Adenine-uridine base pairing at the water-solid-interface

Article abstract of DOI:10.1021/ja990447b

The formation of the base pair adenine-uracil at a water-solid interface, at an immobilized monolayer of adenine disulfide with adenine groups exposed to the very surface, respectively, is shown here. To overcome the steric hindrance of tightly packed ade

Full text of DOI:10.1021/ja990447b

J. Am. Chem. Soc. 2000, 122, 87-95
87
Adenine-Uridine Base Pairing at the Water-Solid-Interface
Michael Weisser, Josua Ka1shammer, Bernhard Menges, Jin Matsumoto,†
Fumio Nakamura,† Kuniharu Ijiro,† Masatsugu Shimomura,† and Silvia Mittler*
Contribution from the Max-Planck-Institut fu¨r Polymerforschung, Ackermann Weg 10,
55128 Mainz, Germany, and Research Institute for Electronic Science, Hokkaido UniVersity,
N12W6, Sapporo 060, Japan
ReceiVed February 11, 1999. ReVised Manuscript ReceiVed August 24, 1999
Abstract: The formation of the base pair adenine-uracil at a water-solid interface, at an immobilized monolayer
of adenine disulfide with adenine groups exposed to the very surface, respectively, is shown here. To overcome
the steric hindrance of tightly packed adenine groups in a pure adenine thiolate monolayer on gold, the formation
of self-assembled monolayers out of a binary mixture of the adenine disulfide and CH3- or OH-terminated
thiols are investigated. Electro-chemical investigations, surface plasmon spectroscopy (PSP, plasmon surface
polariton), multimode waveguide-PSP-coupling spectroscopy, contact angle measurements, and spontaneous
desorption time-of-flight mass spectrometry were used to characterize the monolayers. The specific base pairing
was investigated for a variety of monolayer compositions. A specific base pairing was successful for an optimized
mixed adenine/OH-terminated thiol monolayer. Nevertheless unspecific binding is a problem.
Introduction
dimensional geometry was achieved by an air-water interface
on a Langmuir-Blodgett (LB) trough, where the hydrogen
bonding took place in the water phase. An amphiphilic diami-
notriazine was able to selectively bind nucleosides and nucleic
acid bases. In 1997 the base pairing of cytosine with guanosine
at the air-water interface was found as the first two-dimensional
system being biologically relevant.17 In the same year Matsuura
et al.18 have demonstrated the two-dimensional hydrogen
bonding on a solid-air interface via a self-assembled monolayer.
Recently an artificial hydrogen bonding molecular pair was
demonstrated at a solid-hydrophilic organic solvent interface.19
Here we like to demonstrate the possibility of using hydrogen
bonding in a two-dimensional array at the solid-water interface.
Hydrogen bonding in water is especially significant due to its
relevance to biological molecular recognition. The base pair
adenine-uracil was chosen. Therefore a disulfide with two
spacers and an adenine headgroup at each end (Figure 2) was
synthesized for forming self-assembled adenine thiolate mono-
layers exposing adenine groups to the very surface of a solid
metal substrate.20 Uridine as the water soluble derivative of
uracil was investigated for the binding processes. For unspecific
binding tests cytidine was used. Figure 1 demonstrates the
specific recognition and binding via hydrogen bonds of the base
pairs adenine/uracil and guanine/cytosine. The chemical struc-
tures of the water soluble uridine and cytidine used in this study
are shown as well (Figure 1b).
To tailor molecular organization is one of the final goals of
supra molecular chemistry1,2 and is essential for the design of
molecular devices.3,4 Weak intermolecular interactions such as
hydrogen bonds,5 as well as the interactions in typical guest-
host systems such as in the biotin-streptavidin-system6,7 or in
the family of the cyclic molecules such as cylodextrins8-12 or
calixarenes13,14 are well-known architectural tools for the
assembling of molecular organization. Nature has used these
tools in a wide variety to create functionality.15 Cell-cell
recognition or communication via guest-host reactions of
proteins or the most versatile DNA double helix are typical
examples where molecular organization based on specific
intermolecular interactions delivers a very particular biological
functionality and therefore contains well-defined information.15
Hydrogen bonding was studied intensely for the past decades
within three-dimensional geometry, for example, DNA in
solution.15 Two-dimensional arrangements of molecules being
able to develop hydrogen bonds with a fitting partner were
studied for the first time by Kurihara et al. in 1991.16 The two-
* Corresponding author: e-mail: mittler@mpip-mainz.mpg.de.
† Hokkaido University.
(1) Lehn, J. M. Angew. Chem., Int. Ed. Engl. 1989, 27, 89.
(2) Lehn, J. M. Angew. Chem., Int. Ed. Engl. 1990, 29, 1304.
(3) Ringsdorf, H.; Schlarb, B.; Venzmer, J. Angew. Chem., Int. Ed. Engl.
1989, 27, 113.
(4) Kunitake, T. Angew. Chem., Int. Ed. Engl. 1992, 31, 706.
(5) Boschke, F. L., Ed. Hydrogen bonds. In Topics in Current Chemistry;
Springer-Verlag: Heidelberg, 1984; Vol. 120.
(6) Spinke, J.; Liley, M.; Schmitt, F.-J.; Guder, H.-J.; Angermaier, L.;
Knoll, W. J. Chem. Phys. 1993, 99, 7012.
(7) Spinke, J.; Liley, M.; Guder, H.-J.; Angermaier, L.; Knoll, W.
Langmuir 1993, 9, 1821.
(8) Saenger, W. Angew. Chem., Int. Ed. Engl. 1980, 19, 344.
(9) Wenz, G. Angew. Chem., Int. Ed. Engl. 1994, 33, 803.
(10) Nelles, G.; Weisser, M.; Back, R.; Wohlfart, P.; Wenz, G.; Mittler-
Neher, S. J. Am. Chem. Soc. 1996, 118, 5039.
(11) Weisser, M.; Nelles, G.; Wohlfart, P.; Wenz, G.; Mittler-Neher, S.
J. Phys. Chem. 1996, 100, 17893.
(13) Bo¨hmer, V. Angew. Chem., Int. Ed. Engl. 1995, 34, 713.
