Improved stereoselective synthesis of the β-anomer of 1-[3,5-bis-0-(p-chlorobenzoyl)-2-deoxy-D-ribofuranosyl]-5-iodo-2-pyrimidinone

Article abstract of DOI:10.1021/op980204o

The lack of stereochemical control has been a major hurdle in synthesizing β-nucleosides in large scale. This paper reports a study of the effects of different catalysts used in the synthesis of β-nucleosides. The effects of time and temperature on α- and β-anomers are illustrated in this paper. The yield and selectivity of the β-nucleoside have been improved vastly at temperatures between -30 and -40 °C and by using SnCl₄ as the catalyst.

Full text of DOI:10.1021/op980204o

Organic Process Research & Development 1999, 3, 135−138
Technical Notes
Improved Stereoselective Synthesis of the â-Anomer of
1-[3,5-Bis-O-(p-chlorobenzoyl)-2-deoxy-D-ribofuranosyl]-5-iodo-2-pyrimidinone
Ralph Schure, Aye Aye Mar, Brian Pease, Wyeth Jones, Barb Felt, and Mani S. Iyer*
MediChem Research, Inc., 12305 South New AVenue, Lemont, Illinois 60439
Abstract:
this method seems to lack stereochemical control in forming
the â-anomer. Since there are no methods of epimerization11
nor any existing methods12 to synthesize the â-anomer of
1-[3,5-bis-O-(p-chlorobenzoyl)-2-deoxy-D-ribofuranosyl]-5-
iodo-2-pyrimidinone (IPdR, 3) in large scale, we were forced
to reinvestigate conditions which would give a predominantly
â-anomer.
The lack of stereochemical control has been a major hurdle in
synthesizing â-nucleosides in large scale. This paper reports a
study of the effects of different catalysts used in the synthesis
of â-nucleosides. The effects of time and temperature on r- and
â-anomers are illustrated in this paper. The yield and selectivity
of the â-nucleoside have been improved vastly at temperatures
between -30 and -40 °C and by using SnCl4 as the catalyst.
Results and Discussion
In an attempt to find a procedure that increases the â
content of the coupled reaction of 1 with the silylated
2-hydroxy-5-iodopyrimidine, the effects of three acid cata-
lysts were examined. The catalysts were p-nitrophenol,7
trimethylsilyltrifluoromethane sulfonate (trimethylsilyl tri-
flate),5 and stannic chloride.5 The reactions were run initially
at -20 °C for 3 h to examine the effects of the catalysts on
the â/R ratio as well as the total conversion to the R- and
â-anomers. The reaction was assayed by using reversed-
phase (C18 column) HPLC. The results are summarized in
Table 1.
Table 1 indicates that the best overall catalyst was stannic
chloride for coupling of 1 with the silylated 5-iodopyrimi-
dinone 2 to produce the â-anomer of IPdR (3). Good â/R
ratios and excellent conversions were achieved using stannic
chloride as the catalyst. TLC and HPLC assays indicate the
near disappearance of starting materials.
While monitoring the coupling reaction with TLC using
p-nitrophenol as the catalyst, it was observed that after 30
min of the reaction, the major product formed was the
â-anomer, and only a trace amount of the R-anomer was
detected. As the reaction continued, more R-anomer was
produced. At the end of 3 h, the â/R ratio was 1.5/1si.e.,
60% â-anomer and 40% R-anomer. There was still a
significant amount (60%) of unreacted protected sugar
chloride remaining. However, it was obvious from this
experiment that formation of the â-anomer was kinetically
controlled, while the formation of R-anomer was thermo-
dynamically controlled. This strongly suggests that the
product distribution of these coupling reactions is dependent
on thermodynamic and kinetic factors. Therefore, to under-
Introduction
Current interest in synthesizing drugs active against
antiviral1-3 and antitumor4,5 agents have concentrated on the
synthesis of â-nucleosides. Since most ribonuclesides and
2-deoxyribonucleosides exist as a â-anomer in nature,
stereoselective synthesis of the â-nucleosides has become
an enormous challenge to all synthetic chemists. Increasingly,
more drugs are nucleoside based, such as the dideoxynucleo-
side derivatives (e.g., AZT, ddC, ddl, 3TC) used in the
treatment of AIDS.6
The most common nucleosides consist of pyrimidine bases
coupled to a 2-deoxyribose sugar at the 1′ position. The
pyrimidine bases coupled to cytosine and uracil derivatives
have shown the most consistent activity against the virus.7,8
As uracil compounds cannot be deaminated in vivo, there is
a strong interest in their selectivity as well as their ability to
form stable derivatives.9 A method frequently used in the
synthesis is the Vorbruggen coupling (Scheme 1).10 However,
(1) Weygand, F.; Wacker, A.; Dellwag, Z. Naturforsch, Teil B 1952, 7, 19.
