A fluorescent polymeric heparin sensor

Article abstract of DOI:10.1002/chem.200700677

Linear copolymers have been developed which carry binding sites tailored for sulfated sugars. All binding monomers are based on the methacrylamide skeleton and ensure statistical radical copolymerization. They are decorated with o-aminomethylphenylboronates for covalent ester formation and/ or alkylammonium ions for noncovalent Coulomb attraction. Alcohol sidechains maintain a high water solubility; a dansyl monomer was constructed as a fluorescence label. Statistical copoly merization of comonomer mixtures with optimized ratios was started by AIBN (AIBN = 2,2′-azoisobutyronitrile) and furnished water-soluble comonomers with an exceptionally high affinity for glucosaminoglucans. Heparin can be quantitatively detected with an unprecedented 30 nM sensitivity, and a neutral polymer without any ammonium cation is still able to bind the target with almost micromolar affinity. From this unexpected result, we propose a new binding scheme between the boronate and a sulfated ethylene glycol or aminoethanol unit. Although the mechanism of heparin binding involves covalent boronate ester formation, it can be completely reversed by protamine addition, similar to heparin's complex formation with antithrombin III.

Full text of DOI:10.1002/chem.200700677

FULL PAPER
DOI: 10.1002/chem.200700677
A Fluorescent Polymeric Heparin Sensor
Wei Sun, Heinz Bandmann, and Thomas Schrader*[a]
Abstract: Linear copolymers have been
developed which carry binding sites
tailored for sulfated sugars. All binding
monomers are based on the methacry-
lamide skeleton and ensure statistical
radical copolymerization. They are dec-
orated with o-aminomethylphenylboro-
nates for covalent ester formation and/
or alkylammonium ions for noncova-
lent Coulomb attraction. Alcohol side-
chains maintain a high water solubility;
a dansyl monomer was constructed as a
fluorescence label. Statistical copoly-
merization of comonomer mixtures
with optimized ratios was started by
a neutral polymer without any ammo-
nium cation is still able to bind the
target with almost micromolar affinity.
From this unexpected result, we pro-
pose a new binding scheme between
the boronate and a sulfated ethylene
glycol or aminoethanol unit. Although
the mechanism of heparin binding in-
volves covalent boronate ester for-
mation, it can be completely reversed
by protamine addition, similar to hepa-
rinꢀs complex formation with anti-
thrombin III.
AIBN
(AIBN=2,2’-azoisobutyroni-
trile) and furnished water-soluble co-
monomers with an exceptionally high
affinity for glucosaminoglucans. Hepa-
rin can be quantitatively detected with
an unprecedented 30 nm sensitivity, and
Keywords: boronic acids · copoly-
mers
· fluorescence titrations ·
glucosaminoglucans · heparin
Introduction
to monomeric sugar species.[2] Recently, several groups
pointed out that a-hydroxycarboxylates represent excellent
candidates for cyclic boronate ester formation,[3] and Strong-
in and Kataoka utilized this interaction to develop chemical
receptors for neuraminic acid.[4]
Sugar recognition under physiological conditions by artificial
receptors poses the double challenge of forming strong in-
teractions with OH groups in competitive buffered water
and thereby achieving high selectivity among the large
number of structurally related carbohydrates. Significant
progress has been made by introducing boronic acid deriva-
Heparin is a highly sulfated polysaccharide and is widely
known as an anticoagulant as a result of its inhibitory com-
plex formation with antithrombin III—it is therefore com-
monly used in surgery and in postoperative treatment.[5]
Serum concentration must be monitored in submicromolar
concentrations during cardiopulmonary surgery and even
lower in long-term anticoagulant therapy of DVT (deep
venous thrombosis; 0.1 mm). Conventional methods involve
the classical measurement of the activated clotting time
(ACT), activated partial thromboplastin time (aPTT), po-
tentiometric assays, and protamine complexation.[6] Hepa-
rinꢀs chemical structure, although of high polydispersity, is a
constant repeat of a 1,4-glycosidic sugar dimer, carrying hy-
droxycarboxylates (iduronic acids) and hydroxysulfates (glu-
cosamines). In serum it is often accompanied by less sulfat-
ed anionic sugars, such as chondroitine sulfate and hyaluron-
ic acid.
À
tives with internal B N bonds, which carry a fluorescent
probe for internal charge transfer (ICT, Shinkai) or photoin-
duced electron transfer (PET, James).[1] However, by far the
majority of these elegant sensor systems has been restricted
[a] W. Sun, Dipl.-Ing. H. Bandmann, Prof. Dr. T. Schrader
Institut für Organische Chemie
Universität Duisburg, Essen
Universitätsstrasse 5, 45117 Essen (Germany)
Fax : (+49)201-183-4252
E-mail: thomas.schrader@uni-due.de
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author: Exact polymeri-
zation conditions, NMR spectroscopic and GPCdata as well as
stacked plots of the NMR titration with model compound 10,
11B NMR, H,H-, and C,H-correlation spectra of the glucosamine sul-
fate and its complex with phenylboronic acid/piperidine. Full details
on the protamine back titration as well as representative binding
curves and Job plots of the fluorescence titrations are also included.
