Triplex selective 2-(2-naphthyl)quinoline compounds: Origins of affinity and new design principles

Article abstract of DOI:10.1021/ja034181r

A novel competition dialysis assay was used to investigate the structural selectivity of a series of substituted 2-(2-naphthyl)quinoline compounds designed to target triplex DNA. The interaction of 14 compounds with 13 different nucleic acid sequences and

Full text of DOI:10.1021/ja034181r

Published on Web 05/21/2003
Triplex Selective 2-(2-Naphthyl)quinoline Compounds:
Origins of Affinity and New Design Principles
Jonathan B. Chaires,*,† Jinsong Ren,‡ Maged Henary,§ Oliwia Zegrocka,§
G. Reid Bishop,| and Lucjan Strekowski*,§
Contribution from the Department of Biochemistry, UniVersity of Mississippi Medical Center,
2500 N. State Street, Jackson, Mississippi 39216-4505, Key Laboratory of Rare Earth Chemistry
and Physics, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences,
Changchun, Jilin 130022, China, Department of Chemistry, Georgia State UniVersity,
Atlanta, Georgia 30303, and Department of Chemistry and Biochemistry,
Mississippi College, Clinton, Mississippi 39058
Received January 15, 2003; E-mail: jchaires@biochem.umsmed.edu; lucjan@gsu.edu
Abstract: A novel competition dialysis assay was used to investigate the structural selectivity of a series
of substituted 2-(2-naphthyl)quinoline compounds designed to target triplex DNA. The interaction of 14
compounds with 13 different nucleic acid sequences and structures was studied. A striking selectivity for
the triplex structure poly dA:[poly dT]2 was found for the majority of compounds studied. Quantitative analysis
of the competition dialysis binding data using newly developed metrics revealed that these compounds
are among the most selective triplex-binding agents synthesized to date. A quantitative structure-affinity
relationship (QSAR) was derived using triplex binding data for all 14 compounds used in these studies.
The QSAR revealed that the primary favorable determinant of triplex binding free energy is the solvent
accessible surface area. Triplex binding affinity is negatively correlated with compound electron affinity
and the number of hydrogen bond donors. The QSAR provides guidelines for the design of improved triplex-
binding agents.
Introduction
Triplex DNA (Figure 1) is of intense interest as a target for
small molecule therapeutic agents.1-6 Interest in triplex DNA
as a target lies in two major areas. First, certain polypurine
sequences within genomic DNA can form an intramolecular
triplex structure called H-DNA.2 Although the precise biological
role of H-DNA is not definitively known, it could function as
a regulatory signal for the control of gene expression. If so,
small molecules that could selectively recognize H-DNA might
be useful therapeutic agents as modulators of gene expression.
Second, there is interest in the use of triplex forming oligo-
nucleotides (TFOs) in the “antigene” therapeutic strategy.7-12
† Department of Biochemistry, University of Mississippi Medical Center.
‡ Key Laboratory of Rare Earth Chemistry and Physics, Changchun
Institute of Applied Chemistry, Chinese Academy of Sciences.
§ Department of Chemistry, Georgia State University.
| Department of Chemistry and Biochemistry, Mississippi College.
(1) Jenkins, T. C. Curr. Med. Chem. 2000, 7, 99-115.
(2) Soyfer, V. N.; Potaman, V. N. Triple-Helical Nucleic Acids; Springer: New
York, 1996.
Figure 1. Structures of a dT-dA-dT triplet (top) and LS-8 (bottom).
(3) Shafer, R. H. Prog. Nucleic Acid Res. Mol. Biol. 1998, 59, 55-94.
(4) Vasquez, K. M.; Glazer, P. M. Q. ReV. Biophys 2002, 35, 89-107.
(5) Mergny, J. L.; Duval-Valentin, G.; Nguyen, C. H.; Perrouault, L.; Faucon,
B.; Rougee, M.; Montenay-Garestier, T.; Bisagni, E.; Helene, C. Science
1992, 256, 1681-1684.
In this strategy, TFOs can be used to selectively target precise
sequences within the genome by formation of an intermolecular
triplex. Such triplex formation can interfere with biological
function. Small molecules that selectively bind to such triplex
(6) Thuong, N. T.; Helene, C. Angew. Chem., Int. Ed. Engl. 1993, 32, 666-
690.
(7) Helene, C.; Thuong, N. T.; Harel-Bellan, A. Ann. N. Y. Acad. Sci. 1992,
660, 27-36.
(8) Helene, C.; Giovannangeli, C.; Guieysse-Peugeot, A. L.; Praseuth, D. Ciba
Found. Symp. 1997, 209, 94-102.
(10) Chan, P. P.; Glazer, P. M. J. Mol. Med. 1997, 75, 267-282.
(11) Casey, B. P.; Glazer, P. M. Prog. Nucleic Acid Res. Mol. Biol. 2001, 67,
163-192.
(9) Giovannangeli, C.; Helene, C. Antisense Nucleic Acid Drug. DeV. 1997, 7,
413-421.
(12) Knauert, M. P.; Glazer, P. M. Hum. Mol. Genet. 2001, 10, 2243-2251.
9
7272
J. AM. CHEM. SOC. 2003, 125, 7272-7283
10.1021/ja034181r CCC: $25.00 © 2003 American Chemical Society

Triplex Selective 2-(2-Naphthyl)quinoline Compounds
A R T I C L E S
structures may serve as “enhancers” that stabilize the TFO-
duplex complex, amplifying their biological effects.
interaction is also stronger for compounds with a single alkyl
group attached to the N4 atom of the quinoline in comparison
to quinolin-4-amines substituted with two alkyl groups. The
primary R-N4H function, as in LS-8, has low steric require-
ments which allow for efficient conjugation of the electron pair
on the N4 atom with the quinoline system, thereby increasing
basicity of ring N1 atom of the quinoline.18 By contrast,
quinolin-4-amines substituted at N4 with two alkyl groups (that
is, containing a sterically hindered secondary R2N4 function)
show a decreased conjugation and, concomitant decreased
basicity of the quinoline N1 atom.
Ethidum bromide, a classic intercalator, was found to bind
to triplex structures, but with a lower affinity than for duplex
DNA.13,14 The Helene group first attempted the rational design
of triplex-selective intercalating agents with BePI (7H -8
-methylbenzo [e] pyrido[4,3-b]indole),5 and subsequently re-
ported improved selectivity with new compounds based on that
structure.15,16 Another strategy for the design of triplex selective
agents was based on naphthylquinoline compounds.17 Three
essential elements were considered in that design strategy. First,
compounds should be cations to complement the high negative
charge density of the triple helix. Second, compounds should
have aromatic surface areas, which can optimally stack on the
crescent-shaped base triplet (Figure 1). Finally, compounds
should have an unfused aromatic system with torsional flex-
ibility, since the bases within the triplet are propeller twisted.
The compound designated LS-8 (Figure 1) fulfilled these design
principles, and proved to be a highly selective triplex interca-
lator. Extensive studies with LS-8 suggested that it binds to
triplex DNA by intercalation.17 Both the quinoline ring nitrogen
atom (pKa ) 7.1) and the terminal amino group of the side chain
(pKa > 8) are protonated in the intercalation complex,18 yielding
a dication with a favorable polyelectrolyte contribution to the
binding free energy. Molecular modeling studies17 are consistent
with intercalation of the quinoline portion of LS-8 between base-
pairs of the original duplex within the triplex, stacking of the
2-naphthyl substituent with bases of the third DNA strand, and
location of the aminoalkyl side chain in the minor groove.
By using footprinting experiments19-24 and Tm measure-
ments17,18,24 it was shown that quinolin-4-amines containing an
aryl group at position 2 and an aminoalkyl moiety at the N4
atom bind selectively to triplex DNA in the presence of the
corresponding duplex DNA. Analogues of LS-8 with a smaller
aromatic substituent than the 2-naphthyl group showed a
decreased triplex/duplex binding selectivity, apparently because
the smaller ring system does not stack optimally with bases of
the third strand of the triplex.
Limited studies conducted so far with substituted quinolin-
4-amines suggested that additional structural features of the side
chain, such as its length, bulkiness, hydrophobicity, and the
presence of heteroatoms may play an important role in the
interaction of these compounds with nucleic acids. These aspects
were addressed in the design of new quinolines for this work
(Figure 2). Studies of the nucleic acid binding of these
compounds were facilitated by the use of a newly developed
competition dialysis assay that allows for a quantitative
comparison of ligand interactions with 13 different nucleic acid
structures and sequences. The systematic studies reported here
reveal specific compounds with improved triplex binding relative
to the parent LS-8, and further allow for a detailed quantitative
structure-affinity relationship to be derived.
The competition dialysis method is now firmly established
as an important tool for rapidly screening ligand-nucleic acid
interactions.25-29 In the competition dialysis method, a test
ligand of interest is dialyzed against an array of nucleic acid
structures. At equilibrium, the free ligand concentration is
identical for all structures, and the amount bound to each
structure directly and quantitatively indicates the binding affinity.
The preference of the ligand for a given nucleic acid structure
is unambiguously identified. Competition dialysis offers distinct
advantages over thermal denaturation and footprinting methods
that have been widely used in studies of structural selectivity.
First and foremost, a large number of different structures and
sequences may be simultaneously compared in the competition
dialysis experiment, in contrast to footprinting and thermal
denaturation experiments where only duplex and one other
structure (triplex, for example) are typically compared. Second,
competition dialysis is firmly grounded in equilibrium thermo-
dynamics, and measures binding directly. In contrast, interpreta-
tion of thermal denaturation curves is often difficult because of
the complexities of the underlying statistical mechanical mech-
anisms involved in nucleic acid melting reactions.30,31 In
particular, ligand redistribution can produce complex, multipha-
sic melting curves that are not easily interpreted. Finally, the
competition dialysis method is rapid compared to both thermal
denaturation and footprinting studies, and is amenable to high
sample throughput. While the existing competition dialysis
method provides a firm, quantitative measure of ligand structural
The interaction of 2-arylquinolin-4-amines with nucleic acids
is strongly affected by the structure of the side chain at the N4
atom of the quinoline. In general, binding is enhanced for
derivatives containing a terminal amino function, such as Me2N
in LS-8, that is protonated under binding conditions. The
(13) Mergny, J. L.; Collier, D.; Rougee, M.; Montenay-Garestier, T.; Helene,
C. Nucleic Acids Res. 1991, 19, 1521-1526.
(14) Scaria, P. V.; Shafer, R. H. J. Biol. Chem. 1991, 266, 5417-5423.
(15) Escude, C.; Nguyen, C. H.; Kukreti, S.; Janin, Y.; Sun, J. S.; Bisagni, E.;
Garestier, T.; Helene, C. Proc. Natl. Acad. Sci. U S A 1998, 95, 3591-
3596.
(16) Pilch, D. S.; Waring, M. J.; Sun, J. S.; Rougee, M.; Nguyen, C. H.; Bisagni,
E.; Garestier, T.; Helene, C. J. Mol. Biol. 1993, 232, 926-946.
(17) Wilson, W. D.; Tanious, F. A.; Mizan, S.; Yao, S.; Kiselyov, A. S.; Zon,
G.; Strekowski, L. Biochemistry 1993, 32, 10 614-10 621.
(18) Strekowski, L.; Gulevich, Y.; Baranowski, T. C.; Parker, A. N.; Kiselyov,
A. S.; Lin, S. Y.; Tanious, F. A.; Wilson, W. D. J. Med. Chem. 1996, 39,
3980-3983.
(19) Cassidy, S. A.; Strekowski, L.; Wilson, W. D.; Fox, K. R. Biochemistry.
1994, 33, 15 338-15 347.
(20) Cassidy, S. A.; Strekowski, L.; Fox, K. R. Nucleic Acids Res. 1996, 24,
4133-4138.
(21) Keppler, M. D.; McKeen, C. M.; Zegrocka, O.; Strekowski, L.; Brown,
T.; Fox, K. R. Biochim. Biophys. Acta 1999, 1447, 137-145.
(22) Keppler, M.; Zegrocka, O.; Strekowski, L.; Fox, K. R. FEBS Lett. 1999,
447, 223-226.
(25) Ren, J.; Chaires, J. B. Biochemistry 1999, 38, 16 067-16 075.
(26) Ren, J.; Chaires, J. B. Methods Enzymol. 2001, 340, 99-108.
(27) Ren, J.; Bailly, C.; Chaires, J. B. FEBS Lett. 2000, 470, 355-359.
(28) Ren, J.; Chaires, J. B. J. Am. Chem. Soc. 2000, 122, 424-425.
(29) Ren, J.; Qu, X.; Dattagupta, N.; Chaires, J. B. J. Am. Chem. Soc. 2001,
123, 6742-6743.
(23) Chandler, S. P.; Strekowski, L.; Wilson, W. D.; Fox, K. R. Biochemistry.
1995, 34, 7234-7242.
(24) Strekowski, L.; Gulevich, Y.; Van Aken, K.; Wilson, W. D.; Fox, K. R.
Tetrahedron Lett. 1995, 36, 225-228.
(30) McGhee, J. D Biopolymers. 1976, 15, 1345-1375.
(31) Crothers, D. M. Biopolymers 1971, 10, 2147-2160.
9
J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003 7273

