A new process for antineoplastic agent clofarabine

Article abstract of DOI:10.1021/op049884n

Clofarabine is a promising DNA polymerase inhibitor currently in clinical trials for a variety of liquid and solid tumor indications. The efforts for development of a new manufacturing process for clofarabine are presented. This new process allows for the reliable and efficient production of drug substance in high anomeric excess and high overall purity, without using chromatography. The high anomeric selectivity is achieved by reacting 2-chloroadenine with 1-bromo-2-deoxy-2-fluoro-3,5-di-O-benzoyl-α-D-ribofuranose (4) and potassium tert-butoxide in a mixture of three solvents. Following crystallization, anomeric ratios exceeding 50 (β/α) are achieved. Deprotection and additional crystallization afford a clofarabine drug substance containing less than 0.1% of the α-anomer.

Full text of DOI:10.1021/op049884n

Organic Process Research & Development 2004, 8, 889−896
|
A New Process for Antineoplastic Agent Clofarabine
William E. Bauta,* Brian E. Schulmeier, Brian Burke, Jose F. Puente, William R. Cantrell, Jr., Dennis Lovett,†
,
†
†,§
†,§
†,§
†
†
†
‡
‡
James Goebel, Bruce Anderson, Dumitru Ionescu, and Ruichao Guo
Department of Process Chemistry and Department of Analytical Chemistry, ILEX Products, Inc., 14805 Omicron DriVe,
San Antonio, Texas 78245, and Ash SteVens, Inc., 18655 Krause AVenue, RiVerView, Michigan 48192
Abstract:
Clofarabine is a promising DNA polymerase inhibitor currently
in clinical trials for a variety of liquid and solid tumor
indications. The efforts for development of a new manufacturing
process for clofarabine are presented. This new process allows
for the reliable and efficient production of drug substance in
high anomeric excess and high overall purity, without using
chromatography. The high anomeric selectivity is achieved by
Figure 1.
reacting 2-chloroadenine with 1-bromo-2-deoxy-2-fluoro-3,5-
di-O-benzoyl-r- -ribofuranose (4) and potassium tert-butoxide
ongoing. The original synthesis of clofarabine,2b-c now over
D
1
0 years old, involved chromatography and was not suf-
in a mixture of three solvents. Following crystallization, ano-
meric ratios exceeding 50 (â/r) are achieved. Deprotection and
additional crystallization afford a clofarabine drug substance
containing less than 0.1% of the r-anomer.
ficiently cost-effective and scalable to meet our needs.
Therefore, we sought to develop a new process, which would
avoid chromatography, produce drug substance of high purity
based on modern analytical methods, and provide clear cost
advantages over the published synthesis.
Results and Discussion
Bromosugar 4, which is readily prepared as the 1-R
anomer with HBr from commercially available 2-deoxy-2-
Introduction
Adenosine-related antimetabolites, such as cladribine (2-
CDA) 1 and fludarabine (2-F-araA) 2 have proven to be
useful chemotherapeutic agents for the treatment of various
leukemias, including those which have become resistant to
4
â-fluoro-1,3,5-tri-O-benzoyl-1-R-D-arabinofuranose, has been
used to synthesize various nucleoside analogues.5 The
anomeric selectivity observed in these reactions is dependent
upon solvent polarity. Solvents with lower dielectric con-
stants favor the formation of the â-anomer in the condensa-
tion products. This is probably due to the suppression of
1
alkylating agents (Figure 1). Clofarabine 3, a “second
generation” drug in this class, is protected against adenosine
deaminase-catalyzed deamination by the 2-chloro substituent
on the adenine ring and rendered more hydrolytically stable
at the anomeric position by the 2′-fluoro substituent.
Inhibition of DNA repair by clofarabine in leukemic
2
N
dissociative (S 1-type) mechanisms, which can lead to both
â- or R-anomeric products (Scheme 1). The effect of solvent
polarity on the anomeric outcome of alkylations with 4 has
3
lymphocytes has also recently been reported. Clinical trials
for clofarabine against a variety of cancers are currently
6
been studied by Howell and co-workers. In the Howell
study, where halogenated solvents were compared, it was
found that higher anomeric beta-selectivity was associated
with solvents of lower dielectric constant, which would
disfavor the less selective S 1 mechanism. Notably, this
N
strategy obviates the use of Lewis acids.
*
To whom correspondence should be addressed. E-mail: wbauta@ilexonc.com.
ILEX Products, Inc.
Ash Stevens, Inc.
†
‡
§
Current addresses: Mr. Burke, Eli Lilly, Inc.; Mr. Schulmeier and Mr.
Puente, Cerilliant, Inc.
7
|
Dedicated to Dr. Arthur F. Lewis on the occasion of his birthday.
In the earlier literature synthesis of clofarabine by
Montgomery and co-workers, 2,6-dichloropurine was used
(1) For examples see: (a) Keating, M. K.; O’Brien, S.; Lerner, S.; Koller, C.;
Beran, M.; Robertson, L. E.; Freireich, E. J.; Estey, E.; Kantarjian, H. Blood
1
998, 92, 1165-1171. (b) Rai, K. R.; Peterson, B. L.; Appelbaum, F. R.;
Kolitz, J.; Elias, L.; Shepherd, L.; Hines, J.; Threatte, G. A.; Larson, R. A.;
Cheson, B. D.; Schiffer, C. A. N. Engl. J. Med. 2000, 343, 1750-1757. (c)
Zinzani, P. L.; Bendani, M.; Magagnoli, M.; Albertini, P.; Rondelli, D.;
Stefoni, V.; Tani, M.; Tura, S. Hematologica 2000, 85, 1135-1139. (d)
McLauhlin, P.; Hagemeister, F. B.; Romaguera, J. E.; Sarris, A. H.; Pate,
O.; Younes, A.; Swan, F.; Keating, M.; Cabanillas, F. J. Clin. Oncol. 1996,
(3) (a) Yamauchi, T.; Nowak, B. J.; Keating, M. J.; Plunkett, W. Clin. Cancer
Res. 2001, 7, 3580-3589. (b) Kantargian H. M.; Gandhi, V. V.; O’Brien,
S.; Giles, F.; Coertes, J.; Kozuch, P.; Du, M.; Plunkett, W.; Rios, M. B.;
Freireich, E. J.; Estey, E. H.; Keating, M. J. Blood 2001, 98, 214b Abstract
4568.
(4) Tann, C. H.; Brodfuehrer, P. R.; Brundidge, S. P.; Sapino, S.; Howell, H.
G. J. Org. Chem. 1985, 50, 3644-3647.
1
4, 1262-1268.
(
2) (a) Parker, W. B.; Shaddix, S. C.; Chang, C.-H.; White, E. L.; Rose, L. M.;
Brockman, R. W.; Shortnacy, A. T.; Montgomery, J. A.; Secrist, J. A., III;
Bennet, L. L., Jr. Cancer Res. 1991, 51, 2386-2394. (b) Montgomery, J.