In the past few years, the Anslyn group combined the
ortho-aminomethylphenylboronate motif with alkylammoni-
um groups for additional electrostatic interactions, and final-
ly presented a heparin sensor fixed onto a wide aromatic
platform.[7] This host molecule was shown to detect heparin
Chem. Eur. J. 2007, 13, 7701 – 7707
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7701

in serum samples with very high affinity, down to 0.1 mm
concentrations. It was argued, that both the hydroxycarbox-
ylates and the sulfates contribute to the binding event, with
selectivity resulting from Coulomb attraction between sul-
fates and ammonium ions, and an additional boronic acid
ester formation, structural features of which were not speci-
fied.
In a new concept, originally devoted to protein surface
recognition,[8] we have now developed copolymers with
binding sites tailored for sulfated sugars. Binding monomers
were all based on the methacrylamide skeleton, carrying o-
aminomethylphenylboronates (ester formation) and/or alky-
lammonium ions (Coulomb attraction) as well as alcohol
side chains to maintain high water solubility. A dansyl mo-
nomer was constructed as a fluorescence label. We found
that heparin can be quantitatively detected with an unprece-
dented 30 nm sensitivity and that a neutral polymer without
any ammonium cation is still able to bind the target with
almost micromolar affinity. From this unexpected result, we
propose a new binding scheme between the boronate and a
sulfated ethylene glycol or aminoethanol unit.
Figure 1. a) Functional comonomer structures. 1: dansyl, 2: boronic acid,
3: ammonium, 4: dodecyl, 5: amino-alcohol unit. b) Schematic represen-
tation of a selected copolymer (6).
mers were highly fluorescent powders with good water solu-
bility except for those containing the alkyl monomer. All
structural elements are visible and afford the expected inte-
1
grals in their H NMR spectra. Molecular weights were de-
termined at ꢀ120 kD (MW) by GPC(standards: polyethy-
lene oxide, polyethyleneglycol).[12] A detailed summary of
all examined copolymers can be found in Table 1.
Table 1. Polymer compositions and molecular weights, determined by
aqueous GPC.
Results and Discussion
Polymer/
monomer
1
2
3
4
5
Mw
Mn
R
G
N
G
ACHTREUNG
Synthesis: Preparation of the comonomer units is very
straightforward and involves one- to three-step procedures.
The dansyl unit is directly coupled via its chloride to ethyl-
enediamine, afterwards the second free amine reacts with
methacryloyl chloride yielding 1. For 2 and 3 mono-Boc-pro-
tected ethylenediamine (Boc=tert-butoxycarbonyl) is first
converted into its methacrylamide. Subsequent Boc depro-
tection furnishes 3, the final reductive amination of which
with o-formylboronic acid leads to 2.[9] Direct amidation
with methacrolyl chloride also affords the alkyl and the al-
cohol building blocks 4 and 5 (Figure 1). These monomers
were subjected to conventional radical copolymerizations
with AIBN (AIBN=2,2’-azoisobutyronitrile). In selected
cases, copolymerization parameters were determined by the
Fineman–Ross method[10] and found to be close to 1.0, en-
suring a statistical copolymerization.[11] The resulting poly-
6
7
8
9
0.3
0.5
0.7
0.8
1.0
1.0
1.0
1.0
2.0
–
2.0
2.0
–
–
–
0.7
–
116000 30000
129000 52000
126000 49000
4.0
4.0
4.0
n.d.[a]
n.d.[a]
[a] n.d.=not determined.
Binding studies: For a systematic study, selectivities of our
new hosts were always tested against the whole sugar series
by beginning with neutral dextran and ending with ovalbu-
min, a typical acidic protein. Direct comparison should thus
yield the selectivity for heparin, which carries most sulfate
groups (low molecular weight heparin (LMWH) was used in
this study, with a mean mass of 3 kD). Polymers were pre-
pared tightly packed with both comonomers, that is, boro-
nates and ammonium ions (6), but as a reference, a neutral
7702
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 7701 – 7707

A Fluorescent Polymeric Heparin Sensor
FULL PAPER
Table 2. Association constants determined by fluorescence titrations in
25 mm aqueous HEPES buffer between anionic biomolecules and poly-
meric hosts 6–8.
polymer was also synthesized, lacking any cationic group
(7). Its heparin affinity should reveal the importance of
Coulomb attraction versus boronate ester formation. Poly-
mer 8 dilutes both binding sites with an excess of amino-al-
cohol groups with the aim of generating more selectivity
and polymer 9 finally incorporates the unpolar dodecyl tail,
to probe the importance of hydrophobic interactions. All in-
vestigations were carried out in aqueous solution, containing
HEPES buffer (HEPES=4-(2-hydroxyethyl)-1-piperazine-
ethanesulfonic acid, 25 mm, pH 7.1). Initial experiments
demonstrated that the statistically distributed dansyl units
exhibited strong changes in their fluorescence emission in-
tensity, especially when mixed with binding partners of high
Polymers
Glycan/protein
dextran hyaluronic acid chondroitin heparin ovalbumine
6
7
8
3”103[a] 2”103
1:2[b]
1:9
<1”102 <1”102
4”106
1:7
3”107
1:6
1”106
1:3
6”103
1:4
4”105
1:5
2”104
1:2
–
–
<1”102 <1”102
2”106
1:10
2”107
1:8
7”104
1:5
–
–
[a] Errors are standard deviations and were calculated at 14–41%.
