Radical cyclization approach to cyclonucleosides

Article abstract of DOI:10.1002/ejoc.200500398

Efficient methodologies based on consecutive radical reactions for the preparation of cyclonucleosides 5 and 13 are reported. The reactions were performed on modified thymidine and 2′-deoxyadenosine substrates using (TMS)₃SiH as the reducing ag

Full text of DOI:10.1002/ejoc.200500398

FULL PAPER
Radical Cyclization Approach to Cyclonucleosides
Maria Luisa Navacchia,[a] Antonio Manetto,[a,b] Pier Carlo Montevecchi,[b] and
Chryssostomos Chatgilialoglu*[a]
Keywords: Cyclization / Kinetics / Nucleosides / Organosilanes / Radical reactions
Efficient methodologies based on consecutive radical reac-
aerobic conditions. The photochemical 5Ј-O-desilylation of
tions for the preparation of cyclonucleosides 5 and 13 are re- 12 has been obtained in 56% yield affording the cyclopurine
ported. The reactions were performed on modified thymidine
and 2Ј-deoxyadenosine substrates using (TMS)3SiH as the
reducing agent. The protected 5Ј-carbaldehyde 3 afforded
13. The reaction mechanisms have been studied in some de-
tail using a variant of the radical clock methodology. The C5Ј
radical 6 or 15, generated by addition of the (TMS)3Si· radical
the cyclonucleoside 4 in 85% yield and in a diastereoiso- to the corresponding aldehyde, undergoes a 6-exo-trig cycli-
meric ratio 4a/4b = 3:7. The mono-desilylation at the 5Ј-O
position of these cyclonucleosides has been successfully
achieved by UV irradiation affording quantitatively deriva-
tives 5a and 5b. The protected 5Ј-carbaldehyde 10 afforded
the cyclonucleoside 12 in 70–75% yield as a single dia-
stereoisomer (5ЈR) either in deoxygenated solution, followed
by in situ oxidation of the reaction mixture by chloranil, or in
zation on the base moiety prior to termination. The rate con-
stants for both 6-exo-trig radical cyclizations have been esti-
mated to be close to 105 s–1 at 86 °C, the cyclization on the
purine moiety being 4–5 times faster than that on the pyrimi-
dine group.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
Introduction
Cyclopurine and cyclopyrimidine lesions are observed
among the decomposition products of DNA, when exposed
to ionising radiations or to some antitumor agents.[1,2] Two
examples of interest to us are 5Ј,6-cyclo-5,6-dihydrothymid-
ine (1) and 5Ј,8-cyclo-2Ј-deoxyadenosine (2). Apart from
the usual glycosidic bond in these moieties there is an ad-
ditional base-sugar linkage between the C6 position of pyr-
imidine or C8 position of purine and the C5Ј position of the
2Ј-deoxyribose. Lesions 1 and 2 may be formed in different
diastereoisomeric forms, differing in the configuration at
the C5Ј position in both cases, as well as in the configura-
tion of C5 and C6 in the pyrimidine derivative. From a
mechanistic point of view, it has been verified that the C5Ј
radical, initially generated by hydrogen abstraction, intra-
molecularly attacks the double bond of the base moiety to
form a cyclonucleotide as the final product, after oxidation
or reduction (Scheme 1).[3,4]
Scheme 1. Examples of 5Ј,6-cyclopyrimidines and 5Ј,8-cyclopurine
lesions.
starting from N6-benzoyl-2Ј-deoxyadenosine in an overall
yield of Ͻ10%.[6] The preparation of the (5ЈR)-isomer of 2
was achieved by two additional steps, involving inversion of
configuration at the C5Ј position.[7]
Synthesis of modified 2Ј-deoxynucleosides containing
specific DNA lesions and their incorporation into a defined
sequence of oligonucleotides has been an outstanding ap-
proach to investigate the biological consequences. Synthetic
oligonucleotides that contain the modified nucleosides 1[5]
or 2[6,7] as well as similar cyclopurine[7] and cyclopyrimi-
dine[8] moieties were also prepared. Recent studies have
shown that the chemical synthesis of these lesions and their
incorporation on specific sites of DNA are of considerable
importance in order to investigate, in detail, the biochemi-
cal and biophysical features of the double helix damage.[9,10]
In our laboratory, a synthetically useful radical cascade
process has been developed that allows the conversion of 8-
bromo-2Ј-deoxyadenosine to 2 in a one-pot procedure.
Using γ irradiation as a source of solvated electrons in
The synthesis of
1 in a diastereoisomeric ratio
(5ЈS,6S,5R)/(5ЈS,6S,5S) = 1:2.7 and in 61% yield was ac-
complished by Bu3Sn·-mediated intramolecular cyclization
of the corresponding 5Ј-carbaldehyde.[5] The (5ЈS)-isomer
of 2 was prepared by the same research group in seven steps
[a] ISOF, Consiglio Nazionale delle Ricerche
Via P. Gobetti 101, 40129 Bologna, Italy
Fax: +39-051-639-8349
E-mail: chrys@isof.cnr.it
[b] Dipartimento di Chimica Organica “A. Mangini”, Università
di Bologna,
Viale Risorgimento 4, 40136 Bologna, Italy
4640
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejoc.200500398
Eur. J. Org. Chem. 2005, 4640–4648

Radical Cyclization Approach to Cyclonucleosides
FULL PAPER
water, a diastereoisomeric ratio (5ЈR)/(5ЈS) = 6:1 in 70%
overall yield (based on the recovered starting bromide) was
obtained,[3] whereas UV irradiation in acetonitrile afforded
a ratio (5ЈR)/(5ЈS) = 1.7:1 in 65% overall yield.[4] However,
attempts to obtain the targeted phosphoramidite required
for the synthesis of oligonucleotides failed due to the diffi-
culties in differentiating the two secondary hydroxyl groups
by the appropriate protective group. In this paper, we de-
scribe the cyclization of 5Ј-carbaldehydes 3 and 10 using
(TMS)3SiH as the mediator[11] under a variety of condi-
tions. Our main objective was to obtain a procedure for the
preparation of some of the diastereoisomers of 1 and 2,
which overcomes the limitations of the existing approaches,
mainly due to low yield multiple-step synthesis, or to dia-
stereomeric mixtures that are difficult to separate, or to dif-
ficulties in differentiating the two secondary hydroxyl
groups by appropriate protective groups. We also report
kinetic studies of C5Ј radical cyclizations using free-radical
clock methodology.[12]
Results and Discussion
Reaction of (TMS)3SiH with 5Ј-Carbaldehyde 3: Alde-
hyde 3 is easily obtained from 3Ј-O-(tert-butyldiphenylsilyl)-
thymidine using Dess–Martin periodinane in anhydrous
CH2Cl2 as described previously.[5,8a] The key cyclization re-
actions were tested employing (TMS)3SiH and AIBN as the
radical initiator. In a typical experiment, a deoxygenated
solution of 3 in benzene (0.01 ) was treated with 5 equiv.
of (TMS)3SiH and 0.2 equiv. AIBN, and the resulting mix-
ture was stirred at 80 °C until the starting aldehyde was
consumed. After evaporation of the solvent, 1H NMR spec-
troscopic analysis showed two diastereoisomers of cyclonu-
cleoside 4a and 4b as the only products (Scheme 2). These
compounds were isolated by flash chromatography in a 25
and 60% yield, respectively, and fully characterized. Sub-
sequently, a 20 m solution of each of them in an 8:3
CH2Cl2/CH3OH mixture was exposed to UV irradiation for
30 min at room temperature using a low-pressure mercury
lamp (5.5 W).[13] The quantitative formation of the mono-
desilylated products 5a and 5b was obtained.
