ
ꢁ
(dt, 1H, H-2ꢁ, J2 ,3 = 4 Hz and J2 ,F = 52 Hz), 5.11 (t,1H, 5ꢁ-OH, J = 4 Hz),
ꢁ
ꢁ
ꢁ
4.44 (dd, 1H, H-3ꢁ, J3 ,F = 18 Hz), 3.32–3.38 (m, 2H, 5ꢁ-CH2), 1.22 (s, 3H,
ꢁ
4ꢁ-CH3). Anal. Calcd. For C11H15 F N6 O3·1.5 H2O: C, 40.62; H, 5.58; N,
25.83. Found: C, 40.49; H, 5.26; N, 25.79.
9-(2-Deoxy-2-fluoro-4-C-methyl-β-D-arabinofuranosyl)guanine (22).[13]
A
suspension of 21 (143 mg, 0.8 mmol) in H2O (10 mL) was warmed to 70◦C
to dissolve solid before being cooled to 30◦C. Solid adenosine deaminase
(22 mg, 33 units, type II: crude powder Sigma) was added, and the solution
was stirred at room temperature. After 2.5 hours, the solution became

4ꢁ-C-Methyl-2ꢁ-fluoro Arabino Nucleosides
675
cloudy, and the stirring was continued for 68 hours. The resulting milky
mixture was filtered to remove crude 22 as a white solid (77 mg). The
product containing filtrate was applied directly to a strong cation exchange
resin in the H+ form (30 mL, AG 50W-X4,100–200 mesh) equilibrated in
H2O. Elution with 0.25 N NH4OH yielded 22 containing small amounts of
UV active impurities. These impurities were removed by preparative TLC on
silica gel (Analtech GF, 10 × 20 cm, 1000 µ) using 9:2 MeCN/1N NH4OH as
eluent to provide more crude 22 (66 mg) as a white solid. Both solids were
combined in hot water, and the solution was diluted to 50 mL. This solution
at room temperature was applied to a XAD-4 resin column (100–200 mesh,
1 × 8.5 cm) equilibrated in H2O. Elution with H2O was continued followed
by 9:1 H2O/MeOH, when 22 appeared in the eluate. The pooled product
containing fractions were evaporated and the residue was triturated with
EtOH (25 mL) to give pure 22 (96 mg, 67%) as a white solid: m.p. 270◦C
(dec.); TLC 9:2 MeCN/1N NH4OH, Rf 0.55; HPLC 100%, tR = 8.3 minutes,
85:15 NH4H2PO4 (0.01 M, pH 5.1)/MeOH; MS m/z 300 (M+H)+; HRMS
m/z 322.09177 (M+Na)+, Calcd. 322.09220 (M+Na)+; UV λmax pH 1, 256
1
(13.1), 280 (sh), pH 7, 251 (14.5), 276 (sh), pH 13, 263 (12.1); H NMR
(DMSO-d6) 10.66 (s, 1H, 3-NH), 7.88 (d, 1H, H-8, J = 1 Hz), 6.50 (s, 2H,
2-NH2), 6.14 (dd, 1H, H-1ꢁ, J1 ,2 = 5 and J1 ,F = 13 Hz), 5.90 (d, 1H, 3ꢁ-OH, J
ꢁ
ꢁ
ꢁ
= 5 Hz), 5.23 (dt, 1H, H-2ꢁ, J2 ,3 = 4 Hz and J2 ,F = 52 Hz), 5.17 (t, 1H, 5ꢁ-OH,
ꢁ
ꢁ
ꢁ
J = 4 Hz), 4.45 (dt, 1H, H-3ꢁ, J3 ,F = 18 Hz), 3.32–3.48 (m, 2H, 5ꢁ-CH2), 1.16
ꢁ
(s, 3H, 4ꢁ-CH3). Anal. Calcd. For C11H14 F N5 O4 : C, 44.15; H, 4.72; N, 23.40.
Found: C, 43.93; H, 4.69; N, 23.19.
2-Chloro-6-methoxy-9-(2-deoxy-2-fluoro-4-C-methyl-β-D-
arabinofuranosyl)purine (23). To a solution of 18 (76 mg, 0.14 mmol)
in anhydrous MeOH (5 mL) at room temperature was added dropwise
0.5 N NaOCH3 in MeOH (280 µl). The solution was stirred for 3 hours,
neutralized to pH 6 with glacial acetic acid, and evaporated. The residue was
purified by preparative TLC on silica gel using 9:1 CHCl3/MeOH as solvent.