A.; Shortnacy-Fowler, A, T.; Clayton, S. D.; Riordan, J. M.; Secrist, J. A.,
III. J. Med. Chem. 1992, 35, 399-401. (c) A newer process, also from
(5) (a) Montgomery, J. A.; Shortnacy, A. T.; Carson, D. A.; Secrist, J. A., III.
J. Med. Chem. 1986, 29, 2389-2392. (b) Brundidge, S. P.; Howell, H. G.;
Sapino, C., Jr.; Chou, H.-T. U.S. Patent 4,879,377, 1989. (c) Vermishetti,
P.; Howell, H. G.; Walker, D. G.; Brodfuehrer, P. R.; Shih, K. M. U.S.
Patent 435,853, 1990.
(6) Howell, H. G.; Brodfuehrer, P. R.; Brundidge, S. P.; Benigni, D. A.; Sapino,
C., Jr. J. Org. Chem. 1988, 53, 85-88.
2
,6-dichloropurine, has been reported: Montgomery, J. A.; Fowler, A. T.;
Secrist, J. A., III. Patent WO 01/60383 A1 (2001). (d) Carson, D. A.;
Wasson, D. B.; Esparza, L. M.; Carrera, C. J.; Kipps, T. J.; Cottam, H. B.
Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 2970-2974.
(7) Vorbruggen, H.; Ruh-Pohlenz, C. Handbook of Nucleoside Synthesis; Wiley
Inter-Science: NY, 2001.
1
0.1021/op049884n CCC: $27.50 © 2004 American Chemical Society
Vol. 8, No. 6, 2004 / Organic Process Research & Development
•
889
Published on Web 09/24/2004

Scheme 1
Table 2. Screen of lower polarity solvents
conv ratio
entry
base
solvent additivea
time
(%)
5/6
1
2
3
4
5
6
7
KOt-Bu THFb
KOt-Bu t-BuOH
none
none
none
none
18 hours
9 days
9 days
9 days
54
62
14
13
83
57
78
5.3
8.8
8.7
0.8
1.0
9.6
0.8
c
Cs
2
CO
3
t-BuOH
NaOt-Bu t-BuOH
KOt-Bu t-BuOH
KOt-Bu t-BuOH
NaOt-Bu t-BuOH
TBABd 9 days
CaH
TBAB
2
9 days
9 days
a
Additive (1 equiv) was used. Dielectric constant ) 7.58.12 Dielectric
b
c
1
2 d
constant ) 12.47.
Tetrabutylammonium bromide.
be counterproductive to our goal of achieving high anomeric
selectivity, as outlined in Scheme 1. Therefore, we concluded
that heterogeneous mixtures would have to be used in order
to attain both reactivity and anomeric selectivity.
The reactions were run in 5 mL vials at ambient
temperature. We analyzed each reaction by HPLC to
determine the overall conversion and anomeric ratio. In all
cases, the reactions were stirred as heterogeneous mixtures.
The variables studied were solvent, base, and additives. Table
1 summarizes the effectiveness of various bases and additives
in acetonitrile.
Table 1. Base and additive screening reactions of
bromosugar 4 with 2-chloroadenine in MeCN
a,c
convb
ratio
entry
base
NaH
additive
(%)
5/6
86d
1
2
3
4
5
6
7
8
9
1
1
1
1
none
none
none
none
none
none
none
1.3
0.9
0.9
1.2
2.0
1.3
2.5
4.2
2.0
2.8
1.8
1.3
3.8
DBU
TMG
66
52
52
70
89
68
90
88
64
84
49
89
e
K CO
2
3
Acetonitrile (MeCN) is a solvent commonly used, in
conjunction with bases such as sodium hydride, in the
Cs CO
2
3
NaOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
NaOt-Bu
NaOt-Bu
7
coupling steps of nucleoside syntheses. Sodium hydride gave
CaH
CsBr
2
only a modest preference for the beta anomer 5 (5/6 ) 1.3),
whereas the amidine and guanidine bases (DBU and N,N,N′,N′-
tetramethylguanidine (TMG)) afforded the alpha anomer 6
2
0
1
2
3
Cs
TBAB
TBAB
2
COf
3
2
c
in preference to 5 (entries 1-3, Table 1). Potassium
CaH
2
10
carbonate gave an anomeric ratio similar to sodium hydride
1
0b
(5/6 ) 1.2), while cesium carbonate
gave a slightly
a
Reactions run in MeCN for 18 h, except entry 12, which was run for 16 h.
b
improved ratio (2) and higher overall conversion (70%)
(entries 4-5, Table 1). A marked increase in the conversion
was observed with sodium tert-butoxide (89%), although the
anomeric ratio remained similar to sodium hydride (1.3). A
higher ratio (2.5) was observed with potassium tert-butoxide
2
-Chloroadenine (1 equiv) was used. Conversion ) (area 5 + 6)/(area 2-Cl-
c
1
2
d
adenine + 5 + 6)*100. MeCN dielectric constant ) 35.94.
The percent
conversion could not be determined because 2-chloroadenine could not be
integrated properly. e N,N,N′,N′-Tetramethylguanidine. f Tetrabutylammonium
bromide.
as the nucleophile in the condensation step, and the chlorine
at the 6-position was selectively displaced with ammonia
1
1
(entry 7, Table 1).
A number of additives were also explored in acetonitrile
2
b,5a
under pressure.
For safety and practicality reasons, we
(Table 1, entries 8-13). The best results were observed with
chose to have the amino group incorporated as part of the
purine nucleophile. Examination of the literature revealed
a mixture of calcium hydride (we previously found that
calcium hydride alone was a poor base for the reaction) and
potassium tert-butoxide (5/6 ) 4.2, entry 7, Table 2). The
calcium hydride had a beneficial effect by removing trace
amounts of water from the solvent, as demonstrated in
that 2-chloroadenine has been selectively N-alkylated at N
9
8
with an allylic tosylate. Therefore, we concluded that the
amino group at C did not require a protecting group.
6
An array of parallel reactions between 2-chloroadenine
and bromosugar 4, affording anomeric benzoylated nucleo-
sides 5 and 6, was undertaken to evaluate various solvent,
additive, and base candidates (Scheme 2). We reasoned that
all but the most polar solvents would result in heterogeneous
reaction mixtures due to the limited solubility of 2-chloro-
adenine. However, the use of highly polar solvents would
(
10) Potassium carbonate and cesium carbonate have been used in various
nucleoside syntheses with variable anomeric selectivities. See: (a) Seela,
F.; Gumbiowski, R. HelV. Chim. Acta 1991, 74, 1048. (b) Reitz, A. B.;
Rebarchak, M. C. Nucleosides Nucleotides 1992, 11, 1115. (c) Seela, F.;
Bourgeois, W. Synthesis 1989, 912. (d) Seela, F.; Bindig, U.; Driller, H.;
Kaiser, K.; Kehne, A.; Rosemeyer, H.; Steker, H. Nucleosides Nucleotides
9
1987, 6, 11. (e) Rosemeyer, H.; Seela, F. J. Org. Chem. 1987, 52, 5136. (f)
Wang, Y.; Fleet, G. W. J.; Wilson, F. X.; Storer, R.; Myers, P. L.; Wallis,
C. J.; Doherty, O.; Watkin, D. J.; Vogt, K.; Witty, D. R.; Peach, J. M.
Tetrahedron Lett. 1991, 32, 1675. (g) Seela, F.; Gumbiowski, R. Liebigs
Ann. Chem. 1992, 679.
