DNA duplexes and triplex-forming oligodeoxynucleotides incorporating modified nucleosides forming stable and selective triplexes

Article abstract of DOI:10.1039/c1ob06411h

We have previously reported DNA triplexes containing the unnatural base triad G-PPI·C3, in which PPI is an indole-fused cytosine derivative incorporated into DNA duplexes and C3 is an abasic site in triplex-forming oligonucleotides (TFOs) introduced by a propylene linker. In this study, we developed a new unnatural base triad A-ψ·C^R1 where ψ and C^R1 are base moieties 2′-deoxypseudouridine and 5-substituted deoxycytidine, respectively. We examined several electron-withdrawing substituents for R1 and found that 5-bromocytosine (C ^Br) could selectively recognize ψ. In addition, we developed a new PPI derivative, PPI^Me, having a methyl group on the indole ring in order to achieve selective triplex formation between DNA duplexes incorporating various Watson-Crick base pairs, such as T-A, C-G, A-ψ, and G-PPI^Me, and TFOs containing T, C, C^Br, and C3. We studied the selective triplex formation between these duplexes and TFOs using UV-melting and gel mobility shift assays.

Full text of DOI:10.1039/c1ob06411h

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2012, 10, 1007
www.rsc.org/obc
PAPER
DNA duplexes and triplex-forming oligodeoxynucleotides incorporating
modified nucleosides forming stable and selective triplexes†
Takashi Kanamori, Yoshiaki Masaki, Masahiro Mizuta, Hirosuke Tsunoda, Akihiro Ohkubo, Mitsuo Sekine*
and Kohji Seio*
Received 18th August 2011, Accepted 19th October 2011
DOI: 10.1039/c1ob06411h
We have previously reported DNA triplexes containing the unnatural base triad G-PPI·C3, in which
PPI is an indole-fused cytosine derivative incorporated into DNA duplexes and C3 is an abasic site in
triplex-forming oligonucleotides (TFOs) introduced by a propylene linker. In this study, we developed a
new unnatural base triad A-y·CR1 where y and CR1 are base moieties 2¢-deoxypseudouridine and
5-substituted deoxycytidine, respectively. We examined several electron-withdrawing substituents for R1
and found that 5-bromocytosine (CBr) could selectively recognize y. In addition, we developed a new
PPI derivative, PPIMe, having a methyl group on the indole ring in order to achieve selective triplex
formation between DNA duplexes incorporating various Watson–Crick base pairs, such as T-A, C-G,
A-y, and G-PPIMe, and TFOs containing T, C, CBr, and C3. We studied the selective triplex formation
between these duplexes and TFOs using UV-melting and gel mobility shift assays.
Introduction
DNA triplexes are supramolecular structures formed by sequence-
specific recognition of DNA duplexes by triplex-forming oligonu-
cleotides (TFOs).1 Many studies have been reported on the
application of DNA triplex formation to gene regulation,2 gene
correction,3 gene detection,4 and DNA nanotechnology.5
TFOs can bind to DNA duplexes in either of the two orien-
tations: parallel and anti-parallel modes. In a parallel triplex,
TFOs recognize DNA duplexes that form Hoogsteen-type base
pairs1 where thymine (T) and protonated cytosine (C+) recognize
adenine (A) and guanine (G) in the major groove of the duplex,
respectively (Fig. 1A, B). In this paper, the Watson–Crick base
pairs in the duplex are depicted with hyphens (e.g., T-A, C-G),
while the Hoogsteen base pairs and other interactions between a
DNA duplex and TFO are represented using dots (e.g., A·T and
G·C+). Note that in natural-type DNAs, there are no Hoogsteen-
type base pair structures that allow binding with the pyrimidine
bases in duplexes. Thus, when natural-type DNA duplexes and
TFOs are used, duplexes that can bind to TFOs are restricted to
those with homopyrimidine–homopurine sequences, and the TFO
should be composed of pyrimidine bases.
A
number of studies for developing unnatural nucleic
acids to overcome this sequence limitation have been
reported.6–8 For example, Fox et al. reported that all four
Fig. 1 The four base triads introduced in this study. A) Canonical
C-G·C+, B) T-A·T, C) A-y·CR1 (R1 = F, Cl, Br, CN), and D) G-PPIR2·C3
(R2 = H, Me, OMe).
