Minor groove intercalation of Δ-[Ru(Me2phen)2dppz]2+ to the hexanucleotide d(GTCGAC)2

Article abstract of DOI:10.1039/b105689c

The metal complex Δ-[Ru(Me₂phen)₂dppz]²⁺ (Me₂phen = 2,9-dimethyl-1,10-phenanthroline; dppz = dipyrido-[3,2-a:2′,3′-c]phenazine) has been synthesised and its binding to the hexanucleotide d(GTCGAC)₂ studied by ¹H NMR spectroscopy. The crystal structure of rac-[Ru(Me₂phen)₂dppz](PF₆)₂ ·2H₂O has been determined. Addition of Δ-[Ru(Me₂phen)₂dppz]²⁺ to d(GTCGAC)₂ induced significant broadening of the Me₂phen and dppz resonances and large upfield shifts of the dppz resonances, consistent with the metal complex binding through intercalation of the dppz ligand. Also indicative of intercalation are the observed upfield shifts of the T₂ and G₄ imino protons and the 173°C increase in the melting temperature of the hexanucleotide upon addition of the metal complex. NOE cross-peaks from the Me₂phen protons were only observed to hexanucleotide minor groove H1′ and H4′/H5′/H5″ protons in NOESY spectra of the hexanucleotide with added metal complex. In addition, NOEs were also observed between the H13/14 protons of the dppz ligand and the hexanucleotide major groove T₂methyl and sugar H2′/H2″ protons. From the combined NMR data it is concluded that the dppz-based ruthenium(II) complex Δ-[Ru(Me₂phen)₂dppz]²⁺ intercalates from the minor groove of the hexanucleotide mini-duplex d(GTCGAC)₂.

Full text of DOI:10.1039/b105689c

View Article Online / Journal Homepage / Table of Contents for this issue
Minor groove intercalation of ꢀ-[Ru(Me2phen)2dppz]2؉ to the
hexanucleotide d(GTCGAC)2 †
Antun Greguric,a,b Ivan D. Greguric,b Trevor W. Hambley,c Janice R. Aldrich-Wright*b and
J. Grant Collins*a
a School of Chemistry, University College, University of New South Wales,
Australian Defence Force Academy, Canberra, 2600, Australia. E-mail: g-collins@adfa.edu.au
b Department of Chemistry, University of Western Sydney, Macarthur, PO Box 555
Campbelltown, NSW, Australia
c Centre for Heavy Metals Research, School of Chemistry, University of Sydney, Sydney, NSW,
2006, Australia
Received 28th June 2001, Accepted 14th January 2002
First published as an Advance Article on the web 4th February 2002
The metal complex ∆-[Ru(Me2phen)2dppz]2ϩ (Me2phen = 2,9-dimethyl-1,10-phenanthroline; dppz = dipyrido-
[3,2-a:2Ј,3Ј-c]phenazine) has been synthesised and its binding to the hexanucleotide d(GTCGAC)2 studied by 1H
NMR spectroscopy. The crystal structure of rac-[Ru(Me2phen)2dppz](PF6)2ؒ2H2O has been determined. Addition
of ∆-[Ru(Me2phen)2dppz]2ϩ to d(GTCGAC)2 induced significant broadening of the Me2phen and dppz resonances
and large upfield shifts of the dppz resonances, consistent with the metal complex binding through intercalation of
the dppz ligand. Also indicative of intercalation are the observed upfield shifts of the T2 and G4 imino protons and
the 17 ЊC increase in the melting temperature of the hexanucleotide upon addition of the metal complex. NOE cross-
peaks from the Me2phen protons were only observed to hexanucleotide minor groove H1Ј and H4Ј/H5Ј/H5Љ protons
in NOESY spectra of the hexanucleotide with added metal complex. In addition, NOEs were also observed between
the H13/14 protons of the dppz ligand and the hexanucleotide major groove T2methyl and sugar H2Ј/H2Љ protons.
From the combined NMR data it is concluded that the dppz-based ruthenium() complex ∆-[Ru(Me2phen)2dppz]2ϩ
intercalates from the minor groove of the hexanucleotide mini-duplex d(GTCGAC)2.
by linear dichroism that [Ru(phen)3]2ϩ (which has been shown
Introduction
to bind from the minor groove)19 was capable of semi- or quasi-
The use of inert transition metal complexes for DNA recog-
intercalation, and that the binding geometry of [Ru(phen)3]2ϩ
nition and DNA mediated electron transfer has attracted
was similar to that of [Ru(phen)2dppz]2ϩ.20
considerable interest over the past decade, and has been the
In the absence of a crystal structure, NMR spectroscopy is
subject of a variety of recent review articles.1–3 Of the DNA
the best technique available to characterize the DNA binding of
binding complexes that have been studied, those based upon the
complexes such as [Ru(phen)2dppz]2ϩ. However, NMR studies
dipyrido[3,2-a:2Ј,3Ј-c]phenazine (dppz) ligand have probably
of oligonucleotide binding by dppz complexes have been
difficult, due to: the lack of DNA binding sequence specificity;
attracted the most interest. Complexes such as [Ru(phen)2-
dppz]2ϩ (phen = 1,10-phenanthroline) bind DNA strongly
broad resonances (due to intermediate exchange kinetics); and
through intercalation of the dppz ligand.4–6 Furthermore, the
the abundance of overlapping resonances in the aromatic
luminescent properties of the dppz-based complexes have led to
region. Dupureur and Barton have used selective deuteration of
their application as DNA ‘light switches’ and probes for long
the phen and dppz ligands of ∆- and Λ-[Ru(phen)2dppz]2ϩ to
range DNA mediated electron transfer studies.7–13 Despite the
simplify the spectra of the metal complex bound to a hexa-
considerable interest and the number of studies that have used
nucleotide.14,15 This approach resulted in the assignment of
dppz-based complexes, the mode of DNA binding remains
unresolved and the centre of controversy.