The stereochemistry of the two isomers was deduced
from NMR spectroscopic analysis. The assignment of the
(5ЈS,6S,5S) configuration to the isomer 4a came from (i) the
large J5Ј,6 = 9.2 Hz indicating that H5Ј and H6 are in a
trans-diaxial arrangement, (ii) NOE on H2Ј (2%) and H3Ј
(6%) upon irradiation of H6, and (iii) NOE on H4Ј (8%)
observed only upon irradiation of H5. The assignment of
the (5ЈR,6S,5S) configuration to the isomer 4b came from
(i) the small J5Ј,6 = 2.0 Hz indicating that H5Ј and H6 are
in an equatorial/axial arrangement, (ii) NOEs on H2Ј (2%),
H5Ј (5%) and H6 (4%) upon irradiation of H3Ј, and (iii)
Scheme 2. Product studies of aldehyde 3 with (TMS)3SiH under
free radical conditions and subsequent photochemical deprotec-
tion.
that attacks intramolecularly the double bond of the thy-
mine moiety to give radical 7 (6-exo-trig cyclization). Hy-
drogen abstraction from the silane gives product 4 and
(TMS)3Si· radical, thus completing the cycle of this chain
reaction (Scheme 3). Attempts to obtain the hydrosilylated
product 8, using high concentrations of (TMS)3SiH, did not
work, suggesting that the hydrogen abstraction step is very
slow. Indeed, when the above-described typical experiment
was carried out in the presence of 0.1  PhSH, the 1H
NMR spectrum showed a quantitative formation of the re-
duction product 8, which was isolated by flash chromatog-
raphy in 95% yield.
Under these conditions, thiol acts as an effective hydro-
gen donor and the resulting thiyl radical is able to abstract
hydrogen from the silane thus completing the cycle of this
chain reaction [Equation (1)].[14]
RS· + (TMS)3SiH Ǟ RSH + (TMS)3Si·
(1)
A rate constant kc for the 6-exo-trig cyclization can be
irradiation of H5 caused an enhancement on H6 (4%), indi- obtained, provided that conditions can be found in which
cating that H5 and H6 are in an equatorial/axial and not in the intermediate radical 6 is partitioned between the two
a trans-diaxial arrangement, which could be deduced by the reaction channels (Scheme 3), that is, the reaction with a
J5,6 value of 7.6 Hz.
hydrogen atom donor and the 6-exo-trig cyclization.[12] This
From a mechanistic point of view, the addition of the scenario can be achieved by replacing PhSH with BuSH,
(TMS)3Si· radical to aldehyde 3 affords the C5Ј radical 6 which is a weaker hydrogen donor. A series of experiments
Eur. J. Org. Chem. 2005, 4640–4648
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4641

M. L. Navacchia, A. Manetto, P. C. Montevecchi, C. Chatgilialoglu
FULL PAPER
Table 1. Kinetic data for the reaction of aldehyde 3 with (TMS)3-
SiH/BuSH in benzene.[a]
T [°C]
Initiator
kH/kc [–1][b]
25
55
80
111
142
Et3B
853 160 (3)
330 29 (4)
152 14 (5)
91 5 (4)
tBuONNOtBu
AIBN
PhC(O)OOtBu[c]
tBuOOtBu[d]
35 4 (3)
[a] Conditions: Aldehyde 3 0.01 , (TMS)3SiH 0.05 , BuSH con-
centration in the range 0.6–40 m and initiator 0.2 equiv. [b] Errors
correspond to the standard deviation; the number of experiments
performed for each temperature are given in parentheses. [c] In tol-
uene. [d] In o-xylene.
along with Equation (3) yields log(A/–1 s–1) = 10.0 and Ea
= 8.4 kcalmol–1 for the cyclization reaction in Scheme 3.
The value of kc can be calculated as 7×103 s–1 at 25 °C.[18]
log kH/kc(–1) = –(1.82 0.23) + (6.50 0.37)/θ
(3)
Radical 6 should be nearly planar in the two possible
conformers that give rise to the (5ЈS)- and (5ЈR)-isomers.
We suggest that the two conformers around the C4Ј–C5Ј
bond have similar stabilities with a low interconversion bar-
rier. Therefore, the 4a/4b = 30:70 ratio reflects that of the
rate constants for the cyclization reaction of the two con-
formers. Both cyclizations occur with defined stereochemis-
try affording exclusively the chair conformation of the ring
formed (6Ǟ7). The pre-exponential factor is within the ex-
pected range. The stereoselectivity observed for the transfer
of hydrogen atom to radical 7 by both silicon hydride and
thiols could be due to preference of attack from the less
hindered side of the ring. Indeed, the anti rule has been
successfully applied to many cases of cyclic radicals.[19]
It is also interesting to compare the present results with
the findings of Cadet and co-workers from similar reactions
using Bu3SnH as the reducing agent.[5,8a] They reported
that both aldehydes 3 and 9 afford analogous cyclonucleo-
sides exclusively having S-configuration at the C5Ј position
and that aldehyde 3 affords two diastereoisomers (5R/5S =
1:2.7) (Scheme 4). The difference in chemical behavior ex-
hibited by (TMS)3SiH and Bu3SnH prompted us to re-in-
vestigate the reaction of aldehyde 3 with Bu3SnH under the
previously reported conditions. 1H NMR spectroscopic
analyses on the reaction mixture revealed the presence of
three cyclonucleosides in a ratio of 15:65:20. Two diastereo-
isomers having S-configuration at the C5Ј position and dif-
fering in the configuration C5 (5R/5S = 1:4.3) are the re-
ported ones,[5] whereas the third one is the diastereoisomer
Scheme 3. Chemical studies on the fate of C5Ј radical 6 under a
variety of experimental conditions.
was conducted in which the aldehyde 3 was treated with a
mixture of (TMS)3SiH/BuSH in known concentrations at
various temperatures in the presence of a radical initiator.
1H NMR spectroscopy and LC/MS were used to analyze
the reaction mixtures. The concentration of BuSH in the
range 0.6–40 m was chosen so that the ratio [8]/[4] ranged
between 1:1 and 5:1. The ratio [8]/[4] varied in the manner
expected with a change in the BuSH concentration, whereas
the diastereoisomeric ratio 4a/4b = 30:70 was nearly con-
stant. Since the thiol concentration during the reaction re-
mained essentially constant under our experimental condi-
tions (pseudo-first-order conditions),[15] the relation [Equa-
tion (2)] is obeyed.[12]
[8]/[4] = (kH/kc)[BuSH]
(2)
The kH/kc values reported in Table 1 were obtained as
the average of different experiments. Linear regression
analysis of a log (kH/kc) vs. 1/T plot yields the relative
Arrhenius parameters given by Equation (3) where θ =
2.3RT kcalmol–1 and the errors represent the standard devi-
ation.
