Adrenaline recognition in water

Article abstract of DOI:10.1002/chem.200400100

Host molecule 1 displays a high affinity in water towards catecholamines and especially related structures such as β-blockers with extended aromatic π-faces (up to 7 × 10³M^-1 for each single complexation step or 5 × 10⁷ M<

Full text of DOI:10.1002/chem.200400100

FULL PAPER
Adrenaline Recognition in Water
Oliver Molt, Daniel R¸beling, Gerhard Sch‰fer, and Thomas Schrader*[a]
Abstract: Host molecule 1 displays a
high affinity in water towards catechol-
amines and especially related struc-
tures such as b-blockers with extended
aromatic p-faces (up to 7î103 mÀ1 for
each single complexation step or 5î
107 mÀ2 for both steps). The amphiphilic
structural design leads to an extensive
self-association of host molecules
through their aromatic flanks. Above a
cmc (critical micelle concentration) of
3î10À4 m, host 1 forms micelles that
produce a favorable microenvironment
for hydrophobic interactions with the
included guest molecules. Electrostatic
attraction of the ammonium alcohol by
the phosphonate anions is thus com-
bined with hydrophobic contributions
between the aromatic moieties. Ionic
hydrogen bonds with polar OH or NH
groups of the guest enforce the non-co-
valent interactions, and finally lead to
increased specificity. Both its affinity
and its selectivity towards adrenergic
receptor substrates are greatly en-
hanced if the receptor molecule 1 is
transferred from water into a lipid
monolayer. Catecholamines and b-
blockers lead to drastically different ef-
fects at concentrations approaching the
micromolar regime. Especially b-block-
ers with minute structural changes can
be easily distinguished from each
other. In both cases, extensive hydro-
phobic interactions with a self-associat-
ed and/or self-organized microenviron-
ment are largely responsible for the ob-
served high efficiency and specificity.
Keywords: adrenergic receptor
¥
beta-blockers ¥ hydrophobic effect ¥
molecular recognition ¥ monolayers
Introduction
Results and Discussion
Adrenaline is the lead compound of a whole class of cate-
cholamine neurotransmitters and mediates signal transduc-
tion across cell membranes.[1] It is a small, highly polar mol-
ecule, which is bound very shortly and efficiently by its natu-
ral receptor. This recognition eventually leads to a confor-
mational change within the transmembrane helices of the
receptor and in turn triggers the activation of the G protein
on the cytosolic side of the membrane.[2] A deep binding
pocket is needed to provide a sufficiently hydrophobic envi-
ronment for complete desolvation of the charged hormone.
It is flanked by several polar amino acid residues, which are
specifically engaged in electrostatic and hydrogen bond in-
teractions with their polar guest.[3] Artificial receptor mole-
cules, which are designed to mimick this binding mode for
adrenaline without the 40 kD protein, must find a way to
create both the hydrophobic microenvironment and an
array of convergent binding sites for efficient interaction
with the guests× functional groups.
State of the art: Most artificial catecholamine receptor mol-
ecules are not biomimetic at all; recently bipyridinium/gold
nanoparticle arrays,[4] phenyl boronates,[5] and pyrazole-con-
taining cryptands have been developed.[6] Bioorganic ap-
proaches include RNA aptamers[7] or copper-containing
redox enzymes.[8] In the past years, our group has presented
several successive generations of adrenaline binders based
on a general recognition motif for amino alcohols, which
features xylylene bisphosphonate dianions.[9] The hydropho-
bic contribution comes mainly from macrocyclic cavities
with nonpolar walls, which carried the amino alcohol recog-
nition element at the bottom and a catechol-affinity moiety
at the top.[10] Thus, adrenaline derivatives and b-blockers
have been included in polar organic solvents, in one case
with 50% water.[11] However, never was biomimetic adrena-
line recognition achieved by bisphosphonates in pure water.