4259, Nagatsuta, Midori-ku, Yokohama, Japan. E-mail: kseio@bio.
titech.ac.jp, msekine@bio.titech.ac.jp; Fax: +81-45-924-5144; Tel: +81-45-
924-5136
† Electronic supplementary information (ESI) available: Synthetic
schemes, procedures, 1H, 13C and 31P NMR spectra of newly synthesized
compounds.. See DOI: 10.1039/c1ob06411h
base pairs (A-T, G-C, C-G, and T-A) could be recog-
nized by TFOs containing four different unnatural bases in-
cluding BAU [2¢-aminoethoxy-5-(3-aminoprop-1-ynyl)uridine],6
This journal is
The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 1007–1013 | 1007
©

View Article Online
MeP (3-methyl-2-aminopyridine),9 APP (6-(3-aminopropyl)-7-
methyl-3H-pyrrolo[2,3-d]pyrimidin-2(7H)-one),6 and S [N-(4-(3-
acetamidophenyl)thiazol-2-yl-acetamide)].10 They designed APP
and S as unnatural nucleobases that recognize C and T in the
duplexes, respectively. In addition to this approach, the recognition
of pyrimidine bases in duplexes by modified bases in TFOs can
be theoretically achieved by incorporating unnatural bases into
the DNA duplexes. This concept was first proposed by Trapane
and Ts’o et al. who reported the design of various base triads11
and later reported experimental evidence of the formation of the
triad adenine (A)-pseudouracil (uracil-5-yl, y)·cytosine (C) (Fig.
3C).12 Similarly, Switzer et al. reported another A-y·A triad.13 In
addition, Ganesh et al. reported DNA duplexes incorporating 5-
aminouracil (U#),14,15 which is recognized by adenine in TFOs.12
We have reported a nucleotide triad with PPI, an indole-fused
cytosine derivative, incorporated into DNA duplexes as shown in
Fig. 1D (R2 = H) and Fig. 2.16 PPI can form a Watson–Crick
base pair with guanosine in DNA duplexes, and its indole ring,
which extrudes into the major groove, can selectively interact with
an abasic site in TFOs containing a propylene linker (C3). This
triad is stabilized by stacking interactions between PPI and the
upstream and downstream nucleobases of C3 in TFOs without
forming hydrogen bonds. These results clearly suggest that the
combined use of unnatural pyrimidine residues such as y, U#, and
PPI incorporated into DNA duplexes and appropriately designed
TFOs is a promising strategy for expanding the base triad codes
of DNA triplexes.
type base triads and developed a set of four base triads capable
of sequence-selective triplex formation. As a result, we found that
the combined use of CBr and PPIMe gave the best result in terms of
stability and selectivity upon DNA triplex formation.
Results and discussion
Design of CR1 and properties of the corresponding nucleosides
We designed the 5-substituted cytosine analog CR1, which can be
incorporated into TFOs by standard phosphoramidite chemistry
and can recognize the A-y base pair in the duplexes. Ts’o et al.12
previously reported that the cytosine base (C) can bind to A-
y to give the triad A-y·C. However, in our case, C itself could
not be used in TFOs because our system was designed to use its
protonated form (C+) for recognizing G in the duplex, as shown in
Fig. 1A. This combination forms a very stable G·C+ Hoogsteen
base pair. Thus, we decided to develop less basic analogs of C,
which could form stable base pairs with y, as shown in Fig. 1C
and 3D, but could not form stable Hoogsteen-type base pairs with
G because they would remain in an unprotonated form under
weakly acidic conditions (Fig. 3B).
Fig. 3 Relative stabilities among A) C-G·C+ matched triad, B) C-G·CR1
mismatched triad, C) A-y·C mismatched triad, and D) A-y·CR1 (R1 = Br,
Cl, F, or CN) matched triad indicated by the inequality signs.
Fig. 2 3D structure of the DNA triplex incorporating a G-PPI·C3 triad
and the extracted structure of the G-PPI·C3 triad obtained by the MD
simulation. Here, PPI and C3 are shown in CPK style.
For this purpose, we designed several cytosine analogs (CR1)
having electron-withdrawing groups such as cyano, fluoro, chloro,
and bromo at the 5-position of the cytosine ring. By this
modification, the basicity at N3 of the cytosine ring would be
decreased, thereby avoiding protonation at this position. We
expected that this unprotonated CR1 would bind more tightly to
y than to G because y·CR1 formed two hydrogen bonds (Fig.
3D), while G·CR1 had only one (Fig. 3B). In addition, the triplex
incorporating y·CR1 would be more stable than that incorporating
y·C because of the stacking interactions between the R1 group and
the nearby bases (Fig. 3C and 3D).
As the next step, we aimed to develop a set of four base triads
including T-A·T, C-G·C+, A-y·CR1, and G-PPIR2·C3, where CR1
and PPIR2 are a 5-substituted cytosine derivative and an indole-
fused cytosine derivative, respectively (Fig. 1C, D). In theory,
different 4n n-mer TFOs and 4n DNA duplexes that form DNA
triplexes in a sequence-selective manner can be designed using the
four base triads.