The rate constant kH for the reaction of the α-silyloxyl
secondary carbon-centered radical 6 with BuSH and its
temperature dependence are unknown, although they are
needed in order to obtain the Arrhenius expression for kc
from Equation (3). However, Newcomb and co-workers
have determined rate constants very close to 2.0×107 –1 s–1
for the reactions between octadecanethiol and primary
alkyl or α-methoxyl secondary alkyl radicals in THF at
30 °C.[16] The Arrhenius parameters for the reaction of pri-
mary alkyl radicals with tBuSH are log(A/–1 s–1) = 8.15
and Ea = 1.86 kcal/mol.[17] The combination of this data Scheme 4. Product studies reported by Cadet and co-workers.[5,8a]
4642 © 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.eurjoc.org
Eur. J. Org. Chem. 2005, 4640–4648

Radical Cyclization Approach to Cyclonucleosides
FULL PAPER
5b having R-configuration at the C5Ј position. Therefore,
the cyclization of the Bu3Sn adduct radical occurred with
a diastereoisomeric ratio (5ЈS)/(5ЈR) = 80:20 whereas the
cyclization of the corresponding silyl-substituted radical 6
occurred with a ratio (5ЈS)/(5ЈR) = 30:70, The inverted dia-
stereoselectivity observed for the two reducing agents is
probably due to the different spatial shapes of the (TMS)3-
Si and Bu3Sn groups.[11]
The photochemical deprotection of tris(trimethylsilyl)si-
lyl ethers is also worth a comment. It has been reported
that (TMS)3SiCl can be used for the protection of primary
and secondary alcohols.[13] These silyl ethers are stable un-
der the usual conditions employed in organic synthesis for
the deprotection of other silyl groups and can be depro-
tected using photolysis at 254 nm, in yields ranging from 62
to 95%. Our findings point out the potentiality of this
method if combined with radical reactions. Although a sim-
ple hydrosilylation/deprotection combination is formally
equivalent to the ionic reduction of carbonyl moieties, the
use of an aldehydic function in consecutive radical reactions
followed by deprotection could be a new approach for the
formation of new stereogenic centers based on the hindered
properties of the (TMS)3Si-group.[11]
Reaction of (TMS)3SiH with Aldehyde 10: Aldehyde 10
is prepared from DMSO-based oxidation of the 3Ј-O-pro-
tected derivative of N-benzoyl-2Ј-deoxyadenosine according
to the Moffatt method.[20] The cyclization reaction was ini-
tially tested using the same conditions reported above for
aldehyde 3, i.e., a deoxygenated solution of 10 in fluoroben-
zene (0.01 ) was treated with 5 equiv. of (TMS)3SiH and
0.2 equiv. AIBN at 85 °C. After 2 h a low conversion of the
starting aldehyde was observed. However, when the AIBN
was used in stoichiometric amounts all the aldehyde was
consumed.[21] 1H NMR spectroscopic and LC/MS analysis
of the crude reaction mixture showed the formation of cy-
clonucleosides 11 and 12 in a 1:1 ratio and in 70% overall
yield together with some minor unidentified products
(Scheme 5). Attempts to separate the cyclonucleosides 11
and 12 by flash chromatography were unsuccessful. How-
ever, when the mixture of 11 and 12 was treated by tetra-
chloro-1,4-benzoquinone (chloranil) in refluxing xylene,[22]
compound 11 was quantitatively oxidized to 12, which can
be isolated and fully characterized. The stereochemistry of
the above-mentioned cyclonucleosides was deduced from
Scheme 5. Product studies of the reaction of aldehyde 10 with
(TMS)3SiH using 1 equiv. of AIBN in reflux fluorobenzene and
subsequent oxidation by chloranil in reflux xylene.
duct. Further support for the assigned stereochemistry of
the above-mentioned cyclonucleosides can be obtained by
the excellent agreement of 1H NMR spectroscopic data
with those reported by Cadet and co-workers for com-
pounds 13 and 14.[6] In their synthetic approach, compound
13 was obtained by the sequential treatment of 14 with sele-
nium oxide in refluxing 1,4-dioxane and NaBH4 in meth-
anol.
Scheme 6. Photolysis of 12 by 254 nm light in CH2Cl2/MeOH for
90 min.
From a mechanistic point of view, the addition of the
NMR analysis. The assignment of the (5ЈS,8R) configura- (TMS)3Si· radical to aldehyde 10 affords the C5Ј radical 15
tion to isomer 11 came from NOEs on H2Ј (2%) and H3Ј that attacks the adenine moiety intramolecularly to give the
(5%) upon irradiation of H8. The assignment of the 5ЈS aminyl radical 16 (Scheme 7).[3] Radical 15 should be nearly
configuration to isomer 12 came from the large J4Ј,5Ј coup- planar and the two possible conformers give rise to the
ling of 6.0 Hz (J4Ј,5Ј = 6.0 Hz and J4Ј,5Ј = 0.0 Hz are re- (5ЈR)- and (5ЈS)-isomers. The steric hindrance between the
ported for the 5ЈS- and 5ЈR-isomers of 5Ј,8-cyclo-2Ј-deoxy- adenine and OSi(TMS)3 moieties favors one of these con-
adenosine, respectively)[3] and the absence of NOEs on H3Ј formers. The cyclization occurs with defined stereochemis-
upon irradiation of H5Ј and vice versa.
try, affording exclusively the chair conformation in the rings
Photolysis with 254 nm light of a 20 m solution of 12 formed (15Ǟ16). Hydrogen abstraction from the silane
in an 8:3 CH2Cl2/CH3OH mixture afforded the 5Ј-O-desi- gives product 11 and (TMS)3Si· radical. However, the large
lylated compound 13 and the 2Ј,5Ј-dideoxy cyclic derivative quantity of oxidized 12 is not straightforward. Beckwith el
14 in 56% and 14% yields, respectively (Scheme 6). Pro- al.[21] recently reported that the Bu3SnH-mediated intra-
longed photolysis of the reaction mixture did not show any molecular-homolytic substitution requires large amounts of
changes, suggesting that compound 14 is a direct photopro- initiator, as in the present work. Based on their investiga-
Eur. J. Org. Chem. 2005, 4640–4648
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4643

M. L. Navacchia, A. Manetto, P. C. Montevecchi, C. Chatgilialoglu
FULL PAPER
Scheme 7. Chemical studies on the fate of C5Ј radical 15 under a variety of experimental conditions.
tion, we suggest that the oxidized product 12 derives from intercept the C5Ј radical prior to cyclization. As shown in
the reaction of radical 16 with radical 17, which was gener- Equation (1), the resulting thiyl radical abstracts hydrogen
ated from the decomposition of AIBN. Assuming that 2kt from the silane, thus completing this chain reaction.
Ϸ 109 –1 s–1 and [R·] Ϸ 10–8 , we calculated, from the
A series of experiments was also conducted in which al-
distribution of the products, kSiH Ϸ 2×102 s–1, which is in dehyde 10 was treated with (TMS)3SiH/BuSH in order to
excellent agreement with the reported rate constants of provide the conditions in which the rate constant kc for the
Ar2N· radicals with (TMS)3SiH.[23]
6-exo-trig cyclization could be obtained (Scheme 7). In par-
When the above-described experiment was carried out in ticular, deoxygenated fluorobenzene 0.01  solutions of 10
the presence of PhSH (0.1 ) and a small amount of AIBN, containing 0.05  (TMS)3SiH, 0.02–0.1  BuSH and 0.2
1
the H NMR spectrum showed a quantitative formation of equiv. AIBN were refluxed at 86 °C. 1H NMR spectroscopy
the reduction product 18 that was isolated by flash and LC/MS were used to analyze the reaction mixtures. The
chromatography in a 95% yield (Scheme 7). Under these relative concentrations of 18, 11, and 12 varied in the ex-
conditions, the hydrogen donor is the thiol that is able to pected manner as the concentration of BuSH was changed.
Scheme 8. Proposed mechanism for the reaction of aminyl radical 16 with molecular oxygen. Structure 22 represents a transition state.
4644
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4640–4648

Radical Cyclization Approach to Cyclonucleosides
FULL PAPER
tively. Chemical shifts are expressed in ppm (δ) and coupling con-
stants in Hertz (Hz). LC/MS analyses were performed with an Ag-
ilent 1100 HPLC system and an Esquire 3000 Plus Bruker mass
spectromer. LC analyses were performed with a Zorbax C8 column
(4.6×150 mm, 5 µ) with a linear gradient acetonitrile/water from
50:50 to 95:5 in 20 min at a flow rate 0.6 mL/min, detection at λ =
260 nm. Column chromatography was performed by the method of
Still using Merck 230–400 mesh ASTM silica gel 60. Analytical
thin-layer chromatography (TLC) was performed using Merck 60
F254 0.2 precoated silica gel plates. Compounds were visualized
Analysis of the data in the usual manner [cf. Equation (2)
in the previous section] allowed a kH/kc = 28.6 2.4 –1 to
be obtained, as an average of three independent experi-
ments (errors corresponds to the standard deviation).
A rate constant kc = 3.5×105 s–1 at 86 °C is calculated
by taking kH = 1.0×107 –1 s–1.[17] For comparison, an ab-
solute rate constant kc = 1.6×105 s–1 at 22 °C is obtained
by pulse radiolysis for the corresponding unprotected radi-
cal in water.[3] Moreover, a kc = 7.7×104 s–1 at 86 °C is cal-
culated for the cyclization of radical 6, which is nearly five using ultraviolet light or by heating plates previously immersed in
an ammonium molybdate/ceric ammonium sulfate/sulfuric acid
mixture. Solvents were freshly distilled prior to use. All other rea-
gents were used as received.
times slower than the cyclization of radical 15.
