Toward synthetic adrenaline receptors: Strong, selective, and biomimetic recognition of biologically active amino alcohols by bisphosphonate receptor molecules

Article abstract of DOI:10.1021/jo971297v

Xylylene bisphosphonates represent a new class of artificial receptor molecules for alkylammonium ions (Schrader, T. Angew. Chem., Int. Ed. Engl. 1996, 35, 2649-2651). Molecular recognition takes place in a 1:1 chelate- binding mode, and an almost ideal array of short, linear hydrogen bonds is created that guarantees maximum electrostatic and hydrogen-bond interactions. The host molecule, which was designed to imitate the natural adrenergic receptor, is selective for 1,2- and 1,3-amino alcohols due to formation of an additional cooperative hydrogen bond between the phosphonate anion and the hydroxyl groups. Biologically important amino alcohols such as glucosamine, 1-aminosorbitol, ephedrine, and the β-blocker propranolol are bound in DMSO with K(a) values between 60 000 and 130 000 M^-1. Secondary amines are complexed at least as strongly as their primary counterparts. The phosphonate ester groups allow lateral recognition of the substate. This could be demonstrated for adrenaline model compounds that were recognized by phosphonates carrying extended aromatic ester groups for Π,Π-interactions.

Full text of DOI:10.1021/jo971297v

264
J . Org. Chem. 1998, 63, 264-272
Tow a r d Syn th etic Ad r en a lin e Recep tor s: Str on g, Selective, a n d
Biom im etic Recogn ition of Biologica lly Active Am in o Alcoh ols by
Bisp h osp h on a te Recep tor Molecu les†
Thomas Schrader*
Department of Organic Chemistry and Macromolecular Chemistry, Heinrich Heine-Universita¨t
Du¨sseldorf, Universita¨tsstrasse 1, D-40225 Du¨sseldorf, Federal Republic of Germany
Received J uly 15, 1997X
Xylylene bisphosphonates represent a new class of artificial receptor molecules for alkylammonium
ions (Schrader, T. Angew. Chem., Int. Ed. Engl. 1996, 35, 2649-2651). Molecular recognition takes
place in a 1:1 chelate-binding mode, and an almost ideal array of short, linear hydrogen bonds is
created that guarantees maximum electrostatic and hydrogen-bond interactions. The host molecule,
which was designed to imitate the natural adrenergic receptor, is selective for 1,2- and 1,3-amino
alcohols due to formation of an additional cooperative hydrogen bond between the phosphonate
anion and the hydroxyl groups. Biologically important amino alcohols such as glucosamine,
1-aminosorbitol, ephedrine, and the â-blocker propranolol are bound in DMSO with Ka values
between 60 000 and 130 000 M-1. Secondary amines are complexed at least as strongly as their
primary counterparts. The phosphonate ester groups allow lateral recognition of the substate.
This could be demonstrated for adrenaline model compounds that were recognized by phosphonates
carrying extended aromatic ester groups for π,π-interactions.
Sch em e 1. Non cova len t In ter a ction s betw een
In tr od u ction
Nor a d r en a lin e a n d th e â-Ad r en er gic Recep tor
During the past decade, enormous research activities
have been directed toward a better understanding of
structure and function of the adrenergic receptor family,2
notably when in 1984 the role of the G-proteins was
discovered.3 Every year, more than 3000 publications
appear dealing exclusively with biochemical and medici-
nal aspects of research on adrenaline and noradrenaline.
This highlights the great importance of adrenaline for
signal transfer and the related control of biological
processes in the human body. Many pharmaceuticals act
as agonists or antagonists for the â-adrenergic receptor,
e.g., the well-known â-blockers.4 However, this receptor
located in the membrane of adrenal gland cells has never
been isolated in a pure form and has hence proven elusive
to X-ray structure analysis.5 The sequence of several
human subtypes is known and, on the basis of site-
directed mutagenesis experiments as well as on molec-
ular modeling studies, some very similar proposals on
the mechanism of neurotransmitter binding have evolved.6
The main noncovalent interactions are summarized in
Scheme 1. Electrostatic and hydrogen-bond interactions
with an aspartate carboxylate bind the ammonium
+
functionality; in addition, the NH3 protons experience
π-cation stabilization because they are surrounded by
three electron-rich aromatic residues. Each of the ali-
phatic and catechol hydroxyls is hydrogen-bonded to a
serine OH, while the aromatic ring is buried in a deep
cleft flanked by two phenylalanine aromatic rings that
are involved in double π-stacking. Although several
structures have been synthesized that bind catechola-
mine neurotransmitters,7 most of them recognize only one
single characteristic part of the molecule by utilizing
interactions that are far from biomimetic.8
It would be of great value for biochemists as well as
pharmacologists to gain access to a small defined model
of the adrenergic receptor that imitates the natural
receptor-ligand interactions. To this end, we designed
a new receptor structure for biomimetic recognition of
* To whom correspondence should be addressed. Fax: Int. + 211/
811-4788. E-mail: schrat@iris-oc2.oc2.uni-duesseldorf.de.
† Synthetic Adrenaline Receptors. 2. For part 1, see ref 1.
X Abstract published in Advance ACS Abstracts, December 1, 1997.
(1) Schrader, T. Angew. Chem., Int. Ed. Engl. 1996, 35, 2649-2651.
(2) Strader, C. D.; Fong, T. M.; Tota, M. R.; Underwood, D. Annu.
Rev. Biochem. 1994, 63, 101-132.
(3) Gilman, A. G. Cell 1984, 36, 577-579. Nobel lecture: Angew.
Chem., Int. Ed. Engl. 1995, 34, 1406.
(4) Craig, C. R.; Stitzel, R. E. Modern Pharmacology, Little, Brown
& Co.: Boston, 1990.
(5) Recently, electron crystallography has been carried out on two-
dimensional rhodopsin crystals: Schertler, G. F. X.; Villa, C.; Hen-
derson, R. Nature 1993, 362, 770-772.
(6) Mutagenesis experiments: Ostrowski, J .; Kjelsberg, M. A.;
Caron, M. C.; Lefkowitz, R. J . Annu. Rev. Pharmacol. Toxicol. 1992,
32, 167-183. Molecular modeling studies: Trumpp-Kallmeyer, S.;
Hoflack, J .; Bruinvels, A.; Hibert, M. J . Med. Chem. 1992, 35, 3448-
3462.
