A new 4,9-diazapyrenium intercalator for single- and double- stranded nucleic acids: Distinct differences from related diazapyrenium compounds and ethidium bromide

Article abstract of DOI:10.1039/b103214n

The new ligand 6 is structurally related to ethidium bromide (EB) and other diazapyrenium compounds (1,2) studied earlier, but exhibits significant differences. In contrast to ligands such as 1 and 2 the new 6 is stable over a large pH range. At concentra

Full text of DOI:10.1039/b103214n

View Article Online / Journal Homepage / Table of Contents for this issue
A new 4,9-diazapyrenium intercalator for single- and double-
stranded nucleic acids: distinct differences from related
diazapyrenium compounds and ethidium bromide†
2
a
b
a
b
˘
Ivo Piantanida, Barbara Susanne Palm, Mladen Zinic´ and Hans-Jörg Schneider*
a Laboratory of Supramolecular and Nucleoside Chemistry,
Department of Chemistry and Biochemistry, Rudjer Boskovic Institute, HR 10000,
Zagreb, POB 1016, Croatia. E-mail: zinic@rudjer.irb.hr
b FR Organische Chemie der Universität des Saarlandes, D 66041 Saarbrücken, Germany.
E-mail: ch12hs@rz.uni-sb.de
Received (in Cambridge, UK) 9th April 2001, Accepted 22nd June 2001
First published as an Advance Article on the web 9th August 2001
The new ligand 6 is structurally related to ethidium bromide (EB) and other diazapyrenium compounds (1, 2)
studied earlier, but exhibits significant differences. In contrast to ligands such as 1 and 2 the new 6 is stable over a
large pH range. At concentrations > 10Ϫ4 mol dmϪ3 6 forms a dimer, which is characterised by NMR analysis. The
association constants of 6 are larger with purine than with pyrimidine nucleotides and are the same for AMP, ADP
or ATP, with an absence of any charge effects. Surprisingly, under conditions close to saturation of intercalation
binding sites association of 6 with single-stranded poly U and poly A is 100 to 1000 times more efficient than that
of EB or analogue 2. Also the affinity of 6 toward poly U is significantly higher than toward poly C. CD spectra
indicate that stacking of 6 with poly U induces a strong helicity in the usually disordered structure of this single
strand. The binding of 6 to poly C is pH dependent, as a consequence of the known formation of a poly CHϩ–poly
CHϩ double strand at pH < 5. With double-stranded polynucleotides the binding affinity of 6 and EB is similar
for RNA homopolymers. Striking fluorescence differences, however, are observed with complexes of 6–poly GC
(decrease of emission intensity of 6 by 80%), and 6–poly AU (increase by 100%). Similar effects are also observed
for the DNA polynucleotides. At higher ligand to phosphate ratios (r > 0.2) 6 shows with double strands electro-
static binding in addition to intercalation, with ensuing opposite effects in fluorescence emission. The biphasic
melting profiles of poly dA–poly dT with 6 differ sharply from those observed with 1 and EB and are in line with
the dependence of binding modes obvious from the fluorescence studies. Melting analyses show with calf thymus
DNA and with poly dA–poly dT an increase (∆∆T ) in melting temperature with 6 about twice as high as that with
1 or EB. With the RNA model poly A–poly U the ∆T m values of 6 at low ionic strength are about seven times higher
than those observed with 2 or with EB. In contrast to other 4,9-diazapyrenium compounds which show with poly
A–poly U a destabilising and a stabilising step, 6 gives only positive ∆T values at both pH = 5.0 and pH = 7.0.
Extensive supporting data for this paper are available as supplementary material† in electronic form, allowing
re-evaluation of experimental results with suitable programs for curve fitting.
Recently we have reported on the interaction of a series of
Introduction
new 4,9-diazapyrenium derivatives (1, 2, Chart 1) with nucleo-
The search for new nucleic acid intercalators1 is stimulated by a
tides and with double-stranded (ds) nucleic acids.10,11 Some of
variety of reasons. One can couple such intercalators, e.g. to
these 4,9-diazapyrenium derivatives exhibit in vitro antitumor
oligonucleotides, or to catalytic units for cleaving nucleic acids,
activity.12 However, a more detailed examination was hampered
for the development of stable triple helices2 or of synthetic
by their instability at neutral pH.11 In the present work,
nucleases,3 exploiting the relatively high association constants
we describe the synthesis via intermediates 3 to 5 of a novel
4-methyl-2,7-diamino-5,10-diphenyl-4,9-diazapyrenium anal-
ogue 6 (Scheme 1) which is stable over a wide pH range (pH =
3–10), including physiological pH conditions. Although 6 is
of intercalators. With new systems, capable of binding mono-
or oligonucleotides, one can also hope to reach a deeper under-
standing of intercalation mechanisms.4 The recently discovered
relatively selective interaction of ethidium bromide (EB) with
structurally very similar to ethidium bromide EB it possesses
particular RNA, HIV–related sequences5 stirs new interest in
several features distinct from classical intercalators. The larger
the development of new aromatic systems with possibly higher
aromatic surface could enhance the intercalation strength,
selectivity. Besides ethidium bromide, related compounds
the 5,10-diphenyl substituents are expected to be oriented
with extended aromatic systems, such as 2,7-diazapyrenium and
orthogonal to the diazapyrene plane, which should sterically
2,7-diazaperopyrenium6 cations, as well as acridinium cations,7
restrict orientation possibilities upon intercalation. We report
have been studied in detail. For systems which exhibit photo-
on the surprising spectroscopic properties of the new inter-
induced cleavage of double-stranded polynucleotides,8 it was
calation complexes of 6, the binding to nucleotides and on
affinities toward single- and double-stranded DNA/RNA poly-
nucleotides, based on UV/Vis, fluorescence and CD measure-
found that small structural variations strongly influence the
binding affinity to polynucleotides.6,9
ments as well as on melting experiments and viscometry.
The data are compared to the ones obtained with EB and the
† Electronic supplementary information (ESI) available: NMR data,
fluorimetric titration data, thermal melting curves. See http://
www.rsc.org/suppdata/p2/b1/b103214n/
previously reported ones of two closely related 4,9-diazapyr-
enium derivatives 1 and 211 (Chart 1).