Recently, we found that flat receptor molecules for bisami-
dinium drugs with optimized proportions are able to carry
two guest molecules at the same time.[12] By extensive stack-
ing interactions, they maximize the hydrophobic attractions
and simultaneously create a good environment for strong
Coulomb interactions. The affinity for bisamidines is compa-
rable to that of DNA, which binds them in a 1:1 complex of
Ka =106 mÀ1 in its minor groove. It appears that nature often
uses this effective combination of powerful salt bridges and
[a] Dr. O. Molt, D. R¸beling, G. Sch‰fer, Prof. Dr. T. Schrader
Department of Chemistry, Philipps-Universit‰t Marburg
Hans-Meerwein-Str., 35032Marburg (Germany)
Fax : (+49)6421-28-28917
E-mail: schradet@mailer.uni-marburg.de
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2004, 10, 4225 4232
DOI: 10.1002/chem.200400100
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4225

FULL PAPER
hydrophobic forces for efficient recognition of small highly
polar substrates (e.g., sugar-binding proteins).[13] Would it be
possible to bind two adrenaline molecules by a bisphospho-
nate-based receptor? We decided to optimize the catechole
recognition elements and to improve the pre-organization of
our macrocyclic receptor cavities. This was indeed realized
by the use of tolane spacers for the walls,[14] and by the in-
troduction of a second bisphosphonate moiety in the top
part instead of the former aromatic diamide. It stretches the
whole macrocyle by repulsion of its four negatively-charged
phosphonate anions, and ensures superior interactions with
the catechol by ionic hydrogen bonds.[15]
with a well defined geometry. Each bisphosphonate moiety
receives the amino alcohol of one guest in a chelate fashion,
and additionally forms hydrogen bonds with the catechole
of the other guest. If the second guest molecule was orient-
ed parallel to the first one at the beginning of the calcula-
tion, it would always rotate inside the complex, until it final-
ly reached the antiparallel orientation; this ternary complex
was reproducibly found as that unique conformation, which
was by far the lowest in energy (Figure 2). Subsequent mo-
lecular dynamics calculations came to the same result, more-
over, host and guest are still mobile enough in the complex
for a favorable entropy balance.[17]
Modeling: Modeling experiments confirmed the expected
stiffness of the macrocyclic skeleton; the host alone is calcu-
lated in the shape of a twisted rectangle with linear, stiff
sidewalls (Figure 1). In a subsequent molecular dynamics
calculation only the bridging p-xylylene bisphosphonates
showed some mobility. The four phosphonate monoanions
are all pointing towards the solvent, away from each other.
In order to receive guest molecules, only a small amount of
energy is needed to produce an open conformation.[16]
The results from Monte-Carlo simulations for such com-
plexes with adrenaline were especially intriguing; the 2:1
complex was indeed calculated to be the most stable one,
Figure 2. The new receptor molecule 1 with two bisphosphonate moieties
and the result of a Monte-Carlo simulation of the 2:1 binding mode with
noradrenaline (MacroModel 7.2, Amber*, 3000 steps, water). Top right:
subsequent molecular dynamics calculation (10 ps, 258C).
Synthesis: A modular approach reduces the synthesis to the
alternating connection of two building blocks. The tolane
sidewall can be prepared by two successive Sonogashira cou-
plings,[18] whereas the phosphonate-modified p-xylylene di-
bromide is accessible by sequentiel NBS (N-bromosuccini-
mide)-bromination and the Arbuzov reaction. Although the
rectangle can be synthesized in one step by a four compo-
nent reaction, yields remain low, because a byproduct, pre-
sumably the double-sized macrocycle, is also formed. We
therefore chose the esterification of two monoprotected
Figure 1. Monte-Carlo simulation (MacroModel 7.0, Amber*, water,
3000 steps) and subsequent molecular dynamics calculation of the free
host 1 in water (10 ps, 258C, no constraints).
4226
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4225 4232

Adrenaline Recognition in Water
4225 4232
tolane carboxylic acids with the xylylene dibromide head
group to afford the U-shaped precursor. After deprotection,
cyclization with Cs2CO3 (cesium effect) followed.[19] Final
mild dealkylation with LiBr furnishes the desired receptor
molecule in a 16% overall yield (Figure 3).[20] The lithium
tetraphosphonate is soluble in polar media such as methanol
and water.
Figure 4. Top: molecular mechanics calculation of a hexameric host ag-
gregate (MacroModel 7.2, water, 3000 steps). Note the similarity to lipid
bilayers with polar headgroups and nonpolar vertical ethinylaryl chains.
Bottom: surface pressure p [mNmÀ1] plotted against surfactant concen-
tration cS [M] at 258C.
muir film balance prove the formation of micelles above a
cmc (critical micelle concentration) of 3î10À4 m (Figure 4
top). Evidently, host 1 lowers the surface pressure or in-
creases the surface tension. We tentatively explain this un-
usual behaviour by the concentration of negatively-charged
phosphonates at the air/water interface, which leads to an
ion-pair reinforced hydrogen-bond network.
Figure 3. Sequential synthesis of macrocycle 1 from xylylene bisphospho-
nate 2 and tolane building blocks 3.
