Synthesis of 1′,2′-cis-Nucleoside analogues: Evidence of stereoelectronic control for SN2 reactions at the anomeric center of furanosides

Article abstract of DOI:10.1021/ja104429y

We are reporting a highly diastereoselective route to 1′,2′- cis-nucleoside analogues in the d-ribo, d-lyxo, d-xylo, and d-arabinoside series. Five-membered ring lactols undergo highly selective N-glycosidation reactions in the presence of dimethylboron bromide with different silylated nucleobases. Stereoelectronic control plays a crucial role for the observed induction, and the products are proposed to be formed through S_N2 "exploded" transition states. This approach shows great potential considering its simplicity and selectivity for the synthesis of nucleoside analogues, an important class of molecules in medicinal chemistry.

Full text of DOI:10.1021/ja104429y

Published on Web 08/12/2010
Synthesis of 1′,2′-cis-Nucleoside Analogues: Evidence of
Stereoelectronic Control for SN2 Reactions at the Anomeric
Center of Furanosides
Michel Pre´vost, Olivier St-Jean, and Yvan Guindon*
Institut de Recherches Cliniques de Montre´al (IRCM), Bio-organic Chemistry Laboratory, 110
aVenue des Pins Ouest, Montre´al, Que´bec, Canada H2W 1R7, Department of Chemistry,
UniVersite´ de Montre´al, C.P. 6128, succursale Centre-Ville, Montre´al, Que´bec, Canada H3C
3J7, and Department of Chemistry, McGill UniVersity, 801 Sherbrooke Street West, Montre´al,
Que´bec, Canada H3A 2K6
Received May 21, 2010; E-mail: yvan.guindon@ircm.qc.ca
Abstract: We are reporting a highly diastereoselective route to 1′,2′-cis-nucleoside analogues in the D-ribo,
D-lyxo, D-xylo, and D-arabinoside series. Five-membered ring lactols undergo highly selective N-glycosidation
reactions in the presence of dimethylboron bromide with different silylated nucleobases. Stereoelectronic
control plays a crucial role for the observed induction, and the products are proposed to be formed through
SN2 “exploded” transition states. This approach shows great potential considering its simplicity and selectivity
for the synthesis of nucleoside analogues, an important class of molecules in medicinal chemistry.
Introduction
Endogenous nucleosides play key roles in various aspects of
life, including cell division and viral replication. It is therefore
not surprising that nucleoside analogues have been designed to
interfere with these processes to provide antiviral1 and antitu-
mor2 drugs. Nucleoside analogues also show promising activity
against emerging viruses and novel pharmaceutical targets, such
as protein kinases (e.g., VEGF and FFGR), involved in
deregulated signaling cascades in various diseases.3 The 1′,2′-
cis arrangement between the nucleobase attached at the anomeric
center and the heteroatom at C-2′ is found in numerous
biologically relevant nucleoside analogues (Figure 1).2a,4 The
synthesis of this motif (1′,2′-cis) is still problematic in spite of
past achievements. A general strategy giving access to this class
of molecule, regardless of the relative stereochemistry of
different groups attached at C-3 and C-4 positions of the
furanose ring, would be a great asset for synthetic and medicinal
chemists.
Figure 1. FDA approved drugs with 1′,2′-cis relative stereochemistry.
delivered on the most hindered face of the furanose moiety, a
generally higher energy approach than the one leading to the
1′,2′-trans product.5 The latter can be prepared by taking
advantage of the anchimeric participation of C-2 neighboring
groups (OAc or OBz) to block the syn face of the glycosyl
acceptor.6 The 1′,2′-cis N-glycosidations have been achieved
by SN2 displacements of 1,2-trans-halosugars, but the latter are
not trivial to generate diastereoselectively and useful selectivities
can only be reached by extensive optimization of the reaction
conditions.7 Other strategies to provide 1′,2′-cis N-glycosidation
involve activation of anomeric thioethers8 or carboxyl9 groups
by various Lewis acids (AgOTf, BiBr3, TMSOTf, NBS, NIS,
SiF4). However, to the best of our knowledge, none of these
The current paradigm governing the synthesis of nucleosides
implies that the nucleobase (purine or pyrimidine) be linked to
a glycosyl acceptor in the presence of a Lewis acid. In this
particular strategy, the nucleophilic heterocycle has to be
(1) (a) Howell, H. G.; Brodfuehrer, P. R.; Brundidge, S. P.; Benigni, D. A.;
Sapino, C., Jr. J. Org. Chem. 1988, 53, 85. (b) De Francesco, R.;
Carfi, A. AdV. Drug DeliVery ReV. 2007, 59, 1242. (c) De Clercq, E.
Med. Res. ReV. 2009, 29, 571.
(2) (a) Sabini, E.; Ort, S.; Monnerjahn, C.; Konrad, M.; Lavie, A. Nat.
