Optical, redox, and DNA-binding properties of phenanthridinium chromophores: Elucidating the role of the phenyl substituent for fluorescence enhancement of ethidium in the presence of DNA

Article abstract of DOI:10.1002/chem.200902823

The phenanthridinium chromophores 5-ethyl-6-phenylphenanthridinium (1), 5-ethyl-6-methylphenanthridinium (2), 3,8-diamino-5-ethyl- 6methylphenanthridinium (3), and 3,8diamino-5-ethyl-6-(4-N,N-diethylaminophenyl) phenanthridinium (4) were characterized by their optical and redox properties. All dyes were applied in titration experiments with a randomsequence 17mer DNA duplex and their binding affinities were determined. The results were compared to well-known ethidium bromide (E). In general, this set of data allows the influence of substituents in positions 3, 6, and 8 on the optical properties of E to be elucidated. Especially, compound 4 was used to compare the weak electron-donating character of the phenyl substituent at position 6 of E with the more electrondonating 4-N,N-diethylaminophenyl group. Analysis of all of the measurements revealed two pairs of chromophores. The first pair, consisting of 1 and 2, lacks the amino groups in positions 3 and 8, and, as a result, these dyes exhibit clearly altered optical and electrochemical properties compared with E. In the presence of DNA, a significant fluorescence quenching was observed. Their binding affinity to DNA is reduced by nearly one order of magnitude. The electronic effect of the phenyl group in position 6 on this type of dye is rather small. The properties of the second set, 3 and 4, are similar to E due to the presence of the two strongly electron-donating amino groups in positions 3 and 8. However, in contrast to 1 and 2, the electron-donating character of the substituent in position 6 of 3 and 4 is critical. The binding, as well as the fluorescence enhancement, is clearly related to the electron-donating effect of this substituent. Accordingly, compound 4 shows the strongest binding affinity and the strongest fluorescence enhancement. Quantum chemical calculations reveal a general mechanism related to the twisted intramolecular charge transfer (TICT) model, Accordingly, an increase of the twist angle between the phenyl ring in position 6 and the phenanthridinium core opens a nonradiative channel in the excited state that depends on the electron-donating character of the phenyl group. Access to this channel is hindered upon binding to DNA.

Full text of DOI:10.1002/chem.200902823

DOI: 10.1002/chem.200902823
Optical, Redox, and DNA-Binding Properties of Phenanthridinium
Chromophores: Elucidating the Role of the Phenyl Substituent for
Fluorescence Enhancement of Ethidium in the Presence of DNA
Christa Prunkl,[a] Markus Pichlmaier,[b] Rainer Winter,[b] Vladimir Kharlanov,[c]
Wolfgang Rettig,*[c] and Hans-Achim Wagenknecht*[a]
Abstract: The phenanthridinium chro-
mophores 5-ethyl-6-phenylphenanthri-
dinium (1), 5-ethyl-6-methylphenan-
thridinium (2), 3,8-diamino-5-ethyl-6-
methylphenanthridinium (3), and 3,8-
diamino-5-ethyl-6-(4-N,N-diethylami-
nophenyl)phenanthridinium (4) were
characterized by their optical and
redox properties. All dyes were applied
in titration experiments with a random-
sequence 17mer DNA duplex and their
binding affinities were determined. The
results were compared to well-known
ethidium bromide (E). In general, this
set of data allows the influence of sub-
stituents in positions 3, 6, and 8 on the
optical properties of E to be elucidat-
ed. Especially, compound 4 was used to
compare the weak electron-donating
character of the phenyl substituent at
position 6 of E with the more electron-
group. Analysis of all of the measure-
ments revealed two pairs of chromo-
phores. The first pair, consisting of 1
and 2, lacks the amino groups in posi-
tions 3 and 8, and, as a result, these
dyes exhibit clearly altered optical and
electrochemical properties compared
with E. In the presence of DNA, a sig-
nificant fluorescence quenching was
observed. Their binding affinity to
DNA is reduced by nearly one order of
magnitude. The electronic effect of the
phenyl group in position 6 on this type
of dye is rather small. The properties
of the second set, 3 and 4, are similar
to E due to the presence of the two
groups in positions 3 and 8. However,
in contrast to 1 and 2, the electron-do-
nating character of the substituent in
position 6 of 3 and 4 is critical. The
binding, as well as the fluorescence en-
hancement, is clearly related to the
electron-donating effect of this sub-
stituent. Accordingly, compound
4
shows the strongest binding affinity
and the strongest fluorescence en-
hancement. Quantum chemical calcula-
tions reveal a general mechanism relat-
ed to the twisted intramolecular charge
transfer (TICT) model. Accordingly, an
increase of the twist angle between the
phenyl ring in position 6 and the phe-
nanthridinium core opens a nonradia-
tive channel in the excited state that
depends on the electron-donating char-
acter of the phenyl group. Access to
this channel is hindered upon binding
to DNA.
strongly
electron-donating
amino
Keywords: charge transfer · DNA ·
dyes/pigments · fluorescent probes ·
intercalations
donating
4-N,N-diethylaminophenyl
[a] Dipl.-Chem. C. Prunkl, Prof. Dr. H.-A. Wagenknecht
Institute for Organic Chemistry
[c] Dr. V. Kharlanov, Prof. Dr. W. Rettig
Institute of Chemistry
University of Regensburg
Universitꢀtstraße 31
Humboldt University Berlin
Brook-Taylor-Str. 3
93053 Regensburg (Germany)
Fax : (+49)941-943-4617
E-mail: achim.wagenknecht@chemie.uni-regensburg.de
12489 Berlin (Germany)
Fax : (+49)30-2093-5574
E-mail: rettig@chemie.hu-berlin.de
[b] Dipl.-Chem. M. Pichlmaier, Prof. Dr. R. Winter
Institute for Inorganic Chemistry
University of Regensburg
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.200902823.
Universitꢀtstraße 31
93053 Regensburg (Germany)
3392
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 3392 – 3402

FULL PAPER
Introduction
incorporation of the phenanthridinium heterocycle of E as
an artificial DNA base through DNA building-block chemis-
try,[18] and applied these conjugates in charge-transfer ex-
periments[17] and for the investigation of DNA–protein com-
plexes.[19]
Planar polycyclic aromatic heterocycles ability to intercalate
between two adjacent base pairs in duplex DNA was first
proposed by Lerman.[1] Among such intercalators, charged
heterocycles are the most potent with respect to noncova-
lent stacking interactions. It is believed that the electrostatic
energy plays an important role, not only for binding to nu-
cleic acids, but also for the intercalation process.[2,3] Accord-
ingly, quinolizinium and acridizinium dyes have been devel-
oped as powerful fluorescent probes that light up upon in-
tercalative binding to duplex DNA.[4,5] 3,8-Diamino-5-ethyl-
6-phenylphenanthridinium (Ethidium, E) represents the
most prominent example of these positively charged hetero-
cyclic intercalators for DNA (Scheme 1). It is widely used as
a fluorescent staining agent for nucleic acids due to the sig-
nificant emission increase upon binding to DNA or RNA, as
intense research in the 1960ꢂs and 1970ꢂs revealed.[6] The
phenanthridinium compounds, including E, had their origin
in the search for antitumor,[7] as well as antiviral drugs,[8]
and most importantly for trypanocidal drugs.[9,10] However,
the application in human therapy is not reasonable due to
the significant mutagenic and carcinogenic properties of
E.[11] Thus, research has been focused on the development
of phenanthridinium intercalators for fluorescent DNA ana-
lytics. Mono- and bifunctional derivatives have been synthe-
sized to modulate the binding properties of E.[12] The photo-
chemical reactivity of E[13] has been applied to study photo-
induced energy-transfer[14] and charge-transfer (CT) process-
es in DNA.[15–17] We developed a protocol for the synthetic
Understanding the unique intercalation and emission
properties of E has been the aim of a number of intense
studies.[10,20–24] This is of high importance for the develop-
ment of new and improved fluorescent intercalators for mo-
lecular diagnostics with DNA and RNA. It has been pro-
posed that enhancement of the emission of E is due to a
proton transfer to the excited state of E, which is prevented
upon intercalation into duplex DNA.[20] It became clear that
the amino groups in positions 3 and 8 of the phenanthridini-
um heterocycle of E are involved in this proton transfer.