Self association: Contrary to expectations, the NMR signals
remained sharp only in methanolic solution, and pure water
produced extremely broad ™mountains∫. Although a rela-
tively low value of 130mÀ1 was determined for the self-asso-
ciation constant by dilution titration, the aggregates are
thermally stable up to 908C. Since shifts occurred mainly in
the tolane region, the host molecules probably aggregate
through their stiff unpolar aromatic sidewalls forming vesi-
cle-like structures with a hydrophobic interior. The results
from modeling experiments suggest a relatively dense pack-
ing of receptor molecules in their thermodynamically most
stable twisted form, with close contacts between their tolane
sidewalls. In these structures, hydrophobic cavities between
neighbouring host molecules offer enough room for the po-
tential insertion of guest molecules, which are too large to
be accomodated within the macrocycle×s cavity. These inter-
molecular cavities may explain the extraordinary affinity of
the self-associated receptor molecule 1 for unpolar b-block-
ers (see Figure 4). Surface tension measurements on a Lang-
Binding experiments: Addition of catecholamines to the
new receptor molecule resulted in large upfield shifts of the
guest molecules, especially in the aromatic region. Job plots
indicate a clean 2:1-stoichiometry, as expected from the
modeling studies (see Figure 5 and Table 1).[21]
We performed NMR titrations with increasing amounts of
host compound and obtained smooth binding curves.[22]
They allowed an excellent fit in accordance with the deter-
mined complex stoichiometry by nonlinear regression, with
standard deviations in the range of 2 10%. No cooperativity
could be found in any case, that is, both guest molecules are
bound independent of each other with exactly the same
binding constant. Therefore, for ease of comparison we
always use the 1:1 association constants for each single bind-
ing step in [mÀ1] (Table 2); the overall 2:1 association con-
stants are also given in [mÀ2]. When we increased the solvent
polarity from pure methanol over methanol/water (1:1) to
pure water, a remarkable dependence of the 1:1 binding
Chem. Eur. J. 2004, 10, 4225 4232
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4227

T. Schrader et al.
FULL PAPER
Table 1. ESI-MS data for the complex formation betwen host 1 and nor-
adrenaline 6 in [D4]MeOH. Peak values shown for 1:1 and 2:1 complex
wih noradrenaline (from methanol).
Species
m/z Found Assignment
m/z Calculated
free host 1
583.1126
586.1119
589.1172[
[14À +2H+]2À
583.0923
586.0964
589.1005
667.6293
673.6374
676.6415
758.1744
764.1826
[14À +H+ +Li+]2À
14À +2Li+]2À
1:1 complex 667.6610
673.6583
[14À +NA+ +H+]2À
]
[14À +NA+ +2Li+ÀH+2À
]
676.6658
[14À +NA+ +3Li+À2H+2À
]
2:1 complex 758.2089
764.2065
[ 14À +2NA+ +2Li+À2H+2À
]
]
[14À +2NA+ +4Li+À4H+2À
catecholamine molecules at the same time. Well-defined
host guest complexes with aggregated receptor species have
been observed earlier with macrocyclic amphiphilic host
compounds.[24]
The opposite orientation of both guest molecules inside
the tight complex is strongly supported by NOESY (Nuclear
Overhauser and Exchange Spectroscopy) measurements.
Contrary to the NOE (Nuclear Overhauser Effect) effects
in noradrenaline, those of the host guest complex are all
positive, which indicates a decelerated rotation.[25] One addi-
tional NOE occurs between proton a and e of the guest.
Since the distance is much too large for an intramolecular
NOE (4.5 ä), this must be an intermolecular crosspeak be-
tween the two bound guest molecules. A short distance of
2.5 ä is indeed found in the 2:1 complex structure, but only
if both guest molecules are oriented antiparallel to each
other (Figure 6). Another indication for the formation of
stable 2:1 complexes is found in the ESI-MS (electrospray
ionisation mass spectrometry) spectrum, which produces a
strong molecular ion peak for this preferred stoichiometry,
in addition to a strong peak corresponding to the 1:1 com-
plex. Interestingly enough, the analysis of both complex ion
peak series reveals that in each guest, up to two protons are
replaced by lithium cations. These must be two acidic phe-
nolic hydrogen atoms, which will be bound to the phospho-
nates by strong lithium chelate salt bridges. Thus, experi-
Figure 5. Top: job plot for complex formation between host 1 and nor-
adrenaline 6 in [D4]MeOH (proton a: CH2-NH3Cl, proton b: -CH(OH)-);
bottom: ESI-MS spectrum for the same complexation with strong molec-
ular ion peaks for the free host, as well as its 1:1 and 2:1 complex with
noradrenaline (from methanol).
constant was revealed for noradrenaline 6. Whereas it
amounts to 4000mÀ1 in MeOD, it drops to ~700mÀ1 in
MeOD/D2O (1:1); however, transition to pure water again
leads to a marked increase up to 1200mÀ1! In the same
order, the NMR shifts of the
guest molecules are also falling
Table 2. Binding constants [Ka (1:1) and Ka (2:1)] as well as Ddsat values in complexes between host 1 and various
guest molecules by NMR titrations in D2O at 2 78C. For each guest, up to five independent CH proton signals
were evaluated. For details, see Supporting Information.
and rising, both in the aromatic
region (hydrophobic interac-
tions), and in the amino alcohol
moiety (electrostatic interac-
tions) (in each case from
[b]
[b]
No.