In this paper, we report the triplex formation between TFOs
containing various modified cytosine bases CR1 (R1 = Br, F, Cl,
and CN) and DNA duplexes incorporating y. In addition, we
describe the interaction of new PPI derivatives, PPIR2 (R2 = CH3
and OCH3), in DNA duplexes with C3 in TFOs. Finally, we
combined these unnatural base triads with canonical Hoogsteen-
To confirm these expectations, we first checked the basicity of
the deoxynucleosides dCBr (1a),17,18 dCCl (1b),18,19,20 dCCN (1c),21,22
and dCF (1d)23 having CR1 as the aglycon (Fig. 4). The reported pKa
1008 | Org. Biomol. Chem., 2012, 10, 1007–1013
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Fig. 5 Sequences of hairpin DNAs incorporating y and PPI derivatives
(left) and TFOs incorporating CR1 derivatives (right).
Stability and selectivity of triplexes of hairpin oligode-
oxynucleotides incorporating w and TFOs incorporating CR1
Using these oligonucleotides, UV-melting temperatures of various
DNA triplexes were measured. The data are shown in Box-1 in
Table 1, where the underlined values are those of the matched base
triads C-G·C+ and A-y·CR1. Initially, we checked the stability
of matched HP-aw/TFO-F and mismatched HP-aw/TFO-C in
order to clarify the effect of the electron-withdrawing group at the
5-position of cytosine. As shown in Box-1 in Table 1, the triplex
with TFO-C (C at Z position) bound to HP-aw (A-y base pair at
the X-Y positions) had a Tm of 42 ◦C, and the triplex with TFO-
F bound to HP-aw demonstrated the same affinity (Tm = 42 ◦C).
These results suggested that the electron-withdrawing fluoro group
at the 5-position of cytosine did not decrease the hydrogen bond
energy between y and CF.
Fig. 4 Structures of the nucleosides and their phosphoramidites having
CR1 and PPIR2 bases.
values of dCF and dCBr are 2.724 and 2.8,25 respectively. The pKa of
dCCN was determined to be 1.7 by a UV-titration experiment (see
Supporting Information†). Considering the previous result26 that
pKa of deoxycytidine derivatives is proportional to the Hammett
constants of the substituents at the 5-position, dCCl was also
expected to be less basic than dC. These pKa values indicated that
when incorporated in TFOs, these modified deoxycytidines would
favor the unprotonated forms under the conditions used for DNA
triplex formation (pH 5.4) and form a weak hydrogen bond with
the guanine residue in the duplex, as shown in Fig. 3B. To confirm
this hypothesis, we studied the stability and selectivity of A-y·CR1
base triads, where CR1 is CF, CCl, CBr, or CCN, by measuring the Tm
values of the DNA triplexes incorporating them.
Next, we checked the stability of the mismatched triplex HP-
cg/TFO-F. TFO-F had a binding affinity for HP-cg giving Tm
=
42 ◦C, which was 10 ◦C lower than that of the HP-cg/TFO-
C triplex (Tm = 52 ◦C). This result is consistent with the above
mentioned hypothesis that the less basic CF, which does not favor
the protonated form under mild acidic conditions, cannot form a
stable Hoogsteen-type base pair with G (Fig. 3B).
Subsequently, we incorporated CCN, CCl, and CBr into TFOs at
the Z position. These modified residues were expected to bind
more securely than CF to HP-aw because of the stacking effect
of their larger substituents (CN, Cl, Br > F). The HP-aw/TFO-
◦
CN triplex had a Tm value of 45 C, which was higher than that
of the HP-aw/TFO-F triplex (Tm = 42 ◦C). In addition, TFO-
CN showed a stronger binding affinity for HP-aw than for HP-
cg, (DTm = +7 ◦C, Table 1). Similarly, TFO-Cl and TFO-Br also
showed higher affinities for HP-aw (Tm = 48 ◦C, both TFOs) than
HP-cg (Tm = 45 ◦C, both TFOs).
Synthesis of phosphoramidite units and oligodeoxynucleotides for
the study of A-w·CR1 base triads
To measure the thermal stability of the DNA triplexes, we
synthesized hairpin-type DNA duplexes, HP-cg and HP-aw (Fig.
5), incorporating C-G and A-y base pairs, respectively, at the X-Y
positions. These hairpin DNAs were synthesized on an automated
synthesizer using commercially available phosphoramidite units.
We also synthesized TFO-F, TFO-Cl, TFO-Br, TFC-CN, and
TFO-C, where the residues at the Z position are CF, CCl, CBr,
CCN, and C, respectively. In these HP and TFO combinations,
nucleobases at positions X, Y, and Z are arranged to form the
base triad X-Y·Z.
Among the 5-substituted deoxycytidine, CCN had the highest
selectivity for the A-y pair over the C-G pair (DTm = 7 ◦C),
while both CCl and CBr were less selective (DTm = 3 ◦C). However,
considering the affinity for the A-y base pair independently,
both CCl and CBr triplexes showed a higher Tm (48 ◦C) than
that of the CCN triplex (45 ◦C). Because TFO-Cl and TFO-Br
showed identical Tm profiles, we used only CBr for the subsequent
studies.
