Locked nucleosides based on oxabicyclo[3.2.1]octane and oxabicyclo[2.2.1]heptane skeletons

Article abstract of DOI:10.1021/jo100194z

(Figure Presented) Intramolecular nitrone cycloaddition (INC) reaction on a D-glucose derived substrate carrying an allyl group at C-I and an enose-nitrone at C-5 or an aldehyde-nitrone at C-I and vinyl group at C-4 furnished a tricyclo[6.2.1.0^2,6]undecane or a tricyclo[5.2.1.0 ^2,6]decane ring structure. These tricycles were converted to bicylic nucleosides with oxabicyclo[3.2.1]octane and oxabicyclo[2.2.1]heptane rings in three steps. An oxabicyclo[3.2.1]octane ring compound could alternatively be formed by RCM reaction between C-1-allyl and C-4-vinyl moieties and transformed to nucleoside analogues through a nucleophilic substitution reaction. Participation of a neighboring benzyl ether substituent in one case paved the way for an enantiodivergent synthesis.

Full text of DOI:10.1021/jo100194z

pubs.acs.org/joc
Locked Nucleosides Based on
Oxabicyclo[3.2.1]octane and
Oxabicyclo[2.2.1]heptane Skeletons
Ramprasad Ghosh, Joy Krishna Maity, Basudeb Achari,
and Sukhendu B Mandal*
Department of Chemistry, Indian Institute of Chemical
Biology (a unit of CSIR), 4, Raja S. C. Mullick Road,
Jadavpur, Kolkata 700 032, India
sbmandal@iicb.res.in
Received February 4, 2010
FIGURE 1. Strategies for locked bicyclic nucleosides synthesis.
and solid-state conformations could be very similar, various
conformationally locked bicyclic nucleosides (10,20-linked,1
10,30-linked,2 20,30-linked,3 20,40-linked,4 30,40-linked,5 and
30,50-linked6) have been targeted for synthesis. These have
restricted geometrical shape, are potentially useful as inhi-
bitors of certain enzymes,7 and are building blocks of
oligonucleotides.8
We have been working for some time on the syntheses of
various nucleoside analogues from the cheap and readily
available 1,2:5,6-di-O-isopropylidene-R-D-glucofuranose
exploiting the intramolecular nitrone cycloaddition (INC)
reaction as a synthetic tool.9 As a part of our program to
construct a new class of C(1)fC(5)-linked bicycles, we
envisioned two strategies (Figure 1), which entailed inclusion
of an olefin functionality at C-5 and an aldehyde unit at C-1
(path A), or an allyl moiety at C-1 and an aldehyde group at
Intramolecular nitrone cycloaddition (INC) reaction on a
D-glucose derived substrate carrying an allyl group at C-1
and an enose-nitrone at C-5 or an aldehyde-nitrone at C-1
and vinyl group at C-4 furnished a tricyclo[6.2.1.02,6]-
undecane or a tricyclo[5.2.1.02,6]decane ring structure.
These tricycles were converted to bicylic nucleosides with
oxabicyclo[3.2.1]octane and oxabicyclo[2.2.1]heptane
rings in three steps. An oxabicyclo[3.2.1]octane ring
compound could alternatively be formed by RCM reac-
tion between C-1-allyl and C-4-vinyl moieties and
transformed to nucleoside analogues through a nucleo-
philic substitution reaction. Participation of a neighbor-
ing benzyl ether substituent in one case paved the way for
an enantiodivergent synthesis.
(2) Kværnø, L.; Wightman, R. H.; Wengel, J. J. Org. Chem. 2001, 66,
5106–5109.
(3) Nielsen, P.; Pfundheller, H. M.; Olsen, C. E.; Wengel, J. J. Chem. Soc.,
Perkin Trans. 1 1997, 3423–3434.
(4) (a) Enderlin, G.; Nielsen, P. J. Org. Chem. 2008, 73, 6891–6894. (b)
Obika, S.; Nanbu, D.; Hari, Y.; Andoh, J.; Morio, K.; Doi, T.; Imanishi, T.
