Anthraquinone photonucleases: Mechanisms for GG-selective and nonselective cleavage of double-stranded DNA

Article abstract of DOI:10.1021/ja953714w

Anthraquinone derivatives 2AQA2(HEt₂) and 2AQC2(HEt₂) were examined as light-activated agents that initiate DNA cleavage. The substituents control the electronic configuration of the lowest excited state of the anthraquinone. 2AQC2(HEt₂) has a lowest nπ* excited state and can react by electron transfer or hydrogen atom abstraction. 2AQC2(HEt₂) has a lowest excited state of ππ* or intramolecular charge-transfer character and reacts only by electron transfer. Spectroscopic evidence indicates that both quinones bind to double-stranded DNA by intercalation with essentially the same affinity. Picosecond time-resolved laser spectroscopy shows that single electron transfer from the DNA bases to either bound quinone occurs rapidly and to the same extent. Irradiation of either intercalated 2AQA2(HEt₂) or 2AQC2(HEt₂) followed by treatment with hot piperdine leads to equally effective cleavage of DNA at the 5'-G of GG steps. These findings indicate that electron transfer from a DNA base to the excited quinone is the dominant path for the GG-selective DNA cleavage. At high concentrations, where some quinone is free in solution, irradiation of 2AQC2(HEt₂), but not 2AQA2(HEt₂), leads to nonselective spontaneous cleavage of DNA. This second path to DNA cleavage is identified as direct hydrogen atom abstraction from the deoxyribose backbone by excited, nonintercalated 2AQC2(HEt₂).

Full text of DOI:10.1021/ja953714w

VOLUME 118, NUMBER 10
MARCH 13, 1996
© Copyright 1996 by the
American Chemical Society
Anthraquinone Photonucleases: Mechanisms for GG-Selective
and Nonselective Cleavage of Double-Stranded DNA
David T. Breslin and Gary B. Schuster*
Contribution from the School of Chemistry and Biochemistry, Georgia Institute of Technology,
Atlanta, Georgia 30332
ReceiVed NoVember 3, 1995X
Abstract: Anthraquinone derivatives 2AQA2(HEt2) and 2AQC2(HEt2) were examined as light-activated agents that
initiate DNA cleavage. The substituents control the electronic configuration of the lowest excited state of the
anthraquinone. 2AQC2(HEt2) has a lowest nπ* excited state and can react by electron transfer or hydrogen atom
abstraction. 2AQC2(HEt2) has a lowest excited state of ππ* or intramolecular charge-transfer character and reacts
only by electron transfer. Spectroscopic evidence indicates that both quinones bind to double-stranded DNA by
intercalation with essentially the same affinity. Picosecond time-resolved laser spectroscopy shows that single electron
transfer from the DNA bases to either bound quinone occurs rapidly and to the same extent. Irradiation of either
intercalated 2AQA2(HEt2) or 2AQC2(HEt2) followed by treatment with hot piperdine leads to equally effective
cleavage of DNA at the 5′-G of GG steps. These findings indicate that electron transfer from a DNA base to the
excited quinone is the dominant path for the GG-selective DNA cleavage. At high concentrations, where some
quinone is free in solution, irradiation of 2AQC2(HEt2), but not 2AQA2(HEt2), leads to nonselective spontaneous
cleavage of DNA. This second path to DNA cleavage is identified as direct hydrogen atom abstraction from the
deoxyribose backbone by excited, nonintercalated 2AQC2(HEt2).
Introduction
possesses significant practical advantages. In particular, pho-
tonucleases can be triggered by exposure to light. Light is an
attractive cofactor since it is easy to manipulate.
We previously reported the efficient nicking of supercoiled
plasmid DNA by water-soluble, UV-light-activated anthraquino-
ne derivatives.3 These photonucleases are catalytic, being
regenerated by reaction with oxygen, and therefore may initiate
coincident site double-strand cleavage of DNA.4 Absorption
and circular dichroism spectra indicate that some of these
anthraquinone derivatives form intercalative complexes with
The design and investigation of artificial nucleases,1 small
molecules capable of cleaving duplex or single-stranded DNA,
is an area of exciting research that has numerous important
biochemical and biomedical applications. While most of these
agents are activated thermally, there is an increasing emphasis
on photoactivated cleavage agents2 since this methodology
X Abstract published in AdVance ACS Abstracts, February 15, 1996.
(1) A recent review of chemical nucleases is available: Sigman, D. S.;
Mazumder, A.; Perrin, D. M. Chem. ReV. 1993, 93, 2295-2316.
(2) Photonucleases: Sitlani, A.; Long, E. C.; Pyle, A. M.; Barton, J. K.
J. Am. Chem. Soc. 1992, 114, 2303-2312. Nielsen, P. E.; Hiort, C.;
Sonnichsen, S. H.; Buchardt, O.; Dahl, O.; Norden, B. J. Am. Chem. Soc.
1992, 114, 4967-75. Hertzberg, R. P.; Dervan, P. B. Biochemistry 1984,
23, 3934. Liang, G.; Gannett, P.; Shi, X.; Zhang, Y.; Chen, F.-X.; Gold,
B. J. Am. Chem. Soc. 1994, 116, 1131. Ranganathan, D.; Patel, B. K.;
Mishra, R. K. J. Chem. Soc., Chem. Commun. 1994, 107. Riordan, C. G.;
Wei, P. J. Am. Chem. Soc. 1994, 116, 2189. Takasaki, B. K.; Chin, J. J.
Am. Chem. Soc. 1994, 116, 1121. Schnaith, L. M. T., Hanson, R. S.; Que,
L., Jr. Proc. Nat. Acad. Sci. U.S.A. 1994, 91, 569. Nagai, K.; Carter, B. J.;
Xu, J.; Hecht, S. M. J. Am. Chem. Soc. 1991, 113, 5099. Nicolau, K. C.;
Maligres, P.; Shin, J.; de Leon, E.; Rideout, D. J. Am. Chem. Soc. 1990,
112, 7825. Neyhart, G. A.; Grover, N.; Smith, S. R.; Kalsbeck, W. A.;
Fairley, T. A.; Cory, M.; Thorp, H. H. J. Am. Chem. Soc. 1993, 115, 4423.
(3) Koch, T.; Ropp, J. D.; Sligar, S. G.; Schuster, G. B. Photochem.
Photobiol. 1993, 58, 554-558.
(4) Armitage, B.; Yu, C.; Devadoss, C.; Schuster, G. B. J. Am. Chem.
Soc. 1994, 116, 9847-9859.
0002-7863/96/1518-2311$12.00/0 © 1996 American Chemical Society

2312 J. Am. Chem. Soc., Vol. 118, No. 10, 1996
Breslin and Schuster
Chart 1
Figure 1. Phosphorescence spectra of 2AQA2(HEt2) and 2AQC2(HEt2)
recorded at 77 K in a 75% PBS-25% glycerol and ethanol, respectively.
