Assessment of the sequence dependency for the binding of 2-aminonaphthyridine to the guanine bulge

Article abstract of DOI:10.1016/S0968-0896(03)00026-9

We have studied the sequence dependent binding of 2-amino-1,8-naphthyridine derivative 1 to a single guanine bulge. The free energy changes for the binding to a guanine bulge with different sequence contexts (5′X_Y3′/3′X′GY′5′) were determined by a curve

Full text of DOI:10.1016/S0968-0896(03)00026-9

Bioorganic & Medicinal Chemistry 11 (2003) 2347–2353
Assessment of the Sequence Dependency for the Binding of
2-Aminonaphthyridine to the Guanine Bulge
Kazuhiko Nakatani,a,b,* Souta Horie,a Takashi Murase,a Shinya Hagiharaa
and Isao Saitoa,*
aDepartment of Synthetic Chemistry and Biological Chemistry, Faculty of Engineering, Kyoto University, Kyoto 606-8501, Japan
bPRESTO, Japan Science and Technology Corporation (JST), Kyoto 606-8501, Japan
Received 31 October 2002; revised 5 December 2002; accepted 21 December 2002
Abstract—We have studied the sequence dependent binding of 2-amino-1,8-naphthyridine derivative 1 to a single guanine bulge.
The free energy changes for the binding to a guanine bulge with different sequence contexts (50X_Y30/30X0GY050) were determined
by a curve fitting of the thermal denaturation profile of DNA in the presence and absence of 1. The data showed that (i) the binding
of 1 to a guanine bulge is stronger for those flanking the G–C base pair than A–T base pair, (ii) the guanine 30 side to 1 in the
complex is especially effective for the complex stabilization, and (iii) the increase of Tm in the presence of 1 is not a good estimate
for the sequence dependent binding. The most efficient 1-binding was observed for the sequence of G_G/CGC. Molecular modeling
simulations suggested that stacking interaction between the 30 side guanine and 1 is the molecular basis for the strong binding to
G_G/CGC.
# 2003 Elsevier Science Ltd. All rights reserved.
Introduction
We have recently demonstrated that DNA intercalators
possessing hydrogen-bonding groups fully com-
plementary to those of nucleotide bases are a unique
class of compounds that bind to a bulge structure and,
hence, stabilize the duplex with high base
selectivity.12,16,18À20 2-Amino-1,8-naphthyridine deriva-
tive 1 having an alignment of hydrogen bonding groups
in a donor–donor–acceptor orientation selectively binds
to a guanine bulge in duplex DNA (Fig. 1a). In contrast
to ethidium bromide,6 1 strongly stabilized a guanine
bulge duplex, but only weakly did other bulges and a
fully matched duplex.12 We have evaluated the drug
binding by thermal denaturation of a guanine bulge-
containing duplex. Thus, the melting temperature (Tm)
of the duplexes increased with increasing the degree of 1
binding. Structure–activity relationship (SAR) studies
for 1 binding to a guanine bulge showed that (1) an
alignment of hydrogen bonding groups fully com-
plementary to that of a guanine is essential for a strong
binding and (2) at least two condensed aromatic rings
are necessary for the efficient binding. These observa-
tions suggest that stacking of the hydrogen-bonded pair
produced from 1 and a bulged guanine with the flanking
base pairs significantly contributes in energy to stabilize
the complex. We here report the overall free energy
change for the 1-binding to the guanine bulge with
A bulge is the site where one or more extra nucleotide
bases remain unpaired in the duplex of DNA and
RNA.1À3 The bulged site in DNA is not only a crucial
structural element for the recognition by DNA repair
proteins,4,5 but also exhibits a remarkable reactivity
toward the drug interactions. Ethidium bromide6 and
other intercalators,7À10 intercalate into the site of a
nucleotide bulge more strongly than does into the stan-
dard Watson–Crick base pairs. Thermodynamic stabili-
zation of a nucleotide bulge is thought to be a molecular
basis of the mutagenic properties of many intercalating
agents.11 One-electron oxidation of duplex DNA by
external oxidants proceeds preferentially at a guanine
bulge compared with the normal guanine in a G–C base
pair.12À14 Covalent modification of guanine by a DNA
alkylating intercalator occurred selectively near the
bulge possibly due to a preferential drug binding to a
bulged site.15 Besides these biological aspects, the site of
a nucleotide bulge provides an important clue for the
mutation detection with small molecular probes.16,17
*Corresponding author. Tel.: +81-75-753-4813; fax: +81-75-753-
5676; e-mail: nakatani@sbchem.kyoto-u.ac.jp
0968-0896/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00026-9

2348
K. Nakatani et al. / Bioorg. Med. Chem. 11 (2003) 2347–2353
different flanking sequences. The data showed that the
binding of 1 to a guanine bulge was the most efficient
for the sequence of G_G/CGC, where the bound 1 was
stacked by both 50 and 30 side guanine in the complex.