With the aim of optimizing the yield of cyclonucleoside
12, we performed the reaction in the presence of oxygen. A
solution of 10 in fluorobenzene (0.01 ) was treated with 5
equiv. of (TMS)3SiH and stoichiometric amounts of AIBN
Reaction of 5Ј-Carbaldehyde 3 with (TMS)3SiH: AIBN (3 mg,
0.02 mmol) and (TMS)3SiH (0.175 mL, 0.5 mmol) were added to a
at 85 °C under air. Complete conversion of the starting al- 0.01  solution of aldehyde 3[5] (48 mg, 0.1 mmol) in C6H5 (10 mL).
The solution was refluxed under argon for 1 h. Analysis by LC/
MS and 1H NMR spectroscopy of the crude reaction mixture after
evaporation of the solvent showed the formation of compounds 4a
and 4b in 30:70 ratio and in 85% overall yield. Column chromatog-
raphy on silica gel by gradient elution with n-hexane/ethyl acetate
led to separation of the two compounds.
dehyde was observed within 2 h with the formation of cy-
clonucleoside 12 as the sole product in 75% yield. The
mechanism that we conceived for the reaction is outlined in
Scheme 8. We have previously shown by kinetic studies
using pulse radiolysis techniques that the unprotected ami-
nyl radical 16 reacts reversibly with oxygen.[3] Aminyl radi-
cal 16 is in resonance with the structure 19, where the un-
paired electron is placed at the C4 position, and addition to
molecular oxygen can afford peroxyl radical 20. This radical
should undergo a facile hydrogen migration via the chair-
type transition state 22 to generate the C6 radical 21, which
eliminates the HOO· radical, thus affording product 12. Hy-
drogen abstraction from silane by the HOO· radical should
regenerate the (TMS)3Si· radical, thus completing this chain
reaction.[24]
(5ЈS,6S,5S)-3Ј-O-(tert-Butyldiphenylsilyl)-5Ј-O-[tris(trimethylsilyl)si-
lyl]-5Ј,6-cyclo-5,6-dihydrothymidine (4a): 1H NMR (400 MHz,
CDCl3): δ = 7.25–7.80 (m, 10 H, Ph), 7.2 (s, 1 H, NH), 5.91 (d,
J1Ј,2Ј = 6.0 Hz, 1 H, H-1Ј), 4.67 (dd, J2Ј,3Ј = 2, J2ЈЈ,3Ј = 7.0 Hz, 1 H,
H-3Ј), 4.47 (d, J4Ј,5Ј = 5.0 Hz, 1 H, H-4Ј), 3.58 (dd, J4Ј,5Ј = 5.0, J5Ј,6
= 9.2 Hz, 1 H, H-5Ј), 3.00 (dd, J5,6 = 3.2, J5Ј,6 = 9.2 Hz, 1 H, H-
6), 2.73 (dq, J5,6 = 3.2, J5,Me = 7.2 Hz, 1 H, H-5), 2.27 (ddd, J2Ј,2ЈЈ
= 14, J1Ј,2Ј = 6.0, J2Ј,3Ј = 2 Hz, 1 H, H-2Ј), 1.46 (dd, J2Ј,2ЈЈ = 14,
J2ЈЈ,3Ј = 7.0 Hz, 1 H, H-2ЈЈ), 1.22 (d, J = 7.2 Hz, 3 H, Me), 1.05 (s,
9 H, tBuSi), 0.15 (s, 27 H, MeSi) ppm; NOE experiments: irradia-
tion at δ = 3.00 (H-6) ppm caused enhancements of signals at δ =
1.46 (H-2ЈЈ, 2%), 2.73 (H-5, 9%) ppm and 4.67 (H-3Ј, 6%) ppm
whereas no enhancement of the signal at δ = 3.58 (H-5Ј) ppm was
found; irradiation at δ = 3.58 (H-5Ј) ppm caused enhancements of
signals at δ = 4.47 (H-4Ј, 8%) ppm whereas no enhancement of the
signal at δ = 2.73 (H-5) ppm was found. 13C NMR (100.6 MHz,
CDCl3): δ = 0.86 (Me), 11.68 (Me), 19.34 (q), 26.89 (Me), 36.25
(CH), 42.17 (CH2), 54.79 (CH), 68.53 (CH), 71.64 (CH), 84.35
(CH), 86.65 (CH), 127.88 (CH), 128.16 (CH), 129.91 (CH), 130.15
(CH), 133.40 (q), 134.65 (q), 135.82 (CH), 135.91 (CH), 150.59
(CO), 172.43 (CO) ppm. MS (ESI): 749 [M + Na+].
Conclusions
We have disclosed two short and efficient synthetic se-
quences, based on consecutive radical reactions followed by
photochemical desilylation, for the preparation of cyclonu-
cleosides 5 and 13. The C5Ј radicals, generated by addition
of the (TMS)3Si· radical to the corresponding 5Ј-carbal-
dehyde, are the key intermediates in these transformations.
The 6-exo-trig cyclization is effective with both pyrimidine
and purine derivatives, although the nature of the base
plays an important role in the stereochemical outcome. Our
findings can furnish a molecular basis for forthcoming ex-
periments involving the role of these conformationally re-
stricted structures in DNA damage.
The research described in this article has demonstrated
the feasibility of radical reactions for the preparation of
5Ј,6-cyclopyrimidine and 5Ј,8-cyclopurine nucleosides start-
ing from easily available modified derivatives. We envisage
that this approach can be extended to other pyrimidine and
purine nucleosides.
(5ЈR,6S,5S)-3Ј-O-(tert-Butyldiphenylsilyl)-5Ј-O-[tris(trimethylsilyl)si-
lyl]-5Ј,6-cyclo-5,6-dihydrothymidine (4b): 1H NMR (400 MHz,
CDCl3): δ = 7.4–7.70 (m, 10 H, Ph), 7.1 (s, 1 H, NH), 6.30 (d, J1Ј,2Ј
= 7.0 Hz, 1 H, H-1Ј), 4.55 (d, J4Ј,5Ј = 2.0 Hz, 1 H, H-4Ј), 4.38 (dd,
J2Ј,3Ј = 2.5, J2ЈЈ,3Ј = 7.0 Hz, 1 H, H-3Ј), 3.18 (t, J4Ј,5Ј = J5Ј,6 = 2 Hz,
1 H, H-5Ј), 3.14 (dd, J5,6 = 7.5, J5Ј,6 = 2.0 Hz, 1 H, H-6), 2.64
(quintuplet, J5,6 = J5, Me = 7.5 Hz, 1 H, H-5), 2.07 (ABXY, JA,B
=
14, J1Ј,2Ј = 7.0, J2Ј,3Ј = 2.5 Hz, 1 H, H-2Ј), 1.95 (ABX, JA,B = 14,
J2ЈЈ,3Ј = 7.0 Hz, 1 H, H-2ЈЈ), 1.14 (d, J = 7.5 Hz, 3 H, Me), 1.05 (s,
9 H, tBuSi). 0.15 (s, 27 H, MeSi) ppm; NOE experiments: irradia-
tion at δ = 2.64 (H-5) ppm caused enhancements of signals at δ =
3.14 (H-6, 4%) and 1.14 (Me, 1.5%) ppm whereas no enhancement
of the signal at δ = 3.18 (H-5Ј) ppm; Irradiation at δ = 4.38 (H-
3Ј) ppm caused enhancements of signals at δ = 1.95 (H-2ЈЈ, 2%),
3.18 (H-5Ј, 5%), 3.14 (H-6, 6%), and 4.55 (H-4Ј, 4%) ppm. 13C
NMR (100.6 MHz, CDCl3): δ = 0.94 (Me), 9.30 (Me), 19.28 (q),
26.92 (Me), 35.05 (CH), 40.14 (CH2), 54.99 (CH), 69.40 (CH),
73.74 (CH), 85.22 (CH), 85.96 (CH), 128.07 (CH), 130.20 (CH),
Experimental Section
General: 1H and 13C NMR spectra were recorded with a Varian
Mercury 400 spectrometer at 400 MHz and 100.6 MHz, respec-
Eur. J. Org. Chem. 2005, 4640–4648
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4645

M. L. Navacchia, A. Manetto, P. C. Montevecchi, C. Chatgilialoglu
FULL PAPER
133.33 (q), 133.86 (q), 136.04 (CH), 136.08 (CH), 154.05 (CO), tion of three diastereoisomeric cyclization products in a ratio 5a/
170.48 (CO) ppm. MS (ESI): 749 [M + Na+].
