Hydrophilic cholesterol-binding molecular imprinted polymers

Article abstract of DOI:10.1016/S0040-4039(01)00045-4

Novel hydrophilic molecularly imprinted polymers (MIPs) were prepared using acryloyl-containing hydrophilic monomers (including cholesteryl acrylate (CA) and acryloyl-6-amino-6-deoxy-β (or γ)-cyclodextrin), and their steroid-binding properties were analyzed in 2-PrOH by HPLC. The MIPs bound 38-50 μmol of cholesterol/g, while the corresponding nonimprinted control polymers (containing no CA) bound only 6-9 μmol cholesterol/g. The sterol-binding selectivity was illustrated with estrone, which was bound by MIPs in the range 5-36 μmol/g.

Full text of DOI:10.1016/S0040-4039(01)00045-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 1839–1841
Hydrophilic cholesterol-binding molecular imprinted polymers
Ning Zhong, Hoe-Sup Byun and Robert Bittman*
Department of Chemistry and Biochemistry, Queens College of The City University of New York, Flushing,
NY 11367-1597, USA
Received 21 December 2000; revised 5 January 2001; accepted 9 January 2001
Abstract—Novel hydrophilic molecularly imprinted polymers (MIPs) were prepared using acryloyl-containing hydrophilic
monomers (including cholesteryl acrylate (CA) and acryloyl-6-amino-6-deoxy-b (or g)-cyclodextrin), and their steroid-binding
properties were analyzed in 2-PrOH by HPLC. The MIPs bound 38–50 mmol of cholesterol/g, while the corresponding
nonimprinted control polymers (containing no CA) bound only 6–9 mmol cholesterol/g. The sterol-binding selectivity was
illustrated with estrone, which was bound by MIPs in the range 5–36 mmol/g. © 2001 Elsevier Science Ltd. All rights reserved.
Molecularly imprinted polymers (MIPs)1 that selec-
tively recognize small molecules have a wide range of
pharmaceutical, analytical, and biological applica-
tions.2–4 The imprinting method is based on the tem-
plate approach in which a target molecule (the
template) recognizes its own recognition site by inter-
acting with complementary functional groups of appro-
priate polymerizable monomers. Removal of the
template yields polymers that are arranged complemen-
tary to the template’s structure, resulting in selective
recognition of the target molecule. Most MIPs prepared
to date feature a high degree of crosslinking in the
polymer network, which enables the cavities to retain
their shape after template removal, giving rise to good
template binding selectivity;1 however, a high content
of crosslinking agent makes it difficult to bind water-
soluble compounds5 and to remove imprinted
molecules.6 In this report we describe the preparation
of water-soluble MIPs in which some of the crosslink-
ing agent N,N%-diacryloylpiperazine (DAPA) is
replaced with a hydrophilic monomer, ethyl 2-hydroxy-
methylacrylate (EHMA).7 Cholesteryl acrylate (CA)
template with or without CD derivatives as monomers
were used to form novel hydrophilic MIPs in aqueous
media via the covalent imprinting methodology. The
MIPs bind selectively to sterols in 2-PrOH after the
template is removed. A possible application of such
polymers is to modulate the flux of cholesterol between
cell membranes and lipoproteins, thus altering choles-
Figure 1.
Scheme 1.
Abbreviations: MIPs, molecularly imprinted polymers; CD, cyclodextrin; EHMA, ethyl 2-hydroxymethylacrylate; AA, acrylic acid; CA, cholesteryl
acrylate; PAA, N-propylacrylamide; AABCD, acryloyl-6-amino-6-deoxy-b-CD; AAGCD, acryloyl-6-amino-6-deoxy-g-CD; DAPA, N,N%-diacry-
loylpiperazine; TMEDA, N,N,N%,N%-tetramethylethylenediamine.
Keywords: cholesterol; cyclodextrin.
* Corresponding author. Tel.: 718-997-3279; fax: 718-997-3349; e-mail: robert bittman@qc.edu
–
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)00045-4

1840
N. Zhong et al. / Tetrahedron Letters 42 (2001) 1839–1841
Table 1. Polymer compositionsa
Polymer
CA
AABCD
AAGCD
AA
PAA
EHMA
P1
0.20
0.20
0.00
0.20
0.20
0.00
0.54
0.00
0.00
0.00
0.54
0.00
0.00
0.00
0.00
0.54
0.00
0.00
0.00
0.54
0.74
0.00
0.00
0.00
0.00
0.00
0.00
0.00
14.00
14.74
14.00
14.00
14.00
14.00
0.00
P2
P3b
P4
P5
P6b
0.00
a All concentrations are in mmol. Porogens were H2O and THF (400 and 88 mmol, respectively). Initiators were TMEDA and (NH4)2S2O8 (0.56
and 0.44 mmol, respectively). The crosslinking agent was DAPA (1.50 mmol).
b Nonimprinted polymers.