Furthermore, the 30 side guanine to the intercalator in
the complex is especially effective for the stabilization of
the complex.
the base pair flanking the guanine bulge.21 Oligomers
having G–C and/or C–G base pairs at the both 50 and 30
side of a guanine bulge showed a pronounced stability
compared with those having at least one A-T base pair
flanking to a guanine bulge. The lowest Tm of 17.6 ꢀC
under the conditions was observed for oligomer A_A/
TGT, where the bulged guanine was located between
two thymine bases.
In the presence of 1, the denaturation curves of the
guanine bulge-containing duplexes were shifted to a
higher temperature range with keeping the sigmoidal
shape unchanged. Increase of Tm (ꢀTm) was 11.8 and
11.2 ꢀC for G_G/CGC and A_G/TGC, respectively,
whereas ꢀTm obtained for other oligomers was about in
the same range (6.7–8.3 ꢀC). Thus, a sequence depen-
dence of the stabilization of a guanine bulge by 1 could
not be clearly deduced form the analysis of ꢀTm.
Results and Discussion
Sequence Dependence for the stabilization of a guanine
bulge by 1
The sequence dependence for the stabilization of a gua-
nine bulge by 1 binding was examined for the duplexes
with ten different flanking sequences (Fig. 1b). These
duplexes are shown by a general formula of 50X_Y30/
30X0GY050 (Table 1). For the duplex containing cytosine
in the position of X and/or Y (e.g., C_G/GGC, C_C/
GGG, and G_C/CGG), the site of a guanine bulge can
not be defined due to an unavoidable migration of the
bulge immediately next to the designated position (e.g.,
C_G/GGC$_CG/GGC). Therefore, the data for these
sequences are average of all possible bulge structures. In
the study reported here, naphthyridine derivative 1
possessing two-methylene linker was used. This deriva-
tive showed an improved thermal stability and stronger
binding compared to the three methylene derivative
reported previously.12 All of the thermal denaturation
curves of guanine bulge-containing duplexes showed a
sigmoidal shape (Fig. 2). Melting temperatures of these
duplexes were determined as the first derivative of the
curve (Table 2). As expected, thermal stability of the
guanine bulge-containing duplexes strongly depends on
Estimation of the free energy change
Since the binding of naphthyridine to the guanine bulge
was intrinsically weak (ꢁ104 MÀ1), isothermal titration
calorimetry and other techniques determining an accu-
rate free energy change for the binding are not applic-
able to the 1 binding to the G bulge. We have estimated
the free energy change for the binding of the naphthyr-
idine to the G bulge by subtracting the free energy
change for the formation of bulge-containing duplex
from the overall free energy change obtained in the
presence of the drug and two single stranded DNAs.
Free energy changes for each process were obtained by
nonlinear least-square method to fit the thermal dena-
turation curves to a two-state model with ꢀH, ꢀS, and
extinction coefficients of single and double stranded
Table 1. Oligomers used in the studies
A_A/TGT:
A_G/TGC:
C_C/GGG:
C_G/GGC:
G_A/CGT:
G_C/CGG:
G_G/CGC:
G_T/CGA:
T_G/AGC:
T_T/AGA:
50-TCCA A_A CAAC-30
30-AGGT TGT GTTG-50
50-TCCA A_G CAAC-30
30-AGGT TGC GTTG-50
50-TCCA C_C GAAC-30
30-AGGT GGG CTTG-50
50-TCCA C_G CAAC-30
30-AGGT GGC GTTG-50
50-TCCA G_A CAAC-30
30-AGGT CGT GTTG-50
50-TCCA G_C GAAC-30
30-AGGT CGG CTTG-50
50-TCCA G_G CAAC-30
30-AGGT CGC GTTG-50
50-TCCA G_T CAAC-30
30-AGGT CGA GTTG-50
50-TCCA T_G CAAC-30
30-AGGT AGC GTTG-50
50-TCCA T_T CAAC-30
30-AGGT AGA GTTG-50
Figure 1. (a) An illustration of a hydrogen bonding pair of 1 with
guanine. (b) A schematic representation of the binding of 1 to a gua-
nine-bulge containing duplex in a sequence of 50X_Y30/30X0GY050.

