Direct effect of heavy ions and electrons on 2′-deoxyguanosine in the solid state

Article abstract of DOI:10.1039/a801069b

Attempts have been made in this work to gain insights into the mechanisms of the formation of degradation products arising from the exposure of 2′-deoxyguanosine (dGuo) in the solid state to O⁷⁺ heavy ions of 10.6 MeV u^-1 (LET ≈ 500 keV μm^-1). The main decomposition products of dGuo have been isolated by reversed-phase high performance liquid chromatography and characterized by extensive spectroscopic (¹H and ¹³C NMR, mass spectrometry, UV) measurements. Reaction mechanisms, involving the transient formation of sugar and purine radical, are proposed to explain the generation of the heavy ion-mediated degradation products. Another major objective of the present work is the comparison of heavy ion-induced modifications of 2′-deoxyguanosine with those produced by lower LET radiation. For this purpose, the samples of 2′-deoxyguanosine were exposed in the solid state to electrons of 2 MeV (LET ≈ 0.18 keV μm^-1). It may be inferred from the results of the qualitative and semi-quantitative comparison that the modifications of the sugar moiety are more efficiently induced by heavy ions than by electrons.

Full text of DOI:10.1039/a801069b

View Article Online / Journal Homepage / Table of Contents for this issue
Direct effect of heavy ions and electrons on 2Ј-deoxyguanosine in the
solid state
Marina Gromova, Robert Nardin and Jean Cadet*
Département de Recherche Fondamentale sur la Matière Condensée, SCIB/Laboratoire
‘Lésions des Acides Nucléiques’, CEA/Grenoble 17, Avenue des Martyrs, F-38054 Grenoble
Cedex 9, France
Attempts have been made in this work to gain insights into the mechanisms of the formation of
degradation products arising from the exposure of 2Ј-deoxyguanosine (dGuo) in the solid state to O7؉
heavy ions of 10.6 MeV u؊1 (LET ≈ 500 keV ìm؊1). The main decomposition products of dGuo have
been isolated by reversed-phase high performance liquid chromatography and characterized by
extensive spectroscopic (1H and 13C NMR, mass spectrometry, UV) measurements. Reaction
mechanisms, involving the transient formation of sugar and purine radical, are proposed to explain
the generation of the heavy ion-mediated degradation products. Another major objective of the
present work is the comparison of heavy ion-induced modifications of 2Ј-deoxyguanosine with those
produced by lower LET radiation. For this purpose, the samples of 2Ј-deoxyguanosine were exposed
in the solid state to electrons of 2 MeV (LET ≈ 0.18 keV ìm؊1). It may be inferred from the results of
the qualitative and semi-quantitative comparison that the modifications of the sugar moiety are more
efficiently induced by heavy ions than by electrons.
noted that the contribution of the direct effect to the overall
Introduction
degradation pathways of cellular DNA is enhanced with
Further information on the mechanisms of heavy charged
particle-mediated degradation of cellular targets is much
needed since radiation protection in space flights and radio-
therapy involving heavy ions are topics of current interest.
Heavy ions produce a highly non-homogeneous distribution
pattern along the path of the individual particle. It is assumed
that the radiation-induced reactions may depend on the non-
homogeneity of the spatial distribution of the energy within the
target.1
increasing LET.14–16
A second aspect of the present work concerns the study of
the chemical alterations induced within 2Ј-deoxyguanosine in
the solid state by electrons of 2 MeV (LET ≈ 0.18 keV µmϪ1).
This has allowed us to perform a qualitative and semi-
quantitative comparison of the modifications induced within
2Ј-deoxyguanosine by these two types of radiation.
The use of high linear energy transfer (LET) radiation
delivered by accelerators has allowed investigations of DNA
damage at the cellular and molecular level. A detailed under-
standing of the effects of ionizing radiation on cells at the
molecular level requires the application of a wide variety of
physical, chemical and biological approaches. The damage
caused by heavy ions has been extensively investigated using
either biological or biochemical endpoints, including studies on
cell survival, DNA strand breaks, chromosome aberrations
or mutations derived from damage to DNA.2–7 It should also
be noted that the differences in the chemistry associated with
the increase in LET have been assessed using computer
simulation.8–11 On the other hand, there is a paucity of in-
formation on the chemical effects of heavy ions on DNA and
related compounds. A few structural data are available from
EPR studies on heavy ion-induced radicals within DNA and
its components.12,13 In contrast, there is no information on the
final decomposition products arising from the exposure of
DNA components to heavy charged particles, at least in terms
of direct effect.
Our research program is aimed at understanding the effects
of heavy ions at the molecular level through the charac-
terization of the final diamagnetic decomposition products
of DNA model systems. In the present work, attempts were
made to characterize the main degradation products of 2Ј-
deoxyguanosine which arise from the direct effect of O7ϩ
heavy ions of 10.6 MeV uϪ1 (LET ≈ 500 keV µmϪ1). For this
purpose, the samples were irradiated in the solid state in order
to concentrate on the direct effect of heavy ions. It should be
Results and discussion
Characterization of the heavy ion-mediated decomposition
products of 2Ј-deoxyguanosine
The separation of the radiation-induced degradation products
of 2Ј-deoxyguanosine (dGuo) was achieved on various semi-
preparative and analytical reversed-phase silica gel columns
using combinations of water and methanol as the mobile
phases. The HPLC profile of the dGuo degradation products
upon exposure to a dose of 8 MGy is illustrated in Fig. 1. The
intact 2Ј-deoxyguanosine was firstly removed (see Experimental
section). The main modified products of dGuo isolated by
HPLC were identified by detailed UV, 1H and 13C NMR
measurements together with mass spectroscopic analysis. The
compounds were characterized as the following (see also
Fig. 2): 1, 2-deoxy--ribono-1,4-lactone; 2, (5ЈR)-5Ј,8-cyclo-
2Ј-deoxyguanosine; 3, guanine; 4, 9-(2-deoxy-α--threo-pento-
1,5-dialdo-1,4-furanosyl)guanine; 5, 9-(2-deoxy-β--erythro-
pento-1,5-dialdo-1,4-furanosyl)guanine; 6, 9-(2-deoxy-α--
threo-pentofuranosyl)guanine; 7, 2-amino-1,9-dihydro-9-(tetra-
hydro-4-hydroxyfuran-2-yl)-(2R-trans)-6H-purin-6-one; 8, 9-(2-
deoxy-α--erythro-pentofuranosyl)guanine; 9, 9-(2-deoxy-
β--threo-pentofuranosyl)guanine; 10, 5Ј,8-cyclo-2Ј,5Ј-di-
deoxyguanosine; 11, 8-oxo-7,8-dihydro-2Ј-deoxyguanosine;
12, 2Ј,5Ј-dideoxyguanosine; 13, 5Ј-O-(2-deoxy-α--erythro-
pentofuranosyl)-2Ј-deoxyguanosine; 14, 9-ethenylguanine;
15, 9-(2,3-dideoxy-3,4-didehydro-β--erythro-pentofuranosyl)-
guanine; 16, 9-(5-hydroxymethylfuran-2-yl)guanine.
The radiation-induced degradation products of dGuo may
J. Chem. Soc., Perkin Trans. 2, 1998
1365

View Article Online
be grouped into several classes according to the nature of
the chemical modifications: fragmentation products (1, 3),
anhydronucleosides (2, 10), a nucleoside with a modified purine
ring (11), nucleosides modified within the sugar moiety (4–9,
12, 14–16), and the dimeric product 13.
Fragments of nucleoside. One of the fastest eluting HPLC
degradation products was assigned to 2-deoxy--ribono-1,4-
1
lactone (1) on the basis of its H NMR and MS features. This
product was already shown to be produced upon exposure of
2Ј-deoxyguanosine to γ-rays in oxygen-free aqueous solution.17
More recently 1 was also isolated after γ-irradiation of 2Ј-
deoxycytidine in frozen aqueous solution.18 It should be noted
that 1 was obtained upon exposure of thymidine in the solid
state to O7ϩ heavy ions (work submitted for publication).
