C5′-adenosinyl radical cyclization. A stereochemical investigation

Article abstract of DOI:10.1021/jo060197z

A variety of substituted 2′-deoxyadenosin-5′-yl radicals 3 were generated under different reaction conditions. Radicals 3 underwent intramolecular cyclization onto the C8-N7 double bond of the adenine moiety leading to aminyl radicals (5′S,8R)-4 and (5′R,8R)-4 and, eventually, to the corresponding cyclonucleosides 5 and 6. The effect of the solvent, the nature of the substituents, and the generation method of radicals 3 on the stereoselectivity of the C5′-radical cyclization have been considered. The observed increase of the (5′S)/(5′R) ratio by increasing the bulkiness of the R₁ group is explained in terms of steric repulsion between R₁ and the purine moiety which favors the C5′-endo conformation, whereas the effect of the water solvent in promoting the (5′R)-stereoselective cyclization is ascribed to intermolecular hydrogen bonding stabilizing the C5′-exo confomation.

Full text of DOI:10.1021/jo060197z

C5′-Adenosinyl Radical Cyclization. A Stereochemical Investigation
Maria Luisa Navacchia,† Chryssostomos Chatgilialoglu,† and Pier Carlo Montevecchi*,‡
ISOF, Consiglio Nazionale delle Ricerche, Via P. Godetti 101, 40129 Bologna, Italy, and Dipartimento di
Chimica Organica “A. Mangini”, UniVersita` di Bologna, Viale Risorgimento 4, 40136 Bologna, Italy
monteVec@ms.fci.unibo.it
ReceiVed January 30, 2006
A variety of substituted 2′-deoxyadenosin-5′-yl radicals 3 were generated under different reaction
conditions. Radicals 3 underwent intramolecular cyclization onto the C8-N7 double bond of the adenine
moiety leading to aminyl radicals (5′S,8R)-4 and (5′R,8R)-4 and, eventually, to the corresponding
cyclonucleosides 5 and 6. The effect of the solvent, the nature of the substituents, and the generation
method of radicals 3 on the stereoselectivity of the C5′-radical cyclization have been considered. The
observed increase of the (5′S)/(5′R) ratio by increasing the bulkiness of the R1 group is explained in
terms of steric repulsion between R1 and the purine moiety which favors the C5′-endo conformation,
whereas the effect of the water solvent in promoting the (5′R)-stereoselective cyclization is ascribed to
intermolecular hydrogen bonding stabilizing the C5′-exo confomation.
Introduction
generated by 1,6-radical translocation (1,6-RT) from the C8
position. C8-Radicals were in turn produced from the corre-
sponding 8-bromo derivative 1a by γ-radiolysis3 (Scheme 1,
method A) or photolysis (Scheme 1, method B).4,5 Thus, radicals
3a generated in water under γ-radiolysis conditions gave the
two diasteromeric isomers (5′R)-6a and (5′S)-6a in a (5′S)/(5′R)
) 15:85 ratio (Table 1, entry 1).3 A similar (5′S)/(5′R) ) 18:82
Both exogenous agents, i.e., ionizing radiations, UV-rays and
mutagenic chemical agents, and endogenous agents, as reactive
oxygen species (ROS), can be responsible for the radical damage
of DNA. Among ROS, highly diffusible hydroxyl radicals (•OH)
can cause chemical modification to DNA either by radical
addition to the base or hydrogen abstraction from one of the
five positions of the sugar unit. For steric reasons, the H5′
hydrogen atom appears to be the most accessible one. The
resulting C5′ radical can repair itself by hydrogen abstraction
from glutathione, or can lead to DNA chemical modification
or strand breakage.1
Among decomposition products of DNA exposed to ionising
radiations, Dizdaroglu and co-workers reported the formation
of cyclonucleotides characterized by an additional bond between
the sugar unit and the base moiety.2 These lesions are due to
the intermediacy of C5′ radicals through intramolecular cy-
clization onto the C8 position of the adenine unit. It has been
reported that the radical cyclization is (5′R)-stereoselective, with
a (5′S)/(5′R) ) 35:65 ratio.
(1) For reviews, see: (a) Cadet, J.; Douki, T.; Gasparutto, D.; Ravanat,
J.-L. Radiat. Phys. Chem. 2005, 72, 293-299. (b) Cadet, J; Douki, T.;
Ravanat, J.-L. In Redox-Genome Interactions in Health and Disease; Fuchs,
J., Podda, M., Packer, L., Eds.; Dekker: New York, 2003; pp 143-189.
(c) Dizdaroglu, M.; Jaruga, P.; Rodriguez, H. In Critical ReViews of
OxidatiVe Stress and Aging: AdVances in Basic Science Diagnostics and
InterVention; Cutler, R. G., Rodriguez, H., Eds.; World Scientific: London,
2003; Vol. 1, pp 165-189. (d) Dizdaroglu, M.; Jaruga, P.; Birincioglu,
M.; Rodriguez, H. Free Radic. Biol. Med. 2002, 32, 1102-1115. (e)
Chatgilialoglu, C.; O’Neill, P. Exp. Gerontol. 2001, 36, 1459-1471. (f)
Pogozelski, W. K.; Tullius, T. D. Chem. ReV. 1998, 98, 1089-1107. (g)
Pratviel, G.; Bernadou, J.; Meunier B. Angew. Chem., Int. Ed. Engl. 1995,
34, 746-769. (h) Goldberg, I. H., Acc. Chem. Res. 1991, 24, 191-198.
(2) Dizdaroglu, M.; Jaruga, P.; Rodriguez, H. Free Radic. Biol. Med.
2001, 30, 774-784. Dirksen, M. L.; Blakely, W. F.; Holwitt, E.; Dizdaroglu,
M. Int. J. Radiat. Biol. 1988, 54, 195-204. Dizdaroglu, M.; Dirksen, M.
L.; Jiang, H. X.; Robbins, J. H. Biochem. J. 1987, 241, 929-932.
(3) (a) Chatgilialoglu, C.; Guerra, M.; Mulazzani, Q. C. J. Am. Chem.
Soc. 2003, 125, 3839-3848. (b) Flyunt, R.; Bazzanini, R., Chatgilialoglu,
C.; Mulazzani, Q. C. J. Am. Chem. Soc. 2000, 122, 4225-4226.
(4) Jimenez, B. L.; Encinas, S.; Miranda, A. M.; Navacchia, M. L.;
Chatgilialoglu, C. Photochem. Photobiol. Sci. 2004, 3, 1042-1046.
(5) Romieu, D.; Gasparruto, D.; Cadet, J. J. Org. Chem. 1998, 63, 5245-
5249.
The same (5′R)-stereoselectivity was found in model studies
concerning the fate of C5′-radical of 2′-deoxyadenosine (3a)
* To whom correspondence should be addressed. Fax: (+39) 051 209 3654.
† Consiglio Nazionale delle Ricerche.
‡ Universita` di Bologna.
10.1021/jo060197z CCC: $33.50 © 2006 American Chemical Society
Published on Web 05/16/2006
J. Org. Chem. 2006, 71, 4445-4452
4445

Navacchia et al.