A R T I C L E S
Chaires et al.
Figure 2. Structures of the naphthylquinoline compounds used in this investigation.
selectivity, we describe here new data analysis tools that make
it an even more powerful and exacting measure of structural
selectivity.
4-hydroxy-2-(2-naphthyl)quinoline with 2-(dimethylamino)ethyl
chloride (not shown).
The synthesis of MHQ-13, OZ-91, and OZ-97 by the
reaction of 1 with a diamine was conducted with a large excess
of the diamine to decrease the formation of a bis-quinoline
byproduct.
Results
Synthesis of Quinoline Ligands. Readily available 4-chloro-
2-(2-naphthyl)quinoline32 (1) was a starting material for the
preparation of quinoline derivatives containing an unsubstituted
2-naphthyl group (Scheme 1). A nucleophilic displacement of
chloride from 1 by the reaction with a primary amine in the
presence of a catalytic amount of SnCl4 furnished all MHQ
quinolin-4-amines, OZ-91, and OZ-97. This one-step approach
could not be used for the synthesis of OZ-115 and OZ-124
because the corresponding amines are not readily available.
These quinoline derivatives were prepared from hydroxyalkyl-
amino-substituted compounds 2,3 which were transformed into
chlorides 4,5 by the reaction with thionyl chloride followed by
treatment of 4,5 with an amine. In the synthesis of OZ-153,
compound 1 was treated with 4-hydroxyaniline and the resultant
phenol derivative 6 was subjected to a Mannich reaction with
formaldehyde and morpholine. The 4- alkoxyquinoline MHQ-
12 was synthesized by alkylation of a sodium derivative of
A different approach was used to synthesize quinoline OZ-
85H that contains a hydroxyalkyl-functionalized naphthyl
substituent (Scheme 2). Thus, a lithium 2-(dimethylamino)-
ethylamide mediated cyclization 33-35 of a Schiff base 7 derived
from 2-(trifluoromethyl)aniline and 6-methoxynaphthone fur-
nished methoxynaphthyl-substituted quinoline 8. Demethylation
36 of 8 by the reaction with BBr3 followed by alkylation of the
resultant naphthol 9 with 1-bromo-6-(tert-butyldimethylsiloxy)-
hexane gave 10. Compound OZ-85H was obtained by desily-
lation of 10. Due to high efficiencies of the first transformations
7 f 8 f 9 it was not necessary to rigorously purify the
(33) Strekowski, L.; Mokrosz, J. L.; Honkan, V. A.; Czarny, A.; Cegla, M. T.;
Wydra, R. L.; Patterson, S. E.; Schinazi, R. F. J. Med. Chem. 1991, 34,
1739-1746.
(34) Strekowski, L.; Wydra, R. L.; Cegla, M. T.; Czarny, A.; Harden, D. B.;
Patterson, S. E.; Battiste, M. A.; Coxon, J. M. J. Org. Chem. 1990, 55,
4777-4779.
(35) Strekowski, L.; Patterson, S. E.; Janda, L.; Wydra, R. L.; Harden, D. B.;
Lipowska, M.; Cegla, M. T. J. Org. Chem. 1992, 57, 196-201.
(36) Greene, T. W.; Wuts, P. G. M. ProtectiVe Groups in Organic Synthesis;
Wiley: New York, 1999.
(32) Strekowski, L.; Zegrocka, O.; Windham, C.; Czarny, A. Org. Proc. Res.
DeV. 1997, 1, 384-386.
9
7274 J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003

Triplex Selective 2-(2-Naphthyl)quinoline Compounds
A R T I C L E S
Scheme 1
Figure 3. (A) Results of competition dialysis experiments using LS-8, the
parent naphthylquinoline compound. (B) Normalized competition dialysis
data, resulting from transformation of the data shown in (A) by dividing
all Cbound values by Cmax) 12.5 µM. The normalized data emphasizes the
relative binding affinity for LS8 for the different nucleic acid structures
and is used for calculation of the specificity sum.
Table 1. Nucleic Acid Samples Used in Competition Dialysis
Experiments
λ
ꢀ
monomeric
unit
conformation
nucleic acid
poly dT
(nm) (M-1cm-1
)
single-strand
pyrimidine
single-strand
purine
264
257
260
260
260
8520
8600
nucleotide
nucleotide
base pair
base pair
base pair
Scheme 2
poly dA
duplex DNA
C. perfringens
(31%GC)
calf thymus
(42%GC)
M. lysodeikticus
12476
12824
13846
(72%GC)
poly dA: poly dT
[poly (dAdT)]2
[poly (dGdC)]2
poly rA:poly rU
260
262
254
260
260
254
12000
13200
16800
14280
12460
16060
17200
39267
base pair
base pair
base pair
base pair
base pair
base pair
triplet
duplex RNA
DNA-RNA hybrid poly rA: poly dT
Z DNA
triplex DNA
tetraplex DNA
[Br-poly (dGdC)]2
poly dA: (poly dT)2 260
(5′T2G20T2)4 260
tetrad
Key: λ, wavelength; ꢀ, molar extinction coefficient at the wavelength λ,
expressed in terms of the monomeric unit specified.
ligand accumulates in those solutions with preferred binding
structures or sequences. Table 1 lists the nucleic acid structures
used in this study. The data of Figure 3A show unequivocally
that LS-8 binds preferentially to the poly dA-(poly dT)2 triplex
structure. LS-8 binds with lesser affinity to DNA duplex forms,
to an RNA:DNA hybrid structure, and to a parallel-stranded
tetraplex structure. LS-8 shows little or no binding to single-
stranded DNA, to RNA or to left-handed Z DNA. Figure 3B
shows the normalized binding profile, in which the amount
bound to each structure was normalized relative to the maximum
amount bound to the triplex form. Figure 3B shows that the
relative binding to the triplex structure is at least 2.5 times
greater than to any other structure or sequence. The normalized
affinity profile will be used later for a quantitative assessment
of binding selectivity.
intermediate products 7-9. On the other hand, the intermediate
compound 10 and the final product OZ-85H were obtained in
an analytically pure form and thoroughly characterized by
spectral methods and elemental analysis. The structures of the
remaining quinolines used in this study (Figure 2) were also
fully consistent with the obtained analytical data.
DNA Binding Analysis. Figure 3A shows the results of a
competition dialysis experiment using LS-8, the parent naph-
thylquinoline.17 In the competition dialysis experiment, the
different nucleic acid structures and sequences are dialyzed
against a common free ligand solution. At equilibrium, more
Figure 4 shows the comparative binding profiles for all 14
naphthylquinoline compounds designed and synthesized for this
study. Figure 4 contains a wealth of quantitative data. What is
9
J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003 7275