(14) Huisman, B.-H.; Kooymann, R. P. H.; van Veggel, F. C. J. M.;
Reinhoudt, D. N. AdV. Mater. 1996, 8, 561.
(15) Hoppe, W.; Lohmann, W.; Markl, H.; Ziegler, H. Biophysik;
Springer-Verlag: Heidelberg, 1982.
(16) Kurihara, K.; Ohto, K.; Honda, Y.; Kunitake, T. J. Am. Chem. Soc.
1991, 113, 5077.
(17) Shimomura, M.; Nakamura, F.; Ijiro, K.; Taketsuna, H.; Tanaka,
M.; Nakamura, H.; Hasebe, K. J. Am. Chem. Soc. 1997, 119, 2341.
(18) Matsuura, K.; Ebara, Y.; Okahata, Y. Langmuir 1997, 13, 814.
(19) Motesharei, K.; Myles, D. C. J. Am. Chem. Soc. 1998, 120, 7328.
(20) Ulman, A. Chem. ReV. 1996, 96, 1533.
(12) Weisser, M.; Nelles, G.; Wenz, G.; Mittler-Neher, S. Sens. Actuators,
B 1997, 38-39, 58.
10.1021/ja990447b CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/18/1999

88 J. Am. Chem. Soc., Vol. 122, No. 1, 2000
Weisser et al.
e
e
Figure 1. (a) Specific recognition and binding of the base pairs
adenine/uracil and guanine/cytosine. (b) Chemical structure of uridine
and cytidine.
Materials
11-mercapto-1-undecanol (OH(CH2)11SH, Aldrich, OHT) and decane
thiol (CH3(CH2)9SH, Aldrich, CH3T) (Figure 2) were used without
further purification.
Synthesis of 11-[2-(9-Adeninyl)propionyloxy]undecyl Disulfide
(A). Nuclear magnetic resonance (NMR) spectra were recorded with a
JEOL EX-400 spectrometer (1H, 400 MHz). Chemical shifts are
reported in parts per million (ppm) downfield of tetramethylsilane
(TMS). Fast atom bombardment (FAB) mass spectra were recorded
with a JEOL HX110 mass spectrometer with a cesium ion source and
positive ion detection. Fourier transform infrared (FTIR) spectra were
recorded with a JASCO FTIR-300 spectrometer. The samples were
prepared as KBr pellets.
Figure 2. Synthesis scheme for 11-[2-(9-adeninyl)propionyloxy]-
undecyl disulfide and the structure formula of the OHT and the CH3T.
residual solid was washed with benzene while stirring. The fine white
product was filtered and washed with benzene to give p-nitrophenyl
ester (3) in 75% yield. The product was immediately used for a next
reaction without further purification.
Pyridine was freshly distilled from CaH2. Dimethylformamide (DMF)
was freshly distilled. All other reagents were of the highest available
commercial quality and used without further purification.
11-[2-(9-Adeninyl)propionyloxy]undecyl Disulfide (4). A mixture
of p-nitrophenyl ester (1.0 g, 2.7 mmol), imidazol (0.44 g, 6.4 mmol)
and 11-mercapto-1-undecanol (0.68 g, 3.2 mmol) in a mixture of 10
mL of N,N-dimethylformamide (DMF) and 10 mL of dichloromethane
was stirred for 2 days at 25 °C. The precipitate was collected and
washed with dichloromethane, ethanol, water, and ethanol. The product
was recrystallized in ethanol to give 11-[2-(9-adeninyl)propionyloxy]-
undecyl disulfide (4) as a mixture with a thioester compound as a side
product, 11-[2-(9-adeninyl)propionylthio]-1-undecanol (50 wt %) (13%
yield). The product was used without further purification due to a
difficulty of further separation and considerably small probability of
Ethyl 3-(9-Adeninyl)propionate (1).21,22 A mixture of adenine (10
g, 74 mmol), metal sodium (74 mg, 32 mmol), and ethyl acrylate (24
mL, 0.22 mol) in 200 mL of anhydrous ethanol was refluxed for 2 h.
The precipitate was collected and recrystallized in ethanol to give ethyl
3-(9-adeninyl)propionate (1) in 77% yield: mp 170-171 °C; 1H NMR
(CDCl3) d 1.20-1.23 (t, J ) 7.2 Hz, 3 H), 2.91-2.93 (t, J ) 3.2 Hz,
2 H), 4.11-4.16 (q, J ) 7.6 Hz, 2 H), 4.47-4.51 (t, J ) 6.4 Hz, 2 H),
5.49 (s, 2 H), 7.87 (s, 1 H), 8.35 (s, 1 H); IR 1720, 1680, 1609, 1205,
651 cm-1
.
3-(9-Adeninyl)propionic Acid (2). The ester (1) (13 g, 55 mmol)
in 250 mL of 5 N HCl was refluxed for 3 h. Neutralization of the
hydrochloride salt of (2) by sodium carbonate solution gave 3-(9-
adeninyl)propionic acid (2) in 67% yield: mp 285-288 °C; 1H NMR
(DMSO-d) d 2.84-2.87 (t, J ) 7.2 Hz, 2 H), 4.33-4.36 (t, J ) 6.8
Hz, 2 H), 6.96 (s, 2 H), 8.04 (s, 1 H), 8.13 (s, 1 H); IR 1703, 1617,
1
the thioester binding to the gold surface: mp 153-155 °C; H NMR
(DMSO-d) d 1.21-1.23 (m, 28 H), 1.37-1.46 (m, 8 H), 2.67-2.70 (t,
J ) 7.6 Hz, 4 H), 2.78-2.82 (t, J ) 7.6 Hz, 4 H), 3.96-3.99 (t, J )
6.4 Hz, 4 H), 4.39-4.42 (t, J ) 7.2 Hz, 4 H), 6.99 (s, 4 H), 8.05 (s,
2H), 8.14 (s, 2H); IR 2919, 2850, 1727, 1645, 1594, 1199 cm-1; MS
786 (M+).
1410, 645 cm-1
.
