Nucleoside Analogues: Synthesis from Strained Rings

Article abstract of DOI:10.1002/ijch.201500090

Nucleoside analogues are widely employed as bioactive compounds against cancer and viral infections. Consequently, it is important to develop efficient synthetic methods to access them with high efficiency and structural diversity. Herein, we present a full account of our work on the synthesis of nucleoside analogues via annulations of donor acceptor aminocyclopropanes and aminocyclobutanes. Thymine- and uracil-derived diester cyclopropanes were accessed from the corresponding nucleobases via vinylation and rhodium-catalyzed cyclopropanation, and were then used in (3+2) annulations with aldehydes, ketones and enol ethers. The obtained analogues could be transformed into important hydroxymethyl derivatives. Thymine and fluoro-uracil-derived diester cyclobutanes obtained from the nucleobases via vinylation and (2+2) cycloaddition could also be used in a (4+2) annulation with aldehydes. Finally, purine-derived diester cyclopropanes could be accessed using the condensation of nucleobases with chloromethyl ethylidene malonates, but annulation reactions with this class of substrates were not successful.

Full text of DOI:10.1002/ijch.201500090

Full Paper
DOI: 10.1002/ijch.201500090
Nucleoside Analogues: Synthesis from Strained Rings
Sophie Racine, Jꢀrꢀmy Vuilleumier, and Jꢀrꢁme Waser*[a]
Abstract: Nucleoside analogues are widely employed as bio-
active compounds against cancer and viral infections. Con-
sequently, it is important to develop efficient synthetic meth-
ods to access them with high efficiency and structural diver-
sity. Herein, we present a full account of our work on the
synthesis of nucleoside analogues via annulations of donor
acceptor aminocyclopropanes and aminocyclobutanes. Thy-
mine- and uracil-derived diester cyclopropanes were ac-
cessed from the corresponding nucleobases via vinylation
and rhodium-catalyzed cyclopropanation, and were then
used in (3+2) annulations with aldehydes, ketones and
enol ethers. The obtained analogues could be transformed
into important hydroxymethyl derivatives. Thymine and
fluoro-uracil-derived diester cyclobutanes obtained from the
nucleobases via vinylation and (2+2) cycloaddition could
also be used in a (4+2) annulation with aldehydes. Finally,
purine-derived diester cyclopropanes could be accessed
using the condensation of nucleobases with chloromethyl
ethylidene malonates, but annulation reactions with this
class of substrates were not successful.
Keywords: annulation · catalysis · cyclopropanes · nucleosides · synthetic methods
1. Introduction
Nucleoside analogues (Figure 1) are widely employed as
bioactive compounds against a broad diversity of illness-
es, such as HIV and different types of cancer. Neverthe-
less, the emerging resistance towards nucleoside-based
antiviral and anticancer agents is alarming.[1] Therefore,
the development of new bioactive nucleoside analogues is
urgently needed. Nucleosides exhibiting biological activi-
ty have structurally diverse cores, including the most
common tetrahydrofuranyl amines, such as the FDA ap-
proved drugs azidothymidine (1) and didanosine (2), and
more rare six-carbon-based scaffolds, such as the antiviral
compounds anydrohexitol-g (3) and cyclohexenyl-g (4).
Carbonucleoside analogues, such as the antiviral com-
pounds aristeromycin (5), abacavir (6), and cyclohexenyl-
g (4), are also of great interest. In comparison with the
corresponding tetrahydrofuranyl amine derivatives, they
display increased metabolic stability.
Despite the fact that nucleoside analogues are of high
interest and have been extensively studied, there are only
a few methods for accessing them with high diversity and
efficiency.[2] Nucleoside analogues are generally synthe-
sized using the Vorbrꢂggen reaction.[2a] This transforma-
tion consists of the substitution of a leaving group on the
ribose by a nucleobase, providing compounds derived
from natural ribose. This method allows variation of the
nucleobase. However, if unnatural modifications of the
ribose core are desired, multistep procedures are usually
required.[3] A frequently used approach to access carbo-
nucleoside analogues is to use the Vince lactam,[3a,b]
[a] S. Racine, J. Vuilleumier, J. Waser
Ecole polytechnique fꢀdꢀrale de Lausanne
Institut des sciences et ingꢀnierie chimiques
SB ISIC LCSO, EPFL
CH-1015 Lausanne (Switzerland)
Tel.: +41(0)216939388
Fax: +41(0)216939700
e-mail: jerome.waser@epfl.ch
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/ijch.201500090.
Figure 1. Bioactive nucleoside analogues.
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&1
&
ÞÞ
These are not the final page numbers!