However, there are some serious doubts about this early
proposal. Especially with respect to the current model of
the intercalated structure of E, it is not clear why the two
amino groups, which point out of the DNA duplex in the di-
rection of the water layer, should be shielded from deproto-
nation. Early studies have shown that the exocyclic primary
amino groups of E are not mandatory for the intercalation
reaction, but that their presence adds stability to the com-
plex with a nucleic acid.[21] NMR spectroscopy studies by
Leupin et al. using 3- and 8-deaminoethidium revealed that
the removal of the 8-amino group had little effect on the
binding to DNA, whereas the removal of the 3-amino group
greatly affected the intercalation.[25] The most recent study
by Tor and co-workers showed the influence of such modifi-
cations on the electronic structure of E.[26] This study includ-
ed 5-ethyl-6-phenylphenanthridinium (3,8-dideaminoethidi-
um, 1) that was also previously described by others.[24,27–29]
On the other hand, several theoretical chemical calcula-
tions have shown that the quaternization at position 5 adds
significantly to the intercalation properties of phenanthri-
dines.[22–24] However, variation of the quarternizing group at
position 5 has little or no effect on the intercalation proper-
ties.[10,21] The stabilization energy of the cationic intercala-
tors is considerably larger than those of the uncharged
equivalents. The electron-poor phenanthridinium core stacks
with the relatively electron-rich DNA bases and interaction
is dominated by dispersion energy. The CT character seems
to be important.[22] Our studies with phenanthridinium incor-
porated as an artificial DNA base into oligonucleotides
have also shown that the emission of this chromophore is
slightly diminished in a G–C environment compared with a
T–A environment, which might indicate a CT interaction,
especially with the electron-rich guanines inside DNA.[18]
According to the mentioned studies,[10,22–24] it looked rea-
sonable to synthesize phenanthridinium chromophores as
derivatives of E to experimentally elucidate the role of the
substituents with respect to the photophysical and DNA-
binding properties, as well as the concomitant fluorescence
enhancement. Early and basic steps in the synthesis of phe-
nanthridinium derivatives were made by Walls and
Morgan,[30] Watkins et al.,[27,31] and Berg.[32] A variety of
chemical modifications have been introduced synthetically
Scheme 1. Ethidium (E), the phenanthridinium chromophores 1–4, and
the DNA sequence that were used in this study.
Chem. Eur. J. 2010, 16, 3392 – 3402
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3393

H.-A. Wagenknecht, W. Rettig et al.
to the amino functions at positions 3 and 8.[24,26–29,33,34]
Herein, we present the characterization of 1, 5-ethyl-6-meth-
ylphenanthridinium (2), 3,8-diamino-5-ethyl-6-methylphe-
nanthridinium (3), and 3,8-diamino-5-ethyl-6-(4-N,N-diethyl-
aminophenyl)phenanthridinium (4) with respect to their op-
tical and redox properties. Titration experiments with these
intercalators and a 17mer DNA duplex (Scheme 1) allow
the effects of the substituents in positions 3 and 8, but espe-
cially in position 6 of the core heterocycle on the DNA-
binding properties to be elucidated. Phenanthridinium 3 is
structurally similar to desphenyl dimidium, which was pub-
lished by Waring and Wakelin.[10] Derivative 4 was prepared
to compare the weak electron-donating character of the
phenyl substituent at position 6 of E with the stronger elec-
tron-donating 4-N,N-diethylaminophenyl group, which is
similar to the published ethyl phenidium, but with the 3-
amino group.[10] Additionally, quantum chemical calculations
were applied to investigate the effect of conformational re-
laxation on the efficiency of fluorescence enhancement
upon intercalation into DNA. Accordingly, the difference in
the fluorescence enhancement that was observed for E and
4 can be rationalized by a twisted intramolecular charge-
transfer (TICT)-like model.
mentally determined extinction coefficients for chromo-
phore 1 (Table 1) are in good agreement with the published
values.[34] In contrast, the phenanthridinium dyes 3 and 4
Table 1. Characterization of the chromophores E and 1–4 by absorption
and fluorescence spectroscopy and summary of the titration experiments
with DNA; sodium phosphate buffer (10 mm, pH 7.0, T=208C) was used
for all measurements.
Absorption
Without DNA
With DNA[a]
lmax lmax Dlmax lmax
[nm] [mÀ1 cmÀ1
Fluorescence
Without DNA
With DNA[a]
lmax Dlmax
[nm] [nm]
e
FF
]
[nm] [nm]
[nm]
615
[b]
E[c] 285
479
1[d] 251
322
51950
5420
44250
7640
4880
42460
6640
3830
50830
4640
63310
8180
–
0.01[e]
601
14
522
43
[b]
–
420
407
0.24[f]
420
0
325
378
3
1
377
[b]
2
250
315
369
–
0.35[f]
0.02[e]
407
583
0
320
369
5
0
[b]
3[g] 282
463
–
599
616
16
15
503
40
[b]
4[h] 281
486
–
(0.0001)[i] 601
ACHTUNGTRENNUNG
528
42
[a] 213.6 mm DNA-bp for absorption, 109.3 mm DNA-bp for fluorescence.
[b] Not determined due to DNA absorption. [c] Published values:[33] 285
(56800), 478 (5680); and 480 nm (6300mÀ1 cmÀ1).[10] [d] Published
values:[29] 251 (43040), 321 (7700), 377 nm (4770mÀ1 cmÀ1). [e] Fluores-
cence quantum yield determined with rhodamine 101 as a reference[36]
(FF =1.0). [f] Fluorescence quantum yield determined with quinine sul-
fate as a reference[36] (FF =0.546). [g] Compared with published values
for desphenyl dimidium:[10] 470 nm (4220mÀ1 cmÀ1). [h] Compared with
published values for ethyl phenidium:[10] 430 nm (5.380mÀ1 cmÀ1). [i] Fluo-
rescence quantum yield determined relative to E, but given relative to
rhodamine 101.
Results
Optical properties of chromophores 1–4: The phenanthridi-
nium chromophores 1–4 were characterized by UV/Vis ab-
sorption and fluorescence spectroscopy and the spectra were
compared with those of the commercially available and
well-described E. With respect to the subsequent titration
experiments with DNA, the optical characterization was
performed in aqueous buffer solution (10 mm sodium phos-
phate, pH 7.0). The absorption spectra revealed two distinct
classes of chromophores (Figure 1).
showed absorption maxima at 282/463 nm (3) and 281/
486 nm (4) that are similar to E (285/479 nm). Moreover,
chromophores 3 and 4 display absorption bands in the range
460–490 nm due to the CT character, similar to E. The ex-
tinction coefficient for chromophore
3
at 463 nm
(4640mÀ1 cmÀ1) is clearly lower than the corresponding value
of chromophore 4 at 486 nm (8180mÀ1 cmÀ1) and E at
479 nm (5420mÀ1 cmÀ1).
The steady-state fluorescence spectra also reveal two dif-
ferent classes of chromophores. Dyes 1 and 2 exhibit emis-
sion spectra with maxima at 420 and 407 nm, respectively
(Figure 2). For both compounds, Stokes shifts of about
40 nm are observed. In contrast, the fluorescence of chro-
mophores 3 and 4 show larger bathochromic shifts (in the
range of 130 nm), which are characteristic for E. The emis-
sion maxima were observed at 599 nm (3), and 616 nm (4).