The synthesis scheme for 11-[2-(9-Adeninyl)propionyloxy]undecyl
disulfide is shown in Figure 2. The materials were dissolved in ethanol
(Chromasolv, Riedel-de-Haen).
4-Nitrophenyl 3-(9-Adeninyl)propionate (3). p-Nitrophenyl tri-
fluoroacetate (7.1 g, 30 mmol) was added to a suspension of 3-(9-
adeninyl)propionic acid (2) (5.0 g, 24 mmol) in 150 mL of dry pyridine.
The mixture was stirred for 2 days at 25 °C to give a clear solution.
After vacuum distillation of pyridine at 60 °C to complete dryness the
Experimental Setup
(A) Immobilization of the Adenine. The self-assembled monolayers
were formed on gold films, vacuum-evaporated at a pressure of 5‚10-6
mbar onto cleaned LaSFN9 (Schott) substrates. After preparation, the
(21) Lira, E. P.; Huffman, C. W. J. Org. Chem. 1966, 31, 2188.
(22) Overberger, C. G.; Inaki, Y. J. Polym. Sci., Polym. Chem. Ed. 1979,
17, 1739.

A/U Base Pairing at the Water-Solid-Interface
J. Am. Chem. Soc., Vol. 122, No. 1, 2000 89
dc
gold-coated substrates were allowed to cool under vacuum for ap-
proximately 30 min and were then immediately placed in solution of
the appropriate self-assembling adenine disulfide/alkylthiol binary
mixture or stored under argon. All adsorption processes were performed
in ethanol (Chromasolve, Riedel-de-Haen) with a concentration of 0.5
mM. The mole fraction of the adenine in the OH- or CH3-terminated
thiol was varied from 0 to 1 by mixing corresponding volumina taking
the side product of the thioester into account. The uridine as the binding
base to adenine and cytidine for testing of unspecific binding were
solved in milli-Q water (Millipore) as 0.5 mM solutions.
c ) ∆nd
dn
(1)
(D) Contact Angle Measurement. Advancing and receding contact
angles of water on the films were measured using a contact angle
microscope (Kru¨ss G-1) under ambient conditions, while the volume
of the drop is increased or decreased at the minimum rate required for
movement of the water/air/solid triple point.
(E) Spontaneous Desorption Time-of-Flight Mass Spectrometry
(SDMS). The measurements were performed using a linear time-of-
flight (TOF) mass spectrometer in high vacuum at a pressure of about
1 × 10-6 mbar. The self-assembled monolayers were deposited on 50
nm thick gold films evaporated on glass slides that were covered first
with a 2 nm chromium film to increase the mechanical stability. Atomar
and molecular ions from the sample were released by spontaneous
desorption, a secondary ion process in which the sample is not
bombarded by particles from an external source.27 Primary ions of
adsorbates are field-desorbed from the edges of an acceleration grid
located in front of the sample. These ions are accelerated toward the
sample gaining keV energies. They finally sputter secondary ions from
the sample, and these are analyzed with the TOF mass spectrometer.
Spectra of negative secondary ions are recorded using the simulta-
neously emitted secondary electrons from the sample surface as trigger
particles. An acceleration voltage of 9.5 kV was applied to the sample.
The recording time of a spectrum was 20-40 min. Within one spectrum
mass peaks of interest were integrated and normalized with the number
of start events with one or more corresponding stop events. This leads
to a relative ion yield that allows the comparison of the intensities of
equivalent peaks in different spectra.
(F) Multimode waveguide-PSP-coupling. To determine the thick-
ness and the refractive index of the SAM independently from each
other the recently published method of multimode waveguide-PSP-
coupling was used.23,28-36 An ion-exchanged buried multimode glass
channel waveguide was coated with a 40 nm thick gold film to enable
the coupling of the waveguide modes into the surface plasmon. Due to
the resonant PSP coupling the dispersion of the effective refractive
index of the waveguide mode and therefore the imaginary part of the
effective refractive index is altered, which leads to a change in the
transmitted light intensity of the waveguide modes. This intensity
change can be back-calculated via Fresnel theory to an average thickness
and an average refractive index of the SAM.23,37
The preparation of the waveguide surface plasmon (PSP, plasmon
surface polariton)-coupling device is described elsewhere.23
(B) Electrochemical Measurements. Electrochemistry was per-
formed on a NIKKO KEISOKU NPGFZ-2501-A potentio-galvanostat.
Working electrodes were typically 1 cm2 evaporated gold (2000 Å) on
tin-indium-oxide (ITO) glass. Monolayers of OHT and adenine
thiolate20 were prepared on the electrodes in the same manner as that
described above. A platinum wire was used as a counter electrode. A
saturated calomel electrode (SCE) was used as reference in all
electrochemical experiments. The potential was initiated at 0.00 V and
cycled at 50 mV/s between -1.45 and 0.00 V in 0.5 M aqueous KOH
solution. Deionized water, which was purified by passage through a
milli-Q filtration system, was used for all solution preparation. Solutions
were degassed with N2 for 15 min prior to the measurements.
(C) Surface Plasmon Spectroscopy (PSP, Plasmon Surface
Polariton). PSP measurements were performed to study the optical
film thickness of the SAM with respect to the concentration of the
adenine in the binary mixture of the self-assembly solution. Here the
Kretschmann configuration24 was used with a 50 nm gold film
evaporated onto a substrate, which is then optically matched to the
base of a 90° LaSFN9 glass prism (n )1.85 at λ ) 632.8 nm). Thus,
the plasmon surface polaritons are excited at the metal/dielectric
interface, upon total internal reflection of the laser beam (HeNe, λ )
632.8 nm, power 5 mW) at the prism base. By varying the angles of
incidence of the laser beam, a plot of reflected intensity as a function
of the angle of incidence is obtained. The reflected intensity shows a
sharp minimum at the PSP resonance angle which depends on the
precise architecture of the metal/dielectric interface and is defined by
the matching condition for energy and momentum between the
evanescent photons and the plasmon surface polariton. Adsorption
processes occurring at the gold interface were followed in real time by
selecting an appropriate angle of incidence and monitoring the reflected
intensity as a function of time. Knowledge of the form of the resonance
curve allows this intensity to be interpreted as a shift in the angle of
resonance.