Struct. Biol. 2003, 10, 513. (b) Otova´, B.; Hrdy´, J.; Votruba, I.; Holy´,
A. Anticancer Res. 2009, 29, 1295. (c) De Clercq, E. Nat. ReV. Drug
DiscoVery 2007, 6, 1001.
(3) (a) Klein, R. S.; Lenzi, M.; Lim, T. H.; Hotchkiss, K. A.; Wilson, P.;
Schwartz, E. L. Biochem. Pharmacol. 2001, 62, 1257. (b) Rumjahn,
S. M.; Yokdang, N.; Baldwin, K. A.; Thai, J.; Buxton, I. L. O. Br. J.
Cancer 2009, 100, 1465.
(5) Zhu, X.; Kawatkar, S.; Rao, Y.; Boons, G. J. J. Am. Chem. Soc. 2006,
128, 11948.
(6) (a) Ness, R. K.; Fletcher, H. G., Jr.; Hudson, C. S J. Am. Chem. Soc.
1950, 72, 2200. (b) Niedballa, U.; Vorbru¨ggen, H. J. Org. Chem. 1974,
39, 3672. (c) Vorbru¨ggen, H.; Ruh-Pohlenz, C. Handbook of Nucleo-
side Synthesis; John Wiley & Sons: New York: 2001.
(4) (a) Chuang, R. Y.; Chuang, L. F. Nature 1976, 260, 549. (b) Bonate,
P. L.; Arthaud, L.; Cantrell, W. R.; Stephenson, K.; Secrist, J. A., III;
Weitman, S. Nat. ReV. Drug DiscoVery 2006, 5, 855.
9
10.1021/ja104429y  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 12433–12439 12433

A R T I C L E S
Pre´vost et al.
Scheme 1. Synthesis of 1′,2′-cis-Nucleosides from Lactols (R1 ) H)
Figure 2. Four D-furanose scaffolds.
protocols give access to high 1′,2′-cis selectivity in the four
Me2BBr16 to generate mixtures of cyclic bromoethers. Subse-
quent addition of persilylated pyrimidine bases (B(TMS)x) led
to highly selective 1′,2′-cis-nucleosides in good yields for all
four D-furanose scaffolds (Figure 2). We are proposing that the
selectivity observed is the consequence of faster stereospecific
displacements of 1,2-trans- versus 1,2-cis-bromoethers, which
are undergoing rapid in situ anomerization (Scheme 1). This
work provides evidence against the general assumption that in
situ anomerization glycosidation strategies28,36 can only lead
to high anomeric selectivity in six-membered rings5,17 and
suggests that stereoelectronic effects are significant for SN2
displacements at the anomeric position of five-membered ring
systems.
D-furanose scaffolds to generate modified nucleosides (Figure
2).
Expanding the scope of our retrospective analysis led us to
consider stereoselective C-, O-, and N-glycosidation reactions
that do not involve neighboring group assistance. These
displacement reactions are generally proposed to proceed
through two limit mechanisms: SN1 mechanisms involving
oxocarbenium intermediates,10 generally accessed by activation
of the anomeric center by strong Lewis acids, or SN2 displace-
ments of an anomeric leaving group (e.g., bromide).1a Other
intermediate scenarios involve solvent-separated ions pairs
(SSIP),11 displacements of contact ions pairs (CIP),12 or SN2-
like mechanisms with oxocarbenium character, as suggested in
the “exploded”13 SN2 transition state model.14 The relative
energy of these transition states is strongly influenced by the
nature of the Lewis acid, the nucleophile, the counterion (i.e.,
leaving group), and the nature of the glycosyl donor.8b,15
Results and Discussion
Investigation of the Activation of Tetrahydrofuran Acetals
with Me2BBr. Initially, we were interested in studying N-
glycosidations of halosugars under kinetic control to develop
experimental conditions leading to high 1′,2′-cis selectivities.
We first studied the feasibility of generating cyclic bromoethers
in a single step from readily available methyl furanosides using
Me2BBr, a Lewis acid that has been reported to convert acetals
and ketals to bromoethers at low temperature (Scheme 1, R1 )
Me).18 This strategy was appealing because it would provide
an efficient alternative to other elegant protocols for the
installation of an activating group at the anomeric position.8,10
In order to determine if the desired bromoether products were
formed after the Lewis acid activation step, trapping experiments
were performed at low temperature using thiols in the presence
of a hindered base. Accordingly, 1,2-trans tetrahydrofuran acetal
1a was treated with Me2BBr at -78 °C. Thiophenol and DIEA
were subsequently added to the reaction mixture at low
temperature. Rather than forming the desired cyclic acetal,
acyclic thioacetal 2 was obtained as a 11:1 (1,2-anti:1,2-syn)
ratio (Table 1, entry 1). When performing the reaction at a higher
temperature (0 °C), the diastereoselectivity dropped to 1.3:1 (1,2-
anti:1,2-syn) and thiofuranosides 3 was observed in trace
amounts (Table 1, entry 2). The acyclic thioacetal was also
obtained from tetrahydrofuran 1,2-cis-acetal 1b, but with
opposite diastereoselectivity in a 1:10 (1,2-anti:1,2-syn) ratio
(Table 1, entry 3). Our preliminary results at low temperature
clearly indicated a selective activation of the ring oxygen by
the Lewis acid with subsequent cleavage of the C-O endocyclic
bond. In order to favor a chemoselective complexation of the
The general and highly diastereoselective N-glycosidation
strategy reported herein gives access to 1′,2′-cis-nucleoside
analogues. Five-membered ring lactols (R1 ) H, Scheme 1),
with an electronegative substituent at C-2, were treated with
(7) (a) Glaudemans, C. P. J.; Fletcher, H. G, Jr. J. Org. Chem. 1963, 28,
3004. (b) Shen, T. Y.; Lewis, H. M.; Ruyle, W. V. J. Org. Chem.