The emission spectra have been normalized.
DNA. Time-resolved laser spectroscopic studies indicated that
two processes are initiated when the anthraquinone-DNA
complexes are irradiated: (i) photoinduced electron transfer from
a neighboring nucleic acid base to the excited anthraquinone
which was directly detected and (ii) hydrogen atom abstraction
from a deoxyribose unit which was inferred by analysis of the
decay kinetics.4 DNA cleavage can occur from either pathway,
and the initial spectroscopic observations were interpreted to
indicate that the hydrogen atom transfer route was the most
effective.
Further investigation has revealed that these monosubstituted
anthraquinone derivatives cleave double-stranded DNA selec-
tively at the 5′-G of a GG step.5 Very recently, photonuclease
activity with similar base-pair selectivity has been reported for
three other compounds including two naphthalimide derivatives6
and riboflavin, a naturally occurring imide.7 This pattern of
selective DNA cleavage is attributed to electron-transfer reac-
tions initiated by an electronically excited state.6a
Anthraquinone derivatives 2AQA2(HEt2) and AQC2(HEt2)
(Chart 1)8 were examined to provide insight into the reaction
mechanism of the excited quinones with DNA that leads to
strand cleavage. The excited states of both 2AQA2(HEt2) and
2AQC2(HEt2) are capable of the single electron oxidation of
purine and pyrimidine bases, but only 2AQC2(HEt2) is expected
to efficiently abstract a hydrogen atom. This investigation
reveals that electron transfer is the dominant pathway for the
GG-selective cleavage of DNA by intercalated anthraquinone9
and that nonsequence-selective spontaneous cleavage of DNA
results from reaction with nonintercalated quinone.
For example, ketones that have predominant nπ* character
generally intersystem cross to triplets with high rates, abstract
hydrogen atoms rapidly, and have structured phosphorescence
spectra. In contrast, ketones with predominant ππ* or charge-
transfer character often fluoresce in fluid solution, do not react
rapidly with hydrogen atom donors, and have structureless
phosphorescence spectra. Significantly, the rates of electron-
transfer reactions of excited ketones depend primarily on
energetic factors and are essentially independent of the specific
electronic configuration of the state. We examined the spectral
properties and excited state reactivity of 2AQA2(HEt2) and
2AQC2(HEt2) to assess the electronic configuration of their
excited states.
(a) Absorption and Luminescence of 2AQA2(HEt2) and
2AQC2(HEt2). A strong indication of the distinct electronic
characteristics of 2AQA2(HEt2) and 2AQC2(HEt2) is found in
their absorption spectra. In aqueous solutions, 2AQC2(HEt2)
absorbs with λmax ) 320 nm, which is identical with that of
unsubstituted anthraquinone. In contrast, the absorption maxi-
mum of 2AQA2(HEt2) is found at 375 nm with tailing into the
visible region. This 55 nm bathochromic spectral shift in
absorption maximum clearly signals the participation of low
lying charge-transfer excited states in 2AQA2(HEt2).
The emission spectra of 2AQA2(HEt2) and 2AQC2(HEt2) are
also dramatically different. 2AQC2(HEt2) is not emissive at
room temperature in an aqueous PBS buffer solution, but under
identical conditions, 2AQA2(HEt2) fluoresces, with φfl ) 0.02
and a lifetime of 1.2 ns. At 77 °K, the emission spectrum of
2AQA2(HEt2) broadens to include a phosphorescence compo-
nent with a lifetime of 44 ( 8 µs (in an ethanol glass) (Figure
1), but neither the fluorescence or phosphorescence shows any
vibronic structure. The unstructured emission of 2AQA2(HEt2)
is typical of a ketone triplet with a lowest energy ππ* or charge-
transfer excited state. Strong phosphorescence of 2AQC2(HEt2)
is observed in a frozen 75% PBS-25% ethylene glycol solution
at 77 °K. This spectrum (Figure 1) presents vibronic structure
with a characteristic quinone carbonyl group stretching fre-
quency of 1600 cm-1. Coupling of carbonyl stretching modes
to the electronic state are typical of ketones with lowest nπ*
excited states. Electron-donating substituents on anthraquinones
often introduce a charge-transfer excited state of lower energy
Results
(1) Photochemical and Photophysical Properties of 2AQA2-
(HEt2) and 2AQC2(HEt2). One of the hallmarks of carbonyl
group photochemistry is how dependent some properties are
on the electronic configuration of the lowest energy excited state.
(5) Armitage, B.; Schuster, G. B. Submitted for publication.
(6) (a) Siato, I.; Takayama, M.; Sugiyama, H.; Nakatani, K. J. Am. Chem.
Soc. 1995, 117, 6406-6407. (b) Siato, I.; Takayama, M. J. Am. Chem.
Soc. 1995, 117, 5590-5591.
(7) Ito, K.; Inoue, S.; Yamamoto, K.; Kawanishi, S. J. Biol. Chem. 1993,
268, 13221-13227.
(8) The acronym 2AQC2(HEt2) fully describes the molecular structure
as follows: 2 refers to the position of substitution on the AQ skeleton; AQ
indicates that the molecule is a derivative of anthraquinone; C indicates
that the substituent is linked to the ring system by a carboxamide group; 2
refers to the number of methylene groups in the chain; and (HEt2) indicates
that the terminal ammonium group is diethyl substituted. Other derivatives
bearing sulfonamide or N-amide linkages are designated with S or A,
respectively, in place of C.
(9) We have recently completed NMR studies that further support DNA
intercalation of single-chained monocationic anthraquinone derivatives
(manuscript in preparation).

Anthraquinone Photonucleases
J. Am. Chem. Soc., Vol. 118, No. 10, 1996 2313
Figure 2. Photoreduction of the 0.2 mM anthraquinone derivatives by photolysis at 350 nm in isopropyl alcohol and triethylamine solutions
monitored by UV-vis spectroscopy: (a) 2AQA2(HEt2) in isopropyl alcohol at t ) 0, 20, 40, 60, 80, 120, 160, and 180 min; (b) 2AQC2(HEt2) in
isopropyl alcohol at t ) 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 min; (c) 2AQA2(HEt2) in triethylamine at t ) 0, 1, 2, 3, 4, 5, 6, 7, 8, and 9 min; (d)
2AQC2(HEt2) in triethylamine at t ) 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 min.
than the nπ* state of the parent compound.10 Indeed, emissive
properties similar to those observed for 2AQA2(HEt2) have been
attributed to charge-transfer states of substituted 2-aminoan-
thraquinones.11,12
spectroscopy (Figure 2). The two quinones were irradiated
under identical conditions in a Rayonet reactor. Absorption
spectra recorded as the reactions progressed exhibit clean
isosbestic points, demonstrating the direct and quantitative
conversion of the anthraquinone to a single photoproduct which
was identified by comparison with authentic material as the
corresponding hydroquinone (eq 1).