(
8) Obara, T.; Shuto, S.; Yasuyoshi, S.; Snoeck, R.; Andrei, G. et al. J. Med.
Chem. 1996, 39, 3847-3852.
1
(
9) Structural assignments for 5 and 6 were confirmed by 2D H NMR COSY
(11) Potassium tert-butoxide was reported to be an effective base for the
condensation of purines with 2-deoxy-2,2-difluoro-D-ribofuranosyl-3,5-
dibenzoate. See: Chou, T.-S.; Jones, C. D. U.S. Patent 5,821,357, 1993.
(12) Reichardt, C. SolVents and SolVent Effects in Organic Chemistry, 2nd ed.;
VCH: Weinheim, 1988.
and NOESY experiments and by the conversion of 5 to authentic clofarabine.
See Supporting Information for details. The authors acknowledge Professor
Judith Walmsley (University of Texas, San Antonio) for 2D NMR
experiments.
890
•
Vol. 8, No. 6, 2004 / Organic Process Research & Development

Scheme 2
Table 3. Screen of binary solvent mixturesa
Table 4. Screen of hydroxide bases
solvent
ratio
time conv ratio
(h) (%) 5/6
solvent
time conv ratio
entry
solvent
additive
entry base
solvent
ratio additive (h) (%) 5/6
1
2
3
4
5
6
MeCN/t-BuOHb
MeCN/t-BuOH
MeCN/t-BuOH
MeCN/t-BuOH
MeCN/t-BuOH
MeCN/t-BuOH
MeCN/t-BuOH
1:1 none
1:1 CaH
9 days 83 10.3
1
2
3
NaOH MeCN/t-BuOH
KOH MeCN/t-BuOH
KOH MeCN/t-BuOH
1:1
1:1
1:1
none
none
CaH
2
(1 equiv)
16
42
42
60
1.2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
18
60
60
60
60
16
18
18
60
60
60
60
60
14
14
14
14
24
24
24
24
24
24
24
24
24
24
83 13.2
89 5.5
88 9.5
77 11.6
69 12.7
32 1.0
68 11.0
46 14.7
9:1 CaH
7:3 CaH
3:7 CaH
1:9 CaH
1:1 CaH
1:1 CaH
1:1 CaH
1:1 CaH
19:9 CaH
2:5 CaH
1:1 CaH
2:5 CaH
1:2 CaH
1:2 CaH
1:2 CaH
1:4 CaH
1:2 CaH
1:2 KBr
1:2 KCl
1:2 CuI
4
KOH MeCN/
1:2
CaH
2
25
79 14.2
tert-amylOH
(2 equiv)
c
7
d
8
9
MeCN/DCM
MeCN/PhMe
8
15.2
f
NRg NR
81 11.5
71 8.7
67 8.2
69 8.1
67 8.1
81 9.1
53 16.6
60 18.0
58 18.9
83 12.5
81 12.7
81 12.5
sodium tert-butoxide and the addition of TBAB in combina-
tion with either sodium or potassium tert-butoxide were
deleterious to anomeric selectivity. However, the selectivity
observed with cesium carbonate was better in tert-butanol
than in acetonitrile.
The preliminary data strongly indicated that potassium
tert-butoxide was the preferred base and that a move to
solvents of lower polarity improved anomeric ratio, while
sacrificing percent conversion. We also observed that potas-
sium cation had a beneficial effect on anomeric stereo-
selectivity, while sodium and amine-based cations signifi-
cantly decreased anomeric stereoselectivity.
To reconcile the conflicting requirements for both conver-
sion and anomeric ratio, we next examined tert-butoxide base
in mixtures of two solvents (Table 3). The solvents tested
were acetonitrile, tert-butyl alcohol, tert-amyl alcohol,
dichloromethane (DCM), 1,2-dichloroethane (DCE), THF,
and toluene (PhMe).
In entries 1-6 (Table 3), we see the expected trend in
anomeric ratio and conversion as the reaction is run with
potassium tert-butoxide in varying ratios of acetonitrile and
tert-butyl alcohol. Dichloromethane and toluene both gave
exceedingly poor conversion with acetonitrile and potassium
tert-butoxide (entries 8 and 9, Table 3). It was also not
surprising that sodium tert-butoxide (entry 7, Table 3) gave
essentially no selectivity in a mixture of acetonitrile and tert-
butanol (entry 6, Table 3). Of the other solvent mixtures,
the highest anomeric ratios (in some cases exceeding 20)
were observed with potassium tert-butoxide in DCE and
either tert-butanol or tert-amyl alcohol (entries 16-18, Table
3). Conversions in these cases were in the 50-60% range.
Various other additives were also studied (entries 19-28,
Table 3). Both KBr and KCl had a surprisingly beneficial
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
MeCN/tert-amylOH
THF/t-BuOH
THF/t-BuOH
THF/t-BuOH
THF/t-BuOH
h
MeCN/DCE
DCE/t-BuOH
DCE/tert-amylOH
DCE/tert-amylOH
MeCN/tert-amylOH
MeCN/tert-amylOH
MeCN/tert-amylOH
MeCN/tert-amylOH
MeCN/tert-amylOH
MeCN/tert-amylOH
MeCN/tert-amylOH
MeCN/tert-amylOH
MeCN/tert-amylOH
MeCN/tert-amylOH
3
1.2
1:2 ZnBr
1:2 LiBr
1:2 NaI
1:2 CsI
1:2 MgBr
1:2 Cu(OAc)
2
NRe NA
21 1.4
75 0.7
85 1.7
2
0
1
NA
5 only
2
a
b
Reactions were run with KOt-Bu unless otherwise indicated. t-BuOH
1
2
c
dielectric constant ) 12.47, MeCN dielectric constant ) 35.94.
NaOt-Bu
was used. d Dichloromethane dielectric constant ) 8.93.
12
12
e
None of the alpha
No reaction. 1,2-
f
g
h
anomer was detected. Toluene dielectric constant ) 2.38.
1
2
Dichloroethane dielectric constant ) 10.37.
experiments where water was deliberately added to reactions
with and without CaH . Not surprisingly, calcium hydride
2
13
increased the anomeric ratio when combined with both
sodium and potassium tert-butoxide, although the results for
potassium were still superior. The conversions in reactions
with calcium hydride were also significantly improved.