SCHEME 1. Chemical Studies on the Fate of 2′-Deoxyadenosin-5′-yl Radicals 3 Generated by 1,6-Radical Translocation
(1,6-RT) under a Variety of Experimental Conditionsa
a Method: (A) γ-rays, water; (B) hν, water or acetonitrile; (C) (Me3Si)3SiH, AIBN, fluorobenzene or acetonitrile, sealed tube, 85 °C; (D) (Me3Si)3SiH,
AIBN, refluxing fluorobenzene.
ratio was found when 1a was photolyzed in water solution
(Table 1, entry 11), while a (5′S)/(5′R) ) 37:63 ratio has been
reported in acetonitrile solvent (Table 1, entry 12).4
In sharp contrast with the above results, indicating a large
preference for the (5′R) configuration, we have recently reported6
that radicals 3b generated by tris(trimethylsilyl)silyl radical
addition to the 2′-deoxyadenosine-5′-carbaldehyde 9 (Scheme
2, method E), led to a mixture of products (5′S,8R)-5b and (5′S)-
6b in a completely (5′S)-stereoselective manner (Table 1, entry
2).
The discrepancy in results prompted us to investigate the
factors affecting the stereoselectivity of the C5′-radical cycliza-
tion. The effect of the R, R1, and R2 substituents, the nature of
the solvent, and the generation method of C5′-radicals 3 have
been considered. As reported here, we found that the stereo-
selectivity only depends on the C5′-substituent hindrance and
the nature of the solvent.
2′-deoxyadenosine-5′-carbaldehyde 9 (Scheme 2, method F) and
by 1,6-RT from the corresponding 2′-deoxyadenosin-8-yl radi-
cals 2 (Scheme 1). These latter were in turn obtained from the
8-bromo derivatives 1 both by photolysis of the C8-Br bond
(Scheme 1, method B) and by bromine atom abstraction by silyl
radicals (Scheme 1, methods C and D).
Tributyltin Radical Addition to the 2′-Deoxyadenosine-
5′-carbaldehyde 9. According with the methodology of Ueda7
for the cyclization of adenosine-5′-carbaldehyde and Cadet8 for
the cyclization of thymidine-5′-carbaldehyde, radicals 3c were
generated by addition of tributyltin radicals to the carbonyl
oxygen atom of the 5′-carbaldehyde 9. The reaction was carried
out in refluxing fluorobenzene under argon atmosphere for 2 h
(Scheme 2, method F). The reaction mixture was filtered on
1
silica gel column and analyzed by HPLC-MS and H NMR.
Cyclonucleosides (5′S,8R)-5d and (5′S)-6d and the 2′-deoxy-
adenosine derivative 7d were found as the exclusive reaction
(6) Navacchia, M. L.; Manetto, A.; Montevecchi, P. C.; Chatgilialoglu,
C. Eur. J. Org. Chem. 2005, 4640-4648.
Results and Discussion
(7) Sugawara, T.; Otter, B. A.; Ueda, T. Tetrahedron Lett. 1988, 29,
75-78.
The origins of the observed stereoselectivity of C5′-radical
cyclization were extrapolated from the behavior exhibited by
radicals 3 generated both by tributyltin radical addition to the
(8) Romieu, A.; Gasparutto, D.; Molko, D.; Cadet, J. J. Chem. Soc.,
Perkin Trans. 1 1999, 1257-1263.
4446 J. Org. Chem., Vol. 71, No. 12, 2006

C5′-Adenosinyl Radical Cyclization
TABLE 1. Stereochemical outcome of the Cyclization of 2′-Deoxyadenosin-5′-yl Radicals 3
products, relative yields (%)
5′S/5′R
entry
substituents
R ) R1 ) R2 ) H
radical 3
methoda
solvent
H2O
PhF
PhF
PhF
PhF
MeCN
PhF
MeCN
MeCN
MeCN
H2O
MeCN
MeCN
MeCN
MeCN
MeCN/H2O
MeCN/H2O
(5′S)-6
(5′S)-5
(5′R)-6
(5′R)-5
7
ratio
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
3a
3b
3c
3e
3e
3e
3f
A
E
F
15
52
32
39
60
65
55
70
38
47
7
24
h
h
52
25
25
b
48
60
39
b
b
b
b
b
b
b
b
b
b
b
b
b
85
b
b
4
7
8
7
8
32
38
31
41
g
b
b
b
4
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
8
15:85c
100:0d
100:0e
90:10e
90:10e
90:10e
90:10e
90:10e
55:45e
55:45e
18:82f
37:63f
50:50e,g
50:50e,g
75:25e
30:70e
30:70e
R ) COPh, R1 ) (Me3Si)3Si, R2 ) TBDMS
R ) COPh, R1 ) Bu3Sn, R2 ) TBDMS
R ) COPh, R1 ) R2 ) TBDMS
R ) COPh, R1 ) R2 ) TBDMS
R ) COPh, R1 ) R2 ) TBDMS
R ) H, R1 ) R2 ) TBDMS
R ) H, R1 ) R2 ) TBDMS
R ) R1 ) H, R2 ) TBDMS
R ) R1 ) R2 ) H
D
C
C
C
C
C
C
B
B
B
B
B
B
B
14
33
27
38
22
30
15
b
17
30
30
30
15
15
3f
3g
3a
3a
3a
3a
3g
3f
R ) R1 ) R2 ) H
R ) R1 ) R2 ) H
R ) R1 ) R2 ) H
R ) R1 ) H, R2 ) TBDMS
R ) H, R1 ) R2 ) TBDMS
R ) R1 ) R2 ) H
g
18
60
60
3a
3g
R ) R1 ) H, R2 ) TBDMS
a A: 8-bromo-2′-deoxyadenosine 1, γ-radiolysis. B: 8-bromo-2′-deoxyadenosine 1, photolysis. C: 8-bromo-2′-deoxyadenosine 1, (Me3Si)3SiH, AIBN,
85 °C, sealed tube. D: 8-bromo-2′-deoxyadenosine 1, (Me3Si)3SiH, AIBN, refluxing solvent. E: aldehyde 9, (Me3Si)3SiH, AIBN, 85 °C. F: aldehyde 9,
Bu3SnH, AIBN, 85 °C. b Not detected. c Reference 3. d Reference 6. e Present work. f Reference 4. g Calculated by extrapolation at zero time. h Time dependent
yield.
SCHEME 2. Products from Tributyltin-Promoted Reaction of Aldehyde 9
products in 60:32:8 ratio (Table 1, entry 3). Their formation
can be easily explained through the intermediacy of the
tributyltin hydride-promoted radicals 3c as outlined in Scheme
2. Hydrogen atom abstraction from Bu3SnH led to 7c, whereas
cyclization onto the C8-N7 double bond gave the radical 4c,
leading in turn to 5c by H-atom abstraction and 6c by oxidation,
J. Org. Chem, Vol. 71, No. 12, 2006 4447

Navacchia et al.
possibly through a 2-cyanoisopropyl radical-mediated reaction.9
Hydrolysis of the labile O-Sn bond during the workup
eventually afforded compounds 5d, 6d, and 7d.
conditions that ensured the absence of air. Under these non-
oxidative conditions aminyl radicals 4e led to both cyclonucleo-
sides 6e and 5e, the first one possibly through the above-
mentioned 2-cyanoisopropyl radical promoted reaction,9 the
second one by H-atom abstraction from [(Me3Si)3SiH].
1H NMR and HPLC-MS analyses of the reaction mixture
showed the cyclonucleosides (5′S,8R)-5e, (5′S)-6e, (5′R,8R)-
5e, and (5′R)-6e and the 2′-deoxyadenosine derivative 7e in a
39:39:4:4:14 ratio (Table 1, entry 4), this suggesting that the
cyclization is (5′S,8R)-stereoselective.
Compound (5′S,8R)-5e could be separated by column chro-
matography and characterized by spectral analysis. The structure
of the minor isomer (5′R,8R)-5e was assigned on the basis of
the HPLC-MS spectrum and the 1H NMR analysis which
indicated the specific H1′ signal at δ 6.10.
