Quadracyclic adenine: A non-perturbing fluorescent adenine analogue

Article abstract of DOI:10.1002/chem.201103419

Fluorescent-base analogues (FBAs) comprise a group of increasingly important molecules for the investigation of nucleic acid structure and dynamics as well as of interactions between nucleic acids and other molecules. Here, we report on the synthesis, detailed spectroscopic characterisation and base-pairing properties of a new environment-sensitive fluorescent adenine analogue, quadracyclic adenine (qA). After developing an efficient route of synthesis for the phosphoramidite of qA it was incorporated into DNA in high yield by using standard solid-phase synthesis procedures. In DNA qA serves as an adenine analogue that preserves the B-form and, in contrast to most currently available FBAs, maintains or even increases the stability of the duplex. We demonstrate that, unlike fluorescent adenine analogues, such as the most commonly used one, 2-aminopurine, and the recently developed triazole adenine, qA shows highly specific base-pairing with thymine. Moreover, qA has an absorption band outside the absorption of the natural nucleobases (>300 nm) and can thus be selectively excited. Upon excitation the qA monomer displays a fluorescence quantum yield of 6.8 % with an emission maximum at 456 nm. More importantly, upon incorporation into DNA the fluorescence of qA is significantly less quenched than most FBAs. This results in quantum yields that in some sequences reach values that are up to fourfold higher than maximum values reported for 2-aminopurine. To facilitate future utilisation of qA in biochemical and biophysical studies we investigated its fluorescence properties in greater detail and resolved its absorption band outside the DNA absorption region into distinct transition dipole moments. In conclusion, the unique combination of properties of qA make it a promising alternative to current fluorescent adenine analogues for future detailed studies of nucleic acid-containing systems. Copyright

Full text of DOI:10.1002/chem.201103419

FULL PAPER
DOI: 10.1002/chem.201103419
Quadracyclic Adenine: A Non-Perturbing Fluorescent Adenine Analogue
Anke Dierckx,[a] Francois-Alexandre Miannay,[a] Nouha Ben Gaied,[b] Søren Preus,[c]
Markus Bjçrck,[a] Tom Brown,[b] and L. Marcus Wilhelmsson*[a]
Abstract: Fluorescent-base analogues
(FBAs) comprise a group of increasing-
ly important molecules for the investi-
gation of nucleic acid structure and dy-
namics as well as of interactions be-
tween nucleic acids and other mole-
cules. Here, we report on the synthesis,
detailed spectroscopic characterisation
and base-pairing properties of a new
environment-sensitive fluorescent ade-
nine analogue, quadracyclic adenine
(qA). After developing an efficient
route of synthesis for the phosphorami-
dite of qA it was incorporated into
DNA in high yield by using standard
solid-phase synthesis procedures. In
DNA qA serves as an adenine ana-
logue that preserves the B-form and, in
contrast to most currently available
FBAs, maintains or even increases the
stability of the duplex. We demonstrate
that, unlike fluorescent adenine ana-
logues, such as the most commonly
used one, 2-aminopurine, and the re-
cently developed triazole adenine, qA
shows highly specific base-pairing with
thymine. Moreover, qA has an absorp-
tion band outside the absorption of the
natural nucleobases (>300 nm) and
can thus be selectively excited. Upon
excitation the qA monomer displays
a fluorescence quantum yield of 6.8%
with an emission maximum at 456 nm.
More importantly, upon incorporation
into DNA the fluorescence of qA is
significantly less quenched than most
FBAs. This results in quantum yields
that in some sequences reach values
that are up to fourfold higher than
maximum values reported for 2-amino-
purine. To facilitate future utilisation of
qA in biochemical and biophysical
studies we investigated its fluorescence
properties in greater detail and re-
solved its absorption band outside the
DNA absorption region into distinct
transition dipole moments. In conclu-
sion, the unique combination of prop-
erties of qA make it a promising alter-
native to current fluorescent adenine
analogues for future detailed studies of
nucleic acid-containing systems.
Keywords: DNA
·
DNA solid-
phase synthesis · duplex stability ·
fluorescence · fluorescent adenine
analogue
Introduction
tor the dynamics of DNA/RNA. However, the use of fluo-
rescence to study nucleic acids requires some form of label-
ling with a fluorescent marker since the natural nucleobases
are virtually non-fluorescent (Ff ~10ꢀ4).[1] For this reason,
there is an ongoing effort to develop different types of fluo-
rescent probes with properties suited for various purposes.
Most available dyes are tethered covalently to the nucleic
acids by a linker, for example, fluorescein, rhodamine, the
Alexa-dyes and the Cy-dyes. Since these fluorophores reside
outside of the base stack, these are referred to as external
modifications. A second class, namely internal modifications,
comprise fluorescent nucleic acid base analogues (FBAs).[2]
These molecules resemble one of the natural nucleobases,
show at least some base pairing to one of the natural bases
through hydrogen bonds, do not seriously perturb the natu-
ral DNA or RNA structure, and most importantly, are sig-
nificantly fluorescent. Having the fluorescent probe inside
the base stack has prominent advantages compared to exter-
nally attached bulky adducts, which can significantly perturb
the DNA duplex structure and also interfere with the inter-
action with other molecules. Furthermore, FBAs in contrast
to external probes, can be firmly stacked and hence report
on the local structure and dynamics of the DNA duplex
rather than the dynamics of the probe moiety itself.
Fluorescence is an invaluable tool for researchers studying
the structure and dynamics of nucleic acids and their inter-
actions with other biomolecules. The main advantage of
fluorescence over other standard techniques, such as X-ray
crystallography, NMR spectroscopy, and electrophoresis, is
its high sensitivity, versatility as well as the ability to moni-
[a] A. Dierckx,+ Dr. F.-A. Miannay,+ M. Bjçrck,
Prof. Dr. L. M. Wilhelmsson
Department of Chemical and Biological Engineering
Physical Chemistry, Chalmers University of Technology
41296 Gothenburg (Sweden)
Fax : (+46)31-772-3858
E-mail: marcus.wilhelmsson@chalmers.se
[b] Dr. N. Ben Gaied, Prof. Dr. T. Brown
School of Chemistry, University of Southampton
SO17 1BJ Southampton (UK)
[c] S. Preus
Department of Chemistry, University of Copenhagen
Universitetsparken 5, 2100 Copenhagen (Denmark)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201103419.
Chem. Eur. J. 2012, 18, 5987 – 5997
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5987

Over recent decades several research groups have focused
on the development of FBAs with the above-mentioned
properties. However, synthesising the optimal FBA has
proved challenging. Instead, each class of FBA holds specif-
ic advantages depending on the kind of experiment one
wants to perform. Further development and characterisation
of novel FBAs will hopefully lead to a more complete un-
derstanding of the relationship between their structural and
fluorescence properties.[3] The base-discriminating analogues
designed by Saito and co-workers,[4] the pyrimidine ana-
logues designed by Tor et al.[5] and the cytosine analogue
pyrrolo dC[6] as well as its derivatives[7] are just a few exam-
ples of thoroughly studied FBAs. Other classes, such as the
pteridines designed by Hawkins et al. comprise the guanine
analogues 3-MI[8] and 6-MI and the adenine analogues 6-
MAP and DMAP[9] as their most promising members.[10]
The size-expanded DNA alphabet designed by Kool et al.
seems practical for evaluating steric effects in DNA duplex
structures among other things.[11] Another unique group of
FBAs is the tricyclic cytosine analogues tC and tCO, the
emission of which is virtually insensitive to their microenvir-
onment.[12] In addition tCO, together with the non-emissive
cytosine analogue tCnitro have been shown to function as the
first nucleobase FRET pair, and yield detailed distance and
orientation information about the DNA duplex.[13] In con-
trast to this latter class, all other FBAs show sensitivity to
their immediate surroundings.
other molecules. For instance the fluorescent adenosine ana-
logue 6MAP has been used to analyse the pre-melting tran-
sitions of DNA A tracts.[20] FBAs are now also starting to
appear in the field of nanotechnology. A first report of this
involves the cytosine analogue tCO sensing the local stability
of self-assembling DNA hexagons.[21] Finally, FBAs are also
finding applications in the RNA field. This is nicely illustrat-
ed by the use of 2-AP as a probe for research on small
RNAs and RNA elements that control gene expression.[22]
Here we present the detailed and optimised experimental
procedure for the synthesis of the new fluorescent adenine
analogue, quadracyclic adenine (qA, Figure 1). The synthesis
Figure 1. Structure of the qA monomer in the putative qA–T base pair.