A R T I C L E S
Chaires et al.
Figure 5. Comparative binding free energy difference plots. (A) The
difference in triplex binding free energy (∆∆GTriplex,LS8) relative to the
binding of the parent compound LS825,26 for each of the naphthylquinoline
derivatives. Negative value of ∆∆GTriplex,LS8 indicate higher triplex binding
affinity relative to LS-8; positive values indicate weaker affinity. (B) The
difference in triplex binding free energy (∆∆GTriplex,CT DNA) relative to
binding to duplex calf thymus DNA. All of the naphthylquinoline derivatives
bind preferentially to the triplex structure over duplex calf thymus DNA.
Figure 4. Comparative binding of 14 naphthylquinoline derivatives to 13
nucleic acid structures.
Table 2. Selected Binding Properties of Naphthylquinoline
Compounds
The magnitude of (∆∆GTriplex,
DNA) is a quantitative measure of the
CT
preference for triplex.
-1
compd
Cbound, µM
K
app/105, M
∆G
Triplex
∆G
CTDNA
SS
Cmax/SS
LS8
12.5
19.4
11.9
11.8
13.2
1.7
12.9
13.6
17.7
12.5
16.5
21.8
24.7
0.9
2.00
3.48
1.88
1.87
2.14
0.23
2.09
2.21
3.08
2.00
2.83
4.10
4.91
0.12
-7.08
-7.41
-7.05
-7.04
-7.13
-5.84
-7.11
-7.14
-7.34
-7.09
-7.29
-7.50
-7.61
-5.45
-6.09
-5.98
-5.68
-4.18
-5.49
0
-5.79
-6.48
-5.98
-6.52
-6.38
-6.63
-6.66
0
3.06
1.97
2.11
1.32
1.77
1.00
1.97
3.88
2.18
5.33
3.49
3.14
3.05
1.50
4.09
9.83
5.64
8.93
7.49
1.71
6.58
3.50
8.12
2.35
4.75
6.95
8.10
0.59
binding to triplex DNA, the absorbance spectrum of LS-8
undergoes a red shift from a maximum near 325 nm for the
free form to a maximum near 360 nm for the bound form (Figure
S3A, Supporting Information). A binding isotherm was con-
structed by recording the absorbance of a fixed concentration
of LS-8 (5 µM) while varying the triplex concentration from
0.1 µM to 0.4 mM. Analysis of the binding isotherm 37 yielded
MHQ9
MHQ11
MHQ12
MHQ13
MHQ14
MHQ15
MHQ17
OZ85H
OZ91
a binding constant of 2.3 ((0.4) × 105 M-1. The value of Kapp
)
OZ97
2.0 × 105 M-1 obtained by competition dialysis (Table 2) is in
excellent agreement with the more rigorously determined
binding constant. The close agreement between Kapp values and
spectrophotometrically determined binding constants was previ-
ously demonstrated in our laboratory for the triplex binding of
a cyanine dye 28 and for the binding of the indolocarbazole NB-
506 to calf thymus DNA.27
OZ115
OZ124
OZ153
immediately apparent in Figure 4 is that all naphthylquinolines
studied show preferential binding to the poly dA-(poly dT)2
triplex structure. There are quantitative differences among these
naphthylquinolines, which will be discussed in detail in later
sections.
Figure 5 shows comparative binding data for the compounds
used in this study. Data are shown as differences in binding
free energy relative to two different references. Figure 5A shows
the difference in triplex binding free energy of each compound
relative to the parent compound, LS-8. This difference is simply
∆∆GTriplex, LS8 ) ∆GTriplex - (-7.08), where ∆GTriplex values
are listed in Table 2 and -7.08 is the binding free energy of
LS-8. Figure 5A emphasizes those compounds with lesser or
greater binding free energy relative to LS-8, with negative values
signifying more favorable binding and positive values less
favorable binding. The data of Figure 5A show that compounds
MHQ-9, OZ-85H, OZ-97, OZ-115, and OZ-124 all bind more
tightly to triplex DNA than does LS-8. In contrast, MHQ-14
and OZ-153 bind less tightly.
Apparent binding constants for each structure or sequence,
K
app, may be calculated from competition dialysis data such as
shown in Figures 3A and 4. 26 The simple relationship is Kapp
)
Cb/(Cf)(Stotal - Cb), where Cb is the amount of ligand bound,
Cf is the free ligand concentration and Stotal is the total nucleic
acid concentration. By virtue of the experimental design used
in the competition dialysis experiment, Cf ) 1 µM and Stotal
)
75 µM (expressed in terms of the monomeric unit of the nucleic
acid, i.e., nucleotides, base pairs, triplets or tetrads). Table 2
shows the calculated Kapp values for triplex binding, which range
from 0.12 × 105 M-1 for OZ-153 to 4.91 × 105 M-1 for OZ-
124. We estimate, from replicate experiments, that Kapp values
contain errors in the range of 10-15%. Triplex binding free
energies may be calculated by using the standard relationship
∆GTriplex ) -RTlnKapp (Table 2).
Thermal denaturation experiments were done using selected
compounds to confirm the trends shown in Table 2 and in Figure
5A (Figure S4, Supporting Information). Melting was studied
To validate Kapp values, a specrophotometric titration 37 study
was done for the interaction of LS-8 with triplex DNA. Upon
(37) Qu, X.; Chaires, J. B. Methods Enzymol. 2000, 321, 353-369.
9
7276 J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003

Triplex Selective 2-(2-Naphthyl)quinoline Compounds
A R T I C L E S
for the poly dA:(poly dT)2 triplex alone and in the presence of
LS-8, MHQ-9, MHQ-12, MHQ-14, and OZ-124. Compounds
were added at a molar ratio of 0.25 (mol compound/mol triplet),
and ∆Tm was determined by difference relative to poly dA:
(poly dT)2 alone. In the absence of added ligand, the triplex
melted at 47.0 °C, while the duplex melted at 76.3 °C. Figure
S4 shows that the increase in triplex melting is linearly related
to Kapp, with ∆Tm ) 18.1 + 4.5 Kapp/105 (R ) 0.926; P )
0.023). These data confirm the trends in triplex binding shown
in Table 2 and in Figure 5A. Such confirmation is not
unexpected, since we previously have directly correlated
competition binding data with data from thermal denaturation
studies, spectrophotometric binding studies, and enzymatic
assays.25-29
binding to all structures in the assay, and a complete lack of
selectivity. By way of illustration, the data for LS-8 in Figure
3 may be used to calculate SS as follows. In Figure 3A, Cmax
is seen to be the amount bound to triplex and equals 12.4 µM.
The amounts bound to the other species range from 0 to 6 µM.
Normalizing the data of Figure 3A with Cmax ) 12.4 µM yields
the graph in Figure 3B. Summation of the normalized binding
data in Figure 3B yield SS ) 3.06. Values of SS for all
compounds studied are listed in Table 2 and are shown in Figure
6A. SS values range from 1 to 5.33. Several compounds show
improved selectivity relative to the parent compound LS-8, with
SS < 3.0. For comparison, an average value of SS ) 4.5 ( 2.0
was determined from competition dialysis data obtained for 126
compounds representing a wide variety of chemical classes and
DNA binding modes (Chaires, J. B.; Ren, J., in preparation).
All but one of the naphthylquinoline compounds studied here
show better than average selectivity as judged by SS values
Although ∆∆GTriplex, LS8 provides a quantitative measure of
triplex affinity relative to the parent compound, it contains no
information about selectivity toward the nucleic acid array used.
To evaluate selectivity, the difference ∆∆GTriplex,
)
CTDNA
One limitation of the SS metric is that it does not contain
information about compound binding affinity. This limitation
may be circumvented by calculating the ratio Cmax/SS where
∆GTriplex - ∆GCTDNA was calculated. This is the difference
between the binding free energies of each naphthylquinoline
for triplex and calf thymus DNA, where calf thymus DNA is
taken to represent a standard duplex reference form. The
magnitude of ∆∆GTriplex, LS8 provides a measure of the selectivity
of triplex binding over duplex, with larger magnitudes indicating
greater selectivity. Figure 5B shows the results. All of the
difference values are negative, indicating that for all compounds
binding to triplex is favored over binding to duplex calf thymus
DNA. Compounds MHQ-14 and OZ-153 show the largest
differences, arising from the fact that neither compound binds
appreciably to duplex DNA. Note, however, that both MHQ-
14 and OZ-153 bind less tightly to triplex DNA than does the
parent compound LS-8 (Figure 5A). For these two compounds,
greater selectivity was gained at the expense of binding affinity.
The combined data of Figure 5A and B provide great insight
into the behavior of the naphthylquinolines. Inspection of the
data shows, for example, that MHQ-12 binds to triplex DNA
essentially as well as LS-8, but is more selective for triplex
over duplex DNA as signified by its ∆∆GTriplex, CTDNA value.
MHQ-9 and OZ-124 both bind more tightly to triplex DNA
compared to LS-8, but their selectivity for triplex over duplex
is not improved relative to the parent compound. The competi-
tion dialysis data permit quantitative conclusions to be made
about relative affinity and selectivity.
Cmax is the maximal amount bound as defined above. This ratio
embodies both affinity and selectivity. Cmax directly measures
compound affinity. If SS ) 1, the maximal value of Cmax/SS
will be obtained, whereas if SS ) 13 (no selectivity), the
minimal value of the ratio will result. From results of competi-
tion dialysis studies on 126 compounds (Chaires, J. B.; Ren, J.,
in preparation), Cmax/SS was found to range from 0.06 to 9.8,
with an average value of 2.4 ( 2.2. Values of Cmax/SS for the
compounds studied here are shown in Table 2 and Figure 6B.
Several compounds, notably MHQ-9, MHQ-12, OZ-85H, and
OZ-124, show values of Cmax/SS g 8.0, indicative of both high
selectivity and affinity. All of these are greatly improved over
the parent compound LS8 with Cmax/SS ) 4.1.
To attempt to understand the molecular determinants of triplex
binding affinity for these naphthylquinoline compounds, a
quantitative structure-affinity relationship (QSAR) was derived
from the experimental binding constants (Table 2) and computed
molecular descriptors. A full discussion of the QSAR construc-
tion is provided as Supporting Information. The best three-term
QSAR to emerge was as follows
log Kapp ) 0.00264(( 0.99965)SASA -
0.693(( 0.125)EA - 0.196(( 0.02)HBa + 4.66(( 0.44)
Even more exacting measures of selectivity and affinity may
be derived from the competition dialysis data. These are
described for the first time here. Normalized competition dialysis
data, such as shown in Figure 3B for LS-8, may be used to
calculate the specificity sum, SS. SS is the sum of the normalized
amounts bound to each nucleic acid species i in the assay
(1)
N ) 14, R ) 0.959; RMSE ) 0.130; F ) 49.84; P )
0.0001
In this relationship, log Kapp is the logarithm of the apparent
binding constant (Table 2), SASA is the total solvent accessible
surface area in Å2, EA is electron affinity in eV, and HBa is
the number of hydrogen bond acceptors. The physical meaning
of this is as follows. As SASA increases, log Kapp increases in
magnitude, indicating higher affinity for triplex DNA. Increases
in the magnitudes of EA and HBa result in decreasing binding
affinity. Increasing the solvent accessible surface areas of
naphthylquinoline compounds results in higher affinity for the
triplex. Greater electron affinity and more hydrogen bond
acceptors reduce the affinity of naphthylquinolines for triplex
DNA.
Cb, i
SS )
∑
Cmax
i
where Cb, is the amount bound and Cmax is the maximum
i
amount bound to any species. The index i ranges from 1 to 13
in the current version of the assay, corresponding to the 13
different nucleic acid structures and sequences used.
SS can thus range from 1 to 13, with a value of 1 indicative
of absolute selectivity with compound binding to only one
structure. A value of SS ) 13, in contrast, would indicate equal
9
J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003 7277