In keeping with the idea that solvents of lower polarity
would suppress the R-anomer formation, we investigated
some reactions in THF and tert-butanol. The results are
summarized in Table 2. Despite the higher dielectric constant
of tert-butanol compared to THF, the former solvent gave
better anomeric selectivity. As observed in acetonitrile,
(
13) In parallel experiments where 1% (v/v) water had been deliberately added
to the solvent, the anomeric ratio increased by 36% when CaH was used.
and no added water, the anomeric ratio
(15) (a) Piskala, A.; Masojidkova, M.; Saman, D. Collect. Czech Chem. Commun.
1996, 61, S23-S25. (b) Seela, F.; Bourgeois, W. Synthesis 1990, 10, 945-
950. (c) Dziewiszek, K.; Schinazi, R. F.; Chao, T. C.; Su, T. L.; Dzik, J.
M.; Rode, W.; Watanabe, K. A. Nucleosides Nucleotides 1994, 13, 77-94.
(d) Harayama, T.; Yanada, R.; Taga, T.; Machida, K.; Cadet, J.; Yoneda,
F. J. Chem. Soc., Chem. Commun. 1986, 19, 1469-1471. (e) Chavis, C.;
Dumont, F.; Wightman, R. H.; Ziegler, J. C.; Imbach, J. L. J. Org. Chem.
1982, 47, 202-206.
2
In a comparable reaction with CaH
increased by 113%.
2
(
14) An anomeric ratio of 10 (â/R) has been reported in the reaction of 4 with
N
6
-benzoyl adenine has been reported in MeCN/CH Cl using sodium
2
2
hydride. See: Vemishetti, P.; Howell, H. G.; Walker, D. G.; Brodfuehrer,
P. R.; Shih, K.-M. Eur. Pat. EP0428109, 1991.
Vol. 8, No. 6, 2004 / Organic Process Research & Development
•
891

Table 5. Screen of ternary solvent mixtures with calcium hydride additive
solvent
ratio
time
(h)
conv
(%)
ratio
5/6
entry
base
solvent
1
2
3
4
5
6
7
8
9
0
1
2
3
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOt-Bu
KOtert-amyl
MeCN/t-BuOH/DCE
MeCN/t-BuOH/DCE
MeCN/t-BuOH/DCE
MeCN/t-BuOH/DCE
MeCN/tert-amylOH/DCM
MeCN/tert-amylOH/DCM
MeCN/tert-amylOH/DCM
MeCN/t-BuOH/tert-amylOH
MeCN/t-BuOH/tert-amylOH
MeCN/DCE/tert-amylOH
MeCN/t-BuOH/tert-amylOH
MeCN/t-BuOH/tert-amylOH
MeCN/tert-amylOH/PhMe
1:9:5
1:11:2
6:6:5
11:1:3
1:2:3
1:3:3
1:2:1
15:2:13
9:2:19
1:2:2
1:3:2
1:2:2
1:2:2
60
60
60
60
10
10
10
60
60
14
20
20
10
55
65
66
32
63
69
83
84
80
74
67
72
68
16.9
14.9
11.7
4.4
8.5
13.2
17.7
11.2
14.0
21.4
18.6
17.8
12.7
1
1
1
1
effect on anomeric selectivity, somewhat comparable in
magnitude to calcium hydride (compare entries 20-21 with
entry 10, Table 3). However, some metal salts (Cu, Na, Li,
Zn, and Mg) had detrimental effects on the reaction (entries
Table 6. Bench scale reactions between 4 and
2-chloroadenine to afford 5 and 6
scale
ratio ratio HPLC
(g of
time slurry yield 5/6
5/6
assay
entry 2-chloroadenine) (h) (°C) (%) crudea final area (%)
b
22-28, Table 3). The reaction did not occur at all with
magnesium or zinc bromide and only to a very slight extent
in the presence of copper salts. Under the same solvent
conditions, cesium iodide gave high conversion but only a
1
2
3
4
5
6
7
8
9
0
22.9
28.8
28.8
29.2
29.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
18.0
18.6
18.4
6.0 reflux 53
5.0 reflux 55
27.9
26.3
81.5
58.8
37.1
93.8
95.9
98.6
98.1
96.8
97.5
96.1
96.7
97.3
96.6
97.9
97.9
97.2
96.5
95.3
17.5 reflux 54c 21.6
5.0 reflux 13
25.9 163.6
1
.7 anomeric ratio. The benefits of potassium as a counterion
21.0 reflux 32c
28.3
22.9
19.3
17.4
14.6
21.3
21.2 116.5
21.9 102.5
19.0
23.6
20.4
94.9
53.1
39.6
51.4
80.8
39.3
12
are borne out by these data.
21.5 22.5
20.5 25.2
19.0 23.9
19.0 27.1
19.2 22.3
48
52
49
50
59
Another interesting observation is that sodium and potas-
sium hydroxide can be used as bases in acetonitrile/tert-butyl
alcohol or acetonitrile/tert-amyl alcohol (Table 4). The use
1
7
of hydroxide bases in nucleoside syntheses is precedented
11
19 reflux 41
19 reflux 49
1
1
1
2
3
4
and has been accomplished with phase transfer catalysis in
polar solvents.
Potassium hydroxide was superior to sodium hydroxide
in promoting selectivity (entries 1-2, Table 4). The addition
of the drying agent calcium hydride was also beneficial
22 26
17.5 23.2
19 21.9
45
46
48
87.5
55.2
39.0
15
a
b
Represents ratio after the n-BuOAc/heptane precipitation. Represents ratio
after one MeOH slurry. c Partial deprotection observed during reslurry.
(entries 3-4, Table 4). These results are interesting in part
because the base generates at least 1 equiv of water during
the course of the reaction. Both tert-butanol and tert-amyl
alcohol were suitable cosolvents.
Mixtures of three solvents were explored next, all in the
2
presence of CaH (Table 5). As expected, higher proportions
of acetonitrile generally led to higher conversions at the
expense of anomeric ratio (for example, compare entries 2
and 4). Smaller proportions of acetonitrile gave generally
better results in combination with tertiary alcohols (entries
(5/6) improved without resorting to chromatography. The
problem was solved by filtering most of the 2-chloroadenine;
then the solvent was exchanged for n-butyl acetate, which
effectively dissolved the product mixture. Addition of
heptane precipitated crude 5, giving a slight increase in
anomeric ratio and allowed for ready manipulation of the
solid. We were unable to find satisfactory recrystallization
conditions for 5 and eventually opted for a slurry procedure
in methanol, which could be conducted either at reflux or at
room temperature (see below). Anomeric ratios of >30:1
(â/R) were routinely achieved for isolated 5. Anomeric ratio
could be further improved by secondary methanol slurry,
after which anomeric ratios >50:1 (â/R) were routinely
observed.
An important consideration in performing the methanol
slurry procedure was the presence of adventitious base. For
example, in entry 4 (Table 6), only 13% yield of protected
clofarabine was isolated. The remaining mass balance was
clofarabine and the corresponding O5′-benzoate 7 (Figure 2).