Table 2. Properties of polymersa
Polymer
Water solubility
Cholesterol bound (mmol/g of MIP)c Estrone bound (mmol/g of MIP)c
(g/100 mL)b
Initial cholesterol concentration
Initial estrone concentration
0.5 mg/mL
1.0 mg/mL
1.0 mg/mL
P1 (CA+AABCD+EHMA)
P2 (CA+AA+EHMA)
P3 (AA+EHMA)
P4 (CA+AAGCD+EHMA)
P5 (CA+AABCD+PAA)
P6 (PAA)
5.0
5.0
5.0
5.0
1.3
0.6
2494
1591
293
3092
2795
593
4395
3893
693
5092
4193
992
2492
595
295
3698
25911
1394
a Templates were removed by treating with 5N NaOH at 5°C for 7 days.
b Data are the average of two measurements.
c The results shown are the mean of three independent measurements. The amounts of the bound ligands per g of polymer were calculated
according to A×C×103/(2×MW) (mmol/g). Absorption A was calculated according to A=(C−y)/C×100%. C is the initial ligand concentration
(mg/mL). y was obtained from y=bx, the standard curve equation (b=0.1931 for cholesterol, b=0.1475 for estrone), x is ligand peak area.13
terol transport and absorption. Many other synthetic
receptors that encapsulate cholesterol and other apolar
compounds, such as double-decker cyclophanes, would
have limited biological utility because of their very low
aqueous solubility.8
was prepared to compare with the non-PAA imprinted
P1. Polymer P6, in which monomer PAA was used
instead of monomers AA and EHMA, was synthesized
without the template, and is to be compared with the
nonimprinted polymer P3. Polymer P4, in which
monomer AAGCD was used instead of monomer
AABCD, was synthesized like P1 (Table 1).
The syntheses of monomers CA (Fig. 1)9 and PAA10
are straightforward. EHMA was prepared as described
previously.7 AABCD11 and AAGCD12 were prepared
as outlined in Scheme 1. The polymers were prepared
by (1) formation of monomeric complexes between CA
and acryloyl-CD and (2) polymerization using a low
degree of crosslinking agent and 1:1 THF/H2O as poro-
gens. A low degree (9.2 mol%) of crosslinking agent
(DAPA) was used to prepare polymers P1–P6 (Table
1). The content of crosslinking agents employed in
previous studies of cholesterol-imprinted polymers was
much higher, e.g. 83 mol% crosslinking agent was used
to prepare molecularly imprinted CDs4 and 95 mol%
was used in other covalently imprinted polymers.2 Poly-
mer P1 was prepared using the covalent imprinting
technique with CA as the template and AABCD as one
of the monomers. Polymer P2 was also prepared using
the covalent imprinting technique, but AABCD was
replaced by AA. Polymer P3, in which AA was used
instead of CA and/or AABCD, was prepared as a
control for P1 and P2 (Table 1). Polymer P5, in which
monomer PAA was used instead of monomer EHMA,
Before the template was removed P1, P2, P4, and P5
were practically insoluble in water. Cleavage of the
template enhanced the water solubility of polymers P1,
P2, and P4 (5.0 g/100 mL, Table 2). The water solubil-
ity of P5 was lower (1.3 g/100 mL), perhaps because of
the introduction of the amide-containing monomer
PAA instead of the ester-containing monomer EHMA.
During template hydrolysis, the esters are converted to
free acids, but the amide is stable. The free acids and
hydroxymethyl groups present in the matrix of P1–P4
apparently contribute to the water solubility properties
of the polymers.
The binding data of the MIPs to cholesterol and
estrone (see structures, Fig. 1) are summarized in Table
2. In MIPs P1, P2, P4, and P5, 38–50 mmol cholesterol
was bound per gram of polymer. The corresponding
nonimprinted control polymers (P3 and P6) contained
only 6–9 mmol cholesterol bound per gram of polymer.