1965, 835. (c) Keller, F.; Tyrrill, A. R. J. Org. Chem. 1966, 31, 1289.
(d) Stevens, J. D.; Ness, R. K.; Fletcher, G., H J. Org. Chem. 1968,
33, 1806. (e) Noyori, R.; Hayashi, M. Chem. Lett. 1987, 16, 57. (f)
Kadir, K.; Mackenzie, G.; Shaw, G. J. Chem. Soc., Perkin Trans. 1
1980, 2304.
(8) (a) Nokami, J.; Osafune, M.; Ito, Y.; Miyake, F.; Sumida, S.-i.; Torii,
S. Chem. Lett. 1999, 1053. (b) Crich, D.; Pedersen, C. M.; Bowers,
A. A.; Wink, D. J. J. Org. Chem. 2007, 72, 1553.
(9) (a) Mukaiyama, T.; Yamada, M.; Suda, S.; Yokomizo, Y.; Kobayashi,
K. Chem. Lett. 1992, 1401. (b) Hanessian, S.; Conde, J. J.; Lou, B.
Tetrahedron Lett. 1995, 36, 5865. (c) Kim, K. S.; Fulse, D. B.; Baek,
J. Y.; Lee, B.-Y.; Jeon, H. B. J. Am. Chem. Soc. 2008, 130, 8537.
(10) (a) Sammakia, T.; Smith, R. S. J. Am. Chem. Soc. 1994, 116, 7915.
(b) Larsen, C. H.; Ridgway, B. H.; Shaw, J. T.; Smith, D. M.; Woerpel,
K. A. J. Am. Chem. Soc. 2005, 127, 10879.
(11) Streitwieser, A.; Jayasree, E. G. J. Org. Chem. 2007, 72, 1785.
(12) (a) Zhang, Y.; Reynolds, N. T.; Manju, K.; Rovis, T. J. Am. Chem.
Soc. 2002, 124, 9720. (b) Shenoy, S. R.; Woerpel, K. A. Org. Lett.
2005, 7, 1157. (c) Crich, D.; Li, L. J. Org. Chem. 2007, 72, 1681.
(13) Concerted associative transition states with long breaking and forming
bonds. The anomeric center is considered to be pseudo-sp2-hybridized
with participation of the endocyclic oxygen to stabilize the developing
positive charge. In these transition states, bond forming and bond
breaking are considered to be occurring in a symmetrical manner.
(14) (a) Crich, D.; Chandrasekera, N. S. Angew. Chem., Int. Ed. 2004, 43,
5386. (b) Krumper, J. R.; Salamant, W. A.; Woerpel, K. A. J. Org.
Chem. 2009, 74, 8039. (c) Walvoort, M. T. C.; Lodder, G.; Mazurek,
J.; Overkleeft, H. S.; Code´e, J. D. C.; van der Marel, G. A. J. Am.
Chem. Soc. 2009, 131, 12080.
(16) Other boron activating agents, such as c-Hex2BI, were considered.
Similar trends were observed with compound 15, but the scope of
this activating agent remains to be investigated.
(15) (a) Ichikawa, S.; Hayashi, R.; Hirano, S.; Matsuda, A. Org. Lett. 2008,
10, 5107. (b) Yang, M. T.; Woerpel, K. A. J. Org. Chem. 2009, 74,
545. (c) Dinkelaar, J.; de Jong, A. R.; van Meer, R.; Somers, M.;
Lodder, G.; Overkleeft, H. S.; Code´e, J. D. C.; van der Marel, G. A.
J. Org. Chem. 2009, 74, 4982.
(17) (a) Ellervik, U.; Magnusson, G. J. Am. Chem. Soc. 1994, 116, 2340.
(18) (a) Guindon, Y.; Ogilvie, W. W.; Bordeleau, J.; Cui, W. L.; Durkin,
K.; Gorys, V.; Juteau, H.; Lemieux, R.; Liotta, D.; Simoneau, B.;
Yoakim, C. J. Am. Chem. Soc. 2003, 125, 428. (b) Morton, H. E.;
Guindon, Y. J. Org. Chem. 1985, 50, 5379.