(b) Photoreactions of 2AQA2(HEt2) and 2AQC2(HEt2)
with Isopropyl Alcohol and Triethylamine. Anthraquinone
derivatives having energetic nπ* excited states generally abstract
a hydrogen atom from isopropyl alcohol to form ketyl radicals.13
Further bimolecular reaction of these radicals often leads to
irreversible photoreduction of the ketone. Corresponding car-
bonyl compounds having ππ* or charge-transfer excited states
react slowly, if at all, with isopropyl alcohol.14 In contrast to
the photoreduction of excited carbonyl compounds by direct
hydrogen atom abstraction, the efficiencies of photoreduction
by electron transfer from triethylamine are generally independent
of the electronic configuration of the excited state.15
The relative rates of photoreduction of 2AQA2(HEt2) and
2AQC2(HEt2) with isopropyl alcohol were assessed by compar-
ing the quantum yields for formation of the hydroquinones at
low conversion. This analysis reveals that 2AQC2(HEt2) reats
with the alcohol more than 1 order of magnitude faster than
2AQA2(HEt2) (Figure 3). In contrast, in triethylamine solvent,
the photoreduction rates of 2AQA2(Et2) and AQC2(HEt2) are
nearly identical. Further, the relative efficiency for photore-
duction of 2AQA2(HEt2) and 2AQC2(HEt2) by triethylamine
is not affected when the reaction is carried out in phosphate-
buffered aqueous solution. However, the quantum yield ratio
(Φ2AQA2(HEt )/Φ2AQC2(HEt )) for photoreduction by isopropyl
The photoreduction of 2AQA2(HEt2) and 2AQC2(HEt2) in
neat, N2-purged isopropyl alcohol was monitored by UV-vis
(10) For a review see: Diaz, A. N. J. Photochem. Photobiol. A 1990,
53, 141-167.
(11) Allen, N. S.; Harwood, B.; McKellar, J. F. J. Photochem. 1978, 9,
559-564.
(12) Allen, N. S.; Harwood, B.; McKellar, Greenhalgh, C. W. J.
Photochem. 1979, 10, 193-197.
(13) Hamanoue, K.; Nakayama, T.; Sasaki, H.; Ikenaga, K.; Ibuki, K.
Bull. Chem. Soc. Jpn. 1992, 65, 3141-3148.
(14) Porter, G.; Suppan, P. Trans. Faraday Soc. 1965, 61, 1664-1667.
(15) Cohen, S. G.; Cohen, J. I. J. Chem. Soc. 1967, 89, 164-165.
2
2

2314 J. Am. Chem. Soc., Vol. 118, No. 10, 1996
Breslin and Schuster
Figure 3. Rates of photoreduction of 2AQA2(HEt2) and 2AQC2(HEt2) in (a) isopropyl alcohol and (b) triethylamine solutions. The corresponding
hydroquinone concentrations of 2AQA2(HEt2) and 2AQC2(HEt2) were monitored at 430 and 390 nm, respectively.
alcohol is reduced further to <0.01 with this change in solvent.
Similar behavior reported for the photoreduction of aminoben-
zophenones16 was attributed to a solvent effect on the relative
contributions of nπ* and charge-transfer states. The reduced
reactivity of 2AQC2(HEt2) in the aqueous solution is similarly
ascribed to greater participation by a charge-transfer state due
to solvent stabilization.
From the photochemical evidence, it is clear that 2AQA2-
(HEt2) is a poorer hydrogen atom abstractor than 2AQC2(HEt2).
Conversely, both anthraquinone derivatives react with triethyl-
amine by electron transfer with approximately equal efficiency.
On this basis, it also seems likely that the lowest excited state
of 2AQC2(HEt2) has predominant nπ* character and that of
2AQA2(HEt2) has mostly ππ* or charge-transfer character.
(2) Ground State Association Complexes of 2AQA2(HEt2)
and 2AQC2(HEt2) with DNA. An important question to
Figure 4. Induced CD spectra of 20 µM 2AQA2(HEt2) and 2AQC2-
address is the nature of the associative interaction between the
(HEt2) complexed with 200 µM calf thymus DNA in PBS solution.
ground state of 2AQC2(HEt2) and 2AQA2(HEt2) with DNA.
To examine this issue, equilibrium binding constants of the
complexes were determined. In addition, the structure of the
AQ-DNA complex was analyzed by the electronic absorption
and induced circular dichroism spectroscopy.
Table 1. Binding of 2AQC4(H3) to Synthetic DNA Duplexes
5′-TATAGGTATATGGATAT-3′ 5′-TATAGCTATATCGATAT-3′
3′-ATATCCATATACCTATA-5′ 3′-ATATCGATATACGTATA-5′
oligo-I: Kassoc ) 1.4 × 105 M-1
Oligo-II: Kassoc ) 1.6 × 105 M-1
The equilibrium binding constants of 2AQA2(HEt2) and
2AQC2(HEt2) with calf thymus DNA in 10 mM phosphate
buffer solution were determined by an equilibrium dialysis
method. The concentration of free quinone was determined
directly or indirectly by fluorescence intensity measurements
in 10 mM phosphate solution. Scatchard analysis of these data
using the McGhee-Von Hippel equation reveals that 2AQA2-
(HEt2) and 2AQC2(HEt2) have similar binding constants, Ka )
9.1 × 105 and 15.7 × 105 M-1 in 10 mM phosphate buffer,
respectively.
The absorption spectra of both 2AQA2(HEt2) and 2AQC2-
(HEt2) are changed when they are complexed with CT-DNA.
For instance, 2AQC2(HEt2) undergoes a 3 nm bathochromic
and shows a 23% hypochromicity upon complexation when
compared with its spectrum in PBS solution. Similarly,
2AQA2(HEt2) exhibits a slightly larger red shift (6 nm) and a
25% hypochromicity under the same conditions. The induced
CD spectra of 2AQA2(HEt2) and 2AQC2(HEt2) with CT-DNA
are shown in Figure 4. The complex of either quinone with
the DNA induces a negative CD band in the region where only
the anthraquinone absorbs. The intensity of the 2AQA2(HEt2)
band is approximately twice that of 2AQC2(HEt2).
As mentioned above, GG-selective cleavage has been ob-
served for these anthraquinone derivatives. At low quinone
concentration this might be due to selective binding at GG sites.
This possibility was examined by measuring the binding
constants for 2AQC4(H3) (see Chart 1 for structure) with the
two oligonucleotides shown in Table 1. Oligo-I contains two
GG steps but is otherwise the same as Oligo-II, which contains
none. The association constants in PBS buffer show that there
is no overwhelming preference for binding at GG steps.