Guest molecules[a]
Ka (1:1) [MÀ1
]
Ka (2:1) [MÀ2
]
Ddsat [ppm][c]
Stoichiometry[d]
5
6
7
8
serotonin
noradrenaline
adrenaline
1620mÀ1 Æ4%
1250mÀ1 Æ6%
1230mÀ1 Æ4%
870mÀ1 Æ4%
530mÀ1 Æ25%
<1
2.6î106 mÀ2 Æ4%
1.5î106 mÀ2 Æ 6%
1.5î106 mÀ2 Æ 4%.
7.6î105 mÀ2 Æ 4%
2.8î105 mÀ2 Æ 25%
0.59
0.20
0.26
0.33
0.07
2:1
2:1
2:1
2:1
2:1
0.2ppm
in
MeOD
over
0.05 ppm in MeOD/D2O to
0.3 ppm in pure D2O).[23] We
explain this unusual behavior
with the observed aggregation
of host molecules, which takes
place only in pure water. Evi-
dently, under these circumstan-
ces the host creates an increas-
ingly hydrophobic microenvir-
onment for the approaching
guest, which facilitates the
dopamine
9
acetylcholine
ethanolamine
catechole
alprenolol
propranolol
atenolol
2-phenylethylamine
l-tyrosine methyl ester
glycine
GABA
d-glucose
10
11
12
13
14
15
16
17
18
19
<1
<1
<1
7100mÀ1 Æ 8%
4290mÀ1 Æ11%
830mÀ1 Æ8%
1500mÀ1 Æ2%
130mÀ1 Æ47%
<1
4.9î107 mÀ2 Æ8%
1.10
0.93
1.71
0.89
0.19
2:1
2:1
2:1
1:1
1:1
1.8¥107 mÀ2 Æ11%
6.7¥105 mÀ2 Æ8%
<1
<1
<1
<1
<1
[a] As hydrochloride salts. [b] Errors are calculated as standard deviations from the nonlinear regression.
docking/inclusion of even two [c] Largest shifts from selected CH protons. [d] From job plots and curve-fitting of the titration curves.
4228
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org Chem. Eur. J. 2004, 10, 4225 4232

Adrenaline Recognition in Water
4225 4232
among the most efficient binders kown to date.[27] However,
it is much more selective than most other synthetic recep-
tors (see below).
Selectivity: A total of 14 neurotransmitters or structurally
closely related guest molecules were titrated with the new
host. All catecholamines 6 8 gave an excellent 2:1 fit and
similar association constants in the range of 900 1300mÀ1.
The b-blockers alprenolol 12 and propranolol 13 bind espe-
cially tightly with 4000 and 7000mÀ1, respectively, probably
because the extended aromatic rest is capable of optimizing
its hydrophobic interactions with the nonpolar surrounding
in the aggregates (Figure 7). In this case, a terminal polar
Figure 7. Energy-minimized structure illustrating the efficient inclusion of
three alprenolol molecules into a tetrameric host aggregate (MacroModel
7.2, water, 1000 steps).
Figure 6. Top: schematic drawing of the complex with intermolecular
NOE crosspeaks between the two noradrenaline guest molecules and the
additional lithium salt bridges between catechole and phosphonates.
Bottom: the arrow indicates protons a and e of the noradrenaline guest,
which produce a crosspeak only in the complex. The intermolecular dis-
tance betweeen a and e is much smaller (circle) than the intramolecular
(arrow). All NOE×s in the complex are positive, strong and reciprocal.
group is detrimental for efficient binding, as atenolol only
reaches a moderate affinity towards 1. Binding experiments
with 1 in aqueous NaCl (10mm) demonstrate that electro-
static attraction is important, but strongly supported by hy-
drophobic interactions, and the Ka value for alprenolol de-
creases only slightly from 7100 to 2600mÀ1.
mental evidence has been found for the expected phospho-
nate catechol interaction in the complex.