1808
J. Chem. Soc., Perkin Trans. 2, 2001, 1808–1816
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b103214n

View Article Online
Chart 1
Results and discussion
Synthesis
2,7-Diamino-5,10-diphenyl-4,9-diazapyrene 3 was prepared
according to the previously published procedure.13 The syn-
thesis of 6 from 3 is outlined in Scheme 1. The amino groups of
3 were protected using benzyloxycarbonyl chloride in DMF–
NaHCO3 suspension at room temperature giving 4. Heating
of 4 with an equimolar amount of methyl trifluoromethane-
sulfonate (methyl triflate) in chlorobenzene, followed by sub-
sequent in situ amino deprotection by trifluoromethanesulfonic
acid at room temperature gave 5 as the triflate salt. The
attempted preparation of the 4,9-dimethyl derivative of 5 failed
despite the use of 2–5 molar excess of methyl triflate and
prolonged heating. Under the same reaction conditions
dimethylated analogue 2 was prepared in high yield,10 which
points to a pronounced deactivation effect of the 2,7-benzyl-
oxycarbonylamino substituents in 4. To improve the water
solubility of 5, the triflate anion was exchanged with hydrogen-
sulfate giving, after recrystallization from DMSO–dichloro-
methane, a dark blue precipitate of 6. The hygroscopic nature
of the precipitates hampered elementary analysis; however, the
structures of 5 and 6 were confirmed by (1D and 2D) NMR and
ES-MS data.
Spectroscopic properties of 6
1H-NMR spectra. Changing the concentration of 6 from
1 × 10Ϫ4 to 5 × 10Ϫ3 mol dmϪ3 all signals exhibited upfield
shifts by up to 0.7 ppm (supplementary data),† indicating
self-stacking. Assuming predominant formation of a dimer, the
stability constant Ks,dim. of ≈400 100 dm3 molϪ1 is calculated14
(Fig. 1). The association constant is similar to those calculated
for phenanthridinium and acridinium dimers (EB, proflavine
Ks,dim. ≈ 180 and ≈ 350 dm3 molϪ1, respectively).15 The signals of
6 protons at 5 × 10Ϫ3 mol dmϪ3 concentration were assigned on
the basis of NOESY and ROESY (suppl ementary data)† cross
peaks. At this concentration 6 is considerably self-stacked and
intermolecular ROESY interactions were observed between the
H6 proton of the diazapyrenium ring and the o-protons of the
10-phenyl substituent, and also between H1 of the diazapyr-
enium ring and the o-protons of the 5-phenyl substituent. The
observed intermolecular ROESY cross peaks suggest the form-
ation of a stacked dimer with offset ADAP molecules and the
second molecule rotated by 180Њ around the longer axes (Fig. 2).
Scheme 1 Synthesis of 2,7-diamino-5,10-diphenyl-4-methyl-4,9-diaza-
pyrenium hydrogensulfate (6).
[λmax/nm, (ε/dm3 molϪ1 cmϪ1): 236, (55800); 351 (9000); 389
(10000)] and 2 [λmax/nm, (ε): 243 (45400); 335 (12800); 400
(24000)] the amino substituents of 6 induce pronounced
red shifts of absorption maxima in UV/Vis spectra of the
4,9-diazapyrenium ring [λmax/nm (ε): 266 (37700); 368 (5050);
517 (5750)]. Absorption of 6 in buffered solution (pH = 5 and 7;
Electronic absorption and fluorescence spectra. In comparison
with the previously prepared 4,9-diazapyrenium derivatives10
1
J. Chem. Soc., Perkin Trans. 2, 2001, 1808–1816
1809

View Article Online
Fig. 1 Experimental (᭹) and calculated (——) data of ∆δ (ppm) for
proton H6 according to dimerisation equation.
Fig. 3 Formation of 4,9-diazapyrene pseudobase.
which is obviously due to the influence of the amino substi-
tuents. Only addition of GMP yielded total quenching of 6
emission. Similar differences in fluorimetric response were
reported for proflavine and explained by the more pronounced
electron donating properties of guanine compared to other
nucleic bases.17
Stability constants (Ks) were calculated by processing fluori-
metric titration data with the program SPECFIT;18 the best fit
for a 1 : 1 stoichiometry of complexes was obtained in all cases.
Titration of 6 with UMP gave too small changes of emission
intensity for accurate calculation of the binding constant.
However, a constant similar to that of CMP can be estimated.
Compared to previously reported binding constants for 2
(nucleotide, log Ks: AMP, 1.67; ADP, 1.74; ATP, 1.78; GMP,
1.66; CMP, <1)10 and EB (GMP, 2.01; AMP, 1.92),19 binding
with 6 is somewhat stronger (CMP, 1.61; GMP, 2.30; AMP,
2.16; ATP, 2.39) but similar to 110 (CMP, 1.37; GMP, 2.11;
AMP, 2.21; ATP, 2.24).
Comparing 2 and 6 with EB it appears that neither the
presence of two positive charges (2) nor the larger aromatic
surface (6) leads to a considerable increase of binding strength.
Similar binding affinity of 6 for AMP, ADP and ATP with 2, 3
and 4 negative charges respectively, and larger Ks values for
purinic than for pyrimidinic nucleotides point to the aromatic
π–π stacking interactions between 6 and nucleobases as the
dominant interaction in the complexes. As in related cases20
there is a striking absence of nucleotide charge effects.
Fig. 2 Observed intra- (left) and inter-molecular (right) ROESY inter-
actions in possible ADAP dimer with indicated intermolecular H6–Hortho
(C10-phenyl) and H1–Hortho (C5-phenyl) interactions.
0.01 mol dmϪ3) is found to be linearly concentration dependent
up to 8 × 10Ϫ5 mol dmϪ3. In the concentration range 8 × 10Ϫ5
–
2 × 10Ϫ4 mol dmϪ3 a small hypochromicity was observed,
indicating self-stacking of 6, in accord with the self-
association constant Ks,dim. ≈ 400 100 dm3 molϪ1 determined
by 1H-NMR.
The fluorescence emission intensity of 6 is significantly
lower than those of analogues10 1 and 2, with the emission
maxima strongly red-shifted by almost 200 nm (λmax = 613 nm).
Excitation spectra of 6 agree with the absorption spectra.
Fluorescence intensity is found to be linearly concentration
dependent up to 5 × 10Ϫ6 mol dmϪ3
.
Addition of 20% of D2O to an aqueous solution of 6 induced
a 30% increase in fluorescence intensity. A similar effect was
previously found for EB and explained by an amino group
NH–ND exchange which decreases the quenching by amino
proton transfer to water molecules in the excited state.16
Interactions of 6 with single-stranded RNA polynucleotides
Single-stranded RNA polymers of homogenous structure (poly
A, G, C, U) are good models for testing the nucleic base
dependent affinity of selected molecules and also the spectro-
scopic response of the studied compound induced by binding to
polymers. Most single-stranded (ss) RNA polynucleotides
(except poly U) form rather well organised helical structures in
water at ionic strength I = 0.01–0.1 mol dmϪ3, due to stacking
of nucleobases.21
An interesting feature of 6 is the absence of pH dependent
formation of 5-hydroxy-4,9-diazapyrene pseudobase which one
finds with all previously prepared derivatives including 1 and 2
(Fig. 3).10 Monitoring the changes in UV spectra with variation
of pH (pH range 3–10) only a slight variation of ε can be
observed (less than 5%) between pH 4.5 and 7. Amino groups
are protonated at pH < 4 as is apparent from pH dependent
absorbance changes. The similar chemical shift pattern of
UV/Vis titrations. Additions of ss polymers (poly A, G, C, U)
at pH 5 and 7 induce bathochromic shifts in the spectra of 6
for 20–40 nm and hypochromic effects up to 50%. Similar
changes observed earlier for most of the classical intercalators
were explained by their intercalative binding mode.22 In most
titration experiments isosbestic points are observed, suggesting
that only two spectroscopically active species are present at
equilibrium. Only in the titration with poly A at pH = 5, is
a clear deviation from the isosbestic point observed, due to
protonation of poly A. Under such conditions a double helix
of poly AHϩ–poly AHϩ is partially formed,21 resulting in two
different complexes with 6.