(3) Emission and Time-Resolved Absorption Spectroscopy
of 2AQA2(HEt2), 2AQC2(HEt2), and 2AQC4(H3) with DNA.
In order to understand the interaction of the excited states of
DNA-intercalated 2AQA2(HEt2) and 2AQC2(HEt2) with CT-
DNA, we examined the steady state emission and the transient
(16) Cohen, S. G.; Cohen, J. I. J. Phys. Chem. 1968, 72, 3872-93.

Anthraquinone Photonucleases
J. Am. Chem. Soc., Vol. 118, No. 10, 1996 2315
Figure 5. Transient absorbance spectra obtained from the flash photolysis of (a) 2AQA2(HEt2) and (b) 2AQC4(HEt2) (1.5 mM) complexed with
7.5 mM calf thymus DNA in air-saturated PBS solution. Spectra were recorded at 0 (solid), 50 (dashed), 100 (dotted), and 2000 (dot-dash) ps
delays after the excitation pulse.
absorption spectra of the intermediates formed upon pulsed
irradiation of the intercalated quinones.
The fluorescence of 2AQA2(HEt2) is almost completely
quenched by addition of a 10-fold excess of CT-DNA. Under
the conditions of this experiment, the residual fluorescence
observed from 2AQA2(HEt2) probably results from the 4% of
the quinone that is not bound at this DNA concentration (0.2
mM). Dynamic quenching of the quinone fluorescence by the
DNA is negligible under these conditions because of the short
lifetime of the singlet state of 2AQA2(HEt2) and the low
concentration of DNA. Similarly, 90% of the phosphorescence
intensity of 2AQC2(HEt2) in a 75% PBS-25% ethylene glycol
glass is quenched by CT-DNA. In this experiment, ap-
proximately 82% of 2AQC2(HEt2) is bound to DNA based on
Figure 6. Absorption spectra of 2AQA2(HEt2)•- and 2AQA(HEt2)•-
(0.2 mM) in DMF (1 cm pathlength).
solution binding data in the absence of ethylene glycol.
Time-resolved laser flash photolysis experiments (355 nm,
are commonly observed for quinone radical anions.17 Since the
extinction coefficients of the 2AQC4(H3) and 2AQA2(HEt2)
radical anions in DMS solution are nearly identical, the relative
yield of DNA-bound quinone radical anions formed can be
estimated by comparing their absorbancies in the laser flash
photolysis experiment. On this basis, it appears that irradiations
of DNA-bound 2AQA2(HEt2) and 2AQC4(H3) give nearly
identical yields of their respective radical anions. In both cases,
as expected in the heterogeneous environment of DNA, the
transient radical anions decay following complex kinetic
processes.
20 ps) were carried out to help characterize the reactions initiated
by irradiation of the anthraquinones when they are intercalated
in DNA. Because of the higher concentrations required for these
experiments and the relatively low aqueous solubility of
2AQC2(HEt2), its more soluble analog 2AQC4(H3) was used.
This change is not expected to affect the functional properties
of the excited state quinone.
Pulsed irradiation of a solution containing 2AQA2(HEt2) or
2AQC4(H3) and CT-DNA leads to the instantaneous (less than
20 ps) generation of transient species (Figure 5). These species
were identified as the corresponding radical atoms of the
anthraquinones by their independent generation with potassium
superoxide in DMF. The radical anions formed in DMF are
stable for several minutes, and their absorption spectra are shown
in Figure 6. The spectra of the radical anions produced with
superoxide are not identical with the DNA-bound transients.
We ascribe these spectral shifts to solvent-induced effects that
(4) DNA Cleavage Induced by Irradiation of 2AQA2-
(HEt2) and 2AQC2(HEt2). The selectivity and efficiency of
(17) The absorption maximum of the anthraquinone radical anion is
difficult to predict since it depends on solvent polarity and hydrogen
bonding. Hamanoue, K.; Yokohama, K.; Kajiwara, Y.; Nakayama, T.;
Teranishi, H. Chem. Phys. Lett. 1985, 113, 207-212. Hamanoue, K.;
Nakayama, T.; Suguira, K.; Teranishi, H.; Washio, M.; Tagawa, S.; Tabata,
Y. Chem. Phys. Lett. 1985, 118, 503-506.

2316 J. Am. Chem. Soc., Vol. 118, No. 10, 1996
Breslin and Schuster
Molecular orbital calculations suggest that the GG radical cation
is a thermodynamic sink for an electron hole in duplex DNA.6a
Dioxygen is required to generate GG-specific DNA cleavage
by these irradiated quinones. Piperdine-treatable DNA cleavage
is completely inhibited upon nitrogen purging before photolysis.
Studies of other light-induced DNA cleaving agents have
identified singlet oxygen (1O2),21 formed by energy-transfer
sensitization, or superoxide (O2•-), generated by electron
transfer,22 as participants in damaging DNA. We investigated
their possible role in the 2AQC2(HEt2)-induced GG-selective
photocleavage of DNA. The lifetime of (1O2) increases ca. 10-
fold when the solvent is changed from H2O to D2O,23 and this
typically leads to an increased DNA cleavage efficiency in cases
1
where O2 participates.24 We found that this solvent has no
effect on the amount of DNA cleavage initiated by irradiation
1
of the 2AQC2(HEt2) complex. Thus, O2 appears to play no,
or at most a minor, part in DNA strand scission. Similarly, if
reactions of O2•- initiated the DNA cleavage, this process would
be inhibited by the enzyme superoxide dismutase (SOD). We
observe no change in the efficiency of 2AQC2(HEt2)-induced
DNA cleavage when SOD is present. On this basis we conclude
•-
that O2 plays little or no role in this cleavage reaction.25
Although nearly all (>96%) of the 2AQA2(HEt2) and
2AQC2(HEt2) is bound to DNA under the conditions of the
cleavage experiment, it is unclear whether the observed GG
selectivity originates from free or bound quinone. Since both
2AQA2(HEt2) and 2AQC2(HEt2) intercalate in double-stranded
DNA, increasing the concentration of the quinone to a value
greater than the number of intercalation sites will ensure that a
significant fraction of the quinone is free in solution.
Figure 7. Autoradiogram comparing the photocleavage of a 3′ 32P
labeled 248 base pair restriction fragment by intercalated anthraquino-
nes: lane 1, uncomplexed DNA dark control; lane 2, uncomplexed
DNA irradiated; lane 3, irradiated DNA with 10 µM 2AQC2(HEt2)
after piperdine treatment; lane 4, irradiated DNA with 10 µM 2AQA2-
(HEt2) after piperdine treatment; lane 5, lane 3 before piperdine
treatment; lane 6, lane 4 before piperdine treatment; lane 7, Maxam
Gilbert G sequencing lane.
