![FIG. 2. Lower curve: the relative DEA cross section of thymine [10(b)], shifted by 0.2 eV to lower energy. Upper curve: A model of the electron capture cross section of DNA as a function of electron energy based on the resonance energies of the bases and their widths determined in gas-phase scattering studies. The curve has been shifted to higher energy by 0.41 eV, normalized to the SSB data and a linearly increasing background (dashed line) added. Closed squares: SSB yields from Fig. 1.](/figures/fig-2-lower-curve-the-relative-dea-cross-section-of-thymine-cqn6s0lx.webp)
University of Nebraska - Lincoln University of Nebraska - Lincoln
DigitalCommons@University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln
Paul Burrow Publications Research Papers in Physics and Astronomy
August 2004
DNA Strand Breaks Induced by 0–4 eV Electrons: The Role of DNA Strand Breaks Induced by 0–4 eV Electrons: The Role of
Shape Resonances Shape Resonances
Frédéric Martin
Université de Sherbrooke
Paul Burrow
University of Nebraska-Lincoln
, pburrow1@unl.edu
Zhongli Cai
Universite´ de Sherbrooke
Pierre Cloutier
Universite´ de Sherbrooke
Darel Hunting
Universite´ de Sherbrooke,
See next page for additional authors
Follow this and additional works at: https://digitalcommons.unl.edu/physicsburrow
Part of the Physics Commons
Martin, Frédéric; Burrow, Paul; Cai, Zhongli; Cloutier, Pierre; Hunting, Darel; and Sanche, Léon , "DNA Strand
Breaks Induced by 0–4 eV Electrons: The Role of Shape Resonances" (2004).
Paul Burrow Publications
. 2.
https://digitalcommons.unl.edu/physicsburrow/2
This Article is brought to you for free and open access by the Research Papers in Physics and Astronomy at
DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Paul Burrow Publications by
an authorized administrator of DigitalCommons@University of Nebraska - Lincoln.
DNA Strand Breaks Induced by 0– 4 eV Electrons: The Role of Shape Resonances
Fre
´
de
´
ric Martin,
1
Paul D. Burrow,
2
Zhongli Cai,
1
Pierre Cloutier,
1
Darel Hunting,
1
and Le
´
on Sanche
1,
*
1
Faculty of Medicine, Universite
´
de Sherbrooke, Quebec, J1H 5N4, Canada
2
Department of Physics and Astronomy, University of Nebraska-Lincoln, Lincoln, Nebraska 68588-0111, USA
(Received 15 March 2004; published 3 August 2004)
Collisions of 0– 4 eV electrons with thin DNA films are shown to produce single strand breaks. The
yield is sharply structured as a function of electron energy and indicates the involvement of
shape
resonances in the bond breaking process. The cross sections are comparable in magnitude to those
observed in other compounds in the gas phase in which
electrons are transferred through the
molecule to break a remote bond. The results therefore support aspects of a theoretical study by Barrios
et al. [J. Phys. B 106, 7991 (2002)] indicating that such a mechanism could produce strand breaks in
DNA.
DOI: 10.1103/PhysRevLett.93.068101 PACS numbers: 87.50.Gi, 34.50.Gb, 34.80.Ht, 87.14.Gg
Introduction.—Many investigations during the last
century have been devoted to the study of alterations
induced by high-energy radiation in biological systems,
more particularly within living cells and the DNA mole-
cule. The biological effects of such radiation are usually
not produced by the impact of the primary quanta, but
rather by the secondary species generated along the ra-
diation track [1]. As these species further react within
irradiated cells, they can cause mutagenic, genotoxic, and
other potentially lethal DNA lesions [2], such as single-
and double-strand breaks (SSBs and DSBs).