Full Paper
which after opening, delivers a cyclopentenylamine pre-
cursor for carbonucleoside derivatives. Further elegant
work to access nucleoside analogues with a cyclopentyl
core has been developed during the last several years,
proceeding via the Pauson-Khand reaction,[3c] ring-closing
methathesis,[3d] or desymetrization of meso compounds,
such as cyclopentadiene or diols.[3e–3g] Although these
novel synthesis strategies increased the diversity of acces-
sible scaffolds, they are often substrate specific and their
scope is typically narrow.
Consequently, we decided to develop a straightforward
synthetic method for the synthesis of nucleoside ana-
logues with broad structural diversity. Our goal was to
take advantage of the strain release upon ring-opening of
small rings, such as cyclopropanes and cyclobutanes.
Indeed 28 kcalmolÀ1 of energy is released during the ring
opening of cyclopropane, which can be used to initiate
chemical transformations. Donor-acceptor (DA) cyclopro-
panes and DA cyclobutanes, in particular, are versatile
substrates, as selective CÀC bond cleavage is possible
under mild conditions.[4] Thanks to the induced bond po-
larization, they can act as formal 1,3-dipoles and are suit-
able substrates for annulation with dipolarophiles, provid-
ing an effective method for the synthesis of highly substi-
tuted five- or six-membered rings. Nitrogen-based DA cy-
clopropanes have been extensively studied in our group.[5]
During these studies, we discovered that the protecting
group on the nitrogen atom is key in the success of the
annulation reactions. In particular, phthalimide-substitut-
ed DA cyclopropanes and cyclobutanes were excellent
precursors for the synthesis of tetrahydrofuranyl and tet-
rahydropyranyl amines and cyclopentyl and cyclohexyl
amines via formal (4+2) and (3+2) annulations
(Scheme 1A). To apply this methodology to the synthesis
of nucleoside analogues, deprotection of the phthalimide
group is required to introduce the nucleobase. Unfortu-
nately, the removal of the phthalimide on tetrahydrofur-
anyl amines was not possible. Furthermore, even if the
deprotection worked for cyclopentyl amines, the poor
convergence of this strategy led us to design a new
method for the synthesis of nucleoside analogues.
The new strategy consisted of having the nucleobase di-
rectly linked to the small ring, serving as the donor group
(Scheme 1B). With thymine and uracil-substituted DA cy-
clopropanes and cyclobutanes, five- or six-membered ring
nucleoside analogues become accessible through (3+2)
or (4+2) annulations with dipolarophiles, such as ke-
tones, aldehydes, or enol ethers, in a single step.[6] Herein
we would like to give a full account of this work, includ-
ing details on the synthesis of the cyclopropanes, reaction
optimization, and scope of the annulation reactions. First
attempts towards the synthesis of purine-based nucleoside
analogues will be also described. At this stage, the nu-
cleoside analogues are only obtained as racemic mixtures,
but this could also be an advantage if both enantiomers
are needed for biological testing.
2. Synthesis of Nucleoside Analogues via
Annulation Reactions
We planned to synthesize the required pyrimidine cyclo-
propanes via cyclopropanation of the corresponding en-
amides. Therefore, a robust synthesis of vinyl pyrimidines
was needed. When we started our work, there was only
a few reports of vinylation methods applied to pyrimidine
nucleobases: N(1) alkylation followed by elimination;[7]
Chan-Lam coupling using trimethoxyvinyl silane as the
vinyl source, along with tetrabutyl ammonium fluoride
(TBAF) and a stoichiometric amount of Cu(II);[8] and nu-
cleophilic attack of TMS-protected thymine on vinyl ace-
tate in the presence of trimethylsilyl trifluoromethane sul-
fonate (TMSOTf).[9] In our hands, the preparation of
N(1) vinyl thymine 10a following these reported proce-
dures failed (Scheme 2). With dibromoethane, we were
not able to isolate the desired product 8 (Scheme 2A).
This is in agreement with a later report from Nawrot and
coworkers, who showed that alkylation of thymine under
these conditions only led to bicyclic side products.[10]
Even after intensive optimization, the desired products
could be obtained only in 10% yield. Some conversion
was observed in the Chan-Lam coupling, leading to a mix-
ture of N(3) and N(1) vinylated products 9 and 10a,
which were isolated in 24% and 13% yield, respectively,
with N(3) vinylated thymine (9) as major product
(Scheme 2B). In addition to the low yield of the N(1)
vinyl thymine (10a), the separation of the regioisomers
was laborious, prohibiting the use of this reaction for the
preparation of larger amounts of N(1) vinyl thymine
(10a). In conclusion, the reported procedures were not
Scheme 1. Strategy to access nucleoside analogues from nitrogen-
substituted strained rings.
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&2
&
ÞÞ
These are not the final page numbers!

Full Paper
chloroethane (DCE), chloroform, methanol (MeOH),
ethanol (EtOH), acetonitrile (ACN), tetrahydrofuran
(THF), dioxane, dimethylformamide (DMF), and water.
The solubility of thymine (7a) was highest in hot water
and hot mixtures of alcohol (MeOH or EtOH) and water
(608C), followed by dioxane and DMF (both at 608C), in
which no complete dissolution was observed. Attempts at
palladium coupling in dioxane were not successful, but
better results were obtained in DMF. Na2PdCl4 or PdCl2,
with or without LiCl, were not suitable catalysts for the
vinylation of thymine (7a) (Table 1, entries 1 to 3): only
traces of N(1) vinyl thymine (10a) were detected (a mix-
ture of N(3) and N(1) vinyl thymine 9 and 10a is ob-
served with PdCl2 (entry 1)). With Pd(OAc)2, only 3%
conversion was observed, but selectively to N(1) vinylthy-
mine (10a) (entry 4). No dependence on the equivalents
of vinyl acetate (11) used was observed (entries 5 and 6).
Further investigation with increased amounts of
Pd(OAc)2 showed that higher loading in palladium led to
better conversion (entries 7 and 8). Inspired by the work
of Toti et al.[9] and Nawrot et al.,[10] who reported the alky-
lation or vinylation of thymine (7a) in the presence of
TMSOTf, we performed the reaction with a stoichiometric
amount of TMSOTf. The conversion increased from 3%
to 17% with 4% of Pd(OAc)2 (entry 9). When the
amount of TMSOTf was increased to 2.4 equivalents and
the reaction time was extended to 24 h, the reaction con-
version reached 32% and 51%, respectively (entries 10
and 11). Finally, after adding an additional base (triethyl-
amine), the conversion went to 83% and the isolated
yield reached 79% of the sole N(1) vinyl-thymine (10a)
(entry 12). Fortunately, the vinylation also worked well
Scheme 2. Attempts for the preparation of N(1) vinyl thymine
using reported procedures.
suitable to prepare enough N(1) vinyl thymine (10a) for
the further development of our methodology.
In a precedent report from our group, we described the
preparation of vinyl phthalimide, maleimide, and succini-
mide via palladium cross coupling.[5e] For this transforma-
tion, the palladium(II) sources used were sodium tetra-
chloropalladate or palladium dichloride with lithium chlo-
ride. However, these transformations were performed in
vinyl acetate (11) as the solvent, in which thymine (7a) is
not soluble. Solubility experiments with different cosol-
vents were conducted with dichloromethane (DCM), di-
Table 1. Screening of reaction conditions for N(1) vinylation of thymine (7a).
Entry
Catalyst
Loading[a]
Additive[b]
Equiv. 11
Time
Ratio[c]
1
2
3
4
5
6
7
8
9
PdCl2
Na2Pd2Cl4
PdCl2
4
4
4
4
4
4
7
20
4
4
–
–
LiCl
–
–
–
–
–
TMSOTf
TMSOTf[e]
TMSOTf[e]
TMSOTf[e]
1.2
1.2
1.2
1.2
1.6
2
1.2
1.2
1.2
1.2
1.2
1.2
10 h
10 h
24 h
24 h
10 h
10 h
10 h
10 h
10 h
10 h
24 h
24 h
<2%[d]
<2%
<2%
3%
3%
3%
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
6%
11%
17%
32%
51%
83%
10
11
12
4
4
[e]
NEt3
Reaction conditions: Thymine (7a) (0.1 g, 0.8 mmol), palladium catalyst, and additives in DMF (2 mL) were stirred in a flame-dried flask
under nitrogen, and heated to 708C. [a] In mol%. [b] 1.2 equiv. [c] NMR ratio (10a/(7a+10a)) of the integration of the C=CÀH thymine sig-
nals. [d] Mixture of N(1) and N(3) vinyl thymine. [e] 2.4 equiv. additive was used.
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&3
&
ÞÞ
These are not the final page numbers!