The cyclonucleoside (5′S)-6d has been previously reported.6
Compound 5d was isolated by silica gel column chromatography
and fully characterized by spectral analysis and chemical
evidences. The (5′S,8R) configuration was easily assigned on
the basis of the 1H NMR spectrum. The value of J4′,5′ ) 4.5 Hz
coupling constant is typical for the (5′S)-configuration, whereas
a coupling constant J4′,5′ ) 0-1.0 Hz value would have been
expected for the (5′R)-isomer.10 On the other hand, the large
J5′,8 ) 7.5 coupling constant is indicative for protons in a trans
diaxial orientation, in agreement with the (8R)-configuration.
NOE experiments supported the (5′S,8R) configuration. Irradia-
tion on the H8 signal caused a significant NOE enhancement
of the H3′ signal (3%), besides smaller enhancements of H2′
(1%) and H5′ (1%) signals. Moreover, irradiation on the H5′
signal caused an enhancement of H8 (1%) and H4′ (3.5%), but
not H3′. A further support to the structure of 5d came from air
oxidation in chloroform solution which led quantitatively to 6d.
Attempts to aromatize the five-membered ring of the purine
unit of 5d with chloroanil, a chemoselective oxidizing agent
for hydroaromatic rings, unexpectedly led to the ketone 8.
Thermal Reaction of the 8-Bromo-2′-deoxyadenosine 1
with Silyl Radicals. Reactions of the 8-bromo derivatives 1
with silyl radicals were generally carried out at 85 °C in a sealed
tube in air-saturated fluorobenzene or acetonitrile solutions in
the presence of a 5-fold excess of tris(trimethylsilyl)silane [(Me3-
Si)3SiH] and equimolar amounts of AIBN (Scheme 1, method
C).
A further support to the structure of the above compounds
(5′S,8R)-5e and (5′R,8R)-5e came from the oxidation of the
reaction mixture with chloroanil, leading to the disappearance
of 5e and the formation of a mixture of (5′S)-6e, (5′R)-6e and
1
7e in the expected 78:8:14 ratio, as determined by H NMR
analysis.
Similar results were obtained from the reaction of 1f carried
out in fluorobenzene. HPLC-MS and 1H NMR analysis showed
the presence of (5′S)-6f, (5′R)-6f, and 7f as the exclusive
products in 55:7:38 ratio (Table 1, entry 7). Also in this case
the cyclization was found to be (5′S)-stereoselective with a (5′S)/
(5′R) ) 90:10 ratio. Structural assignment arose from HPLC-
1
MS and H NMR analysis. The (5′S) and (5′R) configurations
of the chiral center were assigned based on the values of
coupling constant J4′,5′ ) 6.0 and 0 Hz, respectively. Unfortu-
nately, compounds (5′S)-6f and (5′R)-6f were not fully char-
acterized due to the impossibility of obtaining pure samples by
column chromatography.
As suggested by results, the stereochemistry of the cyclization
of radicals 3 does not depend on their generation method, since
radicals 3b and 3c, directly generated by addition to the aldehyde
9 of silyl and stannyl radicals, respectively, and radicals 3e and
3f, indirectly generated by 1,6-RT, cyclize in the same 5′S-
stereoselective manner (Table 1, entries 2-5, 7).
Also the polarity of the solvent does not play any role, as
shown by the results obtained from the reaction of 1e and 1f
carried out in acetonitrile. Compounds (5′S)-6e,f, (5′R)-6e,f, and
7e,f were found as the exclusive reaction products, with the
same (5′S)/(5′R) ) 90:10 ratio (Table 1, entries 6 and 8).
On the contrary, the reaction of both 1a and 1g carried out
in acetonitrile led to a mixture of cyclized products (5′S)-6a,g
and (5′R)-6a,g in an almost nonstereoselective manner, irrespec-
tive of the steric hindrance of the C3′-OR2 substituent (Table
1, entries 8 and 9). The overall results led us to the conclusion
that the stereoselectivity is determined by the bulkiness of the
C5′-OR1 substituent, since the (5′S)/(5′R) ratio passed from
100:0 to 90:10 and 55:45 by decreasing the steric demand of
the R1 substituent from Bu3Sn or (Me3Si)3Si to TBDMS and H
(Table 1, entries 2-10).
The reaction of 1e gave a 60:7:33 mixture of cyclonucleosides
(5′S)-6e and (5′R)-6e and the 2′-deoxyadenosine derivative 7e
(Table 1, entry 5), as determined by 1H NMR and HPLC-MS
analysis of the reaction mixture. The (5′S)-configuration of the
major isomer was assigned on the basis of the J4′,5′ ) 6.5 Hz
value, whereas the (5′R) configuration of the minor isomer was
assigned on the basis of J4′,5′ ) 0 Hz.10 Subsequent silica gel
column chromatography allowed us to separate a pure sample
of (5′S)-6e, which was fully characterized. The formation of
reaction products can be easily explained as follows: AIBN-
promoted tris(trimethylsilyl)silyl radicals11 abstract the bromine
atom from 1e; the resulting C8-radicals 2e can undergo hydrogen
abstraction, leading to the 2′-deoxyadenosine derivative 7e, or
1,6-RT to C5′-radicals 3e. Radicals 3e can cyclize onto the C8-
N7 double bond leading to aminyl radicals 4e, then to cyclo-
nucleosides 6e by dioxygen-promoted oxidation. The cyclization
was found to be strongly (5′S)-stereoselective, with a (5′S)/(5′R)
) 90:10 ratio (Scheme 1).
In principle, the 2′-deoxyadenosine derivative 7e could be
also derived from radicals 3e by H-atom abstraction, but this
possibility seems to be unlikely. In fact, we have reported6 that
analogous radicals 3b are incapable of abstracting a hydrogen
atom under the reaction conditions employed (see Table 1, entry
2).
The role of dioxygen in promoting the formation of 6e was
highlighted when the reaction of 1e was repeated in refluxing
fluorobenzene under argon atmosphere (Scheme 1, method D),
Photolysis of the 8-Bromo-2′-deoxyadenosine 1. As men-
tioned above, some of us have recently reported the photolysis
of 1a carried out in acetonitrile.4 HPLC-MS analysis of the
reaction mixture detected the cyclonucleosides 6a in a (5′R)-
stereoselective mode with a (5′S)/(5′R) ) 37:63 ratio (Table 1,
entry 12).
(9) Beckwith, A. L. J.; Bowry, V. W.; Bowman, W. R.; Mann, E.; Parr,
J.; Storey, J. M. D. Angew. Chem., Int. Ed. 2004, 43, 95-98.
(10) For J4′,5′ values of 5′,8-cyclo-2′-deoxyadenosines see refs 3, 4, 15,
and 16.
(11) Chatgilialoglu, C. Organosilanes in Radical Chemistry; Wiley:
Chichester, 2004.
This finding seems to be in contrast with our suggestion about
the role of the C5′-OR1 substituent, since in the case of R1 )
4448 J. Org. Chem., Vol. 71, No. 12, 2006

C5′-Adenosinyl Radical Cyclization
TABLE 2. Product Relative Yields (%) from Photolysis of 1a and
1g in Acetonitrile at Different Irradiation Times
Similar results were obtained by photolysis of 1g in aceto-
nitrile. 1H NMR analyses at different irradiation times showed
the presence of the cyclonucleosides (5′R)-6g and (5′S)-6g, the
2′-deoxyadenosine derivative 7g, and an unknown compound
B in different relative yields. Also in this case, the (5′R)/(5′S)
ratio was found to increase with the irradiation time (Table 2).