K. Nakatani et al. / Bioorg. Med. Chem. 11 (2003) 2347–2353
2349
DNAs as variable parameters.22,23 Since the binding of
naphthyridine to the G bulge is concentration dependent,
the above estimated free energy is a function of the frac-
tion of the drug–DNA complex in the total DNA and
drug concentrations. To obtain the actual free energy
change for the binding of naphthyridine to the G bulge, the
fraction of the complex under the given drug concen-
tration should be determined.
Figure 2. Thermal denaturation profiles of guanine bulge-containing duplexes (4.77 mM) in the absence (open square) and presence (open circle) of 1
ꢀ
ꢀ
(200 mM). The changes of the absorbance (ꢀABS) defined as Abs(T C)ÀAbs(4 C), were plotted against temperature. The absorbance at 260 nm was
measured in 10 mM sodium cacodylate buffer (pH 7.0) containing 0.1 M NaCl. Temperature was increased from 4 to 70 ꢀC at a rate of 1 ꢀC/min.
The absorbance was measured with an interval of 1 ꢀC. Measurements were carried out at least three times. The average of three data sets of
denaturation profiles was used for the plots. The calculated curves obtained for each measured profile were shown with a solid red line. Data points
are shown in every 2 ꢀC between 4 to 60 ꢀC for clarity.

2350
K. Nakatani et al. / Bioorg. Med. Chem. 11 (2003) 2347–2353
equilibriums could be described with the free energy
changes of ꢀGbulge, ꢀGspecific, ꢀGduplex+1, ꢀGsingle+1
and ꢀGduplex for each equilibration of eqs 1–5,
Table 2. Melting temperature of guanine bulge-containing duplexes
in the absence and presence of 1
,
Duplex
Tm (À)a,b
Tm (+)a,b
ꢀTm
c
respectively.
A_A/TGT
A_G/TGC
C_C/GGG
C_G/GGC
G_A/CGT
G_C/CGG
G_G/CGC
G_T/CGA
T_G/AGC
T_T/AGA
17.6Æ0.1
25.6Æ0.236.8
35.3Æ0.1
25.3Æ0.27.7
Æ0.0
11.2
ssDNA1 þ ssDNA2 ! Gb-DNA þ ꢀGbulge
Gb-DNA þ 1 ! Gb-DNAÁ1 þ ꢀGspecific
dsDNA þ 1 ! dsDNAÁ1 þ ꢀGduplexþ1
ssDNA þ 1 ! ssDNAÁ1 þ ꢀGsingleþ1
ssDNA1 þ ssDNA3 ! dsDNA þ ꢀGduplex
ð1Þ
ð2Þ
ð3Þ
ð4Þ
ð5Þ
42.1Æ0.26.8
43.2Æ0.1
31.7Æ0.1
41.1Æ0.4
41.9Æ0.1
Æ0.1
36.5Æ0.1
6.7
7.8
7.1
11.8
6.7
23.9Æ0.0
34.0Æ0.4
30.1Æ0.3
24.4Æ0.230.1
28.9Æ0.1
37.2Æ0.28.3
Æ0.1
61.38.3Æ0.22
8.0
aTm(À) and Tm(+) are melting temperatures (ꢀC) of a duplex (4.77
mM) in the absence and presence of 1 (200 mM), respectively, in 10 mM
sodium cacodylate buffer (pH 7.0) containing 0.1 M NaCl. Tempera-
ture was increased at a rate of 1 ꢀC/min.
bMeasurement for each duplex was carried out at least three times.
The average of three denaturation profiles was analyzed by a curve
fitting.
cꢀTm=Tm(+)ÀTm (À).