Measurements on the qA monomer were performed with R1 =R2 =H.
For qA–T incorporated into DNA, R1, R2 and R3 represent the rest of
the DNA structure.
One of the most popular and widely applied FBAs is the
adenine analogue 2-aminopurine (2-AP). The high quantum
yield as a monomer in water (0.68) is decreased dramatically
upon incorporation into nucleic acids.[14] This sensitivity of
the emission to the microenvironment has been successfully
used in many applications, of which some examples are
given below. Even though it has found use in numerous
studies, 2-AP causes a moderate destabilisation of the DNA
duplex.[15] Furthermore, it shows base-pairing to both thy-
mine and cytosine and hence is not an ideal adenine ana-
logue.[16] It also shows a considerable amount of dynamics
inside DNA.[17] These properties can be a substantial draw-
back in studies requiring a high level of detail and a low
level of perturbation to the DNA duplex.
Of the numerous studies involving FBAs over the past de-
cades, only a few will be highlighted here to illustrate the
importance of these fluorophores (for more examples see
recent reviews by Sinkeldam et al.[2b] and Wilhelmsson
et al.[2a]). Many applications involve the use of the change in
emission upon interaction of another molecule with DNA or
RNA. This can yield information on conformational changes
both in the protein and the DNA/RNA. An illustration is
the elegant study of the kinetics of promoter binding and
open complex formation upon interaction of bacteriophage
T7 RNA polymerase with DNA containing 2-AP modified
promoters.[18] Another example, which generated informa-
tion on the dissociation of the multimeric form of the HIV-
1 integrase complex upon binding to DNA, involves the
guanine analogue 3-MI.[19] Other applications focus on the
structure of the DNA itself instead of interactions with
of qA (the 2’-deoxyriboside of 2,6-dihydro-2,3,5,6-tetraazaa-
ceanthrylene; b anomer) has been reported by Matteucci
et al. but no experimental procedure or analytical data were
described.[23] Furthermore, we have found this nucleobase
analogue to be emissive and therefore we now report on its
fluorescent properties and present a detailed photophysical
characterisation of qA both as a monomer and as a constitu-
ent of single- and double-stranded DNA. As a monomer,
qA shows a lowest energy absorption band well-resolved
from the natural DNA bases and we have located the transi-
tion dipole moments over this absorption band. Surprising
results were obtained when investigating the quantum yields
upon exciting over these different absorption transitions.
Furthermore, we find that qA is moderately fluorescent as
a monomer in water (Ff =6.8%, lex =337 nm) compared to
other adenine analogues, such as 2-AP.[14] However, upon in-
corporation into DNA, quantum yields reach maximum
values, which are higher than those corresponding to 2-
AP.[14] We also report on the non-perturbing properties of
qA in DNA duplex systems. On average qA slightly stabilis-
es DNA duplexes and CD measurements suggest that the
DNA duplex adopts the B-form. Furthermore, qA is a selec-
tive adenine analogue since it shows specific base pairing to
thymine only, in contrast to other fluorescent adenine ana-
logues, such as 2-AP and triazole adenine.[16,24] To the best
of our knowledge qA is the only fluorescent adenine ana-
logue with such favourable properties.
5988
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5987 – 5997

An Emissive Adenine Analogue
FULL PAPER
Results and Discussion
three steps with an overall yield of 36% (Scheme 1). Con-
densation of 6-chloro-7-deaza-7-iodopurine with the chloro
sugar by using sodium hydride[26] gave the desired com-
pound (2) in 73% yield. The major isomer was consistent
with the b anomer by using proton NMR spectroscopy.
Moreover, the final product that was later incorporated into
oligonucleotides gave stable duplexes, further confirming
the correct isomer. The sugar–base condensation step was
followed by a Stille coupling. In the original work Matteucci
et al.[23] used trimethylstannyl-N-(Boc)aniline obtained by
ortho-functionalisation of the N-(Boc)aniline.[27] In our
hands this step was low yielding and we attempted to over-
come this problem using tributyltin chloride. Our first at-
tempts, following the original procedure, gave very low
yields and changing the catalyst, the solvent and the temper-
ature did not improve the situation.
Synthesis of the quadracyclic 2’-deoxyadenosine analogue,
qA: The synthesis of the quadracyclic analogue of 2’-deoxy-
adenosine (qA; 2’-deoxyriboside of 2,6-dihydro-2,3,5,6-tet-
raazaaceanthrylene; b anomer) has previously been reported
by Matteucci et al. but no experimental procedures were de-
scribed.[23] Here we report the synthesis of the correspond-
ing phosphoramidite monomer (Scheme 1) and oligonucleo-
tides containing qA (see the Supporting Information).
The synthesis of the qA phosphoramidite (8) was initiated
with the conversion of the 2’-deoxyribose into 3,5-ditoluoyl-
1-chloro-deoxyribose (predominantly a anomer; 1)[25] in
A procedure was then followed that employed a combina-
tion of cesium fluoride and cuprous salts to enhance the re-
action rate.[28] The use of CuI and highly polar solvents
(NMP or DMF) in the Stille reaction is well-documented.
The rate accelerating effect of CuI is attributed to a prelimi-
nary transmetallation reaction with the organostannane that
generates a more reactive organo–copper intermediate. In
addition, the Bu3SnCl by-product can be efficiently trans-
formed into insoluble Bu3SnF, which is easily removed from
the reaction mixture by filtration. In our case, a high tem-
perature was needed to consume all the starting material to
give a respectable yield of 4 (60%). Formation of compound
5 was achieved by using a mixture of DABCO and DBU.
Under these conditions partial deprotection of the tert-butyl-
AHCTUNGTREGoNNUN xycarbonyl protecting group occurred. Therefore, the Boc-
intermediate was not isolated but was fully deprotected by
using a 25% solution of TFA in DCM. Deprotection of the
toluoyl group was then performed by using a solution of
NaOMe (0.5m) in MeOH, which gave the free qA nucleo-
side (6) in 61% yield. It was crucial to purify this nucleoside
by column chromatography in order to achieve successful
tritylation in the next step. Purified compound 6 was tritylat-
ed by using dimethoxytrityl chloride in pyridine to give (7)
in 68% yield. This was finally converted to the qA phos-
phoramidite monomer (8) in 79% yield.
Spectroscopic characterisation of the qA monomer
Steady state absorption and fluorescence spectra: The steady
state absorption and emission spectra of the quadracyclic
adenine (qA) monomer excited at 337 nm in water and
methanol are shown in Figure 2. The absorption spectrum is
characterised by a sharp peak at 335 nm with an extinction
coefficient of 5000mꢀ1 cmꢀ1 (in water) and a shoulder at
300 nm; this indicates the presence of several different tran-
sitions (S0!S2 and S0!S3, respectively, as discussed below)
and a lower energy spectral feature above 370 nm. The ab-
sorption of qA at 335 nm is red-shifted compared to the nat-
ural nucleobases, which absorb up to 300 nm, and therefore
allows for easy selective excitation. This is an advantage
compared to, for example, the fluorescent triazole adenine,
Scheme 1. Synthesis of the fluorescent quadracyclic phosphoramidite mo-
nomer (8). i) acetyl chloride, MeOH; ii) 1-toluoyl chloride, pyridine, 08C;
iii) acetyl chloride/acetic acid/H2O, acetic acid (36% for 3 steps); iv) 7-
chloro-7-deaza-6-iodo-purine, NaH, CH3CN, 508C (73%); v) PdACHTNUTRGEN(UNG PPh3)4,
CuI, CsF, DMF, 1008C (60%); vi) SnBu3Cl, tert-butyllithium, THF,
ꢀ788C to ꢀ208C (65%); vii) DBU, DABCO, DMF, 758C; viii) 25%
TFA/DCM (46% over 2 steps); ix) NaOMe/MeOH, (61%); x) DMTrCl,
pyridine (68%); xi) PClACHTUNGTRENNUNG[(OCH2)2CN]NACHTUTGNRENN(GUN iPr)2, DIPEA, CH2Cl2 (79%).
Chem. Eur. J. 2012, 18, 5987 – 5997
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5989

L. M. Wilhelmsson et al.
Figure 2. a) Normalised absorption, and b) emission spectra of the qA
monomer in water (black) and methanol (grey) when excited at 337 nm
at 258C. Inset: lowest energy absorption band of qA outside the region
of DNA absorption.