Extrapolation at zero time of the linear relationship (R/S) vs
time gave a (5′S)/(5′R) ) 50:50 ratio (Figure 1 and Table 1,
entry 14). As for compound A, compound B could not be
isolated from the reaction mixture and it remains uncharacter-
ized. However, 1H NMR spectral analysis allowed to establish
a structural similarity between A and B (see the Experimental
Section).
time
R/S
R/S
(min) (5′R)-6a (5′S)-6a
A
ratio (5′R)-6g (5′S)-6g
B
ratio
15
30
50
100
140
180
48
48
47
29
25
13
30
25
20
7
5
2
22 1.6
26 2.0
33 2.4
64 3.9
70 5.0
85 6.5
49
50
48
48
44
41
38
34
27
21
9
1.2
12 1.3
18 1,4
25 1.8
35 2.1
The effect of the bulkiness of the C5′-OR1 substituent in
favoring the (5′S)-cyclization was definitively proved by the
photolysis of 1f in acetonitrile, which led to a mixture of the
cyclized products (5′S)-6f and (5′R)-6f and the 2′-deoxyadeno-
sine derivative 7f in a 52:18:30 ratio (Table 1, entry 15). It is
noteworthy that the relative yields of the cyclized products (5′S)-
6f and (5′R)-6f were found unchanged with the irradiation time
up to 90 min, in contrast with the behavior exhibited by the
cyclonucleosides 6a and 6g which lead to compounds A and
B, respectively, under the reaction conditions. As a matter of
fact, we can infer that the presence of the hydroxy group in the
C5′-position is essential for the bromine-promoted decomposi-
tion of the cyclonucleosides 6a and 6g.
Effect of the Bulkiness of the C5′-OR1 Substituent. The
observed effect of the bulkiness of the C5′-OR1 substituent can
be explained in terms of steric and stereoelectronic factors. The
formation of the C5′-C8 bond occurs when the orbital
containing the unpaired electron and the purine moiety form
an angle of 109°.12 This situation is achieved when the C5′
radical and the purine base are in a pseudo-diaxial orientation3a
(Scheme 3).
In this conformation, the purine base can be syn (a) or anti
(b) with respect to the sugar ring. The cyclization in the anti
conformation is expected to lead to the (8R)-configuration
through a chair transition state: this would be preferred with
respect to the syn cyclization, leading to the (8S)-configuration
through a boat transition state.3 In agreement, we always found
the cyclized products 5 in the (8R)-configuration. Therefore,
we can confidently assume that the cyclization occurs in the
anti conformation.
Analogously, the C5′-OR1 group can be inside (C5′-endo
conformation) (c) or outside (C5′-exo conformation) (d) with
respect to the sugar unit. In the absence of steric effects, the
cyclization can occur in both C5′-endo and C5′-exo conforma-
tions, leading to radicals (5′S,8R)-4 and (5′R,8R)-4, respectively,
in a nonstereoselective mode. In contrast, steric effects due to
the steric repulsion between bulky R1 substituents and the purine
moiety should favor the C5′-endo conformation, from which
the (5′S,8R)-4 radical is formed.
FIGURE 1. (5′R)/(5′S) ratio vs time from photolysis of 1.0 mM
acetonitrile solutions of 1a and 1g.
H a lack of stereoselectivity should be expected. This disagree-
ment prompted us to reinvestigate this reaction.
Photolyses of 1.0 mM deaerated acetonitrile solutions of 1a
were carried out using a 125 W medium-pressure mercury lamp.
The reaction mixtures were monitored at different irradiation
times by both HPLC-MS and 1H NMR. In all cases 1H NMR
analysis showed the presence of the cyclonucleosides (5′R)-6a
and (5′S)-6a, 2′-deoxyadenosine 7a, and an unknown compound
(A) as the exclusive reaction products in time-depending relative
yields. In particular, the (5′R)/(5′S) ratio (Table 2 and Figure
1) increased with the irradiation time from 1.6 at 15 min to 6.5
at 180 min. Obviously, none of these values can be representa-
tive of the authentic (5′S)/(5′R) stereoselectivity. We reasoned
that the correct stereoselectivity must be evaluated from the
linear relationship (R/S) vs time by extrapolation at zero time.
As a result, we calculated a (5′S)/(5′R) ) 50:50 ratio (Figure 1
and Table 1, entry 13). This value strengthens our suggestion
that in the absence of steric effects played by the C5′-OR1
substituent the cyclization is nonstereoselective. Also, 1H NMR
analysis showed the relative yield of compound A to increase
with the irradiation time at the expense of the cyclized products
(5′S)-6a and (5′R)-6a (Table 2). We concluded that the cyclized
products (5′S)-6a and (5′R)-6a are not stable under the reaction
conditions, both leading to A, but at different rates.
A brief investigation about the formation of A was carried
out. Unexpectedly, photolysis of a 1.0 mM acetonitrile solution
of a 1:1 mixture of (5′S)-6a and (5′R)-6a did not lead to
compound A at all. In the presence of 0.2 mM bromine, the
above solution was found to be stable in the dark within 2 h,
whereas under photolytic conditions almost complete decom-
position occurred within 1 h with formation of compound A as
the major product. Attempts to obtain a pure sample of A by
either silica gel or C-18 reversed-phase column chromatography
failed. Even though the analysis of a ca. 80% purity sample
allowed us to obtain a confident 1H NMR spectrum (see
Experimental Section), an unambiguous mass spectrum could
not be obtained by HPLC-MS analysis. Therefore, the structure
of compound A and its formation mechanism remain unknown.
Effect of the Water Solvent. Despite our considerations,
radicals 3a generated from the 8-bromo derivative 1a both by
γ-radiolisys3 and photolysis4,5 in water have been reported to
cyclize in a (5′R)-stereoselective mode (Table 1, entries 1 and
10). Moreover, in an earlier work, (5′R)-5′,8-cyclo-2′-deoxy-
adenosine 6a was identified as one of the principal products of
γ-radiolysis in water of 2′-deoxyadenosine 7a.13 In this case,
(12) Beckwith, A. L. J.; Easton, C. J.; Seredis, A. K. J. J. Chem. Soc.,
Chem. Commun. 1980, 482. Baldwin, J. E. J. Chem. Soc., Chem. Commun.
1976, 734.
J. Org. Chem, Vol. 71, No. 12, 2006 4449

Navacchia et al.
SCHEME 3. Conformational Effects on the Stereochemistry
of the 2′-Deoxyadenosin-5′-yl Radical Cyclization
mixture gave the cyclonucleosides 6g in the same (5′S)/(5′R)
) 30:70 time-independent ratio (Table 1, entry 17).
To explain the role of water in favoring the (5′R)-cyclization,
it might be suggested that in this protic solvent the C5′-exo
conformation of radicals 3a,g is preferred due to the formation
of intermolecular hydrogen bonds between the C5′-OH hydrogen
atom and the oxygen atom of the sugar ring and/or the N1
nitrogen atom (Scheme 3e).
The Case of the C5′-Radical of Adenosine (Ribo Form).
In light of our suggestion on the role of the water solvent and
the steric demand of the C5′-substituent, the hitherto unrational-
ized results reported in the literature about the stereoselectivity
of the adenosin-5′-yl radical cyclization can be explained. Ueda
and co-workers7 reported that C5′ radicals directly generated
in benzene solvent from the C5′-aldehyde through a tributyltin
hydride mediated reaction led to the corresponding cyclonucleo-
side in a (5′S)-stereoselective mode with a (5′S)/(5′R) ) 100:0
ratio. Based on our theory, this can be explained in terms of
steric effects played by the bulky tributyltin group which favors
the C5′-endo conformation. On the other hand, an opposite
(5′R)-stereoselectivity was found in water by generating C5′-
radicals from both adenosine14 and 8-bromoadenosine15 under
γ-radiolysis conditions. In both cases we can suggest that the
water solvent stabilizes the C5′-exo conformation through
intermolecular hydrogen bonding.