Tetrahedron Lett. 1998, 39, 5401–5404. (c) Koshkin, A. A.; Singh, S. K.;
Nielsen, P.; Rajwanshi, V. K.; Kumar, R.; Meldgaard, M.; Olsen, C. E.;
Wengel, J. Tetrahedron 1998, 54, 3607–3630. (d) Kumar, S.; Hansen, M. H.;
ꢀ
Albæk, N.; Steffansen, S. I.; Petersen, M.; Nielsen, P. J. Org. Chem. 2009, 74,
6756–6769. (e) Srivastava, P.; Barman, J.; Pathmasiri, W.; Plash-Kevych, O.;
Wenska, M.; Chattopadhyaya, J. J. Am. Chem. Soc. 2007, 129, 8362–8379.
(5) Freitag, M.; Thomasen, H.; Christensen, N. K.; Petersen, M.; Nielsen,
P. Tetrahedron 2004, 60, 3775–3786.
€
(6) (a) Tarkoy, M.; Bolli, M.; Schweizer, B.; Leumann, C. Helv. Chim.
The conformation of the sugar moiety in nucleosides is
believed to play a crucial role in modulating biological
activity. However, the attempt to correlate a specific type
of sugar conformation to bioactivity is beset with the twin
problem that the pentose ring is quite flexible in solution and
its conformation can differ sharply in the solid state. Con-
sequently, any structure-activity study based on solid-state
conformation would be imperfect. Thus, aiming at the
generation of nucleoside analogues for which both solution
Acta 1993, 76, 481–510. (b) Agrofoglio, L. A.; Challand, S. R. Acyclic,
Carbocylic and L-Nucleosides; Kluwer, Academic Publisher: Dordrecht, The
Netherlands, 1998; pp 174-284. (c) Lin, T. S.; Luo, M. Z.; Liu, M. C.; Pai, S. B.;
Dutschman, G. E.; Cheng, Y. C. J. Med. Chem. 1994, 37, 798–803. (d) Shaikh,
K. I.; Kumar, S.; Lundhus, L.; Bond, A. D.; Sharma, P. K.; Nielsen, P. J. Org.
Chem. 2009, 74, 1557–1566.
(7) Russ, P.; Schelling, P.; Scapozza, L.; Folkers, G.; De Clercq, E.;
Marquez, V. E. J. Med. Chem. 2003, 46, 5045-5054 and references cited
therein.
(8) Meldgaard, M.; Wengel, J. J. Chem. Soc., Perkin Trans. 1 2000,
3539-3554 and references cited therein.
(9) (a) Roy, B. G.; Maiti, J. K.; Drew, M. G. B.; Achari, B.; Mandal, S. B.
Tetrahedron Lett. 2006, 47, 8821–8825. (b) Tripathi, S.; Roy, B. G.; Drew,
M. G. B.; Achari, B.; Mandal, S. B. J. Org. Chem. 2007, 72, 7427–7430.
(c) Shing, T. K. M.; Elsley, D. A.; Gillhouley, J. G. J. Chem. Soc., Chem.
Commun. 1989, 1280–1282. (d) Gallos, J. K.; Stathakis, C. I.; Kotoulas, S. S.;
Koumbis, A. E. J. Org. Chem. 2005, 70, 6884–6890.
(1) (a) Plash-Kevych, O.; Chatterjee, S.; Honchanenko, D.; Pathmasiri,
W.; Chattopadhyaya, J. J. Org. Chem. 2007, 72, 4716–4726. (b) Loiseleur, O.;
Ritson, D.; Nina, M.; Crowley, P.; Wagner, T.; Hanessian, S. J. Org. Chem.
2007, 72, 6353–6363.
DOI: 10.1021/jo100194z
r
Published on Web 03/02/2010
J. Org. Chem. 2010, 75, 2419–2422 2419
2010 American Chemical Society

Ghosh et al.