For effective recognition, electrostatic interactions and
the hydrophobic effect are essential. Evidence for this is
provided by cutting the catcholamine skeleton into two
halves. Neither ethanolamine 10 (ion pairing) nor catechol
11 (hydrophobic attraction) alone show any affinity for the
new host. Likewise, the amino acid neurotransmitters gly-
cine 17 and GABA 18 are not bound by the receptor mole-
cule. Even with an ester-protected carboxylate and free am-
monium cation complexes with amino acids (e.g., 16) reach
Ka values of only one tenth of those for catecholamines. Ac-
cordingly, the new host requires a guest structure with a slim
amino alcohol on one end and an aromatic group on the
The natural adrenergic receptor (as its synthetic models)
binds only one adrenaline guest molecule at a time. A host
capable of binding two guest molecules simultaneously has a
much higher efficiency. The relative amount of bound nor-
adrenaline in a 2:1 complex with Ka =1200mÀ1 corresponds
to that in a 1:1 complex with Ka =2500mÀ1; this is true at
least up to a host/guest ratio of 1:1. For alprenolol with Ka =
7000mÀ1 in the 2:1 complex this even compares with a Ka
value of ~14000mÀ1 in the 1:1 complex.[26] The affinity of
our new host for adrenaline derivatives in water places it
Chem. Eur. J. 2004, 10, 4225 4232
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4229

T. Schrader et al.
FULL PAPER
other end, preferably with polar substituents. These condi-
tions are also met by serotonin 5, which again binds very
well. Polar substituents with hydrophobic faces alone are
not sufficient for complexation: no binding occurs with free
glucose, although it fits into the cavity of 1. Small aromatic
amines such as phenethylamine 15 are recognized, but only
in the form of a 1:1 complex. Large upfield shifts (0.9 ppm)
clearly indicate their complete inclusion in the host×s cavity
(See Scheme 1).
per stearic acid, a highly sensitive doped monolayer evolves,
which distinguishes not only between catecholamines and b-
blockers as such, but also between structurally related com-
pounds of the same class.[33] Moderate increases in the p-A
isotherm [1 3 ä2/molecule] are characteristic of catechol-
amine binding; this is similar to observations from macrocy-
clic and tweezer-type bisphosphonate receptors examined
before (Figure 8, Table 3). Negative controls revealed that
In summary we have found a host molecule that binds
catecholamines and especially related structures, such as b-
blockers with extended aromatic p-faces with high affinity
in water. A combination of electrostatic attraction of the
ammonium alcohol and hydrophobic contributions in the ar-
omatic moiety are essential. Ionic hydrogen bonds with
polar OH or NH groups of the guest enforce the non-cova-
lent interactions and finally lead to increased specificity.[28]
Interestingly, the aggregation of host molecules through
their aromatic flanks seems to produce a favorable microen-
vironment for hydrophobic interactions with the included
guest molecules. These observations prompted us to incor-
porate the new receptor molecule into a monolayer of stea-
ric acid at the air/water interface.[29]
Langmuir film balance: Despite its highly charged tetra-
anionic nature, the new host molecule is amphiphilic enough
to produce a marked linear increase in the pressure/area (p-
A) diagram of stearic acid on a Langmuir film balance; this
indicates the embedding process.[30] Contrary to former ex-
periments with a less polar macrocyclic bisphosphonate, the
picture of a connected Brewster angle microscope remained
smooth without any appearance of patches.[31] Large arrays
or domains of the self-aggregating host molecules seem to
be avoided, probably because of self-repulsion among the
surrounding negative charges. Latest bioanalytical results
show that even the natural adrenergic receptors only form
dimeric structures.[32] With only a 0.4 equivalent of receptor
Figure 8. Top: pressure-area-isotherms of stearic acid (S) and receptor 1
in monolayer over water with noradrenaline (NA), adrenaline (Adr) and
dopamine (Dop). Reference: pure stearic acid over noradrenaline.
Bottom: pressure-area-isotherms of stearic acid (S) and receptor 1 in
monolayer over water, with alprenolol (Alp), atenolol (Aten) and pro-
pranolol (Prop). Reference: pure stearic acid over propranolol.
Scheme 1. Guest molecules for host 1, tested in solution and at the air/water interface.
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4230
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4225 4232

Adrenaline Recognition in Water
4225 4232
must clearly surpass 105 mÀ1 and might be used for new b-
blocker sensoring devices.[34] Thus, the new microenviron-
ment in the monolayer leads to high selectivity for minute
structural changes in the analytes.[35]A complete list of
changes in the p/A diagrams induced by various related
guest molecules is summarized in Table 3 (see also Figure 9.
Table 3. A0 and DA0 values of stearic acid monolayers with embedded re-
ceptor 1 over different subphases. A0 =apparent total area of one stearic
acid molecule in the liquid condensed phase; DA0 =A0 (monolayer with 1
over guest soln.) A0 (monolayer with 1 over water), that is, net influence
of the guest. Monolayer: 0.2equiv. of receptor molecule 1 per stearic
acid molecule; guest molecules at 10À4 m.