When the concentration of 2AQC2(HEt2) is 10 µM and the
DNA is 50 µM in base pairs, calculation from the binding
constant indicates that every fifth intercalation site is occu-
piedand that the concentration of free quinone is less than 0.5
µM. Photolysis of the complex at this concentration yields the
usual pattern of GG-selective cleavage after treatment with
piperidine when analyzed by gel electrophoresis (lane 4, Figure
7). When the concentration of 2AQC2(HEt2) is increased to
50 µM, in principle, every intercalation site in the DNA will
be occupied. Under these conditions, with identical absorbed
light doses, the piperdine-induced cleavage efficiency is un-
changed and the pattern of GG-selective cleavage is maintained
(data not shown). When the concentration of 2AQC2(HEt2) is
increased further to 100 µM, there are considerably more
quinone molecules than there are intercalation binding sites and
the concentration of nonintercalated 2AQC2(HEt2) must be at
least 75 µM. At this high AQ-DNA ratio (lane 2, Figure 8),
there is a significant amount of spontaneous nonselective
cleavage of the DNA. On this basis, it appears that the
piperidine-requiring GG-selective cleavage of 2AQA2(HEt2) and
2AQC2(HEt2) is initiated by irradiation of quinone bound to
the DNA by intercalation. Since even at the lowest quinone
concentration examined intercalation must occur at sites other
than GG, these findings, too, indicate that the GG-selective
the DNA cleavage by 2AQA2(HEt2) and 2AQC2(HEt2) was
examined by polyacrylamide gel electrophoresis of a radiola-
beled restriction fragment consisting of 248 base pairs generated
by EcoRI/PVuII digestion of pA8G2 plasmid. The samples were
irradiated in a Rayonet photoreactor equipped with 350 nm
lamps. After irradiation was complete, the samples were
divided. One aliquot received no treatment and was analyzed
for spontaneous DNA backbone cleavage. The other portion
was treated with hot piperidine to visualize alkaline labile
damage. The autoradiograms recorded after photolysis of the
DNA-anthraquinone mixtures are shown in Figure 7.
Close inspection of Figure 7 reveals that irradiation of either
2AQA2(HEt2) or 2AQC2(HEt2) causes DNA cleavage only after
treatment with piperdine. Alkaline labile cleavage is typical
of oxidative damage to a nucleic acid base.18 Cleavage is
usually spontaneous when it is initiated by hydrogen atom
abstraction from the deoxyribose backbone.19 Further, it is clear
from analysis of Figure 7 that both 2AQA2(HEt2) and 2AQC2-
(HEt2) cause cleavage primarily at 5′-G of GG steps in the DNA
sequence.20 Similar DNA cleavage patterns have been reported
recently for riboflavin and other imides that were attributed to
an electron-transfer reaction that yields a GG radical cation.6
(21) While 1O2 damage has been recorded in many instances, it has
recently been described for an anthraquinone derivative: Reszka, J. K.;
Bilski, P.; Chignell, C. F.; Hartley, J. A.; Khan, N.; Souhami, R. L.;
Mendonca, A. J.; Lown, J. W. J. Photochem. Photobiol. B: Biol. 1992, 15,
317-35.
(22) Hertzberg, R. P.; Dervan, P. B. Biochemistry 1984, 23, 3934-3945.
(23) Merkel, P. B.; Kearns, D. R. J. Am. Chem. 1972, 94, 1029-1030.
(24) The D2O-H2O effect has been clearly demonstrated in the meth-
ylene blue sensitized DNA cleavage. Schneider, J. E., Jr.; Phillips, J. R.;
Pye, Q.; Maidt, M. L.; Price, S.; Floyd, R. A. Arch. Biochem. Biophys.
1993, 301, 91-97. Floyd, R. A.; West, M. S.; Eneff, K. L.; Schneider, J.
E. Arch. Biochem. Biophys. 1989, 273, 106-111.
(18) Modified bases like 8-oxoguanine are released from duplex DNA
only after piperdine treatment. Tchou, J.; Kasia, H.; Shibutani, S.; Chung,
M.-H.; Laval, J.; Grollman, A. P.; Hishimura, S. Proc. Natl. Acad. Sci.
U.S.A. 1991, 88, 4690-4694.
(19) Stubbe, J.; Kozarich, J. W. Chem. ReV. 1987, 87, 1107-1136.
(20) There are also some weaker signals in Figure 7 indicating cleavage
at the G of some 5′-GA-3′ steps.
(25) We have previously observed that 1O2 and O2- are not involved in
DNA cleavage with other anthraquinone derivatives (ref 5).

Anthraquinone Photonucleases
J. Am. Chem. Soc., Vol. 118, No. 10, 1996 2317
releasing and is expected to alter the nature of the lowest excited
state of the quinone so that it has predominant ππ* or charge-
transfer character.27 The spectroscopic and chemical evidence
described above clearly supports this view. Thus, reversing the
structure of the linking group has altered the electronic con-
figuration of the reacting excited state. Of particular relevance
to this work are the experiments that demonstrate that excited
2AQC2(HEt2) is reduced effectively by hydrogen atom abstrac-
tion from isopropyl alcohol and by electron transfer from
triethylamine, but that excited 2AQA2(HEt2) reacts only by
electron transfer from triethylamine. This difference in chemical
properties provides a solid basis for analyzing whether hydrogen
atom abstraction from deoxyribose or electron transfer from a
DNA base is the reaction that initiates GG strand cleavage.
Thermodynamic considerations play a key role in assessing
the likelihood that excited 2AQA2(HEt2) and 2AQC2(HEt2) will
react with DNA by electron transfer. The oxidation potentials
of the nucleic acid bases have been determined.28 Guanine has
the lowest oxidation potential and, as expected, the purines are
more easily oxidized than are the pyrimidines. The reduction
potentials of 2AQA2(HEt2) and 2AQC2(HEt2) are -0.68 and
-0.58 V (vs SCE), respectively. Substitution of these values
into the Weller equation29 (eq 2) reveals that electron transfer
from any of the DNA bases to the singlet or triplet excited state
of 2AQA2(HEt2) or 2AQC2(HEt2) is highly exothermic and
therefore feasible on thermodynamic grounds and likely to be
rapid.
∆Get ) -(ED/D- - EA/A+ + ∆E0,0)
(2)
The spectroscopic data indicate rapid reaction of the excited
anthraquinones with DNA. The emissions of both 2AQA2-
(HEt2) and 2AQC2(HEt2) are quenched when they are com-
plexed with DNA. Laser flash photolysis of 2AQA2(HEt2) or
2AQC2(HEt2) bound to DNA leads to the formation of their
radical anions in less than 20 ps, confirming the thermodynamic
prediction that electron transfer will be spontaneous. The
specific DNA base that serves as the source of the electron
cannot be identified from this experiment, but critically, the
excited state electron transfer pathway for both quinones is
firmly established.