Due to the optical differences between chromophores 1
and 2, and dyes E, 3, and 4, different standards for the de-
termination of fluorescence quantum yields had to be used
(see the Supporting Information).[35,36] Chromophores 1 and
2 are both highly fluorescent in aqueous buffer solution,
with quantum yields of 0.24 (1) and 0.35 (2). Compounds 3
and 4 exhibit strong fluorescence quenching; the quantum
yields are 0.01 (E) and 0.02 (3). The fluorescence quantum
Figure 1. UV/Vis extinction spectra of the chromophores 1–4 in compari-
son with E (c=23 mm in sodium phosphate buffer (10 mm, pH 7.0, T=
208C)).
In comparison with E, the phenanthridinium dyes 1 and 2
exhibit altered absorption maxima at 251/322/377 nm (1)
and 250/315/369 nm (2). It is noteworthy that our experi-
3394
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 3392 – 3402

Phenanthridinium Chromophores
FULL PAPER
tials of both chromophores were calculated to be approxi-
mately 0.6 eV. This value is slightly lower than that of E
(0.7 eV), which means that the dyes 3 and 4 are less easily
reduced than E.
Titration experiments of chromophores 1–4 with DNA: To
compare changes in the optical properties of chromophores
1–4 with E upon binding to DNA, we performed titration
experiments with a representative synthetic 17mer DNA
duplex with a random sequence (Scheme 1). To avoid dilu-
tion of the analyte solutions, the titrant solutions contained
DNA (for absorption: 357 mm DNA-bp, for fluorescence:
192 mm DNA-bp) as well as the corresponding chromophore
(for absorption: 23 mm, for fluorescence: 2.4 mm) in the same
concentration as in the titrated solution. Aliquots (500 mL)
of the chromophore solutions in Na-Pi-buffer were titrated
in 5, 10, 20, 50, and 100 mL steps. Measurements exclude the
region of the DNA absorption below 300 nm. Since the in-
teraction with DNA also leads to significant changes in the
absorption spectra of the chromophores, the excitation
wavelengths for the fluorescence titrations were set to the
corresponding isosbestic points, which were determined for
each chromophore by spectrophotometric titrations.
Figure 2. Normalized UV/Vis absorption (gray lines) and fluorescence
(black lines) spectra of chromophores 1–4 in comparison with E. Dashed
lines: normalized absorption (c=23 mm); solid lines: normalized fluores-
cence (c=2.4 mm), in sodium phosphate buffer (10 mm, pH 7.0, T=
208C).
yield of 4 was too low to be determined with rhodamine 101
as the reference. To obtain a rough estimation, the quantum
yield was determined relative to E and calculated to be
about 100 times lower in aqueous buffer solution.
The following titration with E serves as a reference: The
UV/Vis spectrum of E, in the absence of DNA, showed an
absorption with a maximum at 479 nm, which is typical for
“free” E in water (Figure S1 in the Supporting Informa-
tion).[38] With increasing amounts of DNA, the maximum
shifts to 522 nm, which is characteristic for intercalated
E.[22,39] The isosbestic point at 510 nm shows the transition
between the unbound and the DNA-bound form of the dye.
Upon excitation at 510 nm, the corresponding fluorescence
of E in water (Figure S3 in the Supporting Information) is
weak and has a maximum at 615 nm. With increasing DNA
concentration, the emission increases significantly and the
maximum shifts to 601 nm.
Subsequent titrations with 1–4 show two pairs of chromo-
phores, as already pointed out in the previous sections with
respect to the optical and the redox properties. The absorp-
tion spectra of 1 (maxima at 322, 364 and 377 nm; Figure 3)
and 2 (maxima at 315, 356 and 369 nm; Figure S2 in the
Supporting Information) show a decrease in the extinction
with DNA, accompanied by a rather small shift (3–5 nm)
and a broadening of the signals. When excited at 388 (1) or
378 nm (2), the fluorescence maxima can be observed at 420
or 407 nm, respectively. The fluorescence of both chromo-
phores is quenched significantly in the presence of DNA (1:
Figure 4, 2: Figure S4 in the Supporting Information), but
the quenching is not accompanied by a shift of the corre-
sponding maxima.
Redox characterization: To characterize chromophores 1–4
electrochemically, we measured the reduction potential by
cyclic voltammetry. It is important to note that the reduction
potentials of compounds 2 and 3 were obtained irreversibly.
Chromophores 1 and 2 show reduction potentials of À1.34
and À1.36 V, respectively (Table 2). When compared with E
(À1.53 V), these potentials reflect the absence of the elec-
Table 2. Electrochemical characterization of E and 1–4.
1
Reduction
E = [V][a]
E00 [eV]
E* [eV][b]
2
+
C
E /E
À1.53
2.2
3.2
3.2
2.3
2.2
0.7
1.8
1.9
0.6
0.6
+
C
C
C
C
1 /1
À1.34
+
2 /2
À1.36[c]
À1.67[c]
À1.62
+
3 /3
+
4 /4
[a] Measured by cyclic voltammetry in MeCN (50 mm NBu4PF6 as the
supporting electrolyte) versus Fc+/Fc as internal standard (v=
ACTHNUTRGNEUNG
200 mVsÀ1). [b] Excited-state potentials, calculated according to E*(X+/
+
[37]
X·)=E00 +E
= ACHTUNGTRENNUNG(X /X ). [c] Irreversible potentials.
C
2
1
tron-donating amino substituents, making 1 and 2 more
readily reducible at less negative potentials. By using an E00
value of 3.2 eV for both chromophores, the reduction poten-
tials of the excited state can be estimated as 1.8 eV (1) and
1.9 eV (2).[37] Hence, both chromophores exhibit excited-
state redox potentials that are significantly larger than that
of E.
In contrast, the phenanthridinium chromophores 3 and 4
showed reduction potentials of À1.67 and À1.62 V, respec-
tively, which are more negative than the corresponding
value for E. When the E00 values of 2.3 eV (3) and 2.2 eV
(4) were included in the calculation, the excited-state poten-
In contrast to 1 and 2, chromophores 3 and 4 exhibit an
E-like behavior in the titration experiments with DNA. The
absorption maximum of phenanthridinium 3 shifts from
463 nm in the unbound form to 503 nm in the bound form
(Figure 3); the absorption maximum of 4 shifts from 486 to
528 nm (Figure S2 in the Supporting Information). The iso-
sbestic points at 492 or 510 nm, respectively, support the
Chem. Eur. J. 2010, 16, 3392 – 3402
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3395

H.-A. Wagenknecht, W. Rettig et al.
both cases, the emission intensity increases significantly
when the chromophores are excited at 492 (3; Figure 4) or
510 nm (4; Figure S4 in the Supporting Information). This
fluorescence enhancement is accompanied by a shift of the
corresponding maximum, in the case of 3 from 599 nm with-
out DNA to 583 nm with DNA, in the case of 4 from 616 to
601 nm.
In addition, an analysis of the spectrophotometric titra-
tions according to McGhee and von Hippel[40] (Figure S5 in
the Supporting Information) was used to determine the as-
sociation constants (K) between the chromophores 1–4 and
DNA, as well as the binding-site sizes (n; Table 3). For the
Table 3. Binding constants and binding-site sizes of the chromophores E
and 1–4 as determined from spectrophotometric titrations with DNA in
sodium phosphate buffer (10 mm, pH 7.0, T=208C).