[b] Stoichiometries (bold) were determined from Job plots.
ing number of sulfate groups (Table 2). This trend was
common for all polymers 6–9. Heparin was always bound
one to five orders of magnitude tighter than chondroitin sul-
fate or hyaluronic acid (with KD values down to 30 nm).
Comparison between polymer 6 and 7, which differ mainly
in their cation content, reveals the importance of both bind-
ing sites for efficient sulfated sugar recognition. As a work-
ing model, it may be assumed that electrostatic contacts are
most favorable at the isolated carboxylate or O-sulfate at
the back of the heparin dimer in the 5-position, while the
front with its hydroxysulfate prefers interaction with a bor-
onic acid. Intriguingly, polymer 7 without any cationic bind-
ing site is still able to bind heparin with almost micromolar
affinity; however, it loses all of its binding power if the sul-
fate groups are removed as in dextran. Contrary to 6, it is
also much more selective with respect to chondroitin and
ovalbumin. Experiments with physiological salt loads sug-
gest that heparin binding by 6 relies strongly on Coulomb
interactions (Ka decreases by one order of magnitude),
while 7 predominantly uses covalent bonding (Ka unaffect-
ed). Polymer 8 retains a high affinity towards heparin, but is
much more selective than 6, thus confirming the dilution
concept. The tight packing of cationic groups in 6 facilitates
efficient Coulomb interactions with any anionic binding site
along a saccharide strand or on a protein surface, whereas
dilution with inert moieties forces both complex partners to
an induced-fit process in order to find complementary func-
tionalities. We assume a wrapping mechanism of the poly-
mer around the sugar guest, assuring a maximum number of
(non)covalent interactions, and thereby explaining the high
stoichiometry factors of five to eight LMWH molecules per
polymer. Polymer 9 was unfortunately producing insoluble
complexes with most glucosaminoglycans, which precipitat-
ed from aqueous solution.
Figure 2. a) Fluorescence quenching by addition of heparin (LMWH) to
polymer 6 in 25 mm HEPES buffer. The Ka value obtained from this
curve was calculated at 2.8”107 mÀ1 (stoichiometry HEP/6 6:1). b) Pro-
posed wrapping mechanism with favorable interactions between polymer
and heparin. The B···N interaction was not specified, it is most likely a
hydrated species.[15]
affinity (Figure 2). Intriguingly, the acidic protein produced
substantial fluorescence quenching, while most (albeit not
all) sugar experiments lead to a marked increase in fluores-
cence emission intensity. Fluorescence titrations were subse-
quently carried out at 510 nm and the resulting binding iso-
therms were analyzed by a standard 1:1 algorithm.[13] Molec-
ular weights were referenced to heparin (3 kD) and stoichio-
metries were determined separately by Job plots.[14]
Model studies: From the above-detailed observations, we
conclude that a remarkably stable complex must be formed
between a boronic acid and a sulfated aminoethanol or
glycol. To gain deeper insight into this interaction, we exam-
ined the model compound 10, a true heparin fragment, with
a 1:1 complex of phenylboronic acid and piperidine (Fig-
ure 3a). Addition of increasing amounts of the neat sulfated
sugar to the tetrahedral boronic acid amine complex result-
Thus dextran, a simple polyhydroxy sugar, was bound
only weakly, whereas affinities steadily rose with an increas-
Chem. Eur. J. 2007, 13, 7701 – 7707
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7703

T. Schrader et al.
rin appear in a highly cooperative fashion, leading to the ob-
served free binding enthalpy of ꢀ7.5 kcalmolÀ1. Anslyn
et al. argue along the same lines when they explain the dras-
tic efficiency increase for their second-generation heparin
receptor with enlarged cavity.[7]
Reversibility and calibration: The biological effect of hepa-
rin can be reversed by addition of protamine, a cationic pro-
tein, which is known to sequester the anionic polysugar and
release antithrombin III. To show that our heparin binding
is also a fully reversible process in spite of the formation of
covalent boronate ester bonds, we first prepared the com-
plex between heparin and polymer 1 and subsequently
added increasing amounts of protamine. Figure 4 demon-
Figure 3. a) Model complexation mixture, consisting of d-glucosamine 2-
sulfate sodium salt (10) and the preformed 1:1 aggregate from phenylbor-
onic acid and piperidine. b) Reference compound 11. c) Proposed struc-
ture of the preferred complex. Note that the boronic acid moiety forms a
seven-membered ring with the 3-OH group and the 2-aminosulfate.