The overall free energy change (ꢀGbulge-overall) for the
.
formation of Gb-DNA 1 from two ssDNAs and 1 is
described with a combination of ꢀGbulge, ꢀGspecific
ꢀGduplex+1, and ꢀGsingle+1 by eq 6,
,
ꢀGbulge-overall ¼ꢀGbulge þ ꢀÁꢀGspecific
ð6Þ
ð7Þ
þ ꢁÁꢀGduplexþ1 À ꢂÁꢀGsingleþ1
À1
ꢀ ¼ ½Gb-DNAÁ1ŠÁ½Gb-DNA0Š
where a indicates a molar fraction of the complex Gb-
.
DNA 1 to the initial concentration of Gb-DNA pro-
duced from two ssDNAs ([Gb-DNA0]), whereas b and g
.
are molar fractions of the complexes dsDNA 1 and
.
ssDNA 1 to the concentrations of Gb-DNA and
ssDNA, respectively. This equation shows that the
overall energy change reduces, as the nonspecific bind-
ing of 1 to ssDNA becomes stronger. The overall free
energy gain by the binding of 1 to dsDNA in the pre-
sence of 1 (ꢀGduplex-overall) could be described with the
free energy changes regarding the nonspecific bindings
of 1 to ssDNA and dsDNA by eq 8.
Figure 3. Illustrations of the binding of 1 to a single stranded DNA
(ssDNA), a fully matched double stranded DNA (dsDNA), and a
guanine bulge-containing duplex (Gb-DNA). Hybridizations of
ssDNA1 with ssDNA2and ssDNA1 with ssDNA3 produces Gb-
.
DNA and dsDNA, respectively. ssDNA 1 represents the complex of 1
with ssDNA1, ssDNA2, and ssDNA3.
ꢀGduplex-overall ¼ ꢀGduplex þ ꢁÁꢀGduplexþ1
ð8Þ
À ꢂÁꢀGsingleþ1
We considered five equilibriums for the thermodynamic
analysis of 1 binding to a guanine bulge (Fig. 3). These
include (i) a formation of DNA containing a guanine
bulge (Gb-DNA) from two single stranded DNAs
(ssDNA), (ii) a specific binding of 1 to the guanine bulge
With a combination of eqs 6 and 8, the free energy
change for the specific binding of 1 to a guanine bulge
(ꢀGspecific) could be described with the experimentally
obtainable energy terms by eq 9,
.
producing a complex (Gb-DNA 1), (iii) a formation of a
complex (dsDNA+1) by a nonspecific binding of 1 to a
ꢀÁꢀGspecific ¼ ꢀGbulge-overall À ꢀGbulge
double stranded DNA (dsDNA), (iv) a nonspecific
binding of
À ðꢀGduplex-overall À ꢀGduplex
Þ
ð9Þ
1 to ssDNA producing a complex
(ssDNA+1), and (v) a formation of dsDNA. We
assumed that a nonspecific binding of 1 to a duplex
region of Gb-DNA is indistinguishable in the energy
from that to a fully complementary duplex. These five
Since the nonspecific binding of 1 to a double stranded
region is much weaker than the specific biding of 1 to a

K. Nakatani et al. / Bioorg. Med. Chem. 11 (2003) 2347–2353
2351
guanine bulge, we assumed that the initial concentration
of bulge-containing duplex [Gb-DNA0] is equal to a
sum of concentrations of Gb-DNA in a free state and a
respectively. With the values of the free energy changes
thus obtained, a and ꢀGspecific for the specific binding
of 1 to a guanine bulge at the temperature of T could be
obtained by solving the eq 13.
.
complex Gb-DNA 1. It is important to note that under
these estimations the deviation of the calculated data
from the actual value will be larger as weaker the spe-
cific binding. The concentrations of Gb-DNA in a free
and a complex state are described with a and [Gb-
DNA0] by eqs 10 and 11.
Sequence dependence for the binding of 1 to the Gua-
nine bulge. Curve fittings of the thermal denaturation
profiles shown in Figure 2 to a two-state model provides
ꢀGbulge and ꢀGbulge-overall for the formation of the
guanine bulgeÀ1 complex with ten flanking sequences.