[a]
Compound
Kꢃ105 [mÀ1
5.0Æ0.3[c,d]
]
n[b]
E
1
2
3
4
1.6Æ0.1[c]
2.8Æ0.2
0.46Æ0.05[d]
0.67Æ0.05
3.1Æ0.4[e]
7.8Æ1.6[f]
2.7Æ0.1
1.5Æ0.1[e]
Figure 3. Representative titration experiments of chromophores 1 and 3
with DNA followed by UV/Vis absorption spectroscopy. Upper graph:
1.9Æ0.1[f]
spectrophotometric titration of double-stranded (ds)-DNA to
1 (c=
[a] K is the binding constant (in bp). [b] n is the binding-site size (in bp).
[c] Compared with published values for E at higher salt concentration
23 mm) in sodium phosphate buffer (10 mm, pH 7.0, T=208C); lower
graph: spectrophotometric titration of ds-DNA to 3 (c=23 mm) in sodium
phosphate buffer (10 mm, pH 7.0, T=208C). Arrows indicate changes in
the intensity of the bands upon increasing DNA concentrations.
(1m NaCl):[10] K=1.99ꢃ104 mÀ1
,
n=1.90. [d] Similar to published
values.[43] [e] Compared with published values for desphenyl dimidium at
higher salt concentration (1m NaCl):[10] K=1.13ꢃ104 mÀ1
n=1.73.
,
[f] Compared with published values for ethyl phenidium at higher salt
concentration (1m NaCl):[10] K=9.63ꢃ103 mÀ1, n=1.87.
titration of reference compound E with DNA, a binding
constant of 5.0ꢃ105 mÀ1 and a binding-site size of 1.6 (in bp)
was observed. Both values are in good agreement with the
results published by Zimmermann and Pauluhn[41] and Hor-
owitz and Hud,[42] but are higher than the values reported
by Waring and Wakelin,[10] probably due to a lower salt con-
centration in our experiments. The binding constants for
chromophores 1–4 again indicate two groups of chromo-
phores. For chromophores 3 and 4, the binding constants are
in the 105 mÀ1 range (3.1ꢃ105 mÀ1 for 3, 7.8ꢃ105 mÀ1 for 4).
Remarkably, chromophore 4 exhibits a stronger binding to
DNA than the reference E. The values found for chromo-
phores 1 and 2 (5ꢃ104 and 7ꢃ104 mÀ1), however, are about
one order of magnitude lower than that of E. Binding-site
sizes were found to be in the range of 1.5–2.8 (in bp) for all
chromophores, which is consistent with an intercalative
binding mode according to the neighbor exclusion model.
Figure 4. Representative titration experiments of chromophores 1 and 3
with DNA followed by fluorescence spectroscopy. Upper graph: fluori-
metric titration of ds-DNA against 1 (c=2.4 mm) in sodium phosphate
buffer (10 mm, pH 7.0, T=208C); lower graph: fluorimetric titration of
ds-DNA against
3 (c=2.4 mm) in sodium phosphate buffer (10 mm,
Discussion
pH 7.0, T=208C). Arrows indicate changes in the intensity of the band
upon increasing DNA concentrations.
The optical, redox, and DNA-binding properties of the phe-
nanthridinium dyes 1–4 reveal the electron-donating effects
of the substituents in positions 3, 8, and especially 6, on the
electron-poor phenanthridinium core system. It became evi-
dent from our optical and electrochemical characterization,
as well as from the titration experiments with DNA, that
idea of a single transition from unbound molecules to DNA-
bound species.
The fluorescence spectra of the phenanthridiniums 3 and
4, in the presence of DNA, also show E-like profiles. In
3396
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 3392 – 3402

Phenanthridinium Chromophores
FULL PAPER
the dyes can be divided into two pairs. The first pair (1 and
2) lacks the two exocyclic amino groups in positions 3 and 8,
chromophore 2 additionally lacks the phenyl group in posi-
tion 6. Compared with E, both chromophores have signifi-
cantly altered optical, electrochemical, and DNA-binding
properties. As a result of the lack of CT contribution by the
amino groups, both the absorption and the emission of 1
and 2 occur at shorter wavelengths. The emission intensities
in aqueous buffer solutions are higher (FF =0.24 for 1 and
FF =0.35 for 2) compared with E (FF =0.01) and the
Stokesꢂ shifts are small (ca. 40 nm). It is important to point
out that the minor difference of the optical properties be-
tween both chromophores can be attributed to the rather
weak electron-donating character of the phenyl group in po-
sition 6 of 1, which is absent in 2. The observation that these
differences between 1 and 2 are rather small indicates that
electron donation by the phenyl group in position 6 is signif-
icantly weaker than the two exocyclic amines in positions 3
and 8.
In the DNA titration experiments with dyes 1 and 2, a
strong fluorescence quenching was observed that was nearly
identical for both chromophores (Figure 5, top). The electro-
chemical potentials of both chromophores give a possible
explanation for this observation. As mentioned in the previ-
ous section, the excited-state potentials can be estimated to
have values of approximately 1.8 (1) and 1.9 eV (2). The po-
tential for the oxidation of guanine (G+·/G) is 1.3 V versus
NHE.[43] By using a correction value of À0.63 V from the
NHE to Fc/Fc+ (Fc=ferrocene) reference,[44] the corre-
sponding guanine potential is approximately 0.7 V versus
Fc/Fc+. This means that both phenanthridinium dyes, 1 and
2, are potentially able to photooxidize guanines in the
DNA. Such photoinduced processes between organic dyes
and the bases are often observed[45] and have been widely
used to study CT in DNA.[46]
In contrast to the first pair, the second pair of chromo-
phores (3 and 4) exhibits similarities with E due to the pres-
ence of the two amino groups in positions 3 and 8, yielding
more electron-rich compounds. It is important to emphasize
that the observed differences between 3, 4, and E can be as-
signed solely to the different substituents at position 6. Dye
3 lacks the weak electron-donating phenyl group completely,
whereas dye 4 has the significantly stronger electron-donat-
ing 4-N,N-diethylaminophenyl group. The increasing elec-
tron-donating character of the substituents in position 6
orders the chromophores by their optical properties. The ab-
sorption maxima shifts from 463 (3) through 479 (E) to
486 nm (4), the emission maxima shifts from 599 (3),
through 615 (E) to 616 nm (4), and the fluorescence quan-
tum yield decreases from 0.02 for 3 to 0.01 for E; the fluo-
rescence intensity of chromophore 4 is about 100 times
lower than for E (in fact, FF of 4 was too low to be deter-
mined by standard methods). Clearly, the CT contribution
in these dyes increases with the electron-donating character
of the substituent in position 6. This will be further discussed
based on quantum chemical calculations in the next section.
In contrast to the optical properties, the redox potentials
of chromophores 3 and 4 do not follow such a clear trend.
Both potentials (À1.67 V for 3; À1.62 V for 4) were more
negative than that of E. It is generally assumed that the pho-
toexcited state of E cannot oxidize guanines[47] and, based
on the calculated excited-state potential of 0.6 eV as men-
tioned in the previous section, neither compound 3 nor 4
should be able to photooxidize guanines in DNA.
Luedtke et al. recently reported that the aromatic nitro-
gen and some carbon atoms of E have surprisingly high
electron densities.[26] This means that electron donation by
the two exocyclic amines of E has a significantly larger
effect compared to the electron-withdrawing influence of
the endocyclic iminium core. The results of our experimen-
tal study on the substituent effects in chromophores 1–4 pre-
sented herein, strongly support this interpretation. Com-
pared to chromophores 1 and 2, the most notable differen-
ces between the chromophores 3, E, and 4, are the signifi-
cant bathochromic shift of the absorption spectra and the re-
markable increase in the fluorescence intensity that is
observed during the titration experiments with DNA
(Figure 5, bottom). When excited at the corresponding iso-
sbestic point, as revealed by the spectrophotometric titra-
tions, the dye 3 exhibits an approximately 7-fold increase in
fluorescence intensity, which is lower than the 10-fold in-
crease observed with E; in contrast, dye 4 shows a 21-fold
increase. There is clearly a relationship between the sub-
stituent in position 6 of the phenanthridinium core and both
the quantum yield in buffer (without DNA) and the fluores-
cence enhancement upon addition of duplex DNA. With in-
creasing electron-donating character of the substituent in
position 6, the fluorescence enhancement is more strongly
pronounced in the presence of DNA, whereas the basic fluo-
rescence without DNA is quenched more strongly. This can
Figure 5. Titration curves for the spectrofluorimetric titrations of ds-
DNA to E and chromophores 1–4; c=2.4 mm in sodium phosphate buffer
(10 mm, pH 7.0, T=208C).