JOCNote
SCHEME 1. An Approach to the Tricyclic Decane System in 9
SCHEME 2. Synthesis of Bicyclic Nucleoside 12
SCHEME 3. Synthesis of Tricyclic Undecane Ring in 17
C-4 (path B) followed by INC reaction with N-alkyl hydro-
xylamine to synthesize hitherto unknown tricyclic ring
systems. In another strategy, a ring closing metathesis reac-
tion between an allyl group at C-1 and an olefin unit at
C-5 (path C) could produce a locked bicyclic ring. It was
important to ensure that the generated C (1)-substituent
involved in ring closure be β oriented so that cyclization
with the C-5 olefin substituent may proceed in the desired
manner to achieve our target. Once the desired isoxazolidino
bicyclic rings are derived, conversion to their corresponding
nucleosides having bicyclo[3.2.1]octane and bicyclo[2.2.1]-
heptane could follow through cleavage of the isoxazolidine
rings and construction of the nucleoside base. For the RCM
reaction product, nucleophilic attack by a nucleoside base on
an appropriate center of the furanose moiety could afford
locked bicyclic carbanucleoside analogues. The importance
of these nucleosides lacking glycosidic linkages lies in the fact
that they are metabolically stable toward cleavage by phos-
phorylases and hydrolases.
Synthesis of Oxabicyclo[2.2.1]heptane Nucleoside (Path A). I.
Preparation of the Tricyclic Decane System 9 via INC Reaction.
Toward the insertion of an aldehyde group in β-orientation and
eventual transformation to nitrone, introducing β-cyano moiety
at C-1 was necessary. For this, removal of acetonide protection
from the D-glucose-derived product 110 by acid treatment and
subsequent acetylation were carried out, affording 2 as an
anomeric mixture. This was treated with trimethylsilyl cyanide
in the presence of borontrifluoride-etherate to produce the
cyanide 3 exclusively in 71% yield. The peak at 2244 cm-1 in its
IR spectrum clearly indicated the presence of the cyano group,
while the β-orientation was evident from the appearance of the
H-1 signal as singlet at δ 5.47 (coupling constant J1,2 = 0 Hz) in
the 1H NMR spectrum. Acetyl deprotection of 3with potassium
carbonate in methanol furnished the desired product 4 (88%)
along with the minor ester derivative 5 (8%), separated easily by
column chromatography (Scheme 1). Protection of C-2 OH of 4
by benzylation delivered 6 (87%), which was treated with
DIBAL-H in dry DCM to yield the corresponding aldehyde 7
as hydrate. Without purification this was reacted with N-benzyl
hydroxylamine in refluxing ethanol to produce the tricyclic
product 9 through the in situ generated enose-nitrone 8 in good
yield (60% in two steps). The absence of upfield proton and
carbon signals expected for a methylene group in the NMR
spectra clearly excluded the alternate structure. The structure of
9 was elaborated by 2D NMR spectroscopy. The energetically
favored cis ring juncture stereochemistry is expected in this
tricyclo[5.2.1.02,6]decane system. The signals for H-2 (δ 3.01)
and H-9 (δ 3.85) indeed show a characteristic NOESY relation-
ship in agreement with the observed distance between them
˚
(2.38 A) in the energy-minimized structure (obtained using
Chem Office 6.0).
II. Transformation of 9 to Oxabicyclo[2.2.1]heptane
Nucleoside Analogue 12. Transfer hydrogenolysis of 9 with
Pd/C/cyclohexene (Scheme 2) removed the isoxazolidine
ring as well as benzyl protections to produce the bicyclic
trihydroxyamino compound 10 (93%), which was coupled
with 5-amino-4,6-dichloropyrimidine to furnish 11 (48%).
Construction of purine ring on 11 by treatment with triethyl
orthoformate smoothly afforded the desired bicyclic chloro-
purine nucleoside 12 (55%).
Synthesis of Oxabicyclo[3.2.1]octane Nucleoside (Path B). I.
Construction of the Tricyclic Undecane System 17 by INC
Reaction. For this goal, the synthon 1311 was reacted with allyl
trimethylsilane in the presence of boron trifluoride-etherate
to smoothly furnish the C-1 allylated product 14 (72%).
Deacetylation to 15 followed by vicinal diol cleavage generated
an aldehyde, which without purification was treated with
N-benzyl hydroxylamine in refluxing ethanol to furnish
(Scheme 3) the tricyclic product 17 through nitrone 16.