Conclusions
2′-Deoxyadenosin-5′-yl radicals 3 were generated both di-
rectly by tributyltin radical addition to the carbonyl oxygen atom
of the aldehyde 9 and by 1,6-RT from the corresponding C8-
radicals. These latter were in turn generated starting from the
corresponding 8-bromo derivatives 1 both under thermal condi-
tions through bromine atom abstraction by silyl radicals and
by photolysis of the C8-Br bond. In all cases radicals 3
underwent intramolecular cyclization onto the C8-N7 double
bond to aminyl radicals (5′S,8R)-4 and (5′R,8R)-4, finally
leading to the cyclonucleosides (5′S)-6 and (5′R)-6 and, under
anaerobic conditions, (5′S,8R)-5 and (5′R,8R)-5 (Scheme 1 and
Table 1). The (5′S)/(5′R) ratio was found independent of the
polarity of the solvent (fluorobenzene or acetonitrile), the
method of generation of radicals 3, and the nature of the C3′-
OR2 substituent (R2 ) H or TBDMS). Instead, it strongly
depends on the steric hindrance of the C5′-OR1 substituent (R1
) H, TBDMS, or Bu3Sn). The increase of the (5′S)/(5′R) ratio
by increasing the bulkiness of the R1 group is explained in terms
of steric repulsion between the R1 group and the purine moiety
which disfavors the C5′-exo conformation of the radical 3
(Scheme 3).
radicals 3a were directly generated by H5′-atom abstraction by
hydroxyl radicals.
Since the (5′S)/(5′R) ratio was found to be independent of
the irradiation time, it represents the real (5′S)/(5′R) cyclization
ratio. In this light, it might be concluded that the observed (5′R)-
stereoselectivity should be actually due to some effect played
by the water solvent.
On the contrary, radicals 3 generated by photolysis in water
solvent cyclized in a (5′R)-stereoselective manner. The effect
of the water is explained assuming that intermolecular hydrogen
bonding between a water molecule, the C5′-OH and the oxygen
atom of the sugar ring, or the N1 atom of the purine unit,
stabilizes the C5′-exo confomation (Scheme 3).
A further support to the water effect came from the photolysis
1
of 1a carried out in a 1:1 acetonitrile/water mixture. H NMR
analysis showed the formation of the cyclonucleosides 6a in a
(5′S)/(5′R) ) 30:70 ratio (Table 1, entry 16), which is between
the 50:50 ratio found in neat acetonitrile and 18:82 ratio in neat
water (Table 1, entries 12 and 10). As in the reaction carried
out in neat water, the (5′S)/(5′R) ratio was found independent
of the irradiation time, and no compound A was formed at all.
Similarly, the photolysis of 1g in a 1:1 acetonitrile/water
Experimental Section
General Methods. HPLC analyses were performed on a C18
and C8 column (4.6 × 150 mm, 5 µm) with a linear gradient water/
acetonitrile at a 0.6 mL/min flow rate, detection at λ 260 nm.
(14) Raleigh, J. A.; Fuciarelli, A. F. Radiat. Res. 1987, 35-44.
(15) Chatgilialoglu, C.; Duca, M.; Ferreri, C.; Guerra, M.; Ioele, M.;
Mulazzani, Q. C.;Strittmatter, H.; Giese, B. Chem Eur. J. 2004, 10, 1249-
1255.
(13) Cadet, J.; Berger, M. Int. J. Radiat. Biol. 1985, 47, 127-143.
Mariaggi, N.; Cadet, J.; Teoule, R. Tetrahedron 1976, 32, 2385-2387.
4450 J. Org. Chem., Vol. 71, No. 12, 2006

C5′-Adenosinyl Radical Cyclization
Starting Materials. 8-Bromo-2′-deoxyadenosine (1a) was bought
and used as it was. N-Benzoyl-O3′-(tert-butyldimethylsilyl)-5′-oxo-
2′-deoxyadenosine (9) was synthesized as we previously reported
in the literature.6
0.10 (s; 6H); 13C NMR (100.6 MHz, CDCl3) δ -4.5, -4.45, 18.3,
26.0, 40.9, 63.4, 74.0, 88.9, 90.7, 121.0, 127.5, 150.0, 151.4, 154.4;
MS (ES+) 444 (M + 1); MS2 (444) 214. Anal. Calcd for C16H26-
BrN5O3Si: C, 43.24; H, 5.90; Br, 17.98; N, 15.76. Found: C, 43.10;
H, 5.90; Br, 18.00; N, 15.70.
Synthesis of 8-Bromo-O3′,O5′-bis(tert-butyldimethylsilyl)-2′-
deoxyadenosine (1f). Commercial 8-bromo-2′-deoxyadenosine (330
mg, 1.0 mmol) was dissolved in dry DMF (10 mL). Imidazole (200
mg, 3.0 mmol), (dimethylamino)pyridine (30 mg, 0.25 mmol), and
tert-butyldimethylsilyl chloride (TBDMSCl) (375 mg, 2.5 mmol)
were added to the solution. The reaction mixture was allowed to
stir at room temperature for 3 h, and then water was added (10
mL). The resulting mixture was extracted with ethyl acetate (2 ×
10 mL). The organic layer was dried (Na2SO4), the solvent
evaporated under reduced pressure, and the residue purified on silica
gel column. The elution with ethyl acetate/hexane 80:20 yielded
the desired product as a yellow foam (530 mg, 0.95 mmol, 95%
yield): 1H NMR (400 MHz, CDCl3) δ 8.24 (s; 1H); 6.34 (t, J1′,2′
) J1′,2′′ ) 7.0 Hz; 1H), 6.14 (br s; 2H), 4.86 (ddd; J2′,3′ ) 4.0 Hz;
J2′′,3′ ) 5.2 Hz; J3′,4′ ) 3.6 Hz; collapsing to dd upon irradiation at
δ 2.22; J3′,4′ ) 3.6 Hz, J2′′,3′ ) 5.2 Hz; collapsing to dd upon
irradiation at δ 3.94, J2′,3′ ) 4.0 Hz; J2′′,3′ ) 5.2 Hz; 1H), 3.90-
3.95 (m; 2H), 3.63-3.69 (m; 2H), 2.22 (ddd, J2′,2′′ ) 12.0 Hz, J2′,3′
) 4.0 Hz, J1′,2′ ) 7.0 Hz; 1H), 0.95 (s; 9H), 0.90 (s; 9H), 0,00 (s;
6H), -0,05 (s; 6H); 13C NMR (100.6 MHz, CDCl3) δ - 5.3, -5.2,
-4.5, -4.45, 18.3, 18.6, 26.1, 37.0, 62.8, 72.5, 86.5, 88.0, 120.7,
128.5, 151.1, 152.3, 154.4; MS (ES+) m/z 560 (M + 1), 582 (M
+ 23); MS2 (560) 216; MS (ES-) 558 (M - 1). Anal. Calcd for
C22H40BrN5O3Si2: C, 47.30; H, 7.22; Br, 14.30; N, 12.54. Found:
C, 47.45; H, 7.20; Br, 14.35; N, 12.50.