(7) (a) Behr, J .-P.; Lehn, J .-M.; Vierling, P. Helv. Chim. Acta 1982,
65, 1853-1867. (b) Bernardo, A. R.; Stoddart, J . F.; Kaifer, A. J . Am.
Chem. Soc. 1992, 114, 10624-10631. (c) Ishizu, T.; Hirayama, J .;
Noguchi, S. Chem. Pharm. Bull. 1994, 42, 1146-1148. (d) Campayo,
L.; Bueno, J . M.; Ochoa, C.; Navarro, P.; J imenez-Barbero, J .; Pepe,
G.; Samat, A. J . Org. Chem. 1997, 62, 2684.
S0022-3263(97)01297-8 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/23/1998

Toward Synthetic Adrenaline Receptors
J . Org. Chem., Vol. 63, No. 2, 1998 265
Sch em e 2. Syn th esis of P h osp h on a te Recep tor s 1-3
substituents for lateral recognition of the catechol. Fur-
thermore, chiral ester alcohols may lead to discrimination
between enantiomers of optically active catecholamines.
According to molecular modeling, complexation of alkyl-
ammonium ions by the p-xylylene phosphonate 2 leads
to a chelate structure with very strong electrostatic
interactions, enhanced by a network of short and almost
linear hydrogen bonds (Figure 1).11 Each of the three
positively polarized ammonium hydrogen atoms is lo-
cated exactly above the xylylene phenyl ring at a calcu-
lated distance of 3.5 Åsvery similar to the situation in
the natural receptor, where three electron-rich aromatic
residues are actively involved in π-cation stabilization.12
F igu r e 1. Complex of p-xylylene bisphosphonates with nor-
adrenaline according to force-field calculations.11
alkylammonium ions. As stated above, the new host
should be capable of a triple binding mode, i.e., exert
attraction for the guest by means of electrostatic, hydro-
gen-bond, and π-cation interactions. One of the simplest
solutions to this problem is a p-xylylene bisphosphonate
that not only imitates the ammonium-aspartate interac-
tion but also amplifies it due to its bidentate character
(Figure 1).9 The great advantage of the phosphonate over
a carboxylate stems from the fact that at neutral pH
phosphonates are fully dissociated due to their much
lower pKa values (1.8 vs 4.8), a prerequisite for molecular
recognition at physiological conditions.10 Another con-
venient feature of the phosphonate is the additional ester
functionality that can at a later stage serve to introduce
Resu lts a n d Discu ssion
The phosphonates can be prepared by the classic route
employing the Michaelis-Arbuzow reaction of xylylene
halides with trialkyl phosphites (Scheme 2).13 Basic
hydrolysis of the obtained tetraalkyl bisphosphonates
affords the respective bisphosphonic acids, which are
subsequently converted into their tetrabutylammonium
salts. We developed a shortcut for the direct transforma-
tion of alkylphosphonates into their tetrabutylammonium
salts. An aqueous solution of equimolar amounts of
tetrabutylammonium hydroxide and a dialkyl phospho-
nate are heated for 1 week at ∼100 °C. During this time,
the equilibrium is shifted slowly but steadily toward the
product side because the resulting phosphonate anion is
resonance-stabilized. After 1 week, spectroscopically
pure product is obtained in quantitative yield (Scheme
2).
(8) Exceptions are some recent ditopic receptors for dopamine: (a)
Kimura, E.; Fujioka, H.; Kodama, M. J . Chem. Soc., Chem. Commun.
1986, 1158-1159. (b) Schmidtchen, F. P. Z. Naturforsch., C. Biosci.
1987, 42, 476-485. (c) Hayakawa, K.; Kido, K.; Kanematsu, K. J .
Chem. Soc., Perkin Trans. 1 1988, 511-519. (d) Paugam, M.-F.;
Valencia, L. S.; Bogess, B.; Smith, B. D. J . Am. Chem. Soc. 1994, 116,
11203-1120. Paugam, M.-F.; Biens, J . T.; Smith, B. D.; Christoffels,
A. J .; de J ong, F.; Reinhoudt, D. N. J . Am. Chem. Soc. 1996, 118, 9820-
9825.
In the 1H NMR spectra of equimolar mixtures of
aliphatic primary ammonium chlorides with phosphonate
receptors 2 and 3, large downfield shifts of 0.5-0.8 ppm
(9) A bisphosphonate and a tetraphosphate have recently been used
for strong glucoside binding in acetonitrile: (a) Das, G.; Hamilton, A.
D. J . Am. Chem. Soc. 1994, 116, 11139-11140. Das, G.; Hamilton, A.
D. Tetrahedron Lett. 1997, 38, 3675-3678. (b) Anderson, S.; Neidlein,
U.; Gramlich, V.; Diederich, F. Angew. Chem., Int. Ed. Engl. 1995,
34, 1596-1600. Neidlein, U.; Diederich, F. J . Chem. Soc., Chem.
Commun. 1996, 1493-1494. (c) Recently a neutral phosphonate-
calix[4]arene was presented that utilizes PdO‚‚‚HN+ hydrogen bonds
for complexation of simple amines: Delangle, P.; Dutasta, J .-P.
Tetrahedron Lett. 1995, 36, 9325. (d) Polyammonium compounds have
been reported as artificial phosphate receptors: Furuta, H.; Magda,
D.; Sessler, J . L. J . Am. Chem. Soc. 1991, 113, 978-985. Hosseini, M.
W.; Lehn, J .-M.; Mertes, M. P. Helv. Chim. Acta 1983, 66, 2454-2466.
(10) (a) Natural examples for phosphate-ammonium interactions
are found in many enzymes that recognize phosphates, e.g., the inositol
1,4,5-trisphosphate receptor: Potter, B. V. L.; Lampe, D.; Angew.
Chem., Int. Ed. Engl. 1995, 34, 1933. (b) Polyammonium cations are
known to complex the phosphate groups of DNA and RNA and thereby
stabilize their tertiary structure: Saenger, W. Principles of Nucleic
Acid Structure; Springer-Verlag: New York, 1984; pp 343-346.