ꢀGduplex and ꢀGduplex-overall were obtained from the
similar analysis of a fully matched duplex of d(TCC
AGG CAA C)/d(GTT GCC TGG A) and were À16.2and
À16.7 kcal/mol at 277.15 K, respectively. Accordingly,
the contribution of the non-specific binding of 1 to
ssDNA and dsDNA in the overall free energy change is
determined to be À0.5 kcal/mol. Under the experi-
mental conditions, we estimated that the initial concen-
tration of guanine bulge-containing duplex ([Gb-
DNA0]) is equal to the concentration of a single stran-
ded DNA (4.77Â10À6 M) and the initial drug concen-
tration ([10]) is 2.00Â10À4 M. With these data in hand,
ꢀGspecific and Kspecific at 277.15 K were calculated
(Table 3). The flanking sequences we examined could be
divided into three groups regarding the magnitude of
the free energy gain for the 1 binding. As we expected,
the strongest binding of 1 was observed for the sequence
of G_G/CGC with the free energy change (ꢀGspecific) of
À5.3Æ0.3 kcal/mol among other guanine bulge con-
taining sequences. The apparent equilibrium association
constant (Kspecific) for 1 binding to G_G/CGC was
15.2Æ4.0Â103 MÀ1, that is in a good agreement with
3.4Â104 MÀ1 previously obtained by DNase I footprint
titration within an experimental variability.12 The
marked difference of ꢀGspecific between G_G/CGC and
A_A/TGT by 0.8 kcal/mol showed that the strong
binding of 1 to the sequence is due to the stacking of 1
by two flanking guanine bases, but is not due to the
stacking by two purines. The second group of the
sequences is those containing G base 30 side to 1 in the
complex. The guanine base immediately next to the 30
side of 1 in the complex (Y¼G) has strong effects on the
stabilization of the complex. Substitution of the 30 side
G in G_G/CGC to A in G_A/CGT, C in G_C/CGG,
and T in G_T/CGA resulted in a loss of the free energy
change for the complex formation by ca. 0.7 kcal/mol.
A strong stabilization of 1 by the 30 side guanine is in a
½Gb-DNAÁ1Š ¼ ꢀÁ½Gb-DNA0Š
ð10Þ
ð11Þ
½Gb-DNA0Š ¼ ½Gb-DNAŠ þ ½Gb-DNAÁ1Š
½Gb-DNAŠ ¼ ð1 À ꢀÞÁ½Gb-DNA0Š
An association constant for the binding of 1 to a
guanine bulge (Kspecific) could be described by eq 12,
À1
À1
Kspecific ¼ ½Gb-DNAÁ1ŠÁ½Gb-DNAŠ Á½1Š
¼ ꢀÁð1 À ꢀÞÀ1 Áð½10Š À ꢀÁ½Gb-DNA0ŠÞÀ1
ð12Þ
where [10] is an initial concentration of 1. Thus, a free
energy change for the 1 binding to a guanine bulge is
described as a function of a with experimentally
obtainable energy terms by eq 13.
ꢀGspecific ¼fꢀGbulge-overall À ꢀGbulge
À ðꢀGduplex-overall À ꢀGduplexÞgÁꢀÀ1
¼ ÀRTÁlnfꢀÁð1 À ꢀÞÀ1 Áð½10Š
ð13Þ
À ꢀÁ½Gb-DNA0ŠÞÀ1
g
ꢀGbulge and ꢀGbulge-overall were obtained by a curve-
fitting analysis of a guanine bulge-containing duplex in
the absence and presence of 1, respectively, at the initial
concentration of [10]. ꢀGduplex and ꢀGduplex-overall were
similarly obtained by a curve-fitting analysis of a fully
matched duplex in the absence and presence of 1,
Table 3. Free energy changes and association constants for the binding of 1a
50X_Y30
30X0GY050
ꢀTm (ꢀC)
ꢀGbulge
ꢀGbulge-overall
a
ꢀGspecific
Kspecific (10 M
3
À1
)
.