1
diazapyrenium protons in H NMR spectra at pD 2.4 and 6.2
also exclude, formation of pseudobase. Apparently, the pres-
ence of the electron donating 5,10-diphenyl and 2,7-diamino
substituents of 6 strongly increases its stability to pseudobase
formation in neutral and weakly basic conditions.
Interactions of 6 with nucleotides in aqueous media
Fluorimetric titration of 6 with nucleotides exhibited inter-
esting nucleic base specific spectroscopic changes. Addition of
AMP, ATP and CMP in excess does not quench fluorescence
emission down to zero as in the case of analogues 1 and 2,10
1810
J. Chem. Soc., Perkin Trans. 2, 2001, 1808–1816

View Article Online
Table 1 Stability constants (Ks) and ratios (n)a ([bound compound] : [polynucleotide]) calculated for UV/Vis and fluorimetric titrations of 6 or EB
with ss polynucleotides at pH = 5.0 (buffer citric acid (0.01 M), I = 0.025 M) and pH = 7.0 (buffer Na cacodylate, I = 0.02 M)
poly A
poly G
poly C
poly U
pH
n
log Ks
n
log Ks
n
log Ks
n
log Ks
b
b
6
UV
6
Fluo
EB
UV
5
7
5
7
5
7
0.6
0.8
0.6
1
0.3
0.5
4.5
4.8
4.7
4.5
3.8
3.1
0.5
1c
1c
1c
1c
1c
4.5
3c
1
1
1
>5d
>5d
3.9b
4.4b
3c
1
>5d
b
b
2.9c
2.6c
3.3c
<3c
0.5
0.5
0.5
4.1
1
3.3b
3.9
1c
1c
<3c
a Accuracy of n 10–30%, consequently log Ks values vary within the same order of magnitude. b Due to formation of different types of complexes
only estimation of cumulative log Ks was possible. c Small spectroscopic changes allowed only estimation of n and log Ks. d Linear change of
absorbance with [polynucleotide] allowed only estimation of log Ks.
Vis, respectively). Consequently, in fluorescence titration first
spectroscopic changes are observed at ratio r (c6/cpolynucl.) << 1
where the intercalative mode of binding prevails. In line with
that, similar stability constants are obtained by UV/Vis and
fluorescence measurements at pH = 7 where protonation of
poly C is negligible.
Stability constants Ks calculated from titration data of 6 with
poly A and poly G at large excess of either polynucleotide
(r << 1) were of the same order of magnitude (Table 1). Differ-
ent behaviour was observed, however, when 6 was in excess
(r > 1); addition of poly A produced small but significant
quenching while addition of poly G had no effect. Taking into
account the UV/Vis titration experiments where changes of
absorbance were quite different for poly A and poly G one can
assume formation of two different complexes with poly A and
Fig. 4 UV/Vis titration of 6, c = 2 × 10Ϫ5 mol dmϪ3 with poly U; pH = 7
(PIPES buffer, c = 0.01 mol dmϪ3).
only one with poly G.
Addition of poly U (pH 5 and 7) induced rather small but
significant fluorescence quenching (7–9%) of 6 with excess of
the compound (ratio r > 1). With excess of poly U (10–100
times) dominant fluorescence quenching (15%) of 6 was
observed. It was not possible to clearly differentiate the con-
centration range where only one binding mode prevails. There-
fore, the cumulative stability constant was calculated, being
significantly larger than for poly C (pH = 7).
UV/Vis titration data of 6 with poly U (Fig. 4) point to
surprisingly high complex stability. The linear dependence of
absorbance changes vs. concentration of poly U observed did
not allow calculation of Ks and n by the Scatchard equation.
The end value at the 6 : poly U ratio of 1 suggests a value of
n ≈ 1 and Ks > 105 dm3 molϪ1. Similar results are obtained for
titration of 6 with poly A at pH = 7. Some 100–1000 times
higher concentrations of poly A and poly U, respectively, need
to be added to the solution of EB or 2 in order to induce
spectroscopic changes comparable to those of 6 under the
same experimental conditions. This clearly shows much weaker
binding of EB and 2 compared to 6 (Table 1). Ks values for 6
and EB for the complexes with poly G and poly C are com-
parable except for the Ks for 6 and poly C at pH = 5 which is
close to two orders of magnitude higher than that at pH = 7
(Table 1). The latter can be explained by protonation of poly C
at pH = 5 and formation of a double helix21 which increases
the electrostatic interactions of 6 and polyphosphates (see the
following paragraph on non-intercalative interactions with
double-stranded polynucleotides).
CD spectra. It is known that intercalation of some flat
aromatic molecules into single-stranded polynucleotides
induces large chirality changes and consequently significant
effects in their CD and LD spectra.7,23 The CD and LD changes
are useful for determination of mobility and orientation of
intercalated dye in double helical polynucleotides.24 Most of the
ss polynucleotides (poly A, G, C) form helical structures in
water due to stacking interactions between adjacent bases and
it is possible to monitor their conformational changes using CD
spectroscopy.23
The variation of ratio r (c6/cpoly U) from 0 to 0.8 induces a
large increase of CD intensity of poly U spectra (Fig. 5). The
observation of the isosbestic point at 272 nm suggests form-
ation of only one complex at equilibrium with free polymer.