From a Fresnel simulations to the resonance curve for bare gold
surfaces, it is possible to obtain the dielectric constant and the thickness
of the gold layer. Addition of a thin layer to the surface of the gold
typically shifts the position of the resonance to a higher angle, and a
simulation of this second resonance curve determines the optical
thickness, (nd), of the layer. Although plasmon surface polariton
measurements allow the determination of an average optical thickness
of an adsorbed film, accurate conversion of this optical thickness to a
geometrical thickness requires knowledge of the refractive index of
the film, a parameter which depends on both the molecular composition
of the film and the packing density. In practice, it is not possible to
distinguish between a thin film with a high refractive index and a film
twice as thick but with half the refractive index contrast in the medium.
For the data analysis here a refractive index of n ) 1.5 was chosen for
both the adenine thiolate and the diluting thiols, OHT and CH3T. The
data for the pure adenine thiolate and the pure OHT and CH3T
monolayers were evaluated using a concentration-dependent refractive
index measurement (dn/dc) of the materials in ethanol25 and calculate
the surface concentration c in [mol/cm2].26
Monolayer Characterization
Measurements. To find the optimum binding conditions at
the solid-water interface of the uridine the adenine thiolate is
laterally diluted by the OHT or CH3T.6,7 Due to the bulky
headgroup of the adenine disulfide, monolayers of the pure
adenine thiolate might not be very ordered and the structure of
the film not very well understood.38 Therefore, the mole fraction
x of the adenine within the self-assembly solutions was
systematically varied from pure OHT or CH3T solutions (x )
0) to pure adenine disulfide solutions (x ) 1).
The length of the anchor groups of the adenine disulfide and
the diluting thiols was chosen to achieve via van der Waals
(27) Voit, H.; Schoppmann, C.; Brandl, D. Phys. ReV. B 1993, 48, 17517.
(28) Lavers, C. R.; Wilkinson, J. S. Sens. Actuators, B 1994, 22, 75.
(29) Harris, R. D.; Wilkinson, J. S. Sens. Actuators, B 1995, 29, 261.
(30) Weiss, M. N.; Srivastava, R.; Groger, H.; Lo, P.; Luo, S.-F. Sens.
Actuators, A 1996, 51, 211.
(31) Ctyroky, J.; Homola, J.; Skalsky, M. Opt. Quantum Electron. 1997,
29, 301.
(32) Mouvet, C.; Harris, R. D.; Maciag, C.; Luff, B. J.; Wilkinson, J.
S.; Piehler, J.; Brecht, A.; Gauglitz, G.; Abuknesha, R.; Ismail, G. Anal.
Chim. Acta 1997, 338, 109.
(33) Brecht, A.; Gauglitz, G. Anal. Chim. Acta 1997, 347, 219.
(34) Homola, J.; Ctyroky, J.; Skalky, M.; Hradilova, J.; Kolarova, P.
Sens. Actuators, B 1997, 38-39, 286.
(35) Ctyroky, J.; Homola, J.; Skalsky, M. Electron. Lett. 1997, 33, 1246.
(36) Weisser, M.; Menges, B.; Mittler-Neher, S. SPIE 1998, 3414, 250.
(37) TRAMAX, University of Jena, Institut of Theoretical Optics, Max
Wien Platz 1, 07743 Jena, Germany.
(23) Weisser, M.; Menges, B.; Mittler-Neher, S. Sens. Actuators, B, in
press.
(24) Kretschmann, E. Opt. Commun. 1972, 6, 185.
(25) Becker, A.; Ko¨hler, W.; Mu¨ller, B. Ber. Bunsen-Ges. Phys. Chem.
1995, 99, 600.
(26) Miller, C. E.; Meyer, W. H.; Knoll, W.; Wegner, G. Ber. Bunsen-
Ges. Phys. Chem. 1992, 96, 869.
(38) Qian, J.; Hentschke, R.; Knoll, W. Langmuir 1997, 13, 7092.

90 J. Am. Chem. Soc., Vol. 122, No. 1, 2000
Weisser et al.
of the molecule are desorbed: the molecular peak of the adenine
thiolate at m/z ) 393 u is missing as well as fragments which
would arise from the thioester. The CN- fragment of the
heterocyclic purin ring at m/z ) 26 u can be seen by the
inversion of the intensity of the pump oil peaks at m/z ) 25 or
26 u, respectively. The deprotonated adenine (A-) is found at
m/z ) 134 u. A frequent fragment of the adenine thiolate is the
deprotonated carboxyethyle (A + EtCOO-, m/z ) 206 u). The
peak at m/z ) 229 u of the AuS- fragment shows the covalent
bond between the adenine compound and the gold. A series of
small mass peaks arise from small fragments from the alkyl
chain and the adenine. They are not discussed any beyond this
point, because they are common in all the spectra under
investigation and yield no further information.
The pure CH3T monolayer shows similar low mass peaks of
fragments from the alkyl chains. The lightest specific fragment
is the decan thiolate ion (S(CH2)9CH3-) at m/z ) 173 u. Stronger
is the peak at m/z ) 370 u, the gold decane thiolate ion (Au +
CH3T-). Unusually high is the peak at m/z ) 543 u, which can
be ascribed to a gold didecanthiolate ion (Au + (CH3T)-2). The
fragment at m/z ) 229 u is found in this spectrum as well,
pointing to the covalent bond between the CH3T and the gold
surface.
Figure 3. Cyclic voltammograms initiated at 0.00V for the reductive
desorption of OHT (- - -) and adenine thiolate (s) monolayers from
Au surfaces.
Table 1. Charge (Q), Surface Coverage (Γ) and Area per Molecule
(A) Determined from Electrolysis of OHT and Adenine Thiolate
Monolayers Assembled on Aua
thiolate
Q [µC/cm2]
Γ [mol/cm2]
A [Å2]
OHT
adenine
73.13 ( 7.00
62.05 ( 6.00
7.6 ( 0.7
6.4 ( 0.6
22 ( 3
26 ( 3
a The data were not corrected for surface roughness of the gold.
interactions between these chains a well-ordered mixed
monolayer.39-41
The spectrum of the binary mixture shows all of the
characteristic peaks of both adsorbed species.