A R T I C L E S
Chaires et al.
Table 3. Properties of Triplex-Binding Compounds
compd
SS
Cmax/SS
∆G
duplex
∆G
triplex
ethidium
BePI
coralyne
berberine
R-naphthoflavone
2,6 anthraquinone
LS8
5.05
6.39
4.34
1.59
1.43
5.08
3.06
1.32
1.93
1.38
5.22
3.55
3.67
2.55
4.08
8.93
-6.4
-6.5
-6.3
-4.7
0
-6.1
-6.1
-4.2
-6.5
-6.8
-7.5
-6.6
-6.5
-6.7
-7.1
-7.0
MHQ12
Figure 6. Values of SS and Cmax/SS for naphthylquinoline compounds.
Discussion
The discovery of small molecules that selectively bind to
triplex DNA is of intense interest. A new competition dialysis
assay developed in our laboratory provides the most rigorous
method available for rapidly screening the structural selectivity
of small molecules. New methods of data analysis are described
here that make competition dialysis an even more powerful tool
for measure the structural selectivity of new ligands. By using
that assay, we show here that a new generation of naphthyl-
quinoline compounds represent some of the best triplex-selective
molecules synthesized to date.
Figure 7. Comparison of the results of competition dialysis experiments
on MHQ12 and several compounds reported to be selective for triplex
structures.
The new naphthylquinolines are based on the parent com-
pound LS-8 (Figure 2). LS-8 was designed as a triplex binder
using three design principles.17 First, it is a dication, and its
positive charges should interact with the negatively charged
triplex structure with a large, favorable polyelectrolyte contribu-
tion to its binding free energy. Second, it has a crescent shaped
aromatic ring surface that should stack well on the base triplets
within the triplex structure (Figure 1). Finally, its ring system
is unfused, providing the flexibility to adapt to any propeller
twists with the triplex structure. An unanticipated design element
emerged from initial studies on LS-8. Molecular modeling
studies revealed that the alkylamine substituent on LS-8 fits
snugly into the minor groove of the triplex, providing additional
stabilizing interactions.17 Jenkins and co-workers had long
recognized the influence of side chains and groove occupancy
on intercalator stability,38-41 and subsequently demonstrated that
substituted anthraquinones with substituents that occupied two
triplex grooves showed improved triplex stability and selectiv-
ity.42 The new naphthylquinoline compounds synthesized and
studied here (Figure 2) were designed to explore the effects
resulting from modifications of groove interactions on triplex
binding.
MHQ-15, and OZ-85H all show significantly lower SS values
and significantly higher Cmax/SS values when compared to the
parent compound LS-8 (Figure 6). MHQ-12 stands out as the
compound with the greatest improvement in triplex selectivity,
while maintaining affinity. Indeed, MHQ-12 stands out as one
of the best triplex binders synthesized to date when compared
to other compounds (Table 3, Figure 7). Table 3 lists binding
properties for a variety of triplex binding compounds, whose
structures are shown in Figure 8. MHQ-12 has the lowest SS
value and the highest Cmax/SS value among these compounds
(Table 3), indicating combined high selectivity and affinity.
Berberine and R-naphtholflavone both have SS values ap-
proaching that found for MHQ-12, but have much lower binding
affinity. Coralyne has higher triplex affinity than does MHQ-
12, but is less selective. Figure 7 shows comparative normalized
competition dialysis data for the compounds listed in Table 3.
MHQ-12 clearly stands out for its nearly exclusive binding to
triplex.
Why is MHQ-12 a better triplex binder than LS-8? The
possible reason is somewhat subtle. MHQ-12 differs from LS-8
only at one position, the terminal atom in the R1 substituent
(Figure 2) where oxygen was substituted for nitrogen. That
substitution lowers the energy barrier for rotation of the side
chain, facilitating its reorientation within the minor groove. In
vacuo, semiempirical computations using the AM1 method (not
shown) show that the N f O substitution lowers the bond
rotation energy barrier by 4-5 kcal mol-1. One explanation
for the greater triplex selectivity of MHQ-12 relative to LS-8,
therefore, is that it results from more facile reorientation of the
alkylamine chain within the minor groove of the triplex. That
Several of the new naphthylquinolines show greatly improved
triplex affinity and selectivity. MHQ-9, MHQ-12, MHQ-13,
(38) McKenna, R.; Beveridge, A. J.; Jenkins, T. C.; Neidle, S.; Denny, W. A.
Mol.Pharmacol. 1989, 35, 720-728.
(39) Agbandje, M.; Jenkins, T. C.; McKenna, R.; Reszka, A. P.; Neidle, S. J.
Med. Chem. 1992, 35, 1418-1429.
(40) Neidle, S.; Jenkins, T. C. Methods Enzymol. 1991, 203, 433-458.
(41) Tanious, F. A.; Jenkins, T. C.; Neidle, S.; Wilson, W. D. Biochemistry
1992, 31, 11 632-11 640.
(42) Haq, I.; Ladbury, J. E.; Chowdhry, B. Z.; Jenkins, T. C. J. Am. Chem. Soc.
1996, 118, 10 693-10 701.
9
7278 J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003

Triplex Selective 2-(2-Naphthyl)quinoline Compounds
A R T I C L E S
to have a reduced charge of +1 relative to the dicationic parent
compound. It therefore will have a lower polyelectrolyte
contribution to its binding free energy.43,44 The polyelectrolyte
contribution for the interaction of charged ligands with nucleic
acids (∆Gpe) is given by the equation ∆Gpe ) RT(ZΨ)-
ln[M+],43,45,46 where R is the gas constant, T is the temperature,
Z is the charge on the ligand, Ψ is the extent of counterion
condensation, and M+ is the monovalent cation concentration
(0.2 M in the competition dialysis experiment). For duplex
DNA, with 2 phosphates per 3.4 Å, Ψ ) 0.88,43,47 whereas for
triplex DNA, with 3 phosphates per 3.4 Å, Ψ ) 0.923. Loss of
a +1 charge would thus reduce the polyelectrolyte contribution
to the binding free energy by about 0.9 kcal mol-1, which can
account for most of the 1.2 kcal mol-1 difference in triplex
binding free energy observed between LS-8 and MHQ-14
(Table 2, Figure 5). These data reinforce the importance of
ligand charge as a design principle for targeting triplexes.
The reduced affinity of OZ-153 arises from another source.
OZ-153 is a trication, and should otherwise have a more
favorable polyelectrolyte contribution to its binding free energy
than LS-8, yet it binds worse by 1.6 kcal mol-1 (Table 2, Figure
5). As is seen in Figure 2, however, OZ153 has a bulky,
branched side chain that would be difficult to fit into the minor
groove of the triplex. The substantial steric hindrance presented
by that side chain most probably results in greatly reduced
binding affinity.
The origins of naphthylquinoline triplex affinity were further
explored by construction of a QSAR, which is shown above
(eq 1). Because the compounds used in this study shared a
common intercalating ring structure, the derived QSAR reflects
primarily the energetic contributions of the groove-binding
substituents. Details of the QSAR construction are presented
as Supporting Information, which also describes demonstrations
of the stability and validity of the derived QSAR. The primary
positive contribution to triplex binding affinity revealed by the
QSAR is the total solvent accessible surface area (SASA) of
the side chain. This finding is fully consistent with previous
thermodynamic studies from this laboratory that show that
binding of both intercalators and groove binders is driven by a
large, favorable free energy contribution from the hydrophobic
effect.48-50 A substantial binding free energy contribution results
from the removal of nonpolar surface area by complex forma-
tion. Such is the case here. Burial of hydrophobic side chains
with greater solvent accessible surface areas within the triplex
groove provides a proportionally larger free energy contribution.
Unexpected properties were revealed by the QSAR that oppose
triplex binding. Triplex binding affinity is negatively correlated
with both electron affinity and the number of hydrogen bond
acceptors. Both of these are important solvation descriptors. 51
Their opposing influence on triplex affinity indicates that the
Figure 8. Structures of triplex binders.
explanation implies that ligand flexibility is an important
component of triplex binding. Ligands that can more easily
change conformation to adapt to elements of the triplex structure
will have more favorable interactions. That conclusion suggests
an important new design element for targeting triplex structures,
ligand flexibility.
OZ-85H shows improved selectivity and slightly improved
affinity for triplex relative to LS-8. OZ-85H is unique among
the naphthylquinolines studied here in that it has a second
pendant chain attached. This chain was positioned so that it
might occupy a second groove within the triplex structure. It is
thus similar to the 2,6-disubstituted anthraquinone compounds
developed by Jenkins and colleagues as triplex selective
agents.39,42 These bind to triplexes by an intercalative “thread-
ing” mode. The addition of a second chain that can occupy an
additional groove produces a molecule with distinctly improved
triplex binding properties, although the improvement is not as
great as is seen for molecules such as MHQ-12 with single
chain modifications. Nonetheless, OZ-85H represents a promis-
ing lead structure, and systematic modification of both of its
pendant chains could yield greatly improved triplex binders.
(43) Record, M. T., Jr.; Anderson, C. F.; Lohman, T. M. Q. ReV. Biophys. 1978,
11, 103-178.
(44) Chaires, J. B. Anticancer Drug Des. 1996, 11, 569-580.
(45) Record, M. T., Jr. Biopolymers 1967, 5, 993-1008.
(46) Record, M. T., Jr.; Mazur, S. J.; Melancon, P.; Roe, J. H.; Shaner, S. L.;
Unger, L. Annu. ReV. Biochem. 1981, 50, 997-1024.
(47) Manning, G. S. Q. ReV. Biophys. 1978, 11, 179-246.
(48) Haq, I.; Ladbury, J. E.; Chowdhry, B. Z.; Jenkins, T. C.; Chaires, J. B. J.
Mol. Biol. 1997, 271, 244-257.
(49) Haq, I.; Jenkins, T. C.; Chowdhry, B. Z.; Ren, J.; Chaires, J. B. Methods
Enzymol. 2000, 323, 373-405.
(50) Ren, J.; Jenkins, T. C.; Chaires, J. B. Biochemistry 2000, 39, 8439-8447.
(51) Karelson, M. Molecular Descriptors in QSAR/QSPR; Wiley-Interscience:
New York, 2000.
MHQ-14 and OZ-153 show dramatically reduced triplex
affinity relative to LS-8. The reasons are different for the two
compounds, but are simple in each case. MHQ-14 was designed
9
J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003 7279