1
ed in formation of a new set of H NMR spectroscopic sig-
nals in the aromatic region, with considerable shift differen-
ces to the starting complex (ortho protons d=0.10 ppm up-
field). This is typical for cyclic boronic acid esters, which are
usually formed in a kinetically slow process on the NMR
spectroscopic time scale.[16] In support of this assumption,
11B NMR spectra furnished two new broadened signals at
lower field (shifted from d=5.80 to 9.30 and 11.7 ppm, re-
spectively).[17] Interestingly, the sugar CH region also dis-
played a new set of NMR spectroscopic signals, with large
downfield shifts of d=0.30–0.90 ppm. COSY experiments
were very difficult to analyze unambiguously because of
substantial overlap of cross peaks. However, a (small) new
a-anomeric signal was found with a d=0.40 ppm downfield
shift, and at least one of the three methine protons was also
drastically shifted downfield. By contrast, the closely related
model-compound 11 with an N-acetyl group instead of the
sulfate did not produce significant complex peaks under the
same conditions. Obviously, the sulfate is important for com-
plexation, and the neighboring hydroxyl groups participate
in cyclic ester formation. Although we were unfortunately
not able to produce a clear NOE correlation due to the
broadened complex peaks, we tentatively suggest formation
of the cyclic seven-membered ester depicted in Figure 3a. A
molecular model can be constructed with small ring strain
and a potential S=O···H+ N stabilization. If this is correct,
À
the five-membered cyclic esters formed with a-hydroxycar-
boxylates can be extended to seven-membered cyclic esters
likewise formed with hydroxamine sulfates.
Figure 4. Reverting heparin binding by 6 upon titration with protamine.
a) Addition of eight aliquots of heparin to polymer 6 (33 nm), followed
by five aliquots of protamine: the fluorescence emission is fully restored.
b) Fluorescence emission quenching of 6 upon addition of increasing
amounts of heparin (30–220 nm). Subsequent protamine addition restores
the original emission intensity.
A closer inspection shows that the new CH signals form
3JH,H couplings among each other and with nonshifted CH
signals. From their distribution, a picture evolves, which fea-
tures all possible cyclic esters, although only the one depict-
ed in Figure 3a can exist in the heparin polymer. Each single
species is present only in a relatively small amount (7–10%)
and the signal ratio between 10 and all complexes furnishes
a virtual association constant of <70mÀ1. However, the re-
spective Ka value for cyclic boronic acid ester formation
with a-hydroxycarboxylates is in the same range
(ꢀ300mÀ1). As a consequence, it must be assumed that mul-
tiple ester formations between the polyboronate and hepa-
strates that the effect is completely reversed, and fluores-
cence emission intensity reaches the starting value. Simulta-
neously, the fluorescence of the heparin–protamine complex
appears at lower wavelengths (see the Supporting Informa-
tion). Consequently, our polymers imitate the reversible
binding mode of heparin found in nature.
7704
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 7701 – 7707

A Fluorescent Polymeric Heparin Sensor
FULL PAPER
1
a yellow solid. Yield: 2.58 g, 94%; H NMR (200 MHz, CDCl3): d=2.66–
2.72 (m, 2H), 2.88–2.92 (m, 8H), 7.19 (dd, 3J=7.61, 5J=0.59 Hz 1H),
7.49–7.60 (m, 2H), 8.17–8.26 (m, 2H), 8.51–8.58 ppm (m, 1H); 13CNMR
(75 MHz, CDCl3): d=41.0, 45.5, 45.6, 115.4, 118.9, 123.3, 128.5, 129.7,
129.8, 130.1, 130.5, 134.9, 152.2 ppm; HRMS (ESI, pos., CH2Cl2): m/z:
calcd for C14H19N3O2S: 293.1148 [M+H+]; found: 293.1157.
The high sensitivity of polymer 6 for heparin was used to
generate a calibration curve for quantitative measurements.
An almost perfect linearity was found for the concentration
range from 30 to 220 nm heparin (Figure 5). This markedly
N-[(5-N,N-Dimethylaminonaphthylsulfonylamino)ethyl]-2-methacryl-
amide (1):[8] 2-Dansylaminoethylamine (2.55 g, 8.68 mmol) and triethyl-
amine (1.33 mL, 9.55 mmol) were dissolved in dichlormethane (150 mL).
Methacryloyl chloride (0.91 mL, 9.55 mmol) in dichlormethane (50 mL)
was added dropwise to this solution. After 5 h, the solvent was distilled
off and the crude product was purified over silica gel by eluting with
ethyl acetate/hexane 1:1 (Rf =0.08) to afford a green-yellow product.