An electrochemical approach for determining the surface
coverage of thiolate monolayers formed at gold electrodes were
reported.42,43 The approach is based on the voltammetric
measurement of the charge passed for the one-electron reductive
desorption of the resulting gold-bound thiolate layer (Au-SR)
in alkaline solution (pH > 11): Au-SR + e- f Au(0) + -SR.
Figure 3 shows the cyclic voltammograms (CV), the current-
voltage (i-E) curves for the OHT and the adenine thiolate
monolayers. The large cathodic wave with a peak current near
-1.30 V and -1.35 V in the first scan of the OHT and the
adenine thiolate layers, respectively, arises from the one-electron
reduction. The charge associated with the electrode reaction (Q)
was determined by integration of the cathodic peak in the first
voltammetric scan. A surface coverage (G) and a molecular area
(A) of thiolate compounds bound to the gold surfaces were
calculated from the charge values and the apparent area of the
electrode. Those values are summarized in Table 1. The surface
coverage (Γ) both of OHT and adenine thiolate layers coincide
within the experimental errors with these of alkanethiolate
layers.42 We therefore assume that the side product of the
thioester is not binding to the gold.
The analogue mass spectra were taken for OHT as the diluting
molecule. Figure 5 shows the mass spectrum of the pure adenine
thiolate SAM (5a, like 3a, for comparison shown again), the
pure OHT SAM (5b) and of a SAM which was self-assembled
out of a 19:2 (adenine/OHT), x ) 0.95, binary solution (5c). In
the low mass regime the spectrum of the OH terminated thiol
is identical with all the other investigated monolayers. The only
characteristic peak is found at m/z ) 203 u, the deprotonated
11-mercapto-1-undecanol ion (OHT-). Gold thiolate ions were
not found. The AuS- peak at m/z ) 229 u is present. The mass
spectrum of the monolayer of the binary mixture depicts the
characteristic peaks of both adsorbed species.
For the PSP measurements monolayers out of binary mixtures
between the diluting thiols and the adenine disulfide were self-
assembled for 8-10 h. For the pure materials the surface density
or the area per molecule, respectively was determined via the
dn/dc measurement and eq 1. For the thickness determination
a refractive index of 1.5 was chosen for all three materials. The
results for the pure materials are summarized in Table 2. The
Mass spectra were taken from SAMs which were self-
assembled out of solutions with varying adenine disulfide
concentration. In Figure 4 the mass spectra for the pure adenine
thiolate SAM (4a), the pure CH3T SAM (4b) and a SAM self-
assembled out of a 19:2 (A/CH3T), x ) 0.95, binary solution
(4c) are shown. In all spectra the gold ions Au- (m/z ) 197 u),
Au2- (m/z ) 394 u), and Au3- (m/z ) 591 u) are found. Besides
these gold signals peaks are found which appear periodically
with the higher gold clusters. For example, at m/z ) 221, 222,
and 223 u peaks are found which belong to Au- with pump oil
contaminations of m/z ) 24, 25, and 26 u.44 The spectrum of
the pure adenine thiolate monolayer shows that only fragments
Table 2. Thickness, dn/dc-values and Surface Density of the Pure
CH3T, OHT, and 11-[2-(9-Adeninyl)propionyloxy]undecyl Disulfide
CH3 T
OHT
adenine thiolate
thickness [Å] (n ) 1.5)
dn/dc [ml/mol]
10 ( 1.5
18.4 ( 1
22 ( 3
13 ( 1.7
26.5 ( 1.2
24 ( 3
28 ( 1.8
70.3 ( 1.6
31 ( 3
area per molecule [Å2]
area per molecule is similar within the errors for the two diluting
thiols (22-24 Å2), the bound adenine thiolate depicts a larger
area per molecule of 31 Å2 because of the large adenine
headgroup. The thicknesses of the pure and the mixed mono-
layers are shown with respect to the mole fraction x of the
adenine in the self-assembly solution in Figure 6. The film
thickness increases very slightly from the pure OHT or CH3T
(x ) 0) to x ) 0.7. From 70 to 100% adenine in the
self-assembly solution the film thickness increases strongly to
the final adenine thiolate monolayer thickness of 28 Å. The
concentration of the adenine in the monolayer is obviously
different from the concentration within the self-assembling
solution.
(39) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am.
Chem. Soc. 1987, 109, 3559.
(40) Hautmann, J.; Klein, M. L. J. Chem. Phys. 1989, 91, 4994.
(41) Fenter, P.; Eisenberger, P.; Liang, K. S. Phys. ReV. Lett. 1993, 70,
2447.
(42) Widrig, C. A.; Chung, C.; Porter, M. D. J. Electroanal. Chem. 1991,
310, 335.
(43) Walczak, M. M.; Popenoe, D. D.; Deinhammer, R. S.; Lamp, B.
D.; Chung, C.; Porter, M. D. Langmuir 1991, 7, 2687.
(44) Wohlfart, P. Ph.D. Thesis, Mainz, Germany, 1997.

A/U Base Pairing at the Water-Solid-Interface
J. Am. Chem. Soc., Vol. 122, No. 1, 2000 91
Figure 4. SDMS spectra of (a) a pure adenine thiolate SAM, (b) a pure CH3T SAM, and (c) a SAM self-assembled out of a 19:2 (adenine/CH3T)
binary solution on gold.
Figure 5. SDMS spectra of (a) a pure adenine thiolate SAM, (b) a pure OHT SAM, and (c) a SAM self-assembled out of a 19:2 (adenine/OHT)
binary solution on gold.