several NOE cross-peaks from the ∆-enantiomer to the hexa-
nucleotide major groove protons, thereby suggesting that
From both photophysical studies and NMR data, Barton
the metal complex intercalates from the major groove.15 No
and co-workers have proposed that [Ru(phen)2dppz]2ϩ inter-
intermolecular NOEs between the hexanucleotide and the
calates from the DNA major groove.14–16 Alternatively, Nordén
Λ-[Ru(phen)2dppz]2ϩ were detected.15
and co-workers, on the basis of the similarity of the binding
In an attempt to observe a greater number of metal complex–
geometry to that of the proven minor groove intercalating agent
hexanucleotide intermolecular NOEs, so as to further charac-
actinomycin D and photophysical studies using T4-DNA, have
terize the DNA binding, we have used the 2,9-dimethyl-
proposed that [Ru(phen)2dppz]2ϩ intercalates from the minor
1,10-phenanthroline analogue of ∆-[Ru(phen)2dppz]2ϩ, i.e.
groove.17,18 Furthermore, Lincoln and Nordén, demonstrated
∆-[Ru(Me2phen)2dppz]2ϩ, as shown in Fig. 1. The incorpor-
ation of the methyl groups onto the phenanthroline ligands has
† Electronic supplementary information (ESI) available: a figure show-
ing the aromatic region of the hexanucleotide with added ∆-[Ru(Me2-
phen)2dppz]2ϩ at a metal complex:duplex ratio of 0.9 (A) and 1.8 (B).
For the crystal structure of [Ru(Me2phen)2dppz]2ϩ the positional
atomic coordinates, bond lengths, and bond angles and details of
least-squares planes calculations are presented in Tables S1 to S4. See
http://www.rsc.org/suppdata/dt/b1/b105689c/
three advantages. (1) The methyl groups provide strong signals
(compared to the aromatic protons) from which intermolecular
NOEs can be observed. (2) The aromatic region is simplified,
due to the reduction of the number of aromatic resonances and
the degree of J-coupling that is observed. (3) The assignment of
the resonances from the bound metal complex is facilitated.
DOI: 10.1039/b105689c
J. Chem. Soc., Dalton Trans., 2002, 849–855
This journal is © The Royal Society of Chemistry 2002
849

View Article Online
fraction was added water (20 mL) to yield fine orange needles
1
(0.99 g, 76%). H NMR (d6-acetone): δ 9.60 (d, J = 8.10 Hz,
1H), 8.94 (d, J = 8.42 Hz, 1H), 8.50 (d, J = 8.10 Hz, 1H), 8.47
(d, J = 8.42 Hz, 1H), 8.43 (dd, J = 6.6, 3.48 Hz, 2H), 8.31
(d, J = 8.80 Hz, 1H), 8.15 (dd, J = 6.6, 3.48 Hz, 2H), 8.04 (d,
J = 8.42 Hz, 1H), 7.88 (d, J = 5.60 Hz, 1H), 7.72 (dd, J = 8.15,
5.4 Hz, 1H), 7.44 (d, J = 8.24 Hz, 1H), 2.18 (s, 3H), 2.11 (s, 3H).
Anal. calc. for [Ru(Me2phen)2dppz](PF6)2ؒH2O: C, 49.88; H,
3.28; N, 10.11%. Found: C, 50.26; H, 3.15; N, 9.81%.
Crystallography
Structure determination. Rac-[Ru(Me2phen)2dppz](PF6)2ؒ
2H2O: C46H38F12N8O2P2Ru, M = 1125.86, monoclinic, space
group P21/n (no. 14), a
= 14.5556(8), b = 16.8090(9),
Fig. 1 Structure and atom numbering of ∆-[Ru(Me2phen)2dppz]2ϩ
.
c = 19.300(1) Å, β = 100.736(1)Њ, V = 4639.4(4) Å, Dc = 1.612 g
cmϪ3, Z = 4, crystal size 0.35 × 0.11 × 0.09 mm, orange needle,
λ(MoKα) = 0.71073 Å, µ(MoKα) = 5.05 cmϪ1, 2θmax = 56.6Њ, T
= 294 K, Nind = 11393 (Rmerge = 0.022), Nobs = 7861 [I > 2.5 σ(I )],
Nvar = 658, residuals R(F) = 0.050, Rw(F) = 0.049, GoF(all) =
The phenanthroline 2-methyl group provides a link to the dppz
ligand due to its closeness to the H10 proton, compared with
the corresponding distance from the 9-methyl group. This
allows the assignment of the Me2phen and the dppz resonances.
In addition, while the methyl groups will affect the intercalation
process to some degree, simple molecular modelling suggests
that the [Ru(Me2phen)2dppz]2ϩ complex can still fully inter-
calate. In this paper we report a 1H NMR study of the binding
of ∆-[Ru(Me2phen)2dppz]2ϩ to the hexanucleotide d(GTC-
GAC)2. As ruthenium() dppz complexes show no sequence
selectivity in their DNA binding,15 it was not possible to study
the binding of ∆-[Ru(Me2phen)2dppz]2ϩ bound to an oligo-
nucleotide at a single site. Consequently, to aid comparison
with earlier NMR studies of the binding of [Ru(phen)2dppz]2ϩ
and closely related complexes to d(GTCGAC)2,14,15,21 we have
used the same hexanucleotide to study the DNA binding of
3.96, ∆ρmin,max = Ϫ0.46, 0.68 e ÅϪ3
.
Data collection, structure solution and refinement. Data were
collected on a Bruker SMART 1000 CCD diffractometer. The
data integration and reduction were undertaken with SAINT
and XPREP,23 and absorption corrections were applied using
SADABS.23 The data reduction included the application of
Lorentz and polarisation corrections. The structure was solved
by direct methods using SHELXS-8624 and refined using full-
matrix least-squares methods with teXsan.25 Hydrogen atoms
were included at calculated sites with isotropic thermal param-
eters based on that of the riding atom. Non-hydrogen atoms
were refined anisotropically except for the partially occupied
solvent sites (assigned as oxygen atoms with occupancies fixed
on the basis of observed peak intensities) and the minor sites of
∆-[Ru(Me2phen)2dppz]2ϩ
.
