Molecular recognition of nucleotides by the guanidinium unit at the surface of aqueous micelles and bilayers. A comparison of microscopic and macroscopic interfaces

Article abstract of DOI:10.1021/ja960991+

Molecular recognition of the guanidinium/phosphate pair was investigated at microscopic interfaces of aqueous micelles and bilayers. Monoalkyl and dialkyl amphiphiles with guanidinium head groups were synthesized and dispersed in water to form micelles and bilayers having guanidinium groups at the aggregate surface. Binding of nucleotides such as AMP to these functionalized aggregates was evaluated by using an equilibrium dialysis (ultrafiltration) method. The observed binding constants of 10²-10⁴ M^-1 are much larger than the corresponding binding constant reported for a monomerically dispersed pair in the aqueous phase (1.4 M^-1) but are smaller than those found at the macroscopic air-water interface (10⁶-10⁷ M^-1). Therefore, the macroscopic interface promotes guanidinium-phosphate interaction more effectively than the microscopic interface. The present finding indicates that the microscopic interface can strengthen hydrogen bonding and/or electrostatic interaction even in the presence of water. Saturation binding phenomena were different between micelles and bilayers. All of the guanidinium groups in fluid micelles are effective for phosphate binding, but part of the guanidinium group in bilayers are not effective probably because of steric restriction.

Full text of DOI:10.1021/ja960991+

8524
J. Am. Chem. Soc. 1996, 118, 8524-8530
Molecular Recognition of Nucleotides by the Guanidinium Unit
at the Surface of Aqueous Micelles and Bilayers. A
Comparison of Microscopic and Macroscopic Interfaces
Mitsuhiko Onda, Kanami Yoshihara, Hiroshi Koyano, Katsuhiko Ariga, and
,
†
Toyoki Kunitake*
Contribution from the Supermolecules Project, JRDC, Kurume Research Park,
2
432 Aikawa, Kurume, Fukuoka 839, Japan
X
ReceiVed March 26, 1996
Abstract: Molecular recognition of the guanidinium/phosphate pair was investigated at microscopic interfaces of
aqueous micelles and bilayers. Monoalkyl and dialkyl amphiphiles with guanidinium head groups were synthesized
and dispersed in water to form micelles and bilayers having guanidinium groups at the aggregate surface. Binding
of nucleotides such as AMP to these functionalized aggregates was evaluated by using an equilibrium dialysis
2
4
-1
(
ultrafiltration) method. The observed binding constants of 10 -10 M are much larger than the corresponding
-
1
binding constant reported for a monomerically dispersed pair in the aqueous phase (1.4 M ) but are smaller than
6
7
-1
those found at the macroscopic air-water interface (10 -10 M ). Therefore, the macroscopic interface promotes
guanidinium-phosphate interaction more effectively than the microscopic interface. The present finding indicates
that the microscopic interface can strengthen hydrogen bonding and/or electrostatic interaction even in the presence
of water. Saturation binding phenomena were different between micelles and bilayers. All of the guanidinium
groups in fluid micelles are effective for phosphate binding, but part of the guanidinium group in bilayers are not
effective probably because of steric restriction.
Introduction
found the binding was energetically similar to binding in
methanol.
6
Molecular recognition in biological systems is achieved
through combined noncovalent interactions such as hydrogen
bonding, electrostatic interaction, and hydrophobic interaction.
These noncovalent interactions are also useful for designing
artificial host molecules. The hydrogen bonding interaction has
Molecular recognition in biological systems usually proceeds
at microscopic interfaces such as cell surface and protein surface.
We have found that complementary hydrogen bonding acts
efficiently for molecular recognition at the air-water interface.
For example, effective binding was observed for the comple-
mentary pairs of diaminotriazine monolayer and barbituric acid
1
-4
been frequently used for preparing specific host molecules.
However, the hydrogen-bond-mediated interactions are not
effective in aqueous systems, because bulk water forms strong
hydrogen bonds with host molecules. Therefore, these host
systems are not directly relevant to biological phenomena.
7
or nucleic acid bases (thymine, etc.), orotate monolayer and
8
nucleic acid bases (adenine etc.), and guanidinium monolayer
9
and phosphate (ATP, etc.). It appears that the presence of the
macroscopic interface is a key to accomplish effective molecular
recognition in contact with the aqueous phase.
5
Nowick et al. investigated specific binding between thymine
and adenine in aqueous micelles by the NMR titration method.
But the thymine functional group is buried in the hydrophobic
core of micelles, and effective hydrogen bonding is realized by
avoiding direct contact of water with the host/guest system.
Bonar-Law also investigated hydrogen bonded association of
a hydrophobic porphyrin-based receptor inside SDS micelle and
It is important to know to what extent the interface must be
macroscopic in order to realize the effective hydrogen bonding.
Aqueous micelles and bilayers provide microscopic interfaces
appropriate for this purpose. We selected the guanidinium/
phosphate pair for molecular recognition. Guanidinium groups
are found at the arginine side chain in proteins. They form
1
0
specific bonds with carboxylates and phosphates and play
*
To whom correspondence should be addressed.
Permanent address: Faculty of Engineering, Kyushu University,
11
important roles in keeping tertiary structures of proteins and
†
in providing anion recognition sites in some enzymes,1
2,13
Fukuoka 812, Japan.
X
Abstract published in AdVance ACS Abstracts, August 1, 1996.
(1) (a) Rebek, J., Jr.; Nemeth, D. J. Am. Chem. Soc. 1986, 108, 5637.
(6) Bonar-Law, R. P. J. Am. Chem. Soc. 1995, 117, 12397.
(7) Honda, Y.; Kurihara, K.; Kunitake, T. Chem. Lett. 1991, 681. (b)
Kurihara, K.; Ohto, K.; Honda, Y.; Kunitake, T. J. Am. Chem. Soc. 1991,
113, 5077.
(8) Kawahara, T.; Kurihara, K.; Kunitake, T. Chem. Lett. 1992, 1839.
(9) (a) Sasaki, D. Y.; Kurihara, K.; Kunitake, T. J. Am. Chem. Soc. 1991,
113, 9685. (b) Sasaki, D. Y.; Kurihara, K.; Kunitake, T. J. Am. Chem.
Soc. 1992, 114, 10994. (c) Sasaki, D. Y.; Yanagi, M.; Kurihara, K.;
Kunitake, T. Thin Solid Films 1992, 210/211, 776.
(10) Springs, B.; Haake, P. Bioorg. Chem. 1977, 6, 181.
(11) Anfinsen, C. B. Science 1973, 181, 223.
(12) (a) Weber, D. J.; Serpersu, E. H.; Shortle, D.; Mildvan, A. S.
Biochemistry 1990, 29, 8652. (b) Serpersu, E. H.; Shortle, D.; Mildvan,
A. S. Biochemistry 1987, 26, 1289.
(
b) Rebek, J., Jr. Acc. Chem. Res. 1990, 23, 399.