9
12434 J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010

SN2 Reactions at the Anomeric Center of Furanosides
A R T I C L E S
Table 1. Formation of Acyclic and Cyclic Thioacetals
Figure 3. Endocyclic and exocyclic activation pathways of pseudoaxial
or pseudoequatorial methyl furanosides.
the anomeric center are properly aligned for endo- and exo-
anomeric effects to take place (Figure 3).17,22 Hence, both
oxygens should display lower basicity because of the electron
delocalization toward the anomeric center.22 In contrast, the
pseudoequatorial conformation only allows donation from the
glycosidic oxygen (Figure 3), as the unshared pair of electrons
on the ring oxygen are not properly oriented to interact with
the glycosidic σ*C-O orbital. Therefore, by only considering
stereoelectronic factors in the analysis, it is expected that the
ring oxygen in the pseudoequatorial conformation is the more
basic. Formation of acyclic products would consequently occur
selectively through C. Given that five-membered rings exist in
different conformations that are close in energy and can rapidly
equilibrate, the more reactive conformation could be easily
accessed, even if it is not the most thermodynamically favored
in the ground state.23 Interestingly, ring-opening studies of
methyl pyranosides, which generally have a single low-energy
chair conformation,23 indicated that the R-anomers were sig-
nificantly less reactive than ꢀ-anomers at low temperatures.21
This scenario would, however, only be valid at low temperatures
for furanosides, in view of the fact that at 0 °C the 1,2-anti
diastereoselectivity of the acyclic products was lost and that
cyclic thioacetals 3 were formed in low yield (Table 1, entry
2). This is suggestive that at higher temperature the reaction
becomes reversible and could consequently provide the ther-
modynamically preferred cyclic thioacetal. The observed high
diastereoselectivities at low temperature are also of particular
interest (Table 1, entries 1 and 3) because they could be the
result of two consecutive substitution reactions occurring with
SN2-like character.
a With 2.0 equiv of Me2BBr at different T (°C), then 3.0 equiv of
DIEA and 2.5 equiv of PhSH. b Ratios were determined by 1H NMR
spectroscopic analysis of unpurified products. c Reaction stirred for 3 h,
before the addition of PhSH.
exocyclic oxygen, we envisioned using lactols as substrates for
these reactions. We were pleased to note that cyclic product 3
was synthesized selectively when lactol 4 was reacted at low
temperature with Me2BBr (Table 1, entry 4). These results are
interesting both at the theoretical and practical levels.
Substitutions of acetals have been shown to involve intermediates
(vide supra) with oxocarbenium character.10 These intermediates
are formed with the stereoelectronic assistance of the acetal oxygen
lone pair that is antiperiplanar to the breaking C-O bond after
Lewis acid activation.19 As illustrated in Figure 3, the exocyclic
oxocarbenium intermediates B and C can be accessed in both the
pseudoaxial and pseudoequatorial conformations, whereas the
endocyclic oxocarbenium intermediate A is only accessible from
the pseudoaxial conformation. Endocyclic or exocyclic C-O bond
cleavage and subsequent nucleophilic attack by a bromide would
generate cyclic bromoethers through A and acyclic bromoethers
through B or C (Figure 3).20,21 Displacement of these bromoethers
by the thiol nucleophile would then give thioacetals 2 or 3 (Table
1, entries 1-3).
The selectivity for the formation of acyclic versus cyclic
products is likely to be determined at the complexation step.21
In the pseudoaxial conformation, the carbon-oxygen bonds at
Further studies will be undertaken to determine if these
reactivity and selectivity trends hold with more substituted
acetals and to establish if the ring-opening reaction occurs
through SN2 displacements. As for the successful realization of
(19) (a) Deslongchamps, P. Stereoelectronic Effects in Organic Chemistry;
Pergamon Press: Oxford, England, 1983. (b) Deslongchamps, P.; Dory,
Y. L.; Li, S. Can. J. Chem. 1994, 72, 2021. (c) Deslongchamps, P.;
Jones, P. G.; Li, S.; Kirby, A. J.; Kuusela, S.; Ma, Y. J. Chem. Soc.,
Perkin Trans. 2 1997, 2621.
(21) (a) Guindon, Y.; Anderson, P. C. Tetrahedron Lett. 1987, 28, 2485.
(b) Guindon, Y.; Anderson, P. C.; Yoakim, C.; Girard, Y.; Berthiaume,
S.; Morton, H. E. Pure Appl. Chem. 1988, 60, 1705. (c) Guindon, Y.;
Bernstein, M. A.; Anderson, P. C. Tetrahedron Lett. 1987, 28, 2225.
(22) Praly, J.-P.; Lemieux, R. U. Can. J. Chem. 1987, 65, 213.
(23) (a) Angyal, S. J. Angew. Chem., Int. Ed. Engl. 1969, 8, 157. (b) Angyal,
S. J. AdV. Carbohydr. Chem. Biochem. 1984, 42, 15.
(20) (a) Olsson, R.; Rundstro¨m, P.; Frejd, T. J. Chem. Soc., Perkin Trans.