Under the same conditions addition of EB had no effect. Poly
U and poly T under these experimental conditions, in contrast
to other single-stranded polynucleotides, do not form organised
helical structures in water due to lack of stacking interactions
between bases.21 Therefore, the observed CD effects strongly
suggest an increase of chirality in the 6–poly U complex com-
pared to free polymer. The rational explanation can be the
induction of helicity by intercalation of 6. Structural features
of 6 and uracils accompanied by strong stacking interactions
can provide the driving force for helical organisation of the
complex. In contrast to poly U, addition of 6 as well as EB to
solutions of poly A, poly G and poly C produced a decrease of
CD intensities in accord with a decrease of helicity upon inter-
calation. The effects of 6 in CD spectra of poly A are more
pronounced than those of EB; this observation is in accord
Fluorescence titrations. Addition of almost all polynucleo-
tides quenched fluorescence emission of 6 (with poly A by 30%,
and with poly G by 70%) with a similar trend to that observed
in titrations with nucleotides. Results obtained from titration of
6 with poly C appeared to be strongly pH dependent, which is
not the case for EB (Table 1). Binding of 6 with poly C at
pH = 5 results in a unique fluorescence increase (35%) while at
pH = 7 fluorescence decreases. However, the fluorescence gives
one order of magnitude lower Ks for 6 and poly C at pH = 5
than that determined by UV/Vis measurements. This can be
explained by a much lower electrostatic contribution (between 6
and the phosphates of poly CHϩ–poly CHϩ formed at pH = 521)
to overall binding (intercalative and electrostatic) in the
fluorescence titration due to a lower concentration range
(c6 = 2 × 10Ϫ6 and 2 × 10Ϫ5 mol dmϪ3 for fluorescence and UV/
J. Chem. Soc., Perkin Trans. 2, 2001, 1808–1816
1811

View Article Online
Fig. 5 CD titration of poly U (A, c = 4.5 × 10Ϫ5 mol dmϪ3) and poly A
(B, c = 3 × 10Ϫ5 mol dmϪ3) with 6; ratio r ([6] : [poly X]) shown for each
spectrum; pH = 7, buffer PIPES, 0.01 M.
Fig. 6 Fluorimetric titration of 6, c = 2 × 10Ϫ6 mol dmϪ3 with poly AU
(᭹) and poly GC (᭿), pH = 7 (buffer Na cacodylate, c = 0.02 mol
dmϪ3).
Table 2 Stability constants (Ks) and ratios (n)a ([bound 6] : [poly-
nucleotide]) calculated from fluorimetric titrations of 6 with ds poly-
nucleotides at pH = 5.0 (buffer citric acid (0.01 M), I = 0.025 M) and
pH = 7.0 (buffer Na cacodylate (0.02 mol dmϪ3), I = 0.02 mol dmϪ3
)
and 6–GMP complex (non-fluorescent) it is likely that 6 gives
different fluorescence responses depending on the presence
of A or G in DNA/RNA polymers. Some acridine derivatives
also exhibited a similar nucleic base selective spectroscopic
response; the effect was correlated with the electron donating
properties of guanine.17,25
pH = 5
pH = 7
n
log Ks
n
log Ks
poly A–poly U
poly G–poly C
0.1
6.5
6.8
5.4
0.16
0.16
0.14
0.14
0.47
—
6.5
6.1
5.6
5.9
6.9b
—
For both 6 and EB26 additional, non-intercalative binding
is present at r > 0.2. The non-intercalative binding contribution
is more pronounced for 6–DNA complexes (especially for the
6–poly dAdT complex) than for the RNA complexes. It is inter-
esting to note that addition of poly dAdT at low ionic strength
and ratio r > 0.2 quenches the fluorescence of 6 (Fig. 7, A) while
the emission decrease is not present if the titration is performed
at high ionic strength (I = 0.1 mol dmϪ3) (Fig. 7, B). The latter
suggests that the observed emission decrease is the consequence
of electrostatic interactions between 6 and polymer backbone
phosphates. At higher r ratios (r < 0.05) the intercalation of 6
in poly dAdT prevails, causing an increase in its fluorescence
emission. Binding constants Ks and n values (r < 0.1) deter-
0.16
0.16
0.09
0.21
0.34
poly dA–poly dT
poly dAdT–poly dAdTc
poly dGdC–poly dGdCc
Calf thymus DNAc
6.1
6.1b
6.4b
a Accuracy of n 10–30%, consequently log Ks values vary in the same
order of magnitude. b Cumulative value due to unknown contribution
of non-intercalative mode of binding at ratios r ([6] : [polynucleotide])
> 0.2. c High ionic strength buffers used: pH = 5, citric acid buffer,
I = 0.13 mol dmϪ3; pH = 7, buffer Na cacodylate, I = 0.12 mol dmϪ3
.
with the larger aromatic surface and somewhat stronger bind-
ing of 6 compared to EB.
mined at different ionic strengths (I = 0.01 and 0.1 mol dmϪ3
)
were practically the same (log Ks/n: 5.6/0.14 and 5.7/0.1, respect-
ively) in accord with low sensitivity of intercalative binding to
ionic strength.
Interactions of 6 with double-stranded RNA and DNA
polynucleotides
Fluorescence titrations. Binding constants for 6 and EB with
double-stranded DNA and RNA polymers (Table 2) are mostly
of the same order of magnitude and do not depend significantly
on pH, base composition or structural differences between
homo- and alternating polymers. A striking difference in fluor-
escence response is observed for 6–poly GC and 6–poly AU
complexes. Upon addition of poly GC the emission intensity of
6 decreased by 80% while poly AU yielded an increase of 100%
(Fig. 6). Similar effects are observed for DNA homopolymer
poly dAdT and alternating polymers poly (dAdT)2–poly
(dAdT)2 and poly (dGdC)2–poly (dGdC)2. Considering the dif-
ferent fluorescence properties of 6–AMP complex (fluorescent)
Somewhat larger binding constants were obtained from
titrations of 6 with poly GC, poly dGdC and calf thymus
(CT)-DNA (Table 2). With these polymers both non-
intercalative (electrostatic) and intercalative binding modes
seem to cause quenching of 6. At r > 0.2 already ca. 70% of
emission intensity is quenched, preventing accumulation of
sufficient data points for accurate calculations at r < 0.1 where
the intercalation binding mode is prevalent. Therefore the
calculated values of Ks and n are cumulative, comprising
both binding modes; the latter can also explain the values of
n being larger than those theoretically possible for inter-
calation.
1812
J. Chem. Soc., Perkin Trans. 2, 2001, 1808–1816

View Article Online
Fig. 8 Melting curves of poly dA–poly dT and 6 at low ionic strength I
and pH = 7.0. For measuring conditions see footnotes to Tables 3 and 4
and the Experimental section; the ligand : nucleic acid ratios r are: 0.0;
0.1; 0.2; 0.3; 0.5 (from left to right).
the three compounds and CT-DNA are 6.4 (6), 5.04 (1) and 5.8
(EB).
The melting profiles of poly dA–poly dT with 6 (supple-
mentary material,† Fig. 8) in the buffers of low ionic strength
(I ≈ 0.02 M) show at both pH = 5 and pH = 7 an unusual
behaviour absent with other diazapyrenium monointercalators
like 211 and also EB. The curves are clearly separated into two
transitions with both midpoints lying above the melting tem-
perature of the homopolymer-duplex alone; going from lower
to higher ligand : phosphate ratios r, the absorbance of the
second step becomes more and more pronounced until at r ≈ 0.7
(pH = 5) or r ≈ 0.8 (pH = 7) one sees more or less only one
transition. An increase of both denaturation temperatures is
seen with increasing ratio r in a constant difference between
each other of ≈ 10 ЊC. This is in contrast to the biphasic melting
with some other ligands, due to insufficient saturation
below r = 0.16, as was described earlier11 for poly dA–poly dT
(pH = 5). There the first midpoint increases faster than the
second, merging finally in one single curve. This distinct bi-
phasic behaviour of 6 with poly dA–poly dT below high ratios r
(in low ionic strength buffer) is in line with the fluorescence
titrations with the same polymer (Fig. 7A,B). The first transi-
tions in the melting profiles can be assigned to the intercalative
complex; the latter is supported by the comparison of the ∆T m
values at e.g. r = 0.3 for 6 (I transition 12.5/II transition 22.8 ЊC)
and for EB (14.0 ЊC). Therefore, the second transition should
correspond to the non-intercalative complex. However, at high
ionic strength both 6 and EB give monophasic melting profiles
with similar stabilisation (∆T m values at e.g. r = 0.3: 6 4.0 ЊC
and EB 3.0 ЊC, Tables 4 and 5).