To avoid the formation of 7, the pH of reaction mixtures
was routinely checked and adjusted to pH 5-7 with acetic
acid prior to the precipitation step. Compound 7 was purified
by column chromatography and characterized. The structure
8, 9, 11, and 12) or, better yet, with a combination of tertiary
alcohol and a halogenated solvent (entries 7 and 10).
Through a series of gram-scale experiments, the reaction
parameters were further defined (Table 6). The solvent
volume ratios were slightly adjusted to 1:2:1 MeCN/tert-
amyl alcohol/DCE and 50 °C was chosen as the reaction
temperature. Good conversions were observed at 20 g scale
after 5 h but the reaction could also be run for longer times
without affecting the purity profile or yield.
The isolation and purification of benzoylated clofarabine
5
from the heterogeneous reaction mixture presented a
number of challenges. First, the rather insoluble 2-chloro-
adenine had to be removed along with inorganic byproducts,
then intermediate 5 had to be isolated and its anomeric ratio
892
•
Vol. 8, No. 6, 2004 / Organic Process Research & Development

Table 8. Reactions of 5 with NaOMe to afford 3
anomer
weight
scale
entry (g of 5)
ratio
NaOMe yield 3 HPLCa assayb
(5/6)
(mol %)
(%)
(%)
(%)
1
2
3
4
5
20.0
25.7
10.0
23.4
8.0
49
38.5
87
80
40
4
5
27
11
25
61
64
61
60
68
99.3
99.6
99.7
99.2
99.7
98.5
NDc
98.9
ND
Figure 2.
98.7
Table 7. Comparison of OH resonances for 7 and
a
This is the area assay for 3. b This is the HPLC weight assay for 3 based on
clofarabine (DMSO-d
6
)
a reference standard. The result is not corrected for water or solvent content.
Not determined.
c
3
′-OH
ppm
(DMSO-d
5′-OH
ppm
(DMSO-d )
6
compound
6
)
reaction temperatures are possible under these conditions. It
is also noteworthy that the anomeric ratio (5/6) can be as
low as 30 and still produce anomerically pure clofarabine
clofarabine
7
5.95 (d, 1H, J ) 5)
6.19 (d, 1H, J ) 5)
5.08 (t, 1H, J ) 5)
not observed
(3).
Scheme 3
Conclusions
We have demonstrated the use of 2-chloroadenine and
bromosugar 4, in conjunction with potassium tertiary al-
koxide base and a three solvent system, as a selective route
to protected clofarabine (5) and hence clofarabine (3), after
debenzoylation. The superiority of potassium as a counterion
in this chemistry was clear from the screening experiments.
Significantly, the coupling reaction was run in a heteroge-
neous manner. It was also concluded that the addition of
calcium hydride as a drying agent was beneficial to selectiv-
ity.
Experimental Section
2
Reactions were run under N atmosphere. Melting points
was assigned by NMR on the basis of the 3′-hydroxyl doublet
see Table 7). The relative chemical shift and coupling
constant for this resonance correlate well with the corre-
sponding clofarabine resonance. As expected, the 5′-hydroxyl
triplet is absent.
were obtained on a Buchi B545 melting point apparatus and
are uncorrected. NMR spectra were obtained on a Brucker
250, 400, or 500 MHz Varian instrument. IR spectra were
obtained using KBr plates on a Matson Infinity Gold FTIR
instrument. UV spectra were obtained using a Beckman
DU640 spectrophotometer. HPLC data were collected on
either of two Waters instruments: Waters System 600 Dual
Pump Controller or Waters 515 HPLC pump with Waters
996 Photodiode Array detector running Millenium software.
HPLC methods are described in the Supporting Information.
2-Deoxy-1-r-bromo-2-â-fluoro-3,5-di-O-benzoyl-d-ri-
bofuranose (4). Procedure A. A three-neck round-bottom
flask (3 L) was equipped with a stir bar and nitrogen inlet
adapter and charged with dichloromethane (1.44 L) and
commercially available 2-deoxy-2-â-fluoro-1,3,5-tri-O-ben-
zoyl-1-R-D-ribofuranose (354.3 g, 0.763 mol) at room
temperature to give a solution. HBr (33% in acetic acid) (279
mL, 1.62 mol) was charged (the solution changed from
colorless to golden yellow to orange), the nitrogen line was
removed to ensure that HBr was not lost, and the resultant
mixture was stirred at ambient temperature overnight. The
acid was neutralized with saturated sodium bicarbonate
solution (aq.) (5 × 800 mL portions) using a separatory
funnel, whereupon the pH of the aqueous layer was 7-8.
(
Montgomery and co-workers also reported the formation
of 7 as a byproduct during the aminolysis of 2,6-dichloro-
9
-(3′-acetyl-5′-O-benzoyl-2′-deoxy-2′-fluoro-R-D-arabino-
2b
furanosyl)-9H-purine. Based on our work, it cannot be
determined whether 7 was formed as a kinetic or thermo-
dynamic product. The selective formation of 7 might be
attributed to induction by the vicinal fluorine atom, but
additional work would be required to resolve this question.
The deprotection reaction of 5 to afford clofarabine was
accomplished with catalytic sodium methoxide in methanol
(Scheme 3). The product was crystallized directly from the
reaction mixture by cooling and then recrystallized from
methanol to afford clofarabine. The product thus obtained
contained <0.1% of the corresponding anomer 8 (Table 8).
An authentic sample of clofarabine R-anomer 8 was obtained
by deprotection of 6 under similar conditions. Stereochemical
assignments were confirmed by the COSY and NOESY
spectra (see Supporting Information).
As can be seen from Table 8, methoxide charges of as
little as 4 mol % can be used to carry out the reaction.
However, later work revealed that 20-25 mol % was
preferable on a larger scale, since faster rates and lower
4
The organic phase was separated, dried over MgSO (150
g), and filtered. An HPLC chromatogram revealed that all
the benzoic acid byproduct had been removed. Rotary
Vol. 8, No. 6, 2004 / Organic Process Research & Development
•
893

evaporation of the solvents, followed by pumping under high
vacuum, afforded 4 (313.4 g, 0.740 mol) as a viscous yellow
gum. This material was stored under nitrogen and used
without further purification. The NMR spectrum was con-
mL) was added, followed by potassium tert-butoxide (13.9
g, 0.12 mmol) and CaH (5.0 g, 0.12 mol). The mixture was
2
heated to 48 to 53 °C for 40 min. A 1.1 M solution of
bromosugar (4) (110 mL, 0.12 mol) in 1,2-dichloroethane
(DCE) was added, followed by DCE (24 mL), and the
mixture was heated at 48 to 53 °C for 19 h. In-process
analysis revealed an anomeric ratio of 15 (â/R). After cooling
to 27 °C, DCE (300 mL) was charged, the resultant mixture
was suction filtered through a pad of Celite, and the filter
cake was rinsed with DCE (80 mL). The pH of the filtrate
was measured with wet pH paper and found to be 6. Rotary
evaporation of the filtrate afforded 79.7 g of crude solid.