(11) Molecular Modeling Program: CERIUS2 from Molecular Simu-
lations, Inc. Force field: Dreiding 2.21.
(12) Dougherty, D. A. Science 1996, 271, 163-167.
(13) Tewari, R. S.; Kumari, N.; Kendurkar, P. S. Ind. J . Chem. 1977,
15B, 753-755. Chantrell, P. G.; Pearce, C. A.; Toyer, C. R.; Twaits, R.
J . Appl. Chem. 1965, 15, 460-462.

266 J . Org. Chem., Vol. 63, No. 2, 1998
Schrader
F igu r e 2. Dependence of the change in chemical shift of five
signals of 17 (c ) 0.5 mM) on the concentration of bisphos-
phonate 18 in DMSO-d6 at 20 °C.
F igu r e 3. J ob plot for complex formation between p-xylylene
bisphosphonate 2 and benzylamine (c ) 10 mM).
Ta ble 1. Associa tion Con sta n ts (K1:1) [M-1] fr om NMR
Titr a tion s in DMSO a t 20 °Ca
Ta ble 2. Associa tion Con sta n ts for Am in es vs Am in o
Alcoh ols (K1:1) [M-1] fr om NMR Titr a tion s w ith 3 in
DMSO a t 20 °Ca
benzylamine
(()-4
receptor
103 [M-1
]
103 [M-1
]
1
2
3
0.2
2.8
7.4
103
103
15.5
55.0
amines
benzylamine
L-alanine methyl ester
2-phenylethylamine
[M-1
]
amino alcohols
(()-4
R-D-glucosamine 10
(1S,2R)-norephedrine 6
(R)-propranolol 11
[M-1
]
a
7
10
12
19
55
59
62
66
Because of the strongly hygroscopic character of both titration
partners the DMSO-d6-solution contained about 0.1% of water.
Errors in Ka are standard deviations; they were estimated at e20%
for K e 104 M-1 and at e40% for K e 105 M-1
.
N-benzyl-N-methylamine
a
Because of the strongly hygroscopic character of both titration
partners the DMSO-d6-solution contained about 0.1% of water.
Errors in Ka are standard deviations; they were estimated at e20%
+
and line broadening of the NH3 signals demonstrate
their hydrogen bonding to the phosphonate anions. As
a consequence, the R-methylene protons are markedly
shifted upfield in the range of 0.3-1.0 ppm. On the other
hand, small but distinct complexation-induced shifts
(CIS) of 0.1-0.2 ppm are also observed at the receptor
molecule, namely downfield shifts of the R-methylene
protons and upfield shifts of the methyl esters. Finally,
at a 1:1 stoichiometry a downfield shift of 1-3 ppm is
for K e 104 M-1 and at e40% for K e 105 M-1
.
so that the existence of π-cation interactions remains
speculative. Perhaps the aromatic ring has to be located
in such a way that the Hδ+ points directly toward the
center of electron density, as is the case in the natural
receptor. In comparison, the m-diphosphonate 3 binds
even stronger, although here such a π-cation stabilization
consistently observed for the host’s 31P NMR signal.14
A
+
is not possible because the NH3 group no longer lies
second hint for interaction was found when at higher
concentrations (>10-2 M) the 2:1 complexes of various
alkylammonium chlorides with p-xylylene bisphospho-
nate 2 precipitated from DMSO solution.
exactly above the xylene ring. The advantage of 3 over
2 (i.e., of a m- vs p-xylylene core) according to force-field
calculations is the decreased PO-‚‚‚HN+ distance, which
strengthens both electrostatic attractions and hydrogen
bonds.10 Another indicator for the chelate-binding mode
is the 0.2 ppm downfield shift of the m-xylylene bisphos-
phonate’s lone aromatic proton CH-2 (Scheme 2), which
shows that the receptor conformation is changing when
the alkylammonium ion approaches (induced fit).9a
J ob’s method of continuous variations16 confirms the
postulated 1:1 stoichiometry. Maxima in the range of M
) 0.5-0.55 are observed for guest as well as for host
signals (Figure 3).
When 2-phenyl-2-hydroxyethylamine hydrochloride 4
(a model for the amino alcohol moiety in adrenaline
derivatives) was treated with the bisphosphonate recep-
tor molecules, the OH-signal immediately broadened and
strongly shifted downfield by ∼1.0 ppm. The â-CH signal
shifted upfield, but only by 0.2-0.3 ppm. Obviously, an
We performed NMR titrations of benzylammonium
chloride with phosphonate receptors 1-3 in DMSO-d6.
We measured the CIS of the R-CHN+ proton and ana-
lyzed the binding curves by nonlinear regression methods
(Figure 2).15 The calculated association constants are
summarized in Table 1. The simple monophosphonate/
ammonium interaction is comparatively weak; binding
constants of this kind are typically in the 102 M-1 range.
By contrast, the p-xylylene bisphosphonate binds to the
benzylammonium ion 14 times strongersa very clear
indication for the postulated chelate effect. No change
is observed for the p-xylylene aromatic proton at 7.0 ppm,
(14) No self-association of the simple phosphonates 1-3 could be
detected by simple dilution experiments.
(15) a) Schneider, H. J .; Kramer, R.; Simova, S.; Schneider, U. J .
Am. Chem. Soc. 1988, 110, 6442. (b) Wilcox, C. S. In Frontiers in
Supramolecular Chemistry and Photochemistry; Schneider, H.-J ., Du¨rr,
H., Eds.; VCH: Weinheim, 1991; p 123. We thank Prof. Schneider for
his program to evaluate 1:1 complexation.
(16) (a) J ob, P. Compt. Rend. 1925, 180, 928. (b) Blanda, M. T.;
Horner, J . H.; Newcomb, M. J . Org. Chem. 1989, 54, 4626.

Toward Synthetic Adrenaline Receptors
J . Org. Chem., Vol. 63, No. 2, 1998 267
additional hydrogen bond is formed between the phos-
phonate and the alcohol moiety of 4. This single ad-
ditional hydrogen bond results in a drastic 5-7.5-fold
increase of the binding constant over benzylamine’s Ka
value, reaching 55 000 M-1 for 3 (Table 1). We examined
other biologically important 1,2-amino alcohols (e.g.,
glucosamine 10 as a model for neuraminic acid, ephe-
drine 7 as an agonist of the adrenergic receptor, and
propranolol 11 as a typical â-blocker) and compared them
to the respective amines. The results are summarized
in Table 2.