As with CT-DNA, 6 also shows higher melting temperature
increases with poly dA–poly dT in low ionic strength buffer
compared to the other diazapyrenium monointercalators.11
Even at the lowest r the ∆T m is 5–8 times (6–10 ЊC) higher than
the ∆T m’s of 2 at the highest ratios (e.g. r = 0.3: 6, 12.5/22.8 ЊC
corresponding to 2, 2.7 ЊC); the second transition appears with
a difference of 15–20 ЊC. The observed large stabilisation by 6
can be attributed to the additional binding interactions involv-
ing the amino substituents being absent in 2. The same holds
for the interactions with the RNA model poly A–poly U at all
ratios (Table 3). The ∆T m values of 6 are about seven times
higher than the stabilising transitions in the profiles of the
Fig. 7 Fluorimetric titration of 6, c = 2 × 10Ϫ6 mol dmϪ3 with poly
dAdT, pH = 7, at I = 0.01 M (A) and I = 0.1 M (B).
UV/Vis titrations. Bathochromic shifts of λmax at 520 nm of
25–35 nm and hypochromic effect of up to 50–60% observed
in titrations of 6 with all examined DNA and RNA double-
stranded polymers strongly suggest intercalation of 6 as the
dominant binding mode.22
Melting temperature UV studies with 6
For CT-DNA and poly dA–poly dT, the melting point increases
(∆T m) caused by 6 (in low ionic strength buffers I ≈ 0.02 M)
do not show significant pH dependence between pH 5 and 7
within the error limit ( 0.5 ЊC). This is consistent with the
almost identical melting point increases of poly dA–poly dT
with EB at pH = 5 (r = 0.3: 14.0 ЊC ) and pH = 7 (r = 0.3: 13.5 ЊC )
(at low ionic strength) and the stability constants of 6 with poly
dA–poly dT in both buffers (log Ks = 5.4, pH = 5; 5.6, pH = 7)
and also poly (dAdT)2 (log Ks = 6.1, pH = 5; 5.9, pH = 7), here
at high ionic strength (I = 0.125 M).
A comparison of the melting point increases of CT-DNA by
6 and by its structurally closest analogue 211 at pH = 5.0 in low
ionic strength buffer shows at all ratios r about 5–7 ЊC higher
values of 6 with only one charge but two amino groups (e.g.
r = 0.3: 23.5 ЊC for 6 and 16.1 ЊC for 2). This is in line with the
binding constants of both ligands to CT-DNA (log Ks = 6.4 for
6 and 4.7 for 2). As it is known for EB,27 the amino groups can
give additional interactions with e.g. the 5Ј-oxygen atoms of the
phosphate backbone which is probably the reason for the
increased stabilisation. The diazapyrenium system of 6 with an
additional phenyl ring also causes a difference of 6 ЊC com-
pared to the phenanthridinium moiety of EB (r = 0.3: 17.3 ЊC;
at low ionic strength). The same feature of a better binding of 6
compared to 1 and 2 and to EB is also seen at high ionic
strength (pH = 5); 6 shows the ∆T m-values for CT-DNA are
more than twice as large as those for 111 or those for EB (e.g.
r = 0.3: 6, 6.5 ЊC; 1, 3.0 ЊC; and EB, 3.2 ЊC); the log Ks values of
J. Chem. Soc., Perkin Trans. 2, 2001, 1808–1816
1813

View Article Online
Table 3 ∆T m values (ЊC) of 6 in low ionic strength buffers pH = 5 and pH = 7a
pH
5
r=
0.1
0.2
0.3
0.5
28.8
15.7/26.7
43.6
CT-DNA
14.5
6.8/16.3
28.2
13.9
19.6
10.3/20.3
35.6
23.5
12.5/22.8
39.0
poly dA–poly dT
poly A–poly U
CT-DNA
7
19.4
23.2
>26.5
poly dA–poly dT
poly A–poly U
7.1/16.7
10.9/23.3
11.3/21.6
25.5
13.8/24.3
29.0
17.5/27.9
32.9
a r: molar ratio of ligand : nucleic acid phosphates; CT-DNA: calf thymus DNA; ionic strength I = 0.025 M (pH = 5.0) and I = 0.016 M (pH = 7);
error in ∆T m: 0.5 ЊC.
Table 4 ∆T m values (ЊC) of 6 in high ionic strength buffer pH = 5a
r=
0.1
0.2
0.3
CT-DNA
poly dA–poly dT
poly A–poly U
2.7
1.9
4.2
5.6
3.0
8.4
6.5
4.0
10.6
a See footnotes to Table 3, except: I = 0.125 M.
Table 5 ∆T m values (ЊC) of EB in low and high ionic strength buffers
pH = 5 and pH = 7.0a
pH
5
r = 0.3
I = 0.025 M
I = 0.125 M
CT-DNA
17.3
14.0
36.1
—
3.2
3.0
10.7
—
poly dA–poly dT
poly A–poly U
CT-DNA
7
poly dA–poly dT
poly A–poly U
13.5
29.1
—
—
Fig. 9 Helix length extension (L/L0) vs. ligand : DNA phosphate ratio
r; plot for 6 at pH = 5.0 (low ionic strength).
a See footnotes to Tables 3 and 4.
using the data for the first transition in the melting curves of
poly dA–poly dT attributed to the intercalation mode. At all r
ratios, the selectivity factor was ≈ 3 at pH = 5 and ≈ 2 at pH = 7.
Such a preference of 6 for the poly A–poly U polymer is com-
parable to that of EB (2.4)5 and to the other 4,9-diazapyrenium
compounds11(1.8–2.7).
related diazapyrenium derivative 2 (pH = 5, e.g. r = 0.3: 2, Ϫ4.0/
ϩ5.9 ЊC and 6, 39.0 ЊC). Ethidium bromide with two amino
substituents shows stabilisation similar to that of 6 in both low
and high ionic strength buffers (r = 0.3: EB, 36.1 ЊC compared
to 6, 39.0 ЊC at I = 0.025 M and at I = 0.125 M, EB, 10.7 ЊC and
6, 10.6 ЊC; Tables 3–5).