Ϫ
Experimental
a rotationally disordered PF6 anion. Neutral atom scattering
factors were taken from International Tables.26 Anomalous
dispersion effects were included in Fc;27 the values for ∆fЈ
and ∆fЉ were those of Creagh and McAuley.28 The values for
the mass attenuation coefficients are those of Creagh and
Hubbell.29 All other calculations were performed using the
teXsan25 crystallographic software package of Molecular
Structure Corporation.
Materials
The hexanucleotide d(GTCGAC)2 was obtained from Gene-
Works, South Australia. Ruthenium() chloride hydrate, 1,10-
phenanthroline, 2,9-dimethyl-1,10-phenanthroline hydrate,
potassium hexafluorophosphate, aluminium oxide, activated,
neutral, Brockmann I, Amberlite IRA-400(Cl) ion exchange
resin, D2O (99.96%) and dibenzoyl--tartaric acid were
obtained from Aldrich Chemical Company.
An ORTEP30 plot is shown in Fig. 2. Positional atomic
Ligand synthesis
The dppz ligand was prepared by the method of Dupureur and
Barton.15 A mixture of 1,10-phenanthroline-5,6-dione22 (1.0 g,
4.8 mmol) and 1,2-phenylenediamine (0.6 g, 5.7 mmol) in
ethanol (400 mL) was stirred for 2 h at 50 ЊC and then at room
temperature overnight. The resulting solution was reduced
in volume by rotary evaporation at 50 ЊC to yield a cream-
coloured product. The crude product was left to stand for 8 h,
methanol–water (10:90) was then added, and the product was
filtered and recrystallized from methanol to give a cream solid.
Yield: 0.96 g, 80%.
Fig. 2 ORTEP plot (30% thermal ellipsoids) with partial atom
numbering scheme.
Metal complex synthesis
[Ru(Me2phen)2Cl2] was synthesised as previously described.21
coordinates, bond lengths, and bond angles and details of least-
squares planes calculations are presented in Tables S1–S4 of the
ESI.† Observed and calculated structure factors, thermal
parameters for non-hydrogen atoms, positional and thermal
parameters of the hydrogen atoms, torsion angles, and close
inter- and intra-molecular contacts are available upon request
from the author.
CCDC reference number 168297.
See http://www.rsc.org/suppdata/dt/b1/b105689c/ for crystal-
lographic data in CIF or other electronic format.
[Ru(Me2phen)2dppz]2ϩ was prepared by refluxing 1.0
g
of [Ru(Me2phen)2Cl2] and 0.52 g (1.2 equiv) of dppz in 500 mL
of 75% ethanol for 5 h. The volume was then reduced (50 mL),
the solution cooled, and excess KPF6 added. The resultant
orange precipitate was filtered and washed with water (100 mL)
and then diethyl ether (50 mL). The solid was dissolved in
acetonitrile (20 mL) and applied to the head of a column
(5 × 30 cm) of activated aluminum oxide (neutral Brockmann
1). The orange band was eluted with acetonitrile, and to this
850
J. Chem. Soc., Dalton Trans., 2002, 849–855

View Article Online
Enantiomer resolution
NOESY spectra and the observation of imino protons from the
non-terminal base pairs indicated that the hexanucleotide
adopted a B-type mini double helix in solution.
The enantiomers of [Ru(Me2phen)2dppz]2ϩ were resolved on a
Sephadex SP-C 25 column (40 × 2.5 cm) using 0.1 M disodium
dibenzoyl--tartrate, 0.2 M sodium chloride, and 20% acetone
as the eluent. The enantiomeric purity of the chloride salt was
assayed by CD spectroscopy, with the ∆-enantiomer displaying
negative circular dichroism at 310 and 464 nm.
The resonances from both the metal complex and hexa-
nucleotide become significantly broadened upon addition of
the metal complex to the hexanucleotide at 25 ЊC, indicating
intermediate exchange binding kinetics. At lower temperatures
(<10 ЊC) slow exchange binding kinetics are observed for
some resonances, in agreement with the [Ru(phen)2dppz]2ϩ–
hexanucleotide binding results of Dupureur and Barton.14,15
However, at temperatures below 10 ЊC the resonances are
extremely broad. At 45 ЊC the spectrum of d(GTCGAC)2 with
added ∆-[Ru(Me2phen)2dppz]2ϩ is sufficiently resolved so that
the resonances from both the hexanucleotide and the metal
complex can be assigned (see Fig. 3).
Sample preparation for NMR analysis
The hexanucleotide was dissolved in 0.65 mL of phosphate
buffer (10 mM, pH 7) containing 20 mM NaCl, 0.1 mM EDTA
and a trace of DSS as an internal chemical shift reference. For
experiments carried out in D2O the sample was repeatedly
freeze dried from D2O and finally made up in 99.96% D2O. The
hexanucleotide concentration was determined from the A260
absorbance using an extinction coefficient of 6600 MϪ1 cmϪ1
per nucleotide.31
Instrumental methods
400 MHz 1H NMR spectra were recorded on a Varian
Unityplus-400 spectrometer. NOESY spectra were recorded
by the method of States et al.,32 using 2048 data points in t2 for
256 t1 values with a pulse repetition delay of 1.7 s. DQFCOSY
experiments were accumulated using 2048 data points in t2 for
256–310 t1 values with a pulse repetition delay of 1.7 s. Spectra
recorded in 90% H2O–10% D2O were collected using the
WATERGATE solvent suppression technique of Piotto et al.33
Circular dichroism spectra (CD) were recorded at ambient
temperature on a Jasco 500C spectropolarimeter.
Molecular modelling
Fig. 3 1H NMR spectra of the aromatic region of the free hexanucleo-
tide (A), the hexanucleotide (1.4 mM) with added ∆-[Ru(Me2phen)2-
dppz]2ϩ at a metal complex:duplex ratio of 0.9 (B) and the free metal
complex (C). All spectra were obtained in the same aqueous buffer
[10 mM phosphate (pH 7) containing 20 mM NaCl and 0.1 mM EDTA]
at 45 ЊC.