This was dissolved in butyl acetate (150 mL). Heptane (750
mL) was added to the butyl acetate solution over 1 h, and
the resultant suspension was cooled to 5 to -5 °C for 1 h
and filtered and the filter cake was washed with a mixture
of 5:1 (v:v) heptane/butyl acetate (160 mL) and then with
heptane (100 mL) to afford 40.7 g of crude 5 with an
anomeric ratio of 15 (â/R). A portion (40 g) of this material
was mixed with methanol (400 mL), and the resultant slurry
was stirred (approximately 215 rpm) at 20 to 30 °C for 16
h. The mixture was cooled to 0 to -10 °C, stirred for 60
min, and filtered. The filter cake was washed with methanol
(160 mL) and heptane (164 mL). After vacuum-drying, 5
(30.4 g, 0.059 mol) was obtained in 50% yield with an
anomeric ratio of 80.1 (â/R). This material could be used in
the subsequent step without further purification.
4
sistent with that reported in the literature. Although 4
crystallizes slowly from concentrated dichloromethane solu-
tion, it is possible to obtain the compound as a solid through
seeding.
Procedure B. A 1000-mL flask was charged with
2
(
2
-deoxy-2-â-fluoro-1,3,5-tri-O-benzoyl-1-R-D-ribofuranose
45.3, g, 97.55 mmol), HBr (33 wt % in acetic acid, 35 mL,
04.9 mmol), and CH Cl (200 mL). The flask was capped
2
2
with a rubber septum to prevent escape of HBr. The reaction
mixture was stirred at ambient temperature for 18.5 h. HPLC
analysis showed no starting material remaining. The reaction
mixture was poured into saturated NaHCO (600 mL) with
3
stirring. Stirring was stopped, and the layers were allowed
to separate. The top aqueous layer was discarded. The bottom
organic layer was washed with saturated NaHCO
mL). The aqueous layer pH was 8-8.5 (pH paper). The
organic layer was dried with MgSO and NaHCO . The
3
(aq.) (100
4
3
organic layer was filtered, and the filtrate was combined with
heptane (350 mL). The volume was reduced by rotary
evaporation to remove CH Cl . The resulting emulsion was
2 2
seeded with authentic crystals of 4 while stirring. Crystals
started to form within 1 min. Solids stuck to the sides of the
flask. The solids were scraped from the sides of the flask
and the suspension was stirred at ambient temperature for
Chromatographic Preparation of a Reference Stand-
ard for 5. Benzoylated clofarabine 5 (54.5 g, 0.11 mol) was
mixed with EtOAc (230 mL) and stirred overnight. The
undissolved solid (18.2 g) was removed by filtration, and
the filtrate was chromatographed on a silica gel column (13
45 min. The mixture was filtered, and the flask and filter
cake were washed with the mother liquor. The wet solid was
placed in a vacuum oven (40 °C, 27 inHg) for 3.5 h. The
white solid weighed 36.5 g (88.5% yield).
1
2
Mp ) 73 °C. H NMR 400 MHz (CDCl
3
) δ 4.69-4.86
), 5.60 (d, 1H, J
), 7.04-7.63 (m, 6H,
× 51 cm ) with 2:1 EtOAc/heptane eluent. The TLC on silica
(
)
m, 3H, H4′, H5′), 5.54 (dm, 1H, J ) 18, H
48, H ), 6.64 (d, 1H, J ) 10, H
aromatic H), 8.04-8.13 (m, 4H, aromatic H) ppm. C NMR
3
f
gel (1:1 EtOAc/heptane) gave R values of 0.21 and 0.10
2
1
for 5 and 6, respectively. Fractions (500 mL) were collected
and analyzed by HPLC. Fractions containing >99.9% 5 were
13
1
00 MHz (CDCl
3
) 166.0 (CO), 165.5 (CO), 133.9 (para),
pooled and evaporated to afford pure 5 (22 g, 0.043 mol).
1
133.2 (para), 130.0 (ortho), 129.8 (ortho), 129.4 (ipso), 128.6
Mp ) 159-162 °C. H NMR 250 MHz (CDCl
3
) δ 4.53-
(
meta), 128.5 (ipso), 128.4 (meta), 100.6 (d, JCF ) 191, C
2
),
),
4.57 (m, 1H, H4′), 4.80 (d, 1H, J ) 4, H5′), 5.60 (dd, 1H, J
) 50, J ) 3, H2′), 5.74 (dd, 1H, J ) 15, J ) 3, H3′), 6.48 (s,
8
6
1
7.5 (d, JCF ) 30, C
1
), 84.7 (C
) ppm. F NMR (CDCl
2 Hz) ppm. COSY and NOESY spectra were consistent
4
), 76.2 (d, JCF ) 33, C
) -167.2 (ddd, J ) 50, 22,
3
19
2.5 (C
5
3
2H, NH
2
), 6.56 (dd, 1H, J ) 23, J ) 3, H1′), 7.40-7.67 (m,
6H, aromatic H), 8.05-8.09 (m, 5H, aromatic H and H
8
)
,
1
3
with the assigned structure.
ppm. C NMR (CDCl
3
) 63 MHz δ 63.27 (C5′), 81.12 (C3′
Representative Experimental Procedure for a Parallel
Reaction Experiment (See Tables 1-5). A set of 5-mL
screw cap vials with stir bars was charged with 50 mg (0.29
mmol) of 2-chloroadenine, followed by the indicated base
and anhydrous MeCN (1.0 mL). A stock solution of 4 was
prepared by dissolving 1.25 g in 5.0 mL of anhydrous MeCN.
The stock solution of 4 (600 µL, 0.28 mmol) was added to
each vial at room temperature, and the mixture was stirred
for 2 days. The reactions were analyzed by HPLC.
C
4′), 83.42 (d, JC-F ) 17, C1′), 92.62 ( d, JC-F ) 192, C2′),
5
117.72 (C ), 128.04 (Ph), 128.68 (Ph), 129.28 (Ph), 129.69
(Ph), 129.95 (Ph), 133.36 (Ph), 134.14 (Ph), 140.17 (d, JC-F
) 7, C ), 150.47 (C ), 154.32 (C ), 156.33 (C ), 165.16 (CO),
166.12 (CO) ppm. IR (KBr) 3327m, 1725s, 1643s, 1594s,
1452m, 1352m, 1311m, 1270s, 1178m, 1109s, 1096s,
1070m, 1027m, 923w, 711m, 684w cm . UV (MeOH)
λmax
λmax
512 (MH ) (100), 513 (24), 514 (32), 515 (5) (characteristic
100:32:5 Cl isotope pattern observed for parent ion). Anal.