(2) (a) Chang, S.-K.; Van Engen, D.; Fan, E.; Hamilton, A. D. J. Am.
Chem. Soc. 1991, 113, 7640. (b) Hamilton, A. D.; Van Engen, D. J. Am.
Chem. Soc. 1987, 109, 5035.
(
3) (a) Aoyama, Y.; Tanaka, Y.; Toi, H.; Ogoshi, H. J. Am. Chem. Soc.
988, 110, 634. (b) Aoyama, Y.; Tanaka, Y.; Sugahara, S. J. Am. Chem.
Soc. 1989, 111, 5397.
4) (a) Lehn, J.-M. Pure Appl. Chem. 1994, 66, 1961. (b) Russell, K.
C.; Leize, E.; Van Dorsselaer, A.; Lehn, J.-M. Angew. Chem., Int. Ed. Engl.
995, 34, 209. (c) Marsh, A.; Nolen, E. G.; Gardinier, K. M.; Lehn, J.-M.
Tetrahedron Lett. 1994, 35, 397.
5) (a) Nowick, J. S.; Chen, J. S. J. Am. Chem. Soc. 1992, 114, 1107.
b) Nowick, J. S.; Chen, J. S.; Noronha, G. J. Am. Chem. Soc. 1993, 115,
1
(
1
(
(
7
636. (c) Nowick, J. S.; Cao, T.; Noronha, G. J. Am. Chem. Soc. 1994,
(13) Cotton, F. A.; Day, V. W.; Hazen, E. E., Jr.; Larsen, S.; Wong, S.
T. K. J. Am. Chem. Soc. 1974, 96, 4471.
1
16, 3285.
S0002-7863(96)00991-2 CCC: $12.00 © 1996 American Chemical Society

Guanidinium/Phosphate Pair Molecular Recognition
J. Am. Chem. Soc., Vol. 118, No. 36, 1996 8525
because they are protonated in the wide pH range and can form
hydrogen bonded ion pairs.
Chart 1. Structures of the Amphiphiles Used in This Study
14
The use of the guanidinium group as recognition sites has
been investigated in artificial systems. For example, Lehn and
co-workers reported a bicyclic guanidinium compound as a
15
selective anion receptor, and Anslyn et al. described a bis-
guanidinium compound that increased the rate of imidazole-
catalyzed mRNA hydrolysis by 20-fold in water.1 Hamilton
et al. synthesized an artificial enzyme that possessed two
guanidinium moieties and increased the lutidine-catalyzed
transterification of a p-nitrophenyl-activated RNA analog by
nearly 1000-fold in acetonitrile,17 and Schmidtchen investigated
the phosphate binding ability of a linear ditopic anion host with
6
-1
a guanidinium moiety. Its binding constant in water, 10.6 M ,
1
8
was ca. 7.5 times larger than that of the monomeric system.
We reported extremely effective binding of nucleotides with
9
guanidinium monolayers. In this study, we synthesized
monoalkyl and dialkyl amphiphiles with guanidinium groups
as hydrophilic head groups and examined their binding behavior
toward adenine nucleotides in aqueous micelles and bilayers.
Experimental Section
Materials. Adenosine-5′-monophosphate disodium salt (AMP),
adenosine-5′-diphosphate disodium salt (ADP), and adenosine-5′-
triphosphate disodium salt (ATP) were purchased from Oriental Yeast
Co. Bis(2-hydroxyethyl)iminotris(hydroxymethyl)methane (Bis-Tris)
and 2,6-dichlorophenolindophenol (DCPIP) were purchased from Wako
Pure Chemical Industries. Ultrafilter Molcut II LC (fractionating
molecular weight, 5000) was purchased from Nihon Millipore Kogyo.
Water used in this study was ion-exchanged by Millipore WQ 500
1
1
(
3
8
08.2 °C; H NMR (DMSO-d
6
) δ 0.88 (t, 3H, J ) 6.7 Hz, CH
), 1.51 (m, 2H, NHCH CH ), 2.29 (s, 3H, ArCH
), 7.19 (d, 2H, J ) 8.0 Hz, Ar), 7.48 (d, 2H, J )
3
), 1.25
m, 2H, 11 CH
.02 (m, 2H, NHCH
2
2
2
2
3
),
.2 Hz, Ar), 6.94 and 7.39 (br and m, respectively, 5H, guanidinium).
S: C, 61.79; H, 9.66; N, 9.83. Found:
41 3 3
Anal. Calcd for C22H N O
C, 61.71; H, 9.52; N, 9.85. The p-toluenesulfonate salt was converted
to chloride by an ion exchange on IRA-400 resin (Cl ). Chloride (1):
-
1
H NMR (CDCl
), 1.60 (m, 2H, NHCH
br and m, respectively, 5H, guanidinium).
-Octadecylguanidinium Chloride (3) was obtained by ion-
exchange of 1-octadecylguanidinium p-toluenesulfonate salt to improve
water dispersibility. 1-Octadecylguanidinium p-toluenesulfonate salt
was synthesized by operations similar to those for 1. 1-Octade-
3
) δ 0.88 (t, 3H, J ) 6.6 Hz, CH
3
), 1.25 (m, 22H, 11
(Yamato), and its specific resistance was around 18 MΩ‚cm. Molecular
CH
(
2
2
CH ), 3.11 (m, 2H, NHCH
2
2
), 6.46 and 7.87
structures of amphiphiles used in this study are shown in Chart 1.
Hexadecyltrimethylammonium bromide (2) was purchased from Tokyo
Chemical Industries, Ltd. Dialkyl nonionic amphiphile 4 was provided
1
19
by a research group at Kyushu University. Guanidinium amphiphiles
, 3, 5, and 6 were synthesized as described below. L-Argenine ethyl
1
ester dihydrochloride (Kokusan Chemical Works) for the preparation
of 6 was used as supplied.
cylguanidinium p-toluenesulfonate salt (19% yield): mp 117-118 °C;
1
H NMR (DMSO-d
6
) δ 0.85 (t, 3H, J ) 6.6 Hz, CH
), 1.44 (m, 2H, NHCH CH ), 2.29 (s, 3H, ArCH
H, NHCH ), 7.11 (d, 2H, J ) 8.0 Hz, Ar), 7.47 (d, 2H, J ) 8.0 Hz,
3
), 1.24 (m, 30H,
1
-Tetradecylguanidinium Chloride (1). Methylisothiourea sulfate
1.65 g, 11.9 mmol) was dispersed in methanol (30 mL). Sodium (0.303
g, 10.2 mmol) dissolved in methanol (30 mL) was added slowly at 0
C, and the resulting mixture was stirred for 1 h at 0 °C. Tetradecy-
1
2
5 CH
2
2
2
3
), 3.07 (m,
(
2
Ar), 6.80 and 7.37 (br and m, respectively, 5H, guanidinium). Anal.