Full Paper
on uracil (7b) and 5-fluoro-uracil (7c), with complete re-
gioselectivity and respectively 69% and 57% isolated
yield (products 10b and 10c).
To access the desired cyclopropanes, protection of the
imide nitrogen was performed, followed by cyclopropana-
tion using the corresponding diazomalonate and the rho-
dium Espino catalyst developed Du Bois and coworkers
(Scheme 3).[11] For thymine (7a) and uracil (7b), a tert-bu-
Scheme 4. Unexpected formation of pyridone cyclopropane 14
and tentative mechanism. Reaction conditions: 0.5 mol% Rh2(esp)2,
1.2 equiv. diazodimethylmalonate, CH2Cl2. Yields: 13b: 28% and
14: 64%.[12]
Scheme 3. Synthesis of cyclopropanes 13. Reaction conditions: a)
4 mol% Pd(OAc)2, vinyl acetate, TMSOTf, NEt3, 708C, DMF; b)
Boc2O, DMAP, CH2Cl2 or BzCl DMAP, NEt3; and c) 0.02 mol%
Rh2(esp)2, diazodimethylmalonate, CH2Cl2.
Table 2. Optimization of the (2+2) reaction.
toxycarbonyl protecting group was used. When protection
of fluoro-uracil (7c) was attempted, a mixture of carbo-
nates and carbamate was obtained in poor yield. In this
case, a benzoyl group could be installed selectively on the
nitrogen. The three pyrimidine cyclopropanes 13a, 13b,
and 13c were obtained in good overall yields.
Surprisingly, when the cyclopropanation of vinyl-uracil
12b was performed with 1.2 equivalents instead of an
equimolar amount of the diazomalonate, a new pyrimi-
dine cyclopropane was isolated, containing two malonate
moieties with an intact thymine double bond (Scheme 4).
After extensive spectroscopic investigations, we identified
this compound as 14. The formation of cyclopropane 14
can be tentatively explained by the unusual formation of
epoxide I, instead of cyclopropanation of the thymine
double bond through the reaction with one additional
equivalent of diazomalonate, mediated by the rhodium
Espino catalyst. Then, an intramolecular rearrangement
via opening of the epoxide, followed by the migration of
the Boc group to the oxygen atom would afford pyridone
cyclopropane 14.
Entry
LA
Ratio[a]
1
2
3
4
5
6
7
8
9
FeCl3·Al2O3
Zn(OTf)2
AuCl3
Sc(OTf)3
Sn(OTf)2
Hf(OTf)4
TiCl4
Cu(OTf)2
Yb(OTf)3
In(OTf)3
FeCl3·Al2O3
13%
<2%
<2%
<2%
<2%
dec.
dec.
5%
5%
10
11
10%
35%[b]
Reaction conditions: Vinyl thymine 10a (50 mg, 0.33 mmol),
2 equiv. methylidene malonate 15, dry CH2Cl2, 0.3 M, stirred under
N2. [a] NMR ratio (16a/(10a+16a)) of the integration of the C=CÀ
H signals. [b] 24 hour reaction.
Starting again from vinyl thymine 10a, a (2+2) cyclo-
addition gave access to the thymine DA cyclobutane 16a.
As our groupꢃs reported procedure to synthesize thymine
cyclobutanes through (2+2) cycloaddition using FeCl3 on
aluminium oxide as the catalyst[5i] gave low yield (Table 2,
entry 1), other Lewis acids were screened. Low conver-
sion was obtained with Zn(OTf)2, AuCl3, Sn(OTf)2, and
Sc(OTf)3 (entries 2 to 5). Complete decomposition of the
starting material was observed when Hf(OTf)4 and TiCl4
were used (entries 6 and 7). With In(OTf)3, Cu(OTf)2,
and Yb(OTf)3 as Lewis-acid catalysts, the reaction exhib-
ited low conversion (entries 8 to 10). Nevertheless, using
FeCl3 on aluminium oxide, the conversion was increased
from 13% after 14 hours to 35% after 24 hours (entries 1
and 11). The reaction was shut down when other solvents
were used, such as MeOH, DMF, dioxane, CCl4, or a mix-
ture of DCM/MeOH. ACN allowed the reaction to pro-
ceed, but with lower conversion. Using DCE as the sol-
vent gave roughly similar conversion to the DA cyclobu-
tane 16a.
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&4
&
ÞÞ
These are not the final page numbers!

Full Paper
When the reaction was conducted with freshly pre-
pared FeCl3·Al2O3 and methylidene malonate on a prepa-
rative scale, the isolated yield of 16a reached 43%
(Scheme 5). However, (4+2) annulation was not success-
Scheme 6. Preparation of 5-fluoro uracil cyclobutane 18b.
tions, catalyzed by Lewis acids with ketones, aldehydes,
and enol ethers as partners, to access 5- and 6-membered
nucleoside analogues. Using the conditions described for
the (3+2) reaction with aldehydes and phthalimide-sub-
stituted DA cyclopropanes, the desired five-membered
ring 22a was isolated in low yield with the thymine-substi-
tuted DA cyclopropane 13a (Table 3, entry 1). When the
Lewis acid was changed from FeCl3·Al2O3 to In(OTf)3,
the isolated yield increased to 86% (entry 2). Interesting-
ly, with Hf(OTf)4 as the catalyst, the yield was moderate,
but gave direct access to the deprotected product 22b
(entry 3). However, this yield could not be improved. In
our previous work, when ketones and enol ethers were
used as dipolarophiles with phthalimide cyclopropanes,
the best Lewis acid was SnCl4 at low temperatures.[5a,5b]
Using these conditions with thymine cyclopropane 13a
over two days led to low yield with acetophenone (20)
and moderate yield with TIPS-protected acetophenone
(21) (entries 4 and 5). Increasing the temperature to
À208C resulted in complete conversion after 30 minutes
(entries 6 and 7). The two desired nucleoside analogues
Scheme 5. Synthesis of thymine cyclobutanes. Reaction condi-
tions: a) 0.2 equiv. FeCl3·Al2O3, methylidene dimethylmalonate (15),
DCM, 43%; b) Boc2O, DMAP, ACN, 81%; and c) BzCl, DMAP, NEt3,
DCM, 86%.
ful with this substrate (vide infra). Thymine cyclobutane
16a was therefore protected with a tert-butoxy carbonyl
group to avoid side reactions. However, the Boc group
exhibited high lability and was removed in the presence
of Lewis acids. Using benzoyl as a protecting group was
then examined, as it was expected to be more stable than
tert-butoxycarbonyl group. Benzoylation of 16a gave pro-
tected cyclobutane 18a in 86% yield. Using the same re-
action sequence, fluoro-uracil-derived cyclobutane 18b
could be also obtained in good yield (Scheme 6).
Having the pyrimidine-based DA cyclopropanes 13a-
c and DA cyclobutanes 18a, b in hand, we then examined
the corresponding (3+2) and (4+2) formal cycloaddi-
Table 3. Optimization of the (3+2) annulation.
Entry
Partner[a]
LA
Loading[b]
T
Time
Yield[c]
1
2
3
4
5
6
7
19
19
19
20
21
20
21
FeCl3·Al2O3
In(OTf)3
Hf(OTf)4
SnCl4
SnCl4
SnCl4
5
rt
rt
rt
3 days
1 h
1 h
2 days
2 days
30 min
30 min
27%
86%
49%[d]
6%
52%
60%
60%
20
20
5
10
5
À788C
À788C
À208C
À208C
SnCl4
10
[a] Reaction conditions: 1.2 equiv. 19, 20, or 21, dry DCM, 0.3 M, stirred under N2. [b] In mol%. [c] Isolated yield. [d] Yield of the nucleoside
without the protecting group; deprotection occurred under these conditions.
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&5
&
ÞÞ
These are not the final page numbers!