Chem. Eur. J. 2010, 16, 3392 – 3402
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3397

H.-A. Wagenknecht, W. Rettig et al.
be explained by a TICT-like model as described in the next
section. According to this model, the nonradiative channel
in the excited-state, which leads to the small fluorescence
quantum yield in solution without DNA, is more pro-
nounced with a stronger electron-donating substituent in po-
sition 6. Access to this channel involves a twist of the phenyl
ring against the plane of the phenanthridinium core. The
necessary angular relaxation occurs in solution without
DNA, but is hindered by intercalation in the presence of
DNA, leading to the fluorescence enhancement.
twisted is a sufficiently good electron donor. This has been
highlighted in a rhodamine derivative with a flexible dialkyl-
anilino group, with the expected fluorescence quenching oc-
curring with high efficiency. If the dialkylanilino group is
protonated and transformed into a weak acceptor, the
quenching channel is blocked, and the quantum yield in-
creases by a factor of 290.[53]
This mechanism is closely related to the TICT theory of
excited states in which CT and twisting are combined and
the maximum CT is observed for the perpendicular confor-
mation. In some cases, dual fluorescence can arise, whereas,
in other cases, only the fluorescence quenching of the pre-
cursor state is observed; this could be induced by the relaxa-
tion of the pending donor group towards the perpendicular
conformation.[51,52] Applying this mechanism to the group of
phenanthridinium compounds investigated herein, we can
predict that if the phenyl group in E is exchanged for a
better donor, such as a dialkylanilino (D) group, a low-lying
excited CT state may also come into energetic reach and,
possibly, enhance the fluorescence quenching observed for
E in fluid solutions. Just as in the case of the di- or triphe-
nylmethane dyes (e.g., MG and CV), inhibition of the twist-
ing relaxation of the D group slows down the nonradiative
channel. A simple explanation for the very strong fluores-
cence enhancement observed for 4 in DNA is therefore that
this CT mechanism is partially involved, and that the relaxa-
tion is (partially or completely) blocked in the DNA envi-
ronment.
Additionally, the binding constants that were determined
for chromophores 1–4 with DNA reflect the strong influence
of the electron-donating exocyclic amino groups. Similar to
E, the binding constants of 3 and 4 are in the 105 mÀ1 range
(3.1–7.8ꢃ105 mÀ1), whereas the values for 1 and 2 are re-
duced by nearly one order of magnitude (5–7ꢃ104 mÀ1).
Moreover, parallel to the reduction of the binding affinity of
1 and 2, the binding-site sizes are significantly enhanced
(2.7–2.8 base pairs) compared with those of E, 3, and 4 (1.5–
1.9 base pairs). The binding constants for E, 3, and 4 are in
a similar range to those observed for phenanthridinium–nu-
cleotide conjugates (logK=5.5–6.0).[48] Furthermore, the
binding affinity of chromophores 3, E, and 4 increases con-
comitantly with the enhanced electron-donating character of
the phenyl substituent in position 6. Accordingly, dye 4
shows the strongest binding affinity. Hence, the binding af-
finity, as well as the fluorescence enhancement, is clearly re-
lated to the overall supply of electron density in the phenan-
thridinium heterocycle by the corresponding electron-donat-
ing substituents, not only in positions 3 and 8. It is remark-
able that, according to our studies, the electron-donating
character of the phenyl ring in position 6 of the phenanthri-
dinium core is also crucial, especially in the presence of
DNA. This observation is supported by quantum chemical
calculations.
To test this hypothesis, we performed quantum chemical
calculations on model structures of E, 3, and 4, as well as on
an additional derivative 5 (Scheme 2), which is predicted
from the model to show a stronger enhancement of this CT
quenching mechanism. In the first step, we used semiempiri-
cal AM1/ZINDO calculations to compare the four com-
Quantum chemical calculations: The results showing fluo-
rescence enhancement upon intercalation with DNA in
Figure 5 indicate that the efficiency increases in the order
3<E<4. It is possible that for E a conformational relaxa-
tion of the phenyl plays some role for this efficiency. This
led us to look for a generally applicable mechanism whereby
conformational relaxations of the substituted phenyl groups
play a decisive role in the fate of the excited state.
A long-standing observation is the vastly different fluores-
cence quantum yield of rhodamine dyes (very large quan-
tum yield) and their counterparts in which the bridge be-
tween the dialkylanilino groups is missing, such as in mala-
chite green (MG) or crystal violet (CV) derivatives (ex-
tremely small quantum yield in fluid solvents). In the latter
group, the quantum yield is strongly enhanced if the twisting
relaxation is hindered by the medium (e.g., in highly viscous
solutions).[49,50] Clearly, the possibility of twisting an anilino
group in these compounds creates a strong nonradiative
channel. It has been proposed that the mechanism of this
nonradiative channel is connected to a CT process,[50–52] that
is, the channel is only introduced if the group that can be
Scheme 2. Structures of the model compounds used for the quantum
chemical calculations.
3398
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 3392 – 3402

Phenanthridinium Chromophores
FULL PAPER
pounds. In the second step, the high-level ab initio method
DFT was used to confirm these results. These demanding
calculations could only be performed for one compound; in
this work we chose to investigate compound 4.
Table 4 shows that our calculations for the ground-state
equilibrium geometry indicate a good agreement between
the calculated transition energy and the experimental results
herein for the CT state are significantly lower in polar sol-
vents due to the solvent interaction with the positive charge.
This lowering is larger for the CT than for the LE state due
to the additional dipolar contribution for the CT state. The
consequence is that the CT state becomes energetically
more easily accessible from the LE state and, as a result, an
increase of the fluorescence quenching with increase of sol-
vent polarity is expected. Experimentally, this has been stud-
ied in detail for a derivative of 4, in which the P moiety is
exchanged for a pyridinium group.[55] The quantum chemical
results obtained for this molecule are similar to those of 4,
that is, a CT state with full charge localization can be
reached by angular relaxation to 908 twist.[56]
Table 4. Comparison of experimental and calculated spectral characteris-
tics of some ethidium derivatives.[a]
Experimental structure
E
3
4
–
Calculated model structure
E(M) 3(M) 4(M)
5(M)
exp. fluorescence enhancement
DE01 (exp.) [eV]
10
2.6
2.61
5420
0.29
7
21
2.6
2.53
8180
0.42
67
nd
nd
2.46
nd
0.41
64
The results of the DFT calculations (Figures 6 and 7) for
4(M) confirm the full CT nature of the excited state at the
orthogonal geometry, with full localization of the positive
2.7
2.55
4640
0.24
–
–
–
–
DE01 (eq, calcd) [eV]
extinction coefficient [mÀ1 cmÀ1
oscillator strength f (calcd)
]
equilibrium twist angle f (eq, calcd) [8] 84
S0 barrier to 908 [cmÀ1
DE01 (908, calcd) [eV]
]
22
2.59
0.95
32
137
2.65[b] 2.52[c]
DECTÀLE (908, calcd) [eV][d]
0.01
À0.12
[a] Ground-state geometry optimizations were performed by applying the
AM1 method. Excited-state characteristics (transition energies DE01 and
energy differences derived thereby, as well as oscillator strengths, f) were
calculated by using ZINDO/s for the AM1-optimized ground-state struc-
ture. Determination of the nature of the excited states involved an analy-
sis of the orbitals and the configuration interaction matrix. [b] S1 (LE)
and S2 (CT) are nearly degenerate. [c] S1 (CT), S2 (LE). [d] Energy differ-
ence from S1(CT) to the next higher (LE) state. nd: not determined.
for the absorption band maximum; the calculations also
allow the localized nature of the lowest lying excited-state
S1 to be confirmed. The calculated oscillator strength in-
creases in the same order as the experimental extinction co-
efficients (3(M)<E(M)<4(M)).