Calcd for C26 S: C, 64.55; H, 10.21; N, 8.69. Found: C, 64.53;
H, 10.19; N, 8.52. Chloride (3): H NMR (DMSO-d
J ) 6.4 Hz, CH ), 1.23 (m, 30H, 15 CH ), 1.43 (m, 2H, NHCH
.06 (m, 2H, NHCH ), 7.21 and 7.43 (br and m, respectively, 5H,
guanidinium).
Dihexadecylamine. A mixture of hexadecylamine (61.9 g, 256
mmol), hexadecyl bromide (65.0 g, 213 mmol), Na CO (56.6 g, 534
°
49 3 3
H N O
lamine (2.10 g, 9.85 mmol) dissolved in methanol (20 mL) was added
slowly to the dispersion at room temperature. The whole mixture was
stirred at room temperature for 2 h and then at 40 °C for 42.5 h.
1
6
) δ 0.84 (t, 3H,
CH ),
3
2
2
2
3
2
p-Toluenesulfonic acid monohydrate (p-TsOH‚H
2
O) (3.80 g, 20.0
mmol) was added to the solution. The solution was heated to 60-70
°
°
C, and the precipitate was filtered off. The filtrate was kept at -20
C to give colorless crystals (1.09 g, mp, 98-101 °C). Recrystallization
2
3
mmol), and ethanol (400 mL) was refluxed for 116 h. The solvent
was removed by evaporation to dryness, and the residual solid was
suspended in CHCl
from acetonitrile (100 mL) gave 1-tetradecylguanidinium p-toluene-
sulfonate salt as fine colorless crystals (0.887 g, 21%): mp 106.9-
3
. This was washed with aqueous Na
2 3
CO and water
(
(
14) Hall, N. F.; Sprinkle, M. R. J. Am. Chem. Soc. 1932, 54, 3469.
15) (a) Dietrich, B.; Fyles, D. L.; Fyles, T. M.; Lehn, J.-M. HelV. Chim.
(×2) followed by drying over Na
2
SO . The solvent was removed by
evaporation, and the residue was recrystallized three times from hexane
4
Acta 1979, 62, 2763. (b) Dietrich, B.; Fyles, T. M.; Lehn, J.-M.; Pease, L.
G.; Fyles, D. L. J. Chem. Soc., Chem. Commun. 1978, 934.
to give dihexadecylamine (27.0 g, 27%) as a colorless powder: mp
1
6
1
4
5.7-65.9 °C; H NMR (CDCl
3
) δ 0.88 (t, 6H, J ) 6.7 Hz, 2 CH
), 1.4-1.6 (m, 4H, 2 CH CH N), 2.64 (t,
N); 13C NMR (CDCl
, 75 MHz) δ 14.07, 22.67,
3
),
(16) (a) Kneeland, D. M.; Ariga, K.; Lynch, V. M.; Huang, C.-Y.; Anslyn,
.2-1.4 (m, 52H, 26 CH
H, J ) 7.5 Hz, 2 CH
2
2
2
E. V. J. Am. Chem. Soc. 1993, 115, 10042. (b) Ariga, K.; Anslyn, E. V.
J. Org. Chem. 1992, 57, 417. (c) Smith, J.; Ariga, K.; Anslyn, E. V. J.
Am. Chem. Soc. 1993, 115, 362. (d) Anslyn, E. V.; Smith, J.; Kneeland,
D. M.; Ariga, K.; Chen, F.-Y. Supramolecular Chem. 1993, 1, 201.
2
3
27.41, 29.34, 29.58, 29.60, 29.68 (br), 30.16, 31.91, 50.13; HRMS (EI)
m/z calcd for C32 67N 465.5273, found 465.5272. Anal. Calcd for
O: C, 80.93; H, 14.43; N, 2.95. Found: C, 81.02; H,
4.26; N, 2.97.
-(Dihexadecylcarbamoyl)butylammonium Chloride. A solution
of dihexadecylamine (7.99 g, 17.1 mmol), 5-bromovaleryl chloride (3.50
mL, 24.7 mmol), and triethylamine (3.60 mL, 25.9 mmol) in CH Cl
H
(
17) Jubian, V.; Dixon, R. P.; Hamilton, A. D. J. Am. Chem. Soc. 1992,
1
C
32
H
2 2
67N‚ / H
1
14, 1120.
1
(
(
18) Schmidtchen, F. P. Tetrahedron Lett. 1989, 34, 4493.
19) (a) Okahata, Y.; Ando, R.; Kunitake, T. Ber. Bunsenges. Phys. Chem.
4
1
981, 85, 789. (b) Okahata, Y.; Tanamachi, S.; Nagai, M.; Kunitake, T. J.
Colloid Interface Sci. 1981, 82, 401.
2
2

8526 J. Am. Chem. Soc., Vol. 118, No. 36, 1996
Onda et al.
(
300 mL) was stirred at room temperature for 2.5 h followed by washing
with aqueous Na CO
(×2) and water (×5). After being dried over
Na SO , the solvent was removed by evaporation to give a yellow oil.
7.7 Hz, CH
2.25. Found: C, 77.20; H, 12.74; N, 2.27.
N,N-Dioctadecyl-N′-ethylarginylsuccinamide HCl Salt (6). N,N-
Dioctadecylsuccinamic acid (0.50 g, 0.80 mmol) was dissolved in CH
Cl (150 mL), and diethyl phosphorocyanidate (DEPC) (0.150 mL, 0.99
2
N). Anal. Calcd for C40
H
79NO
3
: C, 77.23; H, 12.80; N,
2
3
2
4
Hexane (300 mL) was added, and the insoluble material was removed
by filtration. The filtrate was evaporated to give a yellow oil, which
2
-
2
was chromatographed on SiO
2
(3:1 hexane/ethyl acetate) to give crude
mmol) was added to the solution at 0 °C. After 15 min arginine ethyl
ester dihydrochloride (0.265 g, 0.97 mmol) and triethylamine (0.60 mL,
4.30 mmol) dissolved in DMF (50 mL) was added slowly. The whole
mixture was stirred at room temperature for 120 h. Solvent was
removed in vacuo, and water was added to the residue. The supernatant
was removed by decantation. The solid residue was dispersed in
5
-bromo-N,N-di(hexadecyl)pentanoylamide (7.68 g, 71%) as an oil:
1
TLC R
f
0.73 (2:1 hexane/ethyl acetate); H NMR (CDCl
), 1.2-1.4 (m, 52H, 26 CH ), 1.4-1.6 (m, 4H,
), 1.7-2.0 (m, 4H, 2 CH ), 2.31 (t, 2H, J ) 7.1 Hz, CH C-
), 3.28 (t, 2H, J ) 7.6 Hz, NCH ),
Br). A mixture of this product and
potassium phthalimide (4.56 g, 24.6 mmol) and N,N-dimethylformamide
DMF) (150 mL) was stirred at 80-90 °C for 21 h. DMF was removed
under vacuum, and the residue was dissolved in CH Cl (500 mL).