Full Paper
Table 4. Screening of conditions for the (4+2) annulation.
23a and 24a, synthesized from the ketone 20 and the enol
ether 21, respectively, were isolated in 60% yields.
As the Boc group on thymine is thermally labile, an ef-
ficient two-step procedure, consisting of the Lewis acid-
catalyzed (3+2) annulation, followed by filtration (to
remove the Lewis-acid catalyst) and tert-butoxycarbonyl
removal at 708C for 18 h was developed (Scheme 7).
When the two-step procedure was applied to the (3+2)
reaction with benzaldehyde (19), the unprotected nucleo-
side 22b was isolated in 87% yield (Scheme 7A). The
two-step protocol gave better yield of nucleoside ana-
logues in the case of acetophenone (20): 94% yield of
23b was obtained instead of 60% of 23a without depro-
tection (Scheme 7B). Carbonucleoside 24b was isolated in
84% yield using the two-step procedure.
Entry
Lewis acid
Loading
Time
Yield[a]
1
2
3
4
5
6
7
8
9
Hf(OTf)4
Hf(OTf)4
Hf(OTf)4
Hf(OTf)4
Hf(OTf)4
FeCl3·Al2O3
FeCl3·Al2O3
FeCl3·Al2O3
Hf(OTf)4
20 mol%
20 mol%
20 mol%
20 mol%
20 mol%
20 mol%
100 mol%
100 mol%
20 mol%
14 h
51%
76%
30 min
15 min
10 min
5 min
14 h
30 min
15 min
15 min
87%
55%[b]
44%[b]
59%
62%
60%[c]
78%[c]
Reaction conditions: 18a (20 mg, 0.050 mmol), 1.5 equiv. benzalde-
hyde (19), 0.05 M in dry DCM, stirred under N2. Then filtration,
NH4OH, EtOH, 16 hours. [a] NMR yields after isolation due to ben-
zamide poisoning (not removable by column), all (4+2) annula-
tions went to full conversion, unless otherwise stated. [b] (4+2) an-
nulation not complete. [c] 0.3 mmol scale.
performed directly after annulation.[13] As benzoyl depro-
tection was quantitative, the optimization focused on the
conditions for the (3+2) annulation (Table 4). With haf-
nium triflate, the two-step procedure gave moderate yield
after 14 hours (entry 1) due to partial decomposition of
the product 25a. When the reaction time was shortened
to 30 min, the product 25a was isolated in an increased
76% yield (entry 2). The optimum time was found to be
15 minutes (entry 3), as the conversion was complete and
the degradation minimized. If the reaction time was
shortened to 10 and 5 minutes (entries 4 and 5), the reac-
tion was not complete and lower yields were obtained.
The reaction with iron trichloride supported on alumina
went to completion only after 14 hours, providing 25a in
moderate yield (entry 6). Therefore, a stoichiometric
amount of FeCl3·Al2O3 was used to shorten the reaction
time. Complete conversion was reached after 30 min, but
the yield was not improved (entry 7). The same observa-
tion was made when the reaction was performed at
a 0.3 mmol scale. The reaction was stopped after comple-
tion (15 minutes in this case), after filtration and depro-
tection 60% yield was obtained (entry 8).
Finally, the (4+2) annulation was performed on larger
scale with 0.3 mmol of cyclobutane 18a with hafnium tri-
flate, providing the nucleoside analogue 25a in 78% iso-
lated yield (entry 9). Unfortunately, the (4+2) annulation
conducted with TIPS enol ether 21 gave only traces of
product, even after 24 hours. With acetophenone 20, the
desired benzoyl-protected nucleoside was obtained in
54% yield after 24 hours, but the conversion was not
complete.
Scheme 7. Optimized conditions for annulation-deprotection.
The (4 +2) reaction was then investigated using thy-
mine cyclobutane 16a with different Lewis acids
(Zn(OTf)2, AuCl3, Sn(OTf)2, Sc(OTf)3 TiCl4, Hf(OTf)4,
In(OTf)3, Cu(OTf)2, FeCl3·Al2O3, and Yb(OTf)3). No con-
version was observed for most of the Lewis acids, except
with In(OTf)3 and Hf(OTf)4: these catalysts gave access
to tetrahydropyranyl amine in low yield due to degrada-
tion. A screening of Lewis acids was also conducted with
Boc-protected thymine cyclobutane 17, including
Zn(OTf)2, AuCl3, Sn(OTf)2, Sc(OTf)3, TiCl4, SnCl4,
Hf(OTf)4, In(OTf)3, Cu(OTf)2, FeCl3·Al2O3, and
Yb(OTf)3. Encouraging results were obtained with
Hf(OTf)4 and FeCl3·Al2O3: the desired nucleoside ana-
logues were isolated in 59% and 56% yield, respectively.
However, partial loss of the Boc protecting group was
again observed. Therefore, we focused our efforts on
these two Lewis acids with benzoyl-protected thymine cy-
clobutane 18a. Even in this case, a partial benzoyl group
removal was observed, thus deprotection using an ammo-
nium hydroxide solution in ethanol was systematically
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&6
&
ÞÞ
These are not the final page numbers!