F igu r e 4. Energy-minimized complex of glucosamine with
bisphosphonate 3.11
Ta ble 3. Associa tion Con sta n ts (K1:1) [M-1] for P r im a r y
vs th e Cor r esp on d in g N-Meth yl Su bstitu ted Secon d a r y
Am in es fr om NMR Titr a tion s w ith 3 in DMSO a t 20 °Ca
103
103
(a) All amino alcohols are bound much stronger than
their simple amine counterparts. An average estimate
of the association constants for the amino alcohols of
60 000 M-1 is five times higher than the average estimate
primary amines
benzylamine
[M-1
]
secondary amines
[M-1
]
7
18
55
62
N-methyl-N-benzylamine
(()-adrenaline 9
(()-halostachine 5
(1S,2R)-ephedrine 7
(R)-propranolol 11
19
16
26
71
66
(()-noradrenaline 8
for simple amines of 12 000 M-1
.
(()-4
(1S,2R)-norephedrine 6
(b) NMR experiments provide further structural infor-
mation and confirm the strong hydroxyl binding: In the
free glucosamine 10 only the OH-1 and OH-3 adjacent
to the ammonium functionality shift during the titration,
whereas OH-4 and OH-6 are pointing away from the
receptor and thus remain untouched (Figure 4). Pro-
pranolol 11 has a diastereotopic O-methylene group,
which is normally conformationally flexible. On addition
of small amounts of the phosphonate, however, these
methylene protons immediately turn chemically non-
equivalent because the OH group is conformationally
a
Because of the strongly hygroscopic character of both titration
partners the DMSO-d6-solution contained about 0.1% of water.
Errors in Ka are standard deviations; they were estimated at e20%
for K e 104 M-1 and at e40% for K e 105 M-1
.
bisphosphonates can be expected to bind especially well
to adrenaline itself as well as to the respective â-blockers.
The moderate binding constant for adrenaline 9 can
be rationalized by intermolecular competition of the
catechol OH groups, which also bind to the receptor
molecule. On phosphonate addition, large shifts of both
relatively acidic phenol protons indicate a second super-
imposed equilibrium, so that evaluation based on a
simple 1:1 stoichiometry of the complex must give
misleading results. Therefore, adrenaline is assumed to
bind to 3 as strongly as the other amino alcohols. This
competition of several groups in the guest for binding
sites on the host indicates that by introduction of a
properly positioned functional group for catechol recogni-
tion the overall binding constant should increase mark-
edly.
The question arises how far the capability of the
bisphosphonate reaches to establish cooperative hydrogen
bonds with remote hydroxyl groups in a substrate. We
studied the binding behavior of 3 with 1,n-amino alcohols
systematically. The results are summarized in Table 4.
While 1,2- and 1,3-amino alcohols are bound very strongly,
a marked drop in binding energy is observed for 1,4- and
1,5-amino alcohols, although even in the latter case a
strong downfield shift of the OH proton accompanied with
+
locked. The same effect is observed even for the NH2
protons. Obviously, each NH is pointing to a different
phosphonate moiety, with only one of the latter being
involved in additional OH binding (Figure 5).
(c) Secondary amines seem to be bound even stronger
than primary amines (see Table 2: benzylamine vs
benzylmethylamine; see also propranolol 11 with its
sterically demanding isopropyl group vs the primary
amino alcohols). We checked this hypothesis for several
pairs of biologically important amino alcohols that differ
only in the extra methyl group of the respective secondary
amine. The results are given in Table 3. With the
exception of 4/5, the secondary amines are indeed com-
plexed at least as strongly as their primary counterparts.
This may indicate that electrostatic interactions contrib-
ute most to the stabilization energy in the complex, so
that loss of one hydrogen bond can be overcompensated
by other factors such as hydrophobic interactions with
the additional alkyl group. Contrary to all artificial
ditopic catecholamine receptors known to date, our

268 J . Org. Chem., Vol. 63, No. 2, 1998
Schrader
still well above those found for simple amines (cf. Table
2), there is ample evidence for the fact that cooperative
hydrogen bonds are formed even at a 1,5-distance of the
binding sites. This is in full agreement with the highly
selective binding of N-amide-protected arginine deriva-
tives by similar bisphosphonate tweezer molecules.17
These strongly bind to the guanidinium cation and
simultaneously form a cooperative hydrogen bond to the
arginine’s amide-NH five atoms away from the guani-
dine. The additional hydrogen bond leads to an increase
in the binding constant by 1 order of magnitude and
demonstrates together with the above-discussed ex-
amples that high degrees of organization can be achieved
in a complex by means of a single cooperative hydrogen
bond to a relatively acidic XH functionality.
However, additional hydrogen bonds can also be coun-
terproductive if they are competitive, as comparison of
the amino diols and the related serine derivative shows
(they represent other biologically important 1,2-amino
alcohols, e.g., the amino acid derivative serine methyl
ester 12 and serinol 13 as a model for sphingosine, Table
4). The 1,3-functionalities in serine methyl ester and
serinol form an internal hydrogen bond producing a six-
membered ring, which has to be broken before OH
binding can occur. Thus, binding to the bisphosphonate
receptor remains moderate. By contrast, 3-amino-2-
hydroxypropanol can form two cooperative hydrogen
bonds with both OH groups, and the binding constant is
even much higher than for the 1,2- or the 1,3-amino
alcohol. Extension of this concept leads to the complex-
ation of amino sugar alcohols. When 1-aminosorbitol was
titrated with 3, all OCH protons shifted upfield and the
binding constant reached 127 000 M-1. Force-field cal-
culations suggest that all OH groups may be involved in
hydrogen bonding if the aminosorbitol chain winds itself
around the receptor (Figure 6).11 It is tempting to draw
parallels to nucleosomes in which the DNA is wound
around histones interacting mainly with its polyphos-
phate anions.18
F igu r e 5. (Top) energy-minimized complex of (R)-propranolol
with bisphosphonate 3.11 (Bottom) 1H NMR signal of the
diastereotopic methylene protons (OCH2) in (R)-propranolol (1)
before addition of 3 and (2) with 1 equiv of 3.