This solution was washed with aqueous Na CO
(×2) and water (×2)
and dried over Na SO . The solvent was removed by evaporation, and
the residue was chromatographed on SiO (3:1 hexane/ethyl acetate)
to give crude 5-phthalimido-N,N-di(hexadecyl)pentanoylamide (6.04
g, 71%) as a colorless wax: TLC R 0.57 (2:1 hexane/ethyl acetate);
) δ 0.88 (t, 6H, J ) 6.6 Hz, 2 CH ), 1.2-1.4 (m,
), 1.4-1.6 (m, 4H, 2 NCH CH ), 1.6-1.8 (m, 4H, 2 CH ),
.34 (t, 2H, J ) 7.0 Hz, CH C(O)), 3.19 (t, 2H, J ) 7.7 Hz, NCH ),
.27 (t, 2H, J ) 7.6 Hz, NCH ), 3.71 (t, 2H, J ) 6.8 Hz, CH N), 7.6-
.9 (m, 4H, Ar). A portion of this product (1.0 g) was refluxed with
3
) δ 0.88 (t,
6
2
H, J ) 6.6 Hz, 2 CH
NCH CH
3
2
2
2
2
2
(
3
O)), 3.19 (t, 2H, J ) 7.7 Hz, NCH
2
2
.42 (t, 2H, J ) 6.6 Hz, CH
2
acetonitrile (100 mL). p-TsOH‚H
2
O (3.03 g, 15.9 mmol) was added
to the dispersion, and the solid was dissolved with sonication. Water
(50 mL) was slowly added to the resulting clear solution. The
precipitates (the crude p-toluenesulfonate) were collected and dissolved
in methanol. Concentrated HCl was added to the solution and the
resulting precipitates were collected by filtration and recrystallized from
(
2
2
2
3
2
4
2
acetonitrile/ethanol (5:1) (7 mL) to give 6 as colorless crystals (0.156
1
g, 23%): mp 168-169 °C; H NMR (CDCl
3
), δ 0.88 (t, 6H, J ) 6.7
and COOCH CH ), 1.4-1.7 (m,
NH), 2.59 (br, 2H, CH CO), 2.70
CO), 3.23 (br, 6H, 2 NCH and CH NHC(NH )NH ), 4.17
(q, 2H, J ) 7.1 Hz, COOCH CH ), 4.42 (br, 1H, NHCHCO), 7.26 (br,
1H, CONH), 6.91 and 7.89 (br each, 5H, guanidinium). Anal. Calcd
for C48 Cl: C, 68.41; H, 11.48; N, 8.31. Found: C, 68.36; H,
11.38; N, 8.06.
f
Hz, 2 CH
8H, 2 NCH
(br, 2H, CH
3
), 1.25 (m, 63H, 30 CH
2
2
3
1
H NMR (CDCl
3
3
2
CH and CHCH CH CH
2
2
2
2
2
5
2
3
7
2H, 26 CH
2
2
2
2
2
2
2
2
2
2
2
2
3
2
2
96 5 4
H N O
hydrazine monohydrate (288 mg, 5.69 mmol) and 95% aqueous ethanol
30 mL) for 16 h. Concentrated hydrochloric acid was added, and the
(
Characterization of the Aggregates. The critical micellar con-
centration (CMC) of monoalkyl amphiphiles 1 and 2 were determined
by a dye method using DCPIP.20 DCPIP shows partial deprotonation
in water at neutral pH. The presence of cationic micelles shifts its
equilibrium to the deprotonated species (blue) because of incorporation
of the dye in micelles and subsequent stabilization of the anionic species.
The CMC values of the amphiphiles were measured as follows. Mixed
solutions of amphiphile 1 or 2 and DCPIP were prepared (1/DCPIP )
0-1/0.02 mM and 2/DCPIP ) 0-5/0.02 mM), and the observed λmax
value was monitored as a function of amphiphile concentration. The
resulting insoluble material was removed by filtration through a Celite
pad. The filtrate was concentrated, and the residue was recrystallized
twice from ethyl acetate to give the title compound (690 mg, 80%) as
1
a colorless powder: mp 71.7-72.7 °C; H NMR (CDCl
3
) δ 0.88 (t,
), 1.4-1.6 (m, 4H,
), 2.37 (br s, 2H, CH C(O)), 3.06 (br
), 3.1-3.3 (m, 4H, 2 CH N), 8.50 (br s, 3H, NH ). Anal.
OCl: C, 73.89; H, 12.90; N, 4.66. Found: C, 73.95;
6
2
H, J ) 6.7 Hz, 2 CH
CH ), 1.7-1.9 (m, 4H, 2 CH
NH
Calcd for C37
H, 12.93; N, 4.63.
-(4-(Dihexadecylcarbamoyl)butyl)guanidinium p-Toluenesulfonate
5). Methylisothiourea sulfate (417 mg, 3.0 mmol) was dissolved in
.44 N NaOH (8 mL) at room temperature. The resulting clear solution
3
), 1.1-1.4 (m, 56H, 28 CH
2
2
2
2
s, 2H, CH
2
3
2
3
H
77
N
2
1
λmax value of DCPIP was ca. 520 nm below CMC and shifted to ca.
(
0
600 nm at sufficiently high concentrations of the amphiphile. The CMC
values were determined from the minimum amphiphile concentration
where the λmax showed a shift to a long wavelength. UV absorption
measurements were performed in water or in 0.5 mM Bis-Tris buffer
(pH 7) at 25 °C with a JASCO UV/VIS/NIR spectrometer equipped
with a JASCO EHC-441 temperature controller. The measurements
were carried out.
was stirred for 10 min followed by addition of 4-(dihexadecylcarbam-
oyl)butylammonium chloride (300 mg, 0.50 mmol) in ethanol (8 mL).
The whole mixture was stirred at 70 °C for 16 h. p-TsOH‚H
g, 7.0 mmol) was added to the mixture followed by further stirring at
room temperature for 6 h. This mixture was extracted with CHCl
The combined extracts were dried over Na SO and concentrated under
reduced pressure. To the residue was added CH CN/THF (1:1) with
2
O (1.34
3
.