A R T I C L E S
Chaires et al.
energetic cost of ligand desolvation is unfavorable for binding.
Overall, the derived QSAR offers a powerful and quantitative
description of the origin of triplex affinity of the naphthylquino-
lines studies here. Increases in the solvent accessible surface
area of the pendant aminoalkyl chain can enhance binding, but
adding substituents to the chain with electron affinity or
hydrogen bond acceptor functionality will oppose binding. The
derived QSAR emphasizes different properties than were
discussed above to explain the triplex affinity of MHQ-12, OZ-
85H, MHQ-14, and OZ-153, because we lacked appropriate
descriptors to account for chain flexibility, threading potential,
of both parallel and antiparallel DNA triple helices.19 Foot-
printing was used to investigate the effect of LS-8 on the
stringency of parallel triple helix formation of a series oligode-
oxynucleotides containing all combinations of bases at a
particular position.23 Footprinting was again used to study the
effect of LS-8 on the formation of intermolecular triplexes of
mixed sequence.20 Footprinting and UV melting studies were
used to investigate the effect of tethered LS-8 on both inter-
and intramolecular triplex stabilization.21 Recently, a novel,
high-throughput molecular beacon assay was used to investigate
the effect of LS-8 on the thermal denaturation of an intramo-
lecular triplex of mixed sequence.58 Collectively, these studies
make LS-8 one of the more extensively studied triplex binders,
and indicate that our studies with poly dA:(poly dT))2 are
reflective of, and consistent with, its selective binding to a wide
variety of triplex structures.
and steric hindrance. These issues are currently under study by
52
use of more sophisticated 4D QSAR
methods that will
incorporate more detailed structural and molecular dynamics
information (Bishop, G. R.; Senese, C. L.; Chaires, J. B.;
Hopfinger, A. J., in preparation).
The derived QSAR offers a powerful means of predicting
the triplex affinity of newly designed naphthylquinoline com-
pounds. The properties of newly designed compounds can be
easily computed, and substitution of these properties into eq 1
allows prediction of their triplex binding affinity. That such an
approach works is shown conclusively in the Supporting
Information, Figure S2, where systematic studies were done in
which each of the naphthylquinolines studied here were omitted
in turn, and the remaining compounds used to construct a QSAR.
In all cases, the QSAR could be used to accurately predict the
known triplex binding affinity of the omitted compound.
We note that the triplex studies described here were limited
to the poly dA:(poly dT)2 triple helix. This is a parallel triplex
that may not be representative of triplex structures in general.
It is, however, widely used in studies of ligand-triplex interac-
tions, and is one of the best characterized triple helix structures
One reviewer expressed reservations about the utility of the
specificity sum (SS) as a metric. The concern raised was that
although the limiting value of 1 and 13 are clear indications of
absolute selectivity or the complete lack of selectivity, respec-
tively, intermediate values are ambiguous. A particular inter-
mediate value of SS could arise from a number of binding
distributions to the different structures. For example, a value
of SS ) 2.0 could arise from equal binding to 2 of the 13
structures, or from maximal binding to one structure, and binding
to 10 others with a relative binding of 0.1. An actual example
may be seen in Figure 7 for the data for ethidium (SS ) 5.0)
and the 2,6-disubstituted anthraquinone (SS)5.08). The values
for SS are similar for these compounds, but the binding
distribution is clearly different. We recognize that SS is a simple
metric, with limited information about the detailed binding
distribution, and that more sophisticated chemometric tools are
needed for a more detailed description of the exact binding
distributions. In our experience with over 126 compounds from
diverse chemical classes that have been studied by this first-
generation competition dialysis assay with 13 structures, we
found an average value of SS ) 4.5 ( 2.0 (Chaires, J. B.; Ren,
J., in preparation). This average reflects that few compounds
exhibit absolute structural selectivity. With reference to this
average value, however, SS provides a clear, quantitative
measure of structural preferences that is operationally useful
for identifying interesting binding behavior. We freely acknowl-
edge the simplicity and limitations of SS, but have found it
useful and an appropriate point of departure for more sophis-
ticated analyses.
53
54,55
from both a structural and thermodynamic
perspective.
Although it is certainly desirable to investigate the interaction
of the naphthylquinolines studied here with additional types of
triplex structures of different strand polarity and that include
guanine and cytosine, such is not possible with this version of
the competition dialysis assay. Antiparallel triplex structures and
triplexes containing G and C require different solution conditions
(pH, divalent cations) than employed here. Although alternate
triplex structures can most certainly be used in competition
dialysis studies,56 different buffer conditions are needed that
require complete reformulation of the assay, with rigorous
verification that all structures included are stable under the
particular ionic conditions used. We intend to continue to
develop the competition dialysis assay to include more structures
of interest along with a range of ionic conditions. Studies of
the interaction of LS-8 with a variety of triplex structures using
a several experimental techniques were previously reported
19,20,23,57,58. These studies lend confidence that our studies with
poly dA:(poly dT)2 are applicable to triplexes in general. DNase
I footprinting was used to show that LS-8 facilitates formation
The studies presented here suggest new design principles for
the development of triplex selective agents. In addition to the
three design principles used in the initial development of LS-8
17
(positive charge, aromatic surface complementary to base
triplet shape, ring flexibility to accommodate propeller twist),
at least two more may be added. First, pendant chains that
occupy one or more of the triplex grooves are important for
both selectivity and affinity. Second, these chains should be
flexible and able to adapt to the triplex groove structure to
optimize interactions within the groove. The studies described
(52) Duca, J. S.; Hopfinger, A. J. J. Chem. Inf. Comput. Sci. 2001, 41, 1367-
1387.
(53) Chandrasekaran, R.; Giacometti, A.; Arnott, S. J. Biomol. Struct. Dyn. 2000,
17, 1011-1022.
here reinforce the importance of groove interactions previously
(54) Ross, P. D.; Howard, F. B. Biopolymers 2003, 68, 210-222.
(55) Plum, G. E.; Pilch, D. S.; Singleton, S. F.; Breslauer, K. J. Annu. ReV.
Biophys. Biomol. Struct. 1995, 24, 319-350.
39,42
emphasized by Jenkins and colleagues,
and introduce the
(56) Alberti, P.; Hoarau, M.; Guittat, L.; Takasugi, M.; Arimondo, P. B.; Lacroix,
L.; Mills, M.; Teulade-Fichou, M.-P.; Vigneron, J.-P.; Lehn, J.-P.; Maillet,
P.; Mergny, J.-L. Triplex- Versus quadruplex-specific ligands and telomerase
inhibition.; Demeunynck, M., Bailly, C. and Wilson, W. D., Ed.; Wiley-
VCH: Darmstadt, 2003; Vol. 1, pp 315-336.
(57) Keppler, M. D.; McKeen, C. M.; Zegrocka, O.; Strekowski, L.; Brown,
T.; Fox, K. R. Biochim. Biophys. Acta 1999, 1447, 137-145.
(58) Darby, R. A.; Sollogoub, M.; McKeen, C.; Brown, L.; Risitano, A.; Brown,
N.; Barton, C.; Brown, T.; Fox, K. R. Nucleic Acids Res. 2002, 30, e39.
9
7280 J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003