For the synthesis of these TFOs, the phosphoramidites of
dCBr (2a), dCCl (2b), and dCCN (2c) were synthesized from
nucleosides 1a–c, as described in the Supporting Information†.
The phosphoramidite of dCF (2d) was synthesized according to
the known procedure.23
Studies using CCN and CBr in combination with the G-PPI·C3 triad
Using the above A-y·CR1 base triad where CR1 is CCN or CBr,
we compared the stability of the four matched base triads
This journal is
The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 1007–1013 | 1009
©

View Article Online
Table 1 Tm values [◦C] of various triplexesa
a n.t. not tested
(T-A·T, C-G·C+, A-y·CR1 and G-PPI·C3) and all possible mis-
matched triads. To this end, we synthesized the hairpin duplexes
HP-ta and HP-gppi and the TFOs TFO-T and TFO-C3 as
shown in Fig. 5. The PPI residue was incorporated using the
previously reported phosphoramidite unit 2e16 (see the Support-
ing Information†) and C3 was introduced using commercially
available 3-(4, 4¢-dimethoxytrityloxy)propyl 2-cyanoethyl N,N-
diisopropylphosphoramidite.
Stability and base recognition of triplexes incorporating PPIR2 and
C3
We attempted to tune the base pairing property of PPI by
introducing a substituent into its indole ring. Before the synthesis,
we used a molecular dynamics simulation of a DNA triplex
incorporating a G-PPI·C3 triad to identify the positions where
the introduction of a substituent reduces the stability of the
mismatched triads by steric repulsion, but does not significantly
reduce the stability of the matched G-PPIR2·C3 triad. The
3D structure obtained by the MD simulation16 is shown in
Fig. 2.
Using the hairpin duplexes and TFOs as mentioned above, we
measured the Tm values of all possible triplexes. The data are
shown in Box-2 in Table 1.
In Box-2 in Table 1, the underlined values are those of the
matched base triads C-G·C+, T-A·T, G-PPI·C3, and A-y·CR1.
The Tm values were compared to evaluate the sequence selectivity
of each hairpin DNA for the individual TFOs and of each TFO for
the different hairpin DNAs. Considering the former selectivities
(see Tm values in Box-2 in Table 1 horizontally), each matched
Tm value is higher than the mismatched ones. In contrast, in the
case of selectivities of TFOs for hairpin DNAs (see Tm values
in Box-2 in Table 1 vertically), it was found that TFO-Br and
TFO-CN did not selectively bind to HP-aw (Tm = 48 and 45 ◦C,
As is evident, the indole ring of PPI fits the space formed by
the C3 linker and the upstream and downstream bases in the TFO
strand. According to the 3D structure, the 5-position (shown by a
line in Fig. 2) could be the potential site for modification because
substitution at this position does not seem to interfere with binding
to the TFO containing C3. Therefore, we designed two deoxynu-
cleosides having PPI derivatives as the aglycon, including dPPIMe
(1f) and dPPIOMe (1g) incorporating a methyl and methoxy group,
respectively, at the 5-position (Fig. 4). The synthetic protocols
for dPPIOMe (1g) and its phosphoramidite have been previously
described27 and dPPIMe was newly synthesized according to a
procedure similar to that in our previous papers,16,27,28 as shown in
Scheme 1.
Briefly, 5-iododeoxycytidine17 was converted into the indole
nucleoside 1f by treatment with 5-methylindole-boronic acid in the
presence of a palladium catalyst and triphenylphosphine trisul-
fonate (TPPTS). DMTr protection followed by phosphitylation
gave the phosphoramidite 2f in 73% yield.
respectively) but instead bound more tightly to HP-gppi (Tm
=
49 ◦C, both TFOs). This is because PPI-related mismatches are
thermally stable (Tm is from 45 ◦C to 49 ◦C). Therefore, we
next attempted to modify the structure of PPI to PPIR2 by the
introduction of a substituent (R2) to suppress the Tm of these
mismatched triads so that all the matched triads became more
stable than the corresponding mismatched ones. For the A-y
pair, we selected CBr, which bound more tightly to A-y than
CCN, expecting that TFO-Br can bind more tightly to HP-aw
than to the modified HP-gppiR2 which incorporates PPIR2 in place
of PPI.
The hairpin duplexes HP-gppiMe and HP-gppiOMe were synthe-
sized using the phosphoramidites 2f and 2g so that the Tm values
of their triplexes with various TFOs could be measured. The
hairpin DNA sequences are shown in Fig. 5 and the Tm data are
listed in Box-3 in Table 1. As expected, the results show that HP-
gppiMe incorporating PPIMe demonstrated lower affinities for the
mismatched TFOs such as TFO-C and TFO-Br than did HP-gppi
We anticipated that the mismatched trip◦lexes incorporating
modified PPIR2 would have a Tm lower than 48 C so that sequence-
selective triplex formation would be achieved in all possible
patterns.