The absence of signals for olefinic protons and the occur-
1
rence of signals for the N-benzylic group in the H NMR
spectrum of 17 proved that cyclization involving the olefin
and the nitrone moieties had indeed taken place. The forma-
tion of the product through the other possible cyclization
(10) Rathbone, E. B.; Woolard, G. R. Carbohydr. Res. 1976, 48, 143–146.
2420 J. Org. Chem. Vol. 75, No. 7, 2010
(11) Kiss, J.; Wyss, P. C. Tetrahedron 1976, 32, 1399–1402.

Ghosh et al.
JOCNote
SCHEME 4. Synthesis of Locked Bicyclic Nucleoside 20
SCHEME 6. Synthesis of Enantiomeric Nucleosides 34 and 35
SCHEME 5. Synthesis of Bicyclic Nucleoside 25
Hydrogenation of the olefin moiety as well as debenzylation
with hydrogen gas in the presence of Pd/C produced 24 (72%).
Protection of the hydroxyl group of this product as triflate and
subsequent nucleophilic attack by 6-chloropurine furnished the
bicyclic nucleoside 25 (45%). The formation of N-9-linked
nucleoside12 was confirmed by the presence of a signal at
δ 129.7 (C-50) in the 13C NMR spectrum. The N-7-linked
nucleoside would have generated instead a signal at δ ∼125.0.
Further, C-80 of 25 resonates at δ 145.1 rather than at δ ∼155.0
as expected for the N-7 isomer.
In a separate experiment, conversion of 23 to 26 by K2CO3
treatment followed by attempted installation of nucleoside
base at the hydroxyl bearing carbon through its triflate to
generate 27 was unsuccessful, possibly due to the hindrance
offered by the β-oriented ring residue to the approach of the
base from the β-side. Therefore, it was decided to insert the
nucleoside base before the formation of the bicyclic ring by
RCM reaction.
II. Nucleosidation and RCM Reaction to Enantiomeric
Oxabicyclo[3.2.1]octane Nucleoside Analogues 34 and 35.
Toward this end, formation of the triflate of 21 followed
by nucleosidation with 6-chloropurine in refluxing DMF
furnished an inseparable mixture of 30 and 31 (Scheme 6).
The crude NMR spectrum of this mixture indicated the
presence of four distinct aromatic proton signals at δ
∼8.0-8.4 attributed to the formation of the two nucleoside
derivatives. The chloro group was replaced by dimethylami-
no group by reaction with dimethylamine formed in situ
from DMF. The formation of the mixture of 30 and 31 from
21 could only be explained by invoking the participation of
the neighboring benzyloxy substituent during elimination of
the triflate substituent. This generated a nonisolable epox-
onium intermediate 29, which suffered nucleophilic attack
by nucleoside base from either side of the epoxide ring. This
mixture (of 30 and 31) without purification was subjected to
mode was not observed. The 1H-1H COSY, 1H-13C HSQC,
1
and H-13C HMBC spectra clearly identified most of the
signals establishing the disposition of the atoms in the
tricyclic framework. The newly generated ring juncture
stereochemistry, which is expected to be cis, could be either
R,R or β,β in the tricyclo[6.2.1.02,6]undecane system. The
˚
distance between H-6 (δ 2.80) and H-9 (δ 4.13) is 2.32 A in
the energy-minimized structure of the β,β-isomer (steric
energy 26.5 kcal/mol) suggesting that their signals should
show correlation in the NOESY spectrum. This was indeed
observed. The other cis structure has 32.2 kcal/mol steric
energy and is not expected to show the NOESY relationship
between H-6 and H-9. Therefore, the formation of this
structure could be ruled out.
II. Conversion of 17 to Oxabicyclo[3.2.1]octane Nucleoside
Analogue 20. Hydrogenolytic cleavage (Scheme 4) of 17
produced 18, which without further purification was con-
verted to 19 (50%) and subsequently to 20 (60%) through the
familiar sequence of reactions, viz. coupling with 5-amino-
4,6-dichloropyrimidine and then conversion to the bicyclic
carbanucleoside analogues.