Full Paper
Having the optimized conditions in hand for the (3+2)
and the (4+2) annulations, we investigated the scope of
the reaction and its limitations. The reaction was exam-
ined with aldehydes first (Scheme 8A). The nucleoside
22c, synthesized from 4-fluoroacetophenone, was ob-
tained in 79% yield and high diastereoselectivity. The
diastereoselectivity was lower in the case of cinnamalde-
hyde, but product 22d was still obtained in 96% yield.
The (3 +2) reaction worked also with the sterically hin-
dered aliphatic isobutyraldehyde, providing 22e in 75%
yield and complete diastereoselectivity. We then exam-
ined the use of 2-hydroxy acetaldehyde derivatives in the
reaction to provide nucleoside analogues 22f–h with the
hydroxymethyl group often required for the bioactivity of
Scheme 8. Scope of the (3+2) annulation. The products were obtained as a single diastereoisomer, unless stated otherwise.
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&7
&
ÞÞ
These are not the final page numbers!

Full Paper
drugs. Unfortunately, this type of aldehyde was more
prompt to polymerize and no conditions to obtain the nu-
cleoside derivatives could be found.
We then turned to ketones as substrates (Scheme 8B).
Aromatic ketones (acetophenone (20) and 4-fluoro aceto-
phenone) afforded the nucleoside analogues 23b and 23c
in high yield and diastereoselectivity. In the case of an al-
kenyl ketone, the product 23d was isolated in good yield
and poor diastereoselectivity. Aliphatic ketones also
worked well in the annulation reaction: the phenethyl-
substituted nucleoside derivative 23e and the bicyclic nu-
cleoside 23f were obtained in 85% and 97% yield, re-
spectively. Using TBS-protected 2-hydroxyacetacetone,
nucleoside 23g bearing a hydroxymethyl group could be
accessed in 38% yield.
Carbonucleoside derivatives were then synthesized
with enol ethers as dipolarophiles (Scheme 8C). 24b–d
were obtained in high yield and diastereoselectivity, with
enol ethers substituted with aromatic groups. When a di-
benzylester thymine cyclopropane was used with enol
ether 21, the yield reached 95% (product 24c). The styr-
enyl carbonucleoside 24e was isolated with 13 :1 diaste-
reoselectivity in 55% yield. One more stereocenter is
formed using a tri-substituted enol ether and the highly
substituted nucleoside analogue 24f was obtained as
a single diastereoisomer in 79% yield.
The annulation reaction was also conducted with 5-
fluoro uracil cyclopropane 13c and uracil cyclopropane
13b under the conditions optimized for thymine cyclopro-
pane 13a (Scheme 8D). For the deprotection of the ben-
zoyl group on the 5-fluorouracil nucleoside, we again
used NH4OH in EtOH, and the mixture was stirred at
room temperature for 18 hours. Nucleoside analogues
synthesized from benzaldehyde bearing uracil and 5-
fluoro-uracil nucleobases 26a and 26b were obtained in
62% and 72% yield. With acetophenone, the yields re-
mained similar, 76% and 71% for 26c and 26d, respec-
tively. Finally, carbonucleoside derivatives 26e and 26f
were isolated in moderate to high yields (81% and 51%
respectively).
We then examined the scope of the (4+2) annulation
with thymine and 5-fluoro uracil cyclobutanes 18a and
18b and aldehydes (Scheme 9). The (4+2) annulation
worked with electron-neutral benzaldehyde, electron-
poor 4-fluoro-benzaldehyde, and thiophene-2-carbalde-
hyde, providing pure compounds 25a, 25b, and 25c in 70
to 79% yield and 7:1 to >20 :1 diastereoselectivity after
column chromatography and recrystallization to remove
the benzamide side product. Aldehydes bearing electron
rich aromatic groups, such as meta-methoxy benzaldehyde
and para-methoxy benzaldehyde, gave the corresponding
nucleosides 25d and 25e in 76% yield. Styryl-substituted
product 25f could also be obtained, but with a loss of ste-
reoselectivity, as previously noticed in the (3+2) reaction.
5-Fluoro-uracil cyclobutane 18b efficiently provided the
Scheme 9. Scope of the (4+2) annulation. Yield and diastereose-
lectivity are given after column chromatography and recrystalliza-
tion.
desired nucleoside analogue 27, with good diastereoselec-
tivity.
To increase the diversity of our compounds and to pro-
vide potentially bioactive molecules, methods to access
the hydroxymethyl-substituted derivatives were exam-
ined. As discussed above, only a-hydroxylated ketones
could be used directly in the annulation reaction. Howev-
er, TBS deprotection of product 23g led to a complex
mixture, containing a major compound tentatively identi-
fied as lactone 28,[14] and the desired free alcohol could
not be accessed (Scheme 10).
Consequently, we investigated modification of the
esters to access the desired hydroxy methyl group. The
Krapcho decarboxylation[15a,b] was first examined. With
different salts (NaCl, LiCl, LiI, and Li2SO4), and DMF or
DMSO as the solvent at high temperatures (from 808C to
1308C), no mono ester derivatives were isolated. Alterna-
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&8
&
ÞÞ
These are not the final page numbers!

Full Paper
good yield, but with lower diastereoselectivity for 30b
and 30d (Scheme 11B).
When carbonucleoside derivatives were submitted to
the saponification decarboxylation reaction, a mixture of
unsaturated five-membered rings was obtained, due to si-
lanol elimination. Consequently, we turned our attention
to benzyl diester 24c, which can be converted to the diac-
ids under milder, neutral conditions. The benzyl protect-
ing groups were smoothly removed under hydrogenolysis
conditions, then the diacid obtained was heated neat at
808C to allow decarboxylation to proceed providing the
monoacid 31 (Scheme 12). In this case, a controlled elimi-
Scheme 10. Attempted deprotection of nucleoside analogue 23g.
tive methods for ester hydrolysis (Ba(OH)2,[15c]
Me3SnOH[15d]) or reduction (DiBAl-H, LiAlH4, Me(Et-
O)2SiH with Zn(OAc)2[15e]) were then investigated. Un-
fortunately, either no reaction was observed or degrada-
tion occurred. However, in MeOH at 708C with an excess
of KOH, the compound 23b underwent saponification of
both dimethylesters, followed by a diastereoselective de-
carboxylation, cleanly providing the carboxylic acid deriv-
ative 29 in high yield and diastereoselectivity
(Scheme 11A). The use of four equivalents of fine pow-
dered and dry KOH, in combination with dry MeOH,
was important to obtain reproducibly high yields and se-
lectivity. Reduction of the carboxylic acid with the boron
dimethylsulfide complex then gave access to 4’-hydroxy-
methyl nucleoside derivatives. This efficient method was
applied to the synthesis of nucleoside analogues 30a–d in
Scheme 12. Preparation of hydroxymethyl carbonucleoside deriva-
tive 32. a) 10% Pd/C, 1 atm H2, EtOH, 578C, then neat, 808, 64%;
b) 5% Pd/C, 1 atm H2, EtOH, 85%; c) BH3·SMe2, THF; and d) TIPSCl,
DMF, imidazole, 55% over two steps.
nation of the silanol was observed. Finally, olefin hydro-
genation, followed by reduction of the carboxylic acid
and protection of the corresponding alcohol, afforded the
protected hydroxymethyl derivatives 32 in an efficient
manner. Protection prior to isolation was preferred, as
the free alcohol was unstable.
We were also interested in the synthesis of 4’-hydroxy-
methyl nucleoside analogues. Thus, we investigated the
transformation of the styrene group of compounds 22b,
24b, and 30b into the corresponding aldehyde, which
could lead, after reduction, to the targeted alcohol
(Scheme 13). Ozonolysis generally led to degradation of
the starting materials. Dihydroxylation followed by oxida-
tive cleavage, using several reported procedures (cat.
OsO4 with NaIO4 and lutidine, cat. OsO4 with PIDA and
lutidine,[16a] cat. OsO4 with NMO and citric acid, followed
by NaIO4[16b]) was also not successful, leading to insepara-
Scheme 13. Attempted synthesis of 4’-hydroxymethyl nucleoside
Scheme 11. Synthesis of hydroxymethyl nucleoside analogues.
analogues.
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&9
&
ÞÞ
These are not the final page numbers!