Triplex Selective 2-(2-Naphthyl)quinoline Compounds
A R T I C L E S
N-[2,2-Dimethyl-3-(dimethylamino)propyl]-2-(2-naphthyl)quino-
lin-4-amine (MHQ-11). This compound was obtained from N,N,2,2-
tetramethyl-1,3-propanediamine; yield 70%; mp 116-118 °C;1H NMR
(CDCl3) δ 1.15 (s, 6 H), 2.44 (s, 6 H), 2.48 (s, 2 H), 3.30 (s, 2 H), 6.88
(s, 1 H), 7.39 (t, J ) 8 Hz, 1 H), 7.49 (m, 2 H), 7.63 (t, J ) 8 Hz, 1
H), 7.73 (d, J ) 8 Hz, 1H), 7.87 (m, 1 H), 7.94 (d, J ) 8 Hz, 1 H),
7.99 (m, 1 H), 8.08 (d, J ) 8 Hz, 1H), 8.28 (d, J ) 8 Hz, 1H), 8.45
(bs, 1 H, exchangeable with D2O), 8.56 (s, 1H); 13C NMR (CDCl3) δ
25.7, 34.5, 48.5, 56.1, 71.6, 95.6, 118.7, 119.9, 124.1, 125.6, 126.0,
126.2, 126.8, 127.7, 128.1, 128.8, 129.0, 130.2, 133.6, 133.7, 138.8,
148.9, 151.5, 158.4; high-resolution EI-MS calcd for C26H29N3 (M+)
m/z 383.2361, observed m/z 383.2360. Anal. Calcd for C26H29N3‚
0.5H2O: C, 79.55; H, 7.70; N, 10.70. Found: C, 79.45, H, 7.77; N,
10.50.
new concept of the importance of flexibility of the substiuents
within the grooves.
Experimental Section
General. Abbreviations: AcOEt, ethyl acetate; BOP, benzotriazol-
yloxytris(dimethylamino)phosphonium hexafluorophosphate; DMF,
N,N-dimethylformamide; DMSO, dimethyl sulfoxide; Et3N, triethyl-
amine; Et2O, diethyl ether; EtOH, ethanol; MeOH, methanol; THF,
tetrahydrofuran.
Synthesis of N-[2-(dimethylamino)ethyl]-2-(2-naphthyl)quinolin-4-
amine 59 (LS-8), 4-chloro-2-(2-naphthyl)quinoline 32 (1), and 4-hydroxy-
2-(2-naphthyl)quinoline p-toluenesulfonate 32have been reported pre-
viously. 1H NMR and 13C NMR spectra were recorded at 400 and 100
MHz, respectively. Melting points (Pyrex capillary) are not corrected.
4-[2-(Dimethylamino)ethoxy]-2-(2-naphthyl)quinoline (MHQ-12).
A mixture of 2-chloroethyldimethylamine hydrochloride (0.22 g, 1.5
mmol), NaH (0.12 g, 5 mmol), and 4-hydroxy-2-(2-naphthyl)quinoline
p-toluenesulfonate (0.33 g, 0.75 mmol) in anhydrous DMF (6 mL) was
stirred and heated to 80 °C under a nitrogen atmosphere for 3 h. After
cooling to 23 °C, a precipitate of NaCl was filtered off, washed with
AcOEt (5 mL), and the organic solution was concentrated on a rotary
evaporator. Silica gel chromatography eluting with pentanes/Et2O/Et3N
(5:4:1) followed by crystallization from Et2O/pentanes gave 0.14 g
(55%) of MHQ-12: mp 52-54 °C; 1H NMR (CDCl3) δ 2.43 (s, 6 H),
2.95 (t, J ) 5.0 Hz, 2 H), 4.43 (t, J ) 5.0 Hz, 2 H), 7.32 (s, 1H), 7.51
(m, 3 H), 7.71 (t, J ) 8 Hz, 1 H), 7.88 (t, J ) 8 Hz, 1 H), 7.97 (m, 2
H), 8.14 (d, J ) 8 Hz, 1 H), 8.21 (d, J ) 8 Hz, 1 H), 8.30 (d, J ) 8
Hz, 1 H), 8.54 (s, 1 H); 13C NMR (CDCl3) _ 46.2, 58.0, 67.1, 98.7,
120.5, 121.8, 125.2, 125.4, 126.3, 126.6, 127.0, 127.7, 128.5, 128.8,
129.3, 130.0, 133.5, 133.9, 137.7, 149.4, 158.5, 162.1; high-resolution
EI-MS calcd for C23H22N2O (M+) m/z 342.1732, observed m/z
342.1723. Anal. Calcd for C23H22N2O: C, 80.67; H, 6.48; N, 8.18.
Found: C, 80.27, H, 6.63; N, 8.07.
General Procedure for MHQ-9, MHQ-11, MHQ-13, MHQ-14,
MHQ-15, MHQ-17, OZ-91, and OZ-97. A mixture of 4-chloroquino-
line 1 (0.29 g, 1 mmol), an amine (1.5 mL) and two drops of SnCl4
was stirred and heated to 130 °C under a nitrogen atmosphere for 4 h.
The resultant dark brown oil was allowed to cool to room temperature
and then quenched with water (20 mL). After extraction with AcOEt,
the organic layer was dried with anhydrous MgSO4 and concentrated
under a reduced pressure. The residue was purified by silica gel
chromatography using AcOEt/pentanes (4:1) as an eluent to yield a
substituted quinolin-4-amine as a pale yellow solid. The quinolinamine
was crystallized from MeOH or AcOEt. In several cases the product
was additionally purified by crystallization of its hydrobromide or
hydrochloride salt. Thus, a solution of a quinolin-4-amine in AcOEt
was stirred and treated dropwise with hydrobromic acid (47%) or
hydrochloric acid (36%). The precipitated salt was crystallized twice
from EtOH or AcOEt. The composition was determined by elemental
analysis.
N-(Trans-2-aminocyclohexyl)-2-(2-naphthyl)quinolin-4-amine Di-
hydrochloride (MHQ-13‚2HCl‚0.5H2O). This compound was obtained
1
from trans-1,2-cyclohexanediamine; yield 40%; mp 232-234 °C; H
NMR (DMSO-d6) _ 1.35 (m, 4 H), 1.76 (m, 2 H), 2.04 (m, 1H), 2.12
(m, 1 H), 3.06 (m, 1 H), 3.72 (m, 1 H), 5.25 (bs, 2 H, exchangeable
with D2O), 6.83 (bs, 1 H, exchangeable with D2O), 7.27 (s, 1H), 7.44
(t, J ) 8 Hz, 1 H), 7.57 (m, 2 H), 7.67 (t, J ) 8 Hz, 1 H), 7.91 (d, J
) 8 Hz, 1H), 7.98 (m, 1H), 8.03 (d, J ) 8 Hz, 1 H), 8.12 (m, 1 H),
8.34 (d, J ) 8 Hz, 1 H), 8.41 (d, J ) 8 Hz, 1H), 8.74 (s, 1 H); 13C
NMR (DMSO-d6) δ 24.3, 30.8, 32.2, 38.5, 53.5, 55.9, 95.4, 118.4,
121.9, 123.5, 125.1, 126.2, 126.4, 127.4, 127.7, 128.5, 129.1, 129.2,
132.9, 133.2, 133.8, 137.5, 148.5, 150.0, 156.6; high-resolution EI-
MS calcd for C25H25N2 (M+) m/z 367.2048, observed m/z 367.2044.
Anal. Calcd for C25H25N2 ‚2HCl‚0.5H2O: C, 66.80; H, 6.28; N, 9.35.
Found: C, 67.10, H, 5.83; N, 9.47.
N-[2-(t-Butoxycarbonylamino)ethyl]-2-(2-naphthyl)quinolin-4-
amine (MHQ-14). This compound was obtained from N-BOC-
ethylenediamine; yield 43%; mp 87-89 °C; 1H NMR (CDCl3) δ 1.48
(s, 9 H), 3.48 (t, J ) 5.0 Hz, 2 H), 3.59 (t, J ) 5.0 Hz, 2 H), 5.02 (bs,
1 H, exchangeable with D2O), 6.33 (bs, 1 H, exchangeable with D2O),
6.91 (s, 1 H), 7.41 (t, J ) 8 Hz, 1 H), 7.49 (m, 2 H), 7.64 (t, J ) 8 Hz,
1 H), 7.86 (m, 2 H), 7.95 (m, 2 H), 8.09 (d, J ) 8 Hz, 1 H), 8.27 (d,
J ) 8 Hz, 1 H), 8.54 (s, 1 H); 13C NMR (CDCl3) δ 28.4, 39.5, 46.0,
80.4, 96.1, 118.1, 120.0, 124.5, 125.4, 126.1, 126.3, 126.8, 127.7, 128.2,
128.7, 129.3, 130.0, 133.5, 133.7, 138.4, 148.7, 150.6, 158.1, 158.2;
high-resolution EI-MS calcd for C26H27N3O2 (M+) m/z 413.2103,
observed m/z 413.2104. Anal. Calcd for C26H27N3O2 ‚0.5H2O: C, 73.91;
H, 6.68; N, 9.94. Found: C, 74.41, H, 6.92; N, 9.79.
N-[3-(Morpholino)propyl]-2-(2-naphthyl)quinolin-4-amine (MHQ-
15). This compound was obtained from N-(3-aminopropyl)morpholine;
yield 59%; mp 143-145 °C; 1H NMR (CDCl3) δ 2.01 (m, 2 H), 2.59
(m, 4 H), 2.65 (m, 2 H), 3.56 (m, 2 H), 3.87 (m, 4 H), 6.97 (s, 1 H),
7.02 (bs, 1 H, exchangeable with D2O), 7.44 (t, J ) 8 Hz, 1 H), 7.50
(m, 2H), 7.66 (t, J ) 8 Hz, 1H), 7.89 (m, 2 H), 7.97 (m, 2 H), 8.10 (d,
J ) 8 Hz, 1H), 8.28 (d, J ) 8 Hz, 1H), 8.55 (s, 1 H); 13C NMR (CDCl3)
δ 23.8, 44.0, 54.1, 58.9, 67.0, 96.5, 118.3, 120.0, 124.2, 125.5, 126.1,
126.3, 126.9, 127.7, 128.2, 128.7, 129.2, 130.3, 133.6, 133.7, 138.5,
148.8, 150.9, 158.4; high-resolution EI-MS calcd for C26H27N3O (M+)
m/z 397.2154, observed m/z 397.2156. Anal. Calcd for C26H27N3O‚
0.5H2O: C, 76.81; H, 6.94; N, 10.34. Found: C, 76.77, H, 7.06; N,
10.27.
N-[3-(Imidazolo)propyl]-2-(2-naphthyl)quinolin-4-amine (MHQ-
9). This compound was obtained from 1-(3-aminopropyl)imidazole;
yield 80%; mp 169-171 °C; 1H NMR (CDCl3) δ 2.25 (m, 2 H), 3.41
(q, J ) 7 Hz, 2 H), 4.12 (t, J ) 7 Hz, 2 H), 5.10 (bs, 1H, exchangeable
with D2O), 6.91 (s, 1H), 6.94 (m, 1 H), 7.12 (m, 1 H), 7.42 (t, J ) 8
Hz, 1H), 7.50 (m, 3 H), 7.67 (t, J ) 8 Hz, 2 H), 7.87 (m, 1 H), 7.96
(m, 2 H), 8.13 (d, J ) 8 Hz, 1 H), 8.24 (d, J ) 8 Hz, 1 H), 8.50 (s, 1
H); 13C NMR (CDCl3) δ 30.2, 40.3, 44.6, 96.9, 118.1, 118.8, 119.1,
124.8, 125.3, 126.2, 126.5, 126.9, 127.7, 128.3, 128.8, 129.4, 130.2,
130.5, 133.5, 133.8, 137.2, 138.1, 148.8, 149.8, 158.2; high-resolution
EI-MS calcd for C25H22N4 (M+) m/z 378.1838, observed m/z 378.1844.
Anal. Calcd for C25H22N4: C, 79.30; H, 5.86; N, 14.81. Found: C,
79.54; H, 5.89; N, 14.78.
N-[2-(1-Methylpyrrolidin-2-yl)ethyl]-2-(2-naphthyl)quinolin-4-
amine Dihydrobromide (MHQ-17‚2HBr‚2.5H2O). This compound
was obtained from 2-(2-aminoethyl)-1-methylpyrrolidine; yield 40%;
1
mp 280-282 °C; H NMR (DMSO-d6) δ 1.74 (m, 1 H), 1.94 (m, 2
H), 2.05 (s, 3 H), 2.34 (m, 2 H), 2.70 (m, 2 H), 2.84 (m, 1 H), 3.43 (m,
1 H), 3.83 (m, 2 H), 7.28 (s, 1H), 7.74 (m, 3H), 8.00 (t, J ) 8 Hz, 1
H), 8.14 (m, 4 H), 8.24 (d, J ) 8 Hz, 1 H), 8.65 (d, J ) 8 Hz, 1 H),
8.72 (s, 1 H), 9.31 (bs, 1 H, exchangeable with D2O), 9.74 (bs, 1 H,
exchangeable with D2O); 13C NMR (DMSO-d6) δ 20.9, 28.2, 29.0, 30.6,
38.6, 55.0, 66.1, 97.6, 116.0, 120.4, 123.1, 124.9, 126.5, 127.3, 127.8,
128.3, 128.8, 129.1, 129.6, 132.2, 133.7, 134.1, 138.3, 141.9, 152.6,
(59) Wilson, W. D.; Zhao, M.; Patterson, S. E.; Wydra, R. L.; Janda, L.;
Strekowski, L. Med. Chem. Res. 1992, 2, 102-110.
9
J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003 7281