(2) Ground State Quinone-DNA Complexes. In order to
assess the nuclease activities of 2AQA2(HEt2) and 2AQC2-
(HEt2), the nature of their association with DNA must be
understood. The equilibrium binding constants with CT-DNA
of the two quinones are nearly identical, suggesting that the
precise structure of the linking group does not affect the
thermodynamics of association. The spectral properties of the
bound quinones provide some additional structural details.
X-ray crystal structures of anthracycline drugs, which contain
an anthraquinone subunit, reveal that they bind to DNA by
intercalation.30 The bathochromic and hypochromic shifts of
the absorptions of bound 2AQA2(HEt2) and 2AQC2(HEt2) as
well as their induced CD spectra indicate that they are also
intercalated. The induced CD spectra of 2AQA2(HEt2) and
2AQC2(HEt2) have identical shapes but different intensities.
Figure 8. Autoradiogram comparing the photocleavage of a 3′ 32P
labeled 248 base pair restriction fragment by nonintercalated an-
thraquinones (total DNA concentration 50 µM base pairs): lane 1,
irradiated with 100 µM 2AQA2(HEt2) added; lane 2, 100 µM 2AQC2-
(HEt2) irradiated with 2AQC2(HEt2) added; lane 3, Maxam Gilbert G
sequencing lane.
cleavage cannot be due to GG-selective binding. The nonselec-
tive spontaneous cleavage observed at high 2AQC2(HEt2)
concentrations must originate with nonintercalated quinone.
Significantly, irradiation of high concentrations of 2AQA2-
(HEt2), where the concentration of nonintercalated quinone must
also be at least 75 µM, does not give any nonselective
spontaneous cleavage (lane 1, Figure 8).
Discussion
(1) Photochemistry and Photophysics of 2AQA2(HEt2)
and 2AQC2(HEt2) in Solution and with DNA. The an-
thraquinone photonucleases are built from three pieces: (1) the
anthraquinone chromophore, (2) a linker group, and (3) a
cationic ammonium side chain. The function of the chro-
mophore is to absorb light and initiate chemical transformations
of the DNA. The cationic side chain serves two purposes: to
increase the water solubility of the compound and to provide
increased binding to DNA by electrostatic attraction. The linker
group provides, at least, a convenient means of tethering the
chromophore and ammonium group.
The carboxamide group of 2AQC2(HEt2) not only links the
ammonium terminus with the anthraquinone ring but also
modifies the properties of the excited state quinone. As a
relatively electron deficient anthraquinone derivative, 2AQC2-
(HEt2) is expected to behave like typical reactive nπ* carbonyl
compounds.26 In contrast, as a derivative of 2-aminoan-
thraquinone, the linking group of 2AQA2(HEt2) is electron
(27) Allen, N. S.; McKellar, J. F. J. Photochem. 1976, 7, 107-111.
(28) Jovanovich, S. V.; Simic, M. G. J. Phys. Chem. 1986, 90, 974-
979.
(29) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259-263. The
exergonicity of electron transfer between the DNA bases and the excited
quinones can be calculated from the base oxidation potentials in ref 26, the
quinone redox potentials, and the 0,0 energies we report, ∆Get (2AQA2-
(HEt2)) ) -1.35, -1.23, -1.17, and -1.14 V for G, A, C, and T,
respectively.
(30) Quigley, G. J.; Wang, A. H.-J.; Ughetto, G.; van der Marcel, G.;
van Boom, J. H.; Rich, A. Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 7204-
7208.
(26) Peters, R. H.; Sumner, H. H. J. Chem. Soc. 1953, 2101-2110.

2318 J. Am. Chem. Soc., Vol. 118, No. 10, 1996
Breslin and Schuster
Since the relative intensities of these bands do not reflect
differences in binding affinity, this difference may indicate a
subtle change in intercalation site or a shift of the chromophore
relative to the center of the intercalation site.31 These differences
in orientation are unlikely to affect the rate of electron transfer
since the quinones are in contact with the bases and π-electron
overlap should not depend strongly on relative orientation. In
contrast, a small change in the structure of the intercalated
quinone may significantly influence the efficiency of hydrogen
atom abstraction since this process is strongly dependent on
the distance and orientation of the carbonyl oxygen atom and
the hydrogen atom.32
Kasai, Cadet, and co-workers33 have recently demonstrated that
guanine radical cations react with water, leading to damage that
is revealed by hot piperidine treatment.18 Migration of a base
radical cation away from the intercalation site of its generation
to a GG trap can proceed downhill thermodynamically for some
base pair sequences.34 This separation in space will slow the
rate of back electron transfer and provide time for GG radical
cations to react with water or O2, accounting for the observed
GG selectivity. Experiments are underway to test this hypoth-
esis.
Conclusions
(3) Selectivity and Efficiency of DNA Cleavage by 2AQA2-
(HEt2) and 2AQC2(HEt2). Comparison of the cleavage
patterns and efficiencies induced by irradiation of 2AQA2(HEt2)
and 2AQC2(HEt2) bound to the restriction fragment in combi-
nation with the spectroscopic and chemical analysis permits
assignment of a reaction mechanism. When the ratio of the
quinone to DNA base pairs is 1:5, both 2AQA2(HEt2) and
2AQC2(HEt2) cleave DNA with equal efficiency and show the
same GG selectivity. Since the chemical and spectroscopic
results indicate that only 2AQC2(HEt2) can abstract a hydrogen
atom, we conclude that the piperidine-activated GG-selective
cleavage is initiated by electron transfer to form a base radical
cation.
The experiments at high quinone concentration reveal a
different pattern of reactivity and selectivity. At high concen-
tration of the quinone, where some is not intercalated, irradiation
of 2AQC2(HEt2) gives spontaneous, nonselective cleavage, but
cleavage by 2AQA2(HEt2) remains GG-selective and still
requires piperidine activation. These findings indicate that the
GG-selective cleavage is initiated by either of the quinones when
intercalated, but only nonintercalated 2AQC2(HEt2) gives
spontaneous nonselective cleavage. This result is consistent with
the chemical and spectroscopic properties of the quinones which
indicate that nonintercalated excited 2AQC2(HEt2) is capable
of direct hydrogen atom abstraction from a deoxyribose unit, a
process that 2AQA2(HEt2) cannot undergo.