1 1998, 785. (b) Olsson, R.; Rundstro¨m, P.; Persson, B.; Frejd, T.
Carbohydr. Res. 1998, 307, 13. (c) Olsson, R.; Berg, U.; Frejd, T.
Tetrahedron 1998, 54, 3935.
9
J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010 12435

A R T I C L E S
Pre´vost et al.
Table 2. N-Glycosidation Induced by Me2BBr
Table 3. N-Glycosidation of C-2-Substituted Lactols
a With 1.8 equiv of Me2BBr at -40 °C, then 30 min at 0 °C, before
b
1
addition of 3.0 equiv of B(TMS)x. Ratios were determined by H NMR
spectroscopic analysis of unpurified products. c Reaction was cooled at
-78 °C prior to the addition of Me2BBr. d Reaction stirred for 16 h.
e Unoptimized yield with silylated 5-fluorouracil.
a With 1.8 equiv of Me2BBr at -40 °C, then 30 min at 0 °C, before
b
1
addition of 3.0 equiv of B(TMS)x. Ratios were determined by H NMR
ratio greater than 20:1 from lactol 12, in spite of having four
substituents on the same face of the ring (Table 2, entry 5).
Stereoselective formation of 1′,2′-cis N-glycosides, regardless
of the C-3 and C-4 relative stereochemistry of the lactols studied,
suggests that the C-2 alkoxy group is the dominant controlling
factor to reach high selectivities. To ascertain this hypothesis,
lactols bearing electronically and sterically different C-2 sub-
stituents were prepared and submitted to our optimized N-
glycosidation conditions (Table 3). With benzyloxy or tert-
butyl(dimethyl)silyloxy group at C-2, high 1′,2′-cis selectivities
were observed (Table 3, entries 1-3). Replacing the alkoxy
group at the C-2 position by a fluorine atom also provided good
1′,2′-cis selectivities (Table 3, entries 4 and 5). Interestingly,
with a methyl substituent at the C-2 position, the selectivity of
the N-glycosidation was completely lost (Table 3, entry 6). High
1′,2′-cis diastereocontrol is therefore only observed with elec-
tronegative atoms (F or O) attached at the C-2 position and is
not influenced by the steric bulk of the alkoxy protecting group.
Investigation of the 1′,2′-cis N-Glycosidation. In order to
understand the origin of the high 1′,2′-cis selectivity, we initiated
various studies to determine the nature of the intermediates
involved in the N-glycosidation reaction. 1H NMR spectroscopic
analysis of solutions of lactols, 30 min after the addition of
Me2BBr, indicated clean formation of cyclic bromoethers
(23-26) with anomeric ratios ranging from 8:1 to 1:1 (1,2-
trans:1,2-cis), as seen in Table 4.
spectroscopic analysis of unpurified products.
the objectives sought after in the present work, these results
indicated that selective formation of cyclic bromoethers under
kinetic conditions can be achieved from lactol derivatives and
that stereoelectronic effects are at work in five-membered rings.
Synthesis of 1′,2′-cis-N-Glycosides from Hemiacetals. The
efficient formation of cyclic thioacetals from lactols, using
Me2BBr, encouraged us to study N-glycosidation reactions with
silylated pyrimidine nucleobases. Solutions of hemiacetals in
CH2Cl2 were cooled to -40 °C, before the addition of Me2BBr
(Table 2). The reaction mixtures were then warmed at 0 °C
and treated with silylated nucleobases. The N-glycosidation was
completed within 2 h in the ribo (5, entries 1 and 2), arabino
(8, entry 3), and xylo (10, entry 4) series; whereas 6 h was
needed to reach full conversion in the lyxo (12, entry 5) series.
This protocol provided 1′,2′-cis-nucleosides in very good
isolated yields from the four D-furanose scaffolds (Table 2).
Both pyrimidine natural nucleobases (thymine and cytosine)
were competent nucleophiles to generate N-glycosides with high
1′,2′-cis selectivity (Table 2, entries 1 and 2).
These results are exciting due to the high 1′,2′-cis diastereo-
control obtained in the four D-furanose scaffolds. It should be
noted that the reaction was very selective even with the most
sterically encumbered glycosyl acceptors. For example, N-
glycoside 13a was obtained in good yield in a diastereomeric
9
12436 J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010

SN2 Reactions at the Anomeric Center of Furanosides
A R T I C L E S
Table 4. Activation of Lactols by Me2BBr
Figure 4. 1H NMR spectra of the anomeric hydrogens resonance region
during N-glycosidation of D-ribofuranoside 5 in CDCl3 (0.1 M), where ti is
the time elapsed after the addition of 1.8 equiv of Me2BBr and tii is the
time elapsed after the addition 3.0 equiv of T(TMS)2.
Scheme 2. N-Glycosidation of Acyclic Acetals
a Conditions: Me2BBr (1.8 equiv), CDCl3 (0.1 M). b Ratios were
1
determined by H NMR spectroscopic analysis of the reaction mixtures.