1
Yield: 2.54 g, 81%; H NMR (200 MHz, CDCl3): d=1.86 (s, 3H), 2.89 (s,
6H), 3.03–3.11 (m, 2H), 3.32–3.41 (m, 2H), 5.28 (s, 1H), 5.37 (t, 3J=
3
5
6.01 Hz, 1H), 5.61 (s, 1H), 6.22 (brs, 1H), 7.19 (dd, J=7.79, J=0.59 Hz,
1H), 7.49–7.61 (m, 2H), 8.21–8.25 (m, 2H), 8.48 ppm (d, 2J=8.80 Hz,
1H); 13C NMR (75 MHz, CDCl3): d=18.6, 43.3, 39.7, 45.6, 115.6, 119.1,
120.3, 123.4, 128.6, 129.6, 129.7, 130.0, 130.6, 134.7, 139.4, 169.2 ppm;
HRMS (ESI, pos. CH2Cl2): m/z: calcd for C18H23N3O3SNa: 384.1352
[M+Na+]; found: m/z: 384.1354.
(2-Aminoethyl)carbamic acid tert-butylester:[19] A mixture of di-tert-butyl
dicarbonate (5.00 g, 22.9 mmol) in chloroform (40 mL) was added drop-
wise to a solution of 1,2-ethylenediamine (12.4 g, 206 mmol) in chloro-
form (100 mL) at 08C. The mixture was then stirred overnight at room
temperature. Chloroform was evaporated and water (100 mL) was added
to the oily product. The insoluble bis-substituted diamine was removed
by filtration. The filtrate was extracted with dichloromethane (3”
100 mL). The organic layer was separated, dried over anhydrous Na2SO4,
and evaporated to dryness to give a yellow, oily product. Yield: 2.90 g
(18.1 mmol, 79%); 1H NMR (300 MHz, CDCl3): d=1.20 (s, 9H), 2.54 (t,
3J=6.09 Hz, 2H), 2.91 (m, 2H), 3.46 (s, 1H), 5.54 ppm (s, 1H); 13CNMR
(75 MHz, CDCl3): d=28.3; 41.6, 43.0, 78.8, 156.2 ppm; HRMS (ESI pos.,
CH2Cl2): m/z: calcd for C7H17N2O2: 183.1109 [M+Na+]; found: 183.1112.
N-{2-[(tert-Butoxycarbonyl)amino]ethyl}methacrylamide:[19] (2-Amino-
ethyl)carbamic acid tert-butylester (2.65 g, 16.5 mmol) and triethylamine
(6.67 mL, 47.8 mmol) were dissolved in chloroform (30 mL). Methacrylo-
yl chloride (1.83 g, 17.5 mmol) was dissolved in chloroform (20 mL) and
added dropwise to the above-described mixture over a period of 2 h at
08C. The reaction mixture was stirred for another 2 h at room tempera-
ture. The organic layer was extracted with water (5”30 mL) and dried
over anhydrous Na2SO4. Chloroform was evaporated and the product
was recrystallized from diethyl ether/hexane 3:5. Another recrystalliza-
tion step from chloroform/diethyl ether/hexane 1:55:55 yielded analytical-
ly pure product as a colorless solid. Yield: 2.26 g (9.90 mmol, 60%);
1H NMR (300 MHz, CDCl3): d=1.20 (s, 9H), 1.71 (s, 3H), 3.00 (t, 3J=
6.09 Hz, 2H), 3.11 (t, 3J=6.09 Hz, 2H), 4.25 (s, 2H), 5.13 (s, 1H),
5.50 ppm (s, 1H); HRMS (ESI pos., CH2Cl2): m/z: calcd for
C11H20N2O3Na: 251.1372 [M+Na+]; found: 251.1369.
Figure 5. Calibration curve displaying a linear correlation between fluo-
rescence emission intensity and heparin concentration from 30 nm to
0.22 mm heparin (25 mm HEPES buffer).
extends the values beyond the lower limits reported to date
for artificial heparin receptors. Due to the built-in fluores-
cence label, the new polymers seem to be ideal materials for
a heparin quantification in medicinal samples. Intravenous
or subcutaneous injection of heparin occurs at dosing levels
as low as 2 UmLÀ1 (800 nm) in surgery or emergency DVT,
and even reaches long-term levels of 0.2 UmLÀ1 (80 nm).
Thus, even the lowest clinically relevant dose can be quanti-
tatively measured with a simple fluorescence assay by using
a cheap copolymer from readily available building blocks.
Conclusion and Outlook
We conclude that fluorescent copolymers derived from rela-
tively simple building blocks can be tailored for the sensitive
and specific detection of sulfated sugars. Specifically, a neu-
tral polymer lacking any cationic groups was shown to inter-
act strongly with heparin, most likely due to formation of
cyclic esters with the aminoethanol and glycol sulfate moiet-
ies. The binding event can be totally reverted by protamine
and quantitatively observed in a medicinally useful concen-
tration range between 30 and 250 nm.