The coordinates for the metal complex were taken from the
crystal structure. The hexanucleotide model was constructed
using HyperChem molecular modelling software34 and the
metal complex was manually docked. Energy minimisation by
Polak–Ribiere conjugate-gradient refinement was carried out
with the metal complex treated as a rigid group, a sound
approximation given the rigidity of each of the ligands and the
strained nature of the complex. Models were investigated with
the metal complex intercalated between each of the unique
base-pair combinations and a variety of orientations of the
complex were used as starting points. Minimisation was con-
tinued until the RMS energy gradient was less than 0.2 kJ ÅϪ1
The resonances from the bound metal complex were assigned
by a combination of two-dimensional NMR experiments. The
relative closeness of the H10 proton to the 2-methyl compared
with the 9-methyl group allows the assignment of these protons
in a NOESY spectrum. Once the methyl groups have been
specifically assigned the remaining Me2phen protons are easily
assigned through their various connectivities in NOESY and
DQFCOSY spectra. Similarly, given the assignment of the H10
resonance the H11 and H12 resonances can then be assigned
through their respective coupling in a DQFCOSY spectrum.
The dppz H13/H14 resonances can then be non-specifically
assigned through their coupling in a DQFCOSY experiment.
Although some of the hexanucleotide resonances still exhibited
considerable exchange broadening at 45 ЊC, nearly all the
resonances from the bound d(GTCGAC)2 could be assigned
through their connectivities in NOESY and DQFCOSY
spectra.36–38
molϪ1
.
Results
Rac-[Ru(Me2phen)2dppz](PF6)2ؒ2H2O crystal structure
The crystal structure analysis confirms that simultaneous
coordination of two Me2phen ligands and the dppz ligand is
possible but there is evidence of significant steric strain arising
from the bulk introduced by the methyl substituents. For
instance, the Ru–N(Me2phen) [2.092(3)–2.125(3) Å] and Ru–
N(dppz) [2.082(3)–2.084(3) Å] bond lengths are significantly
longer than the Ru–N(phen) [2.065(6)–2.073(6) Å] and Ru–
N(dpq) [2.043(5)–2.063(5) Å] bonds lengths observed in the
ꢀ-[Ru(Me2phen)2dppz]2؉–d(GTCGAC)2 interactions –
one-dimensional NMR
closely related complex [Ru(phen)2dpq]2ϩ 35
In addition, the
.
Me2phen ligands are significantly bowed as a result of inter-
actions between the methyl groups and the nitrogen donor
atoms of adjacent ligands. Deviations from the least-squares
planes are in the range of 0.18 Å for both of these ligands. The
methyl groups adjacent to the dppz ligand do not protrude
toward the dppz as far as the nitrogen donors but will impact
somewhat on the extent to which the complex can intercalate.
The metal complex was initially titrated into the hexanucleotide
solution to a metal complex:hexanucleotide duplex ratio of 0.9,
at 25 ЊC. Owing to the overlap and broadness of the resonances,
it was not possible to determine the chemical shift of most of
the resonances from the metal complex and hexanucleotide at
each point in the titration. However, upon titration of the metal
complex up to a metal complex:hexanucleotide ratio of 1.8, at
35 and 45 ЊC (see ESI†), it was observed that the chemical shifts
of the resonances from the metal complex remained constant
( 0.01 ppm). This suggests that the metal complex predomin-
antly exists in the hexanucleotide-bound form throughout the
Assignment of the 1H NMR resonances
The 1H NMR spectrum of the free d(GTCGAC)2 was assigned
by standard techniques.36–38 Analysis of short mixing time
J. Chem. Soc., Dalton Trans., 2002, 849–855
851

View Article Online
Table 1 Change in chemical shift (ppm) for the d(GTCGAC)2 resonances upon addition of ∆-[Ru(Me2phen)2dppz]2ϩ at a metal complex:duplex
ratio of 0.9, in 10 mM phosphate (pH 7) containing 20 mM NaCl at 25 and 45 ЊC. Numbers in parentheses are tentative assignments. ND = not
determined
Hexanucleotide proton
H8/H6
H5/H2
Me
H1Ј
25 ЊC
45 ЊC
25 ЊC
45 ЊC
25 ЊC
45 ЊC
25 ЊC
45 ЊC
G1
T2
C3
G4
A5
C6
Ϫ0.22
Ϫ0.10
Ϫ0.09
Ϫ0.18
0.09
Ϫ0.19
Ϫ0.11
Ϫ0.15
Ϫ0.12
0.02
Ϫ0.15
Ϫ0.48
ND
Ϫ0.08
Ϫ0.34
0.00
Ϫ0.21
Ϫ0.43
Ϫ0.55
Ϫ0.19
Ϫ0.39
Ϫ0.04
0.08
Ϫ0.05
0.00
Ϫ0.14
ND
(0.19)
(Ϫ0.26)
Ϫ0.02
Ϫ0.11
0.00
Table 2 1H NMR chemical shifts of the free ∆-[Ru(Me2phen)2dppz]2ϩ and the hexanucleotide-bound ∆-[Ru(Me2phen)2dppz]2ϩ, in 10 mM
phosphate buffer (pH 7) containing 20 mM NaCl at 45 ЊC. The assignment of the H14 and H13 resonances could be inverted, i.e. H13 = 7.99
and 7.64 ppm
Ligand proton
Free ∆-[Ru]2ϩ complex/ppm
d(GTCGAC)2-bound ∆-[Ru]2ϩ complex/ppm
Change in shift upon binding/ppm
dppz
H14
H13
H12
H11
7.99
8.20
9.22
7.40
7.58
7.64
7.78
8.55
6.89
7.15
Ϫ0.35
Ϫ0.42
Ϫ0.67
Ϫ0.51
Ϫ0.43
H10
Me2phen
9-Methyl
H8
H7
H6
H5
H4
H3
2-Methyl
2.04
7.86
8.75
8.33
8.16
8.31
7.33
1.91
2.01
7.88
8.81
8.37
8.20
8.25
7.33
1.85
Ϫ0.03
0.02
0.06
0.04
0.04
Ϫ0.06
0.00
Ϫ0.06
titration, consistent with the established high DNA binding
affinity of dppz-based ruthenium() complexes.14,15
NMR spectra of the metal complex-bound hexanucleotide,
at a metal complex:hexanucleotide duplex ratio of 0.9, were
recorded over a range of temperatures. At all temperatures,
addition of the metal complex induced significant changes in
the chemical shift of the resonances from the hexanucleotide
(see Table 1). All the aromatic H8 and H6 resonances exhibited
significant upfield shifts (except for A5 and T6 at 45 ЊC). How-
ever, the metal complex binding induced greater shifts for most
of the sugar H1Ј protons. Hexanucleotide binding induced
large upfield shifts for the dppz protons (see Table 2), consistent
with the metal complex binding by intercalation. Alternatively,
only relatively small shifts (≤0.06 ppm) were observed for the
Me2phen protons. This is consistent with selective intercalation
by the dppz ligand. Addition of the metal complex to the hexa-
nucleotide also induced extensive broadening and upfield shifts
of the T2 and G4 imino resonances. At 25 ЊC where both imino
protons could be assigned, upfield shifts of 0.59 and 0.35 ppm
were observed for the T2 and G4 imino resonances respectively.