8
6
4
2
-
1
1
212 nm (ꢀ ) 33 608), λmax 232 nm (ꢀ ) 30 876),
2
3
262 nm (ꢀ ) 17 253). MS m/z (% rel. abundance)
+
6
-Amino-2-chloro-9-(2′-deoxy-2′-fluoro-3′,5′-di-O-ben-
zoyl-â-d-arabinofuranosyl)-9H-purine (5). A jacketed glass
-L reactor, equipped with mechanical stirrer, reflux con-
1
5 5
Calcd for C24H19ClFN O : C, 56.31; H, 3.74; N, 13.68; F,
denser, and temperature probe under stirring (189 rpm) was
charged with 2-chloroadenine (20 g, 0.12 mol) and aceto-
nitrile (100 mL) to give a suspension. tert-Amyl alcohol (200
3.71; Cl, 6.93. Found: C, 56.26; H, 3.78; N, 13.08; F, 3.57;
Cl, 6.72. COSY and NOESY spectra support the assigned
structure.
894
•
Vol. 8, No. 6, 2004 / Organic Process Research & Development

Chromatographic Preparation of a Reference Stan-
dard for 6-Amino-2-chloro-9-(2′-deoxy-2′-fluoro-3′,5′-di-
O-benzoyl-r-D-arabinofuranosyl)-9H-purine (6). A sample
of crude 5 (43.9 g, 0.086 mol) was stirred with EtOAc (250
mL) and filtered to remove a small amount of insoluble
material. The filtrate was chromatographed on a silica gel
heptane (784 mL) was added to generate a solid (33.0 g),
which was filtered and dried. The HPLC analysis of this
material revealed a mixture of clofarabine (3) and 7 (1:2.1).
The mixture was purified by flash chromatography on a silica
gel (2.08 kg) column, eluting with 3:1 (EtOAc/heptane)
(silica gel TLC R (7) ) 0.48 in 4:1 EtOAc/heptane).
f
column (13 × 54 cm), eluting with 1:1 EtOAc/heptane (R
f
Fractions were collected and analyzed by TLC and HPLC.
(
6) ) 0.10). Fractions (500 mL) were analyzed by HPLC,
The combined pure fractions afforded 12.64 g of 7.
1
and those containing >98% pure 6 were pooled and
evaporated to afford 6 (8.27 g, 0.016 mol). Combined
samples from several columns were pooled and further
purified by preparatory HPLC (Waters Nova-Pak Silica
Mp ) 196-198 °C. H NMR 250 MHz (CDCl
3
) δ 4.17-
4.23 (m, 1H, H4′), 4.53-4.67 (m, 3H, H3′, H5′), 5.29 (dt,
1H, J ) 53, J ) 5, H2′), 6.37 (dd, 1H, J ) 15, H1′), 6.19 (d,
1H, J ) 5, OH3′), 6.39 (dd, 1H, J ) 15, 4, H1′), 7.48-7.55
(m, 2H, aromatic H), 7.63-7.69 (m, 1H, aromatic H), 7.91
2
column, 6 µm, 19 × 300 mm ), eluting with EtOAc/MTBE/
heptane (1:1:1) containing 1 mL of Et
3
N per 250 mL of
(br s, 1H, NH
1H, J ) 2, H
(C5′), 73.7 (d, JC-F ) 23, C3′), 80.6 (d, J ) 6, C4′), 81.6 (d,
C-F ) 19, C1′), 95.0 (d, J ) 195, C2′), 114.68 (C ), 128.79
(Ph), 129.27 (Ph), 129.36 (Ph), 133.52 (Ph), 140.1 (C ),
150.19 (C ), 153.40 (C ), 156.85 (C ), 161.98 (CO) ppm.
O/MeOH) λmax 212 nm, λmax 233 nm, λmax 263
2
), 7.96-8.00 (m, 2H, aromatic H), 8.17 (d,
) ppm. C NMR 63 MHz (DMSO) δ 64.1
1
3
EtOAc. Pure fractions were pooled, evaporated, and dried
8
under high vacuum to afford 6 (3.28 g, 6.4 mmol).
1
Mp ) 101 to 104 °C. H NMR 250 MHz (CDCl
3
) δ
J
5
4
.63-4.75 (m, 2H, H5′), 4.93-4.96 (m, 1H, H4′), 5.79 (dm,
J ) 17, H3′), 6.16 (d, 1H, J ) 49, H2′), 6.43 (d, 1H, J ) 14,
1′), 6.73 (s br, 2H, NH ), 7.25-7.77 (m, 8H, aromatic H),
.00-8.09 (m, 3H, aromatic H and H
) ppm. 13C NMR 63
MHz (CDCl ) δ 63.31 (C5′), 77.20 (C1′), 83.97 (C4′), 89.24
d, JC-F ) 36, C3′), 96.50 (d, JC-F ) 188, C2′), 118.88 (C ),
28.04 (Ph), 128.42 (Ph), 128.53 (Ph), 129.28 (Ph), 129.63
),
), 164.97 (CO), 166.10
8
6
4
2
H
8
2
UV (H
nm.
2
1
2
3
8
3
6-Amino-2-chloro-9-(2′-deoxy-2′-fluoro-â-D-arabino-
furanosyl)-9H-purine (3). Clofarabine. A 1-L reactor
equipped with a condenser and overhead stirrer was charged
with protected clofarabine (5) (25.7 g, 84.8 mmol) and
methanol (154 mL) to give a slurry. Sodium methoxide
solution (30% w/w in MeOH) (0.20 mL, 1.05 mmol) was
added, and the reaction was stirred and heated to 33 °C for
7 h, at which point a clear yellow solution resulted. The
reaction was complete by HPLC analysis. The solution was
cooled to room temperature and neutralized with glacial
acetic acid (0.05 mL). The mixture was cooled to -10 °C
for 1 h, and the resultant white solid was suction filtered
and washed with -15 °C methanol (77 mL) to afford 15.2
g of wet product. A loss-on-drying analysis performed on a
small aliquot revealed 25.3% solvent content. The wet cake
was mixed with methanol (310 mL) and heated with stirring
to 63 °C, whereupon the solid dissolved. The solution was
then cooled to -8 to -12 °C and stirred at this temperature
for 1 h, and then the resultant crystals were vacuum filtered.
The cake was washed with cold (-15 °C) methanol (22 mL)
and dried under vacuum to afford clofarabine (3) (9.7 g,
(
1
5
(Ph), 129.76 (Ph), 133.31 (Ph), 133.90 (Ph), 138.92 (C
8
1
6 4 2
50.15 (C ), 154.36 (C ), 156.40 (C
(
1
CO) ppm. IR (KBr) 3343m, 1726s, 1642s, 1593s, 1453m,
345m, 1315m, 1269s, 1179m, 1096s, 1028w, 710s cm .
-
1
UV (MeOH) λmax
28 187), λmax
abundance) 217 (51), 279 (30), 512 (MH ) (100), 514 (42),
1
213 nm (ꢀ ) 32 399), λmax 231 nm (ꢀ
2
)
3
263 nm (ꢀ ) 15 108). MS m/z (% rel.
+
515 (6), 534 (4).