Viscometry
In contrast to the behaviour of all other examined 4,9-
diazapyrenium compounds towards poly A–poly U with one
destabilising and one stabilising step,11 6 gives only positive
values at both pH = 5 and pH = 7 (I ≈ 0.02 M). The poly
A–poly U denaturation curves with 6 (data not shown) exhibit
two phases only at low ratios (r < 0.1) which is due rather to
insufficient saturation than to two distinct binding modes (as
discussed above in the case of poly dA–poly dT). In contrast
to poly dA–poly dT, in the fluorescence titrations with poly
A–poly U a decrease of emission due to the non-intercalative
binding could not be observed. The observed non-intercalative
binding of 6 to poly dA–poly dT could be the consequence of
its more narrow and deeper minor groove compared to that of
poly A–poly U.28
In contrast to CT-DNA and to poly dA–poly dT, the ∆T m
values for 6 and poly A–poly U are about 5–11 ЊC higher at
pH = 5 than at pH = 7 (e.g. r = 0.3: 39.0 ЊC at pH = 5 and
29.0 ЊC at pH = 7, Table 3), although the stability constants
are comparable (log Ks = 6.5 at both pH, Table 2). It is known
that single-stranded poly A at acidic pH21,29 forms stable poly
AHϩ–poly AHϩ duplexes. During thermal denaturation of poly
A–poly U at pH 5, the poly A released is expected to form a
duplex to which the ligand can intercalate. The ∆T m then
reflects a mixed binding to poly A–poly U and poly AHϩ–poly
AHϩ. The pH dependence of the poly A–poly U melting
temperatures is also observed with EB (r = 0.3 : 36.1 ЊC at
pH = 5 compared to 29.1 ЊC at pH = 7, I ≈ 0.02 M, Table 5).
In order to evaluate RNA/DNA selectivity for 6, the ∆T m
poly A–poly U : ∆T m poly dA–poly dT ratio was calculated
Viscometric titrations of 6 with CT-DNA in low salt citric acid
buffer gave a slope of α = 1.1 (Fig. 9); this is close to that of EB
(α = 1.0)5,11 and consistent with monointercalation at ligand to
phosphate ratios r < 0.12.
Conclusions
The results demonstrate that relatively small structural vari-
ations in intercalating ligands can lead to dramatic differences
in their interactions with nucleic acids. This study also shows
how different methods such as UV/Vis, fluorescence and CD
measurements and melting analyses lead, when applicable, to a
consistent picture of binding differences.
Although the binding affinity of 6 toward nucleotides is
similar to that of EB and the previously studied 4,9-diaza-
pyrenium analogue 1, a smaller affinity of 2 toward nucleotides
is probably due to the larger hydration shell of this doubly
charged ligand.
The observed specific fluorescence response of 6 on binding
to double-stranded G–C (quenching) and A–U(T) (emission
increase) polymers could be of interest for development of
sensors for such polynucleotides.
In view of the bulky phenyl substituents placed at both ends
of the short 4,9-diazapyrenium axes of 6 it seems reasonable
to propose threading intercalation into double-stranded poly-
nucleotides as a main binding mode, with the intercalator and
adjacent base pairs’ long axes being parallel. Stability constants
of 6–ds polynucleotide complexes are similar to those of EB,
somewhat larger than found for analogue 1, and almost two
1814
J. Chem. Soc., Perkin Trans. 2, 2001, 1808–1816

View Article Online
orders of magnitude larger than those for the analogue 2. Such
a difference in affinity along with the significantly larger stab-
ilisation of the double strand by addition of 6 than found
for analogues 1 and 2 points to additive interactions of amino
substituents of 6 and EB with polynucleotides. Additional non-
intercalative (electrostatic) interactions of 6 with ds polymers
under conditions close to saturation of polymer (r > 0.2) are
more pronounced for DNA than for RNA derivatives.
Interactions of 6 with ss polynucleotides depend strikingly
on the nucleobase composition and in some cases on pH. The
complex of 6 with poly U formed close to saturation of poly
U (r ≈ 1) has a well organised, possibly helical structure (CD
evidence); it is much more stable than the other complex
dominating at large excess of poly U, where 6 binds on
“isolated” binding sites with an affinity similar to that of EB.
Until now none of the intercalators had exhibited such different
modes of binding exclusively on one ss polymer combined with
a specific change in chiral properties of the polynucleotide.
Fluorimetric properties of 6–poly C are found to be strongly
pH dependent, pointing to the sensitivity of the 6 chromophore
system to electron donor–acceptor surrounding conditions.
UV/Vis and fluorescence titrations were used to determine
the binding affinity of 6 and EB as reference intercalator
toward polynucleotides. In fluorimetric titrations excitation
wavelength of 385 nm was used and changes of emission were
monitored at 615 nm. Where possible, using the titration data
stability constants (Ks) and [bound intercalator] : [polynucleo-
tide phosphate] ratios (n) were calculated according to the
Scatchard equation30 by non-linear least-squares fitting
methods.11 Values for Ks and n given in Tables 1 and 2 are
results of calculations of absolute minima and all had satisfying
correlation coefficients (> 0.999).
Thermal melting curves for DNA, RNA and their complexes
were determined as previously described11 by following
the absorption change at 260 nm as a function of temperature.
The absorbance of the ligands was subtracted from every
curve, and the absorbance scale was normalised. T m values
are the midpoints of the transition curves, determined from
the maximum of the first derivative or graphically by a tangent
method.31 ∆T m values were calculated by subtracting T m of
the free nucleic acid from T m of the complex. Every ∆T m value
reported here was the average of at least two measurements;
the error in ∆T m is 0.5 ЊC.
Viscometric titrations were conducted in an Ubbelohde
micro-viscometer (Schott) as previously described.11 The con-
centration of CT-DNA in low salt citric acid buffer was
5 × 10Ϫ4 mol dmϪ3 in phosphates, the ligand to DNA phosphate
ratio r less than 0.15. The viscosity index α was obtained from
the flow times at varying r according to eqn. (1),32 where t0, tDNA
Experimental
1H-NMR spectra were recorded on Varian-Gemini 300 MHz
and Bruker 500 MHz Avance DRX instruments with tetra-
methylsilane as internal standard for organic solvents and
the acetone signal as external, or acetate buffer as internal,
standard for aqueous media. Chemical shifts (δ) are expressed
in ppm, with numbering of protons according to Fig. 2. Signal
multiplicities are denoted by s (singlet), d (doublet), t (triplet),
pt (pseudotriplet—formed from overlapping double doublet),
q (quartet) and m (multiplet). Electronic absorption spectra
were obtained on a Varian Cary 1 spectrometer using quartz
cuvettes (1 cm). Fluorescence spectra were recorded on a
Perkin-Elmer LS 50 fluorimeter. CD spectra were recorded on
JASCO J-715 spectropolarimeter in quartz cuvettes (1 cm),
total absorbance at the end of titration being Absmax = 1.5.