Differential scanning calorimetry (DSC) measurements were per-
formed with a Seiko SSC/5200H calorimeter equipped with DSC 120
module. Aqueous dispersions of 3/4 (0.5/1.0 mM), 5 (1.0 mM), and 6
(1.0 mM) were prepared, subjected to ultrasonic treatment with a probe-
type sonicator (Bransonic Sonifier Model 250), and sealed (60 µL) in
2
4
3
heating, and the insoluble material was removed by filtration. The
filtrate was concentrated, and the residue was subjected to column
chromatography on alumina (2:1 CH
3
CN/CH
colorless oil (152 mg, 39%): H NMR (CDCl ) δ 0.88 (t, 6H, J ) 6.7
Hz, 2 CH ), 1.4-1.7 (m, 8H, 4 CH ), 2.31
C(O)), 2.35 (s, 3H, ArCH ), 3.1-3.3 (m, 6H, 2 CH
NH), 7.00 (br s, 4H, 2 NH ), 7.17 (d, 2H, J ) 8.0 Hz, Ar),
3
OH) to give 5 as a
1
-1
a Ag sample pan, and scaning was carried out at a rate of 1 °C‚min
from 5 to 70 °C.
3
3
), 1.1-1.4 (m, 52H, 26 CH
br s, 2H, CH
2
2
(
2
3
2
N
Transmission electron microscopy (TEM) observation of aqueous
dispersions of the amphiphiles was performed as follows. Aqueous
dispersion of the amphiphile (0.5-1.0 mM) were negatively stained
by mixing with equal volumes of saturated aqueous uranyl acetate. The
mixture was kept at room temperature for 1 h. Aliquots of the solution
were placed on a carbon-coated copper mesh and dried in vacuo. The
sample was observed with a HITACHI H-600S electron microscope at
and CH
2
2
7
.62 (br s, 1H, NH), 7.72 (d, 2H, J ) 8.0 Hz, Ar). Anal. Calcd for
S: C, 69.36; H, 11.12; N, 7.19. Found: C, 69.18; H, 11.08;
45 86 4 4
C H N O
N, 7.10.
Dioctadecylamine was prepared by the procedures similar to those
of dihexadecylamine in 37% yield: mp 72.0-72.5 °C; 1H NMR
1
2 000-60 000 magnification.
(
1
CDCl
.4-1.6 (m, 4H, 2 CH
Anal. Calcd for C36 75N: C, 72.83; H, 14.48; N, 2.68. Found: C,
2.70; H, 14.41; N, 2.67.
N,N-Dioctadecylsuccinamic Acid. A solution of dioctadecylamine
2.47 g, 4.74 mmol) and succinic anhydride (956 mg, 9.56 mmol) in
THF (100 mL) was stirred at room temperature for 17 h. Solvent was
removed by evaporation, and the residue was dissolved in CH Cl . This
was washed with 1 N aqueous HCl (×2) and dried over Na SO
Evaporation of the solvent and recrystallization of the residue from
CH CN gave N,N-dioctadecylsuccinamic acid (2.26 g, 77%): mp 65.8-
6.5 °C; H NMR (CDCl
m, 60H, 30 CH
3
) δ 0.88 (t, 6H, J ) 6.7 Hz, 2 CH
3
), 1.2-1.4 (m, 60H, 30 CH
2
),
N).
Ultrafiltration and Determination of the Component Concentra-
2
CH N), 2.61 (t, 4H, J ) 7.4 Hz, 2 CH
2
2
tion. Binding of substrate molecules to aqueous aggregates was
H
2
1
evaluated by the equilibrium dialysis method (ultrafiltration method)
which can separate unbound substrates from bound substrates (Figure
). All of the solutions used here were prepared with 0.5 mM Bis-
8
1
(
Tris buffer (pH 7). Amphiphiles and nucleotides were mixed at the
following concentrations: 1/AMP ) 0.5/0-1.0 mM, 1/ADP ) 0.5/
2
2
0
-1.0 mM, 1/ATP ) 0.5/0-1.0 mM, 2/AMP ) 10.0/0-10.0 mM,
2
4
.
3/4/AMP ) 0.5/1.0/0-1.0 mM, 5/AMP ) 0.5/0-1.0 mM, and 6/AMP
)
0.5/0-1.0 mM. The mixture was put into the upper cup of ultrafilter
3
1
6
(
3
) δ 0.87 (t, 6H, J ) 6.6 Hz, 2 CH
), 1.4-1.6 (m, 4H, 2 CH CH N), 2.68 (s, 4H,
C(O)), 3.22 (t, 2H, J ) 7.8 Hz, CH N), 3.31 (t, 2H, J )
3
), 1.2-1.4
(
(
20) Orrin, M. L.; Harkins, W. D. J. Am. Chem. Soc. 1947, 69, 679.
21) Connors, K. A. Binding Constants Measurement of Molecular
2
2
2
C(O)CH
2
CH
2
2
Complex Stability; John Wiley & Sons: New York, 1987; p 310.

Guanidinium/Phosphate Pair Molecular Recognition
J. Am. Chem. Soc., Vol. 118, No. 36, 1996 8527
Molcut II LC (low adsorption type filter with fractionating molecular
weight of 5000). It was kept in a thermostated bath at 25 °C for 1 h
until the set up was thermally equilibrated. Pressure was then applied
to the upper cup to accelerate filtration. The nucleotide concentrations
in the filtrates were determined from absorbance of the adenine ring (ꢀ
3
)
13 × 10 at 260 nm). Concentrations of guanidinium amphiphiles
in the filtrates were determined from the absorption (535 nm) of the
dye mixture that was formed from their reaction with diacetyl and
-naphthol (the V-P reaction). An aqueous sample (2 mL) and 2 N
2
2
1
NaOH (1 mL) were mixed. 1-Propanol solution of 0.025 wt% diacetyl
and 5 wt% 1-naphthol (1 mL) was added, and the absorbance at 535
nm was measured after 30-min reaction.
The concentration of the bound substrates [S]bound was estimated by
subtracting the concentrations of unbound substrate [S]free from initial
substrate concentration [S]initial
.
Figure 1. Scheme of equilibrium dialysis (ultrafiltration).
[S]bound ) [S]initial - [S]free
(1)
The ratio of the bound substrate to the guanidinium group, y, was
estimated from eq 2
[
S]bound [S]initial - [S]free
y )
)
(2)
[H]
[H]
where [H] is the concentration of guanidinium group in aggregates that
is corrected with the amount of the molecularly dispersed amphiphile.
Binding of the substrates to guanidinium amphiphiles is represented
as eq 3 by assuming Langmuir type adsorption (eqs 3 and 4)23
R × [S]
free
y )
(3)
(4)
1
/K + [S]
free
[
S]free
1
1
)
+
× [S]free
Figure 2. Reciprocal plots of the Langmuir binding isotherm for
guanidinium micelles and nucleotides; 25 °C and pH 7.
y
R × K
R
where R and K represent saturation binding ratio of substrates and
binding constant, respectively. The R and K values can be obtained
from the intercept and the slope, respectively, of eq 4.
molecules passed through the filter, and the micellar 1 was
trapped on the filter.