(4) GG-Selectivity by Hole Hopping. The equilibrium
association constant experiments and the observation that GG-
selective cleavage occurs even when essentially all intercalation
sites are occupied rules out selective binding as the cause of
the GG-selective cleavage. Electron transfer to an excited
intercalated quinone generates a base radical cation and the
quinone radical anion. Reaction of the base radical cation to
produce strand cleavage must compete with back electron
transfer, which is expected to be rapid when the radical cation
and radical anion are in direct contact. The time-resolved laser
spectroscopy reveals multiple decay rates for the quinone radical
anion. Some radical anions decay within a few picoseconds of
the laser pulse, others live more than 10 ns. The long-lived
radical anions cannot be in direct contact with a base radical
cation. Either the first-formed base radical cation has been
consumed by a chemical reaction or the radical cation has
migrated away from the intercalation site containing the radical
anion.
2AQC2(HEt2) has lowest energy nπ* excited states and is
capable of reaction by electron transfer or hydrogen atom
abstraction. 2AQA2(HEt2) has lowest energy charge-transfer
or ππ* states and can react only by electron transfer. Electron
transfer from either intercalated quinone forms a base radical
cation and leads, eventually, to GG-selective strand cleavage.
Radical cation migration through double-stranded DNA may
account for the GG-selectivity. Direct hydrogen atom abstrac-
tion by a nonintercalated excited 2AQC2(HEt2) leads to
spontaneous, nonselective cleavage of DNA.
Experimental Section
Materials. The syntheses of anthraquinone-2-carbonyl chloride and
2AQC4(H3) have already been reported,4 and the material obtained by
these procedures gave satisfactory NMR spectra and melting points.
Spectrograde ethylene glycol (Eastman Kodak) was used in the low-
temperature phosphorescence studies. The phosphate buffer used,
unless specified otherwise, is 10 mM phosphate with 100 mM NaCl at
pH ) 7.2 (denoted as PBS). Calf thymus DNA and superoxide
dismutase from bovine erthyocytes were purchased from Sigma Co.
PA8G2 plasmid was kindly supplied by Dr. Peter Nielsen of the
University of Copenhagen. Restriction enzymes EcoRI and PVuII and
DNA polymerase (Klenow fragment) were purchased from New
England Biolabs. [R-32P]dATP was purchased from Amersham. N,N-
Diethylenediamine, 2-aminoanthraquinone, 3-bromopropionyl chloride,
diethylamine, isopropyl alcohol, triethylamine, dimethylformamide, D2O
(99%), and potassium superoxide were purchased from Aldrich
Chemical Co. In all cases the chemicals were used without further
purification. Cellulose dialysis tubing (2000 MW cutoff) was manu-
factured by Spectrum Medical and was washed several times with water
prior to use.
Instrumentation. NMR spectra were recorded on a Varian Gemini-
300 spectrometer. CD spectra of the DNA-ligand complexes were
recorded on a Jasco J-270 instrument. Fluorescence and phosphores-
cence spectra were measured using a Spex-1681 Fluorlog 2. Phos-
phorescence spectra were measured in a 25% ethylene glycol-aqueous
phosphate buffer solvent cooled to a 77 °K glassy matrix. A Cary 1E
spectrometer was used to record UV-vis spectra. A Bioanalytical
Systems 100 potentiostat was used to record cyclic voltammetry traces.
Excited state lifetime measurements were performed using a Photon
Technology LS-1 time-resolved fluorimeter. Steady state irradiations
were conducted in a Rayonet RPR-100 photoreactor. The picosecond
transient spectroscopy laser apparatus has been described elsewhere.35
N-[[2-(N′,N′-Diethylamino)ethyl]amino]-2-anthraquinonecarbox-
amide Hydrochloride (2AQC2(HEt2)) was prepared by suspending
0.25 g of anthraquinone-2-carbonyl chloride in 20 mL of methylene
chloride and adding a solution of 0.11 g of N,N-(diethylamino)-
ethylamine in 10 mL of methylene chloride. The solution was stirred
for 2 h after the addition was complete. The product was isolated by
precipitation of the reaction mixture in 200 mL of diethyl ether.
An attractive explanation for the GG-selective cleavage is
radical cation migration in the oxidized DNA. Recent calcula-
tions indicate that GG steps have particularly low oxidation
potentials.6 The oxygen-dependent cleavage efficiency suggests
that the guanyl radical cation itself or a subsequently formed
chemical intermediate may be trapped by O2. Furthermore,
(33) Kasai, H.; Yamaizumi, Z.; Berger, M.; Cadet, J. J. Am. Chem. Soc.
1992, 114, 9692-9694.
(34) For example, if the quinone [AQ] is intercalated at the 5′ end of
the sequence 5′-[AQ]AGG-3′, the first-formed A•+ will migrate to GG,
leaving neutral A between the radical cation and radical anion.
(35) Zhu, Y.; Koefod, R. S.; Devadoss, C.; Shapley, J. R.; Schuster Inorg.
Chem. 1992, 31, 3505-3506.
(31) Lyng, R.; Rodger, A.; Norde´n, B. Biopolymers 1991, 31, 1709-
1720.
(32) Weisz, A.; Kaftory, M.; Vidavsky, I.; Mandelbaum, A. J. Chem.
Soc. 1984, 18-19.

Anthraquinone Photonucleases
J. Am. Chem. Soc., Vol. 118, No. 10, 1996 2319
Recrystallization from isopropyl alcohol gave 0.28 g (76% yield) of
light beige crystals: mp 181-182 °C; H NMR (DMSOd6) δ 1.23 (t,
anions were prepared by treating 0.2 mM anthraquinone in a N2-purged
DMF solution (5 mL) with 10 mg of KO2. After 5 min, the excess
KO2 was filtered away and the colored solution was transferred to a
sealed cuvette.
1
6 H, J ) 7.3 Hz), 3.17 (q, 4 H, J ) 7.3 Hz), 3.24 (t, 2 H, J ) 6.5 Hz),
3.69 (q, 2 H, J ) 6.5, 5.6 Hz), 7.94 (d, 1 H, J ) 3.3 Hz), 7.96 (d, 1 H,
J ) 3.3 Hz), 8.22 (t, 1 H, J ) 5.6 Hz), 8.23 (t, 1H, J ) 5.6 Hz), 8.28
(d, 1 H, J ) 8.1 Hz), 8.40 (dd, 1 H, J ) 1.7, 8.0 Hz), 8.66 (d, 1 H, J
) 1.7 Hz), 9.32 (t, 1 H, J ) 5.6 Hz); 13C NMR (DMSO-d6) δ 8.37,
34.53, 46.43, 49.58, 125.97, 127.11, 127.29, 127.38, 127.59, 133.17,
133.27, 133.30, 133.37, 134.73, 135.03, 139.02, 165.52, 182.45, 182.55;
HRMS 351.1709 (calcd), 351.1709 (found).