AT, (labs,max =282 nm)[24] or the series of 7-substituted ade-
nines designed by Seela et al., of which the most fluorescent
members have their absorption centred around 280 nm.[29]
As can be seen in Figure 2 the two emission spectra of qA
in water and methanol show a maximum emission peak at
456 and 440 nm, respectively. Furthermore, the spectral
shape and position of the emission are independent of the
excitation energy; this suggests that emission always occurs
from the same energy level (S1; see the Supporting Informa-
tion). qA shows a lower fluorescence in water (Ff =6.8%,
lex =337 nm) than other fluorescent adenine analogues, such
as 2-AP (68%),[14] 6-MAP (39%), DMAP (48%),[9] xA
(44% in methanol),[11a] hxy7c7Ad (27%)[30] or triazole ade-
nine (61%).[24] However, as will be discussed below, qA is
a promising FBA when located inside DNA.
Figure 3. Excitation anisotropy spectrum (black line) and excitation spec-
trum (dashed line) of qA in an H2O/ethylene glycol matrix (1:2 mixture)
at ꢀ1008C. The table shows the results of the TDDFT calculations on
the three lowest energy electronic transitions of qA.
emission transition dipole moment is invariable over this
spectral region (see the Supporting Information). As
a result of this, the anisotropy recorded will give a direct
measure of the angles between a certain absorption transi-
tion moment and the emission transition moment. Conse-
quently, the three different plateaus mentioned above, sug-
gest that there are three distinct absorption transition dipole
moments: one around 400 nm that is parallel to the emission
transition moment (S0!S1), and two others (S0!S2 and
S0!S3, respectively) with different angles relative the con-
stant emission dipole moment. We also noticed a bulge lo-
cated around the second plateau at 340 nm. This observation
might be attributed to vibrational features situated between
the electronic states S2 and S1.
Transition dipole moments: To obtain detailed information
about the transition dipole moments of qA, we recorded the
steady-state excitation spectrum and fundamental anisotro-
py of the qA monomer in an H2O/ethylene glycol matrix
(1:2 mixture) at ꢀ1008C (Figure 3). As in the absorption
spectrum recorded for the free monomer in solution
(Figure 2), this excitation spectrum (Figure 3, dashed line)
shows a sharp peak around 330 nm, a weaker one at 300 nm,
and a longer wavelength absorption transition centred at
390 nm with fine structure due to the vibrational levels. The
fundamental anisotropy, r0 (Figure 3), shows three different
main plateaus between 280 and 450 nm. The lowest energy
plateau is located between 390 and 450 nm (r0 =0.36), the
second occurs around the absorption maximum at 330 nm
(r0 =0.25) and the third around 300 nm (r0 =0.05). The low
energy plateau shows an r0 value close to the theoretical
maximum of 0.4. This suggests that, in this particular spec-
tral region, the absorption and emission transition dipole
moments are parallel.
Theoretical calculations of the lowest singlet energy exci-
tations of qA were performed on the B3LYP/6-31GACHTUNGTRENNUNG(d,p) op-
timised geometry by using TDDFT B3LYP/6-311+G(2d)
including a CPCM solvation shell for H2O as implemented
in Gaussian 03. Above 280 nm these calculations show three
significant transitions centred at 283.15 nm (f=0.1049; S0!
S3), 329.84 nm (f=0.1968; S0!S2) and 349.51 nm (f=0.0594;
S0!S1). Comparing these values to the transitions suggested
by the absorption and excitation spectra as well as the fun-
damental anisotropy, we observe that they match very well,
except that they are somewhat blue-shifted. This kind of
spectral shift between theoretical calculations and experi-
ments is common and has also been observed for other
FBAs.[12a,31] Importantly, the relative positions of the transi-
tions and the trend in the calculated oscillator strengths cor-
respond very well to the intensity of the spectral bands of
both the absorption and excitation spectra of the qA mono-
The reason why this experimental value is slightly lower
could be due to vibrational torsions and slight geometrical
changes in the molecule between the moment of excitation
and emission. Furthermore, as was mentioned in the previ-
ous section, all emission spectra of qA are virtually inde-
pendent of the excitation wavelength; this suggests that the
5990
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5987 – 5997

An Emissive Adenine Analogue
FULL PAPER
mer. This is further evidence for the existence of three dif-
ferent transitions in the low energy band of qA.
wavelengths) the quantum yield decreases. Upon excitation
further into the blue to S3 (lexc =295 nm) the fluorescence
quantum yield in water does not change significantly (6.8
and 6.4%, respectively). To explain these observations we
suggest a hypothesis in which part of the highly excited pop-
ulation could relax and then be trapped in a non-radiative
parasite state. Such an energy trap state could be localised
between the S2 and S1 states. Indeed, the decrease in quan-
tum yield of at least 30% between the S1 and S2 states clear-
ly indicates that a decrease of deactivation from S1 through
the fluorescence path occurs. In conclusion, we suggest a de-
activation pathway of the qA monomer in which at least
30% of the molecules that are excited to higher electronic
states (S2, S3) get trapped in a “parasite state” and the rest
relax to the lowest vibrational level of S1 and then further to
S0 by emission or internal conversion and vibrational relaxa-
tion. It has been observed that the quantum yield of the vir-
tually non-fluorescent natural mononucleotide, dGMP, also
varies with the excitation energy. In that case, populating
higher energy electronic states was suggested to open non-
radiative deactivation pathways.[32]
Fluorescence properties of qA: The fluorescence quantum
yields of qA obtained for different excitation wavelengths
between 295 and 377 nm are shown in Figure 4. Surprisingly
Figure 4. Steady-state absorption spectrum and fluorescence quantum
yields of qA for different excitation wavelengths between 295 and
377 nm in: a) MilliQ water, and b) methanol. All measurements were
performed at 25
N
Structure and stability of DNA duplexes containing qA
the fluorescence quantum yields in water decrease gradually
from 9.8 to 6.4% with excitation at shorter wavelengths (see
the Supporting Information). When exciting at wavelengths
further into the red region even higher quantum yields are
detected. However, these values are difficult to accurately
determine since the excitation energy severely overlaps the
emission spectrum at these wavelengths. Nevertheless, irre-
spective of the excitation wavelength the spectral envelope
of the emission and its position stays virtually the same for
the monomer both in water and in methanol (see the Sup-
porting Information). Furthermore, the time-resolved finger-
print of qA in water is defined by a single time component
of 3.2 ns (c2 =1.684), which is insensitive to excita-
UV duplex melting experiments: In order to investigate the
spectroscopic and structural properties of qA inside DNA,
eleven 10-mer oligonucleotides were synthesised, each with
one qA base incorporated opposite T (Table 1). To evaluate
the effect of both purines and pyrimidines neighbouring qA
at the 3’ and 5’ ends, these eleven sequences have variations
in the bases directly flanking qA. Furthermore, we designed
a twelfth 10-mer DNA duplex containing two qA molecules
in order to investigate the structural influence and the pho-
tophysical properties of qA in this case (sequence CT,TA).
The melting temperatures of the different duplexes con-
taining qA (TmqA) as well as those of the corresponding nat-
tion (295, 330 and 377 nm) and emission wave-
Table 1. Melting temperatures of the qA modified DNA duplexes (TmqA). Melting
temperatures of the corresponding natural duplexes are also listed (TmA) as well as the
corresponding difference in melting temperature with the modified duplexes (DTm).
lengths (456 and 520 nm; see the Supporting Infor-
mation). This suggests that regardless of which state
is being populated upon excitation, the fluorescence
always originates from deactivation of the same
electronic state (S1). Thus, internal conversion from
the high-energy electronic states (S2, S3) occurs fol-
lowed by a deactivation of the excited S1 population
through fluorescence or non-radiative processes.