Pure 23 (43 mg, 93%) was obtained as a clear glass and was used directly in
1
the next step: TLC 9:1 CHCl3/MeOH, Rf 0.48; MS m/z 333 (M+H)+; H
NMR (DMSO-d6) 8.64 (d, 1H, H-8, J = 1 Hz), 6.47 (dd, 1H, H-1ꢁ, J1 ,2 = 5
ꢁ
ꢁ
Hz and J1 ,F = 10 Hz), 5.98 (bs, 1H, 3ꢁ-OH), 5.43 (dt, 1H, H-2ꢁ, J2 ,3 = 4 Hz
ꢁ
ꢁ
ꢁ
and J2 ,F = 52 Hz), 5.30 (t,1H, 5ꢁ-OH, J = 4 Hz), 4.54 (dd, 1H, H-3ꢁ, J3 ,F = 20
Hz), 4.12 (s, 3H, 6-OCH3), 3.48–3.52 (m, 2H, 5ꢁ-CH2), 1.14 (s, 3H, 4ꢁ-CH3).
23α was prepared from 18α as described for 23. Pure 23α (96%) was
obtained as a glass: TLC 9:1 CHCl3/MeOH, Rf 0.48; MS m/z 333 (M+H)+;
ꢁ
ꢁ
1H NMR (DMSO-d6) 8.62 (d,1H, H-8, J = 1 Hz), 6.20 (dd, 1H, H-1ꢁ, J1 ,2
ꢁ
ꢁ
= 5 Hz and J1 ,F = 14 Hz), 6.16 (bs, 1H, 3ꢁ-OH), 5.76 (dt, 1H, H-2ꢁ, J2 ,3
=
ꢁ
ꢁ
ꢁ
5 Hz and J2 ,F = 52 Hz), 5.24 (bt,1H, J = 4.50, 5ꢁ-OH), 4.54 (dd, 1H, H-3ꢁ,
ꢁ
J3 ,F = 20 Hz), 4.12 (s, 3H, 6-OCH3), 3.34–3.38 (m, 2H, 5ꢁ-CH2), 1.24 (s, 3H,
ꢁ
4ꢁ-CH3).

676
K. N. Tiwari et al.
2-Chloro-9-(2-deoxy-2-fluoro-4-C-methyl-β-D-arabinofuranosyl)adenine
(24). A solution of 23 (88 mg, 0.27 mmol) in 60 mL of ethanolic ammonia
(saturated at 5◦C) in a glass-lined stainless steel bomb was heated at 80◦C for
21 hours and evaporated. The residue was purified by silica gel preparative
TLC (Analtech GF, 10 × 20 cm, 1000 µ) with two developments in 5:1
CHCl3/MeOH + 1% NH4OH. Pure 24 (71 mg, 84%) as a white foam
from acetone was ground to a white powder: m.p. 220–222◦C; TLC 9:1
CHCl3/MeOH + 1% NH4OH, Rf 0.29; HPLC 97%, tR = 14 minutes, 4:1
NH4H2PO4 (0.01 M, pH 2.7)/MeOH; MS m/z 318 (M+H)+; UV λmax pH
1
1, 264 (14.7), pH 7, 264 (15.4), pH 13, 264 (15.8); H NMR (DMSO-d6)
8.44 (d, 1H, H-8, J = 1 Hz), 7.86 (s, 2H, 6-NH2), 6.35 (dd, 1H, H-1ꢁ, J1 ,2 = 5
ꢁ
ꢁ
Hz and J1 ,F = 10 Hz), 5.94 (d, 1H, 3ꢁ-OH, J = 6 Hz), 5.37 (dt, 1H, H-2ꢁ, J2 ,3
ꢁ
ꢁ
ꢁ
= 4 Hz and J2 ,F = 52 Hz), 5.25 (t, 1H, 5ꢁ-OH, J = 4 Hz), 4.53 (dd, 1H, H-3ꢁ,
ꢁ
J3 ,F = 20 Hz), 3.48–3.50 (m, 2H, 5ꢁ-CH2), 1.14 (s, 3H, 4ꢁ-CH3). Anal. Calcd.