A similar behavior is found by the contact angle measure-
ments. Figure 7 depicts the advancing and receding contact
angles of water on the CH3T diluted adenine thiolate mono-
layers. High contact angles of around 100° representing
hydrophobic surfaces are found for a wide mole fraction range
of x ) 0 to x ) 0.7. For higher adenine concentrations the
contact angles decrease steeply to 20-40° describing a hydro-
philic surface. The hysteresis between advancing and receding
contact angle shows the same behavior. For the high contact
angle regime the hysteresis is small, about 10°, for the small
contact angle regime the hysteresis is larger. At x ) 0.95 a
maximum of the hysteresis of 38° is found, which decreases
for the pure adenine thiolate monolayer to 19°.
In Figure 8 the contact angle measurements for the OHT
diluted adenine thiolate monolayers are shown. The development
of the contact angles with increasing amount of adenine
corresponds to the hydrophilic nature of the OH-termination.
Starting at very low contact angles (5-12°) for the pure OHT
at x ) 0 the contact angles slightly increase up to x ) 0.82 to
20-40°. The hysteresis is constant from x ) 0 to x ) 0.82 at
around 7° and increases to 18° at x ) 0.95. For the pure adenine
thiolate monolayer the hysteresis reaches the maximum of 19°.
Multimode waveguide-PSP-coupling experiments were per-
formed for adenine thiolate monolayers diluted with OHT only.23
The average geometrical thicknesses and the average refractive
indices of these mixed monolayers are depicted in Figure 9.

92 J. Am. Chem. Soc., Vol. 122, No. 1, 2000
Weisser et al.
Figure 6. Thickness of the mixed monolayers (n ) 1.5) versus mole
fraction of the adenine in the self-assembly solution: for CH3T (0)
and OHT (O).
Figure 8. Advancing (0) and receding (O) contact angle of water
versus the mole fraction of adenine in the self-assembly solution for
OHT as the diluting molecule.
which is reasonable. However, the area per molecule value of
31 ( 3 Å2 or 26 ( 3 Å2, from the electrochemical measure-
ments, respectively, yields a much higher surface density than
that calculated for a stretched molecular arrangement with an
area per molecule of 36 Å2.47 The large hysteresis of the contact
angle measurements for the pure adenine thiolate monolayers
points to a rough surface, caused by an unorganized film. A
modified monolayer architecture has therefore to be taken into
account. Obviously more adenine thiolates are bound to the
surface than adenine headgroups could be exposed to the surface.
A theoretical investigation has shown for cyclodextrin thiols,
molecules with a similar geometry, that they can form mono-
layers consisting of molecules which are stretched and which
expose the active group and in addition molecules which are
filling the space under and between the stretched species.38 The
measured high refractive index of the pure adenine thiolate film
also supports the filling of the space under the large adenine
groups. If the density of the grafted molecules would be
influenced only by the tight packing of the headgroups, the
overall refractive index must be lower, because of the smaller
density of the film under the adenine.
This behavior is one of the two distinct differences between
the two-dimensional arrangements of adenine-containing am-
phiphiles at the air-water interface on LB troughs and adenine
thiolates self-assembled on a metal. In the case of the air-water
interface the molecules are arranged with respect to the size of
their headgroups, whereas in the case of the self-assembly
approach the grafting density is dominated by the binding of
the much smaller thiolate groups. Therefore, it is necessary to
dilute the active molecules of the self-assembled monolayer with
proper inactive molecules. The concentration of the active
species within the self-assembled monolayer, the adenine
thiolate, respectively, then does not necessarily have to be
identical with the concentration in the self-assembly solution.
This seems to be the case with both of the chosen diluting
molecules. All measurements point to a concentration behavior
of the adenine thiolate on the surface which is strongly shifted
to very high adenine concentrations within the self-assembly
solution. Either the thickness, as the contact angles and their
hysteresis show an influence of the adenine thiolate content in
the self-assembled monolayers first from solution mole fractions
of x g 0.8. Therefore the mass spectra were analyzed in a
quantitative way. First of all the number of counts within a
molecular specific peak are integrated for the pure monolayers.
For the pure adenine thiolate monolayer the deprotonated
adenine peak at m/z ) 134 u and the deprotonated carboxy-
Figure 7. Advancing (0) and receding (O) contact angle of water
versus the mole fraction of adenine in the self-assembly solution for
CH3T as the diluting molecule.
Both thickness and refractive index are only slowly increasing
with increasing mole fraction of the adenine disulfide in solution.
Both refractive index and film thickness are close to the values
expected for pure OHT monolayers.20 At very high mole
fractions of adenine in solution the average geometrical thickness
reaches 22 Å and a refractive index n of 1.53.
Discussion
The results from the electrochemical investigations show that
self-assembled monolayers of OHT and adenine thiolate were
formed on the gold surfaces. The smaller surface coverage of
adenine thiolate compared to the OHT is considered to be caused
by the larger headgroup of the adenine compound.
The thiols and the disulfides are covalentely bound to the
gold surface as thiolates.20 All the mass spectra depict the peak
for the AuS- ion and show specific peaks for the self-assembled
monolayers for the pure ones as well as for the mixed ones.44
The OHT and the CH3T are able to dilute the adenine thiolate
monolayers and exhibit the expected surface hydrophilicity or
hydrophobicity. The PSP measurements yield thicknesses for
the pure OHT and CH3T which are in perfect agreement with
the literature.45 The model of the parallel aligned alkyl chains
with a tilt angle of 25-30° to the surface normal can be applied.
An analogue construction of the monolayer for the adenine
thiolate taking into account the geometry of the purine ring46
leads to a thickness of 29 Å. For the PSP data analysis with a
refractive index of 1.5 a layer thickness of 28 Å was found
(45) Ulmann, A. Ultrathin organic films; Academic Press: Boston, 1991.
(46) Solomons, T. W. G. Organic Chemistry; Wiley & Sons: New York,
1994.
(47) Boland, T.; Ratner, B. D. Langmuir 1994, 10, 3845.

A/U Base Pairing at the Water-Solid-Interface
J. Am. Chem. Soc., Vol. 122, No. 1, 2000 93
Figure 9. (a) Average geometrical thickness and (b) average refractive index of the adenine thiolate/OHT mixed monolayers versus mole fraction
of adenine in solution.
