The first direct ammonolysis of 2-thiouracil nucleosides to 2- thiocytosine nucleosides

Article abstract of DOI:10.1016/S0040-4039(00)00712-7

Sugar-peracetylated 1-(β-D-glucopyranosyl)-, 1-(β-D-galactopyranosyl)- , and 1-(β-D-xylopyranosyl)-6-aryl-5-cyano-2-thiouracils 2-4 and the corresponding 1-pyranosyl-2-(pyranosylthio)pyrimidines 5-7 undergo efficient ammonolysis (NH₃/MeOH, 0→23

Full text of DOI:10.1016/S0040-4039(00)00712-7

Tetrahedron Letters 41 (2000) 4757±4761
The ®rst direct ammonolysis of 2-thiouracil nucleosides to
2-thiocytosine nucleosides
Lucjan Strekowski,* Ibrahim M. Abdou,y Adel M. E. Attia{ and Steven E. Patterson
Department of Chemistry, Georgia State University, Atlanta, Georgia 30303, USA
Received 11 February 2000; accepted 24 April 2000
Abstract
Sugar-peracetylated 1-(b-d-glucopyranosyl)-, 1-(b-d-galactopyranosyl)-, and 1-(b-d-xylopyranosyl)-6-aryl-
5-cyano-2-thiouracils 2±4 and the corresponding 1-pyranosyl-2-(pyranosylthio)pyrimidines 5±7 undergo
ecient ammonolysis (NH3/MeOH, 0!23ꢀC, 16 h) to give the corresponding 1-(b-d-pyranosyl)-6-aryl-5-
cyano-2-thiocytosines. # 2000 Elsevier Science Ltd. All rights reserved.
Keywords: 6-aryl-5-cyano-2-thiouracils; 1-(b-d-pyranosyl)-2-thiocytosines.
Unnatural nucleosides continue to draw attention as potential biologically active agents, and
many diverse structures are being designed and synthesized. In particular, it has been reported
recently1 that the reaction of a dianion derived from a 2-thiouracil 1 with 2,3,4,6-tetra-O-acetyl-a-
d-glucopyranosyl bromide or its a-d-galactopyranosyl isomer gives the corresponding N3,S2-
bis(pyranosyl)pyrimidin-4-one 11 (Scheme 1). It has also been claimed that the treatment of these
products with ammonia in methanol furnishes the corresponding N3-pyranosyluracils.
In this paper, we report a reinvestigation of the research discussed above and show that the bis-
nucleosides (the major products) derived from 1 are N1,S2-bis(pyranosyl) derivatives 5 and 6
rather than the claimed N3,S2-isomers 11 and they are accompanied by N1-pyranosylpyrimidines
2 and 3. Xylopyranosylpyrimidines of analogous structures 4 and 7 were also obtained as part of
this work (Scheme 1). Most important, however, in our hands the deacetylation of 2±7 by treat-
ment with ammonia in methanol, in addition to removal of the pyranosyl function from the S2
atom at the pyrimidine of 5±7, was also accompanied by an unprecedented ammonolysis to furnish
N1-(b-d-pyranosyl)-2-thiocytosines 8±10 in high yields. These results are fundamentally di€erent
from those reported, and consistent results were obtained in all cases studied.
* Corresponding author. Tel: 404-651-0999; fax: 404-651-1416; e-mail: lucjan@gsu.edu
y
Present address: School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, GA 30332, USA.
Present address: Chemistry Department, Faculty of Education, Tanta University, Kafr El-Sheikh Branch, Egypt.
{
0040-4039/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)00712-7

4758
Scheme 1.
Our ®ndings have important rami®cations for the synthesis of novel cytosine nucleosides and
analogs. First, in addition to the starting materials 1, a large number of other 5-cyano-2-thiouracils
are readily available.1 3 Second, the products 8±10 are potential substrates for the preparation of
nucleosides that contain a fused heterocyclic system because the 4-amino-5-cyano functionality is
a well known synthon for annulation of the pyrimidine to pyrazolopyrimidines, pyridopyrimidines,
and pyrimidopyrimidines.4
The treatment of a solution of 1 (Ar=Ph, 10 mmol) and KOH (22 mmol) in water (6 mL) with
2,3,4,6-tetra-O-acetyl-a-d-glucopyranosyl bromide (22 mmol) in acetone (30 mL) and stirring the
mixture1 (25ꢀC, 4 h) resulted in the disappearance of 1 and the formation of two products that