1010 | Org. Biomol. Chem., 2012, 10, 1007–1013
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
Fig. 6 Sequences of hairpin DNAs and TFOs incorporating natural and
unnatural bases. B = CBr; 3 = C3; P = PPIMe. The underlined bases are not
identical between HP-1/TFO-1 and HP-2/TFO-2.
Scheme 1 Synthesis of dPPIMe (1f) and its phosphoramidite 2f.
incorporating unsubstituted PPI. Tm decreased from 49 ◦C (HP-
gppi/TFO-C and HP-gppi/TFO-Br) to 45 ◦C (HP-gppiMe/TFO-C
and HP-gppiMe/TFO-Br), which is a decrease of 4 ◦C. In the case of
another mismatched TFO (TFO-T), the◦HP-gppi and HP-gppiMe
triplexes gave identical Tm values of 45 C. Although its affinity
for matched TFO-C3 was also decreased by the incorporation of
the methyl group from 56 ◦C (HP-gppi/TFO-C3) to 53 ◦C (HP-
gppiMe/TFO-C3), HP-gppiMe maintained the largest Tm value with
matched TFO-C3 compared to mismatched TFO-C, TFO-Br, and
TFO-T strands (compare second-row values horizontally). Thus,
we concluded that PPIMe had a higher affinity for the matched TFO
incorporating C3 and lower affinity for the mismatched TFOs.
In the case of HP-gppiOMe, the Tm profiles are similar to those
of HP-gppiMe. Thus, we selected PPIMe for further studies.
HP-1/TFO-2 and HP-2/TFO-1. The triplex HP-1/TFO-2 has
two C-G·C3 and A-y·T mismatches, whereas HP-2/TFO-1 has
G-PPIMe·C and T-A CBr mismatches. As a result, it was found that
the matched triplexes HP-1/TFO-1 and HP-2/TFO-2 showed
very high Tm values of 59 ◦C and 60 ◦C, respectively, while the
two-base-mismatched triplexes HP-1/TFO-2 an◦d HP-2/TFO-1
◦
showed much lower Tm values of 12 C and 28 C, respectively.
These results clearly suggest that TFO-1 bound more strongly to
the matched partner HP-1 than to HP-2 and, likewise, TFO-2
substantially favored HP-2 over HP-1, thereby discriminating be-
tween fully matched and two-base-mismatched hairpin duplexes.
To further evaluate the selective formation of HP-1/TFO-1
and HP-2/TFO-2, the gel shift assay experiment was performed
with these hairpin oligomers and TFOs. The results are shown
in Fig. 7. Sequence-matched triplexes HP-1/TFO-1 (lane 2) and
HP-2/TFO-2 (lane 6) appeared as retarded bands. In contrast,
a sequence-unmatched mixture of HP-1 and TFO-2 (lane 3) or
HP-2 and TFO-1 (lane 5) did not show such a retarded band.
These results clearly suggested that base triads such as C-G·C+,
T-A·T, G-PPIMe·C3, and A-y·CBr could selectively form base pairs
as determined not only by the UV-melting conditions but also by
the gel mobility shift assays.
Selection of four base triads capable of selective recognition
On the basis of their Tm properties, we selected C-G·C+, T-A·T,
and A-y·CBr (Box-2 in Table 1) and G-PPIMe·C3 (Box-3 in Table 1)
as the set of selective base triads for further studies. The Tm data
are extracted from Tables 1 and listed again in Table 2.
As shown in Table 2, the Tm of the matched triplexes
HP-cg/TFO-C, HP-ta/TFO-T, HP-gppiMe/TFO-C3, and HP-
aw/TFO-Br shown in the diagonal cells are larger than those
for the mismatched TFOs shown in the non-diagonal cells. Thus,
it is expected that DNA duplexes with mixed sequences comprised
of C-G, T-A, G-PPIMe, and A-y can be selectively recognized
by TFOs incorporating C, T, C3, and CBr, respectively, in the
appropriate sequences. To test this expectation, we synthesized
the hairpin oligodeoxynucleotides HP-1 and HP-2 incorporating
all four base pairs (T-A, C-G, A-y, and G-PPIMe) and the
corresponding TFO-1 and TFO-2 incorporating T, C, CBr, and
C3, which recognize HP-1 and HP-2, respectively (Fig. 6).
First, we measured the Tm values for the fully matched triplexes
HP-1/TFO-1 and HP-2/TFO-2 and the mismatched triplexes
Fig. 7 Native gel shift assay using HP-1, HP-2, TFO-1, and TFO-2.