The MIPs bound estrone in the range 5–36 mmol/g of

N. Zhong et al. / Tetrahedron Letters 42 (2001) 1839–1841
1841
polymer. The presence of CD, which is known to bind
to steroids,14 may account for the wide range of bound
estrone; the non-CD containing MIPs (such as P2) bound
less estrone (5 mmol/g). Decreasing the initial cholesterol
concentration by one-half (0.5 mg/mL) resulted in a
decrease in cholesterol bound in MIPs by about one-half
(Table 2). P4, which contains a g-CD15 residue, bound
slightly more cholesterol (50 mmol/g) than P1, which
contains a b-CD residue (Table 2). The larger cavity
monomer is apparently more favorable for binding
cholesterol in the polymer matrix. MIP P2, which does
not contain any CD residues, took up less cholesterol
than MIPs P1 and P4, indicating that CD residues allow
MIPs to take up cholesterol more efficiently.
J=4.4 Hz), 4.70 (m, 1H), 2.34 (d, 2H, J=8.0 Hz),
0.90–2.10 (m, 26H); 1.01 (s, 3H), 0.89 (d, 3H, J=6.4 Hz),
0.84 (d, 6H, J=6.4 Hz), 0.66 (s, 3H); 13C NMR l 165.6,
139.6, 130.3, 129.0, 122.7, 74.1, 56.7, 56.1, 50.0, 42.3, 39.7,
39.5, 38.1, 35.8, 31.9, 31.8, 28.2, 28.0, 27.7, 24.3, 23.8, 22.8,
22.6, 21.0, 18.7, 11.8; MS calcd for C30H48O2 (M−H)+ m/z
440.71, found 440.30.
10. Monomer PAA was synthesized in 88% yield as a light-yel-
low liquid by reaction of n-PrNH2 with acryloyl chlor-
ide in CH2Cl2 (rt, 30 min); Rf 0.15 (hexane/EtOAc 1:1); 1H
NMR (CDCl3) l 6.29 (d, 1H, J=16.8 Hz), 6.13 (ddd, 1H,
J=2.8, 10.4, 16.8 Hz), 5.87 (br s, 1H); 5.64 (d, 1H, J=10.4
Hz), 3.29 (q, 2H, J=6.6 Hz), 1.53–1.62 (m, 2H), 0.95 (t,
3H, J=7.2 Hz); 13C NMR l 165.9, 131.2, 126.4, 41.6, 23.0,
21.3, 14.4, 11.6; MS calcd for C6H11NO (M+) m/z 113.16,
found 113.10.
11. Monomer AABCD was obtained as a white powder in 84%
yield by reaction of 6-deoxy-6-amino-b-CD16 with acryloyl
chloride and NaHCO3 in H2O, followed by passage
through a Sephadex G-25 column (eluted with H2O) and
precipitation from MeOH; Rf 0.39 (n-PrOH/H2O/EtOAc/
conc. NH4OH 5:3:1:1); 1H NMR (D2O) l 6.17 (d, 1H,
J=10.2 Hz), 6.07 (d, 1H, J=17.1 Hz), 5.66 (d, 1H, J=10.2
Hz), 4.95 (br s, 7H), 3.13–3.85 (m, ꢀ42H); 13C NMR
(D2O) l 169.37, 130.79, 128.52, 102.78, 102.41, 84.28,
82.01, 73.96, 72.95, 72.67, 72.54, 71.06, 61.15, 60.61, 49.80.
Electrospray MS calcd for C45H74NO35 (M+H)+ m/z
1188.4, found 1188.2.
To examine whether the amide functionality in the
polymers affects the template binding, PAA was intro-
duced into P5 and P6. The amount of bound cholesterol
in P5 (41 mmol/g) was the same as that in the correspond-
ing non-PAA imprinted polymer P1 (43 mmol/g) (Table
2). Nonimprinted P6 bound slightly more cholesterol
than nonimprinted control P3. These data indicate that
the amide functionality did not affect the template
binding in MIPs.
Acknowledgements
This work was supported by NIH Grant HL-16660.
References
12. Monomer AAGCD was prepared from 6-deoxy-6-amino-
g-CD17 in a similar fashion as AABCD; 81% yield, white
powder; Rf 0.38 (n-PrOH/H2O/EtOAc/conc. NH4OH
1
5:3:1:1); H NMR (D2O) l 6.17 (d, 1H, J=10.3 Hz), 6.09
(d, 1H, J=17.1 Hz), 5.66 (d, 1H, J=10.1 Hz), 4.98 (br s,
8H), 3.32–3.92 (m, 48H); 13C NMR (D2O) l 167.02, 128.21,
126.22, 100.11, 99.72, 80.88, 78.88, 78.68, 71.36, 70.74,
70.18, 70.00, 68.49, 58.65, 58.24, 47.33. Electrospray MS
calcd for C51H84NO40 (M+H)+ m/z 1350.5, found 1351.0.