The large upfield shifts and extensive broadening observed for
the imino resonances are again consistent with the metal com-
plex binding by intercalation.15,39 The broadening of the imino,
and amino, resonances precluded the observation of NOE
cross-peaks from these protons in NOESY experiments.
DNA binding by intercalation is also generally characterised
by an increase in the melting temperature of the DNA.15 For
the hexanucleotide, changes to the melting temperature can be
determined from the transition midpoint of the temperature
dependence curve of the resonances from the hexanucleo-
tide. The chemical shift changes reflect the conversion from a
duplex state to the totally base-destacked single state for the
Fig. 4 Melting curves (data points connected by a smoothed line) of
the free d(GTCGAC)2 duplex (1.4 mM) (᭿) and the ∆-[Ru(Me2phen)2-
dppz]2ϩ-bound hexanucleotide duplex (ꢀ). (A) The chemical shift
of the A5H1Ј resonance of the free hexanucleotide and the metal
complex-bound hexanucleotide as a function of temperature. (B) The
A5H8 resonance of the free hexanucleotide and the H10 resonance
(δ ϩ 1.0 ppm) of the metal complex bound to the hexanucleotide as a
function of temperature.
852
J. Chem. Soc., Dalton Trans., 2002, 849–855

View Article Online
hexanucleotide at millimolar concentrations. Fig. 4 shows the
chemical shift of several resonances from the free and metal
complex-bound hexanucleotide as a function of temperature.
The transition midpoint of the temperature dependence curve
of the free hexanucleotide was determined to be 45 ЊC, in
agreement with the results reported by Dupureur and Barton
for the same hexanucleotide.15 Addition of the metal complex
increased the midpoint of the transition of the hexanucleotide
by 17 ЊC, consistent with intercalation and the results of
Dupureur and Barton for ∆-[Ru(phen)2dppz]2ϩ 15
.
ꢀ-[Ru(Me2phen)2dppz]2؉–d(GTCGAC)2 interactions –
two-dimensional NMR
NOESY spectra of the hexanucleotide with added metal com-
plex were recorded at temperatures in the range 3–60 ЊC, using
mixing times ranging from 100 to 350 ms, to obtain a more
detailed picture of the metal complex binding. As in previous
NMR studies of the oligonucleotide binding by dppz com-
plexes, it was not possible to observe all expected intraduplex
NOE cross-peaks expected for a B-type DNA helix at any one
temperature. This is presumably due to selective exchange
broadening of some of the resonances from the hexanucleotide.
However, a considerable number of NOE cross-peaks between
the metal complex and the hexanucleotide could be assigned,
particularly at 45 ЊC (see Fig. 5 and Table 3). At 45 ЊC the
Fig. 6 Expansion of the NOESY spectrum (350 ms mixing time) of
∆-[Ru(Me2phen)2dppz]2ϩ and d(GTCGAC)2, at a metal complex :
duplex ratio of 0.9 at 35 ЊC. NOEs from the Me2phen H8, H7, 2-methyl
and 9-methyl protons to hexanucleotide minor groove protons and
NOEs from the dppz ligand H13/14 protons to hexanucleotide major
groove protons are indicated. The insert shows the NOEs from the
2-methyl (and C3H2Ј) and 9-methyl groups to the H4Ј/H5Ј/H5Љ protons
of the hexanucleotide (indicated by arrows) but at 45 ЊC where the
cross-peaks are more clearly resolved.
hexanucleotide minor groove. No NOEs were observed from
the Me2phen protons to hexanucleotide major groove protons
(H8/H6, TMe, H2Ј/H2Љ and H3Ј). By contrast, intermolecular
NOEs are only observed from the dppz H13/H14 protons to the
hexanucleotide major groove protons, in particular the G4H2Ј,
G1H2Ј/H2Љ and T2Me protons. The NOE data indicate that
∆-[Ru(Me2phen)2dppz]2ϩ intercalates from the minor groove.
In this binding mode the non-intercalating Me2phen ligands
would be located in the minor groove with the dppz ligand
intercalated between the stacked bases and, due to the length of
the dppz ligand, the H13/14 protons project out into the major
groove.
The strongest NOE cross-peaks observed from the metal
complex are to the hexanucleotide G4 protons, indicating that
the metal complex intercalates predominantly on either side of
the G4 base. However, the intermolecular NOEs observed
between the H14 and the G1H2Ј/H2Љ protons indicate that the
metal complex also intercalates between the A5C6 residues,
indicating that ∆-[Ru(Me2phen)2dppz]2ϩ demonstrates little
sequence selectivity. As the metal complex exchanges relatively
rapidly between the various binding sites, the observed inter-
molecular NOEs only represent a time-averaged indication
of the binding of the metal complex with the hexanucleotide.
Consequently, the NOE results only give an indication of the
dominant binding modes and minor binding modes cannot
be discounted.