2-[(2-Methacrylamidoethylamino)methyl]phenylboronic acid (2):[16] N-(2-
Aminoethyl)methacrylamide hydrochloride (1.00 g, 6.07 mmol), 2-formyl-
boronic acid (0.91 g, 6.07 mmol), triethylamine (6.14 g, 60.7 mmol) and
molecular sieves 3 Š) were dissolved/suspended in absolute methanol
(5 mL) under a nitrogen atmosphere. The mixture was stirred for 2 h at
room temperature. Then sodium borohydride (0.23 g, 6.07 mmol) was
added at once. After an additional 1 h stirring, the mixture was filtered
over Celite and the filtrate was cooled in an ice bath and filtered again.
The second major batch of filtrate was concentrated and dried in vacuum
to afford a white product. Yield: 1.41 g (5.40 mmol, 89%); 1H NMR
(300 MHz, MeOD): d=1.97 (s, 3H), 3.05 (t, 3J=6.31 Hz, 2H), 3.61 (t,
3J=6.31 Hz, 2H), 4.08 (s, 2H), 5.42 (m, 1H), 5.72 (m, 1H), 7.16 (m, 3H),
7.20 ppm (m, 1H); 13CNMR (75 MHz, MeOD): d=10.2, 18.7, 38.3, 47.4,
55.3, 121.0, 123.9, 127.7, 128.5, 131.5, 141.0, 142.6, 171.7 ppm; HRMS:
(ESI pos., MeOH): calcd for C15H23BN2O3: m/z: 291.1802
Experimental Section
2-Dansylaminoethylamine:[18]
A solution of dansyl chloride (2.50 g,
9.26 mmol) in dichloromethane 40 mL was added dropwise to 1,2-ethyl-
enediamine (28.8 mL, 270 mmol) in dichloromethane (100 mL) while stir-
ring and cooling at 08C. The mixture was stirred while warming to room
temperature. It was subsequently acidified with HCl (1n) and then ex-
tracted with dichloromethane (3”20 mL). The aqueous layer was basidi-
fied (pH 9) by using NaOH (5n) and again extracted with dichlorome-
thane (2”20 mL). The organic layer was dried over Na2SO4, filtered
through a sinter, and the solvent removed under reduced pressure to give
[MÀ2H+2CH +H+]; found: 291.1878 (dimethyl ester formation).
3
N-(2-Aminoethyl)methacrylamide hydrochloride (3):[20] A solution con-
taining N-{2-[(tert-butoxycarbonyl)amino]ethyl}methacrylamide (2.07 g,
9.04 mmol) in CH2Cl2 (20 mL) and 2m HCl/Et2O (20 mL) was stirred at
Chem. Eur. J. 2007, 13, 7701 – 7707
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7705

T. Schrader et al.
room temperature for 24 h. After this time, the precipitated hygroscopic
salt was filtered and washed with diethyl ether to afford a white solid.
Yield: 1.21 g (7.32 mmol, 81%); 1H NMR (300 MHz, D2O): d=1.89 (s,
3H), 3.13 (t, 3J=5.76 Hz, 2H), 3.53 (t, 3J=6.09 Hz, 2H), 5.46 (s, 1H),
5.72 ppm (s, 1H); 13CNMR (75 MHz, D 2O): d=169.3, 140.4, 119.8, 42.7,
19.1 ppm; HRMS: (ESI pos., MeOH): calcd for C6H13N2O: m/z: 129.1028
[M+H+]; found: 129.1027.
formed by using nonlinear regression. The obtained values hence contain
two simplifying assumptions: 1) The polydisperse polymer was averaged
to a monodisperse host compound with uniform molecular weight corre-
sponding to MW. 2) All steps of a multistep binding event were assumed
to occur with equal affinity, that is without cooperativity. Binding con-
stants are therefore averaged over a broad molecular weight distribution
and diverging consecutive free binding energies.
Protamine back-titration: A heparin stock solution was added in 12 suc-
cessive aliquots to a solution of polymer 6 (4”10À8 m) until a 12-fold hep-
arin excess was reached (5”10À7 m), then a protamine stock solution was
added in eight successive aliquots to the final mixture, until the fluores-
cence emission quenching was almost totally reversed.
Dodecylmethacrylamide (4):
A solution of methacryloyl chloride
(0.31 mL, 3.22 mmol) in dichloromethane (10 mL) was added dropwise
into dodecylamine (500 mg, 2.69 mmol) and triethylamine (0.45 mL,
3.22 mmol) in dichlormethane (50 mL). Then the crude was washed with
NaOH (1n, 3”50 mL) and with HCl (1n, 3”50 mL). The organic layer
was dried over MgSO4 and condensed to give a colorless product. Yield:
640 mg, 97%; 1H NMR (200 MHz, CDCl3): d=0.87 (t, 3J=6.39 Hz, 3H),
1.12–1.52 (m, 20H), 1.96 (s, 3H), 3.25–3.35 (m, 2H), 5.30 (s, 1H), 5.66 (s,
1H), 5.77 ppm (brs, 1H); 13CNMR (75 MHz, CDCl 3): d=14.2, 18.8, 22.8,
27.1, 29.4, 29.6, 29.7, 29.8, 32.0, 40.0, 119.4, 140.3, 168.7 ppm; HRMS
(ESI, pos., CH2Cl2): m/z: calcd for C16H32NO: [M+H+] 254.2478; found:
254.2479.