5b/(5ЈS,5R,6S)[5] = 65:20:15 and in 80% overall yield.
Preparation of 5Ј-Carbaldehyde 10: N5-Benzoyl-3Ј-O-(tert-butyldi-
metylsilyl)-2Ј-deoxyadenosine[25] (470 mg, 1 mmol) was dissolved in
dry DMSO (5 mL). Dicyclohexylcarbodiimide (820 mg, 4 mmol)
and dichloroacetic acid (0.041 mL, 0.5 mmol) in dry DMSO
(1.4 mL) were added to the solution. The mixture was stirred at
room temperature for 1 h and then ethyl acetate (4 mL) was added.
Oxalic acid (250 mg, 2 mmol) was added portionwise followed by
ethyl acetate (4 mL). The solution was left whilst stirring for 1 h at
room temp. The mixture was filtered through celite and the filtrate
was washed with water and extracted with ethyl acetate (3×5 mL).
The organic layer was dried, the solvent evaporated under reduced
pressure and the residue purified on a silica gel column. The elution
with 80:20 diethyl ether: pentane gave the 5Ј-carbaldehyde 10 as a
(5ЈS,6S,5S)-3Ј-O-(tert-Butyldiphenylsilyl)-5Ј,6-cyclo-5,6-dihydrothy-
midine (5a): A 20 m solution of 4a (10 mg, 0.02 mmol) in an 8:3
CH2Cl2/MeOH mixture (25 mL) was photolysed at λ = 254 nm for
30 min to give quantitatively the 5Ј-O-desilylated compound 5a. 1H
NMR (400 MHz, CDCl3): δ = 8.0 (s, 1 H, NH), 7.25–7.80 (m, 10
H, Ph), 6.2 (d, J1Ј,2Ј = 6.0 Hz, 1 H, H-1Ј), 4.55 (dd, J2Ј,3Ј = 3.0,
J2ЈЈ,3Ј = 7.0 Hz, 1 H, H-3Ј), 4.18 (d, J4Ј,5Ј = 4.4 Hz, 1 H, H-4Ј), 3.67
(dd, J4Ј,5Ј = 4.4, J5Ј,6 = 9.3 Hz, 1 H, H-5Ј), 3.01 (dd, J5,6 = 3.6, J5Ј,6
= 9.3 Hz, 1 H, H-6), 2.75 (m, 1 H, H-5), 2.31 (ddd, J2Ј,2ЈЈ = 14.4,
J1Ј,2Ј = 6.0, J2Ј,3Ј = 3.0 Hz, 1 H, H-2Ј), 2.12 (dd, J2Ј,2ЈЈ = 14.4, J2ЈЈ,3Ј
= 7.0 Hz, 1 H, H-2ЈЈ), 1.18 (d, J = 7.2 Hz, 3 H, Me), 1.05 (s, 9 H,
tBuSi) ppm. 13C NMR (100.6 MHz, CDCl3): δ = 11.24 (Me), 19.25
(q), 27.06 (Me), 36.00 (CH), 43.42 (CH2), 53.75 (CH), 62.90 (CH),
69.97 (CH), 84.25 (CH), 86.20 (CH), 127.99 (CH), 128.05 (CH),
130.17 (CH), 130.19 (CH), 133.25 (q), 133.61 (q), 135.93 (CH),
151.18 (CO), 173.33 (CO) ppm. MS (ESI): 503 (M + 23).
1
yellow foam (380 mg, 80%). H NMR (400 MHz, CDCl3): δ = 9.8
(s, 1 H, CHO), 9.10 (br. s, 1 H, NH), 8.74 (s, 1 H, H-2Ј), 8.30 (s, 1
H, H-8), 7.4–8.0 (m, 5 H, Ph), 6.58 (dd, J1Ј,2Ј = 8.0, J1Ј,2ЈЈ = 6.4 Hz,
1 H, H–1Ј), 4.90 (m, 1 H, H-3Ј), 4.45 (d, J3Ј,4Ј = 1.2 Hz, 1 H, H-
4Ј), 2.90 (ddd, J2Ј,2ЈЈ = 14.0, J2Ј,1Ј = 8.0, J2Ј,3Ј = 2.0 Hz, 1 H, H-2Ј),
2.48 (ddd, J2Ј,2ЈЈ = 14.0, J2ЈЈ,1Ј = 6.4, J2ЈЈ,3Ј = 3 Hz, 1 H, H-2ЈЈ), 0.95
(s, 9 H, tBuSi), 0.18 (s, 6 H, MeSi) ppm. MS (ESI): 468 (M + 1).
MS2 (468) 239. The aldehyde 10 contained its hydrated form 10Ј
(10/10Ј ratio 80:20). The ratio 10/10Ј changed by adding a drop of
D2O in the NMR tube, going from an 80:20 to 30:70 ratio within
50 min, then it remained constant over 20 h. The 30:70 ratio should
be considered the equilibrium ratio in CDCl3:D2O (v/v 7:1).
Hydrated 5Ј-Carbaldehyde 10Ј: 1H NMR (400 MHz, CDCl3): δ =
9.20 (br. s, 1 H, NH), 8.70 (s, 1 H, H-2), 8.18 (s, 1 H, H-8), 7.4–
8.0 (m, 5 H, Ph), 6.42 (dd, J1Ј,2Ј = 10.0, J1Ј,2ЈЈ = 5.0 Hz, 1 H, H-
1Ј), 5.16 (d, J4Ј,5Ј = 2 Hz, 1 H, H-5Ј), 4.57 (d, J2Ј,3Ј = 4.8 Hz, 1 H,
H-3Ј), 4.15 (d, J4Ј,5Ј = 2 Hz, 1 H, H-4Ј), 2.91 (m, 1 H, H-2Ј), 2.22
(dd, J1Ј,2ЈЈ = 5.0, J2Ј,2ЈЈ = 13.5 Hz, 1 H, H-2ЈЈ), 0.95 (s, 9 H, tBuSi),
0.18 (s, 6 H, MeSi) ppm. MS (ESI): 486 [M + 1].
(5ЈR,6S,5S)-3Ј-O-(tert-Butyldiphenylsilyl)-5Ј,6-cyclo-5,6-dihydro-
thymidine (5b): A 20 m solution of 4b (10 mg, 0.02 mmol) in an
8:3 CH2Cl2/MeOH mixture (25 mL) was photolysed at λ = 254 nm
for 30 min to give quantitatively the 5Ј-O-desilylated compound 5b.
1H NMR (400 MHz, CDCl3): δ = 7.6 (s, 1 H, NH), 7.4–7.8 (m, 10
H, Ph), 6.42 (d, J1Ј,2Ј = 6.0 Hz, 1 H, H-1Ј), 4.26 (dd, J2Ј,3Ј = 3.0,
J2ЈЈ,3Ј = 6.5 Hz, 1 H, H-3Ј), 4.24 (d, J4Ј,5Ј = 2.4 Hz, 1 H, H-4Ј), 3.20
(dd, J5Ј,6 = 1.5, J5,6 = 7.5 Hz, 1 H, H-6), 2.96 (br. s, collapsing to
a doublet upon irradiation at δ = 3.2, J4Ј,5Ј = 2.4 Hz, 1 H, H-5Ј),
2.88 (br. s, 1 H, OH), 2.70 (quintuplet, J5,6 = J5, CH3 = 7.5 Hz, 1
H, H-5), 2.34 (A part of an ABXY system, JAB = 14.5, J1Ј,2Ј = 6.0,
J2Ј,3Ј = 3.0 Hz, 1 H, H-2Ј), 2.30 (B part of an ABX system, JAB
=
14.5, J2ЈЈ,3Ј = 6.5 Hz, 1 H, H-2ЈЈ), 1.18 (d, J = 7.2 Hz, 3 H, Me),
1.05 (s, 9 H, tBuSi) ppm. 13C NMR (100.6 MHz, CDCl3): δ = 9.16
(Me), 19.25 (q), 27.04 (Me), 34.99 (CH), 39.78 (CH2), 54.80 (CH),
65.56 (CH), 73.22 (CH), 85.52 (CH), 87.52 (CH), 128.12 (CH),
128.18 (CH), 128.21 (CH), 130.31 (CH), 133.52 (q), 136.12 (CH),
154.95 (CO), 171.51 (CO) ppm. MS (ESI): 503 [M + 23].