Ta ble 4. Associa tion Con sta n ts (K1:1) [M-1] for 1,n -Am in o
Alcoh ols fr om NMR Titr a tion s w ith 3 in DMSO a t 20 °Ca
103
103
1,n-amino alcohol [M-1
]
1,n,m-amino alcohol
[M-1
]
1,2-aminoethanol
1,3-aminopropanol
1,4-aminobutanol
1,5-aminopentanol
73
66
21
27
serine methyl ester 12
serinol 13
3-amino-1,2-propanediol 14
1-aminosorbitol 15
6
16
99
127
The above results provide ample evidence for the fact
that 2 and especially 3 represent indeed highly efficient
biomimetic receptor molecules for primary and secondary
amines and that they are selective for 1,2- and 1,3-amino
alcohols. To our surprise, very little synthetic work has
been done on this field of supramolecular chemistry:
a
Because of the strongly hygroscopic character of both titration
partners the DMSO-d6-solution contained about 0.1% of water.
Errors in Ka are standard deviations; they were estimated at e20%
for K e 104 M-1 and at e40% for K e 105 M-1
.
a distinct upfield shift of the corresponding OCH proton
proves hydrogen bonding to the phosphonate anion. We
conclude that due to smaller entropy losses 1,2- and 1,3-
amino alcohols are much better bound by the m-xylylene
bisphosphonate than their higher homologs. Since the
association constants of 1,4- and 1,5-amino alcohols are
(17) Schrader, T. Chem. Eur. J . 1997, 3, 1537-1541.
(18) For the structure of nucleosomes, see: Darnell, J .; Lodish, H.;
Baltimore, B. Molecular Cell Biology; Scientific American Books: New
York, 1990.

Toward Synthetic Adrenaline Receptors
J . Org. Chem., Vol. 63, No. 2, 1998 269
Sch em e 3. Syn th esis of Ar yl Bisp h osp h on a te Recep tor s 18-20
F igu r e 7. Energy-minimized structure of the noradrenaline/
20 complex.11
F igu r e 6. Energy-minimized structure of the 1-aminosorbi-
tol/3 complex.11
Reetz reported about an artificial receptor for amines and
alcohols, based on a crown ether boronate; however, this
system works only in dry nonpolar dichloromethane.19a
Gellman achieved strong binding of amino sugars in
methanolic chloroform with his neutral PdO/SdO-
“bowls”; however, for solubility reasons, binding is also
restricted to organic solvents.19b The above-discussed
NMR titrations with our bisphosphonate receptor mol-
ecules have been carried out in DMSO with small
amounts of watersfor such a competitive solvent binding
constants up to 130 000 M-1 are remarkably high,
especially in view of the important role of hydrogen
bonding. In methanol, m-xylylene bisphosphonate 3
binds the noradrenaline model compound 4 with a Ka of
620 M-1 and glucosamine with a Ka of 490 M-1; even in
water weak binding in observed, indicated by a 0.1 ppm
upfield shift of the CHN proton.
F igu r e 8. Alternating network of charges and hydrogen bonds
in the 2:1 complex between benzylamine and 18; formation of
an aromatic hydrophobic cluster.
To achieve biomimetic recognition of the catechol by
π-stacking interactions, we synthesized the three bisaryl
phosphonates 18-20 (Scheme 3). The tetraalkyl esters
were converted into the corresponding tetrachlorides with
PCl5 or milder SOCl2 and a catalytic amount of DMF.
Subsequent alcoholysis with phenol derivatives furnished
the bismonoesters,20 which were subjected to pH titration
(19) (a) Reetz, M. T.; Niemeyer, C. M.; Harms, K. Angew. Chem.,
Int. Ed. Engl. 1991, 30, 1474. (b) Savage, P. B.; Gellman, S. H. J . Am.
Chem. Soc. 1993, 115, 10448-10449.

270 J . Org. Chem., Vol. 63, No. 2, 1998
Schrader
F igu r e 9. Computer-calculated structure of the complex between 17 and aryl bisphosphonate 20.11
Ta ble 5. Associa tion Con sta n ts for th e Ar om a tic
Bisp h osp h on a tes (K1:1) [M-1] fr om NMR Titr a tion s in
DMSO a t 20 °Ca
tion of the analytically pure 2:1 complex. On the basis
of force-field calculations, we assume that a very stable
alternating network of NH3+ and PO2- groups is formed
because this array guarantees maximum electrostatic as
well as hydrogen-bond interactions.9 It may also explain
the extremely low solubility of these complexes, for all
charges and protons are directed inward while the
aromatic rings can form a hydrophobic cluster that
suppresses solvation by polar solvents (Figure 8). Highly
diluted DMSO-d6 solutions of the complexes still show a
significant upfield shift of the CH2N+ proton of ca. 0.3
ppm, but the NMR titration with benzylamine or 4 gave
only moderate Ka’s (Table 5).
Primary ammonium complexes with 19 and 20 are very
well soluble, but also for these hosts the binding con-
stants remain lower than for the simple p-xylylene
bisphosphonate. Highfield shifts are not detectable in
the aromatic region of the modified receptors or of their
guests. This rules out the presence of π-stacking interac-
tions. Perhaps the low values of Ka originate from
competing self-association of the extended aromatic arms
of the receptor molecules. In addition, free rotation of
the twisted biphenyl and trityl moieties may hinder the
access of guest molecules. In general, π,π-interactions
of electron-rich phenol esters with electron-rich phenyl
rings are relatively weak. Therefore, it will be necessary
to design phosphonate hosts with electron-poor aromatic
moieties in order to achieve a maximum effect for binding
of the catechol by means of a CT complex. To check the
validity of this concept, we compared the association
constants between 18 or 20 and benzylamine or 4 with
those obtained for the related p-nitroaromatic compounds
(R)-(R)-methyl-4-nitrobenzylamine 16 or D-(-)-threo-2-
amino-1-(4-nitrophenyl)-1,3-propanediol (17) (Table 5). In
all cases, small but distinct upfield shifts for both protons
of the nitroarene were observed. Table 5 shows that the
association constants for the electron poor aromatic
benzylamine
(()-4
receptor
103 [M-1
]
(R)-16
f 8.9
103 [M-1
]
(-)-17
f 10.2
18
19
20
2.0
1.1
4.2
7.0
3.4
6.0
f 18.9
a
Because of the strongly hygroscopic character of both titration
partners the DMSO-d6-solution contained about 0.1% of water.
