D.E. Williams et al. / Journal of Inorganic Biochemistry 168 (2017) 55–66
61
that ranges from approximately 290 nm to 410 nm with a maximum
output value at 350 nm. While copper(II) chloride absorbs poorly in
this region, there is major overlap between the lamp output and the ab-
sorption spectrum of complex 4 (λmax = 323 nm) and slight overlap
with the absorption spectrum of DNA (λmax = 259 nm; Fig. S5). Rongoni
and co-workers used Cu-K-Edge X-ray absorption spectroscopy to mon-
itor the oxidation states of copper ions directly bound to DNA. When CT
DNA samples were pre-equilibrated with CuCl2 and then irradiated at
310 nm, the authors demonstrated that approximately 35% of DNA-
bound copper was reduced to copper(I) as a result of one-electron
transfer from DNA to copper(II) [62].
The photo-reduction of Cu(II) can be viewed in the context of DNA
photocleavage yields. As stated, when 38 μM bp of DNA was present
in reactions, irradiating 30 μM of CuCl2 and 10 μM of complex 4 gener-
ated intermediate and high levels of copper(I) respectively (Fig. 4A).
By comparison, 30 μM of CuCl2 produced extremely low levels of DNA
photo-damage (Fig. S6) compared to 2 μM of complex 4, which caused
near-complete DNA degradation (Figs. 2, 3, and S3). The apparent disso-
ciation constant reported for the major Cu(II) binding site in double-he-
lical DNA is in the micromolar range [47], suggesting that copper ions
were directly bound to DNA in the BCS and photocleavage experiments
that employed CuCl2 (Figs. 4A and S6). In the presence of ground state
triplet oxygen (3O2), the reduction of Cu(II) to Cu(I) generates DNA
cleaving reactive oxygen species (ROS), either free hydroxyl radicals
(•OH) [31,43–46] (as in a Cu(I)-based Fenton-type reaction) or Cu(I)-
peroxides [65–68]. If copper(I) had played a role in ROS production
and DNA photocleavage in our experiments, then it might have done
so in an inefficient fashion when bound directly to DNA. The hat-
(COO-)6 ligand of complex 4 is clearly instrumental in photosensitizing
DNA damage.
were markedly altered as a result. A broad 425 nm to 475 nm shoulder
absent in the case of hat-(COO−
6 (3), but present for (Cu(en))3hat-
)
(COO)6 (4) was reduced in intensity (Fig. S9 in Supplementary data).
While it is possible that EDTA also diminished weak interactions be-
tween Na(I) ions and carboxylate groups of 4, the spectral changes illus-
trated in Fig. S9 suggest that photocleavage inhibition by EDTA involves
an interaction between the chelating agent and the copper(II) centers of
4.
The singlet oxygen (1O2) scavenger sodium azide also exhibited a
strong inhibitory effect on DNA photocleavage. Polypyridyl com-
plexes based on copper(II) and other metal ions are capable of
photosensitizing the production of DNA damaging singlet oxygen
through a type 2 energy transfer pathway [11–16,31,32]. While alka-
line- and piperidine labile lesions at guanine bases are the most com-
monly observed reaction products, singlet oxygen also generates
direct DNA strand breaks. Experiments conducted in the presence of
the singlet oxygen scavenger sodium azide and D2O, a solvent that in-
creases the lifetime of singlet oxygen, have shown that 1O2 forms direct
strand breaks at guanines under neutral to near-neutral reaction condi-
tions (pH 7.0 to 7.4), without requiring subsequent treatment with a
base to induce the cleavage [30,32,33,36,69–72]. This being said, we
used D2O in an attempt to confirm 1O2 involvement in our reactions.
However, replacing ddH2O with the D2O failed to enhance complex 4-
sensitized photocleavage yields (Table 1, Fig. S10). This pointed to the
possibility that a scavenging reaction between sodium azide and singlet
oxygen had never occurred. Taking into consideration that azide can
serve as a copper binding ligand in aqueous solution [73,74], the D2O
data suggest that the DNA photocleavage inhibition displayed by sodi-
um azide may have arisen from a disruptive interaction between
azide anions and the copper(II) centers of 4.