The decrease in quantum yield from 377 to
300 nm is stepwise in methanol (Figure 4b) and
more continuous in water (Figure 4a); this suggests
a gradually changing overlap between the electronic
transitions S0!S3, S0!S2 and S0!S1. This indicates
a more efficient non-radiative deactivation to the
ground state when populating the S2 or S3 states
rather than the S1 state. The highest quantum yield
value is obtained when mostly populating the S1
state (lexc =377 nm); when the population ratio is
qA
A
DNA sequence[a,b]
Neighbouring
bases[c]
Tm
Tm
DTm
[8C]
[8C][d]
[8C][d]
5’-d
5’-d
5’-d
5’-d
5’-d
5’-d
5’-d
5’-d
5’-d
5’-d
5’-d
5’-d
N
AA
GA
GG
GC
AC
46.3
48.3
51.9
52.3
52.7
48.5
53.2
48.7
58.2
56.1
50.6
55.6
46.6
48.3
51.5
50.8
50.8
45.3
49.9
45.3
52.9
49.6
43.7
46.4
ꢀ0.3
0.0
0.4
1.5
1.9
3.2
3.3
3.4
5.3
6.5
6.9
9.2
TA
CA
TG
CC
CT
TT
CT,TA
[a] The natural sequences contain an adenine instead of (qA). [b] Duplexes were ob-
tained by annealing the sequences listed to their natural complementary oligonucleo-
tides. [c] Sequences are named by the bases neighbouring qA, in which the purines are
shown in bold and the pyrimidines in italic. [d] Samples were prepared in phosphate
increased between S2 and S1 (i.e., exciting at lower buffer (pH 7.5, 200 mm Na+, 17 mm phosphate).
Chem. Eur. J. 2012, 18, 5987 – 5997
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5991

L. M. Wilhelmsson et al.
ural duplexes (TmA) are listed in Table 1. In general qA has
a stabilising effect on the duplex structure, which translates
to an average rise in melting temperature of approximately
38C. Moreover, the extent of this stabilisation is dependent
on the bases flanking qA, which allows for selecting the de-
sired stability of a qA-containing duplex. When qA is
flanked by two purines (AA, GA and GG), the stability of
the duplex structure on average is equal to that of the natu-
ral duplexes. A pyrimidine flanking qA on the 3’ side (GC
and AC) causes a slight stabilisation of the duplexes (1.78C
on average), whereas a pyrimidine neighbouring on the
5’ side (TA, CA and TG) causes a moderate stabilisation
(3.38C on average). DNA duplexes containing qA flanked
by two pyrimidines have the highest stability (CC, CT and
TT), which is shown by an average rise in melting tempera-
ture of 6.28C compared to the natural duplexes. In general
we can conclude that the stability of the duplexes is in-
creased most when qA is flanked by a pyrimidine on the
5’ side.
When investigating the structural overlap of qA and its
neighbours in sequences CA and CC, an overlap between
the extended ring system of qA and the cytosine flanking it
on the 5’ side is observed. This overlap is not present for
natural adenine in the same sequence (data not shown). The
increased overlap arising when qA instead of natural A is
flanked by a pyrimidine on the 5’ side might explain the
higher stability of such duplexes. The overall stabilising
effect of qA can thus be explained by better p–p stacking
with its neighbours due to the extended ring system com-
pared to natural adenine. Furthermore, the additional rings
extended on C6 and N7 of adenine are well accommodated
in the major groove in contrast to extensions on the C8-po-
sition, which are known to destabilise the duplex struc-
ture.[24,33] It should also be noted that sequence CT,TA con-
taining two qAs shows a rise in melting temperature (9.28C)
that is virtually equal to the linear combination of the stabi-
lisations caused by qA in the sequences CT and TA (9.78C).
The UV melting results determined for qA are very simi-
lar to earlier observations for the tricyclic cytosine ana-
logues tC and tCO, containing a similar two-ring extension
towards the major groove.[12a,34] Both tC and tCO increase
the melting temperature of 10-mer duplexes by about 38C
under similar conditions (phosphate buffer, pH 7.5, [Na+]=
50 mm). Furthermore, as for qA, the most stable duplexes
are formed when they are flanked by a pyrimidine on the
5’ side. Also in the case of tC and tCO, a better overlap be-
tween the extended ring systems and the adjacent 5’ pyrimi-
dine has been suggested as the reason for the increased
duplex stability.[12a,34]
decamer).[15b] Also, another study involving 14-mer duplexes
confirmed the destabilising effect of 2-AP by a drop in melt-
ing temperature of 68C (15 mm sodium citrate, pH 7.25,
50 mm DNA duplexes).[15a] However, other studies have re-
ported a more limited destabilisation of duplexes due to in-
corporation of 2-AP.[17,35] Besides 2-AP, other adenine ana-
logues, such as the size-expanded adenine analogue xA,
cause a destabilisation of 4.2–4.98C when centrally substitut-
ed into 12-mer duplexes (100 mm NaCl, 10 mm MgCl2,
10 mm Na·PIPES, pH 7.0, 5.0 mm DNA).[11c] Also, the pteri-
dine adenine analogues, 6MAP and DMAP, on average
have shown a slight (2.48C) to moderate destabilisation
(4.68C) in 21-mer duplexes (10 mm NaCl, 10 mm Tris,
pH 7.5), respectively.[9] Furthermore, another purine ana-
logue, the pteridine guanine analogue 3-MI, shows an aver-
age drop in melting temperature of 8.38C, which is compa-
rable to a single base mismatch under similar conditions.[10b]
As for qA, the modified 7-deaza-adenine analogues devel-
oped by Seela et al. have also been shown to stabilise DNA
duplexes.[30,36] However, to the best of our knowledge, no re-
ports have been published regarding the fluorescence prop-
erties of these analogues inside DNA.
Circular dichroism: CD spectra of the samples listed in
Table 1 were used to evaluate any structural perturbations
to the B-DNA helix caused by the incorporation of qA
(Figure 5). All duplexes show a typical B-DNA signature,
which in the region between 180 and 300 nm, is character-
ised by a positive band at 275 nm, a negative band at
240 nm, a band that is less negative or positive at 220 nm
and a narrow negative band in the region between 220 and
190 nm, preceded by a large positive peak at 180–190 nm.
Surprisingly there is no specific signal corresponding to the
lowest energy absorption band of qA centred around
337 nm. For other common base analogues, such as tC or 2-
Importantly, qA is a good adenine analogue since it
causes no destabilisation of the DNA structure. On the con-
trary, it slightly stabilises the duplexes depending on the
neighbouring bases. This property distinguishes qA from
most other fluorescent base analogues. For example, the
widely used adenine analogue 2-aminopurine (2-AP) has
been shown to cause a destabilisation of 108C in 10-mer du-
plexes (100 mm KCl, 0.1 mm EDTA, 21 mm DNA duplexes,
Figure 5. CD spectra of the twelve qA modified duplexes. Duplexes are
named as the bases neighbouring qA (purines are shown in bold and pyr-
imidines in italic): AA (black), GG (red), GA (yellow), CT (light blue),
TG (pink), AC (light green), CC (grey), TA (brown), CA (dark blue),
GC (dark green), CT/TA (violet) and TT (orange). Duplexes were pre-
pared at a concentration of 2.5 mm in phosphate buffer (pH 7.5, 200 mm
Na+, 17 mm phosphate) and measured at 258C; m on the De axis refers
to the qA/duplex concentration (2.5 mm).
5992
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5987 – 5997

An Emissive Adenine Analogue
FULL PAPER
qA–X
Table 2. Melting temperatures of sequences containing quadracyclic adenine (Tm
)
aminopurine, a CD band appears for the low
energy transition.[34–35] However, similar observa-
tions to those made here for qA have been report-
ed for the structurally comparable cytosine ana-
logue tCO.[12a] We have not yet found a satisfactory
explanation for this phenomenon.