(2) Belts, R. E.; Visser, D. J. Am. Chem. Soc. 1955, 77, 736.
(3) Belts, R. E.; Viser, D.; Frisch, D. M. Biochem. Pharmacol. 1960, 5, 157.
(4) Heidelberg, C.; Griesbach, L.; Cruz, O.; Schnitter, R. J.; Grunberg, E. Proc.
Soc. Exp. Biol. Med. 1958, 97, 470.
(5) Burchenal, J. H.; Holmberg, A. D.; Fox, J. J.; Hemphill, S. C.; Reppert, J.
A. Cancer Res. 1959, 19, 497.
(6) Mitsuya, H.; Weinhold, K. J.; Furman, P. A.; St. Clair, M. H.; Lehrman,
S. N.; Gallo, R. C.; Bolognesi, D.; Barry, D. W.; Broder, S. Proc. Natl.
Acad. Sci. U.S.A. 1985, 82, 7096.
(7) Prusoff, W. H.; Chen, M. S.; Fischer, P. H.; Lin, T. S. AdV. Ophthalmol.
1979, 38, 3.
(8) Cheng, Y. C.; Hoffman, P. J.; Ostrander, M. AdV. Ophthalmol. 1979, 38,
1976.
(9) Efange, S. M. N.; Alessi, E. M.; Shih, H. C.; Cheng, Y.; Bardos, T. J.
Med. Chem. 1985, 28, 904.
(10) (a) Niedballa, U.; Vorbruggen, H. J. Org. Chem. 1974, 39, 3654. (b)
Vorbruggen, H.; Bennua, B. Chem. Ber. 1981, 114, 1279.
(11) Lipshutz, B. H.; Hayakawa, H.; Kato, K.; Lowe, R. F.; Stevens, K. L.
Synthesis 1994, 1476.
(12) Okasuchi, T.; Kubota, H.; Narasaka, K. Chem. Lett. 1989, 801.
10.1021/op980204o CCC: $18.00 © 1999 American Chemical Society and Royal Society of Chemistry
Published on Web 01/08/1999
Vol. 3, No. 2, 1999 / Organic Process Research & Development
•
135

Scheme 1
Table 1. Effect of catalysts at -20 °C on â-selective
synthesis of 1-[3,5-bis-O-(p-chlorobenzoyl)-2-deoxy-
D-ribofuranosyl]-5-iodo-2-pyrimidinone (IPdR, 3)
equation,
k-1 - λ
-k-1
0
-k1
k1 + k2 - λ
-k2
0
-k-2
k-2 - λ
catalyst
â/R-anomers % â/(R + â) % (R + â)
) 0
(5)
[
]
p-nitrophenol
trimethylsilyl triflate
stannic chloride
1.5/1
0.43/1
1.5/1
60
30
60
40
15
90
yields the particular solutions, exp(-λit), where λi is the
eigenvalue of the equation and i ) 1, 2, or 3. Thus,
λ1 ) 0
stand the product optimization between the R- and â-ano-
mers, a mathematical model was used to estimate the product
distribution due to kinetic and thermodynamic control at any
time, t. This model is meant to be predictive and not used to
interpret the mechanistic aspect of the reaction. In other
words, by controlling the process, it should be possible to
improve and optimize the â-anomer yields. To produce a
simple mathematical model, it was assumed that the reaction
was first order and that the rate-determining step was the
disappearance of the protected sugar chloride S.