Product 3 was identified as guanine. Since guanine (3)
exhibits a low solubility in water, its measured formation, as
inferred from semi-preparative chromatography (Fig. 1A) is
likely to be underestimated.
Cyclic nucleosides and nucleoside modified within the purine
Fig.
1 HPLC separation of the degradation products of 2Ј-
deoxyguanosine induced by O7ϩ heavy ions (10.6 MeV uϪ1) irradiation
(dose 8 MGy). A, HPLC fraction of the modified products eluted
before dGuo; B, HPLC fraction containing the compounds eluted after
dGuo. The intact dGuo was preliminarily eliminated (see Experimental
section). The contaminants guanosine (*) and 2Ј-deoxyinosine (**) are
present in the batch of dGuo.
ring. Compounds 2 and 10 were both characterized as cyclo-
1
nucleosides. The H NMR spectrum of 2, obtained in D2O,
shows the presence of six non-exchangeable protons. Selective
decoupling experiments allowed us to assign all the signals. Two
peculiar features are the lack of the H8 proton and of one of
3
the two 5Ј-methylene protons. Furthermore, the J coupling
constants involving H1Ј and H2Ј on the one hand and H3Ј and
H4Ј on the other hand exhibit very low values (J < 1 Hz). This
is indicative of a cyclic structure for the modified nucleoside, as
the result of the occurrence of severe steric constraints within
O
O
N
N
N
N
NH
NH
HOH2C
H
O
O
1
the sugar moiety.19,20 The comparison of the latter H NMR
N
NH2
N
NH2
HO
O
O
data with those reported for 5Ј,8-cyclo-2Ј-deoxyadenosine19
allows the proposal of the structure of (5ЈR)-5Ј,8-cyclo-2Ј-
deoxyguanosine for 2. The 5ЈS diastereoisomer is not produced,
at least not in a detectable amount. In addition, the cyclic
structure of 2 is in agreement with the electron-impact
spectroscopy–mass spectroscopy (EIS–MS) analysis which
shows a loss of 2 amu (pseudo-molecular ions [M ϩ H]ϩ at m/z
266) with respect to the molecular weight of dGuo.
The second cyclic product 10 was identified as 5Ј,8-cyclo-
2Ј,5Ј-dideoxyguanosine. This compound had already been
obtained upon γ-irradiation of 2Ј-deoxyguanosine in deaerated
aqueous solution.21
HO
OHC
HO
HO
1
2
4
O
N
O
N
N
N
N
NH
NH
OHC
N
NH2
NH2
O
O
HO
HO
5
7
The HPLC peak of 11 was found to contain the well known
8-oxo-7,8-dihydro-2Ј-deoxyguanosine. The attribution of 11
was inferred from the comparison of the 1H NMR features with
earliest reported data.22 This received further confirmation
from the comparison of the chromatographic features of 11
with those of the authentic sample. This was achieved by HPLC
analysis associated with an electrochemical detection.23,24
Nucleosides modified within the sugar moiety. Ten products
(4–9, 12, 14, 15 and 16) which exhibit a modified sugar frag-
ment were isolated and characterized. NMR and MS structural
analyses of these compounds indicate that the guanine residue
is intact. The ion at m/z 152 corresponding to the [Gua ϩ H]ϩ
fragment is observed in the EIS–MS spectra of all the products.
O
O
N
N
N
N
NH
NH
H
N
H3C
H
N
NH2
NH2
O
O
HO
HO
10
12
O
N
O
N
N
N
N
NH
NH
HOH2C
1
It should be added that the H and 13C NMR spectra of the
N
O
B
NH2
H
NH2
CH2
O
A
O
above compounds exhibit the characteristic signals of the H8
proton and the C8 carbon of the guanine moiety.
H
HO
HO
H
The nucleosides may be ranged into three main classes
according to the nature of the modifications. Compounds 4
and 5 were characterized as two C4Ј epimers of the 9-(1,5-
dialdo-1,4-furanosyl)guanine. Nucleosides 6, 8 and 9 result
from isomerization reactions of the sugar fragment of 2Ј-
deoxyguanosine whereas products 15 and 16 show unsaturated
sugar rings.
13
14
O
O
N
N
NH
NH
HOH2C
N
N
N
NH2
N
NH2
O
O
1
Products 4 and 5.—The H NMR spectra of 4 and 5, re-
corded in D2O, exhibit similar features. Both nucleosides have
six non-exchangeable sugar protons. Interestingly, a doublet
(relative intensity = 1) is observed in place of the characteristic
AB pattern (2 × 4 lines) of the ABX system corresponding to
15
16
Fig. 2 Structures of the O7ϩ-mediated decomposition products of
2Ј-deoxyguanosine
1366
J. Chem. Soc., Perkin Trans. 2, 1998

View Article Online
Table 1 400.13 MHz 1H chemical shifts and coupling constants for dGuo degradation products modified within the sugar moiety in D2O
δ
Compound
H1Ј
H2Ј
H2Љ
H3Ј
H4Ј
H5Ј
H5Љ
H8
Other
4
5
6.45
6.41
6.45
6.40
6.36
6.26
6.32
7.12
6.68
—
2.70
2.88
2.69
2.86
2.95
2.51
2.88
5.80
3.36
6.65
2.95
2.55
2.96
2.63
2.59
2.98
2.57
5.28
3.06
—
4.79
4.76
4.82
4.87
4.58
4.61
4.46
—
4.37
4.09
4.57
4.37
4.43
4.22
4.22
—
5.25
5.22
3.99
—
3.82
4.04
—
—
—
4.68
—
—
3.91
—
3.74
3.95
—
—
—
4.68
8.07
8.05
8.05
8.01
8.15
8.10
8.08
8.11
8.00
8.04
—
—
—
6
7
4.06 (H4Љ)
—
—
8
9
12
14
15
16
1.42 (4Ј-Me)
—
—
—
5.44
6.64
6.58
—
J/Hz
1Ј2Ј
Compound
1Ј2Љ
2Ј2Љ
2Ј3Ј
2Љ3Ј
3Ј4Ј
4Ј5Ј
4Ј5Љ
5Ј5Љ
Other
4
5
6
7
6.8
8.3
7.0
6.9
7.6
6.1
6.9
6.9
Ϫ14.4
Ϫ14.2
Ϫ14.7
Ϫ14.8
1.1
6.0
1.7
5.4
5.1
2.4
5.2
1.5
3.1
2.1
3.3
3.5
7.3
4.2
4.4
—
—
—
6.8
—
—
—
Ϫ12.1
—
—
—
—
2Љ4Љ = 1.1; 3Ј4Љ = 1.0;
4Ј4Љ = Ϫ10.0
—
8
9
12
14
15
16
7.8
3.0
6.4
15.9
9.4
—
3.2
8.9
6.7
9.0
3.5
—
Ϫ14.9
Ϫ15.2
Ϫ14.1
1.7
Ϫ17.3
—
6.9
1.0
6.2
—
2.5
3.5
3.0
5.8
4.2
—
2.4
—
3.2
3.5
3.9
—
2.7
—
3.8
4.4
—
—
—
—
5.0
7.2
—
—
—
—
Ϫ12.3
Ϫ11.8
—
—
—
—
4Ј,4Ј-Me = 6.6
—
2Ј4Ј = 2.4; 2Љ4Ј = 2.3
3Ј5Ј < 1
—
the H5Ј and H5Љ protons, at δ 5.25 and δ 5.22 for 4 and 5
respectively. The complete assignment of the proton signals was
achieved by selective decoupling experiments (Table 1). The
1
comparison of the H NMR features with those of authentic
samples17 allows the unambiguous characterization of 5 as 9-
(2-deoxy-β-erythro-pento-1,5-dialdo-1,4-furanosyl)guanine. A
downfield shift of the H4Ј signal (∆δ ϩ0.28) is noted for 4 with
respect to that of 5. In addition, a low scalar coupling between
H2Ј and H3Ј (3J2Ј3Ј = 1.1 Hz) on the one hand, and a relatively
high coupling constant between H1Ј and H2Љ protons (3J1Ј2Љ
=
7.6 Hz) on the other hand are observed. These values are char-
acteristic of the α--threo configuration of 2Ј-deoxypentofuran-
osyl nucleosides.17 Therefore, a likely structure for 4 is a nucleo-
side with an aldehydic function at the 5Ј-position exhibiting
an α--threo configuration, namely 9-(2-deoxy-α--threo-pento-
1,5-dialdo-1,4-furanosyl)guanine. Further support for the
assignment of 4 and 5 as C4Ј epimers of 9-(2-deoxy-1,5-dialdo-
1,4-furanosyl)guanine was gained from the characterization
of the products of the quantitative reaction with sodium
borohydride (NaBH4).25 The reaction consists of the reduction
of the 5Ј-aldehydic function to the 4Ј-hydroxymethyl group.