DNA Cleavage. A 248 base pair EcoR1/PVuII restriction fragment
was prepared from pA8G2, a derivative of pUC19, containing the
sequence 5′-GT4CT2CT2CTGCA-3′‚5′-GA2GA2GA4CTGCA-3′ cloned
into the PstI site. The restriction fragment was radiolabeled at the 3′
end with [R-32P]dATP using the Klenow fragment DNA polymerase.38
Solutions (10 µL) prepared for analysis of photoinduced cleavage had
the following concentrations of reagents: 10 mM sodium phosphate
(pH ) 7.0), 10 µM AQ, 50 µM calf thymus carrier DNA, 2500 cpm
[32P]DNA. The samples were irradiated for ca. 2 h at 350 nm in 1.5
mL plastic microcentrifuge tubes placed on a rotating carousel in a
Rayonet reactor equipped with four bulbs. Control samples were
subjected to the identical procedure but were wrapped in aluminum
foil throughout the irradiation period. The reaction products were
isolated by precipitation with cold ethanol. Inorganic salts were
removed by washing with 80% ethanol. The samples were divided:
some were treated with 1 M aqueous piperdine (30 min at 90 °C), and
the others were analyzed directly. Electrophoresis of the photolysis
mixture was conducted on an 8% denaturing polyacrylamide gel at 65
W constant power. DNA cleavage patterns on the radiograms were
visualized after exposure of photographic film to the dried gel for 1-2
days. A lane containing the Maxam-Gilbert A+G or G sequencing
reaction39 was run simultaneously to determine the precise cleavage
site.
Laser Flash Photolysis Studies of the Anthraquinone Derivative-
DNA Complexes. A solution containing 1.5 mM of the appropriate
anthraquinone and 7.5 mM sonicated self-thymus DNA in PBS was
excited with a 20 ps duration pulse of 355 nm light from a Nd-YAG
laser.9 Transient absorbance spectra were recorded at delays of 0 ps,
50 ps, 100 os, and 2 ns after the excitation laser pulse.
Electrochemical Measurements. A three-electrode cell configu-
ration, consisting of a platinum working electrode, a saturated calomel
reference electrode, and a platinum auxiliary electrode, was used. The
supporting electrolyte was 0.1 M aqueous sodium phosphate buffered
at pH ) 7 which had been N2 purged to remove molecular oxygen.
2-(3-Bromopropionamido)anthraquinone was prepared by reflux-
ing 1.0 g of 2-aminoanthraquinone in 5 mL of freshly distilled
3-bromopropionyl chloride for 2 h. After cooling, the product was
collected by filtration and washed with diethyl ether. Recrystallization
from ethanol gave 1.5 g (94% yield) of yellow crystals: mp 254 °C
1
(dec); H NMR (DMSO-d6) δ 3.05 (t, 2 H, J ) 6.3 Hz), 3.76 (t, 2 H,
J ) 6.3 Hz), 7.91 (d, 1 H, J ) 4.2 Hz), 7.93 (d, 1 H, J ) 2.2 Hz), 8.00
(dd, 1 H, J ) 2.2, 6.4 Hz), 8.14-8.20 (m, 3 H), 8.48 (d, 1 H, J ) 2.2
Hz), 10.79 (s, 1 H); 13C NMR (DMSO-d6) δ 28.71, 39.12, 116.14,
124.09, 127.02, 127.10, 128.40, 128.93, 133.46, 133.48, 134.56, 134.63,
134.68, 134.97, 144.73, 169.78, 181.91, 182.98; HRMS 357.0001
(calcd), 356.9999 (found).
2-[3-(N,N-Diethylamino)acetamido]anthraquinone (2AQA2(HEt2).
A solution of 4 mL of diethylamine and 2-(3-bromopropionamido)-
anthraquinone was heated at reflux for 1 h. The solvent was removed
under reduced pressure, and the remaining solid was dissolved in a
HCl-saturated ethanol solution and then precipitated with 300 mL of
diethyl ether. Recrystallization from isopropyl alcohol gave 0.7 g (65%
1
yield) of yellow crystals: mp 244-245 °C; H NMR (DMSO-d6) δ
1.24 (t, 6 H, J ) 7.2 Hz), 2.97 (t, 2 H, J ) 7.2 Hz), 3.15 (q, 4 H, J )
7.3 Hz), 3.39 (t, 2 H, J ) 7.3 Hz), 7.89 (d, 1 H, J ) 3.9 Hz), 7.92 (d,
1 H, J ) 3.9 Hz), 8.07 (dd, 1 H, J ) 2.1, 6.5 Hz), 8.16 (s, 1H), 8.17-
8.20 (m, 2 H), 8.50 (d, 1 H, J ) 2.1 Hz), 11.02 (s, 1H); 13C NMR
(DMSO-d6) δ 8.47, 30.81, 46.71, 46.73, 116.19, 124.15, 127.00, 127.07,
128.41, 128.86, 133.44, 133.47, 134.52, 134.63, 134.98, 144.70, 169.51,
181.89, 182.97; HRMS 351.1709 (calcd), 351.1717 (found).
Binding Constants. The affinities of the anthraquinone derivatives
for calf thymus DNA were evaluated by an equilibrium dialysis method.
A 100 µL aliquot of approximately 5 mM concentration in base pairs
was placed and sealed in a 5 cm section of dialysis tubing. The dialysis
bags were suspended in solutions containing the quinone at various
concentrations. Each dialysis sample was stirred constantly at room
temperature for 48 h. Preliminary experiments showed that a 24 h
period was sufficient to reach equilibrium. The concentration of free
2AQA2(HEt2) was determined directly by quantitative fluorescence
measurements. 2AQC2(HEt2) was reduced photochemically to the
hydroquinone with isopropyl alcohol under N2 prior to fluorescence
analysis. The concentrations of quinone were determined from
calibration curves. The binding isotherms were fit using the Scatchard
analysis36 and the Mc Ghee-Von Hippel equation.37
Acknowledgment. The authors gratefully acknowledge
funding of this work by the NIH. We are thankful for the advice
and assistance of Dr. Bruce Armitage of this department for
measurement of transient spectra and for designing and measur-
ing the binding constants of the oligonucleotides with AQC4-
(H3). We also thank Dr. Peter Nielsen of the University of
Copenhagen for generously supplying the plasmid used in the
radiolabeling experiments.
JA953714W
Chemical Reduction of the Anthraquinones. Hydroquinones were
prepared by reduction with hydrosulfite. The semiquinone radical
(38) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular Cloning.
Laboratory Manual, 2nd ed.; Cold Spring Harbor Laboratory Press: Cold
Spring Harbor, NY, 1989.
A
(36) Scatchard, G. Ann. N.Y. Acad. Sci. 1949, 51. 660-672.
(37) McGhee, J. D.; von Hippel, P. H. J. Mol. Biol. 1974, 86, 469-489.
(39) Maxam, A. W.; Gilbert, W. Methods Enzymol. 1980, 65, 499.