4759
were separated by silica gel chromatography (hexanes:ether:chloroform, 1:2:2). The structures of
the major and minor products were established as 5a and 2a, respectively, by elemental analyses,
1
IR, and H NMR including COSY and NOESY methods. In particular, the IR spectrum of 2a
shows absorptions for acetoxy carbonyl groups (1720±1750 cm^1), and a strong signal at 1698
cm^1 that is also present in the IR spectrum of 1 and ch0aracteristic for the pyrimidin-4-one.5 In
1
the H NMR spectrum of 2a the anomeric proton H1 of the glucopyranosyl moiety gives a
doublet at ꢀ 5.99 with a coupling constant J(H10±H20) of 10.9 Hz for the diaxial interaction. First,
this large coupling constant is characteristic for a b-glucopyranosyl anomer.6 Second, the chemi-
cal shift is typical for N1-substitution of 2-thiouracils, and the structurally related N3-substituted
2-thiouracils give a resonance for H10 at a much lower magnetic ®eld.7 The N1-substitution in 2a
was con®rmed by an NOE experiment in which the irradiation at ꢀ 5.99 for H10 gave a strong
signal at ꢀ 8.08 corresponding to ortho protons of the adjacent phenyl group. Additional
enhancements were observed at ꢀ 5.35 and 5.15 for H30 and H50, respectively, and all these
interactions were con®rmed by analysis of the NOESY spectrum of 2a. These conclusions derived
from the NOE studies are fully supported by the results of molecular modeling by using a PC
Model program. The conformational analysis of 2a gave two low-energy conformations, syn
(shown) and anti, in both of which the plane of the pyrimidine bisects the sugar moiety while the
phenyl ring is twisted relative to the pyrimidine. In the more stable (by 0.15 kcal/mol) energy-
minimized syn-structure the distances between H10 of the sugar and the ortho protons of the
phenyl are 2.59 and 3.84 A, which is fully consistent with the observed NOE enhancements. By
contrast, such distances are beyond the limit of NOE interactions for the 0alternative N3-isomer of
2a. In the 1H NMR spectrum of 5a a doublet for H10 at ꢀ 5.98, J(H10±H2 )=10.4 Hz, is indicative
of the presence of the b-glucopyranosyl moiety at position N1 of the pyrimidine, as in 2a. Analysis
of the NOESY spectrum of 5a con®rmed this result. The chemical shift for the anomeric proton
H100 of the second b-glucopyranosyl group in 5a at ꢀ 6.50 is too small for an N3-nucleoside, and the
O4-nucleoside can be ruled out on the basis of a strong infrared absorption at 1685 cm^1.
Interestingly, the reaction of 1 (ArˆPh, 1 equiv.) in the presence of KOH (2.2 equiv.) with
increasing amounts of 2,3,4,6-tetra-O-acetyl-a-d-glucopyranosyl bromide, from 0.5 equiv. to 2.2
equiv. required for a consumption of 1, always furnished a mixture of 2a and 5a. It can be suggested
that 2a and an S2-linked mononucleoside are formed initially and both compounds are precursors
to 5a. Consistent with this suggestion are calculated HOMO electron densities for anions derived
from 1 (ArˆPh) and its S-methyl derivative (Fig. 1). The latter compound is a model for the
presumed S2-linked mononucleoside. As can be seen, the sulfur atom in the dianion 1-N1^N3^
and monoanions 1-N1^, 1-N3^ is the most nucleophilic site. A competitive formation of the N1-
nucleoside 2a from each of the monoanions should be possible, as observed. On the basis of the
calculated HOMO electron densities and the favorable heat of formation for MeS1^ the presumed
S2-nucleoside should be easily transformed into the N1,S2-bis-nucleoside 5a, also as observed.
Ammonolysis of mononucleosides 2±4, dinucleosides 5±7 or the mixtures 2/5, 3/6 and 4/7
furnished the corresponding 1-(b-d-pyranosyl)-2-thiocytosines 8±10 in yields of 84±87%. In a
typical experiment a solution of 5a (1 mmol) in MeOH (10 mL) was added at 0ꢀC to a saturated
solution of NH3 in MeOH (25 mL) and the mixture was stirred at 0ꢀC for 4 h and then at 23ꢀC
for an additional 12 h. Concentration followed by silica gel chromatography (MeOH:CHCl3,
1:19) gave 8a. Analysis of the NOESY spectra showed that products 8±10 contain a pyranosyl
moiety linked to N1 of the pyrimidine. The experimental and computer-simulated 13C NMR
spectra of 8±10 closely matched each other. In comparison to the infrared spectra of 2±7,
the 2-thiocytosine nucleosides 8±10 lack absorption at 1680±1690 cm^1 that is characteristic for