Fig. 5 Expansion of the NOESY spectrum (350 ms mixing time) of ∆-
[Ru(Me2phen)2dppz]2ϩ and d(GTCGAC)2, at a metal complex:duplex
ratio of 0.9 at 45 ЊC. The expansion shows the NOE connectivities from
the hexanucleotide base H8/H6 (7.4 to 8.3 ppm) and sugar H1Ј (5.5 to
6.2 ppm) protons and metal complex aromatic protons (6.8 to 8.8 ppm)
to the hexanucleotide sugar H2Ј/H2Љ and T2Me protons and the metal
complex methyl protons (1.3 to 2.8 ppm). NOEs between the metal
complex H13/H14 protons and the hexanucleotide G4H2Ј and G1H2Ј/
H2Љ protons are indicated. The NOEs between the metal complex H13/
H14 protons and the hexanucleotide T2Me protons are shown (in the
dashed box). The insert is an expansion of the NOEs inside the dashed
box but from a NOESY spectrum at 35 ЊC, where the H13/H14 to T2Me
protons NOEs (indicated by arrows) are more clearly observed.
Binding models
hexanucleotide duplex will have started to melt; however, from
the temperature dependence curves shown in Fig. 4 it is con-
cluded that the duplex will still be largely intact. Most of the
intermolecular NOEs observed at 45 ЊC were also detected in
NOESY spectra recorded at 35 ЊC (see Fig. 6), although with
lower resolution due to increased peak broadening. Inter-
molecular NOEs are observed from the Me2phen ligand pro-
tons to the hexanucleotide H1Ј and H4Ј/H5Ј/H5Љ protons. In
particular, to the H1Ј protons of T2, G4, A5 and C6 and to the
H4Ј/H5Ј/H5Љ protons of C3 and G4. The sugar H1Ј and H4Ј/
H5Ј/H5Љ protons are located in (or are most accessible from) the
The metal complex did not bind exclusively at one site, and even
if it had, rotation about the C2 axis of the complex would result
in averaging of contacts at all binding sites except the central
C3G4. Consequently, only broad exchange-averaged resonances
were observed from the metal complex and hexanucleotide,
and therefore, the NMR data did not allow for the determin-
ation of a quantitative structure. However, a series of simple
binding models was constructed to examine the proposed
minor groove intercalation of the ∆-[Ru(Me2phen)2dppz]2ϩ
complex. For intercalation between the bases, the proximity of
J. Chem. Soc., Dalton Trans., 2002, 849–855
853

View Article Online
Table 3 NOE cross-peaks observed between the ∆-[Ru(Me2phen)2dppz]2ϩ and d(GTCGAC)2, at a metal complex:duplex ratio of 0.9, in 10 mM
phosphate (pH 7) containing 20 mM NaCl at 35 and/or 45 ЊC. The relative intensities of the NOEs are indicated (S = strong; M = medium; W = weak;
and VW = very weak)
Metal complex proton
Hexanucleotide protons
H14
H13
H12
G4H2Ј(S), G1H2Ј(W), G1H2Љ(W), T2Me(M)
G4H2Ј(S), G1H2Ј(W), G1H2Љ(W), T2Me(M)
T2Me(VW)
H10
9-Methyl
H8
H7
H3
C3H4Ј(W)
C3H4Ј(M), G4H4Ј/H5Ј/H5Љ(M)
C3H4Ј(M), G4H4Ј/H5Ј/H5Љ(M)
C3H4Ј(M), G4H4Ј/H5Ј/H5Љ(M)
C6H1Ј(M), G4H1Ј(W)
2-Methyl
T2H1Ј(W), G4H1Ј(VW), A5H1Ј(VW), C6H1Ј(M), G4H4Ј/H5Ј/H5Љ(M)
the metal complex 2-methyl protons to the guanine edges of the
base pairs was the limiting factor to the extent that the dppz
ring could be inserted into the base stack. Despite the potential
for steric clashes, it was found that the metal complex could
intercalate into B-type DNA such that the dppz ring just pro-
jected into the major groove. The ability of the metal complex
to fully intercalate is aided by the significantly bowed Me2phen
ligands, as shown in the crystal structure. Minimised energies
for binding at all three unique base-pair combinations differed
over a range of only 10 kJ molϪ1, consistent with the lack of
selectivity. It was also possible to produce minimised models
with different orientations of the metal complex with respect to
the DNA, again indicating a lack of selectivity. Close inter-
proton contacts were observed in these models that were con-
sistent with all observed intermolecular NOE cross-peaks.
Models of G4A5 (see Fig. 7) and A5C6 intercalation from the
minor groove were sufficient to provide explanations for all
observed NOE contacts, but we would not rule out C3G4 inter-
calation. More importantly, all contacts were found to be con-
sistent with intercalation from the minor groove, with only one
observed NOE (H12 to T2Me) consistent with intercalation
from either groove.
Discussion
Despite the considerable interest in the DNA binding of
ruthenium() complexes based on the dppz ligand the mode
of binding remains unresolved. An important aspect of the
DNA binding mode, and a point of considerable controversy,
is the determination of the groove from which the dppz com-
plex intercalates. We have used the ∆-[Ru(Me2phen)2dppz]2ϩ
complex to study this aspect of DNA binding by dppz-based
ruthenium() complexes.
While there is significant strain, due to the methyl groups, it
has been demonstrated that the complex [Ru(Me2phen)2dppz]2ϩ
can be made and resolved. No experiments were specifically
conducted to determine how stable this complex is. However, no
noticeable degradation of the NMR resonances from the metal
complex was observed in samples that were repeatedly used, for
both the free and hexanucleotide-bound metal complex. From
the crystal structure it was observed that the Me2phen ligands
are significantly bowed. This results in the methyl groups
adjacent to the dppz ligand not protruding towards the dppz as
far as may be expected. Hence, while the methyl groups will
impact somewhat on the extent to which the metal complex can
intercalate, they do not inhibit the complex from intercalating
such that the dppz ligand projects into the opposite groove.