Synthesis of Oxabicyclo[3.2.1]octane Nucleoside (Path C). I.
RCM Reaction and Nucleosidation To Produce 25. In this route,
the C-allylated product 21 (Scheme 5) was prepared in 85%
yield by treatment of 1 with allyl trimethylsilane in the pre-
sence of boron trifluoride-etherate. Acetylation of 21 furnished
22, which upon RCM reaction with Grubbs’ catalyst (2nd
generation) afforded the cyclized compound 23 in 70% yield.
(12) (a) Paquette, L. A.; Kahane, A. L.; Seekamp, C. K. J. Org. Chem.
2004, 69, 5555–5562. (b) Kjellberg, J.; Johansson, N. G. Tetrahedron 1986,
42, 6541–6544.
J. Org. Chem. Vol. 75, No. 7, 2010 2421

Ghosh et al.
JOCNote
(0.85 g, 6.87 mmol) was added to a refluxing dry ethanolic
solution of 7 (1.75 g, 5.72 mmol) and the heating was continued
for 5 h. The solvent was evaporated in vacuo to obtain a
residue, which was extracted with CHCl3 (2 ꢀ 25 mL). The
CHCl3 solution was dried (Na2SO4) and the solvent was
evaporated to afford a crude product, which was purified by
silica gel column chromatography with use of EtOAc-petroleum
ether (7:93) to furnish 9 as a sticky gum (1.37 g, 60% yield): [R]25
D
þ69.9 (c 0.50, CHCl3); 1H NMR (CDCl3, 300 MHz) δ 3.01 (d, 1H,
J = 7.8 Hz), 3.24 (s, 1H), 3.31 (q-like, 1H, J = 7.8 Hz), 3.61 (t, 1H,
J = 7.5 Hz), 3.76 (d, 1H, J = 12.3 Hz), 3.85 (br d, 2H, J = 6.9 Hz),
4.09 (t, 1H, J = 8.4 Hz), 4.21 (d, 1H, J = 12.0 Hz), 4.25-4.29 (m,
2H), 4.34 (d, 1H, J = 12.3 Hz), 4.43 (s, 2H), 7.21-7.43 (m, 15H);
13C NMR (CDCl3, 75 MHz) δ 46.8 (CH), 62.0 (CH2), 68.8 (CH2),
70.7 (CH2), 70.8 (CH), 72.4 (CH2), 78.2 (CH), 83.3 (2 ꢀ CH), 85.3
(CH), 127.6 (CH), 127.7 (3 ꢀ CH), 127.8 (2 ꢀ CH), 127.9 (CH),
128.2 (2 ꢀ CH), 128.3 (2 ꢀ CH), 128.4 (2 ꢀ CH), 129.4 (2 ꢀ CH),
136.2 (C), 137.3 (C), 137.4 (C); ESIMS, m/z 466 (M þ Na)þ. Anal.
Calcd for C28H29NO4: C, 75.82; H, 6.59; N, 3.16. Found: C, 75.57;
H, 6.38; N, 3.01.
FIGURE 2. Characteristic NOESY relationship in 32 and 33.
RCM reaction with Grubbs’ catalyst (second generation) to
provide the cyclized derivatives 32 and 33 separated by chro-
matography. Hydrogenation reaction of both compounds
afforded the nucleoside derivatives 34, [R]25 -19.8, and 35,
D
[R]25D þ19.0. Their enantiomeric nature was anticipated from
the optical rotations, which were almost identical in magnitude
but opposite in sign. Besides, they exhibited virtually identical
1H and 13C NMR spectra.