ꢁ
For C11H13Cl F N5 O3 : C, 41.59; H, 4.12; N, 22.04. Found: C, 41.55; H,
4.12; N, 22.06. 24α was prepared from 23α as described for 24. Pure 24α
(75%) was recovered from acetone as a white solid: m.p., dual 105◦C and
205◦C; TLC 9:1 CHCl3/MeOH + 1% NH4OH, Rf 0.35; HPLC 98%, tR = 13
minutes, 4:1 NH4H2PO4 (0.01 M, pH 2.7)/MeOH; MS m/z 318 (M+H)+;
1
UV λmax pH 1, 264 (15.1), pH 7, 264 (16.1), pH 13, 264 (15.9); H NMR
(DMSO-d6) 8.38 (d, 1H, H-8, J = 1 Hz), 7.90 (s, 2H, 6-NH2), 6.09 (dd, 1H,
H-1ꢁ, J1 ,2 = 4 Hz and J1 ,F = 14 Hz), 6.02 (bs, 1H, 3ꢁ-OH), 5.74 (dt, 1H, H-2ꢁ,
ꢁ
ꢁ
ꢁ
J2 ,2 ,F = 4 and 52 Hz), 5.16 (t, 1H, 5ꢁ-OH, J = 4 Hz), 4.50 (dd, 1H, H-3ꢁ,
ꢁ
ꢁ
J3 ,4 = 4 and 20 Hz), 3.32–3.38 (m, 2H, 5ꢁ-CH2), 1.24 (s, 3H, 4ꢁ-CH3). Anal.
Calcd. For C11H13Cl F N5 O3. 0.65·H2O·0.10 C3H6O: C, 40.49; H, 4.48; N,
20.89. Found: C, 40.43; H, 4.35; N, 20.72.
ꢁ
ꢁ
REFERENCES
1. Plunkett, W.; Gandhi, V. Purine and pyrimidine nucleoside analogs. Cancer Chemother. Biol. Response
Modif. 2001, 19, 21–45.
2. Faderl, S.; Gandhi, V.; Kantarjian, H.; Plunkett, W. New nucleoside analogues in clinical develop-
ment. Cancer Chemother. Biol. Response Modif. 2002, 20, 37–58.
3. Parker, W.B.; Secrist, J.A. III; Waud, W.R. Purine nucleoside antimetabolites in development for the
treatment of cancer. Curr. Opin. Investig. Drugs 2004, 5, 592–596.
4. Bonate, P.L.; Arthaud, L.; Cantrell, W.R. Jr.; Stephenson, K.; Secrist, J.A. III; Weitman, S. Discovery
and development of clofarabine: A nucleoside analogue for treating cancer. Nat. Rev. Drug Discov.
2006, 5, 855–63. Review.
5. Nomura, M.; Shuto, S.; Tanaka, M.; Sasaki, T.; Mori, S.; et al. Nucleosides and nucleotides. 185.
Synthesis and biological activities of 4ꢁα-C-branched-chain sugar pyrimidine nucleosides. J. Med.
Chem. 1999, 42, 2901–2908.
6. Kitano, K.; Machida, H.; Miura, S.; Ohrui, H. Synthesis of novel 4ꢁ-C-methyl-pyrimidine nucleosides
and their biological activities. Bioorg. Med. Chem. Lett. 1999, 9, 827–830.
7. Sugimoto, I.; Shuto, S.; Mori, S.; Shigeta, S.; Matsuda, A. Nucleosides and nucleotides. 183. Synthesis
of 4ꢁα-branched thymidines as a new type of antiviral agent. Bioorg. Med. Chem. Lett. 1999, 9, 385–
388.
8. Waga, T.; Nishizaki, T.; Miyakawa, I.; Ohrui, H.; Meguro, H. Synthesis of 4ꢁ-C-methylnucleosides.
Biosci. Biotechnol. Biochem. 1993, 57, 1433–1438.

4ꢁ-C-Methyl-2ꢁ-fluoro Arabino Nucleosides
677
9. Waga, T.; Ohrui, H.; Meguro, H. Synthesis and biological evaluation of 4ꢁ-C-methyl nucleosides.
Nucleosides & Nucleotides 1996, 15, 287–304.