1
The N-glycosidation reactions were then monitored by H
NMR spectroscopy. The 1H NMR resonance of the C-1
anomeric hydrogens of the species present during the N-
glycosidation of 5 (ribo scaffold) are shown in Figure 4.
Addition of Me2BBr provided bromoethers 23a and 23b in less
than 5 min (ti ) 5 min, Figure 4). Silylated thymine T(TMS)2
glycosidic bonds. Although the 1′,2′-cis selectivity is controlled
by the presence of an electronegative substituent (-F, -OR)
at the C-2 position, an N-glycosidation reaction performed on
acyclic acetal 27 was observed to be poorly selective under our
optimized conditions (Scheme 2). The conformational bias
imposed by the ring system was therefore considered in the
analysis. From a mechanistic standpoint, these results will be
rationalized in light of previous findings and hypotheses.
1
was then added at 25 °C, and H NMR spectra were collected
at different reaction times (tii). N-Glycoside 6a (δ 6.26 ppm)
was formed with high 1′,2′-cis selectivity; the minor diastere-
omer 6b (δ 6.08 ppm) was not observed (tii ) 5 min). It is
noteworthy that 1′,2′-cis:1′,2′-trans diastereomeric ratios for the
nucleosides remained constant during the reaction (tii ) 5 min
to 2 h). A similar trend was also noted with D-xylose 10 and
D-lyxose 12.24 From these experiments, one can conclude that
high levels of 1′,2′-cis N-glycosidation selectivity are not the
result of selective formations of 1,2-trans-bromoethers undergo-
ing a SN2-like displacement.
Halosugars have been extensively studied as glycosyl donors,
and different mechanisms have been proposed to explain the
1,2-cis selectivity obtained for various C-, N-, and O-glycosi-
dation reactions.25 Classical SN2 substitution (D)26 and free
oxocarbenium (SSIP, G)10b transition state models often failed
to predict the anomeric selectivities for displacements of
anomeric halides by nucleophiles requiring other intermediate
mechanisms to be invoked, such as “exploded” SN2 transition
Mechanisms and Rationale
Synthesis of 1′,2′-cis-N-Glycosides from Lactols. As reported
herein, we have developed a general strategy to synthesize
modified 1′,2′-cis-nucleoside analogues in the four D-furanose
scaffolds (Table 2). This stereoselective N-glycosidation was
shown to involve two consecutive reactions under kinetic
control. Different lactols were reacted with Me2BBr to give
anomeric mixtures of furanosyl bromides, which were subse-
quently displaced by silylated nucleobases to form the C-N
(25) (a) Lemieux, R. U.; Cromer, R.; Spohr, U. Can. J. Chem. 1988, 66,
3083. (b) Gravier-Pelletier, C.; Ginisty, M.; Le Merrer, Y. Tetrahedron:
Asymmetry 2004, 15, 189. (c) Cociorva, O. M.; Lowary, T. L.
Tetrahedron 2004, 60, 1481. (d) El-Badri, M. H.; Willenbring, D.;
Tantillo, D. J.; Gervay-Hague, J. J. Org. Chem. 2007, 72, 4663. (e)
Pedersen, C. M.; Nordstrøm, L. U.; Bols, M. J. Am. Chem. Soc. 2007,
129, 9222. (f) Ko, Y.-J.; Shim, S.-B.; Shin, J.-H. Org. Lett. 2009, 11,
609.
(26) Hubbard, A. J.; Jones, A. S.; Walker, R. T. Nucleic Acids Res. 1984,
12, 6827.
(24) Results are provided as Supporting Information.
9
J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010 12437

A R T I C L E S
Pre´vost et al.
cleophilic additions of allyltrimethylsilane or silylated enol
ethers to anomeric acetates activated with strong Lewis acids
(TMSOTf, SnBr4, SnCl4, BF3 ·OEt2).10b,31 Excellent diastereo-
selectivity was noted from D-ribose derivatives. This suggested
stabilization of the five-membered oxocarbenium conformations
with the C-3 alkoxy group in the pseudoaxial position to
maximize electrostatic stabilization and, to a lesser extent, with
the C-2 alkoxy group occupying the pseudoequatorial position.
The nucleophilic attack was proposed to occur from the inside
face of low-energy envelope conformers (inside attack) to avoid
eclipsing interactions that are encountered when attacking from
the outside face (outside attack).30 Importantly, a complete loss
of the diastereoselectivity was noted in the D-arabinose series.
In this case, the C-3 and the C-2 alkoxy groups have opposite
directing effects. These results were reproduced within a window
of nucleophilicity parameters using various reagents.32
Figure 5. Different transition states for glycosidation reactions depicted
in the ribose scaffold (CIP, contact ion pair; SSIP, solvent-separated ion
pair; Nu, nucleobases).