Monolayer
Subphase
A0
DA0
[ä2/molecule] [ä2/molecule]
stearic acid
Water
20.5
stearic acid + 1 Water
stearic acid noradrenaline
(~catecholamines)
stearic acid + 1 (R/S)-noradrenaline
stearic acid + 1 (R/S)-adrenaline
stearic acid + 1 dopamine
21.5
21.5
1
0
23.5
22.5
24.5
2
1
3
stearic acid
propranolol
(~b-blockers)
27.5
23.5
29.0
32.5
0
stearic acid + 1 (R/S)-alprenolol
stearic acid + 1 (R/S)-atenolol
stearic acid + 1 (R)-propranolol
À4
1.5
5
Figure 9. Proposed binding modes for catecholamines and b-blockers
within the monolayer.
in all cases no interaction takes place between the guest
molecules and the stearic acid monolayer itself. By contrast,
all b-blockers examined in our study drastically expanded
the pure monolayer by ~6 ä2 per molecule. Subinjection of
the antihypertensives produced very distinct additional
changes in the p-A diagram, which demonstrate the exqui-
site selectivity and high affinity of the immobilized receptor
molecules for structurally related b-blockers. Atenolol, pro-
pranolol, and alprenolol differ only in the substitution pat-
tern of their terminal phenoxy aromatic group. Atenolol
leads to a moderate increase in p-A, whereas propranolol
dramatically expands the monolayer especially in the liquid
phase [=5 ä2/molecule]. Alprenolol finally leads to a strong
negative p-A shift of 4 ä2/molecule.
We explain this divergent behavior with a model, devel-
oped earlier for amphiphilic water-soluble host molecules.
Since guest molecules are subinjected into the aqueous
phase, they are bound by solvated receptor molecules close
to the monolayer. The host molecules× negative charges
become neutralized in part and their lipophilicity increases.
This in turn leads to reincorporation of the whole complex
into the monolayer and explains the moderate increases ob-
served with the three catecholamines; these correspond to
their similar binding constants in water (Table 2). b-Blockers
lack the polar catechol moiety leading to a nonpolar aro-
matic headgroup; this makes them much more amphiphilic
and explains their tendency to insert the hydrophobic heads
into the lipid monolayer. Incorporation of the receptor mol-
ecule into the monolayer leads to a moderate additional in-
crease in p-A, which corresponds to its moderate binding
constant. Propranolol, on the other hand, provides an ex-
panded p-face for more efficient stacking interactions with
the receptor, and seems to form loose hydrophobic aggre-
gates in the lower pressure region, which are only dissolved
as the pressure increases. The smaller and compact alprenol-
ol finally binds so tightly to the host, that its complex is
drawn from the monolayer back into the aqueous subphase.
Since large p-A changes can still be detected far below
10À4 mÀ1, a rough estimation of the respective binding con-
stants at the air/water interface leads to Ka values which
Conclusion
We conclude that the transfer of receptor molecule 1 from
water to the altered microenvironment within a monolayer
greatly enhances both its affinity and its selectivity towards
adrenergic receptor substrates. Catecholamines and b-block-
ers lead to drastically different effects at concentrations ap-
proaching the micromolar regime. Especially b-blockers
with minute structural changes can be easily distinguished
from each other. In both cases, extensive hydrophobic inter-
actions with a self-associated and/or self-organized microen-
vironment are largely responsible for the observed high effi-
ciency and specificity. In the future we intend to incorporate
the new macrocyclic receptor molecules into alternating
layers of receptor and ammonium-stabilized gold nanoparti-
cles attached to an ITO (indium tin oxide) electrode for
electrochemical detection of catecholamine derivatives and
b-blockers (Figure 10).[36]
Figure 10. Schematic illustration of the new proposed array of alternating
layers composed of receptor molecules 1 and cationic gold nanoparticles
on ITO.
Chem. Eur. J. 2004, 10, 4225 4232
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4231

T. Schrader et al.
FULL PAPER
[20] a) R. Karaman, A. Goldblum, E, Breuer, H. Leader, J. Chem. Soc.
Perkin Trans. 1 1989, 765; b) E. Breuer, M. Safadi, M. Chorev, D.
Gibson, J. Org. Chem. 1990, 55, 6147.
[21] a) P. Job, Compt. Rend. 1925, 180, 928; b) M. T. Blanda, J. H.
Horner, M. Newcomb, J. Org. Chem. 1989, 54, 4626.
[1] T. P. Iismaa, T. J. Biden, J. Shine, G Protein-Coupled Receptors,
Springer, Heidelberg, 1995.