The CD signatures of all the modified duplexes
containing qA were compared to those of the corre-
sponding non-modified duplexes, and were shown
to be virtually identical (see the Supporting Infor-
mation). Overall it can be concluded that the incor-
poration of one or several qA molecules preserves
the overall B-DNA structure.
and natural adenine (TmA–X) opposite to guanine, adenine or cytosine (X). Melting
temperatures of matched sequences with qA (TmqA–T) and A (TmA–T) are compared
with the mismatched cases.
qA–T
qA–X
qA
A–T
A–X
A
Seq.[a]
X[b]
Tm
Tm
DTm
[8C]
Tm
Tm
DTm
[8C]
[8C][c]
[8C][c]
[8C][c]
[8C][c]
G
A
C
38.8
34.2
34.1
9.5
14.1
14.2
35.6
32.7
28.7
12.4
15.4
19.3
GA
48.3
48.3
G
A
C
47.6
40.5
39.3
8.5
15.6
16.8
38.4
33.1
31.5
11.1
16.4
18.1
CT
56.1
49.6
G
A
C
37.0
33.5
32.8
11.5
15.0
15.8
31.2
27.3
24.9
12.4
16.3
18.7
TA
48.5
45.3
Base pairing specificity of qA: To investigate the
base-pairing specificity of qA, we synthesised oligo-
nucleotides to combine with three of the modified
sequences (GA, CT and TA) to produce duplexes
containing an adenine, guanine or cytosine instead
of thymine opposite to qA. The sequences were
chosen to evaluate the influence of having purines,
pyrimidines or a combination of both directly neighbouring
mismatched qA. The melting temperatures of the modified
as well as the unmodified mismatched duplexes are listed in
Table 2. For comparison, the melting temperatures of the
matched duplexes with qA and A opposite to thymine are
also shown. When comparing the difference in melting tem-
perature between the modified matched and mismatched
duplexes (TmqA–T and TmqA–X) all three sequences are signifi-
cantly destabilised when qA is placed opposite to G, C or A
instead of thymine. This indicates that qA exhibits selective
base-pairing with thymine, and hence, is a specific adenine
analogue. In contrast, a less specific base-pairing pattern has
been reported for other fluorescent base analogues. The ad-
enine analogues 2-AP and triazole adenine, AT, for example,
are able to form stable base pairs with cytosine and adenine,
respectively, besides the conventional base pairing with thy-
mine.[16,24] Furthermore, the base discriminating fluorescent
analogue BPP can form stable base pairs both with adenine
and guanine.[4]
Interestingly, qA and natural adenine show the same
order of preference for base-pairing partners. When compar-
ing the destabilisations in Table 2, both for qA and A,
a more stable mismatch is formed with guanine compared to
the mismatches with cytosine and adenine. The same order
of thermal stability has been reported before in a study by
Gaffney and Jones for mismatches in a 9-mer oligonucleo-
tide.[37] The fact that qA shows the same base pairing prefer-
ence as natural adenine is an important indication that the
base-pairing pattern upon modification with qA is indeed
unperturbed.
[a] Sequences are named by the bases neighbouring qA and A. Full sequences are
listed in Table 1. [b] Base opposite qA or A in the mismatched sequences. [c] Samples
containing duplexes (2.5 mm) were prepared in phosphate buffer (pH 7.5, 200 mm Na+,
17 mm phosphate).
Table 3. Quantum yields of the mismatched (FqA–X) and matched (FqA–T
)
modified duplexes.
Seq.[a]
X[b]
FqA–T
[%][c,d]
FqA–X
[%][c,d,e]
G
A
C
3.0
1.6
1.6
GA
0.5
0.2
0.2
G
A
C
0.5
0.5
0.5
CT
G
A
C
0.4
0.5
0.4
TA
[a] Sequences are named by the bases neighbouring qA and A. Full se-
quences are listed in Table 1. [b] Base opposite of qA or A in the mis-
matched sequences. [c] Samples containing duplexes (2.5 mm) were pre-
pared in phosphate buffer (pH 7.5, 200 mm Na+, 17 mm phosphate).
[d] Fluorescence quantum yields were determined relative to the refer-
ence quinine sulfate in 0.5m H2SO4 (F=0.55) at 258C, and are averages
of at least two independent measurements. [e] Fluorescence quantum
yields determined at 138C were very similar (maximum 11% difference)
to values presented here.
ally equal to their corresponding mismatched duplexes.
Thus, we suggest that the increase in quantum yield might
be due to the weaker base-pairing between qA and the mis-
matching bases. This will lead to less rigid or less extensive
base-stacking and an increased opening rate for the mis-
matched qA base pairs relative to qA–T. In spite of these
perturbations, the CD spectra of the modified mismatched
and matched duplexes are virtually identical (data not
shown); this indicates that the overall B-DNA structure re-
mains unchanged in these mismatched 10-mers.
The quantum yields of the duplexes containing a qA mis-
match as well as the corresponding matched duplexes are
listed in Table 3. On average the quantum yield increases
threefold for the mismatched sequences compared to the
matched qA–T sequences. We have observed that the aver-
age quantum yield of qA in single strands is higher or virtu-
Spectroscopic and photophysical characterisation of qA in
DNA: The fluorescence quantum yields and lifetimes of qA
incorporated into the twelve single-stranded sequences, as
Chem. Eur. J. 2012, 18, 5987 – 5997
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5993

L. M. Wilhelmsson et al.
Table 4. Fluorescence quantum yields and average fluorescence lifetimes of the twelve single- and double-stranded oligonucleotides containing qA, re-
ferred to as ss and ds, respectively.
Neighbouring
basesss
Ff,ss
[%][b,c]
htiss
Absorption
maxss [nm][b]
Emission
Neighbouring
basesds
Ff,ds
[%][b,c]
htids
Absorption
Emission
[a]
[a]
[ns][b,d]
maxss [nm][b]
[ns][b,d]
maxds [nm][b]
maxds [nm][b]
AA
GG
GA
AC
GC
CA
TA
CT,TA
TG
CC
5.8
5.5
4.4
2.0
1.6
1.0
0.6
0.4
0.4
0.4
0.3
0.2
1.3
2.0
2.0
0.6
1.3
0.8
0.7
0.6
0.6
0.4
0.5
0.7
337
337
337
338
337
338
338
337
337
338
337
338
432
433
432
434
438
433
434
439
437
440
443
434
AA
GG
GC
GA
CA
CT,TA
AC
TA
TG
CT
0.6
0.6
0.6
0.5
0.3
0.3
0.2
0.2
0.2
0.2
0.2
0.2
0.7
0.9
1.4
0.5
0.5
1.3
0.4
0.6
0.6
1.3
0.8
0.7
334
336
339
336
337
337
337
337
334
331
337
336
434
435
453
438
444
457
441
451
453
459
452
452
CT
TT
CC
TT
[a] Full sequences are listed in Table 1. [b] Samples were prepared in phosphate buffer (pH 7.5, 200 mm Na+, 17 mm phosphate). [c] Fluorescence quan-
tum yields were determined relative to the reference quinine sulfate in 0.5m H2SO4 (Ff =0.55) at an excitation wavelength of 337 nm at 258C, and are
averages of three independent measurements. [d] Fluorescence decays were recorded by using an excitation wavelength of 330 nm and monitoring emis-
sion at 456 nm. For each sample, two independent measurements were obtained to confirm the reproducibility of the results. Decays were convoluted by
using the IRF and fitted to a bi- or tri-exponential function. Detailed fitting parameters are presented in the Supporting Information.
well as the corresponding duplexes are listed in Table 4.
Also, the absorption and emission maxima of each sequence
are presented. The emission maxima of qA in both single-
and double-stranded sequences (except duplexes CT and
CT,TA) show a blue-shift compared to the free qA mono-
mer in water (lem,max =456 nm). This is according to expecta-
tions upon introducing qA to a less polar environment since
TDDFT calculations suggest an increased dipole moment in
the excited state.[38] Furthermore, slight variations for single-
stranded samples (432–443 nm) and large shifts (434–
459 nm) for duplexes can be observed between various se-
quences. In contrast, the absorption maxima are virtually
the same for all single-stranded sequences (337–338 nm) and
only slightly red-shifted compared to the free qA monomer
(labs,max =335 nm). For the modified duplexes these numbers
show more variation, and range from 331 to 339 nm, with
slight red-shifts (except sequences CT, AA and TG) com-
pared to the qA monomer. Similar observations have been
made for the FBAs 2-AP[14] and 3-MI.[8] Moreover, red-
shifts are common for intercalating dyes, such as YO,
YOYO[39] and ethidium bromide.[40]
A general decrease in the quantum yields and average
fluorescence lifetimes are observed upon incorporation of
qA into oligonucleotides. The extent of the decrease is
highly dependent on the immediately surrounding bases
(vide infra). Lifetimes of qA inside DNA were fitted to two
or three exponentials, for which detailed fitting parameters
are shown in the Supporting Information. The average life-
times of qA vary between 0.4–2.0 ns in single strands and
between 0.4–1.4 ns in duplexes (Table 4). The average life-
times for all single- and double-stranded molecules (1.0 and
0.8 ns, respectively) are approximately three- and four-times
shorter, respectively, than the lifetime of the qA monomer
in water (t=3.2 ns; c2 =1.684). The observed general de-
creases in quantum yield and average lifetime upon incorpo-
ration into DNA are common features of most fluorescent
base analogues.[2a,6,10a,14,24] A range of theories have been
suggested in the literature to explain this phenomenon. It
has previously been shown that, for example, in single-
stranded systems, the base-stacking effect could produce sig-
nificant charge transfer by the molecular orbital overlap
with the surrounding natural bases.[41]
In the double-stranded structure there is also the effect of
the opposite strand to consider. Particularly in the case of 2-
AP, the results of Ward et al. suggest that this quenching
could be due to base-pairing.[14] Another explanation for the
decrease in fluorescence intensity and the shortening of
fluorescence lifetimes in a DNA-context could be the im-
proved accessibility of a conical intersection (CI) between
the excited state and the ground state[42] induced by the im-
mediate surroundings in the DNA. It has been proven that
such CIs occur in all natural nucleobases and in various flu-
orescent base analogues,[43] such as 2-AP.[44] Even though
this hypothesis needs to be verified by theoretical calcula-
tions on qA, it is possible that in the monomeric form, the
CI between S1 and S0 is not accessible to the excited popula-
tion because of energy barriers that occur on the potential
energy surface. These energy barriers could disappear when
qA is incorporated into DNA systems, and thus a larger part
of the excited population could relax non-radiatively
through the CI. This kind of mechanism could explain the
short time-components found here in the fitting parameters
(see the Supporting Information).