Secondary electrons produced with energies below
20 eV are the most abundant of the secondary species
[3,4], and their energy distribution is further degraded
through ionization and inelastic collision processes. To
understand the effects of radiation in cells, it is therefore
crucial to determine the damage induced by such elec-
trons on vital cellular components, particularly DNA. In
experiments directly exposing DNA on a surface to an
electron beam, Boudaiffa et al. [5] found that SSBs and
DSBs could be produced at much lower energies than
previously observed. The strand break yield as a function
of electron impact energy peaked near the thresholds for
electronic excitation, and the authors suggested that the
process takes place through short-lived core-excited
anion states, whose ‘‘parents’’ are the excited states of
the neutral molecule [6]. Later Pan et al. [7] demonstrated
that the dissociative electron attachment (DEA) process
contributes significantly to these breaks. At still lower
energies, electrons captured temporarily by simple mole-
cules into normally empty valence orbitals (i.e., shape
resonances) can also produce bond breaking through
DEA, if the latter is exothermic [6]. Two properties of
shape resonances are particularly relevant to the present
work. Because the magnitude of the capture cross section
for formation of a resonance varies inversely with the
resonance energy, the cross sections may be very large if
these resonances lie at low energy [6]. Furthermore, the
lifetimes of temporary anion states increase as their
energies decrease [6]. This combination of properties
suggests that, if present, low-lying shape resonances in
DNA could be highly efficient at breaking DNA strands.
Recently, a number of theoretical and experimental in-
vestigations [8–11] on basic DNA constituents have gen-
erated considerable interest [12] in this possibility.
In the present Letter, we show that single strand breaks
in DNA are produced at energies as low as the nominal
zero energy threshold of the electron beam, and that the
yield as a function of energy exhibits a sharp peak at
0:8 0:3eVand a broader feature centered at 2:2eV.
We compare these results to those obtained in gas-phase
DEA [10,11,13,14] and electron transmission spectros-
copy (ETS) [15] experiments on the DNA bases and in
theoretical studies [8,9]. We provide evidence that the
SSBs are initiated in part by electron attachment into
the empty
valence molecular orbitals [16] of the DNA
bases, and we show that the yield is comparable in mag-
nitude to that observed in gas-phase DEA processes, in
which electrons in
orbitals are transferred to a ‘‘re-
mote’’ bond that is subsequently broken. Direct attach-
ment into low-lying
orbitals [16] of the phosphate
group may also play a role, but gas-phase results support-
ing DEA via this process are less conclusive.
Procedure.—In the present experiment, supercoiled
DNA [pGem-3Zf(-), 3197 base pairs] was prepared and
purified as previously described [17]. An aqueous solution
of the molecule as a sodium salt was obtained in the last
step of the purification procedure, where the DNA was
washed with a buffer containing sodium perchlorate.
Thus, the negative charge of the phosphate groups is
counterbalanced by Na
ions. Under a dry nitrogen at-
mosphere, 125 ng of DNA [17] in 10 l of H
2
O was
deposited on a chemically clean tantalum sheet, frozen
at 70
C and lyophilized with a hydrocarbon-free sorp-
tion pump at a pressure of 3 mTorr for two hours.
Lyophilized DNA formed a film with a diameter of 3:5
0:2mm that was directly transferred to an ultrahigh
vacuum chamber (UHV). After a 24 h evacuation, the
VOLUME 93, NUMBER 6
PHYSICAL REVIEW LETTERS
week ending
6 AUGUST 2004
068101-1 0031-9007=04=93(6)=068101(4)$22.50 2004 The American Physical Society 068101-1
DNA was exposed to an electron beam at a background
pressure of 5 10
9
Torr at room temperature, at a fixed
incident electron current of 2.0 nA (current density of
1:6 10
10
electrons s
1
cm
2
) and a constant incident
electron energy, for irradiation times of seconds up to
4 min. The energy resolution of the beam was 0.5 eV full-
width at half-maximum (FWHM). The beam energy was
determined relative to the vacuum level within 0:3eV
by measuring the energy for the onset of current trans-
mitted through the film. The data was recorded in the
energy range 0:1 to 4.1 eVat 0.3 eV intervals. The area of
the electron beam was adjusted to be slightly smaller than
that of the sample.
The SSBs and DSBs were detected by measuring the
transformation induced by electron impact on DNA, from
the initial supercoiled configuration of the molecule to the
circular or linear forms. When a single strand is broken,
the strained energy of the twisted supercoiled configura-
tion is released causing the DNA topological configura-
tion to become circular, whereas a cut of the two adjacent
strands necessarily produces a linear DNA molecule.
Once removed from the UHV chamber, the DNA was
recovered and analyzed as previously described [17].