A R T I C L E S
Chaires et al.
155.3; high-resolution EI-MS calcd for C26H27N3 (M+) m/z 381.2205,
observed m/z 381.2211. Anal. Calcd for C26H 27N3‚2HBr‚2.5H2O: C,
53.07; H, 5.74; N, 7.14. Found: C, 53.24, H, 5.30; N, 6.99.
exchangeable with D2O); 13C NMR (DMSO-d6) δ 25.2, 25.4, 28.5, 32.3,
37.7, 42.5, 54.1, 60.5, 67.8, 97.4, 106.6, 116.1, 120.2, 120.3, 123.3,
125.3, 126.5, 126.8, 127.6, 129.0, 130.5, 133.7, 136.0, 138.2, 152.9,
155.1, 158.6 (two signals overlap). Anal. Calcd for C29H35N3O2‚3HBr‚
0.5H2O: C, 49.09; H, 5.55; N, 5.92. Found: C, 49.14, H, 5.57; N,
5.92.
N-(4-(Dimethylamino)butyl]-2-(2-naphthyl)quinolin-4-amine Di-
hydrobromide (OZ-115‚2HBr‚2H2O). The treatment of 1 with 4-hy-
droxybutylamine followed by workup as described above gave a
hydroxybutyl derivative 2 [yield 70%, mp 132-133 °C (from AcOEt)].
A mixture of 2 (0.2 g, 0.6 mmol), SOCl2 (2 mL), and benzene (5 mL)
was heated under reflux for 4 h and then concentrated on a rotary
evaporator. The residue was treated with aqueous NaOH (5%, 5 mL),
and the mixture was extracted with AcOEt (3 × 10 mL). The extract
was dried over anhydrous MgSO4 and concentrated to give an oily
residue of a chlorobutyl derivative 4. Compound 4 was purified by
crystallization of its hydrobromide 4‚HBr from EtOH/AcOEt (1:1)
[yield 0.17 g (64%), mp 192-195 °C].
A mixture of 4‚HBr (0.17 g, 0.4 mmol), dimethylamine hydrochlo-
ride (0.50 g, 61 mmol), Na2CO3 (0.75 g, 77 mmol), and DMF (10 mL)
was stirred and heated to 50 °C for 10 h. Concentration on a rotary
evaporator followed by treatment of the oily residue with MeOH (5
mL) and then addition of hydrobromic acid to the resultant solution
gave a crystalline salt of OZ-115. This salt was purified by three
crystallizations from EtOH; yield 0.11 g (61%); mp > 250 °C; 1H NMR
(DMSO-d6) δ 1.82 (m, 4 H), 2.78 (s, 6 H), 3.16 (m, 2 H), 3.76 (m, 2
H), 7.24 (s, 1 H), 7.72 (m, 3 H), 8.00 (t, J ) 8 Hz, 1 H), 8.14 (m, 4
H), 8.23 (d, J ) 8 Hz, 1 H), 8.64 (d, J ) 8 Hz, 1 H), 8.70 (s, 1 H),
9.29 (bs, exchangeable with D2O), 9.42 (bs, exchangeable with D2O);
13C NMR (DMSO-d6) δ 21.3, 24.7, 42.2, 42.3, 56.3, 97.4, 116.0, 120.3,
123.1, 124.8, 126.6, 127.4, 127.8, 128.4, 128.9, 129.1, 129.6, 132.4,
133.7, 134.1, 138.4, 152.4, 155.4 (two signals overlap). Anal. Calcd
for C24H27N3‚2HBr‚H2O: C, 52.91; H, 5.87; N, 7.41. Found: C, 52.56,
H, 5.76; N, 7.32.
N-[6-(Cyclopropylamino)hexyl]-2-(2-naphthyl)quinolin-4-am-
ine, Salt with Phosphoric Acid (OZ-124‚2H3PO4‚4H2O). The treat-
ment of 1 with 6-hydroxyhexylamine followed by workup as described
above gave a hydroxyhexyl derivative 3 [yield 59%, mp 122-131 °C
(from MeOH)]. A subsequent treatment of 3 with SOCl2 as described
above, gave a chlorohexyl derivative 5 as an oil. Compound 5 was
purified by crystallization of its hydrobromide 5‚HBr from EtOH (yield
50%, mp 190-195 °C).
N-[3-[N-(3-Aminopropyl)methylamino]propyl]-2-(2-naphthyl)-
quinolin-4-amine Trihydrobromide (OZ-91‚3HBr). This compound
was obtained from N,N-bis(3-aminopropyl)methylamine; yield 64%;
1
mp 232-235 °C; H NMR (DMSO-d6) δ 2.02 (m, 2 H), 2.23 (m, 2
H), 2.84 (s, 3 H), 2.92 (m, 2 H), 3.28 (m, 4 H), 3.83 (m, 2 H), 7.13 (s,
1H), 7.16 (m, 2 H), 7.30 (t, J ) 8 Hz, 1 H), 7.91 (bs, exchangeable
with D2O), 8.02 (t, J ) 8 Hz, 1 H), 8.11 (d, J ) 8 Hz, 1 H), 8.20 (m,
4 H), 8.71 (d, J ) 8 Hz, 1 H), 8.81 (s, 1 H), 9.44 (bs, exchangeable
with D2O), 9.82 (bs, exchangeable with D2O); 13C NMR (DMSO-d6)
δ 22.1, 23.1, 36.5, 40.7, 49.0, 52.8, 53.7, 116.4, 120.7, 123.4, 125.3,
127.4, 127.9, 128.3, 129.1, 129.5, 129.6, 129.7, 129.9, 132.9, 134.4,
134.6, 138.7, 153.1, 155.9. Anal. Calcd for C26H 30N4‚3HBr: C, 35.43;
H, 4.11; N, 6.36. Found: C, 35.43, H, 3.99; N, 6.12.
N-(3-Aminopropyl)-2-(2-naphthyl)quinolin-4-amine Dihydrobro-
mide (OZ-97‚2HBr‚1.5 H2O). This compound was obtained from 1,3-
diaminopropane; yield 61%; mp 280-284 °C; 1H NMR (DMSO-d6) δ
2.13 (m, 2 H), 3.03 (m, 2 H), 3.24 (bs, exchangeable with D2O), 3.85
(m, 2 H), 7.27 (s, 1 H), 7.70 (m, 3 H), 7.95 (bs, exchangeable with
D2O), 7.99 (t, J ) 8 Hz, 1 H), 8.07 (d, J ) 8 Hz, 1 H), 8.18 (m, 3 H),
8.24 (t, J ) 8 Hz, 1 H), 8.71 (d, J ) 8 Hz, 1 H), 8.79 (s, 1 H), 9.37;
(bs, exchangeable with D2O); 13C NMR (DMSO-d6) δ 25.9, 36.7, 40.3,
97.5, 116.2, 120.6, 123.3, 125.1, 126.9, 127.6, 128.8, 128.7, 129.1,
129.2, 129.4, 129.7, 132.6, 134.0, 134.3, 138.5, 152.7, 155.6 (two
signals overlap). Anal. Calcd for C22H 21N3‚2HBr‚1.5H2O: C, 51.17;
H, 5.09; N, 8.14. Found: C, 51.19, H, 4.85; N, 7.96.
N-(2-Dimethylamino)ethyl]-2-[6-[6-hydroxyhexyl)oxy]-2-naph-
thyl]quinolin-4-amine Trihydrobromide (OZ-85H‚3HBr‚0.5H2O).
Condensation of 2-(trifluoromethyl)aniline with 6-methoxy-2-naphthone
and cyclization of the resultant crude Schiff base 7 to quinoline 8 by
treatment with lithium [2-(dimethylamino)ethyl]amide in anhydrous
THF were conducted by using general procedures.33-35 A mixture of
crude compound 8 (0.7 g, 1.9 mmol), BBr3 (1 M in hexanes, 10 mL,
10 mmol), and toluene (10 mL) was stirred at 23 °C for 3 days and
then concentrated on a rotary evaporator. The residue was stirred and
treated dropwise with aqueous NaHCO3 until the mixture reached pH
7. Extraction of the mixture with AcOEt (5 × 10 mL) followed by
concentration of the extract to 10 mL and cooling gave crystalline
1
compound 9 [yield 0.6 g (89%); mp > 250 °C; H NMR (DMSO-d6)
δ 2.78 (s, 6 H), 2.65 (m, 2 H), 3.54 (m, 2 H), 6.95 (bs, 1 H,
exchangeable with D2O), 7.12 (s, 2 H), 7.16 (d, J ) 8 Hz, 1 H), 7.41
(t, J ) 8 Hz, 1 H), 7.63 (t, J ) 8 Hz, 1 H), 7.78 (d, J ) 8 Hz, 1 H),
7.88 (d, J ) 8 Hz, 1 H), 7.92 (d, J ) 8 Hz, 1 H), 8.16 (d, J ) 8 Hz,
1 H), 8.29 (d, J ) 8 Hz, 1 H), 8.60 (s, 1 H), 9.83 (s, 1 H, exchangeable
with D2O)].
A mixture of 9 (0.36 g, 1 mmol), NaH (27 mg, 1.1 mmol), 1 bromo-
6-(tert-butyldimethylsiloxy)hexane (0.33 g, 1.1 mmol), and anhydrous
DMF (6 mL) was stirred at 23 °C for 5 days, then quenched with
aqueous MeOH (2 mL), and concentrated on a rotary evaporator. Silica
gel chromatography of the residue eluting with AcOEt/MeOH (19:1)
gave a silyl derivative 10 [an oil; yield 0.46 g (80%); high-resolution
EI-MS calcd for C35H49N3O2Si (M+) m/z 571.3594, observed m/z
571.3598].
A solution of 10 (80 mg, 0.14 mmol) and Bu4N+F- (1 M in THF,
0.4 mL, 0.4 mmol) in THF (2 mL) was allowed to stand at 23 °C for
1 h and then was concentrated. A solution of the residue in AcOEt
was treated with hydrobromic acid and the resultant precipitate of the
trihydrobromide of OZ-85H was crystallized from EtOH/AcOEt (1:
1): yield 86 mg (86%); mp 224-226 °C; 1H NMR (DMSO-d6) δ 1.41
(m, 2 H), 1.48 (m, 4 H), 1.81 (m, 2 H), 2.94 (s, 6 H), 3.42 (m, 2 H),
3.55 (m, 2 H), 4.12 (m, 2H), 4.17 (m, 2 H), 7.31 (s, 1 H), 7.34 (d, J )
8 Hz, 1 H), 7.50 (s, 1 H), 7.77 (t, J ) 8 Hz, 1 H), 8.02 (t, J ) 8 Hz,
1 H), 8.10 (m, 3 H), 8.18 (d, J ) 7 Hz, 1 H), 8.58 (d, J ) 8 Hz, 1 H),
8.67 (s, 1 H), 9.15 (bs, 1 H, exchangeable with D2O), 9.66 (bs, 1 H,
A mixture of 5‚HBr (0.22 g, 0.56 mmol), cyclopropylamine (0.39
mL, 5.6 mmol), Na2CO3 (0.53 g, 5.0 mmol), and DMF (5 mL) was
stirred and heated to 40 °C for 12 h. After concentration on a rotary
evaporator, the residue of crude OZ-124 was purified by silica gel
chromatography eluting with Et2O/AcOEt (2:1). A solution of the oily
product in MeOH (10 mL) was stirred and treated with H3PO4 (85%,
1 mL). After cooling the resultant precipitate was filtered and
crystallized from MeOH; yield 0.19 g (49%) of OZ-124‚2H3PO4‚4H2O,
1
mp 150-158 °C; H NMR (DMSO-d6) δ 0.15 (m, 2 H), 0.29 (m, 2
H), 1.41 (m, 6 H), 1.77 (m, 2 H), 2.00 (m, 1 H), 2.56 (m, 2 H), 3.45
(m, 2 H), 7.13 (s, 1 H), 7.16 (bs, exchangeable with D2O), 7.42 (t, J )
8 Hz, 1 H), 7.56 (m, 2 H), 7.65 (t, J ) 8 Hz, 1 H), 7.90 (d, J ) 8 Hz,
1 H), 7.97 (m, 1 H), 8.03 (d, J ) 8 Hz, 1 H), 8.10 (m, 1 H), 8.27 (d,
J ) 8 Hz, 1 H), 8.41 (d, J ) 8 Hz, 1 H), 8.72 (s, 1 H), 13C NMR
(DMSO-d6) δ 5.8, 26.6, 27.9, 29.5, 30.1, 42.4, 48.9, 95.1, 118.1, 121.5,
123.6, 125.0, 126.0, 126.1, 126.4, 127.4, 127.7, 128.5, 129.0, 129.2,
132.9, 133.1, 137.5, 148.3, 150.8, 156.3 (two signals overlap); high-
resolution EI-MS calcd for C28H31N3 (M+) m/z 409.2518, observed
m/z 409.2493. Anal. Calcd for C28H31N3‚2H3PO4‚4H2O: C, 49.62; H,
6.71; N, 6.20. Found: C, 49.58, H, 6.41; N, 6.18.
N-[4-Hydroxy-3,5-bis(morpholinomethyl)phenyl]-2-(2-naphthyl)-
quinolin-4-amine Trihydrochloride (OZ-153‚3HCl‚2.5H2O). A mix-
ture of 1 (0.46 g, 1 mmol), 4-aminophenol (0.52 g, 4.8 mmol), and
phenol (0.5 g) was heated to 100 °C for 2 h under a nitrogen atmosphere
9
7282 J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003