6-Amino-2-chloro-9-(2′-deoxy-2′-fluoro-5′-O-benzoyl-
â-D-arabinofuranosyl)-9H-purine (7). A three-neck 1-L
flask was charged with 2-chloroadenine (29.2 g, 172 mmol),
MeCN/tert-amyl alcohol (2:1) (v/v) (440 mL), potassium
2
tert-butoxide (20.3 g, 181 mmol), and CaH (7.24 g, 172
mmol). The mixture was heated for 30 min at 48-58 °C.
Bromosugar (4) (72.96 g, 172 mmol) was added as a solution
in DCE (146 mL). Stirring was continued for 25 h at
approximately 50 °C. The reaction was allowed to cool to
ambient temperature, and DCE (300 mL) was added. The
mixture was cooled to 0-5 °C for 2.5 h and filtered through
Celite, which was washed with DCE (2 × 20 mL). Rotary
evaporation of the filtrate gave a solid, which was dissolved
in butyl acetate (221 mL). Heptane (1550 mL) was then
added to the butyl acetate solution with vigorous stirring over
32.01 mmol) in 64% yield.
1
Mp ) 237 °C. H NMR 500 MHz (DMSO-d
6
) δ 8.27 (d,
1H, J ) 5, H
8
), 7.87 (br s, 2H, NH
2
), 6.32 (dd, J ) 14, J )
5, H1′), 5.95 (d, 1H, J ) 5, OH3′), 5.22 (dt, 1H, J ) 53, J )
5, H2′), 5.08 (t, 1H, J ) 6, OH5′), 4.43 (dm, 1H, J ) 20,
1
1
h. The resultant suspension was cooled in an ice bath for
h and filtered, and the filter cake was washed with butyl
H C
3′), 3.85 (m, 1H, H4′), 3.60-3.72 (m, 2H, H5′) ppm. 13
NMR 126 MHz (DMSO-d
3′), 81.44 (d, J ) 17, C1′), 83.50 (d, J ) 6, C4′), 95.33 (d,
J ) 194, C2′), 117.35 (C ), 140.00 (C ), 150.16 (C ), 153.26
(C ), 156.80 (C ) ppm. IR (KBr) 3330s, 1646s, 1595s,
1507w, 1466w, 1351m, 1307m, 1248w, 1215w, 1038m,
6
) δ 60.34 (C5′), 72.56 (d, J ) 24,
acetate/heptane (1:7) (v/v). Drying under vacuum afforded
7.5 g of white solid. A portion of the solid (57.4 g) was
C
6
5
8
6
suspended in MeOH (570 mL) and heated to reflux for 30
min. The resultant slurry was stirred at ambient temperature
for 14 h. Solids were filtered and washed with MeOH (15
mL) and dried under vacuum to afford 6 (11.4 g, 22 mmol)
with an anomeric ratio of 164:1 (â/R). Concentration of the
combined methanol filtrates afforded 44.7 g of white solid.
This solid was dissolved in butyl acetate (550 mL), and then
4
2
-
1
708w cm . UV (H
2
O) λmax
1
212 nm (ꢀ 22 500), λmax
2
263 nm (ꢀ 15 989). Mass spec. (electrospray) m/e (% rel.
abundance) 170 (32) (-2-chloroadenine), 300 (26), 302 (39),
+
+
303 (M ), 304 (100) (MH ), 305 (21), 306 (41). Elem. anal.
calcd for C10 : C, 39.55; H, 3.65; N, 23.06; F,
H
5 3
11ClFN O
Vol. 8, No. 6, 2004 / Organic Process Research & Development
•
895

1
6
.26; Cl, 11.67. Found: C, 39.24; H, 3.58; N, 22.98; F, 5.93;
Cl, 11.38. Opt. rot. [R] ) 39.93° (c ) 5 mg/mL in DMF).
-Amino-2-chloro-9-(2′-deoxy-2′-fluoro-r-d-arabino-
Mp ) 237 °C. H NMR 250 MHz (DMSO) δ 8.32 (s,
1H, H ), 7.90 (s br, 2H, NH ), 6.20 (dd, 1H, J ) 15, J ) 3,
H
D
8
2
6
1′), 6.02 (d, 1H, J ) 4, OH3′), 5.62 (dt, 1H, J ) 52, J ) 3,
2′), 5.01 (t, 1H, J ) 6, OH5′), 4.36 (dm, 1H, J ) 20, H3′),
furanosyl)-9H-purine (8). A 100-mL three-neck round-
bottom flask was charged with 6 (5.5 g, 11 mmol), followed
by methanol (44 mL) and then NaOMe (30 wt. % in MeOH)
H
4.20-4.27 (m, 2H, H4′), 3.48-3.63 (m, 2H, H5′) ppm. 13
C
NMR (CDCl
(d, JC-F ) 20, C1′), 86.21 (d, JC-F ) 10, C4′), 99.41 (d, JC-F
) 184, C2′), 118.08 (C ), 139.87 (C ), 150.03 (C ), 153.22
(C ), 156.84 (C ) ppm. UV (H O) λmax 210 nm (ꢀ 23 406),
λmax 263 nm (ꢀ 13 765). IR (KBr) 3321s, 1661s, 1599s,
1507w, 1466w, 1354m, 1262m, 1211w, 1102w, 1042m,
3
) δ 60.6 (C5′), 73.3 (d, JC-F ) 23, C3′), 85.9
(0.20 mL, 1.1 mmol). The stirred slurry became essentially
clear within 10 min. The reaction was complete by 2.5 h, at
which point acetic acid (62 µL, 1.1 mmol) was added to
adjust the pH to 9. Heptane (44 mL) was added, and the
resultant slurry was stirred for 20 h. The layers were
separated, and the methanol phase was evaporated. To the
residue was added heptane (26 mL) with stirring, followed
by additional heptane (26 mL) after 25 min. The white solid
was vacuum filtered and washed twice with heptane (10 mL).
The resultant solid (3.75 g) was slurried in heptane (50 mL)
for 20 min and filtered to give 3.57 g of solid. This was
dissolved in methanol (50 mL), and heptane (30 mL) was
added. The solvent was evaporated to roughly half of the
original volume, and the product was vacuum filtered and
washed with heptane (2 ×10 mL). Drying under high
vacuum afforded 8 (2.87 g, 8.98 mole) in 82% yield,
contaminated with 5% (HPLC area) of 3.
5
8
6
4
2
2
1
2
-
1
940m, 840w, 769w cm . Mass spec. (electrospray) m/e (%
rel. abundance) 170 (21) (-2-chloroadenine), 300 (24), 303
(64) (M ), 304 (100) (MH ), 305 (14), 306 (27).
+
+
Supporting Information Available
General experimental methods; chromatograms for Tables
-3, 5, 6, and 8; and characterization data for compounds
-8. This material is available free of charge via the Internet
1
3
at http://pubs.acs.org.
Received for review June 10, 2004.
OP049884N
896
•
Vol. 8, No. 6, 2004 / Organic Process Research & Development