13. For example, the peak area of cholesterol in the P1 sample
(for initial cholesterol concentration of 1.0 mg/mL) was
3.443; therefore, y=0.1931×3.443=0.665; A=[(1.0−
0.665)/1.0]×100%=33.5%. The samples were filtered
through a 0.45 mm Teflon® Cameo filter to remove the
polymers prior to HPLC analysis. Cholesterol was ana-
lyzed on a BioCad Sprint Chromatography System, C18
(Hewlett Packard) column (4.6×100 mm), mobile phase
100% 2-PrOH, flow rate 0.8 mL/min, Sedex Model 55
evaporative light-scattering detector. Estrone was mea-
sured on a HP Model 1050 HPLC, Prodigy (Phenomenex)
silica column (4.6×150 mm), mobile phase 100% 2-PrOH,
flow rate 0.8 mL/min, detection at 254 nm.
1. Wulff, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 1812–
1832.
2. Whitcombe, M. J.; Rodriguez, M. E.; Villar, P.; Vulfson,
E. N. J. Am. Chem. Soc. 1995, 117, 7105–7111.
3. Asanuma, H.; Kakazu, M.; Shibata, M.; Hishiya, T.;
Komiyama, M. J. Chem. Soc., Chem. Commun. 1997,
1971–1972.
4. Hishiya, T.; Shibata, M.; Kakazu, M.; Asanuma, H.;
Komiyama, M. Macromolecules 1999, 32, 2265–2269.
5. The water-in-oil emulsion method was developed to handle
water-soluble compounds at an aqueous–organic interface,
but it is difficult to firmly fix the recognition site on the
polymer surface. See: Yoshida, M.; Ueza, K.; Goto, M.;
Furusaki, S. Macromolecules 1999, 32, 1237–1243.
6. It was noted that base-catalyzed hydrolysis does not cleave
the template from divinylbenzene-based polymers
imprinted with sterol methacrylate esters: Bystro¨m, S. E.;
Bo¨rje, A.; Akermark, B. J. Am. Chem. Soc. 1993, 115,
2081–2083.
14. 1:1 and 1:2 steroid (cortisone, hydrocortisone, proges-
terone, testosterone)/CD inclusion complexes have been
reported: Liu, F.-Y.; Kildsig, D. A.; Mitra, A. K. Pharm.
Res. 1990, 7, 869–873.
7. Byun, H.-S.; Reddy, K. C.; Bittman, R. Tetrahedron Lett.
1994, 35, 1371–1374.
15. Steroid/g-CD complexes: Cserha´ti, T.; Forga´cs, E. Eur. J.
8. (a) Peterson, B. R.; Mordasini-Denti, T.; Diederich, F.
Chem. Biol. 1995, 2, 139–146; (b) Marti, T.; Peterson, B.
R.; Fu¨rer, A.; Mordasini-Denti, T.; Zarske, J.; Jaun, B.;
Diederich, F. Helv. Chim. Acta 1998, 81, 109–144.
9. Monomer CA was prepared by reaction of cholesterol with
acryloyl chloride in pyridine/CH2Cl2 (rt, 30 min); 60%
yield, white powder; Rf 0.40 (hexane/EtOAc 50:1); 1H
NMR (CDCl3) l 6.37 (d, 1H, J=17.2 Hz), 6.06 (dd, 1H,
J=10.4, J=17.2 Hz), 5.79 (d, 1H, J=10.4 Hz), 5.37 (d, 1H,
Pharm. Biopharm. 1998, 46, 153–159.
16. Christian, A. E.; Byun, H.; Zhong, N.; Wanunu, M.; Marti,
T.; Fu¨rer, A.; Diederich, F.; Bittman, R.; Rothblat, G. H.
J. Lipid Res. 1999, 40, 1475–1482. MALDI MS calcd for
C42H71NNaO34 (M+Na)+ m/z 1156.3755, found
1156.2832; MALDI MS calcd for C42H71KNO34 (M+K)+
m/z 1172.3495, found 1172.2473.
17. Hamada, F.; Murai, K.; Ueno, A.; Suzuki, I.; Osa, T. Bull.
Chem. Soc. Jpn. 1988, 61, 3758–3760.