and
2
x
b ( b - 4c
λ2,3
)
(6)
2
where
and
b ) k1 + k2 + k-1 + k-2
This reaction is ordinarily expressed as a competitive
reaction sequence of the type
k-1
â y z S yk2z R
(1)
c ) k1k-2 + k-1k2 + k-1k-2
k1
k-2
Substituting the summation of particular solutions into
eqs 2-4, and applying initial conditions S ) S0, R ) â ) 0
at t ) 0, the general solution is obtained:
where the reactant S proceeds competitively toward products
R and â until a final equilibrium is reached. The time
evolution of the product ratio is determined by two param-
eters: the rate coefficients and the thermodynamics. In eq
1, â is the kinetically controlled product, while R is
thermodynamically stable. Under the condition of low
temperature, the product â is favored, for its low activation
energy speeds up the change of S to â. In contrast, for long
reaction times, the thermodynamic product R tends to
dominate, as it is energetically more stable than the kinetic
product â. At high temperature, the increase of temperature
raises the rate coefficients and facilitates the reaction
equilibrium; this favors the thermodynamic product.
k-1k-2 (k-1 - λ2)(λ2 - k-2
)
2
[S] ) [S]0
+
e-λ t
+
[
λ2λ3
λ2(λ3 - λ2)
(k-2 - λ3)(k-1 - λ3)
(λ3 - λ2)
e-λ t (7)
3
]
k1k-2 k1(λ2 - k-2
)
k1(k-2 - λ3)
e-λ t
+
+
e-λ t (8)
2
3
[â] ) [S]0
[R] ) [S]0
+
[
]
λ2λ3
λ2(λ3 - λ2)
λ3(λ3 - λ2)
k2k-1 k2(k-1 - λ2)
k2(k-1 - λ3)
λ3(λ3 - λ2)
2
3
+
e-λ t
e-λ t (9)
[
]
λ2λ3
λ2(λ2 - λ3)
Based on reaction 1 and assuming first-order kinetics
throughout, the following kinetic equations were set up:
Early on in the reaction, the reaction scheme reduces to
-d[S]
dt
) (k1 + k2)[S] - k-1[â] - k-2[R]
(2)
(3)
(4)
k2
98 R
â &k1
9
S
(10)
d[â]
) k1[S] - k-1[â]
dt
The product R- and â-anomers depend only on the
forward reaction rate. That is, in the early reaction, the
reverse routes may be neglected, since the products of R-
and â-anomers are not significantly accumulated to compete
with the forward reaction rates. Thus, we can show that in
the early stages of the reaction, the following conditions exist:
d[R]
) k2[S] - k-2[R]
dt
By using the matrix method, these coupled differential
equations may be solved analytically.10 The following secular
136
•
Vol. 3, No. 2, 1999 / Organic Process Research & Development

[S] ∼ [S]0e-(k +k )t
(11)
(12)
1
2
k1[S]0
[â] ∼
[R] ∼
[1 - e-(k +k )t
]
1
2
k1 + k2
k2[S]0
[1 - e-(k +k )t
]
(13)
1
2
k1 + k2
For long-term reactions, i.e., t f ∞, eqs 8 and 9 are
replaced by
k1k-2[S]0
[â] ∼
[R] ∼
(14)
(15)
k1k-2 + k-1k2 + k-1k-2
k2k-1[S]0
k1k-2 + k-1k2 + k-1k-2
The concentration of product is invariant, and the system
reaches equilibrium. In this case, the product ratio becomes
k1/k-1 K1
[â]
[R]
)
)
(16)
k2/k-2 K2
Figure 1. Time evolution of reactant S and product r- and
â-anomers at different temperatures.
and depends only on the equilibrium constants K1 and K2.