G_G
A_G
C_G
C_C
T_G
G_C
G_A
T_T
CGC
TGC
GGC
GGG
AGC
CGG
CGT
AGA
TGT
CGA
11.8
11.2
6.7
6.8
8.3
7.1
7.8
8.0
7.7
6.7
À12.0Æ0.2
À11.5Æ0.2
À13.7Æ0.1
À12.6Æ0.1
À12.0Æ0.1
À12.9Æ0.1
À10.5Æ0.2
À9.5Æ0.2
À9.2Æ0.1
À11.1Æ0.2
À16.5Æ0.3
À14.4Æ0.1
À16.6Æ0.1
À15.4Æ0.20.49
À14.7Æ0.4
À15.6Æ0.3
À13.2Æ0.5
À11.9Æ0.20.41
À11.4Æ0.3
À13.3Æ0.1
0.75Æ0.04
0.53Æ0.02
0.51Æ0.02
Æ0.02
À5.3Æ0.3
15.2Æ4.0
Æ0.8
À4.8Æ0.25.7
À4.7Æ0.1
5.3Æ0.4
Æ0.7
À4.7Æ0.24.9
À4.7Æ0.4
0.49Æ0.06
0.48Æ0.05
0.46Æ0.08
Æ0.03
4.9Æ1.6
4.7Æ1.0
4.3Æ1.6
Æ0.8
À4.6Æ0.3
À4.6Æ0.5
À4.5Æ0.23.5
À4.5Æ0.3
A_A
G_T
0.40Æ0.06
0.39Æ0.04
3.4Æ0.9
Æ0.5
À4.4Æ0.23.2
a
À1
À1
.
.
All free energies were reported in kcal mol at 277.15 K. ꢀGduplex-overallÀꢀGduplex=À0.5 kcal mol
.

2352
K. Nakatani et al. / Bioorg. Med. Chem. 11 (2003) 2347–2353
good agreement with our previous studies of the sequence
selective intercalation o0f charg0e neutral intercalators into
the sequence of 50GG3 and 5 CG30.24 The last group of
the sequence contains the A–T base pair in the 30 side of
1 in the complex. The binding of 1 to these sequences
was especially weak. On the basis of these data, the
strong binding of 1 to G_G/CGC is most likely due to a
cooperative stabilization by both 30 and 50 side Gs.
Substitution of one G in the sequence by other bases
dramatically decreased the complex stability.
crude product was suspended in CHCl3. The organic
layer was washed with H2O and dried over MgSO4. The
solvent was evaporated in vacuo and the crude residue
was purified by silica gel column chromatography
(CHCl3/hexane=1/1) to give the title compound (3.8 g,
90%) as white solids: 1H NMR (CDCl3, 400 MHz)
d=5.12(br, 1H), 3.51 (q, 2H, J=6.2 Hz), 2.91 (t, 2H,
J=6.2Hz), 1.43 (s, 9H); FABMS (NBA), m/e 356
[(M+H)+]; HRMS calcd for C14H15O4NF5 [(M+H)+]
356.0921, found 356.0931.
Another important conclusion derived from the data in
Table 3 is that the ꢀTm is not good for an assessment of
the drug binding to DNAs.25 This is especially the case
when the given DNA duplexes for the assessment are
significantly different in Tm from each other. While the
ꢀTm obtained for C_G/GGC in the presence of 1 is
equal to that obtained for G_T/CGA, the binding of 1 to
C_G/GGC is 1.7 fold stronger than the binding to G_T/
CGA. This discrepancy of the data between the ꢀTm
and Kspecific is simply due to a very larger difference in
Tm of the duplex by 12.1 ꢀC between the two oligomers.
3-((tert-Butoxy)carbonylamino)-N-(7-methylpyridinio[3,2-
e]pyridin-2-yl)propanamide (Boc-1). To a solution of the
activated ester (449 mg, 1.26 mmol) in dry DMF (4 mL)
was added 2-amino-7-methyl-1,8-naphthyridine (200
mg, 1.26 mmol) and N,N-diisopropylethylamine (162
mg, 1.26 mmol). The mixture was stirred at 40 ꢀC for 24
h. The solvent was evaporated to dryness and the crude
residue was purified by silica gel column chromato-
graphy (CHCl3/MeOH=50/1) to give Boc-1 (363 mg,
87%) as a pale white solid: 1H NMR (CDCl3, 400 MHz)
ꢃ=8.86 (br, 1H), 8.42(d, 1H, J=9.0 Hz), 8.13 (d, 1H,
J=9.0 Hz), 8.00 (d, 1H, J=8.2Hz), 7.27 (d, 1H, J=8.2
Hz), 5.19 (br, 1H), 3.50 (q, 2H, J=5.9 Hz), 2.74 (s, 3H),
2.70 (t, 2H, J=5.9 Hz), 1.41 (s, 9H); 13C NMR (CDCl3,
100 MHz) ꢃ=171.2, 163.4, 155.9, 154.3, 153.2, 139.2,
136.5, 121.7, 118.6, 114.3, 79.5, 37.5, 36.1, 28.4, 25.5;
FABMS (NBA), m/e 331 [(M+H)+]; HRMS calcd for
C17H23O3N4 [(M+H)+] 331.1770, found 331.1771.