Reaction of 5Ј-Carbaldehyde 10 with (TMS)3SiH: AIBN (15 mg,
0.1 mmol) and (TMS)3SiH (0.175 mL, 0.5 mmol) were added to a
0.01  solution of the aldehyde 10 (50 mg, 0.1 mmol) in C6H5F
(10 mL). The solution was refluxed under argon for 2 h. Analysis
by LC/MS and 1H NMR of the crude reaction mixture after evapo-
ration of the solvent showed the formation of compounds 11 and
12 in a 1:1 ratio and in overall yields of 70%. Attempts to separate
the cyclonucleosides 11 and 12 by flash chromatography were un-
successful. When the mixture was treated by chloranil in refluxing
xylene the compound 11 was quantitatively oxidized to 12.
3Ј-O-(tert-Butyldiphenylsilyl)-5Ј-O-[tris(trimethylsilyl)silyl]thymid-
ine (8): AIBN (3 mg, 0.02 mmol), (TMS)3SiH (0.175 mL,
0.5 mmol) and thiophenol (0.1 mL, 1.0 mmol) were added to a
0.01  solution of the aldehyde 3 (48 mg, 0.1 mmol) in C6H5
1
(10 mL). The solution was refluxed under argon for 1 h. H NMR
spectroscopic analysis of the reaction mixture showed the presence
of compound 8 in 95% yield. Column chromatography led to sepa-
1
ration of a pure sample. H NMR (400 MHz, CDCl3): δ = 8.0 (s,
(5ЈS,8R)-N5-Benzoyl-3Ј-O-(tert-butyldimethylsilyl)-5Ј-O-[tris(trime-
thylsilyl)silyl]-5Ј,8-cyclo-7,8-dihydro-2Ј-deoxyadenosine (11): 1H
NMR (400 MHz, CDCl3): δ = 9.6 (s, 1 H, NH), 7.9 (s, 1 H, H-2),
7.4–7.8 (m, 5 H, Ph), 6.13 (d, J1Ј,2Ј = 6.0 Hz, 1 H, H-1Ј), 5.20 (d,
JNH,8 = 6.0 Hz, 1 H, NH), 4.82 (dd, J5Ј,8 = 7.0, JNH,8 = 6.0 Hz, 1
H, H-8), 4.58 (dd, J2ЈЈ,3Ј = 7.0, J2Ј,3Ј = 2.0 Hz, 1 H, H-3Ј), 4.18 (d,
J4Ј,5Ј = 4.5 Hz, 1 H, H-4Ј), 3.35 (dd, J4Ј,5Ј = 4.5, J5Ј,8 = 7.0 Hz, 1
H, H-5Ј), 2.31 (ABX system, J2Ј,2ЈЈ = 13.5, J2ЈЈ,3Ј = 7.0 Hz, 1 H, H-
2ЈЈ), 2.14 (ABXY system, J2Ј,2ЈЈ = 13.5, J1Ј,2Ј = 6.0, J2Ј,3Ј = 2.0 Hz,
1 H, H-2Ј), 0.9 (s, 9 H, tBuSi), 0.3 (s, 3 H, MeSi), 0.13 (s, 27 H,
Me3Si), 0.1 (s, 3 H, MeSi) ppm; NOE experiments: irradiation at δ
= 4.82 (H-8) ppm caused enhancements at δ = 2.31 (H-2Ј, 2%),
3.35 (H-5Ј, 2%), and 4.58 (H3Ј, 5%) ppm; irradiation at δ = 3.35
(H5Ј) ppm caused enhancements at δ = 4.82 (H-8, 2%), 4.18 (H4Ј,
4.5%) ppm; LC/MS (ESI): 738 [M + Na+], 716 [M + 1]. MS2 (716)
698, 584.
1 H, H-6), 7.3–7.70 (m, 10 H, Ph), 7.2 (s, 1 H, NH), 6.38 (dd, J1Ј,2Ј
= 5.2, J1Ј,2ЈЈ = 9 Hz, 1 H, H-1Ј), 4.35 (d, J2ЈЈ,3Ј = 5.0, 1 H, H-3Ј),
4.08 (br. s, 1 H, H-4Ј), 3.63 (dd, J5Ј,5ЈЈ = 11, J4Ј,5Ј = 2 Hz, 1 H, H-
5Ј), 3.32 (dd, J5Ј,5ЈЈ = 11, J4Ј,5ЈЈ = 2 Hz, 1 H, H-5ЈЈ), 2.16 (dd, J2Ј,2ЈЈ
= 14, J1Ј,2Ј = 5.2 Hz, 1 H, H-2Ј), 1.68 (ddd, J2Ј,2ЈЈ = 14, J1Ј,2ЈЈ = 9,
J2ЈЈ,3Ј = 5.0 Hz, 1 H, H-2ЈЈ), 1.87 (s, 3 H, Me), 1.05 (s, 9 H, tBuSi),
0.15 (s, 27 H, MeSi) ppm. 13C NMR (100.6 MHz, CDCl3): δ = 0.39
[3(CH3)3Si], 0,65 (q), 12.62 (CH3), 26.89 [(CH3)3CSi], 41.16 (C2Ј),
68.01 (C5Ј), 74.05 (C3Ј), 84.90 (C1Ј), 87.90 (C4Ј), 110.64 (C5),
127.91 (CH), 133.19 (q), 133.31 (q), 135.19 (C6), 135.66 (CH),
135.70 (CH), 150.01 (q), 163.53 (q) ppm. MS (ESI): 749 [M + Na+].
Reaction of 5Ј-Carbaldehyde 3 with Bu3SnH:[5] AIBN (0.06 mmol)
and Bu3SnH (0.6 mmol) were added to a 0.01  solution of alde-
hyde 3 (150 mg, 0.3 mmol) in C6H5 (30 mL). The solution was re-
fluxed under argon for 2 h. After evaporation the crude reaction
mixture was eluted on a silica gel column by n-hexane/ethyl acetate
1
from 90:10 to 0:100 to eliminate tin byproducts. H NMR analysis
(5ЈS)-N5-Benzoyl-3Ј-O-(tert-butyldimethylsilyl)-5Ј-O-[tris(trimethyl-
silyl)silyl]-5Ј,8-cyclo-2Ј-deoxyadenosine (12): 1H NMR (400 MHz,
of the residue, after evaporation of the solvent, showed the forma-
4646
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4640–4648

Radical Cyclization Approach to Cyclonucleosides
FULL PAPER
CDCl3): δ = 9.0 (s, 1 H, NH), 7.9 (s, 1 H, H-2), 7.4–7.8 (m, 5 H, 129.10 (CH), 132.94 (CH), 141.87 (q), 149.71 (q), 152.89 (CH),
Ph), 6.46 (d, J1Ј,2Ј = 4.5 Hz, 1 H, H-1Ј), 4.94 (d, J4Ј,5Ј = 6.0 Hz, 1
H, H-5Ј), 4.70 (dd, J2ЈЈ,3Ј = 7.0, J2Ј,3Ј = 4.5 Hz, 1 H, H-3Ј), 4.62 (d,
J4Ј,5Ј = 6.0 Hz, 1 H, H-4Ј), 2.53 (dd, J2Ј,2ЈЈ = 13.0, J2ЈЈ,3Ј = 7.0 Hz,
1 H, H-2ЈЈ), 2.20 (ddd, J2Ј,2ЈЈ = 13.0, J1Ј,2Ј = J2Ј,3Ј = 4.5 Hz, 1 H,
H-2Ј), 0.90 (s, 9 H, tBuSi), 0.30 (s, 3 H, MeSi), 0.13 (s, 27 H,
Me3Si), 0.10 (s, 3 H, MeSi) ppm; NOE experiments: irradiation at
δ = 4.94 (H-5Ј) ppm causes an enhancement at δ = 4.62 (H-4Ј,
6%) ppm but none at δ = 4.70 (H-3Ј) ppm. 13C NMR (100.6 MHz,
CDCl3): δ = –4.58 (Me), –3.95 (Me), 0.78 (Me), 17.93 (q), 25.92
(Me), 46.83 (CH2), 69.63 (CH), 71.08 (CH), 85.68 (CH), 87.16
(CH), 122.53 (q), 127.85 (CH), 129.10 (CH), 132.96 (CH), 134.11
(q), 149.12 (q), 149.28 (q), 151.09 (q), 152.96 (CH), 164.28
(CO) ppm. MS (ESI): 736 [M + 23], 714 [M + 1]. MS2 (714) 514.