Figure 10. Mole fraction ø in the monolayer of two specific fragments
of the adenine thiolate, namely A- (-0-) and A+EtCOO- (-O-) and
two specific fragments of CH3T, namely, Au + CH3T- (-4-) and Au
Figure 11. Mole fraction ø in the monolayer of a specific fragment
of the adenine thiolate, namely, A- (-0-), and a specific fragment of
OHT, namely, OHT- (-O-), of the mixed self-assembled monolayer
versus the mole fraction of adenine in the self-assembly solution.
-
+ (CH3T)2 (-3-), of the mixed self-assembled monolayer versus the
mole fraction of adenine in the self-assembly solution.
time the mole fraction of the diluting molecule decreases. The
sum of the mole fractions of the adenine thiolate and the CH3T
are close to 1 within the experimental errors. A 1/1 mixed
monolayer on the gold can be self-assembled out of an adenine
disulfide solution containing 90% adenine.
A similar picture is found for the OHT-adenine thiolate
mixed monolayers. The results from the mass spectra data
analysis via eq 2 are seen in Figure 11. Also in this case the
specific fragment for the diluting molecule OHT shows up with
a mole fraction ø of 1-0.8 on the gold surface between adenine
concentrations in solution from x ) 0 to x ) 0.7. From x ) 0.7
to the pure adenine disulfide solution the mole fraction ø of the
adenine thiolate on the gold surface increases from 0.2 to 1.
The 1/1 mixed monolayer can be achieved by self-assembling
in an 85% adenine solution. The sum of the mole fractions ø of
both fragments on the gold surface is one within the experi-
mental errors.
Both diagrams clearly depict that the concentrations within
the self-assembly solution and on the gold surface are not
identical. A very high mole fraction of the adenine disulfide in
solution is necessary to find adenine thiolates on the gold
surface. This behavior can be explained by the intermolecular
interactions. It is known, that the van der Waals interaction
between the alkyl chains in self-assembled monolayers gives a
major enthalpic contribution to the order and structure. Estima-
tions show that per CH2 group a cohesive energy of 6.9 kJ/mol
are expected.39-41,45,48-51 In the pure adenine thiolate monolayer
ethyladenine peak at m/z ) 206 u were taken into account. The
CH3T monolayer is characterized by the specific peaks of the
decanthiol ion at m/z ) 173 u, the gold thiolate ion at m/z )
370 u and the gold didecan thiolate at m/z ) 543 u. For the
OHT monolayer only the peak at m/z ) 203 u, the deprotonated
mercaptoundecanal ion. The integrals over these peaks IP100%
are normalized by the total counts of that particular mass
spectrum IT100%. The value IP100%/IT100% corresponds to a mole
fraction of 1: a gold surface which is covered by one thiolate
species only. For the mixed monolayers the specific peaks IPx%
in the mass spectra are also integrated and normalized by the
number of counts of the total spectrum ITx%. The mole fraction
ø of the fragment, and therefore of the thiolate under investiga-
tion in the monolayer is
I
Px%/ITx%
ø )
(2)
IP100%/IT100%
The mole fractions ø calculated by eq 2 for the adenine
thiolate-CH3T mixed monolayer for four specific fragment
peaks of both molecular species versus the mole fraction x of
the adenine in solution are shown in Figure 10. For mole
fractions of the adenine in solution from x ) 0 to x ) 0.8
specific fragments (Au + CH3T - and Au + (CH3T)2-) with a
mole fraction ø on the gold surface of 0.8-1 are found for the
CH3T. The mole fractions of the adenine thiolate specific
fragments (A- and A + EtCOO-) on the gold surface are
increasing slowly to a mole fraction ø of maximal 0.25 at x )
0.82. From x ) 0.82 to x ) 1 the mole fraction ø of the adenine
thiolate fragments on the gold increases to 1, while at the same
(48) Israelachvili, J. N. Intermolecular and surface forces; Academic
Press: London 1992.
(49) Pertsin, A. J.; Grunze, M. Langmuir 1994, 10, 3668.
(50) Whitesides, G. M.; Laibinis, P. E. Langmuir 1990, 6, 87.

94 J. Am. Chem. Soc., Vol. 122, No. 1, 2000
Weisser et al.
Table 3. Thickness of the Adsorbed Uridine or Cytosine Layer,
Respectively, Measured by Surface Plasmon Kinetic Scans before
and after Rinsing with Pure Water after 30 min of Adsorption
Timea
uridine layer
thickness [Å]
cytidine layer
thickness [Å]
before
rinsing
after
rinsing
before
rinsing
after
rinsing
pure gold surface
pure adenine thiolate
pure OHT
pure CH3T
adenine thiolate/OHT
adenine thiolate/CH3T
6
5
3
2
7
3
3
3
1
1
5
1
9
7
6
4
5
-
6
4
3
2
2
-
a The error is (1.5 Å. For the mixed monolayers of adenine thiolate
and OHT or CH3T a mole fraction of x ) 0.9 was chosen.
Figure 12. SDMS spectra of pure adenine thiolate SAM (a) without
and (b) with uridine immersion.
uridine. The binary mixed monolayers of adenine thiolate and
CH3T do not show the peak of the complexed base pair,
independently on the concentration of the adenine in the
monolayer.
Also with respect to the fragments of the uridine with m/z )
111 and 243 u the monolayers show clear differences. Pure
adenine thiolate monolayers and binary mixtures with OHT that
were rinsed with uridine depict both of these mass peaks in the
mass spectra, whereas mass spectra of binary mixtures of
adenine thiolate and CH3T, as well as those of the pure
monolayers of OHT and CH3T all rinsed in uridine solution,
do not show any of these mass peaks.
The adsorption kinetics of uridine and cytosine were followed
for a pure gold surface and a series of pure and mixed
monolayers with PSP spectroscopy. At a given time the water
in the cuvette is exchanged by a uridine- or cytidine-containing
solution. After about half an hour the cuvette is rinsed again
with pure water. During the rinsing a part of the adsorbed
molecules is removed. The thickness results of these kinetic
are summarized in Table 3. The thickness after the adsorption
process and the thickness after the rinsing with pure water after
the adsorption are chosen. Both nucleotides show a strong
adsorption to pure gold surfaces. Thicknesses of up to 9 Å for
cytosine were found. In the overall comparison cytosine seems
to adsorb stronger to any surface than uridine. For all monolayers
the thickness of the adsorbed cytosine layer is larger than that
for the adsorbed uridine layer, with one exception: for the OHT-
diluted adenine thiolate monolayer the adsorbed uridine shows
a higher thickness.