Conclusion
The studies described here showed that the binding of 1
to a guanine bulge is sequence dependent. The mole-
cular basis of the sequence dependence is most likely a
strong stacking stabilization by the flanking G–C base
pairs. For the strong binding to a guanine bulge flank-
ing A–T base pairs, intercalating agents with much
wider aromatic surface is desirable. However, such
intercalators may intercalate more strongly into normal
base pairs than 1 does, resulting in lowering the specifi-
city to the guanine bulge. In order to circumvent these
difficulties, an additional molecular device that binds to
the major groove face of the bound guanine is necessary
to be incorporated into the next generation of the bulge-
targeting intercalators. A guanidium group that can
bind to the N7 and O6 of the G from the major groove
is conceivable for the auxiliary functional group.
3-Amino-N-(7-methylpyridino[3,2-e]pyridin-2-yl)prop-
anamide (1). To a solution of Boc-1 (191 mg, 0.58
mmol) in dry CHCl3 (3 mL) was added ethyl acetate
containing 4 M HCl (1.5 mL) at 0 ꢀC and the mixture
was stirred at room temperature for 0.5 h. The solvent
was evaporated to dryness to give the hydrochloride of
1 (quantitative yield) as a white solid. The hydrochlo-
ride of 1 was dissolved in H2O and extracted into
CHCl3 by the addition of 28% aqueous ammonia solu-
tion. The organic layer was dried over MgSO4 and the
solvent was evaporated in vacuo to give 1 (98.8 mg,
74%) as pale white solids: 1H NMR (CD3OD,
400 MHz) d=8.39 (d, 1H, J=8.9 Hz), 8.25 (d, 1H,
J=8.9 Hz), 8.18 (d, 1H, J=8.2Hz), 7.4 (d, 1H, J=8.2
Hz), 3.02(t, 2H, J=6.4 Hz), 2.71 (s, 3H), 2.67 (t, 2H,
J=6.4 Hz); 13C NMR (CDCl3, 100 MHz) d=173.9,
164.2, 155.5, 140.2, 138.8, 122.8, 119.9, 115.8, 113.6,
40.4, 38.5, 25.0; FABMS (NBA), m/e 231 [(M+H)+];
HRMS calcd for C12H15ON4 [(M+H)+] 231.1246;
found 231.1248.
Experimental
Materials
Reagents and solvents were purchased from standard
suppliers without further purification. 1H and 13C NMR
spectra were measured on a JEOL JNM a-400 (1H
spectra at 400 MHz; 13C spectra at 100 MHz) spectro-
meter. FAB mass spectra were recorded on a JEOL
JMS HX-110 spectrometer.
Measurements of the melting temperature of bulge-con-
taining duplexes. Compound 1 (final concentration of
200 mM) was dissolved in a sodium cacodylate buffer
(10 mM, pH 7.0) containing bulge duplex (4.77 mM,
strand concentration) and NaCl (100 mM). The mixture
was heated for 5 min at 50 ꢀC and cooled slowly to
make sure that the starting oligomer is in a duplex state.
The thermal denaturation profile was recorded on a
JASCO V-550DS spectrometer equipped with a Peltier
temperature controller. The absorbance of the sample
was monitored at 260 nm from 4 to 70 ꢀC with a heating
rate of 1 ꢀC/min.