The reduction products of 4 and 5 were characterized as
9-(2-deoxy-α--threo-pentofuranosyl)guanine and 2Ј-deoxy-
guanosine respectively.
Fig. 3 400.13 MHz 1H NMR spectrum of 2-amino-1,9-dihydro-
9-(tetrahydro-4-hydroxyfuran-2-yl)-(2R-trans)-6H-purin-6-one (7) re-
corded in D2O
H2Ј and H3Ј protons). These observations lead us to propose
for 9 the structure of 9-(2-deoxy-β--threo-pentofuranosyl)-
guanine with a preferential C2Ј exo conformation.
Products 7, 12 and 14.—Fig. 3 shows the 1H NMR spectra of
7 recorded in D2O. Besides the H8 proton, the presence of six
non-exchangeable sugar protons is noted. The signals at δ 6.40,
4.87, 2.86 and 2.63 were assigned to the H1Ј, H3Ј, H2Ј and H2Љ
protons respectively. The values of the chemical shifts of H1Ј,
H2Ј and H2Љ protons are similar to those of the related protons
of 2Ј-deoxyguanosine. A striking feature is the pattern of the
H2Љ proton signal which appears as a doublet of doublet of
pseudotriplet. Selective decoupling experiments show that H2Љ
is not only coupled with H1Ј, H2Ј and H3Ј protons but also
with the proton resonating at δ 4.06. Moreover, the protons
at δ 4.37 and 4.06 are coupled to each other (J = Ϫ10.0 Hz)
and both also to the H3Ј proton. From these observations,
the structure of 7, namely 2-amino-1,9-dihydro-9-(tetra-
hydro-4-hydroxyfuran-2-yl)-(2R-trans)-6H-purin-6-one, may
be rationalized as having a methylene group at position 4Ј.
The observation of a long range scalar coupling between H2Љ
and the H4Љ protons may be explained by the occurrence of an
‘M’ configuration in the sugar moiety with a quasi-planar
geometry.
Products 6, 8 and 9.—The 1H NMR spectra of 6, 8 and
9, recorded in D2O, show a complete set of non-exchangeable
protons of a 2-deoxypentofuranosyl moiety (Table 1). Com-
parison of the 1H NMR features including chemical shifts and
coupling constants with published NMR data of authentic
samples17,26 allows the assignment of products 6 and 8 as 9-(2-
deoxy-α--threo-pentofuranosyl)guanine (6) and 9-(2-deoxy-α-
-erythro-pentofuranosyl)guanine (8) respectively.
A slight upfield shift of the anomeric proton signal
1
(∆δ Ϫ0.12) is observed in the H NMR spectrum of 9 with
respect to the original product. The low magnitude of the scalar
coupling between H2Ј and H3Ј (3J2Ј3Ј = 1.0 Hz) is characteristic
of 2Ј-deoxy-threo-pentofuranosides. On the other hand, the
relatively low magnitude of the coupling between H1Ј and
H2Ј protons (J1Ј2Ј = 3.0 Hz) is indicative of a preferential trans
diequatorial position (β form) for the related protons (as for
J. Chem. Soc., Perkin Trans. 2, 1998
1367

View Article Online
Fig. 4 400.13 MHz 1H NMR spectrum of 2Ј,5Ј-dideoxyguanosine (12)
recorded in D2O
Further support for the proposed structure was provided by
the inspection of the 13C NMR data. The sugar moiety exhibits
only four carbons, in agreement with the loss of the 5-
hydroxymethyl group. Three of the observed signals were
assigned on the basis of a DEPT experiment and by com-
parison of the chemical shifts with those of 2Ј-deoxyguanosine.
Thus, two methinic carbons which resonate at δ 84.1 and 71.1
correspond to C1Ј and C3Ј respectively. In addition, the
Fig. 5 400.13 MHz 1H NMR spectrum of 9-ethenylguanine (14)
recorded in D2O
methylene carbon at 39.3 ppm was assigned to C2Ј. Two-
1
13
1
dimensional heteronuclear ( J C– H) multiquanta correlation
(HMQC) analysis indicates that both protons, supposed to be
H4Ј and H4Љ (δ 4.37 and 4.06), are correlated with the carbon
resonating at δ 75.5. This allows the assignment of the latter
carbon to C4Ј.
In addition, the EIS–MS spectrum of 7 indicates a loss of
30 amu (pseudo-molecular ion [M ϩ H]ϩ at m/z 238). This may
be rationalized in terms of the loss of the hydroxymethyl group
at C4Ј with respect to 2Ј-deoxyguanosine.
Fig.
6
400.13 MHz 1H NMR spectrum of 9-(2,3-dideoxy-3,4-
The signals of the 1H NMR spectrum of 12 (Fig. 4), recorded
in D2O, were assigned by selective decoupling experiments. The
presence of five multiplets, which in each case corresponds to
one proton, was observed for the sugar moiety. In addition,
a doublet with a relative intensity of 3 is noted at δ 1.42. The
latter protons, which are likely to belong to a methyl group,
didehydro-β--erythro-pentofuranosyl)guanine (15) recorded in D2O
This receives further confirmation from the 13C NMR and
EIS–MS analyses of 14. The presence of only two carbons in
the modified sugar moiety, (CH) at δ 125.9 and (CH2) at δ 105.6,
is inferred from the 13C DEPT NMR analysis. The chemical
shifts of the related carbons in the low-field region provide
further support for an ethylenic structure. Furthermore, the
EIS–MS spectrum shows a pseudo-molecular ion [M ϩ H]ϩ at
m/z 178, and an ion [Gua ϩ H]ϩ at m/z 152. This is indicative
of a loss of 90 amu with respect to the molecular weight of
2Ј-deoxyguanosine.
were found to be coupled with the H4Ј (J4Ј,CH = 6.6 Hz). In-
3
terestingly, the characteristic AB pattern of the usual ABX
system of the H5Ј and H5Љ signals of DNA and RNA nucleo-
sides is lacking. Altogether, this is strongly suggestive of a 2Ј,5Ј-
dideoxyguanosine structure for 12.
The 13C NMR analysis provides further support for this
hypothesis: DEPT NMR data show the lack of the C5Ј methyl-
enic signal. Interestingly, the presence of an additional signal,
corresponding to a carbon having an odd number of protons, is
noted in the high field region of the spectrum at δ 17.96. The
low value of the chemical shift of the signal suggests that this
is a methylic carbon. Another significant piece of structural
information was inferred from the EIS–MS spectrum. The
observation of a pseudo-molecular ion [M ϩ H]ϩ at m/z 252
is indicative of a loss of 16 amu with respect to dGuo. In
addition, the presence of an ion [Gua ϩ H]ϩ at m/z 152 shows
that the modification has occurred within the sugar moiety.