Full Paper
ble mixtures of unidentified products, along with decom-
position of the starting materials.[17]
After the synthesis of pyrimidine nucleoside analogues
had been achieved via annulation reactions of strained
rings, we turned our attention to purine nucleoside ana-
logues. As substituents on the cyclopropane, we chose ad-
enine, 6-chloro-2-amino purine and 6-chloro purine. 6-
Chloro-2-amino purine derivatives are interesting, as they
can be used as precursors for both adenine and guanine
derivatives.[18]
To synthesize purine cyclopropanes, we first thought
about using the protocol developed for the pyrimidines.
However, the palladium-catalyzed vinylation was not suc-
cessful with purines. As an alternative approach,
a tandem Michael addition/cyclopropane formation of pu-
rines on chloromethyl-ethylidene malonates was envis-
aged. Indeed, Geen and coworkers reported a single ex-
ample of synthesis of a 6-chloro-2-amino purine-derived
diester cyclopropane using this approach.[19]
As described by Lehnert, chloromethyl-ethylidene mal-
onates can be prepared via a Knoevenagel condensation
of chloroacetaldehyde on malonate derivatives in the
presence of two equivalents of titanium tetrachloride and
four of pyridine.[20] Unfortunately, in our hands, this reac-
tion turned out to give the product in highly fluctuating
yields (from 60% to 0%), as polymerization of the start-
ing material was often observed. Deshmuck and cowork-
ers have described the condensation of trichloroacetalde-
hyde on diethylmalonate using Montmorillonite K-10 in
Ac2O at 1508C for five hours.[21] These conditions were
then examined with chloroacetaldehyde 34, but after only
10 minutes of reaction, complete degradation was ob-
served. When the temperature was lowered to 1108C,
complete conversion was obtained after 1 h to the desired
product 35, together with the aldol dimer 36 of chloroace-
taldehyde. As the two compounds were difficult to sepa-
rate, the reaction conditions were optimized to avoid
dimer formation. At 708C for 7 hours, the reaction was
complete, giving access to the desired ethylidene malo-
nate 35, along with some remaining malonate 33, which
was easily removed by column chromatography
(Scheme 14). The product was isolated up to 61% on
a five gram scale reaction.
Scheme 15. Synthesis of purine cyclopropanes.
with the corresponding pentyl alcohol, followed by the
optimized Knoevenagel reaction.
The obtained alkylidene malonates 35 and 37 were
then submitted to the previously reported conditions for
cyclopropanation (Scheme 15).[19] Purine cyclopropanes
were obtained in moderate to excellent yields, always
with complete regioselectivity. Generally, cyclopropanes
with geminal dimethyl ester 38a, 38b, and 38c were ob-
tained in lower yields. Purification was a major issue, due
to their poor solubility in organic solvents and to their
high polarity. The other three cyclopropanes 38d, 38e,
and 38f with pentyl ester chains were isolated in higher
yields (62 to 93%), as they were well soluble in organic
solvents.
The solubility of dimethyl ester cyclopropanes 38a–
c was poor in all common organic solvents, even if the 6-
chloropurine cyclopropane 38c exhibited a slightly im-
proved solubility. Only partial solubility in hot acetoni-
trile or DMF was observed. In contrast, the purine cyclo-
propanes 38d–f bearing pentyl esters were completely
soluble in dichloromethane at room temperature. To pre-
vent any potential catalyst deactivation or side reaction,
and also to improve its solubility, the free amine of the 6-
chloro-amino-purine cyclopropane 38b was protected
with two tert-butoxycarbonyl groups to give product 39
(Scheme 16). The yield of the protection step is highly de-
pendent on the solvent: in THF or DMSO, the isolated
Furthermore, chloromethyl-ethylidene malonate 37
bearing pentyl esters was synthesized to increase the solu-
bility of the purine cyclopropanes in organic solvents. It
was obtained from the malonic acid through esterification
Scheme 14. Preparation of chloromethyl-ethylidene dimethylmalo-
nate (35).
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&
10&
ÞÞ
These are not the final page numbers!