Electron spray mass spectra (ES-MS) were obtained using
a Varian MAT 711 spectrometer. For chromatographic purifi-
cation of the prepared compounds silica gel HF254 (Merck) for
preparative thin layer chromatography was used.
L/L0 = [(tr Ϫ t0)/(tDNA Ϫ t0)]1/3 = 1 ϩ α*r
(1)
and tr denote the flow times of buffer, free DNA and DNA
complex at ratio r, respectively:
Synthesis
The starting compound 2,7-diamino-5,10-diphenyl-4,9-diaza-
pyrene 3 was prepared by a published procedure.13 The Hygro-
scopic nature of the precipitates hampered elementary analysis.
However, compounds 5 and 6 were prepared according to a
previously well studied procedure,10 structures being confirmed
by (1D and 2D) NMR and ES-MS data.
2,7-Bis(benzyloxycarbonylamino)-5,10-diphenyl-4,9-diaza-
pyrene 4. Starting compound 3 (52 mg, 0.135 mmol) and
NaHCO3 (36 mg, 0.4 mmol) were suspended in dry DMF
(2 ml). To a cooled suspension (0–5 ЊC) a 50% solution of ben-
zyloxycarbonyl chloride was added (0.14 ml, 4 mmol) while
stirring. After 2 hours of stirring at room temperature 5 ml of
dry diethyl ether were added, and the precipitate was collected
and washed with water. Recrystallization from hot methanol
gave 4 (50–61% yield, mp. 280–282 ЊC). 1H NMR (δH/ppm,
DMSO-d6): 5.23 (s, 4H, 2 × OCH2), 7.36–7.49 (m, 10Hbenzyl),
7.67–7.69 and 7.92–7.94 (2m, 10Hphenyl), 8.74 and 8.75 (2s, 4H,
C1-H, C3-H, C6-H, C8-H), 10.55 (s, 2H, 2 × NH).
Solutions
Low ionic strength citric acid buffer contained 0.01 mol dmϪ3
citric acid, adjusted to pH = 5 (I = 0.025 mol dmϪ3); low ionic
strength PIPES buffer contained 0.01 mol dmϪ3 PIPES
(piperazine-1,4-bis(2-ethanesulfonic acid)), adjusted to pH = 7
(I = 0.016 mol dmϪ3). High ionic strength citric acid buffer
contained 0.01 mol dmϪ3 citric acid and 0.1 mol dmϪ3 NaCl,
adjusted to pH = 5 (I = 0.125 mol dmϪ3). For NMR experi-
ments acetate buffer (CD3COOD–NaOD; 0.05 M; pD = 5.4)
was used.
Polynucleotides were purchased as noted: poly A, poly C,
poly U, poly dA–poly dT, poly (dAdT)2–poly (dAdT)2 and poly
(dGdC)2–poly (dGdC)2 (Pharmacia); poly G, poly G–poly C,
poly A–poly U (Sigma), calf thymus (CT)-DNA (Aldrich).
Polynucleotides were dissolved in the respective buffer and
their concentration determined spectroscopically11 as the con-
centration of phosphates.
2,7-Diamino-5,10-diphenyl-4-methyl-4,9-diazapyrenium tri-
flate 5. A solution in chlorobenzene (6 ml) of 4 (100 mg,
0.15 mmol) and methyl triflate (1.6 ml, 0.15 mmol) was heated
under reflux for 2 hours. After cooling to room temperature
triflic acid (0.05 ml, 0.5 mmol) was added to the dark red
suspension and upon treatment of the oily residue in an ultra-
sound bath, the suspension was stirred overnight. Chloro-
benzene was decanted and to the oily residue 5 ml of dry diethyl
ether were added.The blue precipitate was collected, washed
with dry ether and purified by preparative thin layer chrom-
atography on silica gel using 10% methanol in dichloromethane
for elution. Evaporation of solvent and recrystallization from
acetonitrile gave 5 (40–50% yield).1H NMR(δH/ppm, CD3CN):
4.3 (s, 3H, CH3), 5.16 (s, 2H, C7-NH2), 5.52 (s, 2H, C2-NH2),
Methods
All titration data were corrected for dilution, significant ones
are available as supplementary material. Stability constants (Ks)
of 6–nucleotide complexes were calculated by processing fluori-
metric titration data with the program SPECFIT.18
Determination of self-stacking by 1H-NMR experiments was
done by changing the c(6) in acetate buffer. 2D NMR experi-
ments were done in D2O with acetone as external standard.
J. Chem. Soc., Perkin Trans. 2, 2001, 1808–1816
1815

View Article Online
9 C. Coudret and A. Harriman, J. Chem. Soc., Chem. Commun., 1992,
1755; H.-C. Becker and B. Norden, J. Am. Chem. Soc., 1997, 119,
5798; A. Slama-Schwok, M. Rougee, V. Ibanez, N. E. Geacintov,
T. Montenay-Garestier, J.-M. Lehn and C. Helene, Biochemistry,
1989, 28, 3234; H.-C. Becker and B. Norden, J. Am. Chem. Soc.,
1999, 121, 11947; H.-C. Becker and B. Norden, J. Am. Chem. Soc.,
2000, 122, 8344.
7.18 (d, J67 = 1.9 Hz, 1H, C6-H), 7.72–7.75 and 7.87–7.94 (2m,
10Hphenyl), 8.0 and 8.15 (2s, 3H, C1-H, C3-H, C8-H); ES-MS
(m/z) 401.1 (Mϩ), CF3SO3Ϫ 148.7 (MϪ).
2,7-Diamino-5,10-diphenyl-4-methyl-4,9-diazapyrenium
hydrogensulfate 6. To a solution of triflate 5 (16 mg, 0.03 mmol)
in acetonitrile (1 ml), an acetonitrile solution of tetrabutyl-
ammonium hydrogensulfate (175 mg, 5 mmol) was added.
After standing for 30 min at room temperature, the precipitate
formed is collected and washed with 3 × 0.5 ml of dry dichloro-
methane followed by recrystallization from DMSO–dichloro-
methane giving a dark blue precipitate of 6 (76% yield).1H
NMR(δH/ppm, D2O, c(6) = 5 × 10Ϫ3 mol dmϪ3): 3.89 (s, 3H,
CH3), 6.60 (s, 1H, C6-H), 6.85 (d, 2H, H4phenyl), 6.87 (s, 1H, C8-
H), 6.96 (d, 2H, H1phenyl), 7.13 (s, 1H, C1-H), 7.38 (pt, 2H,
H5phenyl), 7.49 (s, 1H, C3-H), 7.54 (pt, 2H, H6phenyl), 7.61 (pt,
2H, H2phenyl), 7.78 (pt, 2H, H3phenyl); ES-MS (m/z) 401.1 (Mϩ),
HSO4Ϫ 96.8 (MϪ).