Addition of the metal complex to the hexanucleotide induced
considerable upfield shifts for the metal complex dppz protons
and the hexanucleotide imino protons. These resonances also
displayed significant broadening, indicating intermediate
exchange binding kinetics. The addition of the metal complex
caused a 17 ЊC increase in the midpoint of the temperature
dependence curve of the resonances from the hexanucleotide.
These observations are in agreement with those reported by
Dupureur and Barton in their hexanucleotide binding study of
∆-[Ru(phen)2dppz]2ϩ 15
and strongly suggest that ∆-[Ru(Me2-
,
phen)2dppz]2ϩ binds the hexanucleotide by intercalation. The
observed pattern of intermolecular NOEs confirms that the
metal complex binds by intercalation, as all the observed NOEs
from the Me2phen ligands are to the minor groove while those
from the terminal protons on the dppz ligand are to the major
groove. Furthermore, from the observed pattern of inter-
molecular NOEs it is concluded that ∆-[Ru(Me2phen)2dppz]2ϩ
intercalates from the minor groove. As the Me2phen ligands are
significantly bowed and the molecular modelling demonstrates
that the ∆-[Ru(Me2phen)2(dppz)]2ϩ complex can fully inter-
calate, we believe that the proposed minor groove binding is not
a consequence of the methyl groups. The minor groove binding
mode conclusion is consistent with the results of our recent
hexanucleotide binding study of ∆- and ∆-[Ru(Me2phen)2-
Fig. 7 Two views of a molecular model of ∆-[Ru(Me2phen)2dppz]2ϩ
bound to the hexanucleotide d(GTCGAC)2, generated using
HyperChem 5.0.34 The model of the hexamer duplex was generated in
HyperChem and the metal complex intercalated between the G4 and A5
residues. Energy minimisation of the hexamer was then carried out to
convergence.
dpq]2ϩ 21
thereby suggesting that ruthenium() polypyridyl
,
complexes favor minor groove binding as originally proposed
by Lincoln et al.17
Using the same hexanucleotide used in this study and the
854
J. Chem. Soc., Dalton Trans., 2002, 849–855

View Article Online
∆-[Ru(phen)2(dppz)]2ϩ complex, Dupureur and Barton pro-
posed major groove binding.15 This conclusion was based on
the observation of an NOE from the H12 of the metal complex
to the A5H8 of the hexanucleotide. In our minor groove binding
model, the metal complex H12 proton is positioned near the
centre of the stacked bases. Consequently, it is possible that an
NOE could be observed from the H12 protons to a major
groove proton. Indeed, we observed a weak NOE from the H12
protons to the T2Me protons.
3 K. E. Erkkila, D. T. Odom and J. K. Barton, Chem. Rev., 1999, 99,
2777.
4 A. E. Friedman, J.-C. Chambron, J.-P. Sauvage, N. J. Turro and
J. K. Barton, J. Am. Chem. Soc., 1990, 112, 4960.
5 R. M. Hartshorn and J. K. Barton, J. Am. Chem. Soc., 1992, 114,
5919.
6 C. Hiort, P. Lincoln and B. Nordén, J. Am. Chem. Soc., 1993, 115,
3448.
7 Y. Jenkins and J. K. Barton, J. Am. Chem. Soc., 1992, 114, 8736.
8 L. S. Schulman, S. H. Bossmann and N. J. Turro, J. Phys. Chem.,
1995, 99, 9283.
Lincoln et al. have suggested that dppz-based complexes bind
DNA in a cooperative manner.40 This indicates that there could
be two metal complexes bound to 50% of the hexanucleotide in
the experiments where the ratio of the metal complex:hexa-
nucleotide duplex is 0.9. However, the distribution of the
bound metal complex should not affect the proposed binding
model. The minor groove binding mode shown in Fig. 7 is
only a qualitative model, based on observed NOEs from two
or more binding sites.
While it may not appear that intercalation would be favoured
in the minor groove on steric grounds, many DNA binding
studies of relatively bulky molecules have demonstrated the
flexibility of DNA. For example, Önfelt et al. reported that the
bis-intercalating complex [µ-c4(cpdppz)2-(phen)4Ru2]4ϩ inter-
calates such that both Ru(phen)2 moieties are in the same
groove.41 For this type of binding to occur the ruthenium bis-
intercalator must thread through the DNA strands, thereby
demonstrating the ‘large amplitude of conformational change’
that can occur in DNA.41
9 C. Moucheron, A. Kirsch-De Mesmaeker and S. Choua, Inorg.
Chem., 1997, 36, 584.
10 C. S. Chow and J. K. Barton, Methods Enzymol., 1992, 212, 219.
11 C. J. Murphy, M. R. Arkin, Y. Jenkins, N. D. Ghatlia, S. H.
Bossman, N. J. Turro and J. K. Barton, Science, 1993, 262, 1025.
12 E. D. A. Stemp, M. R. Arkin and J. K. Barton, J. Am. Chem. Soc.,
1995, 117, 2375.
13 C. G. Coates, L. Jacquet, J. J. McGarvey, S. E. J. Bell, A. H. R.
Al-Obaidi and J. M. Kelly, J. Am. Chem. Soc., 1997, 119, 7130.
14 C. M. Dupureur and J. K. Barton, J. Am. Chem. Soc., 1994, 116,
10286.
15 C. M. Dupureur and J. K. Barton, Inorg. Chem., 1997, 36, 33.
16 R. E. Holmlin, E. D. A. Stemp and J. K. Barton, Inorg. Chem., 1998,
37, 29.
17 P. Lincoln, A. Broo and B. Nordén, J. Am. Chem. Soc., 1996, 118,
2644.
18 E. Tuite, P. Lincoln and B. Nordén, J. Am. Chem. Soc., 1997, 119, 239.
19 M. Eriksson, M. Leijon, C. Hiort, B. Nordén and A. Graslund,
Biochemistry, 1994, 33, 5031.
20 P. Lincoln and B. Nordén, J. Phys. Chem. B, 1998, 102, 9583.
21 J. G. Collins, J. R. Aldrich-Wright, I. D. Greguric and P. A.
Pellegrini, Inorg. Chem., 1999, 38, 5502.