Adsorption measurements of uracil and cytosine on adenine
thiolate/OHT mixed monolayers with x ) 0.9 performed with
the multimode waveguide-PSP-coupling technique yielded a
thickness of 6 ( 2 Å and a refractive index of n ) 1.53 ( 0.01
for uracil. The test for unspecific binding with cytosine leads
to a signal which is not significant within the noise of the
experiment.
the steric hindrance of the adenine headgroups prevent an
optimal packing. Every added diluting molecule (CH3T or OHT)
will bring an enormous enthalpic contribution (ca. 60kJ/mol)
to the monolayer by interdigitating with the adenine thio-
lates.39-41,45,48,49 This process will promote the build in of
diluting molecules until an average area per molecule of 36 Å2,
the needed space for an adenine thiolate, is reached. This surface
density can be achieved at a mixing ratio of 1:1 on the surface.
The results from the multimode waveguide-PSP-coupling
experiments fit perfectly within this picture. The average
geometrical thickness only increases when adenine thiolate is
built into the monolayer at high adenine disulfide concentrations
in solution. Then the thickness reaches a value of 22 Å which
is smaller than the comparable value of 28 Å calculated from
the surface plasmon measurements. The behavior for the
refractive index is pretty similar. As long as the OHT is the
dominant feature within the monolayer a refractive index n of
1.5 fits the data very well. At the high adenine thiolate
concentrations the average refractive index increases to 1.53.
This refractive index is an average over the alkylchain with n
) 1.5 and the conjugated adenine headgroup. Therefore, by
estimating a refractive index of 1.5 for the complete molecule
leads to a too high monolayer thickness (PSP).
Base Pairing
Measurements. For the base pairing self-assembled mono-
layers of all pure species and of various mixtures were prepared.
The specific and unspecific reaction of uridine as well as the
unspecific reaction of cytidine were tested. The mass spectrum
taken from a pure adenine thiolate monolayer which was rinsed
with a uracil solution shows an additional mass peak at m/z )
378 u compared to the mass spectrum of the pure adenine
thiolate monolayer (Figure 12). This mass represents the
deprotonated complex of adenine and uridine. The pure adenine
thiolate monolayer which was rinsed with a cytidine solution
does not show a corresponding peak of an adenine complexed
with the cytosine. A binding of cytosine could also not be
detected in any mass spectra of adenine containing monolayers
rinsed with cytosine, neither mixed with CH3T nor with OHT.
The mass spectra of the mixed monolayers of adenine thiolate
and CH3T or OHT are different from each other with respect
to the mass peak at m/z ) 378 u. The mixed monolayers of
adenine thiolate and OHT show even for very low adenine
concentrations within the monolayer a distinct adenine-uridine
complex peak. Nevertheless, the intensities of the peaks are too
low for a data analysis via eq 2 to quantify the amount of
Discussion
The central question, whether the uridine binds specifically
to the two-dimensional assembled adenine, can clearly be
answered with yes. The presence of the peak in the mass spectra
of the uridine-adenine complex can only exist when uridine is
binding specifically to the adenine at this water-solid interface.
The PSP and multimode waveguide-PSP-coupling technique
data confirm the base-pair reaction. The film thickness of bound
uracil after rinsing is higher than the thickness of the adsorbed
cytosine in the case of the x ) 0.85 adenine thiolate/OHT
monolayer. Nevertheless, there is big tendency for unspecific
(51) Strong, L.; Whitesides, G. M. Langmuir 1988, 4, 546.

A/U Base Pairing at the Water-Solid-Interface
J. Am. Chem. Soc., Vol. 122, No. 1, 2000 95
binding both for uridine and cytosine. On all samples under
investigation (Table 3) both nucleotides adsorb even at surfaces
without any binding sites (gold, CH3T, and OHT), but the mass
never shows a peak due to a complex formed from adenine
and cytosine. Thus, the cytosine adsorption must always be
unspecific. Only in the case where an optimal adenine density
within the monolayer is achieved (x ) 0.9) can the uridine
binding be considered to be specific. A dilution of the active
adenine thiolate is necessary to achieve a proper binding. The
adenine-uracil complex mass peak and the two other specific
uracil fragments are indeed present for a pure adenine thiolate
and an adenine thiolate/OHT monolayer, but the uracil thickness
measured by PSP indicates a specific binding only for the
optimized mixed adenine thiolate/OHT monolayer.
It is totally unclear why the mixed adenine thiolate/CH3T
monolayers do not show a binding. In the mass spectra no
adenine-uracil complex or uracil fragments could be detected,
and the PSP measurements show only very small thicknesses
for uridine and no adsorption at all for cytosine at the mixed
adenine thiolate/CH3T monolayer. For the pure CH3T monolayer
unspecific binding is found mainly for cytidine. There might
be an interaction between the hydrophobic headgroups of the
CH3T and the adenine heads in the active thiolate which hinder
the adsorption of the uracil.
The measured refractive index for uracil is higher than usually
estimated. The nucleosides consist of conjugated units and
therefore have a higher polarizability compared to alkyl chains
leading to an enhanced refractive index.
Conclusions
We have shown the preparation and characterization of
immobilized monolayers containing adenine in a two-dimen-
sional geometry as a solid-water interface. Via self-assembly
of disulfides and thiols onto gold out of pure or binary mixed
solutions the adenine content within the adenine-containing
monolayer could be adjusted. The specific base-pair reaction
between this two-dimensional immobilized adenine and uracil
could be shown by mass spectroscopy and optical methods.
Unspecific binding both of uracil and cytosine was found, too.
This unspecific binding can lead to problems, if such a system
is used in a sensor device.
JA990447B