FC
Figure 6. Qualitative scheme of the ground (S0) and excited states (S1
and T1FC) for 4(M). The energies of the states are calculated by B3LYP
(ground states) and TDB3LYP (excited states), with basis set 6-31+G-
AHCTUNGTERG(NNUN 2df,2p) for the optimized equilibrium (left) and orthogonal (right) struc-
tures of 4(M). Absolute energies are given in Hartree units, energy differ-
If the 908 twisted conformation is considered, either full
CT from the D group to the phenanthridinium group (P) is
possible, or the electron of the donor does not leave D, and
the distribution of the positive charge rests as it is (localized
on P). In this case, the corresponding excited state is termed
locally excited (LE). When the LE and CT states are com-
pared, the positive charge is localized on different parts of
the molecule; hence, the CT process leading from LE to CT
is connected with a change of localization of the positive
charge on the donor D, corresponding to a charge shift to-
wards D. If an excited CT state is within energetic reach on
the S1 hypersurface, it can be populated from the initially
reached equilibrium LE state in which the positive charge is
largely localized on the P moiety and leads to fluorescence
quenching as discussed below.
The large charge shift in the CT state also causes an
energy lowering in polar solvents. This can be explained in
the following way: Any distribution of charges in a positive-
ly charged system, such as 4, can be developed into a multi-
pole expansion,[54] in the simplest case, comprising a monop-
ole and a dipole. The dipolar component leads to an addi-
tional energetic stabilization by polar solvents. We can
therefore expect that the gas-phase energy values calculated
ences in eV. Values in parenthesis are calculated oscillator strengths.
charge on the donor moiety, consistent with the TICT
model. Such a CT state is connected with 1) a strong charge
shift; 2) a vanishing transition moment, causing a very small
oscillator strength as well as emissive rate constant; 3) a
large nonradiative rate constant, causing strong fluorescence
quenching; 4) a very close-lying CT triplet state that can
also cause fast depopulation of the excited singlet state. The
three factors 2), 3), and 4) reduce the fluorescence quantum
yield by intrinsic photophysical processes of the twisted CT
state, factor 1) leads to an energetic lowering of CT in polar
solvents, resulting in a larger population of this state. The
control parameters for the fluorescence quenching (or en-
hancement) are the polarity of the medium and the possibil-
ity of the twist relaxation, which is influenced by the rigidity
of the surroundings.
Figure 6 shows that the relaxation from LE to CT is ener-
getically favorable, and that the energy gap to the ground
state and the first triplet state is reduced, causing enhanced
nonradiative deactivation for CT. Moreover, as expected by
the TICT model, a higher lying triplet state of CT nature is
Chem. Eur. J. 2010, 16, 3392 – 3402
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3399

H.-A. Wagenknecht, W. Rettig et al.
isoenergetic with the CT singlet state at 908. In contrast to
CT, the positive charge remains on the phenanthridinium
fragment (pp* state) for a pure LE state. For the equilibri-
of the medium, allowing the twisting relaxation, and 2) by
the energetic availability of the CT state, which is controlled
by the donor strength and the medium polarity. The fluores-
cence quantum yield reduction in the CT state can occur by
1) the intrinsically very small radiative rate constant of this
state; 2) by an enhancement of the nonradiative rate con-
stant generally observed for TICT states and explainable by
the reduction of the S1–S0 energy gap, and 3) by a triplet
state that is energetically similar. On the basis of the calcu-
lations summarized in Table 4, we can predict that a com-
pound related to 5(M) could give rise to an experimental
fluorescence enhancement even larger than that of 4.
FC
um geometry (S1 state), some charge is transferred to D
due to weak CT p interactions, but the main character is
LE. This can be seen in the frontier molecular orbitals for
this geometry, which are somewhat delocalized (Figure 7).
For the 908 geometry, however, the orbitals involved in the
excitation (Figure 7) are localized on the P and D moieties
and confirm the full CT nature, which causes the vanishing
oscillator strength and large charge shift.
Conclusion
To elucidate experimentally the role of the substituents in
positions 3, 8, and especially 6 of the core heterocycle of E
with respect to the photophysical and DNA-binding proper-
ties, we synthesized and analyzed phenanthridinium chromo-
phores 1–4 experimentally as well as theoretically, and com-
pared them with E. The optical and electrochemical proper-
ties, as well as the DNA titration and binding experiments,
reveal two chromophore pairs. The first pair, consisting of
dyes 1 and 2 and lacking the amino substituents in posi-
tions 3 and 8, exhibits clearly altered optical and electro-
chemical properties than those of E. A significant fluores-
cence quenching of these compounds is observed in the
presence of DNA. Their binding affinity to DNA is reduced
by nearly one order of magnitude. The electronic effect of
the phenyl group in position 6 on this type of phenanthridi-
nium dye is rather small.
The optical and redox properties of the second set of
chromophores, 3 and 4, are similar to E. It is known that the
presence of the two electron-donating amino groups in posi-
tions 3 and 8 of the phenanthridinium heterocycle has a sig-
nificant electronic influence and supplies electron density to
the endocyclic iminium core. However, it is important to
note that, in contrast to compounds 1 and 2, the electron-
donating character of the phenyl ring in position 6 of chro-
mophores 3 and 4 is additionally critical for the characteris-
tic and unique fluorescence behavior of these dyes, especial-
ly in the presence of DNA. The binding, as well as fluores-
cence, is clearly related to the electron-donating character
of this substituent. Accordingly, dye 4 shows both the stron-
gest binding affinity and the strongest fluorescence enhance-
ment. This is a remarkable result and is of potential interest
for the future design of fluorescent probes for nucleic acids.
Quantum chemical calculations reveal a general mechanism
that is based on TICT-like states. A CT state can be reached
by twisting the phenyl ring in position 6 towards 908 with re-
spect to the phenanthridinium core. The properties of this
TICT-like excited state are such that fluorescence quenching
occurs. The energetic availability of the CT state depends on
the electron-donating character of the phenyl group and
changes upon binding to DNA. Moreover, the rigidity of
DNA hinders the formation of CT. This mechanism repre-
Figure 7. The HOMO and LUMO orbitals determining the main elec-
tronic configuration of the S1 excited state for the optimized equilibri-
FC
um (62.68 twist, left panel) and orthogonal (right panel) geometries of
the cation 4(M). For the orthogonal geometry, a full electron leaves the
donor moiety D towards P upon HOMO–LUMO excitation (CT state),
and the positive charge is transferred in the opposite direction towards
D. For the equilibrium geometry, this charge shift is reduced due to orbi-
tal delocalization (left panel).
Comparison of the four compounds listed in Table 4,
using the AM1 and ZINDO methods, indicate that this com-
bined method overestimates the energy of the CT state with
respect to that of the LE state, compared with the DFT
method, however, the energetic distance is still small
enough for thermal population.
In 5(M), the CT state is expected to be further lowered in
energy with respect to the LE state due to the better donor
character of the julolidino group than the dimethylanilino
group in 4(M); the calculations confirm these expectations.