It has been clearly demonstrated that metallointercalators
based on the 9,10-phenanthrenequinone diimine (phi) ligand
intercalate from the major groove.42–45 In particular, as the
complex ∆-[Rh(phen)2phi]3ϩ intercalates from the major
groove,42 the results of this study could indicate that the struc-
ture of the intercalating ligand plays an important role in
determining the DNA binding site. For the phi-based inter-
calators, the long axis of the aromatic rings run parallel to the
long axis of the base-pairs, whereas for the dppz complexes
the long axis of the aromatic rings run perpendicular to the
long axis of the base-pairs. This difference in shape could lead
to the observed difference in the binding modes. However, this
proposal is not consistent with the groove binding preference of
organic drugs that bind DNA by intercalation. For example,
both actinomycin (long axis of the aromatic rings parallel to the
long axis of the base-pairs) and the anthracycline drugs (aro-
matic rings perpendicular to the base-pairs) intercalate from the
minor groove,46 as do most positively charged organic drugs.46
Hence, the difference in the preferred groove for binding by the
metallointercalators may be related to the shape of the inter-
calating ligand, but probably only as a consequence of the steric
bulk provided by the remainder of the octahedral complex.
Haq et al. have demonstrated that the intercalative binding
of [Ru(phen)2dppz]2ϩ with DNA is entirely entropically driven,
with hydrophobic interactions, changes in hydration and the
release of counter ions being the dominant driving forces.47
Hence, while the steric interactions of the non-intercalating
part of the complex with the DNA grooves may affect the bind-
ing preference, the sequestering of the phenanthroline ligands
from the water is also probably important. Rhodium() com-
plexes containing the phi ligand and either aromatic or aliphatic
ancillary ligands have been shown to intercalate from the major
groove.42–45 However, it has not been established whether dppz-
based complexes with non-aromatic ancillary ligands will inter-
calate from the DNA minor groove in a similar fashion to
22 M. Yamada, Y. Tanaka, Y. Yoshimoto, S. Kuroda and I. Shimao,
Bull. Chem. Soc. Jpn., 1992, 65, 1006.
23 SMART, SAINT, SADABS and XPREP. Area detector
control and data integration and reduction software, Bruker
Analytical X-ray Instruments Inc., Madison, WI, USA, 1995.
24 G. M. Sheldrick, SHELXS-86, in Crystallographic Computing 3,
G. M. Sheldrick, C. Krüger and P. Goddard, ed., Oxford University
Press, Oxford, 1985, pp. 175–189.
25 teXsan, Crystal Structure Analysis Package, Molecular Structure
Corporation, Houstan, TX, 1985 and 1992.
26 International Tables for X-Ray Crystallography, vol. 4, Kynoch
Press, Birmingham, 1974.
27 J. A. Ibers and W. C. Hamilton, Acta Crystallogr., 1964, 17, 781.
28 D. C. Creagh and W. J. McAuley, International Tables for
Crystallography, A. J. C. Wilson, ed., Kluwer Academic Publishers,
Boston, 1992, vol. C, Table 4.2.6.8, pp. 219–222.
29 D. C. Creagh and J. H. Hubbell, International Tables for
Crystallography, A. J. C. Wilson, ed., Kluwer Academic Publishers,
Boston, 1992, vol. C, Table 4.2.4.3, pp. 200–206.
30 C. K. Johnson, ORTEP, A Thermal Ellipsoid Plotting Program,
Oak Ridge National Laboratories, Oak Ridge, TN, 1965.
31 T. Maniatis, E. F. Fritsch and J. Sambrook, Molecular Cloning, Cold
Spring Harbor Laboratory Press, New York, 1982.
32 D. J. States, R. A. Haberkorn and D. J. Ruben, J. Magn. Reson.,
1982, 48, 286.
33 M. Piotto, V. Saudek and V. Sklenar., J. Biomol. NMR, 1992, 2, 661.
34 HyperChem, Release 5.01 for Windows Molecular Modelling
System, HyperCube Inc., Ontario, Canada, 1996.
35 J. G. Collins, A. D. Sleeman, J. R. Aldrich-Wright, I. Greguric and
T. W. Hambley, Inorg. Chem., 1998, 37, 3133.
36 R. M. Scheek, R. Boelens, N. Russo, J. H. van Boom and
R. Kaptein, Biochemistry, 1984, 23, 1371.
37 J. Feigon, W. Leupin, W. A. Denny and D. R. Kearns, Biochemistry,
1983, 22, 5943.
38 D. J. Patel, L. Shapiro and D. Hare, J. Biol. Chem., 1986, 261, 1223.
39 J. Feigon, W. A. Denny, W. Leupin and D. R. Kearns, J. Med. Chem.,
1984, 27, 450.
40 P. Lincoln, E. Tuite and B. Nordén, J. Am. Chem. Soc., 1997, 119,
1454.
41 B. Önfelt, P. Lincoln and B. Nordén, J. Am. Chem. Soc., 1999, 121,
10846.
[Ru(Me2phen)2dppz]2ϩ
.
42 S. S. David and J. K. Barton, J. Am. Chem. Soc., 1993, 115, 2984.
43 J. G. Collins, T. P. Shields and J. K. Barton, J. Am. Chem. Soc., 1994,
116, 9840.
44 T. P. Shields and J. K. Barton, Biochemistry, 1995, 34, 15049.
45 B. P. Hudson and J. K. Barton, J. Am. Chem. Soc., 1998, 120, 6877.
46 S. Neidle, DNA Structure and Recognition, IRL Press, Oxford, 1994.
47 I. Haq, P. Lincoln, D. Suh, B. Nordén, B. Z. Chowdhry and
J. B. Chaires, J. Am. Chem. Soc., 1995, 117, 4788.
References
1 D. S. Sigman, A. Mazumder and D. M. Perrin, Chem. Rev., 1993, 93,
2295.
2 B. Nordén, P. Lincoln, B. Akerman and E. Tuite, Met. Ions Biol.
Syst., 1996, 33, 177.
J. Chem. Soc., Dalton Trans., 2002, 849–855
855