Errors in Ka are standard deviations; they were estimated at e20%
for K e 104 M-1 and at e30% for K g 104 M-1
.
to give the bis(tetrabutylammonium) salts. From mo-
lecular modeling studies, we expected that the aromatic
residues should be capable of increased π-stacking in the
order of 18 (â-naphthyl), 19 (m-biphenylyl), and 20
[m-(diphenylmethyl)phenyl]. With 20, we hoped to mimic
the double sandwich-type π-stacking interactions present
in the natural adrenergic receptor; furthermore, we hoped
to achieve complete solvent exclusion from the inner
sphere of the complex by means of the two additional
protecting phenyl rings (solvophobic effect, Figure 7).21
For each receptor molecule, we checked its binding
behavior toward benzylamine and 2-phenyl-2-hydroxy-
ethylamine 4 (Table 5).
Each receptor confirmed the considerably stronger
binding of amino alcohols by means of an additional
hydrogen bond. However, complexes with 18 showed
marked differences to those with 19 and 20. Addition of
various primary alkylammonium chlorides including the
above-mentioned to a solution of the â-naphthylphospho-
nate 18 even in water resulted in spontaneous precipita-
(20) The parent phenol for 20 was conventionally synthesized from
ethyl m-hydroxybenzoate and phenylmagnesium bromide followed by
reduction with zinc/acetic acid: Baeyer, A. Liebigs Ann. Chem. 1907,
354, 167-171.
(21) Schneider, H. J . Angew. Chem., Int. Ed. Engl. 1991, 30, 1419.

Toward Synthetic Adrenaline Receptors
J . Org. Chem., Vol. 63, No. 2, 1998 271
Hz, 1 C), 124.2 (d, J ) 2.3 Hz, 1 C), 127.1 (d, J ) 2.2 Hz, 2 C),
129.4 (d, J ) 5.6 Hz, 2 C), 138.9 (m, 1C); 31P NMR δ 14.89 (s).
Anal. Calcd for C25H48O3P: C, 67.98; H, 10.96; N, 3.17.
Found: C, 67.93; H, 11.19; N, 3.40.
Bis(tetr a bu tyla m m on iu m )d im eth yl p-xylylen ed ip h os-
p h on a te (2): yield 98%; 1H NMR (300 MHz, DMSO-d6) δ 0.93
(t, 24H, J ) 7.3 Hz), 1.30 (tq, 16H, J ) 7.3 Hz), 1.57 (m, 16H),
2.54 (d, 4H, J ) 18.5 Hz), 3.18 (m, 16H), 3.22 (d, 6H, J ) 9.8
Hz), 7.00 (s, 4 H); 13C NMR (75 MHz, DMSO-d6) δ 13.5 (s, 8C),
19.2 (s, 8C), 23.1 (s, 8C), 35.1 (d, J ) 123.7 Hz, 2 C), 50.5 (m,
2 C), 57.4 (m, 8 C), 128.4 (m, 4C), 134.5 (m, 2C); 31P NMR δ
16.86 (s). Anal. Calcd for C42H86N2O6P2‚2H2O: C, 61.88; H,
11.27; N, 3.81. Found: C, 62.04; H, 11.16; N, 3.45.
amino alcohols are consistently higher than those ob-
tained for the simple phenyl-substituted compounds by
a factor of 1.5-4.5 (Figure 9). From this, we conclude
that π-stacking interactions between an electron-rich and
an electron-poor binding partner are indeed a valuable
means to substantially enlarge the affinity for an adrena-
line-type guest even of an open-chain receptor molecule.
We are currently incorporating the xylylene bisphospho-
nate moiety into macrocylic ring systems with a high
degree of preorientation of all functional groups for
strong, selective, and biomimetic adrenaline recognition.
Bis(tetr a bu tyla m m on iu m )bis(2-n a p h th yl) p-xylylen e-
d ip h osp h on a te (18): yield 93%; mp 159-160 °C; H NMR
Exp er im en ta l Section
1
(300 MHz, DMSO-d6) δ 0.91 (t, 24H, J ) 7.3 Hz), 1.28 (tq, 16H,
J ) 7.3 Hz), 1.54 (m, 16H), 2.77 (d, 4H, J ) 18.7 Hz), 3.14 (m,
16H), 7.06 (s, 4 H), 7.25-7.45 (2t, 1d, 6 H), 7.60 (s, 2H), 7.7-
7.8 (3d, 6H); 13C NMR (75 MHz, CDCl3) δ 13.7 (s, 8 C), 19.6
(s, 8 C), 23.8 (s, 8 C), 35.8 (d, J ) 127.5 Hz, 2 C), 58.4 (m, 8 C),
115.8/122.5/123.7/125.6/127.2/127.4/128.6/129.5/133.9/134.4/
Gen er a l Meth od s. DMSO-d6 was purchased from Aldrich
in 99.8% purity. Thin-layer chromatography (TLC) analyses
were performed on silica gel 60 F-254 with a 0.2 mm layer
thickness. Preparative chromatography columns were packed
with Kieselgel 60 (70-230 mesh) from Macherey & Nagel. All
solvents were dried and freshly disitilled before use.
1H NMR Titr a tion s. A solution of the phosphonate recep-
tor (10 equiv in 0.4 mL of DMSO-d6) was added in portions
via microsyringe to a solution of the primary or secondary
amine hydrochloride (∼1-20 mM, 1 equiv in 0.7 mL of DMSO-
d6) in a capped NMR tube. The amine solution contained
∼0.05-1% water; due to its strong hygroscopic effect the
tetrabutylammonium phosphonate solution contained ∼0.3-
0.6% water. Volume and concentration changes were taken
into account during analysis.14
152.1 (m, 26 C); 31P NMR δ 16.65 (s). Anal. Calcd for C60
-
H94N2O6P2‚4H2O: C, 67.12; H, 9.58; N, 2.61. Found: C, 66.95;
H, 9.28; N, 3.20.