The process is dominated by thermodynamics, and the
products are thermodynamically or equilibrium controlled.
The time evolution at different temperatures of S and R-
and â-anomers is depicted in Figure 1. The rates for the
reactions were arbitrarily assigned on the basis of experi-
ments monitored every 30 min up to 3 h using stannic
chloride as the catalyst. In this manner, the curves very
closely fit the experimental data. The mathematical model
extrapolates data beyond the 3-h time period. It can be seen
from the curves that the equilibrium concentration factors
are quite different from the corresponding “early stage”
values. The curves demonstrate how the product composition
changes with time at constant temperature. The product
composition, defined as [â]/[R], is seen to remain constant
at about 10 min during the early stage of reaction but then
drops until it assumes a final equilibrium value between 1000
and 5000 min. The curves also show that a decrease in
temperature will decrease all rate constants and correspond-
ingly alter the equilibrium constants. The principal effect of
decreasing the temperature is to increase the time taken to
reach equilibrium, thereby providing a means for obtaining
the kinetically controlled product â-anomer.
[â]eq ) 0.33
[R]eq ) 0.67
with the R-anomer being the main product. The final state
of the system is exclusively controlled by the values of the
equilibrium constants K1 and K2, and the reactions are said
to be under equilibrium or thermodynamic control, i.e., k2/
k-2 > k1/k-1.
Thus, according to the kinetic equation, the chemical
system will favor the kinetic products at the early stage of
the reaction and then will gradually shift to the thermody-
namic products as the reaction time is increased. This requires
a reversible channel connecting the reactants and products.
As the concentration of the kinetic product â in reaction 1
increases, the reverse rate increases so that S will apparently
produce the thermodynamic product R. Consequently, to
increase â, the reverse reaction must be decreased by either
removing the product â or isolating it in terms of a protective
reaction. Accordingly, the â/R ratio decreases to zero as the
reverse S R R is blocked. This is shown in Figure 1. In this
way, either the kinetic or thermodynamic product may be
controlled. The graphs in Figure 1 demonstrate that the
optimum temperature for obtaining high yields of the
â-anomer is between -30 and -40 °C using stannic chloride
as the catalyst. The reaction time should not exceed 1.5 h.
Table 2 summarizes HPLC results of experiments run at
different temperatures using stannic chloride as the catalyst.
In all cases, the reactions were terminated at 1.5 h with
saturated NaHCO3. Acid catalysis will also affect both the
kinetic and thermodynamic factors of the coupling reaction
by accelerating the rate constants and thereby reducing the
time it takes to reach equilibrium. Complete destruction of
the acid catalyst is essential for maximizing the yields of
the â-anomer.
For example, if at a temperature of -30 °C the reaction
is quenched after 10 min, then the reaction composition will
be
[â] ) 0.88
[R] ) 0.12
with â-anomer as the main product. During this period of
time, events are controlled almost exclusively by the values
of the rate constants k1 and k2. In these circumstances, the
reactions are under kinetic control, i.e., k1 > k2.
In the final stage of the reaction, equilibrium is attained,
and the final composition will be
Vol. 3, No. 2, 1999 / Organic Process Research & Development
•
137

Table 2. Effect of temperature on â-selective synthesis of
1-[3,5-bis-O-(p-chlorobenzoyl)-2-deoxy-D-ribofuranosyl]-
5-iodo-2-pyrimidinone (IPdR, 3)
in parts per million (δ ppm) downfield from tetramethylsilane
(TMS) which had been used as an internal standard.
Silylations. To 1.5 g (6.8 mmol) of 2-hydroxy-5-iodo-
pyrimidine were added 15 mL of hexamethyldisilazane
(HMDS) and 0.30 mL of chlorotrimethylsilane. The mixture
was refluxed for 2 h (with exclusion of moisture), cooled,
and concentrated in vacuo to yield a light brown oil. Removal
of any remaining HMDS was accomplished with 2 × 25-
50 mL of anhydrous xylene.
temp (°C) â/R-anomers % â/(R + â) % (R + â)a % âb
0
-20
-30
-40
0.35/1
1.56/1
4.72/1
6.89/1
25.8
60.9
82.5
87.4
89.5
88.9
92.1
79.3
23.1
54.2
76.0
69.3
a Percent of R- and â-anomers in residue. b Percent of â-anomer in residue.