Reaction of N-Benzoyl-O3′-(tert-butyldimethylsilyl)-5′-oxo-2′-
deoxyadenosine (9) with Bu3SnH. The aldehyde 9 (250 mg, 0.5
mmol) was dissolved in fluorobenzene (25 mL). AIBN (0.25 mmol,
40 mg) and Bu3SnH (0.27 mL, 1.0 mmol) were added, and the
solution was refluxed under argon atmosphere for 2 h. The solvent
was removed under reduced pressure and the crude mixture
analyzed by HPLC-MS and 1H NMR. 1H NMR showed the
exclusive presence of the cyclized products (5′S)-6d5 and (5′S,8R)-
5d and 2′-deoxyadenosine 7d16 in 32:60:8 ratio. Silica gel column
chromatography provided a pure sample of (5′S,8R)-5d as a pale
yellow foam.
(5′S,8R)-N-Benzoyl-O3′-(tert-butyldimethylsilyl)-5′,8-cyclo-
7,8-dihydro-2′-deoxyadenosine [(5′S,8R)-5d]: 1H NMR (400
MHz, CDCl3) δ 9.2 (br s, disappeared on D2O shake; 1H), 7.9 (s;
1H), 7.4-8.0 (m; 5H), 6.10 (d, J1′,2′ ) 6.0 Hz; 1H), 5.49 (d,
disappeared on D2O shake, JNH,8 ) 4.0 Hz; 1H), 4.89 (dd, J5′,8
)
7.5 Hz, JNH,8 ) 4.0 Hz; collapsed to doublet on D2O shake; 1H),
4.67 (dd, J2′′,3′ ) 7.0, J2′,3′ ) 2.0 Hz; 1H), 4.20 (d, J4′,5′ ) 4.5 Hz;
1H), 3.73 (dd, J4′,5′ ) 4.5, J5′,8 ) 7.5 Hz; 1H), 2.35 (1H, J2′,2′′
)
14.0, J2′′,3′ ) 7.0 Hz; 1H), 2.18 (ddd, J2′,2′′ ) 14.0, J1′,2′ ) 6.0, J2′,3′
) 2.0 Hz; 1H), 0.9 (s; 9H), 0.3 (s; 3H), 0.1 (s; 3H); irradiation at
δ 3.73 caused an enhancement of the signal at δ 4.20 (3.5%) at δ
4.89 (1%); irradiation at δ 5.49 caused an enhancement of the signal
at δ 4.67 (3%), δ 3.73 (1%), δ 2.35 (1%); 13C NMR (100.6 MHz,
CDCl3) δ -4.6, -4.5, 18.3, 26.0, 42.0, 69.7, 71.1, 81.9, 86.1, 122.1,
127.9, 128.2, 128.9, 132.8, 133.0, 135.2, 149.6, 160.6, 166.1; MS
(ES+) 470 (M + 1); MS2 (470) 338, MS3 (338) 320. Anal. Calcd
for C23H31N5O4Si: C, 58.82; H, 6.65; N, 14.91. Found: C.58.9;
H, 6.60; N, 14.95.
A solution of 5d (47 mg, 0.10 mmol) and chloroanil (37 mg,
0.15 mmol) in o-xylene (3 mL) was refluxed for 20 min.17 The
solvent was eliminated under reduced pressure and the residue
chromatographed on silica gel column. Gradual elution with
pentane/ethyl acetate gave the ketone 8 (40 mg, 0.09 mmol, 90%).5
Synthesis of 8-Bromo-N5-benzoyl-O3′,O5′-bis(tert-butyldi-
methylsilyl)-2′-deoxyadenosine (1e). 8-Bromo-O3′,O5′-bis(tert-
butyldimethylsilyl)-2′-deoxyadenosine 1f (400 mg, 0.70 mmol) was
dissolved in dry pyridine (7 mL). Benzoyl chloride (92 µL, 0.77
mmol) was added, and the solution was allowed to stir at room
temperature overnight. The reaction mixture was treated with water
(7 mL) and extracted with ethyl acetate (2 × 7 mL). The solvent
was evaporated under reduced pressure and the residue purified
on silica gel column by gradual elution with ethyl acetate/hexane.
The pure product was obtained as a white foam (170 mg, 0.26
mmol, 60%): 1H NMR (400 MHz, CDCl3) δ 8.73 (br s; 1H), 8.02
(s; 1H); 8.00-7.50 (m; 5H), 6.41 (d, J1′,2′ ) 7.0 Hz, 1H), 4.87 (m;
Reaction of 8-Bromo-2′-deoxyadenosine 1 with Tris(trimeth-
ylsilyl)silane [(Me3Si)3SiH]. General Procedure. A 10.0 mM
solution of the appropriate 8-bromo-2′-deoxyadenosine 1 (0.30
mmol), (Me3Si)3SiH (0.46 mL, 15 mmol), and AIBN (50 mg, 0.30
mmol) in fluorobenzene or acetonitrile (30 mL) was kept in a sealed
tube in a thermostatic bath at 90 °C for 2 h. The solvent was
removed under reduced pressure and the residue analyzed by
1H), 3.97 (m; collapsing to t upon irradiation at δ 4.87, J4′,5′
)
J4′,5′′ ) 6.0 Hz; 1H), 3.90 (m; 1H), 3.90 (dd, J5′,5′′ ) 11.2 Hz; J4′,5′
) 6.0 Hz; 1H), 3.71 (dd; J5′,5′′ ) 11.2 Hz; J4′,5′′ ) 6.0 Hz; 1H),
3.52 (dd; J2′,2′′ ) 13.0 Hz; J2′,3′ ) 6.0 Hz; 1H), 2.30 (ddd, collapsing
to dd upon irradiation at δ 6.41 J2′,2′′ ) 13.0 Hz, J2′,3′ ) 4.0 Hz,
J1′,2′ ) 7.0 Hz; 1H), 0.95 (s; 9H), 0.90 (s; 9H), 0.00 (s; 6H), -0,05
(s; 6H); 13C NMR (100.6 MHz, CDCl3) δ -5.2, -4.3, 18.3, 18.6,
26.0, 37.2, 62.8, 72.3, 88.2, 88.2, 128.0, 129.1, 129.1, 130.8, 134.1,
152.4, 162.6, 164.7; MS (ES+) 664 (M + 1); MS2 (664) 320. Anal.
Calcd for C29H44BrN5O4Si2: C, 52.55; H, 6.69; Br, 12.06; N, 10.57.
Found: C, 52.70; H, 6.65; Br, 12.10; N, 10.60.
1
HPLC-MS and H NMR.
From 1e in Fluorobenzene. 1H NMR and HPLC-MS analysis
of the reaction mixture evidenced the presence of cyclonucleosides
(5′S)-6e and (5′R)-6e and 2′-deoxyadenosine 7e16 as the only
detectable compounds in 60:7:33 ratio. Silica gel column chroma-
tography of the crude mixture led to the isolation of a pure sample
of the major isomer (5′S)-6e.