An average decrease in fluorescence quantum yield of
elevenfold was observed upon incorporation of the qA
monoACTHUNTGRENUNGmer (Fwater,lex337nm =6.8%) into single-stranded DNA
(Table 4). Moreover, qA shows a further sensitivity to its mi-
croenvironment as the quantum yields, average lifetimes
and emission maxima of the sequences differ substantially
depending on the bases neighbouring qA. As can be ob-
served in Table 4, qA shows the highest quantum yield in
oligonucleotides when it is surrounded by two purines (AA,
GG, GA), followed by the group of sequences with only one
purine neighbouring qA at the 5’ side (AC, GC) and by the
5994
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5987 – 5997

An Emissive Adenine Analogue
FULL PAPER
group with one purine on the 3’ side (CA, TA, TG). The
lowest quantum yield values were recorded for the group of
sequences (CC, CT and TT) with two pyrimidines flanking
qA.
corded for nucleic acids containing 2-AP,[14] and also slightly
higher compared to 6MAP.[9] In double-stranded DNA the
maximum value recorded for qA (0.6%) is comparable to
that of 2-AP.[14] The fluorescence decays of single- and
double-stranded systems containing qA, just as for 6-MAP
and DMAP, show more complex fitting parameters, and the
dominant time component of those decays is shorter than
that of the free monomer itself.[9] Furthermore, we recorded
fluorescence melting curves for duplex AA; this yielded
a virtually identical melting temperature to that recorded
with UV/Vis melting (Tm,UV =46.38C; Tm,F =46.88C). This
indicates no pre-melting at the qA-modified site and illus-
trates the applicability of qA as a fluorescent thymine-spe-
cific and non-perturbing adenine analogue.
For the corresponding duplexes, a further fourfold de-
crease in average quantum yields can be observed compared
to the single-stranded sequences (Table 4). Also, shorter
average lifetimes were observed for half of the sequences
(AA, GG, GA, AC, CA and TA). As for single strands, the
highest quantum yield values were observed among the
group of sequences with two purines flanking qA (AA, GG,
GA). The GC sequence can also be found among the higher
values, but all other double-stranded sequences show lower
and similar quantum yields (0.2–0.3%). It should be noted
that duplexes (AA, GG and GA) with two purines flanking
qA, having the same stability as natural duplexes, are the se-
quences that on average show the highest quantum yield
both in single- and double-stranded DNA. In contrast, se-
quences CC, TT and CT are the sequences that have the
highest duplex stability and the lowest quantum yields. This
can indicate a more firm stacking of qA with the neighbour-
ing bases in the latter sequences, and hence explain an in-
creased quenching efficiency.
The fact that GG is one of the sequences with the highest
quantum yield is in stark contrast to observations made for
the majority of fluorescent base analogues. Guanine is the
natural base with the lowest oxidation potential and, thus,
the highest susceptibility to oxidation.[45] This means that,
depending on the oxidation potential of the neighbouring
base analogue, G can transfer an electron into it, thereby
quenching the excited state by photo-induced electron trans-
fer (PET). For several common base analogues, such as the
adenine analogues 2-aminopurine,[41c] triazole adenine,[24] the
cytosine analogue tCO[12a] and the base-discriminating ana-
logue BPP,[4] a decreased quantum yield has been observed
due to neighbouring guanines. It has also been suggested
that the pteridine analogues 6MAP, DMAP and 3MI, act as
excited state acceptors, which are preferentially quenched
by purines.[46] We have not measured the redox potentials of
qA and therefore cannot draw any conclusions regarding
possible PET. One should remember though that even if the
redox potential of qA had been determined, structural pa-
rameters, such as stacking with neighbouring bases and rela-
tive orientation, are also expected to play an important role
for PET. Thus, knowledge of the redox potential of the qA
monomer would give a clue to possible PET but could not
tell the full story. However, as for many other FBAs we be-
lieve the quenching of qA inside DNA to be due to a combi-
nation of various mechanisms, all influenced by the sur-
rounding sequence and base stacking. Therefore, we are cur-
rently investigating the details of the reduced lifetimes using
a range of time-resolved techniques to obtain a more com-
plete picture about the complex quenching mechanisms.
Importantly, qA shows significant fluorescence inside
DNA compared to other adenine analogues. The maximum
quantum yield value reached in single strands (5.8%) is ap-
proximately four-times higher than corresponding values re-
Conclusion
In conclusion, quadracyclic adenine, qA, is an environment-
sensitive non-perturbing fluorescent adenine analogue. The
analogue is easy to selectively excite because its lowest
energy excitation maximum (335 nm) is well resolved from
the natural DNA absorption. Surprisingly, the quantum
yield of qA varies with excitation energy, despite the fact
that its emission always occurs from the S1 state. The free
monomer of qA is moderately fluorescent (Fwater,lex337nm
=
6.8%) in comparison with other adenine analogues, but
upon incorporation into oligonucleotides it reaches quantum
yield values of up to four-times higher than corresponding
values recorded for 2-AP. Furthermore, inside double-
stranded DNAs the quantum yield of qA is of the same
order as 2-AP in duplex systems.[14] Most importantly, the
DNA duplex stability is either enhanced or preserved de-
pending on the bases flanking qA. This is in contrast to the
destabilisation caused by other fluorescent adenine ana-
logues, such as 2-AP, 6MAP, DMAP, xA or triazole adenine,
and is therefore a promising feature.[9,11c,14,24] To the best of
our knowledge these properties of qA are unprecedented
for fluorescent adenine analogues. Additionally, the highly
selective base pairing between qA and thymine is an advan-
tageous feature for many purposes. For the most common
adenine analogue 2-AP and also for the recently developed
triazole adenine, a lack of specificity has been reported.[14,24]
In combination, these favourable properties of qA make it
an interesting alternative to 2-AP in fluorescent experiments
that require detailed measurements and a need to maintain
the integrity of the DNA double helix. As a result of the
promising properties found here, several derivatives of qA
are currently being synthesised, and are aimed at even
higher quantum yields but maintain the important ability to
form stable duplexes and specific pairing with thymine.
Experimental Section
General: The synthesis route to the quadracyclic adenine monomer is
presented in the results section (Scheme 1). Detailed synthesis proce-
Chem. Eur. J. 2012, 18, 5987 – 5997
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5995

L. M. Wilhelmsson et al.
dures, compound characterisation, incorporation of qA into oligonucleo-
tides and purification of oligonucleotides are described in Scheme S1 in
the Supporting Information.
late the anisotropy, r, from Equation (1), in which the instrument correc-
tion factor is given by Equation (2) and nd IXY is the excitation spectrum
for which the subscripts X and Y denote polarisation directions of the ex-
citation and emission light, respectively, and H and V refer to horizontal
and vertical, respectively.