Triplex Selective 2-(2-Naphthyl)quinoline Compounds
A R T I C L E S
and then cooled and treated with Et2O (25 mL). The solution was
applied to silica gel column (30 g). Elution with Et2O/Et3N (9:1) gave
compound 6 [0.46 g (79%), mp > 300 °C].
formed using the software package HyperChem 5.01 distributed by
Hypercube, Inc. Starting structures were drawn using the molecular
builder in HyperChem. Initial structure geometries were optimized
semiempirically from different starting conformations selected manually.
Subsequent rotational studies were carried out using the atom positions
of the lowest energy structure selected. Semiempirical single-point
energy and geometry optimization calculations were performed on LS8
and MHQ12 with each assigned a singlet spin multiplicity and a total
charge of +2. For the rotational study, each system was first optimized,
then the dihedral angle between the quinoline and the side chain was
frozen at different angles, while the rest of the system was allowed to
reoptimize. Increments of 45° were used starting from 0.0°. Following
optimization a single-point energy was calculated.
A solution of 6 (0.40 g, 1.1 mmol), morpholine (0.25 mL, 2.8 mmol),
and aqueous formaldehyde (13.3 M, 0.45 mL, 6 mmol) in EtOH (15
mL) was heated under reflux for 24 h. Concentration on a rotary
evaporator was followed by silica gel chromatography of the residue
eluting with Et2O/Et3N (9:1). The product OZ-153 was dissolved in
EtOH (10 mL) and the solution was treated with concentrated
hydrochloric acid (2 mL). The resultant precipitate was filtered and
crystallized from EtOH to give 0.30 g (40%) of OZ-153‚3HCl‚2.5H2O;
1
mp 224-226 °C; H NMR (DMSO-d6) δ 3.27 (m, 8 H), 3.50 (m, 8
H), 4.48 (s, 4 H), 7.14 (s, 1 H), 7.65 (m, 2 H), 7.79 (t, J ) 8 Hz, 1 H),
7.93 (s, 2 H), 8.03 (d, J ) 8 Hz, 1 H), 8.09 (s, exchangeable with
D2O), 8.12 (m, 4 H), 8.39 (d, J ) 8 Hz, 1 H), 8.82 (m, 1 H), 8.94 (d,
J ) 8 Hz, 1 H), 10.38 (bs, exchangeable with D2O), 11.37 (bs,
exchangeable with D2O); 13C NMR (DMSO-d6) δ 50.7, 53.8, 63.2, 99.4,
116.2, 119.7, 120.6, 123.5, 125.0, 126.9, 127.1, 127.7, 128.3, 128.6,
128.9, 129.5, 129.6, 132.2, 134.0, 139.0, 152.9, 154.4, 155.3 (three
signals overlap); Anal. Calcd for C35H36N4O3‚3HCl‚2.5H2O: C, 58.77;
H, 6.21; N, 7.83. Found: C, 59.03; H, 6.30; N, 7.70.
Competition Dialysis. Competition dialysis experiments were done
exactly as previously described.25,26 The nucleic acids used in the test
array are listed in Table 1. Competition dialysis studies were done in
BPES buffer, consisting of 6 mM Na2HPO4, 2mM NaH2PO4, 1mM
Na2EDTA, and 185 mM NaCl, pH 7.0.
Acknowledgment. This work was supported by Grant No.
CA35635 from the National Cancer Institute (J.B.C.). We thank
Dr. David Wilson (Georgia State University) for helpful
comments on the manuscript, Dr. David Magers (Mississippi
College) for advice on the semiempirical calculations, and Dr.
Anton Hopfinger (University of Illinois-Chicago) for advice on
QSAR construction.
Supporting Information Available: Details of the construc-
tion of a Quantitative Structure-Affinity Relationship (QSAR)
and results from binding and thermal denaturation studies. (11
pages with 5 tables and 4 figures). This material is available
free of charge via the Internet at http://pubs.acs.org.
Computational Methods. The energy barrier for side chain rotation
in LS8 and MHQ12 was determined by computational studies using
semiempirical methods. AM1 semiempirical calculations were per-
JA034181R
9
J. AM. CHEM. SOC. VOL. 125, NO. 24, 2003 7283