Nucleosides containing unsaturated sugar moieties (15, 16).—
The H NMR spectrum of 15, recorded in D2O, shows the
1
presence of five non-exchangeable sugar protons (Fig. 6). Three
of them resonate in the low-field region of the spectrum. The
H1Ј proton (doublet of doublets) is coupled to the H2Ј and H2Љ
protons with coupling constants of 9.4 and 3.5 Hz respectively.
Both the H2Ј and H2Љ proton signals (at δ 3.36 and 3.06) con-
sist of 16 lines. The additional splitting of these multiplets is
due to the occurrence of a long range coupling with the H4Ј
proton resonating at δ 6.58. The chemical shift values of
H3Ј and H4Ј protons (δ 5.44 and 6.58) are strongly indicative
of the presence of an ethylenic bond in the sugar ring. The
1
The H NMR spectrum of 14 recorded in D2O is shown in
Fig. 5. In addition to the H8 signal (δ 8.11), three non-
exchangeable proton resonances are observed in the low-field
region (δ 7.12, 5.80 and 5.28) of the spectrum. These chemical
shift values are suggestive of the presence of an ethylenic bond
within the sugar moiety. Interestingly, each of the signals
appears as a doublet of doublets. It should be added that
each of the three protons exhibits scalar coupling with the
two others. Interestingly a high value is observed for the ‘trans’
coupling constant (3J1Ј2Јa = 15.9 Hz) and the ‘cis’ coupling
constant (3J1Ј2Јb = 9.0 Hz) of the ethylenic protons. In contrast,
a low value (2J2Јa–2Јb = 1.7 Hz) is noted for the geminal coupling
constant. These data are strongly suggestive of a 9-ethenyl-
guanine structure for 14.
1
comparison of these results with available published H NMR
data for 1-(2,3-dideoxy-3,4-didehydro-β--erythro-furanosyl)-
thymine28 suggests a structure of 9-(2,3-dideoxy-3,4-didehydro-
β--erythro-pentofuranosyl)guanine for 15.
Additional structural information for 15 was inferred from
1D-13C NMR and 2D-heteronuclear ( J H–
)
correlated
1
1
13
C
(HMQC) NMR analyses. The high values for the chemical
shifts of the C3Ј (δ 100.7) and C4Ј (δ 143.4) signals are
characteristic of ethylenic carbons. Furthermore, the pseudo-
molecular ion [M ϩ H]ϩ at m/z 220 observed in the EIS–MS
spectrum of 15 is indicative of the loss of 48 amu with respect
to the molecular weight of 2Ј-deoxyguanosine.
1368
J. Chem. Soc., Perkin Trans. 2, 1998

View Article Online
Inspection of the 1H NMR data of 16 obtained in D2O (Table
1) shows the presence of four non-exchangeable protons within
the modified sugar moiety. These include two doublets of
relative intensity 1 at δ 6.65 and 6.64 respectively, together
with a broad singlet of relative intensity 2 at δ 4.68 (the half-
height linewidth is about 2 Hz). The two former protons (δ 6.65
and 6.64) are coupled to each other with a coupling constant of
3.5 Hz. In addition, the proton at δ 6.64 is further coupled with
the two protons resonating at δ 4.68. The latter coupling con-
stant which is of very low magnitude (<0.5 Hz) induces a broad-
ening of the related signals. This was inferred from selective
decoupling experiments. The comparison of the latter 1H NMR
features with those of 9-(5-hydroxymethylfuran-2-yl)adenine28
allows us to propose the structure of 9-(5-hydroxymethylfuran-
2-yl)guanine for 16. The signals at δ 6.65 and 6.64 can be
assigned to H2Ј and H3Ј protons respectively. The signal at
δ 4.68 is attributed to the H5Ј and H5Љ protons.
Further support for the above structure is provided by
considering 13C NMR DEPT and heteronuclear multibond
correlation (HMBC) data. The high chemical shift values for
C1Ј (δ 140.2), C2Ј (δ 105.4), C3Ј (δ 110.4) and C4Ј (δ 152.5)
carbons are indicative of the presence of two ethylenic bonds
within the sugar ring. Furthermore, the EIS–MS spectrum of
16 exhibits a pseudo-molecular ion [M ϩ H]ϩ at m/z 248. This
may be rationalized in terms of a loss of 20 amu with respect to
the molecular weight of 2Ј-deoxyguanosine. In addition, the
presence of an ion at m/z 152 [Gua ϩ H]ϩ indicates that the
base moiety is not modified.
Fig. 7 400.13 MHz 1H NMR spectrum of 5Ј-O-(2-deoxy-α--erythro-pentofuranosyl)-2Ј-deoxyguanosine (13) recorded in D2O
J. Chem. Soc., Perkin Trans. 2, 1998
1369

View Article Online
Table 2 400.13 MHz 1H chemical shifts and coupling constants of
5Ј-O-(2-deoxy-α--erythro-pentofuranosyl)-2Ј-deoxyguanosine (13) in
D2O
Table 3 400.13 MHz 13C chemical shifts of sugar carbons of 5Ј-O-(2-
deoxy-α--erythro-pentofuranosyl)-2Ј-deoxyguanosine (13) in D2O
δ
δ
C1Ј
C2Ј
C3Ј
C4Ј
C5Ј
H1Ј
H2Ј
H2Љ
H3Ј
H4Ј
H5Ј
H5Љ
H8
13A
13B
83.6
104.3
38.3
40.4
71.2
70.7
85.5
85.5
66.7
61.3
13A
6.38
5.26
6.38
2.96
2.33
2.86
2.62
1.89
2.57
4.76
4.30
4.69
4.29
4.03
4.20
3.93
3.76
3.88
3.78
3.68
3.83
8.12
—
8.05
13B
dGuo
other hand, the chemical shifts of the other sugar carbons are
less affected with respect to those of 2Ј-deoxyguanosine. This
suggests, in particular, a change in the electronegativity of the
C1Ј for the unit B.
J/Hz
1Ј2Ј 1Ј2Љ
2Ј2Љ
2Ј3Ј 2Љ3Ј 3Ј4Ј 4Ј5Ј 4Ј5Љ
5Ј5Ј
13A
13B
6.5
5.4
dGuo 7.4
6.7 Ϫ14.1
1.0 Ϫ14.7
6.4 Ϫ14.1
6.3 4.5 4.3 4.9 3.2 Ϫ11.3
7.8 2.4 3.7 3.5 4.9 Ϫ12.3
6.1 3.5 3.2 3.7 4.7 Ϫ12.6
Consideration of the 1H and 13C NMR data allows the
proposal of a structure which involves the formation of an
intermolecular bond between C5Ј of A and C1Ј of B. The lack
of scalar coupling between H5Ј, H5Љ (A) protons on the one
hand and the H1Ј (B) proton is indicative of an intermolecular
bond involving an O-glycosidic bridge (α 1ЈB→5ЈA). In contrast,
in agreement with EIS–MS data (vide infra) the occurrence of
a covalent bond between C5ЈA and C1ЈB may be ruled out.
The proposed structure for 13 is 5Ј-O-(2-deoxy-α--erythro-
pentofuranosyl)-2Ј-deoxyguanosine.
Dimeric product 13.—The 1H NMR spectrum of 13 recorded
in D2O is given in Fig. 7. The measurement of the integrated
signals indicates the presence of 15 non-exchangeable protons.
The singlet at δ 8.12 is attributed to the H8 guanine proton.
In addition, the UV absorption of 13 is similar to that of
2Ј-deoxyguanosine. Altogether, this overall information is
indicative of the presence of an unmodified guanine residue
in 13.