Calibration: A 13 mm heparin solution containing polymer 6 (32 nm) was
added in eight aliquots (2 mL, containing 32 nmol heparin) to a solution
of polymer 6 (32 nm, 800 mL) in a cuvette. Final heparin concentrations
ranged from 32 to 256 nm. Fluorescence emission intensity was recorded
for all solutions and a reference solution containing only the fluorescent
polymer. The first eight data points roughly formed a straight line (0–
224 nm).
Methacryloylamino-2-hydroxypropane (5):[21] A solution of methacryloyl
chloride (2.72 g, 26.0 mmol) in dry dichloromethane (40 mL) was added
dropwise to a mixture of 1-aminopropan-2-ol (4.21 g, 56.1 mmol) in dry
dichloromethane (40 mL) at 08Cunder an argon atmosphere. The precip-
itating solid was filtered off and the solvent was removed under reduced
pressure. After purification by chromatography over silica gel, eluting
with dichloromethane/methanol 14:1 v/v (Rf =0.32) a colorless solid was
obtained. Yield: 1.80 g (12.6 mmol; 48%); 1H NMR (200 MHz, CDCl3):
d=1.21 (d, 2J=6.4 Hz, 3H), 1.98 (dd, 4J=1.5, 1.0 Hz, 3H), 2.51 (brs,
1H), 3.18 (ddd, 2J=14.0, 3J=7.5, 5.3 Hz, 1H), 3.51 (ddd, 2J=14.0, 3J=
6.5, 3.0 Hz, 1H), 3.96 (dqd, 3J=7.5, 6.4, 3.0 Hz, 1H), 5.36 (qd, 4J=1.5,
2J=1.4 Hz, 1H), 5.74 (dq, 2J=1.4, 4J=1.0 Hz, 1H), 6.38 ppm (brs, 1H).
Acknowledgement
This work was funded by the DFG (Deutsche Forschungsgemeinschaft).
[1] a) T. D. James, K. R. A. S. Sandanayake, R. Iguchi, S. Shinkai, J.
Am. Chem. Soc. 1995, 117, 8982; b) T. D. James, S. Shinkai, Top.
Curr. Chem. 2002, 218, 159; c) S. L. Wiskur, J. J. Lavigne, H. Ait-
Haddou, V. Lynch, Y. H. Chiu, J. W. Canary, E. V. Anslyn, Org. Lett.
2001, 3, 1311–1314.
Polymerizations:
A solution/suspension containing a combination of
[2] Notable exceptions: a) S. Patterson, B. D. Smith, R. E. Taylor, Tetra-
hedron Lett. 1998, 39, 3111; b) E. Uchimura, H. Otsuka, T. Okano,
Y. Sakurai, K. Kataoka, Biotechnol. Bioeng. 2001, 72, 307.
[3] a) K. Burgess, A. M. Porte, Angew. Chem. 1994, 106, 1218–1220;
Angew. Chem. Int. Ed. Engl. 1994, 33, 1182–1184; b) S. L. Wiskur,
J. J. Lavigne, A. Metzger, S. L. Tobey, V. Lynch, E. V. Anslyn, Chem.
Eur. J. 2004, 10, 3792 À3804.
monomers, 1, 2, 3, 4, and 5, and a catalytic amount of AIBN (about
1 mol% of all the monomers, see the Supporting Information) in DMF
was degassed and stirred for 20–50 h at 608C. The reaction mixture was
diluted with methanol to give a maximum concentration of 5% of the in-
itial monomer mass and was added dropwise to the 10-fold volume of
ethyl acetate. The precipitating solid was collected by filtration, washed
with ethyl acetate and dried in vacuo. Each polymer was routinely char-
1
[4] a) C. Gray, Jr., T. A. Houston, J. Org. Chem. 2002, 67, 5426–5428;
b) Y. Yang, P. T. Lewis, J. O. Escobedo, N. N. St. Luce, W. D. Tre-
leaven, R. L. Cook, R. M. Strongin, Collect. Czech. Chem. Commun.
2004, 69, 1282–1291; c) H. Otsuka, E. Uchimura, H. Koshino, T.
Okano, K. Kataoka, J. Am. Chem. Soc. 2003, 125, 3493.
[5] J. Hirsh, J. E. Dalen, D. Deykin, L. Poller, Chest 1992, 102, 337S-
352S.
acterized by its H NMR spectrum and molecular weight (MW/Mn).
Fluorescence titrations: The general titration procedure follows refer-
ence [21] published for proteins. A solution of the polymer was prepared
in aqueous HEPES buffer (25 mm, pH 7.12, c=5.0”10À9–4.3”
10À6 molLÀ1). The guest stock solutions were prepared by dissolving
them in this buffered polymer solution (c=1.59”10À5–8.3”10À4 molLÀ1
)
to guarantee a constant polymer concentration during the entire titration.