â-Selective Synthesis of 1-[3,5-Bis-O-(p-chlorobenzoyl)-
2-deoxy-D-ribofuranosyl]-5-iodo-2-pyrimidinone (IPdR,
3). To the silylated 2-hydroxy-5-iodopyrimidine was added
50 mL of anhydrous 1,2-dichoroethane. The solution was
cooled to -30 °C with stirring under a nitrogen atmosphere.
The cooling bath used was an acetone-dry ice mixture. Just
enough dry ice was added to the acetone to maintain a
temperature between -30 and -35 °C. After the temperature
was below -30 °C, 2.54 g (6.1 mmol) of the protected sugar
chloride 1 was added. To the cooled solution was added
dropwise a solution of 0.2 mL of SnCl4 in 25 mL of
anhydrous 1,2-dichloroethane while maintaining the tem-
perature below -30 °C. Analysis by TLC (15% ethyl acetate
in dichoroethane) showed that the reaction was essentially
complete in 1.5 h.
Workup of IPdR. After 1.5 h, the mixture was poured
into ice cold saturated NaHCO3 solution (200 mL) and
methylene chloride (300 mL) and stirred for 30 min to
destroy all the acid catalyst. The mixture was filtered, if
necessary, over Celite, the precipitate being washed 3 times
with methylene chloride. The organic layer was washed with
water (200 mL) and then dried over anhydrous MgSO4. The
solvents were removed under reduced pressure. Reversed-
phase HPLC analysis of the residue showed that the major
product was the â-anomer (82.5%), with 17.5% R-anomer
(Figure 3). The pure â-anomer of IPdR (3) was obtained by
crystallization from ethyl acetate, mp 158-160 °C. 1H NMR
(CDCl3, 300 MHz): δ 8.59 (d, J ) 3 Hz, 1H), 8.23 (d, J )
3, 1H), 8.00 (m, 4H), 7.43 (m, 4H), 6.25 (dd, J ) 5.7, 2.1
Hz, 1H), 5.59 (d, J ) 6.6 Hz, 1H), 4.73 (m, 1H), 3.18 (ddd,
J ) 1.5, 4.2, 7.5 Hz, 3H), 2.2 (m, 1H). Anal. Calcd for
C23H17Cl2IN2O: C, 44.90; H, 2.79; Cl, 11.53; I, 20.63; N,
4.55. Found: C, 45.12; H, 2.87; I, 20.38; N, 4.38.
Figure 2. Effect of temperature on â-anomer yields in IPdR.
Table 2 shows the effect of decreasing the temperature
on the â-anomer yields. The yields of â-anomer rapidly
increase as the temperature is decreased. The optimum yields
of â-anomer occur between -30 and -40 °C. Suprisingly,
HPLC analysis indicates that almost all of the starting
materials are reacted, even at -40 °C. 1,2-Dichloroethane
was used as the solvent down to temperatures of -30 °C;
methylene chloride was the solvent used at temperatures
lower than -40 °C. An attempt was made to produce IPdR
(3) at -60 °C. Only a 42% yield of the R- and â-anomers
of IPdR was obtained, with 25.5% of the product being the
â-anomer, approximately 33% unreacted starting materials,
and the rest byproducts. Thus, at -60 °C, the reaction is
incomplete and does not fit the model. The effect of
temperature on yield of â-anomer is shown in Figure 2.
Acknowledgment
We are grateful to Sparta Pharmaceuticals Inc. for their
financial support for this work.
Experimental Section
All melting points are corrected. NMR spectra were run
on a Varian VX-300 NMR. The chemical shifts are reported
Received for review April 16, 1998.
OP980204O
138
•
Vol. 3, No. 2, 1999 / Organic Process Research & Development