In order to investigate the binding ability of amphiphile 1
toward nucleotides, ultrafiltration was applied to aqueous
mixtures of micelle 1 and substrates, AMP, ADP, and ATP.
As the adsorption of the substrates on the filter is negligible,
the substrate concentration in the filtrate is equal to the
concentration of unbound substrates in the original micelle/
substrate mixtures. These experimental data are plotted ac-
cording to eq 4 in Figure 2. All of the plots show linear
relationships indicating the occurrence of the Langmuir type
adsorption. Saturation binding ratio R and binding constant K
were obtained from intercept and slope, respectively, and
summarized in Table 1, together with determination coefficient
Results and Discussion
Recognition of Nucleotides by Guanidinium Micelles.
Critical micellar concentrations (CMC) of monoalkyl am-
phiphiles 1 and 2 were determined as described in the
Experimental Section. The CMC value of amphiphile 2 was
-
4
-4
9
(
.0 × 10 M in water and 4.3 × 10 M in Bis-Tris buffer
pH 7 and 25 °C). The former value is in good agreement with
24,25
the reported values.
The CMC value of amphiphile 1 was
-
4
-5
1
7
.3 × 10 M in water and 5.7 × 10 M in BisTris buffer (pH
and 25 °C). Monoalkyl guanidinium amphiphile 1 showed a
lower CMC value than monoalkyl quarternary ammonium
amphiphile 2. All the binding experiments described as below
were carried out at concentrations above these CMC.
2
R .
The combination of micellar 1 and AMP substrate gives K
3
-1
The ultrafiltration experiments for aqueous substrates in the
absence of the amphiphiles were performed prior to binding
study as a control. The substrates contained in the initial
solutions were detected in the filtrate within (0.5% error in all
cases. As the experimental error is estimated to be ap-
proximately 1.0%, nonspecific adsorption of the substrate on
the filter is negligible. Control experiment for filtration of the
guanidinium 1 was also carried out in the absence of the
substrate. The concentration of 1 detected in the filtrate by V-P
) 1.8 × 10 M and R ) 1.3 ( 0.3. This R value suggests
formation of an equimolar complex from guanidinium and
phosphate. The obtained binding constant is larger than the
value reported for the corresponding molecularly dispersed
-
1 10
2
system in aqueous media (1.4 M ) by a factor of 10 .
Both of complementary hydrogen bonding and electrostatic
attraction are believed to act as binding forces for this system.
In order to estimate the contribution of these interactions, we
conducted a control experiment by using micellar 2 that is not
capable of hydrogen bonding with AMP. The binding constant
-
5
reaction was 5.0 × 10 M and was almost the same as its
CMC value. Therefore, only molecularly dispersed amphiphile
2
-1
in this case was 8.1 × 10 M and was half the value obtained
for 1/AMP. Thus, both of complementary hydrogen bonding
and electrostatic attraction clearly contributed to the guani-
dinium/phosphate recognition on the micellar surface.
(
22) Micklus, M. J.; Stein, I. M. Anal. Biochem. 1973, 54, 545.
(23) Connors, K. A. Binding Constants Measurement of Molecular
Complex Stability; John Wiley & Sons: New York, 1987; p 59.
(
(
24) Klevens, H. B. J. Am. Oil. Chem. Soc. 1953, 30, 74.
25) Ralston, A. W.; Eggenberger, D. N. J. Am. Chem. Soc. 1947, 70,
Subsequently, binding of guanidinium micelle 1 with ADP,
ATP, and adenosine that have two, three, and zero phosphate
9
77.

8528 J. Am. Chem. Soc., Vol. 118, No. 36, 1996
Onda et al.
Table 1. Binding Parameters between Guanidinium Aggregates
Table 2. Characterization of Aqueous Bilayer
and Substrates
DSC
a
-1
Ra
R2
amphiphile substrate
K /M
1
bilayer
Tc/°C ∆H/kJ‚mol- ∆S/J‚mol-1‚K-1 TEM observation
Micelle
/4 (1:2) 46 and 53a
87b
168
3
4
5
6
28
55
fragment (multilayer)
multilamellae vesicle
string (multilayer)
string
3
1
AMP
ADP
ATP
adenosine
AMP
1.8 ( 0.1 × 10 1.3 ( 0.3 0.95
55
4
1
1
1
2
5.4 ( 0.3 × 10 0.4 ( 0.1 0.99
c
4
2.5 ( 0.7 × 10 0.4 ( 0.1 0.98
40
11
35
b
no binding
2
a
A broad peak with two tops was observed. b
T ) 50 °C was
c
8.1 ( 0.1 × 10 1.1 ( 0.2 0.97
assumed for the calculation. c Endothermic peak was not detected from
to 70 °C.
Bilayer
5
2
3
3
5
5
6
/4 (1:2)
AMP
AMP
AMP
AMP
AMP
1.8 ( 0.1 × 10 0.6 ( 0.2 0.98
2
/4 (1:2)
1.1 ( 0.1 × 10 1.2 ( 0.2 0.98 (sonicated)
3
1.0 ( 0.1 × 10 0.7 ( 0.1 0.99
crowding may be involved as additional factors. These factors
might represent unique characteristics of molecular recognition
at the surface of aggregates.
3
1.0 ( 0.1 × 10 0.5 ( 0.1 0.99 (sonicated)
2
3.3 ( 0.1 × 10 0.9 ( 0.2 0.98
c
Monolayer
Characterization of Aqueous Bilayers of Guanidinium
Amphiphiles. The preceding binding study reveals that the
interaction between guanidinium and phosphate is strengthened
at an aqueous microscopic interface of micelles. In order to
extend this finding to other aqueous aggregates, we investigated
binding behavior of nucleotides onto the surface of bilayer
membranes. Bilayers provide large and stable microscopic
interfaces where exchange of amphiphile molecules is much
slower than that in micelles.
6
7
7
AMP
ATP
3.2 × 10
1.0
0.3
7
1.7 × 10
d
Bulk Water
-
guanidinium H
2
PO
4
1.37
a
The ( values are standard deviations from the least squares
analysis. Detectable binding was not observed. c Data from ref 9a.
b
d
Data from ref 10.
groups, respectively, were performed (Table 1). In the case of
adenosine substrate, binding was not detectably observed. This
result endorses that the observed binding of nucleotides to
guanidinium micelles is based on the specific interaction
between guanidinium and phosphate groups. The adenine
moiety, by itself, is not effective for the binding. Results for
ADP and ATP summarized in Table 1 are different from those
of AMP in both of the binding constant and the saturation
binding ratio. The saturation binding ratio R is 0.4 ( 0.1 for
Double-chain amphiphiles 5 and 6 and single-chain am-
phiphile 7 are expected to produce stable bilayer membranes
when dispersed in water, on the basis of the past molecular
design of bilayer-forming compounds. Monoalkyl amphiphile
3
forms a fluid micelle by itself, but it can be incorporated in
a typical bilayer membrane of 4 without disrupting the bilayer
structure.