CT is the lowest excited state at 908, with a gap of 0.12 eV
to the next (LE) state. For E(M), with the phenyl group as a
very weak donor, the CT energy is very high and this mecha-
nism plays a reduced role. In solution, the LE–CT energy
gap will be somewhat modified with respect to our gas-
phase calculations by the influence of the polar solvent.
In summary, five possible factors controlling the fluores-
cence enhancement on the basis of a TICT state can be dis-
cussed: The population of CT is controlled by 1) the rigidity
3400
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 3392 – 3402

Phenanthridinium Chromophores
FULL PAPER
3997; c) D. A. Dunn, V. H. Lin, I. E. Kochevar, Biochemistry 1992,
31, 11620–11625; d) A. M. Brun, A. Harriman, J. Am. Chem. Soc.
1992, 114, 3656–3660; e) S. J. Atherton, P. C. Beaumont, J. Phys.
Chem. 1995, 99, 12025–12029; f) A. I. Kononov, E. B. Moroshkina,
N. V. Tkachenko, H. Lemmetyinen, J. Phys. Chem. B 2001, 105,
535–541; g) P. T. Henderson, E. Boone, G. B. Schuster, Helv. Chim.
Acta 2002, 85, 135–151.
sents an important addition to the previous proposal of pro-
tonation in the excited state and can help to optimize the
performance of the fluorescence enhancement factor of
future fluorescent probes for nucleic acids.
[16] a) S. O. Kelley, R. E. Holmlin, E. D. A. Stemp, J. K. Barton, J. Am.
Chem. Soc. 1997, 119, 9861–9870; b) D. B. Hall, S. O. Kelley, J. K.
Barton, Biochemistry 1998, 37, 15933–15940; c) S. O. Kelley, J. K.
Barton, Chem. Biol. 1998, 5, 413–425; d) C. Wan, T. Fiebig, S. O.
Kelley, C. R. Treadway, J. K. Barton, A. H. Zewail, Proc. Natl.
Acad. Sci. USA 1999, 96, 6014–6019.
Acknowledgements
This work was generously financed by the Deutsche Forschungsgemein-
schaft and by the Graduate College “Sensory Photoreceptors in Natural
and Artificial Systems” (GRK640).
[17] a) N. Amann, R. Huber, H.-A. Wagenknecht, Angew. Chem. 2004,
116, 1881–1883; Angew. Chem. Int. Ed. 2004, 43, 1845–1847; b) L.
Valis, N. Amann, H.-A. Wagenknecht, Org. Biomol. Chem. 2005, 3,
36–38; c) L. Valis, Q. Wang, M. Raytchev, I. Buchvarov, H.-A. Wa-
genknecht, T. Fiebig, Proc. Natl. Acad. Sci. USA 2006, 103, 10192–
10195.
[1] S. L. Lerman, J. Mol. Biol. 1961, 3, 18–30.
[2] a) V. K. Misra, B. Honig, Proc. Natl. Acad. Sci. USA 1995, 92, 4691–
4695; b) S. E. Patterson, J. M. Coxon, L. Stekowski, Bioorg. Med.
Chem. 1997, 5, 277–281; c) C. Medhi, J. B. O. Mitchell, S. L. Price,
ˇ
A. B. Tabor, Biopolymers 1999, 52, 84–93; d) D. Reha, M. Kabelꢄc,
[18] R. Huber, N. Amann, H.-A. Wagenknecht, J. Org. Chem. 2004, 69,
744–751.
[19] M. Bahr, L. Valis, H.-A. Wagenknecht, E. Weinhold, Nucleosides
Nucleotides Nucleic Acids 2007, 26, 1581–1584.
[20] J. Olmsted III, D. R. Kearns, Biochemistry 1977, 16, 3647–3654.
[21] L. P. G. Wakelin, M. J. Waring, Mol. Pharmacol. 1974, 9, 544–561.
[22] M. Le Bret, O. Chalvet, J. Mol. Struct. 1977, 37, 299–319.
[23] C. Mehdi, J. B. Mitchell, S. L. Price, A. B. Tabor, Biopolymers 1999,
52, 84–93.
ˇ
ˇ
ˇ
F. Ryjꢄcek, J. Sponer, J. E. Sponer, M. Elstner, S. Suhai, P. Hobza, J.
Am. Chem. Soc. 2002, 124, 3366–3376.
[3] H. Ihmels, D. Otto, Top. Curr. Chem. 2005, 258, 161–204.
[4] a) G. Viola, F. DallꢂAcqua, N. Gabellini, S. Moro, D. Vedaldi, H.
Ihmels, ChemBioChem 2002, 3, 550–558; b) G. Viola, M. Bressanini,
N. Gabellini, D. Vedaldi, F. DallꢂAcqua, H. Ihmels, Photochem. Pho-
tobiol. Sci. 2002, 1, 882–889.
[5] a) H. Ihmels, K. Faulhaber, K. Wissel, G. Viola, D. Vedaldi, Org.
Biomol. Chem. 2003, 1, 2999–3001; b) A. Granzhan, H. Ihmels, G.
Viola, J. Am. Chem. Soc. 2007, 129, 1254–1267.
[6] a) A. R. Morgan, J. S. Lee, D. E. Pulleyblank, N. L. Murray, D. H.
Evans, Nucleic Acids Res. 1979, 7, 547–569; b) A. R. Morgan, D.
Evans, J. S. Lee, D. E. Pulleyblank, Nucleic Acids Res. 1979, 7, 571–
594.
[7] H. Nishiwaki, M. Miura, K. Imai, R. Ohno, K. Kawashima, Cancer
Res. 1974, 34, 2699–2703.
[8] a) R. V. Guntaka, B. W. Mahy, J. M. Bishop, H. E. Varmus, Nature
1975, 253, 507–511; b) N. W. Luedtke, Y. Tor, Biopolymers 2003, 70,
103–119.
[9] a) J. N. A. Tettey, G. G. Skellern, M. H. Grant, J. M. Midgley, J.
Pharm. Biomed. Anal. 1999, 21, 1–7; b) J. N. A. Tettey, G. G. Skel-
lern, J. M. Midgley, M. H. Grant, A. R. Pitt, Chem.-Biol. Interact.
1999, 123, 105–115.
[10] a) M. J. Waring in Antibiotics III Mechanism of Action of Antimicro-
bial and Antitumour Agents (Eds.: J. W. Corcoran, F. E. Hahn),
Springer, Heidelberg, 1975, pp. 141–165; b) L. P. G. Wakelin, M. J.
Waring, J. Mol. Biol. 1980, 144, 183–214.
[11] a) V. L. Singer, T. E. Lawlor, S. Yue, Mutat. Res. 1999, 439, 37–47;
b) M. Fukunaga, B. A. Cox, R. S. von Sprecken, L. W. Yielding,
Mutat. Res. 1984, 127, 31–37; c) T. Ohta; S. Tokishita, H. Yamagata,
Mutat. Res. 2001, 492, 91–97; d) S. Hoare, R. B. Kemp, M. A. Ka-
derbhai, Biochem. Soc. Trans. 1993, 21, 504.
[12] a) B. Gaugain, J. Barbet, R. Oberlin, B. P. Roques, J.-B. Le Pecq,
Biochemistry 1978, 17, 5071–5078; b) B. Gaugain, J. Barbet, N. Ca-
pelle, B. P. Roques, J.-B. Le Pecq, Biochemistry 1978, 17, 5078–5088;
c) M. M. Becker, P. B. Dervan, J. Am. Chem. Soc. 1979, 101, 3664–
3666; d) K. F. Kuhlmann, C. W. Mosher, J. Med. Chem. 1981, 24,
1333–1337; e) I. G. Lopp, W. D. Wilson, D. W. Boykin, J. Heterocycl.
Chem. 1982, 19, 695–696.