For the structure elucidation of 32 and 33, the ipso proton
signal for the carbon carrying the heterocyclic group was
initially identified through HMBC correlation with the C-80
signal. With 32, this signal (δ 5.11) showed further HMBC
correlation with the peak for C-5 (δ 32.7) but not with those of
the olefinic carbons. The reverse was true for 33 where the ipso
proton signal (δ 5.23) indeed showed correlation with the peak
for C-3 (δ 127.7) besides those for the neighboring ones. The
crucial evidence for the stereochemistry shown came thereafter
from NOESY studies. For 32, the H-7 signal showed a NOESY
relationship (Figure 2) with one of the C-5 proton signals (δ2.25)
while the H-80 (δ 7.87) peak was correlated with H-8 (δ 4.49),
H-6 (δ 4.45), and H-2 (δ 4.61) peaks indicating their cis
orientation. With 33 on the other hand, similar evidence showed
the closeness in space of H-8 (δ 5.23) with H-3 (δ 6.05) as well as
benzylic protons (δ 4.49 and 4.89), and of H-80 (δ 7.99) with H-7
(δ 4.19), H-2 (δ 4.40), and H-6 (δ 4.84).
In conclusion, the work described herein unravels a synthetic
route to a new class of conformationally locked bicyclic carba-
nucleosides with bicyclo[3.2.1]octane and bicyclo[2.2.1]heptane
ring systems. The steps involve the application of both INC and
RCM reactions on easily prepared D-glucose-derived substrates.
The simplicity of the strategy for precursor assembly for INC as
well as RCM reactions makes the procedure both efficient and
useful. The success of the strategy to synthesize enatiomeric pair
of bicyclo[3.2.1]octane nucleoside analogs is an added advantage.
(1R,5S,6R,7S)-7-(6-Chloropurin-9-yl)-8-oxabicyclo[3.2.1]octan-
6-ol (25). Triflic anhydride (0.31 mL, 1.85 mmol) was added to a
solution of 24 (250 mg, 1.34 mmol) in DCM (15 mL) containing
pyridine (0.3 mL) at 0 °C and the solution was stirred for 1 h.
Evaporation of the solvent furnished a residue (250 mg). To the
solution of the residue in dry DMF (15 mL) were added K2CO3
(130 g, 0.94 mmol), 6-chloropurine (183 mg, 1.18 mmol), and
18-crown-6 (235 mg, 0.79 mmol), and the mixture was heated
at 80 °C for 5 h. Usual workup and purification by silica gel
chromatography with EtOAc-petroleum ether (3:2) as eluent
furnished 25 (169 mg, 45%) as a foamy solid: [R]25 þ5.6
D
1
(c 0.52, CHCl3); H NMR (CDCl3, 300 MHz) δ 1.33-1.50 (m,
4H), 1.86-1.91 (m, 2H), 3.86 (m, 1 H), 4.08 (d, 1H, J = 7.6 Hz),
4.13 (d, 1H, J = 7.6 Hz), 4.31 (dd, 1H, J = 6.8, 13.5 Hz), 8.24 (s,
1H), 8.75 (s, 1H); 13C NMR (CDCl3, 75 MHz) δ 19.9 (CH2), 24.7
(CH2), 30.1 (CH2), 67.6 (CH), 73.9 (CH), 77.6 (CH), 80.0 (CH),
129.7 (C), 145.1 (CH), 148.9 (C), 152.3 (CH), 155.0 (C); ESIMS,
m/z 303 (M þ Na)þ for Cl35 and 305 (M þ Na)þ for Cl37. Anal.
Calcd for C12H13ClN4O2: C, 51.34; H, 4.67; N, 19.96. Found: C,
51.21; H, 4.45; N, 19.73.
Acknowledgment. The authors thank CSIR, Govern-
ment of India for providing Senior Research Fellowships
(to R.G. and J.K.M.) and an Emeritus Scientist scheme
(to B.A.). The financial support from the CSIR Network
Project (No. NWP 00036) is gratefully acknowledged.
Supporting Information Available: General and experi-
mental details and copies of 1H and 13C NMR spectra of 3, 6,
9-12, 14, 17, 19-21, 23-25, and 32-35. This material is
available free of charge via the Internet at http://pubs.acs.org.
Experimental Section
(1S,2S,6S,7R,8S,9S)-3-Aza-3-benzyl-8,9-bisbenzyloxy-4,10-
dioxatricyclo[5.2.1.02,6]decane (9). N-Benzyl hydroxylamine
2422 J. Org. Chem. Vol. 75, No. 7, 2010