Table 2 Tm summary of the set of four base triads
TFO-C Z = C -T = T -C3 = C3 -Br = CBr
Conclusions
HP-cg (X-Y = C-G)
HP-ta (X-Y = T-A)
52
18
21
48
45
36
12
14
53
21
45
21
45
48
In this paper, we reported the development of a set of four
base triads that can be incorporated into DNA duplexes and
TFOs to expand the potential for sequence-selective DNA triplex
HP-gppiMe (X-Y = G-PPIMe) 45
HP-aw (X-Y = A-y)
42
This journal is
The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 1007–1013 | 1011
©

View Article Online
formation. We designed base triads G-PPIMe·C3 and A-y·CBr
having unnatural pyrimidine bases such as PPIMe and y, which
formed Watson–Crick type base pairs with G and A in the duplexes
and were selectively bound to C3 and CBr in TFOs, respectively.
As shown in Table 2, the triplexes incorporating one of
the matched base triads, such as G-PPIMe·C3, A-y·CBr, T-A·T,
or C-G·C+, showed higher Tm than any of their mismatched
counterparts. Thus, DNA duplexes having mixed base pairs such
as T-A, C-G, G-PPIMe, and A-y can be combined with TFOs
that recognize them in a sequence-specific manner to generate the
corresponding DNA triplexes.
ammonia at room temperature. The oligonucleotides were purified
on a C18-reverse phase cartridge and the DMTr group was
removed by treatment with 2% CF3COOH/H2O. In the synthesis
of HP-aw, HP-1, and HP-2 containing deoxypseudouridines,
the deprotection conditions were slightly modified as follows to
prevent cyanoethylation at the N1 position of pseudouridine.29
After the chain elongation, the resin was flushed with a 10%
solution of DBU in MeCN to remove the cyanoethyl groups
on the phosphodiester residues and washed with dry acetonitrile.
After the DBU treatment, the oligonucleotides were cleaved and
deprotected by using 28% aqueous ammonia for 12 h. The
oligonucleotides were purified by reverse-phase HPLC using a
gradient of acetonitrile in 30 mM ammonium acetate. If necessary,
anion-exchange HPLC was used. The purified oligonucleotides
were analyzed by MALDI-TOF-MS. The MS data are as follows.
HP-aw [M + H]+ Calcd 11018.86, found 11020.25; HP-gppi [M +
H]+ Calcd 11174.89, found 11174.62; HP-gppiMe [M + H]+ Calcd
11188.91, found 11190.79; HP-gppiOMe [M + H]+ Calcd 11204.90,
found 11207.00; HP-1 [M + H]+ Calcd 11304.88, found 11306.31;
HP-2 [M + H]+ Calcd 11473.94, found 11475.12; TFO-Br [M
+ H]+ Calcd 4228.61, found 4228.67; TFO-Cl [M + H]+ Calcd
4184.66, found 4183.79; TFO-F [M + H]+ Calcd 4168.69, found
4168.51; TFO-CN [M + H]+ Calcd 4175.69, found 4176.16; TFO-
C3 [M + H]+ Calcd 3999.66, found 3998.10; TFO-1 [M + H]+
Calcd 3992.35, found 3995.74; TFO-2 [M + H]+ Calcd 3778.40,
found 3780.14.
In the current set of four base triads, the most stable matched
triad is G-PPIMe·C3, which gave a Tm of 53 ◦C, while the least
stable matched triads are T-A·T and A-y·CBr both of which gave
a Tm of 48 ◦C when incorporated in the nucleotide sequences
used in this study. In addition, as shown in Table 2, the Tm
values of the least stable matched triads T-A·T and A-y·CBr are
◦
only 3 C higher than those of the most stable mismatched base
triads C-G·CBr, G-PPIMe·C, and G-PPIMe·CBr, all of which have
◦
a Tm of 45 C. Therefore, if very stringent single-base mismatch
discrimination is to be achieved, the design and synthesis of new
base analogs capable of further improved base discrimination
might be necessary. On the other hand, these four triads formed
less stable triplex under the neutral condition (pH 7.0) due to
the unprotonated cytosine in the TFO (see the Supplementary
Information†). Analogues of protonated cytosine such as MeP9
might be useful for further application of our artificial triplexes
under the biological conditions.
In contrast, the Tm analyses performed using HP-1/TFO-
1, HP-2/TFO-2, HP-2/TFO-1, and HP-1/TFO-2 and the gel
electrophoretic studies indicated the ability of the current base
triads to discriminate between fully matched and two-base-
mismatched sequences. For example, although the triplex HP-
2/TFO-1 contains a rather stable G-PPIMe·C mismatch, the forma-
tion of this triplex is substantially inhibited by the incorporation
of an unstable T-A·CBr mismatch as demonstrated by the gel
electrophoresis.