[2] Recent GPCR crystal structures: D. Doyle, J. M. Cabral, R. A.
Pfuetzner, A. Kuo, J. M. Gulbis, S. L. Cohen, B. T. Chait, R. Mackin-
non, Science 1998, 280, 69; K. Palczewski, T. Kumasaka, T. Hori,
C. A. Behnke, H. Motoshima, B. A. Fox, I. Le Trong, D. C. Teller, T.
Okada, R. O. Stenkamp, M. Yamamoto, M Miyano, Science 2000,
289, 739 745; R. E. Stenkamp, S. Filipek, C. A. G. G. Driessen,
D. C. Teller, K. Palczewski, Biochim. Biophys. Acta 2002, 1565, 168.
[3] S. Trumpp-Kallmeyer, J. Hoflack, A. Bruinfels, M. Hibert, J. Med.
Chem. 1992, 35, 3448; K. Wieland, H. M. Zuurmond, C. Krasel,
A. P. Ijzerman, M. J. Lohse, Proc. Natl. Acad. Sci. USA 1996, 93,
9276.
[4] M. Lahav, A. N. Shipway, I. Willner, J. Chem. Soc. Perkin Trans. 2
1999, 1925.
[5] S. M. Strawbridge, S. J. Green, J. H. R. Tucker, Chem. Commun.
2000, 2393 2394; b) S. M. Strawbridge, S. J. Green, J. H. R. Tucker,
Phys. Chem. Chem. Phys. 2000, 2, 2393 2394.
[6] L. Lamarque, P. Navarro, C. Miranda, V. J. Arµn, C. Ochoa, F. Es-
carti, E. GarcÌa-EspaÊa, J. Latorre, S. V. Luis, J. F. Miravet, J. Am.
Chem. Soc. 2001, 123, 10560 10570.
[7] C. Mannironi, A. Di Nardo, P. Fruscoloni, G. P. Tocchini-Valentini,
Biochemistry 1997, 36, 9726.
[8] D. R. Shankaran, M.-S. Won, Y.-B. Shim, Anal. Sci. 2001, 17, 9; M.
Del Pilar Taboada Sotomayor, A. A. Tanaka, L. T. Kubota, Anal.
Chim. Acta 2002, 455, 215.
[9] T. Schrader, Angew. Chem. 1996, 108, 2816 2818; Angew. Chem.
Int. Ed. 1996, 35, 2649 2651; T. Schrader, J. Org. Chem. 1998, 63,
264 272.
[22] C. S. Wilcox, in Frontiers in Supramolecular Chemistry (Ed.: H. J.
Schneider), Chemie, Weinheim, 1991, p. 123. Usually the concentra-
tion of the guests varied between 0.38mm (first sample, free guest)
and 0.30mm (last sample, guest with ~9-fold excess of host). The
concentration dependence of all chemical shifts in our guest mole-
cules was carefully checked as independent measurements of the
free guests. In all cases, the chemical shift deviation within the titra-
tion×s concentration range was smaller than 0.002ppm for every
single 1H NMR signal. This dilution-induced uncertainty represents
a fraction of the observed CIS of less than 1% in the worst case.
Thus, it is safe to say, that dilution effects are negligible for the
NMR titrations within this study.
[23] Many earlier studies, for example, by Schneider or Diederich et al.,
show the expected smooth in- and decrease with the solvent compo-
sition: H.-J. Schneider, Angew. Chem. 1991, 103, 1419 1439; Angew.
Chem. Int. Ed. Engl. 1991, 30, 1417 1436; U. Meyer, Coord. Chem.
Rev. 1976, 21, 159.
[24] a) S. Shinkai, S. Mori, H. Koreishi, T. Tsubaki, O. Manabe, J. Am.
Chem. Soc. 1986, 108, 2409; b) Y. Murakami, J. Kikuchi, T. Ohno, O.
Hayashida, M. Kojima, J. Am. Chem. Soc. 1990, 112, 7672.
[25] H. Mo, T. C. Pochapsky, Prog. Nucl. Magn. Reson. Spectrosc. 1997,
33, 1. In most artificial complexes the NOEs are either weak or neg-
ative.
[26] We thank Dr. Stefan Kubik, D¸sseldorf, for providing us the evalua-
tion software to calculate the degree of complexation in 1:1 versus
2:1 complexes.
[10] T. Schrader, M. Herm, Chem. Eur. J. 2000, 6, 47 53.
[11] T. Schrader, M. Herm, O. Molt, Angew. Chem. 2001, 113, 3244
3248; Angew. Chem. Int. Ed. 2001, 40, 3148 3151 ; T. Schrader, M.