Synthesis of 8-Bromo-O3′-(tert-butyldimethylsilyl)-2′-deoxy-
adenosine (1g). 8-Bromo-O3′,O5′-bis(tert-butyldimethylsilyl)-2′-
deoxyadenosine 1f (560 mg, 1.0 mmol) was dissolved in dry THF
dry (20 mL). Trifluoroacetic acid/water 1:1 mixture (10 mL) was
added dropwise at 0 °C to the solution. The mixture was allowed
to stir at 0 °C for 1 h. The reaction mixture was then neutralized
by the addition of NaHCO3 (saturated solution). The resulting
mixture was extracted with ethyl acetate (2 × 50 mL). The solvent
was evaporated under reduced pressure/ and the residue was purified
on silica gel column. The target compound was obtained as a white
foam (310 mg, 0.7 mmol, 70% yield): 1H NMR (400 MHz, CDCl3)
(5′S)-N5-Benzoyl-O3′,O5′-bis(tert-butyldimethylsilyl)-5′,8-cy-
1
clo-2′-deoxyadenosine (5′S)-6e: H NMR (400 MHz, CDCl3) δ
8.74 (1H, s; 1H), 8.0 and 7.5-7.7 (m; 5H), 6.5 (d, J ) 4.5 Hz;
1H), 5.28 (d, J ) 6.5 Hz; 1H), 4.84 (dd, J ) 7.0 and 4.5 Hz; 1H),
4.6 (d, J4′,5′ ) 6.5 Hz; 1H), 2.56 (dd, J ) 13.2 and 7.0 Hz; 1H),
2.27 (dt, Jd ) 13.2, 5.0, Jt ) 4.5 Hz; 1H), 1.0 (s; 9H), 0.9 (s; 9H),
0.35 (s; 3H), 0.29 (s; 3H), 0.07 (s; 3H), 0.04 (3s; 3H); 13C NMR
(100.6 MHz, CDCl3) δ -4.6, -4.4, -4.3, 18.0, 18.9, 25.9, 26.1,
48.8, 66.3, 69.6, 85.9, 87.2, 128.2, 128.4, 129.2, 133.3, 133.3, 143.5,
148,6, 151,1, 154.2, 164.1; MS (ES+) 582 (M + 1), MS2 (582)
δ 8.65 (br s; 1H), 8.26 (s; 1H); 6.49 (br s; 2H), 6.42 (dd, J1′,2′
)
9.5 Hz, J1′,2′′ ) 5.6 Hz; 1H), 4.71 (d, J2′,3′ ) 5.0 Hz; 1H), 4.14 (br
s; 1H), 3.93 (part A of a ABX system, JAB ) 12.8 Hz, JAX ) 1.5
Hz; 1H), 3.74 (part B of a ABX system JAB ) 12.8 Hz, JBX ) 2.0
Hz; 1H), 3.0 (ddd; J2′′,2′ ) 13.0 Hz, J1′,2′ ) 9.5 Hz, J2′,3′ ) 5.0, Hz;
1H), 2.18 (dd, J2′′,2′ ) 13.0 Hz, J1′,2′′ ) 5.6,Hz, 1H), 0.95 (s; 9H),
(16) Zhu, H. F.; Williams, H. J.; Scott, A. I. J. Chem. Soc., Perkin Trans.
1 2000, 2305-2306.
(17) Richard, T. A.; Collins, C. J. J. Am. Chem. Soc. 1939, 61, 147-
148.
J. Org. Chem, Vol. 71, No. 12, 2006 4451

Navacchia et al.
450(40), 382 (100), 318 (30). Anal. Calcd for C29H43N5O4Si2: C,
59.86; H, 7.45; N, 12.04. Found: C, 59.90; H, 7.45; N, 12.00.
The minor isomer (5′R)-6e was characterized by HPLC-MS and
1H NMR analysis of a (5′S)-6e and (5′R)-6e mixture.
From 1g in Acetonitrile. 1H NMR and HPLC-MS analysis of
the reaction mixture showed the presence of products (5′S)-6g,
(5′R)-6g, and 2′-deoxyadenosine 7g18 in a 38:32:30 ratio as the only
detectable compounds. Reaction products were not separable by
column chromatography.
(5′S)-O3′-(tert-butyldimethylsilyl)-5′,8-cyclo-2′-deoxyadeno-
sine (5′S)-6g: 1H NMR (400 MHz, DMSO-d6) δ 8.06 (s; 1H), 6.38
(d, J1′,2′′ ) 4.8 Hz; 1H), 5.1 (d, J4′,5′ ) 6.4 Hz; 1H), 4.75 (dd, J2′,3′
) 7.2 Hz, J2′′,3′ ) 4.0 Hz; 1H), 4.50 (d, J4′,5′ ) 6.4 Hz; 1H), 2.46
(dd, J2′,2′′ ) 13.2, J2′,3′ 7.2 Hz; 1H), 2.06 (ddd,, J2′,2′′ ) 13.2, J1′,2′′
) J2′,3′ ) 4.8 Hz 1H), 0.85 (s, 9H), 0.02 (s; 6H); MS (ES+) 364
(M + 1); MS2 (364), 164 (100), 136 (60).
(5′R)-O3′-(tert-butyldimethylsilyl)-5′,8-cyclo-2′-deoxyadeno-
sine (5′R)-6g: 1H NMR (400 MHz, DMSO-d6) δ 8.10 (s; 1H), 6.44
(d, J1′,2′′ ) 5.0 Hz; 1H), 4.65 (s; 1H), 4.43 (s; 1H), 4.37 (m; 1H),
2.36 (dd, J2′,2′′ ) 13.2, J2′,3′ 7.0 Hz; 1H), 2.0 (ddd, J2′,2′′ ) 13.2,
J1′,2′′ ) J2′,3′ ) 5.0 Hz; 1H), 0.85 (s, 9H), 0.06 (s; 6H); MS (ES+)
364 (M + 1); MS2 (364) 164 (100), 136 (10).
From 1a in Acetonitrile. 1H NMR and HPLC-MS analysis of
the reaction mixture showed the presence of products (5′S)-6a5 and
(5′R)-6a5 and 2′-deoxyadenosine 7a in a 47:38:15 ratio as the only
detectable compounds.
(5′R)-N5-Benzoyl-O3′,O5′-bis(tert-butyldimethylsilyl)-5′,8-cy-
1
clo-2′-deoxyadenosine (5′R)-6e: H NMR (400 MHz, CDCl3) δ
8.76 (s; 1H), 8.0 and 7.5-7.7 (m; 5H), 6.60 (1H, d, J ) 5.0 Hz;
1H), 4.80 (s; 1H), 4.60 (s, 1H), 4.28 (dd, J ) 7.0 e 4.0 Hz; 1H),
2.46 (dd, J ) 13.2 and 7.0 Hz; 1H), 2.27 (m; superimposed to
signals of the 5′S-isomer); HPLC-MS (ES+) 582 (M + 1), MS2
(582) 450(100), 382 (30), 318 (80).
The reaction was repeated in refluxing fluorobenzene. 1H NMR
and HPLC-MS analysis of the reaction mixture evidenced the
presence of cyclonucleosides (5′S,8R)-5e, (5′R,8R)-5e, (5′S)-6e, and
(5′R)-6e and the 2′-deoxyadenosine derivative 7e in 39:4:39:4:14
ratio. Column chromatography of the crude mixture led to the
separation of a pure sample of (5′S,8R)-5e.