Full Paper
Scheme 16. Preparation of the protected purine cyclopropane.
Scheme 17. Competition experiment: 0.2 equiv. In(OTf)3, 2.4 equiv.
benzaldehyde (19), DCM.
yields were typically around 40%, whereas in acetonitrile
the product was obtained in 85% yield.
Attempts to protect the free amine with a benzyl group
After 45 minutes of reaction, complete conversion of the
pyrimidine cyclopropane 13c into the corresponding nu-
cleoside analogue 40 was observed, while the purine cy-
clopropane 38e stayed untouched.
via an SN reaction using a phase transfer reagent[22] or
2
through reductive amination[23] failed, as did protection
with dimethyl carbonate.
Cyclopropanes 38a–f were then examined in the (3+2)
annulation with benzaldehyde (19) as the partner and
a broad range of Lewis acids. The first cyclopropane
tested was the adenine cyclopropane 38a bearing methyl
esters. The screening was initially planned in dichlorome-
thane, which had been the solvent of choice for the (3+
2) annulation of thymine-substituted cyclopropanes. Un-
fortunately, adenine cyclopropane 38a displayed extreme-
ly poor solubility in DCM, THF, and ACN. Even with
slightly more soluble purine cyclopropanes 38b no reac-
tion was observed using In(OTf)3 as the catalyst in sever-
al solvents (DCM, DCE, DCM/MeOH, CCl4, THF, ACN,
toluene, MeOH, benzene, DMF, and DMSO). We hy-
pothesized that the lack of reactivity of this cyclopropane
may be the consequence of catalyst poisoning due to the
coordination of the Lewis acid to the free amine of the
substrate. Therefore, the annulation of 38c was examined
with different Lewis acids (Zn(OTf)2, AuCl3, Sn(OTf)2,
Sc(OTf)3, Hf(OTf)4, In(OTf)3, stoichiometric In(OTf)3,
Cu(OTf)2, FeCl3·Al2O3, and Yb(OTf)3). However, also in
this case, no conversion to the desired nucleoside ana-
logues was observed.
No reaction was observed with pentyl esters-substituted
adenine cyclopropane 38d, nor with the other purines cy-
clopropane 38e and 38f, even though the solubility in
DCM at room temperature was high and a wide range of
Lewis acids were examined (Zn(OTf)2, FeCl3, AuCl3,
AuCl, Sn(OTf)2, Sc(OTf)3, Hf(OTf)4, In(OTf)3, Cu(OTf)2,
and Yb(OTf)3). The observed lack of reactivity could also
be due to the steric hindrance of the pentyl side chains
on the esters, preventing chelation of the malonate. How-
ever, when the soluble protected 6-chloro-2-amino purine
dimethylester cyclopropane 39 was submitted to the con-
ditions for (3+2) annulation with indium triflate as the
catalyst in either DCM or THF, only removal of the tert-
butoxycarbonyl was observed.
From this result, we can conclude that purine cyclopro-
panes 38a–f do not coordinate too strongly with the
Lewis acid. However, purine cyclopropanes 38a–f are not
suitable substrates for the (3+2) annulation, probably
due to their different electronic properties leading to
a too low reactivity. In the future, further modulation of
the DA cyclopropane structure will be required to extend
the developed annulation processes to this important
class of substrates.
3. Conclusion
A short synthesis of pyrimidine DA cyclopropanes and
cyclobutanes was designed based on regioselective vinyla-
tion, followed by cyclopropanation or (2+2) cycloaddi-
tion. Purine DA cyclopropanes were prepared from chlor-
omethyl-ethylidene malonates via a tandem Michael-ad-
dition intramolecular substitution sequence. Pyrimidine
nucleosides analogues were accessed from the DA cyclo-
propanes and DA cyclobutanes via (3+2) and (4+2) an-
nulation reactions, opening a new chemical space for fur-
ther biological investigations. Unfortunately, purine DA
cyclopropanes remained unreactive in the (3+2) annula-
tion reactions.
Acknowledgements
The NCCR chemical biology of the Swiss National Sci-
ence Foundation is acknowledged for financial support.
References
[1] a) Modified Nucleosides in Biochemistry, Biotechnology and
Medicine (Ed.: P. Herdewijn), Wiley-VCH, Weinheim, 2008;
b) S. Broder, Antiviral Res. 2010, 85, 1; c) J. C. Martin,
M. J. M. Hitchcock, E. De Clercq, W. H. Prusoff, Antiviral
Res. 2010, 85, 34; d) T. Cihlar, A. S. Ray, Antiviral Res.
2010, 85, 39; e) G. Antonelli, O. Turriziani, Int. J. Antimi-
To completely exclude catalyst poisoning by the sub-
strate, we performed a competition experiment between
pyrimidine cyclopropane 13c and purine analogue 38e
under the conditions for (3+2) annulation (Scheme 17).
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&
11&
ÞÞ
These are not the final page numbers!