10. Yamaguchi, T.; Tomikawa, A.; Hirai, T.; Kawaguchi, T.; Ohrui, H.; Saneyoshi, M. Antileukemic
Activities and mechanism of action of 2ꢁ-deoxy-4ꢁ-methylcytidine and related nucleosides. Nucleosides
& Nucleotides 1997, 16, 1347–1350.
11. Griffon, J.-F., Shaddix, S.C., Parker, W.B., Al-Madhoun, A. S., Eriksson, S., et al., Synthesis and
biological evaluation of some 4ꢁ-C-(hydroxymethyl)-α- and -β-d-arabinofuranosyl pyrimidine and
adenine nucleosides. Collection Czech. Chem. Commun. 2006, 71(7), 1063–1087.
12. Shortnacy-Fowler, A.T.; Tiwari, K.N.; Montgomery, J.A.; Secrist, J.A. III. Synthesis and biological
activity of 4ꢁ-C-hydroxymethyl-2ꢁ-fluoro-D-arabinofuranosylpurine nucleosides. Nucleosides Nucleotides
& Nucleic Acids 2001, 20, 1583–1598.
13. Chang, J.; Bao, X.; Wang, Q.; Guo, X.; Wang, W.; Qi, X. Preparation of 2ꢁ-fluoro-4ꢁ-substituted
nucleoside analogs as antiviral agents. Chinese Patent Application # CN 2007–10137548, 20070807.
14. Klumpp, K.; Kalayanov, G.; Ma, H.; Le Pogam, S.; Leveque, V.; et al. 2ꢁ-Deoxy-4ꢁ-azido nucleoside
analogs are highly potent inhibitors of hepatitis C virus replication despite the lack of 2ꢁ-α-hydroxyl
groups. J. Biol. Chem. 2008, 283, 2167–2175.
15. Mandal, S.B.; Achari, B. Stereospecific C-β-glycosidation and synthesis of 4,7-anhydro-5,6-
isopropylidene-4 (S), 5(S), 6(R), 7(R)-tetrahydroxyoxocan-2-one. Synth. Commun. 1993, 23,
1239–1244.
16. Gunic, E.; Girardet, J.-L.; Pietrzkowski, Z.; Esler, C.; Wang, G. Synthesis and cytotoxicity of 4ꢁ-C- and
5ꢁ-C-substituted toyocamycins. Bioorg.Med. Chem. 2001, 9, 163–170.
17. Feather, M.S.; Whistler, R.L. Derivatives of 5-deoxy-5-mercapto-d-glucose. Tetrahedron Lett. 1962, 3,
667–668.
18. Ohrui, H.; Kohgo, S.; Kitano, K.; Sakata, S.; Kodama, E.; Yoshimura, K.; Matsuoka, M.;
Shigeta, S.; Mitsuya, H. Syntheses of 4ꢁ-C-ethynyl-β-d-arabino- and 4ꢁ-C-ethynyl-2ꢁ-deoxy-β-d-ribo-
pentofuranosylpyrimidines and purines and evaluation of their anti-HIV activity. J. Med. Chem., 2000,
43, 4516–4525.
19. Siddiqui, M.A.; Driscoll, J.S.; Marquez, V.E.; Roth, J.S.; Shirasaka, T.; Mitsuya, H.; Barchi, J.J. Jr.; Kel-
ley, J.A. Chemistry and anti-HIV properties of 2ꢁ-fluoro-2ꢁ,3ꢁ-dideoxyarabinofuranosylpyrimidines. J.
Med. Chem., 1992, 35, 2195–2201.
20. Kazimierczuk, Z.; Cottam, H.B.; Revankar, G.R.; Robins, R.K. Synthesis of 2ꢁ-deoxytubercidin,
2ꢁ-deoxyadenosine, and related 2ꢁ-deoxynucleosides via a novel direct stereospecific sodium salt
glycosylation procedure. J. Am. Chem. Soc. 1984, 106, 6379–6382.
21. Secrist, J.A. III, Tiwari, K.N.; Shortnacy-Fowler, A.T.; Messini, L.; Riordan, J.M.; Montgomery, J.A.
Synthesis and biological activity of certain 4ꢁ-thio-D-arabinofuranosylpurine nucleosides. J. Med.
Chem. 1998, 41, 3865–3871.