˘
10 I. Piantanida, V. Tomis˘ic´ and M. Zinic´, J. Chem. Soc., Perkin Trans.
2, 2000, 375.
˘
11 B. S. Palm, I. Piantanida, M. Zinic´ and H.-J. Schneider, J. Chem.
Soc., Perkin Trans. 2, 2000, 385.
12 (a) I. Steiner-Bioc˘ic´, Lj. Glavas˘-Obrovac, I. Karner, I. Piantanida,
˘
M. Zinic´, K. Pavelic´ and J. Pavelic´, Anticancer Res., 1996, 16, 3705;
˘
(b) S. Roknic´, Lj. Glavas˘-Obrovac, I. Karner, I. Piantanida, M. Zinic´
and K. Pavelic´, Chemotherapy, 2000, 46, 143.
13 A. E. S. Fairfull, D. A. Peak, W. F. Short and T. I. Watkins, J. Chem.
Soc., 1952, 4700.
14 H.-J. Schneider and A. Yatsimirsky, Principles and Methods in
Supramolecular Chemistry, Wiley, Chichester, 1999.
15 D. B. Davies, L. N. Djimant and A. N. Veselkov, J. Chem. Soc.,
Faraday Trans., 1996, 92, 383.
16 J. Olmsted III and D. R. Kearns, Biochemistry, 1977, 16, 3647.
17 M. G. Badea and S. Georghiou, Photochem. Photobiol., 1976, 24,
417; S. Georghiou, Photochem. Photobiol., 1977, 26, 59.
18 Specfit Global Analysis, a Program for Fitting, Equilibrium and
Kinetic Systems, using Factor Analysis & Marquardt Minimization
M. Maeder and A. D. Zuberbuehler, Anal. Chem., 1990, 62, 2220;
H. Gampp, M. Maeder, C. J. Meyer and A. D. Zuberbuehler,
Talanta, 1985, 32, 257.
Acknowledgements
This work was carried out largely in the Rudjer Boskovic
Institute, Zagreb, and partially in Saarbrücken, and was
supported by the Deutsche Forschungsgemeinschaft, Bonn
(Grants 436KRO113/3 and Schn115/17-1). Dr S. Simova is
thanked for help with the NMR spectra.
19 A. Odani, H. Masuda and O. Yamauchi, Inorg. Chem., 1991, 30,
4484.
˘
˘
20 P. Cudic´, M. Zinic´, V. Tomis˘ic´, V. Simeon, J.-P. Vigneron and J.-M.
Lehn, J. Chem. Soc., Chem. Commun., 1995, 1073.
References
21 C. R. Cantor, P. R. Schimmel, Biophysical Chemistry, Freeman, San
Francisco, 1980, Vol. 3, pp. 1109–1181.
22 G. Dougherty and J. R. Pilbrow, Int. J. Biochem., 1984, 16, 1179.
23 G. D. Fasman, Circular Dichroism and the Conformational Analysis
of Biomolecules, Plenum Press, New York, 1996, Ch. 12, 18.
24 H.-C. Becker and B. Norden, J. Am. Chem. Soc., 1997, 119, 5798 and
refs. cited therein.
1 See e. g. W. D. Wilson, in Nucleic Acids in Chemistry and Biology,
M. Blackburn, M. Gait, Eds.; IRL Press, Oxford, 2nd edition, 1996,
Ch. 8.
2 (a) N. T. Thuong and C. Helene, Angew. Chem., 1993, 32, 697; N. T.
Thuong and C. Helene, Angew. Chem., Int. Ed. Engl., 1993, 32, 666;
(b) E. Kool, Chem. Rev., 1997, 97, 1473.
3 Selected reviews: A. Blasko and T. Bruice, Acc. Chem. Res., 1999, 32,
475; N. H. Williams, B. Tkasaki, M. Wall and J. Chin, Acc. Chem.
Res., 1999, 32, 485; J. A. Cowan, Chem. Rev., 1998, 3, 1067; E. L.
Hegg and J. N. Burstyn, Coord. Chem. Rev., 1998, 173, 133;
E. Kimura and T. Koike, Adv. Inorg. Chem., 1997, 44, 229.
4 J. Fiel, B. G. Jenkins and J. L. Alderfer, Molecular Basis of
Specificity in Nucleic Acid–Drug Interactions, B. Pullman and
J. Jortner, Eds.; Kluwer Academic Publishers, Dordrecht, 1990,
vol. 23.
25 J. Ramstein and M. Leng, Biochim. Biophys. Acta, 1972, 281, 18.
26 M. K. Pal and J. K. Ghosh, Spectrochim. Acta, Part A, 1995, 51,
489; J. B. LePecq and C. Paoletti, J. Mol. Biol., 1967, 27, 87.
27 J. M. Veal and W. D. Wilson, J. Biomol. Struct. Dyn., 1991, 8, 1119.
28 W. D. Wilson, Y.-H. Wang, C. R. Krishnamoorthy and J. C. Smith,
Biochemistry, 1985, 24, 3991.
29 W. Saenger, Principles of Nucleic Acid Structure, Springer, New
York, 1988, p. 305.
30 G. Scatchard, Ann. N. Y. Acad. Sci., 1949, 51, 660; J. D. McGhee and
P. H. von Hippel, J. Mol. Biol., 1974, 86, 469; J. D. McGhee and
P. H. von Hippel, J. Mol. Biol., 1976, 103, 679.
5 W. D. Wilson, L. Ratmeyer, M. Zhao, L. Strekowski and D. Boykin,
Biochemistry, 1993, 32, 4098.
6 A. J. Blacker, J. Jazwinski and J.-M. Lehn, Helv. Chim. Acta,
1987, 70, 1; A. Slama-Schwok, J. Jazwinski, A. Bere, T. Montenay-
Garestier, M. Rougee, C. Helene and J.-M. Lehn, Biochemistry,
1989, 28, 3227.
7 J. Kapuscinski and Z. Darzynkiewicz, J. Biomol. Struct. Dyn., 1987,
5, 127; M. Dourlent and C. Helene, Eur. J. Biochem., 1971, 23, 86.
8 A. J. Blacker, J. Jazwinski, J.-M. Lehn and F. X. Wilhelm, J. Chem.
Soc., Chem. Commun., 1986, 10, 35.
31 W. D. Wilson, F. A. Tanious, M. Fernandez-Saiz, C. T. Rigl,
Evaluation of Drug/Nucleic Acid Interactions by Thermal Melting
Curves from: Methods in Molecular Biology, Vol. 90; Drug–DNA
Interaction Protocols, ed. K. R. Fox, Humana Press Inc., Totowa,
NY, 1998.
32 G. Cohen and H. Eisenberg, Biopolymers, 1969, 845; M. Wirth,
O. Buchardt, T. Koch, P. E. Nielsen and B. Nordén, J. Am. Chem.
Soc., 1988, 110, 93.
1816
J. Chem. Soc., Perkin Trans. 2, 2001, 1808–1816