Full Paper
crob. Agents 2012, 40, 95; f) L. P. Jordheim, D. Durantel, F.
Zoulim, C. Dumontet, Nat. Rev. Drug Discovery 2013, 12,
447.
Vuilleumier, F. de Nanteuil, J. Waser, Org. Lett. 2015, 17,
1030.
[7] a) P. Ciapetti, M. Taddei, Tetrahedron 1998, 54, 11305;
b) H.-J. Gi, Y. Xiang, R. F. Schinazi, K. Zhao, J. Org. Chem.
1997, 62, 88.
[2] a) U. Niedball, H. Vorbrꢂggen, Angew. Chem. Int. Ed. Engl.
1970, 9, 461; b) G. Shaw, R. N. Warrener, M. H. Maguire,
R. K. Ralph, J. Chem. Soc. 1958, 2294; Reviews: c) M. T.
Crimmins, Tetrahedron 1998, 54, 9229; d) H. Vorbrꢂggen, C.
Ruh-Polenz, Handbook of Nucleoside Synthesis, John Wiley
& Sons, New York, 2001; e) Antiviral Nucleosides: Chiral
Synthesis and Chemotherapy (Ed.: C. K. Chu), Elsevier,
Amsterdam, 2003; f) C. Mathe, C. Perigaud, Eur. J. Org.
Chem. 2008, 1489; g) G. Romeo, U. Chiacchio, A. Corsaro,
P. Merino, Chem. Rev. 2010, 110, 3337; h) Chemical Synthe-
sis of Nucleoside Analogues (Ed.: P. Merino), John Wiley &
Sons, Hoboken NJ, 2013; i) O. Boutureira, M. I. Matheu, Y.
Diaz, S. Castillon, Chem. Soc. Rev. 2013, 42, 5056; j) L.
Scagnelli, M. G. Memeo, S. Carosso, B. Bovio, P. Quadrelli,
Eur. J. Org. Chem. 2013, 3835.
[3] a) S. Daluge, R. Vince, Tetrahedron Lett. 1976, 17, 3005;
b) R. Singh, R. Vince, Chem. Rev. 2012, 112, 4642; c) J.
Velcicky, A. Lanver, J. Lex, A. Prokop, T. Wieder, H. G.
Schmalz, Chem. Eur. J. 2004, 10, 5087; d) W. J. Choi, J. G.
Park, S. J. Yoo, H. O. Kim, H. R. Moon, M. W. Chun, Y. H.
Jung, L. S. Jeong, J. Org. Chem 2001, 66, 6490; e) B. M.
Trost, G. H. Kuo, T. Benneche, J. Am. Chem. Soc. 1988, 110,
621; f) B. M. Trost, L. S. Kallander, J. Org. Chem. 1999, 64,
5427; g) L. F. Tietze, C. Stadler, N. Bçhnke, G. Brasche, A.
Grube, Synlett 2007, 485; h) M. J. Mulvihill, J. L. Gage, M. J.
Miller, J. Org. Chem 1998, 63, 3357; i) G. A. Boyle, C. D.
Edlin, Y. F. Li, D. C. Liotta, G. L. Morgans, C. C. Musonda,
Org. Biomol. Chem. 2012, 10, 1870.
[8] P. Arsenyan, A. Petrenko, E. Paegle, S. Belyakov, Mende-
leev Commun. 2011, 21, 326.
[9] K. S. Toti, M. Derudas, C. McGuigan, J. Balzarini, S. Van
Calenbergh, Eur. J. Med. Chem. 2011, 46, 3704.
[10] B. Nawrot, O. Michalak, S. Olejniczak, M. W. Wieczorek, T.
Lis, W. Stec, Tetrahedron 2001, 57, 3979.
[11] a) C. G. Espino, K. W. Fiori, M. Kim, J. Du Bois, J. Am.
Chem. Soc. 2004, 126, 15378; b) F. Gonzalez-Bobes,
M. D. B. Fenster, S. Kiau, L. Kolla, S. Kolotuchin, M. Sou-
meillant, Adv. Synth. Catal. 2008, 350, 813.
[12] Yield calculated from diazomalonate. If calculated with
uracil as the limiting reagent: 13b: 33%, 14: 38%.
[13] a) T. Kametani, K. Kigasawa, M. Hiiragi, K. Wakisaka, K.
Nakazato, K. Ichikawa, K. Fukawa, O. Irino, N. Nishimura,
T. Okada, J. Med. Chem. 1982, 25, 1219; b) J. Yamashita, M.
Yasumoto, S. Takeda, H. Matsumoto, N. Unemi, J. Med.
Chem. 1989, 32, 136; c) G. Y. Song, V. Paul, H. Choo, J.
Morrey, R. W. Sidwell, R. F. Schinazi, C. K. Chu, J. Med.
Chem. 2001, 44, 3985.
[14] NMR data of compound 28 corresponded to the proposed
structure; however, this compound could not be purified for
complete characterization.
[15] a) A. P. Krapcho, G. A. Glynn, B. J. Grenon, Tetrahedron
Lett. 1967, 8, 215; b) A. P. Krapcho, E. G. E. Jahngen, A. J.
Lovey, F. W. Short, Tetrahedron Lett. 1974, 15, 1091; c) N.
Pꢀrez-Hernꢄndez, M. Febles, C. Pꢀrez, R. Pꢀrez, M. L. Ro-
drꢅguez, C. Foces-Foces, J. D. Martꢅn, J. Org. Chem. 2006,
71, 1139; d) K. C. Nicolaou, A. A. Estrada, M. Zak, S. H.
Lee, B. S. Safina, Angew. Chem. Int. Ed. 2005, 44, 1378;
e) S. Das, K. Mçller, K. Junge, M. Beller, Chem. Eur. J.
2011, 17, 7414.
[16] a) K. C. Nicolaou, V. A. Adsool, C. R. H. Hale, Org. Lett.
2010, 12, 1552; b) P. Dupau, R. Epple, A. Thomas, V. Fokin,
K. Barr. Sharpless, Adv. Synth. Catal. 2002, 344, 421.
[17] Attempts to directly reduce the crude product in one-pot
were not successful.
[4] a) H.-U. Reissig, R. Zimmer, Chem. Rev. 2003, 103, 1151;
b) M. Yu, B. L. Pagenkopf, Tetrahedron 2005, 61, 321;
c) C. A. Carson, M. A. Kerr, Chem. Soc. Rev. 2009, 38,
3051; d) F. De Simone, J. Waser, Synthesis 2009, 3353;
e) T. F. Schneider, J. Kaschel, D. B. Werz, Angew. Chem. Int.
Ed. 2014, 53, 5504.
[5] (3+2) Annulation with imido-cyclopropanes: a) F. de Nan-
teuil, J. Waser, Angew. Chem. Int. Ed. 2011, 50, 12075; b) F.
Benfatti, F. de Nanteuil, J. Waser, Chem. Eur. J. 2012, 18,
4844; c) F. Benfatti, F. de Nanteuil, J. Waser, Org. Lett.
2012, 14, 386; d) F. de Nanteuil, E. Serrano, D. Perrotta, J.
Waser, J. Am. Chem. Soc. 2014, 136, 6239; e) E. Serrano, F.
de Nanteuil, J. Waser, Synlett 2014, 25, 2285; Other reac-
tions with imido-cyclopropanes: f) F. de Nanteuil, J. Loup, J.
Waser, Org. Lett. 2013, 15, 3738; g) R. Tejero, A. Ponce, J.
Adrio, J. C. Carretero, Chem. Commun. 2013, 49, 10406;
h) A. R. Rivero, I. Fernandez, M. A. Sierra, Org. Lett. 2013,
15, 4928; (4+2) Annulations with imido-cyclobutanes: i) F.
de Nanteuil, J. Waser, Angew. Chem. Int. Ed. 2013, 52, 9009;
For a review, see: j) F. de Nanteuil, F. De Simone, R. Frei, F.
Benfatti, E. Serrano, J. Waser, Chem. Commun. 2014, 50,
10912.
[18] L. Beigelman, P. Haeberli, D. Sweedler, A. Karpeisky, Tetra-
hedron 2000, 56, 1047.
[19] G. R. Geen, P. M. Kincey, B. M. Choudary, Tetrahedron
Lett. 1992, 33, 4609.
[20] W. Lehnert, Tetrahedron 1973, 29, 635.
[21] A. R. A. S. Deshmuck, D. G. Panse, B. M. Bhawal, Synth.
Commun. 1999, 10, 1801.
[22] A. R. Maguire, W.-D. Meng, S. M. Roberts, A. J. Willetts, J.
Chem. Soc. Perkin Trans. 1 1993, 1795.
[23] Bristol-Myers Squibb Company, Patent: WO2003/99286 A1,
2003; Location in patent: p. 89.
Received: November 10, 2015
Accepted: December 14, 2015
Published online: && &&, 0000
[6] a) S. Racine, F. de Nanteuil, E. Serrano, J. Waser, Angew.
Chem. Int. Ed. 2014, 53, 8484; b) D. Perrotta, S. Racine, J.
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&
12&
ÞÞ
These are not the final page numbers!

Full Paper
FULL PAPERS
S. Racine, J. Vuilleumier, J. Waser*
&& – &&
Nucleoside Analogues: Synthesis
from Strained Rings
Isr. J. Chem. 0000, 00, 1 – 13
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.ijc.wiley-vch.de
&
13&
ÞÞ
These are not the final page numbers!