When the high intensity broad spectrum 290 nm to 410 nm RPR-
3500 Å Rayonet lamps used the photocleavage experiments were re-
placed with a low intensity 390 to 395 nm LED flashlight, complex 4
was still capable of generating DNA photocleavage (Fig. S7). The LED ex-
citation wavelengths were well outside of the range of DNA absorption
(Fig. S5), showing that direct excitation of DNA by light is not required
for complex 4 to photosensitize DNA damage.
Inhibition by DMSO, sodium benzoate, catalase, and SOD, respective-
ly indicates that hydroxyl radicals, hydrogen peroxide, and superoxide
anion radicals participate in complex 4 sensitized DNA photocleavage
(Table 1; Figs. S8 and S10). While hydrogen peroxide and superoxide
anion radicals are relatively unreactive towards nucleic acids [75], hy-
droxyl radicals are powerful biological oxidants that produce direct
DNA strand breaks by abstracting hydrogen atoms from deoxyribose
[75,76]. The reaction of ground state triplet oxygen with copper(I) has
been shown to trigger Cu(II)/Cu(I) redox cycling and DNA cleavage in-
volving the formation of superoxide anion radicals [43–45], hydrogen
peroxide [44], and hydroxyl radicals [43,44,46]. It is therefore conceiv-
able that the hydroxyl radicals responsible for DNA cleavage in our ex-
periments are generated by a process similar to the series of reactions
shown in Fig. 5. It is possible to consider a scenario in which photo-
assisted electron transfer from DNA to complex 4 reduces Cu(II) to
Cu(I). The Cu(I) then reacts with ground state triplet oxygen to produce
superoxide anion radicals that undergo spontaneous dismutation to
form hydrogen peroxide [45]. Hydroxyl radicals, hydroxide anions,
and Cu(II) could subsequently be produced via a Fenton-type reaction.
3.4. Reagent-induced changes in DNA photocleavage
The effects of enzymatic and chemical reagents on the DNA
photocleaving activity of (Cu(en))3hat-(COO)6 (4) were studied next.
Separate reactions consisting of 38 μM bp of pUC19 plasmid and 1 μM
to 2 μM of 4 were pre-equilibrated with either the singlet oxygen
(1O2) scavenger sodium azide, the hydroxyl radical (•OH) scavengers
sodium benzoate and DMSO, the hydrogen peroxide (H2O2) scavenger
catalase, the superoxide anion radical (O•2-) scavenger superoxide dis-
mutase (SOD), EDTA, and D2O. Out of all of the reagents tested, the
most significant effect on DNA photocleavage was exhibited by EDTA.
Single- and double-strand break formation was almost completely
prevented by the metal chelator (Table 1, Fig. S8). UV–visible spectra
were then recorded upon the addition of EDTA to an aqueous solution
of (Cu(en))3hat-(COO)6 (4). The optical properties of the complex
3.5. DNA binding mode analyses
3.5.1. UV–visible spectrophotometry
In order to identify additional factors contributing to the high levels
of DNA photocleavage produced by (Cu(en))3hat-(COO)6 (complex 4),
a series of binding mode studies was undertaken. Polypyridyl com-
plexes are known to interact with DNA either electrostatically, through
Table 1
Average % change in DNA photocleavage induced by scavengers, D2O, and EDTA.
Reagent
Species targeted
Photocleavage change (%)a
EDTA
Sodium azide
DMSO
Catalase
Sodium benzoate
SOD
Cu(II)
1O2
−90
−89
−71
−69
−58
−39
−5
2
2
7
3
8
2
•OH
H2O2
•OH
O•2-
D2O
1O2
1
a
Data are averaged over three trials and error is reported as standard deviation.
Photocleavage gels appear in Supplementary data (Figs. S8 and S10).
Fig. 5. Proposed mechanism for ROS production by copper(I).