164.84 (CO) ppm. MS (ESI): 738 [M + 23], 716 [M + 1]. MS2 (716)
240.
Kinetic Experiments: A solution containing aldehyde 3 or 10
(0.1 mmol), and (TMS)3SiH (0.17 mL, 0.5 mmol) and the appropri-
ate amount of BuSH in a particular solvent was deoxygenated with
argon (30 min). Reactions were initiated either by Et3B at 25 °C or
thermally (AIBN, tert-butyl hyponitrite, tert-butyl perbenzoate or
di-tert-butyl peroxide). Solutions were refluxed for 1 h and ana-
1
lyzed by H NMR spectroscopy and/or LC/MS.
Acknowledgments
Work supported in part by the European Community’s Human
Potential Programme under contract MCRTN-CT-2003–505086
[CLUSTOXDNA] and by the MURST (Rome).
Photolysis of 12: A 20 m solution of 12 (10 mg, 0.02 mmol) in an
8:3 CH2Cl2/MeOH mixture (25 mL) was photolysed at λ = 254 nm
for 90 min to give the 5Ј-O-desilylated compound 13 and dideoxya-
denosine 14 in 56% and 14% yields, respectively.
(5ЈS)-N5-Benzoyl-3Ј-O-(tert-butyldimethylsilyl)-5Ј,8-cyclo-2Ј-deoxy-
adenosine (13): 1H NMR (400 MHz, CDCl3): δ = 9.2 (s, 1 H, NH),
8.6 (s, 1 H, H-2), 7.4–8.0 (m, 5 H, Ph), 6.46 (d, J1Ј,2Ј = 4.8 Hz, 1
H, H-1Ј), 5.46 (d, J4Ј,5Ј = 6.0 Hz, 1 H, H-5Ј), 4.77 (dd, J2ЈЈ,3Ј = 7.0,
J2Ј,3Ј = 4.0 Hz, 1 H, H-3Ј), 4.33 (d, J4Ј,5Ј = 6.0 Hz, 1 H, H-4Ј), 2.47
[1] For recent reviews, see: a) J. Cadet, T. Douki, D. Gasparutto,
J.-L. Ravanat, Radiat. Phys. Chem. 2005, 72, 293–299; b) J.
Cadet, T. Douki, J.-L. Ravanat, Redox-Genome Interactions in
Health and Disease (Eds.: J. Fuchs, M. Podda, L. Packer)
Dekker, New York, 2003, pp. 143–189; c) M. Dizdaroglu, P.
Jaruga, H. Rodriguez, in: Critical Reviews of Oxidative Stress
and Aging: Advances in Basic Science Diagnostics and Inter-
vention, vol. 1 (Eds.: R. G. Gutler, H. Rordiguez) World Scien-
tific, London, 2003, pp. 165–189; d) M. Dizdaroglu, P. Jaruga,
M. Birincioglu, H. Rodriguez, Free Radic. Biol. Med. 2002, 32,
1102–1115; e) C. Chatgilialoglu, P. O’Neill, Exp. Gerontol.
2001, 36, 1459–1471.
[2] For some recent relevant work, see: a) M. Birincioglu, P. Ja-
ruga, G. Chowdhury, H. Rodriguez, M. Dizdaroglu, K. S.
Gates, J. Am. Chem. Soc. 2003, 125, 11607–11615; b) P. Jaruga,
M. Birincioglu, H. Rodriguez, M. Dizdaroglu, Biochemistry
2002, 41, 3703–3711; c) M. Dizdaroglu, P. Jaruga, H. Rodrig-
uez, Free Radic. Biol. Med. 2001, 30, 774–784.
[3] a) C. Chatgilialoglu, M. Guerra, Q. G. Mulazzani, J. Am.
Chem. Soc. 2003, 125, 3839–3848; b) R. Flyunt, R. Bazzanini,
C. Chatgilialoglu, Q. G. Mulazzani, J. Am. Chem. Soc. 2000,
122, 4225–4226.
[4] L. B. Jimenez, S. Encina, M. A. Miranda, M. L. Navacchia, C.
Chatgilialoglu, Photochem. Photobiol. Sci. 2004, 3, 1042–1046.
[5] A. Romieu, D. Gasparutto, J. Cadet, J. Chem. Soc. Perkin
Trans. 1 1999, 1257–1263.
(dd, J2Ј,2ЈЈ = 13.0, J2ЈЈ,3Ј = 7.0 Hz, 1 H, H-2ЈЈ), 2.21 (ddd, J2Ј,2ЈЈ
=
13.0, J1Ј,2Ј = J2Ј,3Ј = 4.0 Hz, 1 H, H-2Ј), 0.90 (s, 9 H, tBuSi), 0.30
(s, 3 H, MeSi), 0.10 (s, 3 H, MeSi) ppm; NOE experiments: irradia-
tion at δ = 5.46 (H-5Ј) ppm causes an enhancement of the signal
at δ = 4.33 (H4Ј, 2.5%) ppm. 13C NMR (100.6 MHz, CDCl3): δ
= –4.77 (Me), –4.59 (Me), 18.07 (q), 25.92 (Me), 46.55 (CH2), 64.76
(CH), 69.37 (CH), 85.60 (CH), 86.36 (CH), 85.96 (CH), 122.62 (q),
128.20 (CH), 129.09 (CH), 133.15 (CH), 133.59 (q), 148.79 (q),
150.03 (q), 152.55 (CH), 165.49 (CO) ppm. MS (ESI): 468 [M + 1].
MS2 (468) 336, 268.
N5-Benzoyl-3Ј-O-(tert-butyldimethylsilyl)-5Ј,8-cyclo-2Ј,5Ј-dideoxy-
adenosine (14): 1H NMR (400 MHz, CDCl3): 9.2 (s, 1 H, NH), 8.65
(s, 1 H, H-2), 7.4–8.0 (m, 5 H, Ph), 6.58 (d, J1Ј,2Ј = 4.8 Hz, 1 H,
H-1Ј), 4.76 (d, J4Ј,5Ј = 4.8 Hz, 1 H, H-4Ј), 4.43 (dd, J2ЈЈ,3Ј = 7.0,
J2Ј,3Ј = 4.0 Hz, 1 H, H-3Ј), 3.58 (part A of a ABX system, J4Ј,5Ј
(AX) = 4.8, JAB = 18 Hz, 1 H, H-5Ј), 3.10 (part B of a ABX system,
J4Ј,5Ј
= 0, JAB = 18 Hz; 1 H, H-5Ј), 2.63 (dd, J2Ј,2ЈЈ = 14.0,
(AX)
J2ЈЈ,3Ј = 7.0 Hz, 1 H, H-2ЈЈ), 2.31 (ddd, J2Ј,2ЈЈ = 14.0, J1Ј,2Ј = J2Ј,3Ј
= 4.0 Hz, 1 H, H2Ј), 0.90 (s, 9 H, tBuSi), 0.30 (s, 3 H, MeSi), 0.10
(s, 3 H, MeSi) ppm. MS (ESI): 452 [M + 1]. MS2 (452) 320. MS3
(320) 302.