The polymer solution was filled into a stirrable cuvette and the stock so-
lution was added stepwise. The samples were excited at a wavelength of
330 nm and the change of the emission intensity at 510 nm was recorded.
[6] a) G. P. Gravlee, L. D. Case, K. C. Angert, A. T. Rogers, G. S. Miller,
Anesth. Analg. 1988, 67, 469–472; b) J. Umlas, R. H. Taff, G.
Gauvin, P. Sweirk, Anesth. Analg. 1983, 62, 1095–1099; c) S. Mathi-
son, E. Bakker, Anal. Chem. 1999, 71, 4614–4621; d) T.-J. Cheng, T.-
M. Lin, T.-H. Wu, H.-C. Chang, Anal. Chim. Acta 2001, 432, 101–
111; e) N. Ramamurthy, N. Baliga, T. W. Wakefield, P. C. Andrews,
V. C. Yang, M. E. Meyerhoff, Anal. Biochem. 1999, 266, 116–124.
[7] a) Z. Zhong, E. V. Anslyn, J. Am. Chem. Soc. 2002, 124, 9014–9015;
b) A. T. Wright, Z. Zhong, E. V. Anslyn, Angew. Chem. 2005, 117,
5825-5828; Angew. Chem. Int. Ed. 2005, 44, 5679–5682.
[8] S. Koch, C. Renner, X. Xie, T. Schrader, Angew. Chem. 2006, 118,
6500–6503; Angew. Chem. Int. Ed. 2006, 45, 6352–6355.
[9] C. W. Gray, Jr., T. A. Houston, J. Org. Chem. 2002, 67, 5426–5428.
[10] M. Fineman, S. D. Ross, J. Polym. Sci. 1950, 5, 259–262.
[11] C. Renner, J. Piehler, T. Schrader, J. Am. Chem. Soc. 2006, 128,
4831–4841.
Stoichiometries were determined as follows: Dextranꢀs exact molecular
weight was known as 100000. For the anionic sugars, LMWH was used as
a reference with a known molecular weight of 3000 D. As the other relat-
ed polymeric sugars consist of very similar disaccharide repeat units,
their (unknown) molecular weight was also set to 3000, and corrected by
the relative ratio between their and heparinꢀs repeat unit. The exact stoi-
chiometry between synthetic host (100 kD) and sugar (3 kD unit) was
calculated from Job plots. Stoichiometries were always between 1:2 and
1:10 (synthetic host/sugar). In this fashion exact ratios could be deter-
mined much more precisely than from virtual disaccharide repeats with
their low molecular weights of 0.6 kD. The repeat units for dextran, hya-
luronic acid, chondroitin sulfate, and heparin have the following molecu-
lar weights: dextran: 324 D, hyaluronic acid: 392 D, chondroitin sulfate:
471 D, and heparin: 587 D.
[12] We thank the analytical department of the Institute of Technical
Chemistry at the University of Duisburg, Essen, for GPC measure-
ments.
[13] H. J. Schneider, R. Kramer, S. Simova, U. Schneider, J. Am. Chem.
Soc. 1988, 110, 6442–6448.
Samples of constant polymer concentration were titrated with increasing
amounts of sugar, and the drastic change in fluorescence emission inten-
sity was detected. From a Job plot, the exact stoichiometry was deter-
mined and used as a correction factor for the host concentration. A sub-
sequent routine fitting procedure to a 1:1 binding isotherm was per-
[14] a) P. Job, Compt. Rend. 1925, 180, 928–930; b) M. T. Blanda, J. H.
Horner, M. Newcomb, J. Org. Chem. 1989, 54, 4626–4636.
7706
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 7701 – 7707

A Fluorescent Polymeric Heparin Sensor
FULL PAPER
ˇ
ˇ
´
[15] L. Zhu, S. H. Shabbir, M. Gray, V. M. Lynch, S. Sorey, E. V. Anslyn,
J. Am. Chem. Soc. 2006, 128, 1222–1232.
[19] T. Reschel, C. Konµk, D. Oupicky, L. W. Seymour, K. Ulbrich, J.
ControlledRelease 2002, 81, 201–217.
[16] M. Maue, T. Schrader, Angew. Chem. 2005, 117, 2305–2310; Angew.
Chem. Int. Ed. 2005, 44, 2265–2270.
[17] K. Djanashvili, L. Frullano, J. A. Peters, Chem. Eur. J. 2005, 11,
4010–4018.
[20] D. Spivak, K. J. Shea, J. Org. Chem. 1999, 64, 4627–4634.
[21] C. Renner, J. Piehler, T. Schrader, J. Am. Chem. Soc. 2006, 128,
4831–4841.
[18] E. L. Doyle, C. A. Hunter, H. C. Phillips, S. J. Webb, N. H. Williams,
J. Am. Chem. Soc. 2003, 125, 4593–4599.
Received: May 3, 2007
Published online: June 28, 2007
Chem. Eur. J. 2007, 13, 7701 – 7707
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7707