The different forms of DNA were separated by gel elec-
trophoresis and the percentage of each form was quanti-
fied by fluorescence.
Exposure response curves were obtained for several
incident electron energies. As an example, the inset of
Fig. 1 shows the dependence of the percentage yields of
circular DNA on irradiation time for 0.6 eVelectrons. The
amount of the linear form of plasmid DNAwas below the
detection limit of 0.2 ng. Thus, induced DSBs are esti-
mated to be less than one per 10
5
electrons. During the
first 20 sec, the percentage of circular DNA increases
linearly with irradiation time, but saturates thereafter,
possibly due to film charging, as seen from changes in
the energy onset of electron current at the film substrate.
Fluorescent microscopy shows that about 10% of the DNA
film surface is covered with clusters, while the rest con-
sists of a loosely packed monolayer of DNA. Charging is
likely to arise from electron trapping within the clusters.
Yields, e.g., DNA SSB or DSB per incident electron at
each incident electron energy, were determined from the
amounts of circular or linear DNA, respectively, resulting
from 10 sec of exposure (i.e., well within the initial linear
regime of the exposure response curves) at the incident
electron current and the known amount of DNA on the
target. The loss of supercoiled DNA mirrors the formation
of circular DNA. The samples were also placed under the
electron gun for 10 sec when a voltage of 2:0V was
applied to the target, resulting in no current at the target.
The absence of formation of circular and linear DNA in
these controls confirms that heat and light from the
electron gun do not cause any strand breaks.
Results and discussion.—Figure 1 shows the depen-
dence of the yields of SSBs and DSBs on incident electron
energy. In total, 80 different films were bombarded and
analyzed to produce these results. The error bars show
the standard deviation from three to eight exposure ex-
periments, each on separately prepared samples. Two
peaks are observed in the yield function of SSBs at
electron energies of 0.8 and 2.2 eV with yields of 1:0
0:110
2
, 7:5 1:510
3
SSB per incident elec-
tron, respectively. The peaked structure in Fig. 1 provides
unequivocal evidence for the role of low-lying temporary
anion states in the bond breaking process. Of the basic
molecular components comprising nucleic acids, only the
resonances of the four DNA bases and the RNA base
uracil have received significant attention over the energy
range of interest. In the gas phase, sharp structures in the
total scattering cross sections of the bases have been
observed [15] over the range from 0.29 to 4.5 eV using
ETS and attributed to temporary electron occupation of
the lowest * valence orbitals. The assignment of these
anion states, whose energies correspond to vertical at-
tachment energies (VAE), is fully supported by
Koopmans’s theorem ab initio calculations, scaled to
resonances in related molecules as described elsewhere
[15].
The DEA process in the DNA bases has received much
recent attention, and the yields of fragment anions in the
gas-phase display significant resonance structure at low
energies [10,13]. Scheer et al. [11] have analyzed the
structures appearing in DEA and ETS studies of the
bases and halo-substituted bases and attributed them to
two mechanisms. The most prominent and narrow struc-
tures were assigned to vibrational Feshbach resonances
[6] arising from mixing between the dipole bound states
of the bases and the anion states associated with the
lowest
valence orbitals. The remaining features were
attributed to vibronic coupling between the valence
and repulsive anion states, giving rise to DEA peaks
FIG. 1. Quantum yield of DNA single strand breaks (SSBs)
and double-strand breaks (DSBs) vs incident electron energy.
The inset shows the dependence of the percentage of circular
DNA (i.e., SSBs) on irradiation time for 0.6 eV electrons.
VOLUME 93, NUMBER 6
PHYSICAL REVIEW LETTERS
week ending
6 AUGUST 2004
068101-2 068101-2
very near the energies of the
resonances. This process
has been well studied in unsaturated chlorocarbons [18].
Experimental results related to the empty orbitals of the
remaining DNA components, the deoxyribose and phos-
phate groups, are sparse. A preliminary ETS study of
trimethylphosphate [19], a surrogate for the DNA phos-
phate group, indicates a broad temporary anion state near
2eV, consistent with calculations of the
lowest unoc-
cupied molecular orbital (LUMO) anion state. Finally,
the fully saturated deoxyribose group, consisting only of
O, C, and H atoms, should also possess only short-lived
anion states.