[6] A. Romieu, D. Gasparutto, D. Molko, J. Cadet, J. Org. Chem.
1998, 63, 5245–5249.
[7] A. Romieu, D. Gasparutto, J. Cadet, Chem. Res. Toxicol. 1999,
12, 412–421.
N5-Benzoyl-3Ј-O-(tert-butyldimethylsilyl)-5Ј-O-[tris(trimethylsilyl)si-
lyl]-2Ј-deoxyadenosine (18): AIBN (3 mg, 0.02 mmol), (TMS)3SiH
(0.175 mL, 0.5 mmol)and thiophenol (0.1 mL, 1.0 mmol) were
added to a 0.01  solution of the aldehyde 10 (50 mg, 0.1 mmol)
in C6H5F (10 mL). The solution was refluxed under argon for 1 h.
1H NMR analysis of the reaction mixture showed the presence of
compound 18 in 85% yield. Column chromatography led to separa-
tion of a pure sample. 1H NMR (400 MHz, CDCl3): δ = 9.10 (s,
1H, NH), 8.8 (s, 1 H, H8), 8.20 (s, 1 H, H-2), 7.5–8.0 (m, 5 H, Ph),
6.47 (dd, J1Ј,2Ј = 7.5, J1Ј,2ЈЈ = 6.0 Hz, 1 H, H-1Ј), 4.58 (m, 1 H, H-
3Ј), 4.0 (ddd, J4Ј,5ЈЈ = 3.0, J4Ј,5Ј = 5.0, J4Ј,3Ј = 3.0 Hz; 1 H, H-4Ј),
3.76 (ABX, JAB = 11.0, J4Ј,5Ј = 5.0 Hz, 1 H, H-5Ј), 3.62 (ABX, JAB
= 11.0, J4Ј,5ЈЈ = 3.0 Hz, 1 H, H-5ЈЈ), 2.75 (ddd, J2Ј,2ЈЈ = 13.0, J2Ј,1Ј
= J2Ј,3Ј = 7.5 Hz, 1 H, H-2Ј), 2.40 (ddd, J2Ј,2ЈЈ = 13.0, J2ЈЈ,1Ј = 6.0,
J2ЈЈ,3Ј = 2.8 Hz, 1 H, H-2ЈЈ), 0.9 (s, 9 H, tBuSi), 0.13 (s, 27 H,
Me3Si), 0.1 (s, 6 H, MeSi) ppm. 13C NMR (100.6 MHz, CDCl3): δ
= –4.43 (CH), 0.64 (CH3), 18.25 (q), 26.01 (CH3), 41.00 (CH2),
67.70 (CH2), 72.50 (CH), 85.17 (CH), 88.48 (CH), 128.10 (CH),
[8] For analogous 5Ј,6-cyclopyrimidines, see: a) E. Muller, D. Gas-
parutto, M. Jaquinod, A. Romieu, J. Cadet, Tetrahedron 2000,
56, 8689–8701; b) E. Muller, D. Gasparutto, J. Cadet, Chem-
BioChem 2002, 3, 534–542.
[9] a) I. Kuraoka, C. Bender, A. Romieu, J. Cadet, R. D. Wood,
T. Lindahl, Proc. Natl. Acad. Sci. USA 2000, 97, 3832–3837;
b) I. Kuraoka, P. Robins, C. Masutani, F. Hanaoka, D. Gaspa-
rutto, J. Cadet, R. D. Wood, T. Lindahl, J. Biol. Chem. 2001,
276, 49283–49288.
[10] a) P. J. Brooks, D. S. Wise, D. A. Berry, J. V. Kosmoski, M. J.
Smerdon, R. L. Somers, H. Mackie, A. Y. Spoonde, E. J. Ack-
erman, K. Coleman, R. E. Tarone, J. H. Robbins, J. Biol.
Chem. 2000, 275, 22355–22362; b) K. Randerath, G. D. Zhou,
R. L. Somers, J. H. Robbins, P. J. Brooks, J. Biol. Chem. 2001,
276, 36051–36057; c) P. Jaruga, J. Theruvathu, M. Dizdaroglu,
P. J. Brooks, Nucleic Acids Res. 2004, 32, Art. No e87.
[11] C. Chatgilialoglu, Organosilanes in radical chemistry, Wiley,
Chichester, 2004. See also: C. Chatgilialoglu in Radicals in Or-
ganic Synthesis, vol. 1 (Eds.: P. Renaud, M. P. Sibi), Wiley-
VCH, Weinheim, 2001, pp. 28–49.
Eur. J. Org. Chem. 2005, 4640–4648
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4647

M. L. Navacchia, A. Manetto, P. C. Montevecchi, C. Chatgilialoglu
FULL PAPER
[12] a) M. Newcomb, Tetrahedron 1993, 49, 1151–1176; b) D.
Griller, K. U. Ingold, Acc. Chem. Res. 1980, 13, 317–323.
[13] M. A. Brook, S. Balduzzi, M. Mohamed, C. Gottardo, Tetra-
hedron 1999, 55, 10027–10040.
[14] C. Chatgilialoglu, M. Ballestri, D. Vecchi, D. P. Curran, Tetra-
hedron Lett. 1996, 37, 6383–6386.
[15] B. P. Roberts, Chem. Soc. Rev. 1999, 28, 25–35.
zation, see: A. K. Saha, W. Schairer, C. Waychunas, C. V. C.
Prasad, M. Sardaro, D. A. Upson, L. I. Kruse, Tetrahedron
Lett. 1993, 34, 6017–6020.
[21] A. L. J. Beckwith, V. W. Bowry, W. R. Bowman, E. Mann, J.
Parr, J. M. D. Storey, Angew. Chem. Int. Ed. 2004, 43, 95–98.
[22] T. A. Richard, C. J. Collins, J. Am. Chem. Soc. 1939, 61, 147–
148.
[16] C. Tronche, F. N. Martinez, J. H. Horner, M. Newcomb, M. [23] V. T. Varlamov, E. T. Denisov, C. Chatgilialoglu, J. Org. Chem.
Senn, B. Giese, Tetrahedron Lett. 1996, 37, 5845–5848.
[17] M. Newcomb, S.-Y. Choi, J. H. Horner, J. Org. Chem. 1999,
64, 1225–1231.
[18] For 5-exo-trig radical cyclizations in uracil derivatives, see: D. J.
Fenick, D. E. Falvey, J. Org. Chem. 1994, 59, 4791–4799.
[19] a) D. P. Curran, N. A. Porter, B. Giese, Stereochemistry of Rad-
ical Reactions, VCH, Weinheim, 1995; b) P. Renaud, in: Radi-
cals in Organic Synthesis, vol. 1 (Eds.: P. Renaud, M. P. Sibi),
Wiley-VCH, Weinheim, 2001, pp. 400–415.
2001, 66, 6317–6322.
[24] a) C. Chatgilialoglu, V. I. Timokhin, A. B. Zaborovskiy, D. S.
Lutsyk, R. E. Prystansky, J. Chem. Soc. Perkin 2 2000, 577–
582; b) A. B. Zaborovskiy, D. S. Lutsyk, R. E. Prystansky, V. I.
Kopylets, V. I. Timokhin, C. Chatgilialoglu, J. Organomet.
Chem. 2004, 689, 2912–2919.
[25] X.-F. Zhu, H. J. Williams, A. I. Scott, J. Chem. Soc. Perkin
Trans. 1 2000, 2305–2306.
Received: June 3, 2005
Published Online: September 7, 2005
[20] Aldehyde 10 was previously prepared by Swern oxidation in
situ for further manipulation without isolation and characteri-
4648
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4640–4648