(5′S,8R)-N-Benzoyl-O3′,O5′-bis(tert-butyldimethylsilyl)-5′,8-
cyclo-7,8-dihydro-2′-deoxyadenosine (5′S,8R)-5e: 1H NMR (400
MHz, CDCl3) δ 8.00 (s; 1H), 8.20 (d; 2H), 7.4-7.6 (m; 3H), 6.17
(d, J ) 6.5 Hz; 1H), 5.20 (dd, J ) 7.8 and 6.0 Hz; 1H), 4.70 (dd,
J ) 7.0 and 2.0; 1H), 4.23 (d, J4′,5′ ) 6.5 Hz; 1H), 3.90 (dd, J )
6.5 and 7.8 Hz; 1H), 2.7 (m; 1H), 2.2 (m; 1H),1.0 (s; 9H), 0.9 (s;
9H), 0.35 (s; 3H), 0.30 (s; 3H), 0.1 (s; 3H), 0.05 (3s; 3H); 13C
NMR (100.6 MHz, CDCl3) δ -4.5, -4.5, -4.3, -4.0, 18.1, 18.2,
25,7, 26.6, 42.1, 69.3, 72,8, 81.9, 86.8, 121.6, 127.9, 129.0, 129.2,
131.1, 133.2, 133.9, 143.9, 160.1, 166.1; MS (ES+) 584 (M + 1);
MS2 (584) 452 (50), 382 (30), 320 (100). Anal. Calcd for
C29H45N5O4Si2: C, 59.66; H, 7.77; N, 11.99. Found: C,59.75; H,
7.80; N, 11.95.
Photolysis of 8-Bromo-2′-deoxyadenosine 1 in Acetonitrile.
General Procedure. A 1.0 mM solution of the appropriate
8-bromo-2′-deoxyadenosine 1 in acetonitrile (10 mL) was irradiated
under argon atmosphere for 30 min, unless otherwise stated, using
a 125 W medium-pressure mercury lamp. The solvent was
eliminated under reduced pressure and the residue analyzed by
1
HPLC-MS and H NMR.
From 1f. HPLC-MS and 1H NMR analysis showed the presence
of products (5′S)-6f, (5′R)-6f, and 7f as the only detectable
compounds in a 52:18:30 ratio.
1
The minor isomer (5′R,8R)-5e showed a H NMR signal at δ
6.10 (1H, d, J ) 6.5 Hz) and a HPLC-MS signal at MS (ES+)
584 (M + 1).
The 39:4:39:4:14 mixture of compounds (5′S,8R)-5e, (5′R,8R)-
5e, (5′S)-6e, (5′R)-6e, and 7e obtained from a repeated reaction in
boiling fluorobenzene was dissolved in o-xylene. Chloroanil (75
mg, 0.30 mmol) was added, the resulting mixture was refluxed for
20 min, and the solvent was eliminated under reduced pressure.
1H NMR and HPLC-MS of the crude showed the exclusive
presence of compounds (5′S)-6e, (5′R)-6e, and 7e in a 78:8:14 ratio.
From 1g. The reaction was monitored at different irradiation
times by HPLC-MS and 1H NMR. In all cases, compounds (5′R)-
6g, (5′S)-6g, 2′-deoxyadenosine 7g, and product B were detected
as the only reaction products. Relative yields are reported in Table
2. The reaction mixture obtained at 3 h irradiation time was
unsuccessfully chromatographed on silica gel column to obtain a
pure sample of product B [1H NMR (400 MHz, DMSO-d6) δ 9.60
(s; disappeared upon D2O shake; 1H), 9.40 (s; disappeared upon
D2O shake; 1H), 8.78 (s; 1H), 7.0 (d, J1′,2′ ) 6.0 Hz; collapsing to
singlet upon irradiation at δ 2.30; 1H), 6.65 (s; 1H), 4.92 (dd, J2′,3′
) 4.5 Hz, J2′′3′ ) 7.0 Hz; collapsing to doublet, J ) 4.5 Hz, upon
irradiation at δ 2.85, collapsing to doublet, J ) 7.0 Hz, upon
irradiation at δ 2.30; 1H), 2.85 (dd, J2′,2′′ ) 14.0, J2′′3′ ) 7.0, 1H),
2.30 (1H, m; 1H), 0.9 (s; 9H), 0.1 (s; 6H)].
1
From 1e in Acetonitrile. H NMR and HPLC-MS analysis of
the reaction mixture showed the presence of cyclonucleosides (5′S)-
6e, (5′R)-6e, and 2′-deoxyadenosine 7e in 65:8:27 ratio as the only
detectable compounds.
1
From 1f in Fluorobenzene. H NMR and HPLC-MS of the
reaction mixture showed the formation of cyclized products (5′S)-
6f and (5′R)-6f and 2′-deoxyadenosine 7f18 in a 55:7:38 ratio as
the only detectable products. Silica gel column chromatography of
the crude mixture led to separation of a fraction mainly containing
cyclized products 6f, which were identified on the basis of HPLC-
From 1a. The reaction was monitored at different irradiation
times by HPLC-MS and 1H NMR analysis. In all cases, compounds
(5′R)-6a, (5′S)-6a, 7a, and product A were detected as the only
reaction products. Relative yields are reported in Table 2. The
reaction mixture obtained at 3 h irradiation time was unsuccessfully
chromatographed on silica gel and RP-C18 column to obtain a pure
sample of product A [1H NMR (400 MHz, DMSO-d6) δ 9.60 (s;
disappeared upon D2O shake; 1H), 9.38 (s; disappeared upon D2O
shake; 1H), 8.70 (s; 1H), 7.50 (br s; disappeared upon D2O shake;
2H), 6.98 (d, J1′,2′ ) 6.0 Hz; collapsing to singlet upon irradiation
at δ 2.40; 1H), 6.57 (s; 1H), 6,00 (br d; disappeared upon D2O
shake; collapsing to singlet upon irradiation at δ 4.72; 1H), 4.72
1
MS and H NMR signals.
(5′S)-O3′,O5′-Bis(tert-butyldimethylsilyl)-5′,8-cyclo-2′-deoxy-
1
adenosine (5′S)-6f: H NMR (400 MHz, CDCl3) δ 8.25 (s; 1H),
6.40 (d, J ) 4.5 Hz; 1H), 5.90 (br s; 2H), 5.20 (d, J ) 6.0 Hz;
1H), 4.83 (dd, J ) 7.0 and 4.5 Hz; 1H), 4.57 (d, J4′,5′ ) 6.0 Hz;
1H), 2.53 (dd, J ) 13.2 and 7.0 Hz; 1H), 2.20 (dt, Jd ) 13.2, Jt )
4.5 Hz; 1H), 0.9 (s; 18 H), 0.1 (s; 12H); HPLC-MS (ES+) 478 (M
+ 1); MS2 (478) 346 (40), 278 (100), 214 (80)].
(5′R)-O3′,O5′-Bis(tert-butyldimethylsilyl)-5′,8-cyclo-2′-deoxy-
1
adenosine (5′R)-6f: H NMR (400 MHz, CDCl3) δ 8.3 (s; 1H),
(br dd, collapsing to dd upon D2O shake; J2′,3′ ) 2.4 Hz, J2′′3′
)
6.50 (d, J ) 4.8 Hz; 1H), 4.73 (s; 1H), 4.57 (s; 1H); MS (ES+)
478 (M + 1); MS2 (578) 346 (20), 278 (100), 214 (70).
From 1f in Acetonitrile. 1H NMR and HPLC-MS analysis of
the reaction mixture showed the presence of cyclonucleosides (5′S)-
6f, (5′R)-6f, and 2′-deoxyadenosine 7f in a 70:8:22 ratio as the only
detectable compounds.
7.0 Hz; collapsing to doublet, J ) 7.0 Hz, upon irradiation at δ
2.40 and to doublet, J ) 2.4 Hz, upon irradiation at δ 2.80; 1H),
2.80 (dd, J2′,2′′ ) 15,0, J2′′3′ ) 7.0; 1H), 2.40 (m; 1H)].
Acknowledgment. This work was supported by the “Min-
istero della Ricerca Scientifica e Tecnologica”, MURST (Rome).
(18) Ogilvie K. K. Can. J. Chem. 1973, 51, 3799-3807.
4452 J. Org. Chem., Vol. 71, No. 12, 2006
JO060197Z