To verify this hypothesis, we have carried out two-
dimensional NMR analyses including HMBC and ROESY
Detailed 1H NMR measurements including homonuclear
1H–1H NMR (COSY) and selective decoupling experiments
indicate that two complete sets of non-exchangeable sugar
protons of 2-deoxyribose type are present in 13. One of the
anomeric protons has a chemical shift value similar to that
of 2Ј-deoxyguanosine (δ 6.38). The other anomeric signal is
significantly upfield shifted (∆δ ca. Ϫ1.12). Let us call A the
sugar fragment with the H1Ј proton which resonates at δ 6.38
and B the second sugar residue. The chemical shifts and
coupling constants of 13 for both sugar moieties are listed in
Table 2. We note that all the proton resonances of the sugar
moiety B are shifted to the upfield region (∆δ Ϫ1.12 for H1Ј;
∆δ Ϫ0.68 for H2Љ; ∆δ Ϫ0.39 for H3Ј). The effect is particularly
noticeable for the H1Ј proton. This is suggestive of the occur-
rence of a structural modification in the vicinity of the C1Ј
carbon. In addition, the measured coupling constants for B are
characteristic of an ‘α’ anomeric configuration.26 This was
inferred from the relatively low values of the couplings between
the ‘trans’ protons J1Ј2Љ = 1.0 Hz; J2Љ3Ј = 2.4 Hz; J3Ј4Ј = 3.7 Hz.
These results suggest a structure that includes one unmodified
base and two sugar fragments.
experiments. The heteronuclear long range correlated (HMBC)
n
1
13
analysis allows us to visualize a ( J H– C) coupling between
the proton H1Ј(B) and the carbon C5Ј(A) on the one hand, and
between the protons H5Ј, H5Љ(A) and the carbon C5Ј(B) on the
other hand. In addition, a correlation between the H1Ј proton
(A) and the C8 and C4 carbons is observed. This result confirms
that the guanine moiety is linked to the sugar residue of A.
Additional support for the proposed structure is provided
from the ROESY analysis, which shows the occurrence of a
dipolar interaction between H1Ј(B) and H5Ј(A). The observ-
ation of an Overhauser effect between H1Ј(B) and H5Ј(A)
protons suggests that they are close to one another in space.
The latter observation is in agreement with the formation of
a bond between the C5Ј(A) and C1Ј(B) carbons of the two
sugar residues through
a heteroatom. Furthermore, the
ROESY NMR analysis provides relevant information on the
configuration of the sugar parts A and B. The dipolar inter-
action between the H1Ј(A) and H4Ј(A) protons is strongly
indicative of a ‘β’ configuration for the sugar moiety A. In this
configuration, the H1Ј and H4Ј protons are located on the same
side with respect to the sugar plane. In addition, a correlation
between the H1Ј(B) and H3Ј(B) protons of the B part is
observed. This provided further support to the above structure
which involves an ‘α’ configuration for the sugar fragment B.
Finally, the EIS–MS of 13, obtained in the positive mode,
exhibits a pseudo-molecular peak [M ϩ H]ϩ at m/z 384. This is
indicative of a gain of 116 amu with respect to the molecular
weight of 2Ј-deoxyguanosine. This can be explained by the
addition of a 2-deoxyribose unit to a 2Ј-deoxyguanosine
molecule.
The 1H NMR spectrum of 13 recorded in [2H6]DMSO
provides additional information on the structure of 13. The
presence of three new signals is observed within the range
5.5 < δ < 4.5, with respect to the spectrum obtained in D2O.
Two doublets at δ 5.31 and 4.98 were assigned to the protons
of hydroxy groups 3Ј-OH of A and B moieties respectively.
However, there is only one triplet signal (δ 4.68) corresponding
to the hydroxy proton of a 4Ј-hydroxymethyl group. The signal
was attributed to the 5Ј-OH proton of (B). The H5Ј and H5Љ
protons of the residue A do not show any additional coupling
with respect to the spectrum recorded in D2O. We can therefore
assume that there is no hydroxy group at position 5Ј of the A
residue.
Proposed reaction mechanisms for some of the heavy ion-induced
degradation products of dGuo
For most of the heavy ion-mediated degradation products of
dGuo, we can propose mechanisms of formation that involve
radical intermediates. The neutral radical centered on the
C1Ј carbon is probably the precursor of 2-deoxy--ribono-
1,4-lactone (1). For the formation of (5ЈR)-5Ј,8-cyclo-2Ј-
deoxyguanosine (2), it is reasonable to propose a mechanism
involving initial hydrogen abstraction from the 5Ј-CH2OH
group. The resulting C5Ј centered radical is able to undergo
an intermolecular addition to the C8 carbon. The mechanism
which is likely to explain the formation of 5Ј,8-cyclo-2Ј,5Ј-
dideoxyguanosine (10) was previously described by Berger and
Additional information on the intermolecular bond was
inferred from 13C NMR experiments. The 13C DEPT spectrum
shows the presence of 10 sugar carbon signals. However, two of
the latter signals exhibit chemical shifts significantly different
from those of 2Ј-deoxyguanosine: (CH) at δ 104.3, and (CH2)
at δ 66.7. Data inferred from the two-dimensional hetero-
1
1
13
nuclear ( J H– C) correlated (HMQC) NMR analysis allow
the assignment of the latter signals to the C1Ј(B) and C5Ј(A)
carbons respectively (Table 3). A notable upfield shift of C5Ј(A)
(∆δ Ϫ5) and C1Ј(B) (∆δ Ϫ20.7) carbon signals is seen. On the
1370
J. Chem. Soc., Perkin Trans. 2, 1998

View Article Online
Cadet.17 It involves initial hydrogen abstraction at carbon C4Ј,
followed by dehydration and subsequent intramolecular cycliz-
ation. The formation of the ubiquitous 8-oxo-7,8-dihydro-2Ј-
deoxyguanosine (11) may be accounted for by the hydration of
the purine radical cation,22 followed by the oxidation of the
resulting neutral radical.
Another important class of dGuo decomposition products
consists of nucleosides bearing modifications within the sugar
moiety. The general mechanism for the radiation-induced
degradation of the sugar moiety of dGuo is likely to involve
initial ionization followed by deprotonation of the transient
sugar radical cation.13 The C5Ј centered radical is considered to
be the precursor of 9-(2-deoxy-β--erythro-pento-1,5-dialdo-
1,4-furanosyl)guanine (5).17 A second nucleoside exhibiting
an aldehydic function at the C5Ј position, namely 9-(2-deoxy-α-
-threo-pento-1,5-dialdo-1,4-furanosyl)guanine (4), was also
isolated. The mechanism of epimerization at the C4Ј position
of 4 is still unclear.
It is likely that the sugar radicals generated by hydrogen
abstraction at carbons C1Ј, C3Ј and C4Ј are the precursors
of furanosidic isomers of 2Ј-deoxyguanosine (6, 8 and 9).
An alternative mechanism for the formation of these epimers of
2Ј-deoxyguanosine would involve hydrogen abstraction at
carbons C4Ј as the initial step. This could induce a labilization
of the C᎐N glycosidic bond with subsequent opening of the
sugar ring. Rearrangement of the sugar moiety of the nucleo-
side through an acyclic intermediate would generate the three
isomers 6, 8 and 9 after hydrogen transfer to the position 4Ј.17
The formation of 2Ј,5Ј-dideoxyguanosine (12) may be
rationalized in terms of initial generation of a sugar radical
centered at carbon C5Ј. The latter radical should be able to
abstract a hydrogen atom from neighboring molecules. For the
other nucleosides modified within the sugar moiety (7, 14, 15
and 16) and the compound 13, there is not enough available
information to propose mechanisms for their formation.
Fig. 8 Dose-rate decomposition of original product (A) and base
release (B) induced by exposure of dGuo in the solid state to O7ϩ
heavy ions and electrons
The formation of 8-oxo-7,8-dihydro-2Ј-deoxyguanosine (11)
is higher in the samples irradiated with electrons than in those
exposed to heavy ions. It has to be noted that 11 is the only
detected product exhibiting this dose-rate dependence (electron
effects > heavy ion effects). In addition, for both irradiation
sources, the formation of 11 reaches a plateau with the increase
in doses. This may be explained by the experimental conditions
associated with the irradiation. The exposure of dGuo to heavy
ions was carried out under vacuum whereas irradiation with
electrons was achieved under a helium atmosphere. In both
cases, traces of oxygen and/or water molecules may be present.