Bis(tetr abu tylam m on iu m )bis(m-biph en ylyl) p-xylylen e-
d ip h osp h on a te (19): yield 92%; 1H NMR (300 MHz, DMSO-
d6) δ 0.92 (t, 24H, J ) 7.3 Hz), 1.29 (tq, 16H, J ) 7.3 Hz), 1.55
(m, 16H), 2.72 (d, 4H, J ) 18.8 Hz), 3.15 (m, 16H), 7.04 (s, 4
H), 7.05-7.60 (m, 18H); 13C NMR (75 MHz, CDCl3) δ 13.7 (s,
8 C), 19.6 (s, 8 C, 23.9 (s, 8 C), 36.04 (d, J ) 127.6 Hz, 2 C),
58.4 (m, 8 C), 119.4/119.8/120.2/121.0/126.6/127.0/127.5/128.6/
129.1/129.6/134.1/141.1/141.7/154.8 (m, 30 C); 31P NMR δ 14.23
(s).
J ob P lot. Equimolar solutions (10-2 M) of receptor 2 and
benzylamine hydrochloride were mixed in various amounts.
1H NMR spectra of the mixtures were recorded, and the
chemical shifts were analyzed by J ob’s method modified for
NMR results.15
Syn th esis of P h osp h on ic Acid Mon o- a n d Diester s by
Ba se Hyd r olysis. Benzylphosphonic acid monoethylester,
p-xylylenediphosphonic acid dimethyl ester, and m-xylylene-
diphosphonic acid dimethyl ester were prepared by base
hydrolysis of the respective di- and tetraesters with 5 N
aqueous sodium hydroxide, followed by acidification with 1 N
HCl.
Bis(tetr abu tylam m on iu m )bis[m-(diph en ylm eth yl)ph en -
yl] p-xylylen ed ip h osp h on a te (20): yield 89%; 1H NMR (300
MHz, DMSO-d6) δ 0.93 (t, 24H, J ) 7.3 Hz), 1.30 (tq, 16H, J
) 7.3 Hz), 1.55 (m, 16H), 2.58 (d, 4H, J ) 18.6 Hz), 3.16 (m,
16H), 5.51 (s, 2H), 6.63 (d, J ) 7.5 Hz), 6.79 (s, 2 H), 6.87 (s,
4 H), 7.00-7.32 (m, 24H); 13C NMR (75 MHz, CDCl3) δ 13.7
(s, 8 C), 19.6 (s, 8 C), 23.9 (s, 8 C), 35.8 (d, J ) 127.6 Hz, 2 C),
56.7 (s, 2 C), 58.5 (m, 8 C), 118.6/121.8/122.6/126.1/127.4/128.1/
128.3/128.6/128.9/129.5/133.9/144.1/144.5/154.5 (m, 42 C); 31
P
Ben zylp h osp h on ic a cid m on oeth yl ester : yield 96%; mp
NMR δ 14.05 (s). Anal. Calcd for C78H110N2O6P2‚4H2O: C,
71.73; H, 9.11; N, 2.15. Found: C, 71.68; H, 8.03; N, 1.99.
1
66 °C; H NMR (90 MHz, DMSO-d6) δ 1.25 (t, 3H, J ) 7 Hz),
3.0 (d, 2H, J ) 22 Hz), 3.85 (dq, 2H, J ) 7 Hz), 7.25 (m, 5 H),
9.8 (s, br, 1H); 31P NMR (DMSO-d6) δ 24.74 (s). Anal. Calcd
for C9H13O3P: C, 53.98; H, 6.55. Found: C, 53.82; H, 6.55.
p-Xylylen ediph osph on ic acid P ,P ′-dim eth yl ester : yield
77%; mp 209-10 °C; 1H NMR (300 MHz, DMSO-d6) δ 3.05 (d,
4H, J ) 20 Hz), 3.50 (d, 6H, J ) 10.5 Hz), 7.20 (s, 4 H), 8.20
(s, br, 2H); 31P NMR (DMSO-d6) δ 26.00 (s). Anal. Calcd for
Dir ect Syn th esis of Tetr a bu tyla m m on iu m Bisp h os-
p h on a te 3. m-Xylylenediphosphonic acid tetramethyl ester
was treated with 2.0 equiv of aqueous tetrabutylammonium
hydroxide and heated to reflux for 1 week. After evaporation
to dryness, the crude product was extracted with chloroform,
dried over magnesium sulfate, filtered, and again evaporated
to dryness. The receptor was further dried at 10-3 mbar over
P2O5.
Bis(tetr a bu tyla m m on iu m )dim eth yl m-xylylen ed ip h os-
p h on a te 3: yield 92%; mp ∼40 °C; 1H NMR (300 MHz,
DMSO-d6) δ 0.94 (t, 24H, J ) 7.3 Hz), 1.31 (tq, 16H, J ) 7.3
Hz), 1.57 (m, 16H), 2.53 (d, 4H, J ) 18.5 Hz), 3.18 (m, 16H),
3.21 (d, 6H, J ) 9.8 Hz), 6.92-6.95 (m, 4 H); 13C NMR (75
MHz, CDCl3) δ 13.8 (s, 8 C), 19.7 (s, 8 C), 24.1 (s, 8 C), 35.7 (d,
J ) 124.0 Hz, 2 C), 51.61 (m, 2 C), 58.46 (m, 8 C), 126.5/126.9/
131.6 (m, 4CH), 137.7 (m, 2C); 31P NMR δ 16.48 (s). Anal.
Calcd for C42H86N2O6P2‚3H2O: C, 60.69; H, 11.16; N, 3.37.
Found: C, 60.54; H, 11.33; N, 3.35.
Syn th esis of th e Dip h osp h on ic Acid Tetr a ch lor id es.
In a dry argon atmosphere the tetraesters were mixed with 2
equiv of PCl5 or with 3-4 equiv of SOCl2 and a catalytic
amount of DMF. After the mixture was heated to 140 °C for
2 h, POCl3 or remaining SOCl2 was distilled off, and the crude
product was recrystallized twice from toluene.
p-Xylylen ed ip h osp h on ic a cid tetr a ch lor id e: yield 55%;
mp 171-173 °C; 1H NMR (90 MHz, CDCl3) δ 3.95 (d, 2H, J )
15 Hz), 7.35 (m, 4H); 31P NMR(CDCl3) δ 44.85 (s).
C
10H16O6P2: C, 40.81; H, 5.48. Found: C, 40.56; H, 5.44.
m-Xylylen ediph osph on ic acid P ,P ′-dim eth yl ester : yield
79%; mp 99-100 °C; 1H NMR (300 MHz, DMSO-d6) δ 3.05 (d,
4H, J ) 21.6 Hz), 3.52 (d, 6H, J ) 10.8 Hz), 7.10-7.24 (m, 4
H), 9.80 (s, br, 2H); 31P NMR (DMSO-d6) δ 25.86 (s).