The formation of bilayer structures was confirmed by
differential scanning calorimetry (DSC) and by transmission
electron microscopy (TEM), prior to binding experiments.
Results of DSC measurements are summarized in Table 2.
Aqueous dispersion of 5 did not show any endothermic peak
in the temperature range examined (5-70 °C), while aqueous
dispersions of 3/4 (1:2) and 6 gave endothermic peaks due to
crystal-liquid crystal phase transition. The phase transition of
bilayer 5 exists probably below 5 °C, as judged from DSC data
1
/ADP and 1/ATP, being much smaller than that of 1/AMP.
Clearly, the two phosphate units in ADP are used in the
equimolar complexation. It is known that two of the three
2
6
phosphate groups in ATP are deprotonated at neutral pH.
Therefore, ATP molecule may bind to two guanidinium groups,
unless additional dissociation of the phosphate group proceeds
due to a pKa shift on the cationic micelle. Or, if all three
phosphate units are dissociated, the third phosphate may not
be used for binding due to the particular structural feature of
the micellar surface. Unfortunately, the observed R value of
1
9a
of other related amphiphiles.
In fact, the melting point of
solid 5 was lower than room temperature (see synthesis of 5).
0
.4 ( 0.1 is not consistent with either of the 1:2 or 1:3
Aqueous dispersion of 3/4 (1:2) showed overlapping two DSC
peaks at 46 and 53 °C. As this peak pattern was observed in
repeated scans, it cannot be ascribed to metastable phases. As
one of the peak temperature, 53 °C is almost the same as that
observed in aqueous dispersion of single component 4 (Tc )
guanidinium/phosphate complexation. In contrast, a guani-
dinium monolayer of 7 clearly formed a 1:3 complex with ATP
at the air-water interface.9 Globular micelles are known to
have fluid hydrophobic cores in the dynamic equilibrium with
dispersed monomeric surfactant species, whereas surface mono-
layers are composed of static molecular organizations when
compressed. This structural difference may be related to
different modes of binding at the surface of globular micelles
and surface monolayers.
a
-
1
55 °C with ∆H ) 55 kJ‚mol ), the mixed bilayer of 3/4 (1:2)
is presumed to possess a phase-separated domain of pure 4 (Tc
-1
) 53 °C). The total ∆H value for the 3/4 mixture, 28 kJ‚mol ,
-1
is significantly smaller than that (55 kJ‚mol ) of pure 4. When
we recalculated ∆H based on the mole of 4 alone, ∆H is still
-
1
Both ADP and ATP showed larger binding constants to the
guanidinium micelle than AMP. Apparently the multiple
interaction in the former cases enhances the K value. However,
the increment is not proportional to the number of binding sites.
42 kJ‚mol and is lower than that of single component 4.
Therefore, mixing of 3 causes considerable disorder in bilayer
4. Aqueous dispersion of 6 showed its phase transition at 40
-
1
°C with ∆H ) 11 kJ‚mol . This enthalpy change is smaller
4
-1
The binding constant for 1/ATP (2.5 × 10 M ) is significantly
than those reported for typical bilayer membranes such as
4
-1
27,28
19a
smaller than that for 1/ADP (5.4 × 10 M ), in spite of the
fact that ATP has a larger number of the possible binding sites.
The free energy gain upon binding appears to be affected by
factors other than the number of binding sites. Structural
matching of micelle surface and nucleotide conformation,
lipophile/hydrophile balance of host and guest, and steric
phospholipids
4 and other synthetic amphiphiles and may
indicate inferior molecular ordering. DSC data show that 3/4
(1:2) and 6 bilayers are in less-ordered crystalline states and 5
is in the liquid-crystalline state at 25 °C where the binding study
was performed.
(
27) Walter, A.; Hastings, D.; Gutknecht, J. J. Gen. Physiol. 1982, 79,
917.
(28) Deamer, D. W.; Bangham, A. D. Biochim. Biophys. Acta, 1976,
443, 629.
(26) (a) Bock, R. M.; Ling, N.-S.; Morell, S. A.; Lipton, S. H. Arch.
Biochem. Biophys. 1956, 62, 253. (b) Alberty, R. A.; Smith, R. M.; Bock,
R. M. J. Biol. Chem. 1951, 193, 425.

Guanidinium/Phosphate Pair Molecular Recognition
J. Am. Chem. Soc., Vol. 118, No. 36, 1996 8529
bilayer of 4 may suppress the interaction with the guanidinium
moiety of 3 by steric hindrance.
The saturation binding, R, is essentially unity for the
guanidinium function on bilayer 6. However, this is not the
case with bilayer 3/4 and 5. This lack of saturation may come
from the guanidinium function that is not exposed to the other
bilayer surface. Thus, we carried out additional binding
experiments for bilayer/AMP mixtures upon sonication. Un-
fortunately, TEM observation did not indicate the presence of
the inner (closed) bilayer surface. The binding experiment for
the sonicated samples showed contrasting results: the binding
ratio essentially reached unity in the case of the bilayer 3/4,
whereas it appeared to remain unchanged for bilayer 5. The
stoichiometric binding of AMP toward single-component bilayer
5
may be suppressed by steric crowding. In contrast, the
Figure 3. Saturation binding of guanidinium bilayers and nucleotides
at 25 °C and pH 7.
guanidinium function on bilayer 3/4 become fully exposed to
AMP upon sonication.
Guanidinium/Phosphate Binding at Microscopic and
Macroscopic Interfaces. We investigated in this study binding
of nucleotides, AMP, ADP, and ATP, toward the guanidinium
function at the surface of micelles and bilayers. Their binding
In the TEM observation, the 3/4 (1:2) mixture showed the
presence of fragments of multibilayer structures and the absence
of the inner aqueous core, although 4 itself is known to form
clear multilamellar vesicles.1
9b
Mixing with monoalkyl 3
2
4
-1
constants were in the range of 10 -10 M . The guanidinium
function must be located at the aggregate surface due to its
hydrophilic nature and the particular molecular structure of the
appears to deteriorate the bilayer structure of 4 and hinder vesicle
formation. Amphiphile 5 showed a string-like multilayer
structure. Amphiphile 6 also showed a string-like structure, but
the multibilayer structure was not obvious. All the amphiphiles
used in this study did not possess the inner water core unlike
vesicles, and the substrate trapping in the inner core need not
be considered in the following binding study.