Sample preparation: DNA samples were prepared in sodium phosphate
buffer (pH 7.5, 200 mm Na+, 17 mm phosphate) and their concentrations
were determined by recording the absorption at 260 nm. The extinction
coefficients of the various sequences were calculated by summation of
the extinction coefficients of the individual natural bases and qA. This
sum was multiplied by 0.9 to correct for base stacking interactions. The
extinction coefficient of qA was determined to be 10000mꢀ1 cmꢀ1 at
260 nm in water. This was obtained by measuring the absorption of a solu-
tion with a known concentration (86.4 mm) of qA. The extinction coeffi-
cients that were used for the different bases at 260 nm were eT =9300,
eC =7400, eG =11800, eA =15300 and eqA =10000mꢀ1 cmꢀ1. From these ex-
tinction coefficients, those for the different modified sequences were cal-
IVV ꢀ GIVH
ð1Þ
ð2Þ
r ¼
IVV þ 2GIVH
IHV
IHH
G ¼
The anisotropy of the immobile qA monomer in H2O/ethylene glycol
(1:2 mixture) glass at ꢀ1008C, also known as the fundamental anisotropy
(r0i) for a certain transition (i) is related to the angle between the absorb-
ing and the emitting transition moments (ai) as given by Equation (3).
culated to be: eGA =96800, eCT =87400, eGC =89600, eCA =92800, eGG
=
93600, eCC =85700, eTA =94500, eAA =99900, eAC =92800, eTG =91400,
eTT =93100 and eCT,TA =103500mꢀ1 cmꢀ1. Extinction coefficients of the un-
modified oligonucleotides were calculated in the same way.
1
5
ð3 cos2 ai ꢀ 1Þ
ð3Þ
roi
¼
UV and fluorescence melting experiments: Samples containing modified
duplex DNA (2.5 mm) were prepared by mixing buffered solutions con-
taining the single-stranded complementary strands, of which the non-
modified strand was in 15% excess. The excess was to assure hybridisa-
tion of all the modified strands for further fluorescence measurements.
For natural DNA duplexes, the same procedure was applied by using
equimolar amounts of single-stranded DNA. For UV melting experi-
ments, duplex samples were heated to 908C, and annealed by being
cooled to 58C at the rate of 18Cminꢀ1. Next, melting curves were record-
ed during heating of the samples to 908C, followed by cooling to 58C at
a rate of 0.58Cminꢀ1. The temperature was kept at 908C for 5 min be-
tween heating and cooling. The melting curves were measured on
a Varian Cary 4000 spectrophotometer with a programmable multi-cell
temperature block by using an absorption wavelength of 260 nm. Melting
temperatures were calculated by using the half maximum and the maxi-
mum of the first derivative of the melting curves. The values presented in
this paper are averages of at least two independent measurements. The
fluorescence melting curves for duplex AA were recorded on a Varian
Cary Eclipse instrument. Samples were annealed by being heated to
908C, followed by cooling to 158C at a rate of 0.58Cminꢀ1. Next, melting
curves were recorded while the samples were heated to 908C, followed
by cooling to 158C also at a rate of 0.58Cminꢀ1. The melting temperature
was calculated as the maximum of the first derivative of the melting
curve.
Time-resolved fluorescence measurements: Time-resolved fluorescence
measurements were performed by using time-correlated single photon
counting (TCSPC). The excitation source was the third harmonic (330 or
295 nm) of a Tsunami mode-locked Ti-sapphire laser running at 80 MHz.
The UV pulses (1–2 ps) were generated in a Tsunami frequency tripler
by using two 0.5 mm type I BBO crystals. The fluorescence of the sam-
ples was spectrally filtered at 456 nm by a monochromator and detected
by a microchannel plate photomultiplier Hamamatsu R3809U-50. The
counts were fed into a multichannel analyser with 4096 channels (Life-
spec, Edinburgh Analytical Instruments) where a minimum of 10000
counts were recorded in the top channel. All fluorescence decays were
recorded in a time window of 10 ns and the spectral resolution of the
monochromator was set from 5 to 11 nm. The intensity data were convo-
luted with the instrument response function (IRF), the full width half
maximum (FWHM) of which is given at 50 ps and fitted to two- or three-
exponential expressions with the program Fluofit Pro v.4 (PicoQuant,
GmbH) for the free monomer in solution and the modified single strands.
For the modified duplexes, a fitting program created by Thomas Gustavs-
son from the Francis Perrin laboratory at CEA-Saclay was used. This
uses a simple normalised analytical expression: F(t)=const+aꢁexp
t1)+(1ꢀa)ꢁbꢁexp(ꢀt/t2)+(1ꢀa)ꢁ(1ꢀb)ꢁcꢁexp(ꢀt/t3)+(1ꢀa)ꢁ(1ꢀb)ꢁ
(1ꢀc)ꢁexp(ꢀt/t4). To describe the fluorescence decays in a quantitative
A
A
ACHTUNGTRENNUNG
G
way, we performed a nonlinear fitting/deconvolution procedure using bi-
and triexponential functions, convoluted with the IRF. The average decay
time with j=1, 2 or 3 is given by Equation ().
Circular dichroism: CD spectra of 2.5 mm duplex samples, which were
prepared as described above, were recorded on a Chirascan CD spectro-
photometer. The spectra were recorded at 258C for a wavelength range
of 200 to 420 nm. Corrections for background contributions were made
and final spectra are averages of at least five measurements at a scan
j
P
aiti
i¼1
hti ¼
ð4Þ
rate of 2 nmsꢀ1
.
j
P
ai
Fluorescence measurements
i¼1
Steady state fluorescence measurements: For all steady-state fluorescence
measurements, a SPEX fluorolog 3 spectrofluorimeter (JY Horiba) was
employed. Buffered solutions of qA-modified DNA duplexes were pre-
pared as described above (pH 7.5, 200 mm Na+, 17 mm phosphate). Sam-
ples containing the monomer in Milli-Q water or in methanol and buf-
fered solutions of single-stranded modified oligonucleotides were set to
an absorption of approximately 0.03 at the excitation wavelength. Quan-
tum yields were determined relative to the quantum yield of quinine sul-
fate (Ff =0.55) in H2SO4 (0.5m) at 258C.[47] For the duplexes and single-
stranded oligonucleotides an excitation wavelength of 337 nm was used
and emission spectra were recorded between 340 and 700 nm. The quan-
tum yield of the qA monomer in Milli-Q water and methanol was record-
ed as a function of seven different excitation wavelengths between 295
and 377 nm. The emission of the quinine sulfate reference was recorded
between 305 and 700 nm.
Acknowledgements
This work was supported by the Swedish Research Council and Olle
Engkvist Byggmꢂstare Foundation. Furthermore, we would like to thank
Thomas Gustavsson from the Francis Perrin Laboratory CEA-Saclay in
France for the use of his software to fit the fluorescence decays of the
modified duplexes.
[1] D. Onidas, D. Markovitsi, S. Marguet, A. Sharonov, T. Gustavsson,
J. Phys. Chem. B 2002, 106, 11367–11374.
[2] a) L. M. Wilhelmsson, Q. Rev. Biophys. 2010, 43, 159–183; b) R. W.
Sinkeldam, N. J. Greco, Y. Tor, Chem. Rev. 2010, 110, 2579–2619.
[3] S. Preus, K. Kilsꢃ, L. M. Wilhelmsson, B. Albinsson, Phys. Chem.
Chem. Phys. 2010, 12, 8881–8892.
Steady-state excitation spectra were recorded by exciting between 240
and 450 nm and monitoring the emission fixed at 456 nm. Glan polarisers
were used to obtain the polarised excitation spectra necessary to calcu-
5996
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5987 – 5997

An Emissive Adenine Analogue
FULL PAPER
[4] A. Okamoto, Y. Saito, I. Saito, J. Photochem. Photobiol. C 2005, 6,
108–122.
[23] C. A. Buhr, M. D. Matteucci, B. C. Froehler, Tetrahedron Lett. 1999,
40, 8969–8970.
[5] a) N. J. Greco, Y. Tor, J. Am. Chem. Soc. 2005, 127, 10784–10785;
b) S. G. Srivatsan, Y. Tor, J. Am. Chem. Soc. 2007, 129, 2044–2053;
c) S. G. Srivatsan, N. J. Greco, Y. Tor, Angew. Chem. 2008, 120,
6763–6767; Angew. Chem. Int. Ed. 2008, 47, 6661–6665; d) S. G. Sri-
vatsan, H. Weizman, Y. Tor, Org. Biomol. Chem. 2008, 6, 1334–
1338.
[6] D. A. Berry, K. Y. Jung, D. S. Wise, A. D. Sercel, W. H. Pearson, H.
Mackie, J. B. Randolph, R. L. Somers, Tetrahedron Lett. 2004, 45,
2457–2461.
[7] a) F. Wojciechowski, R. H. E. Hudson, J. Am. Chem. Soc. 2008, 130,
12574–12575; b) R. H. E. Hudson, R. D. Viirre, Y. H. Liu, F. Wojcie-
chowski, A. K. Dambenieks, Pure Appl. Chem. 2004, 76, 1591–1598.