Syn th esis of th e Tetr a bu tyla m m on iu m Sa lts 1, 2, a n d
18-20. A typical pH-titration experiment was carried out with
the above-described phosphonic acid mono- and diesters and
1.50 M tetrabutylammonium hydroxide in water. When the
equivalence point was reached (normally after addition of
exactly 1 or 2 molar equiv of aqueous tetrabutylammonium
hydroxide), the reaction mixture was evaporated to dryness
and extracted with dry chloroform. After drying over magne-
siumsulfate and filtration, the solvent was removed and the
receptor was further dried at 10-3 mbar over P2O5.
Tetr a bu tyla m m on iu m eth yl ben zylp h osp on a te (1):
yield 98%; mp 48 °C; 1H NMR (300 MHz, DMSO-d6) δ 0.92 (t,
12H, J ) 7.3 Hz), 1.01 (t, 3H, J ) 7.0 Hz), 1.29 (tq, 8H, J )
7.3 Hz), 1.55 (m, 8H), 2.65 (d, 2H, J ) 22 Hz), 3.20 (m, 8H),
3.60 (dq, 2H, J ) 7.0/7.3 Hz), 7.03 (t, 1 H, J ) 7.4 Hz), 7.14 (t,
2 H, J ) 7.4 Hz), 7.21 (d, 2 H, J ) 7.4 Hz); 13C NMR (75 MHz,
DMSO-d6) δ 13.5 (s, 4 C), 16.9 (d, 1 C), 19.2 (s, 4 C), 23.1 (s, 4
C), 36.3 (d, J ) 122.9 Hz, 1 C), 57.4 (m, 4 C), 58.4 (d, J ) 6.2
m-Xylylen ediph osph on ic acid tetr ach lor ide: yield 65%;
mp 148-149 °C; 1H NMR (90 MHz, CDCl3) δ 3.90 (d, 2H, J )

272 J . Org. Chem., Vol. 63, No. 2, 1998
Schrader
15 Hz), 7.30 (m, 4H); 31P NMR (CDCl3) δ 44.50 (s). Anal.
Calcd for C8H8O2P2Cl4: C, 28.27; H, 2.37. Found: C, 28.28;
H, 2.27.
141.7, (s, 24 C), 151.3 (m, 2 C); 31P NMR (CDCl3) δ 22.19 (s).
Anal. Calcd for C32H28O6P2: C, 67.35; H, 4.95. Found: C,
67.30; H, 5.06.
Syn th esis of th e p-Xylylen ed ip h osp h on ic Acid Dia r yl
Ester s. Under argon, p-xylylenediphosphonic acid tetrachlo-
ride was dissolved in dry tetrahydrofuran, and 2 equiv of the
phenol was added, followed at 0 °C by 2 equiv of triethylamine.
The reaction mixture was stirred for 30 min at 0 °C, warmed
to room temperature, stirred for another 2 h, and filtered.
Water and hexane were then added, and the resulting two-
layer system was stirred vigorously overnight. The precipi-
tated product was filtered off, washed, and recrystallized from
methanol.
p -Xylylen ed ip h osp h on ic a cid P ,P ′-b is[m -(d ip h en yl-
m eth yl)p h en yl ester ]: yield 64%; mp 164 °C; 1H NMR (300
MHz, CDCl3) δ 2.91 (d, 4H, J ) 19.7 Hz), 5.53 (s, 2H), 6.70-
7.40 (m, 32H), 7.97 (s, br, 2H); 13C NMR (75 MHz, CDCl3) δ
32.7 (d, J ) 138.0 Hz, 2 C), 56.48 (s, 1C), 116.5, 121.6, 126.0,
126.4, 128.4, 129.4, 129.6, 130.2, 143.3, 145.9 (s, 42 C), 150.2
(m, 2 C); 31P NMR (CDCl3) δ 24.72 (s). Anal. Calcd for
C
46H40O6P2: C, 73.59; H, 5.37. Found: C, 73.49; H, 5.13.
Isola ted 2:1 com p lex betw een 18 a n d ben zyla m m o-
n iu m ch lor id e: Bis(benzylammonium)bis(2-naphthyl) p-xy-
lylenediphosphonate: yield 95%; mp 195-200 °C dec; 1H NMR
(300 MHz, DMSO-d6): 2.83 (d, 4H, J ) 18.5 Hz), 3.83 (s, 4H),
7.06 (s, 4 H), 7.30-7.50 (m, 16 H), 7.65 (d, J ) 1.7 Hz, 2H),
7.75-7.85 (m, 6H), 8.20 (s, br, 6H); 31P NMR (CD3OD) δ 19.71
(s). Anal. Calcd for C42H42N2O6P2: C, 68.83; H, 5.78; N, 3.82.
Found: C, 68.69; H, 5.65; N, 3.62.
p-Xylylen ed ip h osp h on ic a cid P ,P ′-bis(2-n a p h th yl es-
1
ter ): yield 41%; mp 279 °C dec; H NMR (300 MHz, DMSO-
d6) insoluble. Anal. Calcd for C28H24O6P2: C, 64.87; H, 4.67.
Found: C, 62.64; H, 4.48.
p -Xylylen ed ip h osp h on ic a cid P ,P ′-b is(m -b ip h en ylyl
ester ): yield 36%; mp 205-6 °C; 1H NMR (300 MHz, CDCl3)
δ 3.28 (d, 4H, J ) 20.1 Hz), 7.10-7.64 (m, 20H); 13C NMR (75
MHz, CDCl3) δ 33.4 (d, J ) 135.6 Hz), 118.6, 119.4 (s, 4 C),
120.8 (d, 2 C), 122.4, 126.6, 127.7, 128.9, 129.7, 130.0, 139.2,
J O971297V