2,3,4,5,6-Pentafluorophenyl
3-((tert-butoxy)carbonyl-
amino)propanoate. To a solution of Boc-b-alanine (3.0
g, 15.7 mmol) in DMF (10 mL) was added penta-
fluorophenol (2.2 g, 12.0 mmol) and 1-(3-dimethylami-
nopropyl)-3-ethylcarbodiimide hydrochloride (2.8 g,
14.6 mmol). The mixture was stirred at ambient tem-
perature for 24 h. The solvent was evaporated and the

K. Nakatani et al. / Bioorg. Med. Chem. 11 (2003) 2347–2353
2353
Curve fitting
8. White, S. A.; Draper, D. E. Nucleic Acids Res. 1987, 15,
4049.
9. Williams, L. D.; Goldberg, I. H. Biochemistry 1988, 27,
3004.
10. Cortes, C.-J.; Wang, H.-J. Biochemistry 1996, 35, 616.
11. Brockman, H. E.; Goben, W. Science 1965, 147, 750.
12. Nakatani, K.; Sando, S.; Saito, I. J. Am. Chem. Soc. 2000,
122, 2172.
Three data sets of thermal denaturation profiles
obtained by three independent measurements were fitted
to a two-state model with a nonlinear least-square pro-
gram in Sigma Plots (version 2000). The equation used
for the fitting was described in the reference.22
13. Henderson, P. T.; Boone, E.; Schuster, G. B. Helv. Chim.
Acta 2002, 85, 135.
Acknowledgements
14. Boone, E.; Schuster, G. B. Nucleic Acids Res. 2002, 30,
830.
15. Nakatani, K.; Okamoto, A.; Saito, I. Angew. Chem., Int.
Ed. Engl. 1999, 38, 3378.
16. Nakatani, K.; Sando, S.; Saito, I. Nat. Biotechnol. 2001,
19, 51.
17. Jackson, B. A.; Alekseyev, V. Y.; Barton, J. K. Biochem-
istry 1999, 38, 4655.
This work was supported by Grant-in-Aid for Scientific
Research on Priority Areas (C) ‘Medical Genome
Science’ from the Ministry of Education, Culture,
Sports, Science and Technology of Japan.
18. Nakatani, K.; Sando, S.; Saito, I. Bioorg. Med. Chem.
2001, 9, 2381.
References and Notes
19. Nakatani, K.; Sando, S.; Kumasawa, H.; Kikuchi, J.;
Saito, I. J. Am. Chem. Soc. 2001, 123, 12650.
20. Smith, E. A.; Kyo, M.; Kumasawa, H.; Nakatani, K.;
Saito, I.; Corn, R. M. J. Am. Chem. Soc. 2002, 124, 6810.
21. Wang, Y.-H.; Griffith, J. D. Biochemistry 1991, 30, 1358.
22. Petersheim, M.; Turner, D. H. Biochemistry 1983, 22, 256.
23. McDowell, J. A.; Turner, D. H. Biochemistry 1996, 35,
14077.
24. Nakatani, K.; Matsuno, T.; Adachi, K.; Hagihara, S.;
Saito, I. J. Am. Chem. Soc. 2001, 123, 5695.
25. A similar observation regarding a non-parallel relation
between ꢀTm and ꢀG has been reported for the minor groove
recognition. See,Bostock-Smith, C. E.; Searle, M. S. Nucleic
Acids Res. 1999, 27, 1619.
1. Friedberg, E. C.; Walker, G. C.; Siede, W. DNA Repair and
Mutagenesis; ASM Press: Washington, 1995.
2. Streisinger, G.; Okada, Y.; Emrich, J.; Newton, J.; Tsugita,
A.; Terzaghi, E.; Inouye, M. Cold Spring Harbor Symp. Quant.
Biol. 1966, 31, 77.
3. Turner, D. Curr. Opin. Struct. Biol. 1992, 2, 334.
4. Malkov, V. A.; Biswas, I.; Camerini-Otero, R. D.; Hsieh, P.
J. Biol. Chem. 1997, 272, 23811.
5. Wang, Y.-H.; Bortner, C. D.; Griffith, J. J. Biol. Chem.
1993, 268, 17571.
6. Nelson, J. W.; Tinoco, I., Jr. Biochemistry 1985, 24, 6416.
7. Woodson, S. A.; Crothers, D. M. Biochemistry 1988, 27,
8904.