A pronounced and narrow peak appears near 1 eV in the
DEA cross sections of uracil [10(a)], adenine, thymine
[13], and thymidine [14] in the gas phase. Representative
data for thymine [10(b)] are shown in the lower curve of
Fig. 2, but shifted to lower energy by 0.2 eV in order to
match the location of the 0.8 eV peak in the yield of SSBs.
Such resonances might account for the 0.8 eV peak in
Fig. 1, if electron transfer to the backbone could occur
from them and they exist in DNA. However, we note that
the spacing of the two major peaks in the thymine DEA
cross section does not match that of the strand breaks. The
sharp 1 eV peak in uracil has been assigned by Scheer et
al. to a vibrational Feshbach resonance associated with a
dipole bound anion state. Because the wave functions of
such states are very diffuse, they overlap with other sites
in the condensed phase and, in the absence of evidence to
the contrary, we do not expect them to play a role in
surface studies. On the other hand, the
resonances
associated with the LUMOs of thymine, cytosine, and
guanine are sufficiently long lived in the gas phase to
show clear evidence for anionic vibrational motion [15].
The natural width of the lowest, and largest, vibrational
feature observed by ETS is 150 meV in these com-
pounds. This property thus argues for an assignment of
the 0.8 eV peak in Fig. 1 to a DEA process taking place
through occupation of the LUMO
orbitals of one or
more bases.
Insight into the role of the
anion states of the bases
in bond breaking is found in a theoretical study by Barrios
et al. [8]. In this work, a section of DNAwas modeled that
contains a cytosine base, a sugar ring, and the phosphate
group. They find that an anionic potential surface exists
that connects the initial
anion state of the base to a
anion state. The latter leads to rupture of the C-O bond
connecting the phosphate group to the sugar. In other
words, an electron placed on the base will migrate to
the C-O
antibonding orbital as the latter bond length
is stretched, leading to bond rupture. Transport of an
electron from the base to the sugar-phosphate bond
must take place through three saturated bonds. There is
ample precedence for such transfers leading to bond
breaking in gas-phase DEA studies. Pearl et al. [20,21]
observed electron transfer from an ethylenic
anion
state to a C-Cl
anion state through two [20] and four
[21] saturated bonds in two rigid chloronorbornene com-
pounds. In each case, the maximum DEA cross section
for production of Cl
occurs very near the VAE for
formation of the
resonance. The reaction in these
compounds is exothermic because of the large electron
affinity of the chlorine atom. It is instructive to compare
the DEA cross sections in these compounds with the SSB
yields found in the present work. In the chloronorbornene
compound [20], the ratio of the DEA cross section at its
1.1 eV peak to the theoretical maximum reaction cross
section,
2
,is5:4 10
2
, where is the de Broglie
wavelength of the electron at this energy. In the larger
compound [21], in which transfer occurs through four
saturated bonds, this ratio drops to 10
2
. In the DNA
base-sugar-phosphate system involving three saturated
bonds, we would anticipate an intermediate value if the
total DNA electron scattering cross section were entirely
resonant in character. More reasonably, such a value
should be considered an upper limit. As measured here,
the SSB yield per electron of 10
2
in DNA at 0.8 eV is
entirely consistent with this picture. Considering the dif-
ferences in the molecular systems compared above, this
result should not be overinterpreted. However, it offers
clear support for the charge transfer mechanism proposed
by Barrios et al. [8].
To further support the role of the
resonances, we
simulate in an approximate manner the electron capture
cross section as it might appear in DNA owing to the
anion states of the bases. For this purpose, we represent
FIG. 2. Lower curve: the relative DEA cross section of thy-
mine [10(b)], shifted by 0.2 eV to lower energy. Upper curve: A
model of the electron capture cross section of DNA as a
function of electron energy based on the resonance energies
of the bases and their widths determined in gas-phase scatter-
ing studies. The curve has been shifted to higher energy by
0.41 eV, normalized to the SSB data and a linearly increasing
background (dashed line) added. Closed squares: SSB yields
from Fig. 1.
VOLUME 93, NUMBER 6
PHYSICAL REVIEW LETTERS
week ending
6 AUGUST 2004
068101-3 068101-3