Although the different models described above cannot be
invoked to rationalize our results, they provide key elements
for an interpretation of the N-glycosidation strategy presented
in this work. As stated before, the selectivity observed was
clearly linked to the configuration and electronic properties of
the substituent at C-2 and not that of C-3. No difference of
diastereoselectivity was noticed (vide supra) between the
arabino- and the ribo-pentose scaffolds, and control reactions
using allyltrimethylsilane instead of silylated nucleobases, with
Me2BBr as the activating agent, did not lead to the formation
of C-glycosidation adducts.33 These results indicate that free
oxocarbenium ions are not involved in our reaction conditions,
a conclusion in good agreement with reported kinetic evidence
suggesting that free oxocarbenium intermediates are too short-
lived in the presence of strong nucleophiles, such as bromide
ions, to react diastereoselectively.32b,34 The stereoselective
synthesis of bromofuranoside and their resulting contact ion pair
F cannot be proposed either since the bromoether has been
obtained as an anomeric mixture in all cases (Table 4).9b,35
states E14c,27 and contact ion pair (CIP) F9b,35 (Figure 5). Which
of these transitions states is the lowest in energy in a given
reaction? The answer to this question varies depending on the
nature of the nucleophile, proximal substituents, and the
conditions used (solvent, activating agents).
Lemieux, in his seminal papers, has suggested that the
enhanced reactivity of bromopyranosides, relative to bromocy-
clohexanes, could be explained by donation of a lone pair of
the endocyclic oxygen to the anomeric center to provide reactive
ion pair intermediates.28 These intermediates were proposed to
be involved in the rapid anomerization of the bromopyranosides,
and the endo-anomeric effect was postulated to bias their
reactivities toward O-nucleophiles to give R-products selectively.
Interestingly, in situ anomerization involving ion pairs (F) has
only been invoked to explain erosion of selectivities in the
displacement of 1,2-trans-bromofuranosides.29 This lack of
selectivity was generally attributed to the small energy gap
between the different possible reactive five-membered ring
conformers and to the fact that anomeric effects are weak in
these ring systems.17,26,29a
Crich has offered interesting contributions in his studies of
O-glycosidation reactions.8b,14a,30 A contact ion pair mechanism
was proposed to lead to the ꢀ-glycoside product, while a solvent-
separated ion pair (SSIP in G, Figure 5) was invoked for the
formation of the R-glycoside.14a,27 In the remarkable ꢀ-glycosi-
dation of mannoside derivatives, the presence of a benzylidene
was suggested to destabilize the formation of the oxocarbenium
intermediate (G, Figure 5).12c For steric and electronic reasons,
the equilibrium was shifted toward a covalently bound anomeric
trifluoromethane sulfonate. Contact ion pairs and exploded SN2
transitions states were proposed to govern the nucleophilic
displacement of this reactive intermediate (E and F, Figure 5).
Not less remarkable are the contributions from Woerpel that
demonstrated the importance of stereoelectronic effects of
proximal substituents in C-glycosidation reactions. Five-
membered oxocarbenium intermediates were invoked for nu-
The mechanistic hypothesis that is suggested herein does,
however, relate to the in situ anomerization, but in contrast to
the work of Lemieux and others, the selectivities observed are
not proposed to originate mainly from anomeric effects.36 The
rate difference for the SN2 displacement of anomeric 1,2-cis-
or 1,2-trans-bromides would arise from a stereoelectronic effect
involving the C-2 substituent in 1′,2′-cis and 1′,2′-trans predic-
tive exploded SN2 transition states (Figure 6).9 In these transition
states, a developing sp2 character between the endocyclic oxygen
and the anomeric center would stabilize the developing positive
charge at C-1 (Figure 6).37 Given that the resulting intermediates
display oxocarbenium character in five-membered rings, they
are likely to adopt a low-energy envelop conformation, where
(31) (a) Larsen, C. H.; Ridgway, B. H.; Shaw, J. T.; Woerpel, K. A. J. Am.
Chem. Soc. 1999, 121, 12208. (b) Smith, D. M.; Woerpel, K. A. Org.
Lett. 2004, 6, 2063.
(32) (a) Shenoy, S. R.; Smith, D. M.; Woerpel, K. A. J. Am. Chem. Soc.
2006, 128, 8671. (b) Krumper, J. R.; Salamant, W. A.; Woerpel, K. A.
Org. Lett. 2008, 10, 4907. (c) Correlation between selectivities and
nucleophilicities were recently demonstrated in the substitution of
tetrahydropyran acetals; see ref 14b.
(27) Kim, K. S.; Fulse, D. B.; Baek, J. Y.; Lee, B.-Y.; Jeon, H. B. J. Am.
Chem. Soc. 2008, 130, 8537.
(28) Lemieux, R. U.; Hendriks, K. B.; Stick, R. V.; James, K. J. Am. Chem.
Soc. 1975, 97, 4056.
(33) The lactols along with its decomposition products were recovered intact
after a prolonged reaction time of 16 h at room temperature.
(34) (a) Jencks, W. P. Acc. Chem. Res. 1980, 13, 161. (b) Jencks, W. P.
Chem. Soc. ReV. 1981, 10, 345.