In particular, small amounts of water may facilitate the
formation of 8-oxodGuo 11. The observed plateau in the
formation of 11 with increasing doses of radiation may be
explained by the gradual consumption of water in the
irradiated samples. Another possibility to be considered is the
very high susceptibility of 11 to oxidizing agents. It should
be noted that the formation of 11 represents about 10% of
the overall degradation pathway for an applied dose of 2 MGy
of electrons. As a general comment, it should be mentioned
that the yield of the final degradation products is very low. This
may be rationalized in terms of efficient radical recombination
reactions leading to the restitution of dGuo. It may be added
that 8-oxo-7,8-dihydro-2Ј-deoxyguanosine (11) is the main
degradation product of dGuo resulting from the radical
modification of the base moiety. However, other minor
decomposition products such as the formamidopyrimidine
derivative resulting from the opening of the imidazole ring17
may be also generated.
The decomposition products 5, 7, 12 and 14 which exhibit
a modified sugar moiety are generated in very low amounts
upon exposure to low LET electrons. The level of their form-
ation is lower than those of contaminants such as guanosine
and 2Ј-deoxyinosine which are present in the batch of dGuo.
Interestingly, the efficiency of electron-mediated formation of
degradation products 5, 7, 12 and 14 is, at least, three-fold lower
than that of 17O ions. From these results, it may be inferred that
the sugar moiety is a preferential target for the heavy ions in
terms of formation of the final products of modification. These
results are consistent with recent observations made by Becker
et al.13 These authors showed by EPR spectroscopy that a
relatively higher formation of sugar neutral radicals is pro-
duced by exposure of DNA to heavy ions with respect to low
LET radiation. Further work is required to better delineate the
mechanisms of formation of the sugar modified decomposition
products.
Comparative effects of heavy ions and electrons on 2Ј-deoxy-
guanosine
Another main objective of the present work was to compare
the effects of heavy ion and low-LET radiation on 2Ј-
deoxyguanosine in terms of degradation products. For this
purpose, 2Ј-deoxyguanosine was exposed in the solid state
to electrons of 2 MeV having a LET = 0.18 keV µmϪ1. The
latter value is about three orders of magnitude lower than the
LET of O7ϩ used in the present study. The HPLC separation
and the characterization of electron-induced degradation
products of dGuo were performed in the same way as for
the study of the effects of heavy ions. The main decomposition
compounds characterized were: 2-deoxy--ribono-1,4-lactone
(1), guanine (3), 9-(2-deoxy-β--erythro-pento-1,5-dialdo-
1,4-furanosyl)guanine (5), 2-amino-1,9-dihydro-9-(tetrahydro-
4-hydroxyfuran-2-yl)-(2R-trans)-6H-purin-6-one (7), 8-oxo-
7,8-dihydro-2Ј-deoxyguanosine (11), 2Ј,5Ј-dideoxyguanosine
(12) and 9-ethenylguanine (14). It should be noted that the
same decomposition products were already found to be
generated upon exposure of 2Ј-deoxyguanosine to heavy ions
(the same numbering was used for designating the electron-
mediated dGuo decomposition products). However, the
products 5, 7, 12 and 14 which arise from initial ionization
reactions within the sugar moiety of dGuo are generated
in very low yields.
A semi-quantitative estimation of the level of dGuo de-
composition and base release induced by O7ϩ heavy ions and
electrons respectively was inferred from the analysis of the
HPLC elution profiles. Interestingly, both yields of dGuo
decomposition and guanine release are approximately linear
with the applied radiation doses (Fig. 8). The yield of dGuo
decomposition is about three times higher upon exposure to
heavy ions than to electrons. A similar effect is observed for the
release of guanine (ratio close to 4).
Experimental
Chemicals
2Ј-Deoxyguanosine was obtained from Pharma Waldhof
(Düsseldorf, Germany) and used without further purification.
HPLC grade methanol was purchased from Carlo Erba
(Farmitalia, Milan, Italy). D2O (99.96%) and [2H6]DMSO
J. Chem. Soc., Perkin Trans. 2, 1998
1371

View Article Online
Table 4 HPLC separation retention time kЈ values (in min) of heavy ion-induced decomposition products of dGuo
Product
kЈA
0–4.1
6.5
7.56
8.31
kЈB
kЈ1
kЈ2
kЈ3
kЈ4
kЈ5
kЈ6
2
4
—
—
—
—
—
—
—
6.87
8.18
8.18
8.54
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
11.33
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
6.53
6.53
—
10.53
14.00
—
—
—
—
—
—
—
—
17.80
27.30
—
—
—
7.26
8.33
—
—
—
—
—
—
—
5
6
9.40
7
10
10.80
12.06
12.73
—
—
—
—
—
—
8
10
10
—
—
—
—
—
—
9
11
12
13
14
15
16
12.93
14.50
18.75
—
—
—
(99.96%) utilized for NMR analysis were provided by Euriso-
top (St Aubin, France).
Spectroscopic measurements
Ultraviolet absorption spectra were recorded in water with
Hewlett-Packard 8452 diode array spectrophotometer
a
(Hewlett-Packard, Palo Alto, CA). The mass spectra were
obtained in the positive mode using a VG Platform instrument
(Fisons-VG, Manchester, UK) in the ‘electrospray’ ionization
mode. The samples to be analyzed were dissolved in a mixture
of water and acetonitrile (50–50 v/v) to which 0.1% of formic
Irradiation procedures
The samples were irradiated as compressed pellets at room
temperature under vacuum for the heavy ions and under helium
for the electrons. Typically, the pellets were prepared by
pressing ca. 90 mg of 2Ј-deoxyguanosine. They had a diameter
of 1.5 cm and a thickness of about 500 µm (≈51 mg cmϪ2). The
samples were exposed to a beam of O7ϩ heavy ions with an
energy of 10.6 MeV uϪ1 at the GANIL accelerator facility
(Caen, France). The estimated range of the O7ϩ ions, 44 mg
cmϪ2, is slightly smaller than the thickness of most of the
pellets. Therefore, it may be estimated that a small fraction,
i.e. about 15% of each sample was not irradiated. Moreover,
the LET evolution of a 10.6 MeV uϪ1 ion coming to rest is
obviously monitored. The mean LET over the irradiated
fraction of the sample is 5.2 MeV mgϪ1 cm2 (≈520 keV µmϪ1).
Typically, 80% of the irradiated range, the front part, received
LET between 250 and 650 keV µmϪ1. The remaining 20% which
includes the Bragg peak deals with LET between 650 and 1100
keV µmϪ1. Therefore, the mean value of LET in the pellet was
estimated to be of the order of 520 keV µmϪ1. The flux was
continuously estimated using calibrated current measurements
and finally integrated to give the fluence. The samples received
fluences of 2.4 × 1012, 4.8 × 1012, 7.2 × 1012, 9.6 × 1012 and
1.2 × 1013 cmϪ2, which correspond to mean doses over the
irradiated part of the samples of 2, 4, 6, 8 and 10 MGy respect-
ively. For the doses estimation, the main LET was assumed to
be 5.2 MeV mgϪ1 cm2. The irradiation of 2Ј-deoxyguanosine
by electrons was performed with a Van de Graaff accelerator
(École Polytechnique, Palaiseau, France). The samples were
1
acid was added. The cone tension was 50 V. The H NMR
spectra were recorded in the Fourier transform mode using
a Bruker AC 200, a Bruker AM 400 and a Bruker AMX
600 apparatus (Bruker Spectrospin, Wissembourg, France).