-1
component. These values are much greater than that (1.4 M )
found for guanidinium chloride and simple phosphate in bulk
1
0
water. Therefore, the binding interaction is strengthened by
2
4
factors of 10 -10 at the microscopic surface of micelles and
bilayers, and the water molecules near the aggregate surface
provide much less interference for the binding capability of the
guanidinium unit. We investigated previously the binding of
AMP and ATP to the guanidinium unit at the air-water
Nucleotide Recognition on Guanidinium Bilayers. Ultra-
filtration of aqueous bilayer dispersions without substrates was
first conducted as a control experiment. Guanidinium am-
phiphiles 3 in 3/4 (1:2), 5, and 6 were not detected in the
corresponding filtrates, when their aqueous dispersions were
subjected to ultrafiltration. The detection limit of guanidinium
9
interface. Binding of these nucleotides toward monolayer 7
6
7
was stoichiometric, and the binding constants were 10 -10
-1
M . These values are greater than those observed for micelles
2
4
-6
and bilayers by factor of 10 -10 . Thus, the binding efficiency
moiety by V-P reaction is approximately 1 × 10 M. As
these amphiphiles form stable bilayer assemblies, the amount
of molecularly dispersed guanidinium amphiphiles (at CMC)
2
4
is enhanced by factors of 10 -10 each, as the binding
environment is converted from bulk water to microscopic
interface (micelle and bilayer), and to macroscopic interface.
Interpretation of these environmental effect is not straightfor-
ward.
-6
is less than 1 × 10 M, and bilayer aggregates remain on the
filter completely. Therefore, only the unbound substrates are
detectable in the filtrate.
Figure 4 illustrates binding patterns in bulk water, at a
microscopic surface, and at a macroscopic surface, together with
the corresponding binding constants and ∆G values for binding.
The host-guest interaction in bulk water may be defined to
occur at a molecular surface. The free energy of binding is
Binding of AMP on the bilayer of 3/4 (1:2), 5, and 6 was
investigated by equilibrium dialysis (ultrafiltration) similar to
that employed in the micelle system. Ultrafiltration was carried
out just after mixing of bilayer dispersion and AMP. The plots
of bound substrate, y, against free (unbound) substrate [S]free
-1
enhanced by ca. 20 kJ‚mol in each step. The enhanced
(
eq 3) gave saturation curves as shown in Figure 3. It was found
stabilization of the guanidinium/phosphate pair is not yet fully
explicable. A theoretical approach was recently applied to this
that the reliability of the Langmuir plot of these data was
seriously affected by large errors at low substrate concentrations.
Therefore, we used, instead, nonlinear curve fitting to estimate
K and R. The results are included in Table 1. The binding
29
interaction at the interface by Sakurai et al. As illustrated in
Figure 5, ethyl guanidinium cation and phosphate anion were
placed near/at the interface of two different dielectric media
that correspond to hydrocarbon (ꢀ ) 2.0) and water (ꢀ ) 80).
The binding energy profile was obtained by calculating by
reaction field theory the whole system as a function of the
guanidinium phosphate distance, d, and the position of a
guanidinium N relative to the interface, R. Potential minima
were found when the binding pair was exposed to water (d )
2
3
-1
constants are in the range of 10 -10 M and are comparable
to the value observed in the micelle system. However, the
binding constant for bilayers 3, 4, and 6 are smaller than those
for bilayer 5 and micelle 1 by factors of 5-10. This difference
may be related to the microenvironment at the aggregate surface,
since the chemical structure of the receptor guanidinium is
identical. We could not detect the phase transition for aqueous
bilayer 5 at 5 to 80 °C and assumed that the bilayer is in the
liquid crystalline state. Therefore, it is probable that the rate
difference is derived from difference in the physical state of
the aggregates. The guanidinium function at the surface of fluid
aggregate (micelle and bilayer) must be more efficient than that
on the crystalline aggregate surface of 3/4 and 6. As another
probable factor, long oligo(ethyleneglycol) chains in the matrix
0
to -3.4 in Figure 5) indicating favored pairing of the two
functions under these interfacial arrangements. The free energy
gained by the pairing can be comparable to the experimentally
observed binding energy.
Recently, we observed submicron range attraction between
hydrophobic surfaces of monolayer-modified mica in water by
(29) Sakurai, M.; Tamagawa, H.; Furuki, T.; Inoue, Y.; Ariga, K.;
Kunitake, T. Chem. Lett. 1995, 1001.

8530 J. Am. Chem. Soc., Vol. 118, No. 36, 1996
Onda et al.
Figure 4. Typical binding constant (K), and binding energy (∆G) of guanidinium and phosphate system at varying interfaces.
difference of the binding energy among molecular interface,
microscopic interface, and macroscopic interface is consistent
with this presumption. The air-water interface is much larger
and smoother than surfaces of aqueous micelles and bilayers,
that, in turn, provides supramolecular surfaces larger than that
of simple molecular pairs. Bonar-Law reported binding
behavior of porphyrin-based receptor in SDS micelle where
the binding was energetically similar to that in methanol (ꢀ )
8
3
2.6). Interior of micelle might provide lower dielectric media
than hydrocarbon. As for the hydrogen bonding interaction at
interfaces, Nowick and co-workers studied the thymine/adenine
5
pairing in anionic SDS micelles. The effective binding of
aqueous thymine with alkylated adenine was observed in the
hydrophobic core of the micelle, and the binding constant was
comparable to that observed in CHCl3 medium. These binding
sites in the micellar interior, however, cannot be conceived as
existing at the microscopic interface and cannot be directly
compared with our system.
Figure 5. A continuum model of the lipid-water interface for the
29
interaction of guanidinium and phosphate. d: position of guanidinium
nitrogen with respect to the interface (dielectric boundary). Here, the
parameter d is taken to be zero. R: distance between the guanidinium
carbon and the phosphorus atom.
In conclusion, we demonstrated that nucleotides were bound
to the guanidinium-functionalized micelles and bilayers (mi-
croscopic interface) via specific hydrogen bonding and elec-
trostatic attraction. It is revealed that the microscopic interface
strengthens hydrogen bonding and/or electrostatic interaction
even in the presence of water.
the surface forces measurement.30 We proposed on the basis
of these surprising results that the attraction between opposing
surfaces is much enhanced and becomes long-ranged if the
surfaces are sufficiently large, molecularly smooth, and strongly
hydrophobic. This unique feature of the interface between water
and the hydrophobic surface may be related to the enhanced
guanidinium/phosphate interaction, as the theoretical approach
suggests. Larger binding energies are gained if the interacting
pair is buried deeper in the aqueous phase. The observed
Acknowledgment. We are grateful to Dr. Izumi Ichinose,
Kyushu University, for assistance in TEM measurements. We
also appreciate Shinkai Chemirecognics Project, JRDC for
permission to use their DSC instrument.
(30) (a) Kurihara, K.; Kunitake, T.; Higashi, N.; Niwa, M. Langmuir
1
992, 8, 2087. (b) Berndt, P.; Kurihara, K.; Kunitake, T. Langmuir 1992,
8
, 2486.
JA960991+