Thermal denaturation experiments
UV melting experiments were performed with a UV spectrometer
equipped with a temperature controller. All measurements were
conducted in a buffer containing 10 mM MgCl2, 500 mM NaCl,
10 mM sodium cacodylate at pH 5.4. Oligonucleotides were mixed
in a 1 : 1 ratio at a concentration of 2.0 mM. Melting curves were
recorded at 260 nm in a consecutive heating–cooling–heating cycle
(0–90 ◦C) with a temperature gradient of 0.5 ◦C min-1.
At present, the potential application of DNA triplexes in gene
regulation, gene correction, gene detection, and DNA nanotech-
nology are attracting significant interest. However, because the
formation of unnatural base triads such as G-PPIMe·C3 and A-
y·CBr required the incorporation of the unnatural bases PPIMe and
y into the duplexes, these triads cannot be directly used in gene-
targeting technologies such as gene regulation, gene correction,
and gene detection. In contrast, our base triads can be widely used
in technologies utilizing chemically synthesized DNA duplexes
such as DNA nanotechnology and the development of artificial
genes. These applications of the four base triads described herein
will be reported elsewhere.
Gel shift assay
10 pmol of hairpin oligonucleotides (HP-1, HP-2) and 20 pmol
of TFOs (TFO-1, TFO-2) were dissolved in a 20 mL cacodylate
buffer, which is the same as that used in the Tm experiments. The
solution was heated at 95 ◦C for 5 min, and then gradually cooled
to ambient temperature. We used the protocol of the gel shift assay
of triplex analysis previously described by Johnston et al.30 All
samples were analyzed by electrophoresis at 200 V on 20% native
acrylamide gels in 10 mM Mg(OAc)2, 40 mM NaOAc buffer (pH
5.1) at ambient temperature (26 ◦C) for 2.5 h and the same buffer
was used as the running buffer. After the electrophoresis, DNA
R
bands were stained with SYBRꢀ Gold.
Experimental
Acknowledgements
Synthesis of oligodeoxynucleotides containing modified nucleosides
All oligodeoxynucleotides containing modified nucleosides
were synthesized by using a DMTr-ON protocol automated
DNA/RNA synthesizer on a 1 mmol scale using the standard
cycle. Removal of the protecting groups and the cleavage of the
linker from resins were performed for 12–24 h in 28% aqueous
This study was supported by a Grants in aid for Scientific Research
from the Ministry of Education, Culture, Sports, Science and
Technology, Japan. We also thank Center for Advanced Materials
Analysis (Suzukakedai), Technical Department, Tokyo Institute
of Technology, for MALDI-TOF mass analysis.
1012 | Org. Biomol. Chem., 2012, 10, 1007–1013
This journal is
The Royal Society of Chemistry 2012
©

View Article Online
9 S. Hildbrand, A. Blaser, S. P. Parel and C. J. Leumann, J. Am. Chem.
Soc., 1997, 119, 5499.
10 D. Guianvarc’h, R. Benhida, J. L. Fourrey, R. Maurisseb and J. S. Sun,
Chem. Commun., 2001, 1814.
11 T. L. Trapane and P. O. P. Ts’o, J. Am. Chem. Soc., 1994, 116, 10437.
12 T. L. Trapane, M. S. Christopherson, C. D. Roby, P. O. P. Ts’o and D.
Wang, J. Am. Chem. Soc., 1994, 116, 8412.
13 R. Bandaru, H. Hashimoto and C. Switzer, J. Org. Chem., 1995, 60,
786.
Notes and references
1 M. Duca, P. Vekhoff, K. Oussedik, L. Halby and P. B. Arimondo,
Nucleic Acids Res., 2008, 36, 5123; D. P. Arya, Acc. Chem. Res., 2011,
44, 134; V. Malnuit, M. Duca and R. Benhida, Org. Biomol. Chem.,
2011, 9, 326.
2 C. He´le`ne, Anticancer Drug Des., 1991, 6, 569; D. Praseuth, A. L.
Guieysse and C. He´le`ne, Biochim. Biophys. Acta., 1999, 1489, 181; S.
Obika, Chem. Pharm. Bull., 2004, 52, 1399; R. V. Guntaka, B. R. Varma
and K. T. Weber, Int. J. Biochem. Cell Biol., 2003, 35, 22; K. R. Fox,
Curr. Med. Chem., 2000, 7, 17; S. Buchini and C. J. Leumann, Curr.
Opin. Chem. Biol., 2003, 7, 717; B. P. Casey and P. M. Glazer, Prog.
Nucleic Acid Res. Mol. Biol., 2001, 67, 163.
3 J. M. Kalish and P. M. Glazer, Ann. N. Y. Acad. Sci., 2005, 1058, 151; Y.
Liu, R. S. Nairn and K. M. Vasquez, Nucleic Acids Res., 2009, 37, 6378;
D. de Semir and J. M. Aran, Curr. Gene Ther., 2006, 6, 481; C.-M. Liu,
D.-P. Liu and C.-C. Liang, J. Mol. Med., 2002, 80, 620; O. Igoucheva,
V. Alexeev and K. Yoon, Curr. Mol. Med., 2004, 4, 445; J. Y. Kuan and
P. M. Glazer, Methods Mol. Biol., 2004, 262, 173; P. P. Chan and P. M.
Glazer, J. Mol. Med., 1997, 75, 267.