[13] a) L. Benimetskaya, N. V. Bulychev, A. L. Kozionov, A. A. Koshkin,
A. V. Lebedev, S. Y. Novozhilov, M. I. Stockman, Biopolymers 1989,
28, 1129–1147; b) G. Krishnamurthy, T. Polte, T. Rooney, M. E.
Hogan, Biochemistry 1990, 29, 981–988.
[14] a) R. F. Pasternack, M. Caccam, B. Keogh, T. A. Stephenson, A. P.
Williams, E. J. Gibbs, J. Am. Chem. Soc. 1991, 113, 6835–6840;
b) W. J. Jin, Y. S. Wei, C. S. Liu, G. L. Shen, R. Q. Yu, Spectrochim.
Acta, Part A 1997, 53, 2701–2707.
[15] a) P. Fromherz, B. Rieger, J. Am. Chem. Soc. 1986, 108, 5361–5362;
b) S. J. Atherton, P. C. Beaumont, J. Phys. Chem. 1987, 91, 3993–
ˇ
[24] a) D. Reha, M. Kabelꢄc, F. Ryjꢄcek, J. Sponer, J. E. Sponer, M. Elst-
ner, S. Suhai, P. Hobza, J. Am. Chem. Soc. 2002, 124, 3366–3376;
ˇ
ˇ
b) T. Kubar, M. Hanus, F. Ryjꢄcek, P. Hobza, Chem. Eur. J. 2006, 12,
280–290.
[25] W. Leupin, J. Feigon, W. A. Denny, D. R. Kearns, Biophys. Chem.
1985, 22, 299–305.
[26] N. W. Luedtke, Q. Liu, Y. Tor, Chem. Eur. J. 2005, 11, 495–508.
[27] W. J. Firth III, C. L. Watkins, D. E. Graves, L. W. Yielding, J. Hetero-
cycl. Chem. 1983, 20, 759–765.
[28] a) L. W. Yielding, W. J. Firth III, Mutat. Res. 1980, 71, 161–168;
b) M. Fukunaga, L. W. Yielding, W. J. Firth III, K. L. Yielding,
Mutat. Res. 1980, 78, 151–157; c) M. Fukunagu, L. W. Yielding,
Mutat. Res. 1983, 121, 89–94; d) M. Fukunaga, Y. Mizuguchi, L. W.
Yielding, K. L. Yielding, Mutat. Res. 1984, 127, 15–21; e) F. Leng, D.
Graves, J. B. Chaires, Biochim. Biophys. Acta, Gene Struct. Expres-
sion 1998, 1442, 71–81.
[29] L. W. Yielding, K. L. Yielding, J. E. Donoghue, Biopolymers 1984,
23, 83–110.
[30] a) G. Morgan, L. P. Walls, J. Chem. Soc. 1938, 389–397; b) L. P.
Walls, J. Chem. Soc. 1945, 294–300.
[31] T. I. Watkins, J. Chem. Soc. 1952, 3, 3059–3064.
[32] S. S. Berg, J. Chem. Soc. 1963, 14, 3635–3640.
[33] J. Loccufier, E. Schacht, Tetrahedron 1989, 45, 3385–3396.
[34] a) N. P. Gritsan, A. A. Koshkin, A. Y. Denisov, Y. Y. Markushin,
E. V. Cherepanova, A. L. Lebedev, J. Photochem. Photobiol. B 1997,
37, 40–51; b) N. Amann, H.-A. Wagenknecht, Tetrahedron Lett.
2003, 44, 1685–1690; c) S. Rangarajan, S. H. Friedman, Bioorg.
Med. Chem. Lett. 2007, 17, 2267–2273.
[35] J. N. Demas, G. A. Crosby, J. Phys. Chem. 1971, 75, 991–1024.
[36] D. F. Eaton, Pure Appl. Chem. 1988, 60, 1107–1114.
[37] a) A. Weller, Z. Phys. Chem. (Muenchen Ger.) 1982, 133, 93–98;
b) K. Kumar, I. V. Kurnikov, D. N. Beratan, H. D. Waldeck, M. B.
Zimmt, J. Phys. Chem. A 1998, 102, 5529–5541.
[38] G. Cosa, K.-S. Foscaneanu, J. R. N. McLean, J. P. McNamee, J. C.
Scaiano, Photochem. Photobiol. 2001, 73, 585–599.
[39] a) M. J. Waring, J. Mol. Biol. 1965, 13, 269–282; b) J.-B. Le-Pecq, C.
Paleotti, J. Mol. Biol. 1967, 27, 87–106; c) R. L. Letsinger, M. E.
Schott, J. Am. Chem. Soc. 1981, 103, 7394–7396.
[40] J. D. McGhee, P. H. von Hippel, J. Mol. Biol. 1974, 86, 469–489.
Chem. Eur. J. 2010, 16, 3392 – 3402
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3401

H.-A. Wagenknecht, W. Rettig et al.
[41] a) H. W. Zimmermann, Angew. Chem. 1986, 98, 115–131; Angew.
Chem. Int. Ed. Engl. 1986, 25, 115–130; b) J. Pauluhn, H. W. Zim-
mermann, Ber. Bunsen-Ges. 1979, 83, 76–82.
[42] E. D. Horowitz, N. V. Hud, J. Am. Chem. Soc. 2006, 128, 15380–
15381.
[43] S. O. Kelley, J. K. Barton, Chem. Biol. 1998, 5, 413–425.
[44] V. V. Pavlishchuk, A. W. Addison, Inorg. Chim. Acta 2000, 298, 97–
102.
[49] T. Fçrster, G. Hoffmann, Z. Phys. Chem. (Muenchen Ger.) 1971, 75,
63.
[50] a) M. Vogel, W. Rettig, Ber. Bunsen-Ges. 1985, 89, 962–968; b) M.
Vogel, W. Rettig, Ber. Bunsen-Ges. 1987, 91, 1241–1247.
[51] W. Rettig, Angew. Chem. 1986, 98, 969–986; Angew. Chem. Int. Ed.
Engl. 1986, 25, 971–988.
[52] Z. R. Grabowski, K. Rotkiewicz, W. Rettig, Chem. Rev. 2003, 103,
3899–4031.
[45] M. Torimura, S. Kurata, K. Yamada, T. Yokomaku, Y. Kamagata, T.
Kanagawa, R. Kurane, Anal. Sci. 2001, 17, 155–160.
[53] P. Plaza, N. Dai Hung, M. M. Martin, Y. H. Meyer, M. Vogel, W.
Rettig, Chem. Phys. 1992, 168, 365–373.
[46] H.-A. Wagenknecht, Nat. Prod. Rep. 2006, 23, 973–1006.
[47] N. C. Garbett, N. B. Hammond, D. E. Graves, Biophys. J. 2004, 87,
3974–3981.
[48] a) I. Juranovic, Z. Meic, I. Piantanida, L.-M. Tumir, M. Zinic, Chem.
Commun. 2002, 1432–1433; b) L.-M. Tumir, I. Piantanida, P. Novak,
M. Zinic, J. Phys. Org. Chem. 2002, 15, 599–607; c) L.-M. Tumir, I.
Piantanida, I. J. Cindric, T. Hrenar, Z. Meic, M. Zinic, J. Phys. Org.
Chem. 2003, 16, 891–899; d) L.-M. Tumir, I. Piantanida, I. Juranov-
ic, Z. Meic, S. Tomic, M. Zinic, Chem. Commun. 2005, 2561–2563.
[54] C. J. F. Bçttcher in Theory of electric polarization, Elsevier, Amster-
dam, 1973, p. 42.
[55] H. Ephardt, P. Fromherz, J. Phys. Chem. 1993, 97, 4540–4547.
[56] V. Kharlanov, W. Rettig, Chem. Phys. 2007, 332, 17–26.
Received: October 13, 2009
Published online: February 19, 2010
3402
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 3392 – 3402