[8] S. L. Driscoll, M. E. Hawkins, F. M. Balis, W. Pfleiderer, W. R. Laws,
Biophys. J. 1997, 73, 3277–3286.
[9] M. E. Hawkins, W. Pfleiderer, O. Jungmann, F. M. Balis, Anal. Bio-
chem. 2001, 298, 231–240.
[10] a) M. E. Hawkins, Cell Biochem. Biophys. 2001, 34, 257–281;
b) M. E. Hawkins, W. Pfleiderer, F. M. Balis, D. Porter, J. R. Knut-
son, Anal. Biochem. 1997, 244, 86–95.
[11] a) A. T. Krueger, E. T. Kool, J. Am. Chem. Soc. 2008, 130, 3989–
3999; b) H. B. Liu, J. M. Gao, E. T. Kool, J. Am. Chem. Soc. 2005,
127, 1396–1402; c) J. M. Gao, H. B. Liu, E. T. Kool, J. Am. Chem.
Soc. 2004, 126, 11826–11831.
[12] a) P. Sandin, K. Bçrjesson, H. Li, J. Mꢃrtensson, T. Brown, L. M.
Wilhelmsson, B. Albinsson, Nucleic Acids Res. 2008, 36, 157–167;
b) P. Sandin, L. M. Wilhelmsson, P. Lincoln, V. E. C. Powers, T.
Brown, B. Albinsson, Nucleic Acids Res. 2005, 33, 5019–5025.
[13] K. Bçrjesson, S. Preus, A. H. El-Sagheer, T. Brown, B. Albinsson,
L. M. Wilhelmsson, J. Am. Chem. Soc. 2009, 131, 4288–4293.
[14] D. C. Ward, E. Reich, L. Stryer, J. Biol. Chem. 1969, 244, 1228–
1237.
[24] A. Dierckx, P. Dinꢄr, A. H. El-Sagheer, J. D. Kumar, T. Brown, M.
Grøtli, L. M. Wilhelmsson, Nucleic Acids Res. 2011, 39, 4513–4524.
[25] M. Hoffer, Chem. Ber. 1960, 93, 2777–2781.
[26] Z. Kazimierczuk, H. B. Cottam, G. R. Revankar, R. K. Robins, J.
Am. Chem. Soc. 1984, 106, 6379–6382.
[27] J. M. Muchowski, M. C. Venuti, J. Org. Chem. 1980, 45, 4798–4801.
[28] S. P. H. Mee, V. Lee, J. E. Baldwin, Angew. Chem. 2004, 116, 1152–
1156; Angew. Chem. Int. Ed. 2004, 43, 1132–1136.
[29] F. Seela, M. Zulauf, M. Sauer, M. Deimel, Helv. Chim. Acta 2000,
83, 910–927.
[30] F. Seela, M. Zulauf, Chem. Eur. J. 1998, 4, 1781–1790.
[31] S. Preus, K. Bçrjesson, K. Kilsꢃ, B. Albinsson, L. M. Wilhelmsson, J.
Phys. Chem. B 2010, 114, 1050–1056.
[32] F. A. Miannay, T. Gustavsson, A. Banyasz, D. Markovitsi, J. Phys.
Chem. A 2010, 114, 3256–3263.
[33] R. G. Eason, D. M. Burkhardt, S. J. Phillips, D. P. Smith, S. S. David,
Nucleic Acids Res. 1996, 24, 890–897.
[34] K. C. Engman, P. Sandin, S. Osborne, T. Brown, M. Billeter, P. Lin-
coln, B. Nordꢄn, B. Albinsson, L. M. Wilhelmsson, Nucleic Acids
Res. 2004, 32, 5087–5095.
[35] N. P. Johnson, W. A. Baase, P. H. von Hippel, Proc. Natl. Acad. Sci.
USA 2004, 101, 3426–3431.
[36] F. Seela, H. Thomas, Helv. Chim. Acta 1995, 78, 94–108.
[37] B. L. Gaffney, R. A. Jones, Biochemistry 1989, 28, 5881–5889.
[38] J. R. Lakowicz, in Principles of Fluorescence Spectroscopy, 3rd ed.,
Springer, New York, 2006, pp. 205–235.
[39] A. Larsson, C. Carlsson, M. Jonsson, B. Albinsson, J. Am. Chem.
Soc. 1994, 116, 8459–8465.
[40] M. J. Waring, J. Mol. Biol. 1965, 13, 269–282.
[41] a) J. N. Wilson, Y. J. Cho, S. Tan, A. Cuppoletti, E. T. Kool, Chem-
BioChem 2008, 9, 279–285; b) J. M. Jean, K. B. Hall, Proc. Natl.
Acad. Sci. USA 2001, 98, 37–41; c) O. J. G. Somsen, v. A. Hoek,
v. H. Amerongen, Chem. Phys. Lett. 2005, 402, 61–65.
[42] Conical Intersections: Electronic Structure, Dynamics & Spectrosco-
py (Eds.: W. Domcke, D. R. Yarkony, H. Kçppel), World Scientific
Publishing, 2004.
[43] “Quantum Mechanical Studies of the Photophysics of DNA and
RNA Bases”: K. A. M. Kistler, S. Matsika, in Challenges and Advan-
ces in Computational Chemistry and Physics, Vol. 7 (Ed.: T.-S. Y.
Lee, D. M. York), Springer, Heidelberg (Germany), 2009, pp. 285–
339.
[44] a) S. Perun, A. L. Sobolewski, W. Domcke, Mol. Phys. 2006, 104,
1113–1121; b) J. X. Liang, S. Matsika, J. Am. Chem. Soc. 2011, 133,
6799–6808.
[45] S. Fukuzumi, H. Miyao, K. Ohkubo, T. Suenobu, J. Phys. Chem. A
2005, 109, 3285–3294.
[15] a) O. V. Petrauskene, S. Schmidt, A. S. Karyagina, I. I. Nikolskaya,
E. S. Gromova, D. Cech, Nucleic Acids Res. 1995, 23, 2192–2197;
b) D. G. Xu, K. O. Evans, T. M. Nordlund, Biochemistry 1994, 33,
9592–9599.
[16] a) S. M. Watanabe, M. F. Goodman, Proc. Natl. Acad. Sci. USA
1982, 79, 6429–6433; b) L. C. Sowers, Y. Boulard, G. V. Fazakerley,
Biochemistry 2000, 39, 7613–7620; c) S. M. Law, R. Eritja, M. F.
Goodman, K. J. Breslauer, Biochemistry 1996, 35, 12329–12337.
[17] A. Dallmann, L. Dehmel, T. Peters, C. Mugge, C. Griesinger, J.
Tuma, N. P. Ernsting, Angew. Chem. 2010, 122, 6126–6129; Angew.
Chem. Int. Ed. 2010, 49, 5989–5992.
[18] R. P. Bandwar, S. S. Patel, J. Biol. Chem. 2001, 276, 14075–14082.
[19] E. Deprez, P. Tauc, H. Leh, J. F. Mouscadet, C. Auclair, M. E. Haw-
kins, J. C. Brochon, Proc. Natl. Acad. Sci. USA 2001, 98, 10090–
10095.
[20] K. E. Augustyn, K. Wojtuszewski, M. E. Hawkins, J. R. Knutson, I.
Mukerji, Biochemistry 2006, 45, 5039–5047.
[46] M. Narayanan, G. Kodali, Y. J. Xing, M. E. Hawkins, R. J. Stanley, J.
Phys. Chem. B 2010, 114, 5953–5963.
[21] P. Sandin, J. Tumpane, K. Bçrjesson, L. M. Wilhelmsson, T. Brown,
B. Nordꢄn, B. Albinsson, P. Lincoln, J. Phys. Chem. C 2009, 113,
5941–5946.
[47] W. H. Melhuish, J. Phys. Chem. 1961, 65, 229–235.
[22] a) S. D. Gilbert, C. D. Stoddard, S. J. Wise, R. T. Batey, J. Mol. Biol.
2006, 359, 754–768; b) J. F. Lemay, J. C. Penedo, R. Tremblay,
D. M. J. Lilley, D. A. Lafontaine, Chem. Biol. 2006, 13, 857–868.
Received: October 31, 2011
Published online: March 21, 2012
Chem. Eur. J. 2012, 18, 5987 – 5997
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5997