(29) (a) Seela, F.; Winkeler, H. D. J. Org. Chem. 1982, 47, 226. (b) Tann,
C. H.; Brodfuehrer, P. R.; Brundidge, S. P.; Sapino, C., Jr.; Howell,
H. G. J. Org. Chem. 1985, 50, 3644.
(35) Sugimura, H.; Muramoto, I.; Nakamura, T.; Osumi, K. Chem. Lett.
1993, 22, 169.
(30) (a) Crich, D.; Sharma, I. Org. Lett. 2008, 10, 4731. (b) Crich, D.;
Cai, W. J. Org. Chem. 1999, 64, 4926. (c) Crich, D.; Sun, S. J. Am.
Chem. Soc. 1997, 119, 11217. (d) Crich, D.; Vinogradova, O. J. Am.
Chem. Soc. 2007, 129, 11756.
(36) (a) Lemieux, R. U.; Morgan, A. R. J. Am. Chem. Soc. 1963, 85, 1889.
(b) Lemieux, R. U.; Morgan, A. R. Can. J. Chem. 1965, 43, 2214.
(37) Altona, C. Ph.D. Thesis, University of Leiden, 1964.
9
12438 J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010

SN2 Reactions at the Anomeric Center of Furanosides
A R T I C L E S
the bromides during the N-glycosidation would lead to the
selective formation of 1′,2′-cis products because of the difference
in energy between I and J (Scheme 1 and Figure 6). The
anomerization could be facilitated by bromide anions, which
might come from excess Me2BBr or from the formation of
TMSBr in the course of the reaction, as suggested by
Lemieux.28,36 Since the 1′,2′-cis products were formed selec-
tively regardless of the C-3 or C-4 relative configurations or
substitutions, the proposed mechanistic analysis also rationalizes
the selectivity observed in the four D-furanoside series (Table
2) and for the C-2-substituted lactols (Table 3).
Conclusion
We have described a highly stereoselective N-glycosidation
strategy for the synthesis of modified 1′,2′-cis-nucleosides in
all four D-pentose scaffolds and in C-2-substituted five-
membered ring systems. Even the most sterically encumbered
molecules were formed in good yields with excellent selectivity
by in situ additions of silylated nucleobases to cyclic bromo-
ethers obtained by reacting lactols with Me2BBr. We have
suggested a mechanism involving exploded SN2 transition states
to rationalize 1′,2′-cis selectivity at the N-glycosidation step.
The scope of our methodology is currently being expanded with
various nucleophiles to allow the synthesis of novel nucleoside
analogues and other natural products that could serve as
biological and medicinal tools, in addition to being putative
drugs.
Figure 6. Proposed transition states.
the C-1, C-2, and C-4 carbon atoms are coplanar, as illustrated
in Figure 6.5,10b,38
Addition reactions to 1-bromofuranoside should occur from
the inside face (inside attack) to avoid the development of
eclipsing interactions encountered during the rehybridization
through the outside SN2 attack (Figure 6).38 This suggests that
transition states H and K are higher in energy than I and J
(Figure 6). Qualitatively, the relative energy of I and J can then
be evaluated by considering stereoelectronic factors in the
analysis. Transition state I, involving the 1,2-cis-bromide, is
likely to be destabilized by a repulsive electronic interaction
between two eclipsed electronegative atoms (OBn and Br).10b,31
These interactions are not present for the 1,2-anti-bromide in
transition state J, which should also benefit from the σ-hyper-
conjugation stabilization, the carbon-hydrogen bond at C-2
being almost orthogonal to the emerging electron-deficient
π-system.39 The transition state J leading to the 1′,2′-cis product
should therefore be of lowest relative energy. In accordance
with the Curtin-Hammet principle,40 a rapid anomerization of
Acknowledgment. Funding for this research has been granted
from Natural Sciences and Engineering Research Council (NSERC)
and FQRNT. We are thankful to Mohammed Bencheqroun for his
assistance in the preparation of this manuscript. We thank Francine
Be´langer-Garie´py for X-ray analysis, and Dr. Denis Faubert for
the HRMS analysis. A National Cancer Institute of Canada (NCIC)
scholarship to M.P. is also gratefully acknowledged.
Supporting Information Available: Detailed experimental
procedures, proof of structures, and X-ray data are available.
This material is available free of charge via the Internet at http://
pubs.acs.org.
JA104429Y
(38) Smith, D. M.; Tran, M. B.; Woerpel, K. A. J. Am. Chem. Soc. 2003,
125, 14149.
(40) (a) Seeman, J. I. Chem. ReV. 1983, 83, 83. (b) Seeman, J. I. J. Chem.
Educ. 1986, 63, 42. (c) Curtin, D. Y. Rec. Chem. Prog. 1954, 15,
111.
(39) Alabugin, I. V.; Zeidan, T. A. J. Am. Chem. Soc. 2002, 124, 3175.
9
J. AM. CHEM. SOC. VOL. 132, NO. 35, 2010 12439