J values are given in Hz. The 13C NMR analyses were per-
formed in the direct mode on a Bruker AC 200 and a Bruker
AM 400 instrument. 13C NMR measurements in the reverse
mode were carried out on a Varian Unity 400 (Varian, Palo
1
Alto, CA). The H chemical shifts are expressed in ppm with
respect to the HDO line (δ 4.9) for the spectra recorded
in 99.96% D2O at 25 ЊC, whereas tetramethylsilane (TMS)
was used as the external reference in 99.96% [2H6]DMSO.
Calibration of the 13C NMR spectra recorded in 99.96% D2O at
25 ЊC was made with respect to TMS used as the external
reference.
Isolation and characterization of the radiation-induced
decomposition products of dGuo
330 mg of dGuo exposed to O7ϩ heavy ions to an overall dose
of 10 MGy were dissolved in water. Then, the resulting mixture
of the decomposition products was separated by semi-
preparative HPLC using water–methanol (90:10) as the eluent
at a flow rate of 3 ml minϪ1. The fractions which were eluted
before and after intact 2Ј-deoxyguanosine (fractions A and B
respectively) were collected and evaporated to dryness. Both
resulting mixtures were further resolved by semi-preparative
HPLC using water–methanol as the eluent in the relative ratio
95:5 for the fraction A and 83:17 for the fraction B. In both
cases, the flow rate was 3 ml minϪ1 (Fig. 1). For most of the
modified products, further purification was achieved by either
semi-preparative or analytical HPLC. Different chromato-
graphic systems were used: (i) semi-preparative system 1: C18
(10 µm) column, eluent—H2O, flow rate—2 ml minϪ1; (ii) semi-
preparative system 1: C18 (10 µm) column, eluent—H2O–
CH3OH (92.5:7.5), flow rate—3 ml minϪ1; (iii) analytical
system 3: C18 (5 µm) column, eluent—H2O–CH3OH (96:4),
flow rate—1 ml minϪ1; (iv) analytical system 4: C18 (5 µm)
exposed to electron beams of 2 MeV (LET ≈ 0.18 keV µmϪ1
with doses of 2, 4 and 6 MGy respectively.
)
HPLC separation
The semi-preparative HPLC system consisted of a M 6000 dual
pump (Millipore-Waters Associates, Milford, MA) equipped
with a Rheodyne model 7125 injector loop (Berkeley, CA) and
a differential refractometer detector R401 (Millipore-Waters
Associates). The semi-preparative (300 × 7.5 mm id) columns
were home-packed with 10 µm Nucleosil octadecylsilyl silica
gel from Macherey-Nagel (Düren, Germany). The analytical
HPLC system included two 201 pumps (Gilson, Middleton,
WI), a mixer 811 (Gilson, WI), a generator of gradient 720
(Gilson, WI) and an automatic injector SIL-9A Shimadzu
(Touzart & Matignon, Paris, France). The detection of the
compounds was achieved using a UV–VIS variable wavelength
detector L4000 (Merck, Darmstadt, Germany). The Hypersil
5 µm C18 (250–4.6 mm id) analytical column and the Ultromex
3 µm C18 (150–0.46 mm id) analytical column were purchased
from Interchim (Montluçon, France).
column, eluent—H2O–CH3OH (95:5), flow rate—1 ml minϪ1
;
(v) analytical system 5: C18 (5 µm) column, eluent—H2O–
CH3OH (85:15), flow rate—1 ml minϪ1; (vi) analytical system
6: C18 (3 µm) column, eluent—H2O–acetonitrile (97:3), flow
rate—0.9 ml minϪ1. The retention time kЈ values for dGuo
modified products are reported in Table 4. The isolated com-
pounds were then characterized on the basis of different
spectroscopic analyses.
1372
J. Chem. Soc., Perkin Trans. 2, 1998

View Article Online
A similar procedure was adopted for the purification of
360 mg of dGuo exposed to electron beams at a dose of
6 MGy.
(C5Ј); m/z (EISϩ–MS, relative intensity) 152 (23, [Gua ϩ H]ϩ),
166 (64), 248 (100, [M ϩ H]ϩ), 270 (18, [M ϩ Na]ϩ).
(5ЈR)-5Ј,8-Cyclo-2Ј-deoxyguanosine (2). δH(400.13 MHz,
D2O) 6.50 (d, 1, H1Ј), 4.90 (d, 1, H5Ј), 4.77 (t, 1, H4Ј), 4.52
(m, 1, H3Ј), 2.65 (m, 1, H2Ј), 2.24 (m, 1, H2Љ); J1Ј2Ј < 1, J1Ј2Љ 4.9,
J2Ј2Љ Ϫ13.8, J2Ј3Ј 7.5, J2Љ3Ј 4.5, J3Ј4 < 1, J4Ј5Ј 1.2; δC(100.3 MHz,
D2O, DEPT) 88.7 (CH, C4Ј), 84.8 (CH, C1Ј), 69.7 (CH, C3Ј),
64.5 (CH, C5Ј), 43.3 (CH2, C2Ј); m/z (EISϩ–MS, relative
intensity) 152 (100, [Gua ϩ H]ϩ), 266 (47, [M ϩ H]ϩ), 417 (12,
[M ϩ Gua ϩ H]ϩ), 531 (5, [2M ϩ H]ϩ).
9-(2-Deoxy-á-L-threo-pento-1,5-dialdo-1,4-furanosyl)guanine
(4). For δH(400.13 MHz, D2O) and J values, see Table 1. The
structure inferred from 1H NMR analyses was further con-
firmed by chemical reduction with sodium borohydride
(NaBH4).25 For this purpose, compound 4 was solubilized in
ethanol and 1 mg of NaBH4 was subsequently added. The
mixture was left under stirring at room temperature for 1 h,
and then the solution was evaporated to dryness. The resulting
residue was solubilized in methanol and subsequently evapor-
ated to dryness. The latter process was repeated six times in
order to remove NaBH4 as methyl borate. The resulting product
was characterized as 9-(2-deoxy-α--threo-pento-1,5-dialdo-
1,4-furanosyl)guanine (4) on the basis of the comparison of its
chromatographic and spectroscopic features with those of the
authentic sample.17
Conclusions
Sixteen modified products were isolated and characterized
upon exposure of dGuo in the solid state to O7ϩ heavy ions.
Several new modified products were identified. These include:
2-amino-1,9-dihydro-9-(tetrahydro-4-hydroxyfuran-2-yl)-(2R-
trans)-6H-purin-6-one (7), 2Ј,5Ј-dideoxyguanosine (12), 9-
ethenylguanine (14), 9-(2,3-dideoxy-3,4-didehydro-β--erythro-
pentofuranosyl)guanine (15) and 9-(5-hydroxymethylfuran-2-
yl)guanine (16). It has to be noted that all these nucleosides are
modified within the sugar moieties.
The comparison of the effects induced by heavy ions and
electrons (2 MeV) on dGuo in the solid state shows several
noticeable differences. The decomposition of dGuo is about
three times more efficient upon exposure to heavy ions than
to electrons for the same applied dose. A similar effect is ob-
served for the release of guanine (ratio close to 4). Interestingly,
heavy ions induce a significant increase in the amounts of
modifications within the sugar moiety with respect to lower
LET radiation.
Acknowledgements
9-(2-Deoxy-á-L-threo-pentofuranosyl)guanine (6). λmax(H2O)/
nm 252 (shoulder at 272 nm); for δH(400.13 MHz, D2O)
and J values, see Table 1; m/z (EISϩ–MS, relative intensity)
152 (15, [Gua ϩ H]ϩ), 268 (100, [M ϩ H]ϩ), 535 (28, [2M ϩ
Na]ϩ).