4760
the pyrimidin-4-one. Their IR spectra are dominated by strong CN, OH, and NH2 absorptions.
1
Two sets of resonances for OH and NH2, both exchangeable with D2O, are observed in the H
NMR spectra of 8±10 as well.x
The suggested mechanism for 8±10 is given in Scheme 2. Removal of a sugar moiety from S-
nucleosides under certain basic/nucleophilic conditions to give an anion such as 12 is well
known.7 9 In addition to the standard deacetylation at the pyranosyl moiety of 12, the cyano
group of 12 may also undergo an addition reaction with ammonia to generate an amidine 13. The
subsequent intramolecular cyclization of 13 to 14 is consistent with the computational work of
Fig. 1 which shows that the C4 atom in the analogous anion 1-N3^ is electrophilic in nature.
Elimination of hydroxide ion from 14 (sp3-C4!sp2-C4) would generate a less strained system 15
which is the suggested precursor to 8±10.
Scheme 2.
Figure 1. Calculated (PM3-AQ) HOMO electron densities for anions derived from 1 (ArˆPh) and MeS1 by deproto-
nation at the indicated sites. H denotes the calculated heat of formation
x
Products 2±7 and 8±10 were crystallized from 95% EtOH and 95% MeOH, respectively. All compounds gave satisfactory
.
microanalysis results (C, þ0.3; H, þ0.2; N, þ0.2; S, þ0.3) as exempli®ed for 8a 1.5 H2O. Calcd/found: C, 48.92/48.84; H,
5.03/5.10; N, 13.42/13.33; S, 7.67/7.73. The elemental analysis of the remaining compounds 2±10 also showed the presence
1
of crystallization water that was demonstrated independently by the use of H NMR. All yields given below are for
analytically pure products. The yields of 8±10 are for ammonolysis of the respective mixtures 2a/5a, 2b/5b, 3/6 and 4/7.
Compound, yield, mp, a2D5 (c=2 mg/mL, CHCl3 for 2±7ꢀand MeOH for 8±10): 2a 2H2O, 20%, 214±216 C, 38.3;
ꢀ
.
ꢀ
ꢀ
.
.
.
.
2b 2H2O, 30%, 189±192 C, 23.0; 3 2.5 H2O,ꢀ23%, 232±234 C, 36.4; 4 H2O, 20%, 165±168 C, 26.7; 5a 0.5 H2O, 40%,
ꢀ
ꢀ
ꢀ
.
.
.
114±116 C, 45.0; 5b 0.5H2ꢀO, 40%, 118±120 C, 38.0; 6 H2O, 35%, 128±130 C, 23.0; 7 H2O, 35%, 123±125 C, 21.6;
ꢀ
ꢀ
.
.
.
.
8a 1.5H2O, 86%, 190±193 C, 21.0; 8b 2H2O, 86%, 190±193 C, 25.5; 9 1.5H2O, 87%, 192±195 C, 22.6; 10 2H2O, 84%,
196±198ꢀC, 16.8. Selected H NMR spectra (400 MHz, 25ꢀC, CDCl3/TMS): 2a, ꢀ 1.90±2.14 (4s, 12H), 4.03 (m, 2H),
1
4.43 (s, 1H), 5.15 (m, 1H), 5.35 (m, 1H), 5.54 (m, 1H), 5.99 (d, J=10.9 Hz, 1H), 7.64 (m, 3H), 8.08 (m, 2H); 5a, ꢀ 1.83±
2.01 (8s, 24H), 4.25 (m, 6H), 5.11 (m, 4H), 5.55 (m, 2H), 5.98 (d, J=10.4 Hz, 1H), 6.50 (d, J=8.0 Hz, 1H), 7.62 (m,
3H), 8.08 (m, 2H); 8a, ꢀ 3.21 (m, 4H), 3.48 (m, 1H), 3.68 (m, 1H), 4.80 (m, exchangeable with D2O, 4H), 5.37 (d,
J=10.0 Hz, 1H), 7.18 (bs, exchangeable with D2O, 2H), 7.47 (m, 3H), 7.80 (m, 2H).

4761
References
1. Khodair, A. I.; Ibrahim, E. E.; El Ashry, E. S. H. Nucleosides Nucleotides 1997, 16, 433±444, and references cited
therein.
2. Vanden Berghe, D. A.; Vlietinck, A. J. J. Heterocycl. Chem. 1984, 21, 1307±1312.
3. Navio, J. L. G.; Lorente, A.; Soto, J. L. Heterocycles 1982, 19, 305±310.
4. Fikry, R. M. J. Indian Chem. Soc. 1996, 73, 698±699, and references cited therein.
5. Winkley, M. W.; Robins, R. K. J. Org. Chem. 1968, 33, 2822±2827.
6. Abdel-Megeed, M. F.; Salaeh, M. A.; Aly, Y. L.; Abdo, I. M. Nucleosides Nucleotides 1995, 14, 1985±1996.
7. Niedballa, U.; Vorbruggen, H. J. Org. Chem. 1974, 39, 3660±3663.
8. Vorbruggen, H.; Strehlke, P. Chem. Ber. 1973, 106, 3039±3045.
9. Ueda, T.; Nishino, H. J. Am. Chem. Soc. 1968, 90, 1678±1681.