Herm, O. Molt, Chem. Eur. J. 2002, 8, 1485 1499.
[27] see ref. [4 8]. Other efficient catecholamine binders: J.-P. Behr, J.-
M. Lehn, P. Vierling, Helv. Chim. Acta 1982, 65, 1853 1867; B. Es-
cuder, A. E. Rowan, M. C. Feiters, R. J. M. Nolte, Tetrahedron 2004,
60, 291 300.
[12] T. Schrader, T. Grawe, G. Sch‰fer, Org. Lett. 2003, 5, 1641 1644.
[13] Sugar binding proteins: F. A. Quiocho, Pure Appl. Chem. 1989, 61,
1293; N. Sharon, Trends Biochem. Sci. 1993, 18, 221; c) W. I. Weis,
K. Drickhamer, Annu. Rev. Biochem. 1996, 65, 441; see also ref. [3]
(adrenergic receptor); NAD+ recognition in the Rossman fold: M.
Otagiri, G. Kurisu, S. Ui, Y. Takusagawa, M. Ohkuma, T. Kudo, M.
Kusunoki, J. Biochem. 2001, 129, 205 208.
[14] a) T. Kawase, Y. Seirai, H. R. Darabi, M. Oda, Y. Sarakai, K. Ta-
shiro, Angew. Chem. 2003, 115, 1659 1662; Angew. Chem. Int. Ed.
2003, 42, 1621 1624; b) M. M. Haley, J. J. Pak, S. C. Brand, Top.
Curr. Chem. 1999, 201, 82 130.
[28] These groups are present in the catechole and amino alcohol hy-
droxyls, as well as in tryptophane×s pyrrol. They are all fully proto-
nated at neutral pH.
[29] For experimental details, see Supporting Information.
[30] a) M. Higuchi, T. Koga, K. Taguchi, T. Kinoshita, Langmuir 2002,
18, 813; b) D. Y. Sasaki, T. A. Waggoner, J. A. Last, T. M. Alam,
Langmuir 2002, 18, 3714; c) J. Lahiri, G. D. Fate, S. B. Ungashe, J. T.
Groves, J. Am. Chem. Soc. 1996, 118, 2347.
[31] J. Meunier, Colloids Surf. A 2000, 171, 33.
[32] C. Lavoie, J.-F. Mercier, A. Salahpour, D. Umapathy, A. Breit, L.-R.
Villeneuve, W.-Z. Zhu, R.-P. Xiao, E. G. Lakatta, M. Bouvier, T. E.
Hebert, J. Biol. Chem. 2002, 277, 35402; J. Xu, M. Paquet, A. G.
Lau, J. D. Wood, C. A. Ross, R. A. Hall, J. Biol. Chem. 2001, 276,
41310.
[15] Catechole recognition by phosphonates: G. Das, A. D. Hamilton, J.
Am. Chem. Soc. 1994, 116, 11139; C. E. Park, Y.-G. Jung, J.-I. Hong,
Tetrahedron Lett. 1998, 39, 2353.
[33] Synthetic receptors have also been incorporated into self-assembled
monolayers of alkanethiols on gold (SAM×s) and displayed high se-
lectivities in molecular recognition of various guests dissolved in or-
ganic solvents: K. Motesharei, D. C. Myles, J. Am. Chem. Soc. 1998,
120, 7328 7336.
[16] Minimizations (1000 steps), followed by Monte-Carlo simulations
(2000 steps) and subsequent MD calculations (10 ps): MacroModel
7.2, Amber*, water. Solvents are treated as an analytical continuum
starting near the van der Waals surface of the solute (GB/SA model:
W. C. Still, A. Tempczyk, R. C. Hawley, T. Hendrickson, J. Am.
Chem. Soc. 1990, 112, 6127 6129).
[17] Restricted mobility of a guest inside a very tight host macrocycle se-
verely lowers its entropy content, thus good artificial hosts offer
enough space for some rotatory freedom of the bound guest.
[18] S. Takahashi, Y. Kuroyama, K. Sonogashira, N. Hagihara, Synthesis
1980, 627 630.
[34] Other sensoring devices for adrenaline derivatives reach comparable
detection limits of ~10À6 m see ref. [4 8].
[35] A simple lipid-modified bisphosphonate embedded in the stearic
acid monolayer did not produce any changes in the p/A diagram
even at high guest concentrations in the aqueous subphase.
[36] A. N. Shipway, E. Katz, I. Willner, ChemPhysChem 2000, 1, 18.
[19] A. Ostrowicki, E. Koepp, F. Vˆgtle, Top. Curr. Chem. 1992, 161, 37
67.
Received: February 2, 2004
Published online: July 12, 2004
4232
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4225 4232