2-Amino-1,9-dihydro-9-(tetrahydro-4-hydroxyfuran-2-yl)-(2R-
trans)-6H-purin-6-one (7). λmax(H2O)/nm 252 (shoulder at 272
nm); for δH(400.13 MHz, D2O) and J values, see Table 1; δC(100.3
MHz, sugar carbons) 137.2 (C8), 84.1 (C1Ј), 75.5 (C4Ј), 71.1
(C3Ј), 39.3 (C2Ј); m/z (EISϩ–MS, relative intensity) 152 (20,
[Gua ϩ H]ϩ), 238 (100, [M ϩ H]ϩ), 260 (5, [M ϩ Na]ϩ), 475
(15, [2M ϩ H]ϩ).
9-(2-Deoxy-â-D-threo-pentofuranosyl)guanine (9). λmax(H2O)/
nm 252 (shoulder at 272 nm); for δH(400.13 MHz, D2O)
and J values, see Table 1; m/z (EISϩ–MS, relative intensity)
152 (7, [Gua ϩ H]ϩ), 268 (100, [M ϩ H]ϩ), 535 (18, [2M ϩ
H]ϩ).
We thank Emmanuel Balanzat, Benoît Gervais and Jacques
Dural (Centre Interdisciplinaire de Recherche sur les lons
Lourds, Caen, France) for their contribution to the irradiation
of the samples. We also acknowledge Daniel Ruffieux (Labora-
toire de Biochimie C, Hôpital Michalon, Grenoble, France)
for mass spectrometry measurements, and Marie-Françoise
Foray and Françoise Sarrazin (CEA-Grenoble, France) for
their assistance in acquiring the 600 MHz (Bruker) and 400
MHz (Varian) NMR spectra. The contribution of Maurice
Berger (CEA-Grenoble, France) to HPLC analysis is also
acknowledged. The CNES (Centre National d’Études Spatiales)
is acknowledged for financial support of this work. M. G.
would like to thank the Commissariat à l’Énergie Atomique
for a fellowship. This work represents partial fulfilment
of the requirements for the degree of Docteur ès Sciences
(M. G.).
2Ј,5Ј-Dideoxyguanosine (12). λmax(H2O)/nm 252 (shoulder at
274 nm); for δH(400.13 MHz, D2O) and J values, see Table 1;
δC(50.3 MHz, D2O, DEPT) 83.0 (C4Ј), 82.8 (C1Ј), 74.9 (C3Ј),
37.9 (C2Ј), 18.3 (4Ј-Me); m/z (EISϩ–MS, relative intensity) 152
(18, [Gua ϩ H]ϩ), 252 (100, [M ϩ H]ϩ), 274 (7, [M ϩ Na]ϩ),
503 (9, [2M ϩ H]ϩ).
5Ј-O-(2-Deoxy-á-D-erythro-pentofuranosyl)-2Ј-deoxyguano-
sine (13). λmax(H2O)/nm 254 (shoulder at 272 nm); for δH(400.13
MHz, D2O) and J values, see Table 2; for δC(50.3 MHz, D2O)
values, see Table 3; m/z (EISϩ–MS, relative intensity) 104 (100),
152 (11, [Gua ϩ H]ϩ), 384 (15, [M ϩ H]ϩ), 274 (7, [M ϩ Na]ϩ),
767 (3, [2M ϩ H]ϩ).
9-Ethenylguanine (14). λmax(H2O)/nm 266; for δH(400.13
MHz, D2O) and J values, see Table 1; δC(50.3 MHz, D2O) 137.0
(CH, C8), 125.9 (CH, C1Ј), 105.6 (CH2, C2Ј); m/z (EISϩ–MS,
relative intensity) 152 (6, [Gua ϩ H]ϩ), 178 (100, [M ϩ H]ϩ),
200 (5, [M ϩ Na]ϩ).
9-(2,3-Dideoxy-3,4-didehydro-â-D-erythro-pentofuranosyl)-
guanine (15). λmax(H2O)/nm 254 (shoulder at 274 nm); for
δH(400.13 MHz, D2O) and J values, see Table 1; δC(50.3 MHz,
143.4 (C4Ј), 137.0 (C8), 100.7 (C3Ј), 83.1 (C1Ј), 34.5 (C2Ј);
m/z (EISϩ–MS, relative intensity): 152 (25, [Gua ϩ H]ϩ), 220
(100, [M ϩ H]ϩ).
References
1 G. Kraft, M. Krämer and M. Scholz, Radiat. Environ. Biophys.,
1992, 31, 161.
2 J. Kieffer, Int. J. Radiat. Biol., 1985, 48, 873.
3 J. Kieffer and E. Schneider, Int. J. Radiat. Biol., 1991, 59, 1415.
4 G. Kraft, Nucl. Sci. Appl., 1987, 3, 1.
5 E. A. Blakely, Radiat. Environ. Biophys., 1992, 31, 181.
6 B. Rydberg, Radiat. Res., 1994, 139, 133.
7 B. Rydberg, Radiat. Res., 1996, 145, 200.
8 V. Michalik, Int. J. Radiat. Biol., 1992, 62, 9.
9 V. Michalik, Radiat. Res., 1993, 134, 265.8.
10 D. T. Goodhead, Int. J. Radiat. Biol., 1994, 65, 7.
11 W. R. Holley and A. Chatterjee, Radiat. Res., 1996, 145, 188.
12 A. Schaefer, J. Hüttermann and G. Kraft, Int. J. Radiat. Biol., 1993,
63, 139.
13 D. Becker, Y. Razskazovskii, M. U. Callaghan and M. Sevilla,
Radiat. Res., 1996, 146, 361.
14 R. Roots, A. Chatterjee, P. Chang, L. Lommel and E. A. Blakely,
Int. J. Radiat. Biol., 1985, 47, 157.
15 J. Stanton, G. Taucher-Scholz, M. Schneider and G. Kraft, Radiat.
Prot. Dos., 1990, 31, 253.
16 A. Chatterjee and W. R. Holley, Adv. Space Res., 1992, 12, 33.
17 M. Berger and J. Cadet, Z. Naturforsch., Teil B, 1985, 40, 1519.
18 A. Shaw and J. Cadet, Int. J. Radiat. Biol., 1976, 70, 1.
19 N. Mariaggi, J. Cadet and R. Téoule, Tetrahedron, 1976, 32,
2386.
20 A. Shaw and J. Cadet, Int. J. Radiat. Biol., 1988, 54, 987.
21 M. Berger and J. Cadet, Chem. Lett., 1983, 435.
22 H. Kasai, Z. Yamaizumi, M. Berger and J. Cadet, J. Am. Chem.
Soc., 1992, 114, 9692.
9-(5-Hydroxymethylfuran-2-yl)guanine (16). λmax(H2O)/nm
248 (shoulder at 275 nm); for δH(400.13 MHz, D2O) and
J values, see Table 1; δC(50.3 MHz, D2O, sugar carbons) 152.5
(C4Ј), 140.2 (C1Ј), 137.5 (C8), 110.4 (C3Ј), 105.4 (C2Ј), 55.7
J. Chem. Soc., Perkin Trans. 2, 1998
1373

View Article Online
23 R. A. Floyd, J. J. Watson, P. K. Wong, D. H. Altimiller and
R. C. Rickard, Free Rad. Res. Commun., 1986, 1, 163.
24 M. Berger, C. Anselmino, J.-F. Mouret and J. Cadet, J. Liq.
Chromatogr., 1990, 13, 929.
28 M. J. Robins, A. J. Roger and R. Mengel, Can. J. Chem., 1977, 55,
1251.
25 C. Decarroz, J. R. Wagner, J. E. van Lier, C. Murali Krishna,
P. Riesz and J. Cadet, Int. J. Radiat. Biol., 1976, 50, 491.
26 J. Cadet, C. Taïeb, M. Remin, W. P. Niemczura and F. E. Hruska,
Biochim. Biophys. Acta, 1980, 608, 435.
27 J. Zemlicka, R. Gasser, J. V. Freisler and J. P. Horwitz, J. Am. Chem.
Soc., 1972, 94, 3213.
Paper 8/01069B
Received 5th February 1998
Accepted 13th March 1998
1374
J. Chem. Soc., Perkin Trans. 2, 1998