14 V. S. Rana, D. Barawkar and K. N. Ganesh, J. Org. Chem., 1996, 61,
3578.
15 V. S. Rana and K. N. Ganesh, Nucleic Acids Res., 2000, 28, 1162; V. S.
Rana and K. N. Ganesh, Org. Lett., 1999, 1, 631.
16 M. Mizuta, J. Banba, T. Kanamori, R. Tawarada, A. Ohkubo, M.
Sekine and K. Seio, J. Am. Chem. Soc., 2008, 130, 9622.
17 D. M. Frisch and D. W. Visser, J. Am. Chem. Soc., 1959, 81, 1756.
18 V. Kumar, J. Yap, A. Muroyama and S. V. Malhotra, Synthesis, 2009,
3957.
19 K. Kikugawa, I. Kawada and M. Ichino, Chem. Pharm. Bull., 1975, 23,
35.
4 V. N. Potaman, Expert Rev. Mol. Diagn., 2003, 3, 481; F. McKenzie, K.
Faulds and D. Graham, Chem. Commun., 2008, 2367.
20 J. Nielse, M. Taagaard, J. E. Marugg, J. H. van Boom and O. Dahl,
Nucleic Acids Res., 1986, 14, 7391.
21 G. Lahoud, V. Timoshchuk, A. Lebedev, M. de Vega, M Salas, K.
Arar, Y.-M. Hou1 and H. Gamper, Nucleic Acids Res., 2008, 36,
3409.
22 Y. E. Safadi, J.-C. Paillart, G. Laumond, A.-M. Aubertin, A. Burger,
R. Marquet and V. Vivet-Boudou, J. Med. Chem., 2010, 53, 1534.
23 C. J. Marasco Jr. and J. R. Sufrin, J. Org. Chem., 1992, 57, 6363.
24 A. K. Ghosh and G. B. Schuster, J. Am. Chem. Soc., 2006, 128,
4172.
5 Y. Chen, S. H. Lee and C. Mao, Angew. Chem., Int. Ed., 2004, 43,
5335; J. Tumpane, R. Kumar, E. P. Lundberg, P. Sandin, N. Gale, I. S.
Nandhakumar, B. Albinsson, P. Lincoln, L. M. Wilhelmsson, T. Brown
and B. Norde´n, Nano Lett., 2007, 7, 3832; Y. H. Jung, K. B. Lee, Y. G.
Kim and I. S. Choi, Angew. Chem., Int. Ed., 2006, 45, 5960.
6 D. M. Gowers and K. R. Fox, Nucleic Acids Res., 1999, 27, 1569; D.
A. Rusling, V. E. Powers, R. T. Ranasinghe, Y. Wang, S. D. Osborne, T.
Brown and K. R. Fox, Nucleic Acids Res., 2005, 33, 3025.
7 D. Chen Meena, S. K. Sharma and L. W. McLaughlin, J. Am. Chem.
Soc., 2004, 126, 70; H. Chen Meena and L. W. McLaughlin, J. Am.
Chem. Soc., 2008, 130, 13190.
8 J.-S. Li, Y.-H. Fan, Y. Zhang, L. A. Marky and B. Gold, J. Am. Chem.
Soc., 2003, 125, 2084; J.-S. Li and B. Gold, J. Org. Chem., 2005, 70, 8764;
J.-S. Li, F.-X. Chen, R. Shikiya, L. A. Marky and B. Gold, J. Am. Chem.
Soc., 2005, 127, 12657; D. Guianvarc’h, J.-L. Fourrey, R. Maurisse, J.-S.
Sun and R. Benhida, Org. Lett., 2002, 4, 4209; D. Guianvarc’h, J.-L.
Fourrey, R. Maurisse, J.-S. Sun and R. Benhida, Bioorg. Med. Chem.,
2003, 11, 2751.
25 G. B. Wisdom and B. A. Orsi, Eur. J. Biochem., 1969, 7, 223.
26 C. J. La Francois, Y. H. Jang and T. Cagin, Chem. Res. Toxicol., 2000,
13, 462.
27 M. Mizuta, K. Seio, A. Ohkubo and M. Sekine, J. Phys. Chem. B, 2009,
113, 9562.
28 M. Mizuta, K. Seio, K. Miyata and M. Sekine, J. Org. Chem., 2007, 72,
5046.
29 J. Ofengand, Biochem. Biophys. Res. Commun., 1965, 18, 192.
30 B. P. Belotserkovskii and B. H. Johnston, Electrophoresis, 1996, 17,
1528.
This journal is
The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 1007–1013 | 1013
©