Synthesis and ligand recognition of paracetamol selective polymers: Semi-covalent versus non-covalent molecular imprinting

Article abstract of DOI:10.1039/b900014c

Three molecular imprinting strategies, each based upon a series of ethylene glycol dimethacrylate (EGDMA) cross-linked co-polymers, have been used to produce materials selective for the commonly used analgesic and antipyretic agent paracetamol (p-acetaminophen or 4-acetamidophenol) (1). The polymers were synthesised using either a semi-covalent imprinting strategy based upon 4-acetamidophenyl-(4-vinylphenyl) carbonate (4) or a non-covalent strategy based on methacrylic acid (MAA) as the functional monomer, or by employing a combination of these strategies. Radioligand binding studies demonstrated low template affinity in polymers offering only a single electrostatic interaction point for recognition via the phenolic residue in the template, whereas binding was substantially increased upon the introduction of a second binding mode, namely interaction at the acetamide moiety. HPLC analyses revealed no imprinting effect in the purely semi-covalent system, and only a minor effect in the purely non-covalent systems. However, a pronounced imprinting effect was demonstrated for polymers prepared by a combination of semi-covalent and non-covalent imprinting. This study illustrates a limitation of both the non-covalent and the semi-covalent strategies when it comes to achieving imprinted selectivity for small and poorly functionalised templates such as paracetamol. Parallels with conclusions from studies with antibodies are discussed.

Full text of DOI:10.1039/b900014c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Synthesis and ligand recognition of paracetamol selective polymers:
semi-covalent versus non-covalent molecular imprinting†
Jenny P. Rosengren-Holmberg, Jesper G. Karlsson, Johan Svenson,‡ Ha˚kan S. Andersson and Ian A. Nicholls*
Received 5th January 2009, Accepted 8th May 2009
First published as an Advance Article on the web 11th June 2009
DOI: 10.1039/b900014c
Three molecular imprinting strategies, each based upon a series of ethylene glycol dimethacrylate
(EGDMA) cross-linked co-polymers, have been used to produce materials selective for the commonly
used analgesic and antipyretic agent paracetamol (p-acetaminophen or 4-acetamidophenol) (1). The
polymers were synthesised using either a semi-covalent imprinting strategy based upon
4-acetamidophenyl-(4-vinylphenyl) carbonate (4) or a non-covalent strategy based on methacrylic acid
(MAA) as the functional monomer, or by employing a combination of these strategies. Radioligand
binding studies demonstrated low template affinity in polymers offering only a single electrostatic
interaction point for recognition via the phenolic residue in the template, whereas binding was
substantially increased upon the introduction of a second binding mode, namely interaction at the
acetamide moiety. HPLC analyses revealed no imprinting effect in the purely semi-covalent system, and
only a minor effect in the purely non-covalent systems. However, a pronounced imprinting effect was
demonstrated for polymers prepared by a combination of semi-covalent and non-covalent imprinting.
This study illustrates a limitation of both the non-covalent and the semi-covalent strategies when it
comes to achieving imprinted selectivity for small and poorly functionalised templates such as
paracetamol. Parallels with conclusions from studies with antibodies are discussed.
advantage, the issue of stability (shelf-life) cannot be ignored. The
Introduction
antibody-like recognition characteristics9–11 and general stability12
Paracetamol (p-acetaminophen or 4-acetamidophenol) (1) (Fig. 1)
has become one of the most frequently used analgesic and
of molecularly imprinted polymers (MIPs) make the development
of paracetamol-selective MIPs attractive for potential use in sensor
antipyretic agents since its introduction in the mid 1950s as a
devices for measurement of blood concentrations, or as an SPE
replacement for phenacetin.1 Although it is considered to be quite
matrix for the extraction of 4-acetamidophenol from blood.
safe, it has become one of the most common drugs involved in ac-
Previous attempts to create polymers capable of recognising
cidental or intentional cases of overdose which, if not treated early,
4-acetamidophenol have relied upon the use of non-covalent im-
may result in severe liver damage.2 Efforts to develop methods for
the rapid determination of paracetamol and its metabolites have
long been a focus for researchers, with antibody-based strategies
initially investigated,3–5 though more recently a range of highly
sensitive instrumental techniques has been used.6–8 While the selec-
tivities generally offered by antibody-based approaches are an
printing strategies and have been only moderately successful.13–15
In this study, our objective was to examine the efficiency of
a series of molecular imprinting systems to determine whether
improved MIP materials for paracetamol could be produced, and
to simultaneously gain insights concerning the factors steering
polymer recognition of small and poorly functionalised systems.
Polymers reported by Yang and Li using 4-vinylpyridine
(4-VP) as a functional monomer demonstrated template affinity,
though only a minor imprinting effect was noted.13 Tan et al.
demonstrated that an increase in both the selectivity and sensitivity
of 4-acetamidophenol imprinted bulk acoustic wave sensors could
be obtained with polymers based upon two equivalents of MAA
in combination with two equivalents of 4-VP (relative to the
template). This polymer performed better than those synthesised
using four equivalents of either one of the functional monomers.14
The relative success of the dual functional monomer system was
attributed to two favorable processes, the ability of the basic 4-VP
to interact more strongly with the weakly acidic phenol residue on
the template, and MAA’s interaction with the amide functionality.
Interestingly, the involvement of both of these functionalities has
been implicated in antibody recognition of 4-acetamidophenol.5
A paracetamol imprinted polymer prepared by the electro-co-
polymerisation of aniline and o-phenylenediamine has also been
examined in conjunction with electrochemical sensor studies.8
Fig. 1 Structures of 4-acetamidophenol (1) and the monomer 4-acetami-
dophenyl-(4-vinylphenyl) carbonate (4).
Bioorganic and Biophysical Chemistry Laboratory, School of Pure and
Applied Natural Sciences, University of Kalmar, SE-391 82 Kalmar, Sweden.
E-mail: ian.nicholls@hik.se; Fax: +46 480 446244; Tel: +46 480 446258
† Electronic supplementary information (ESI) available: Spectra for syn-
thesis products, and data from polymer physical characterization, radioli-
gand binding studies and NMR titrations. See DOI: 10.1039/b900014c
‡ Current address: Department of Chemistry, University of Tromsø, N-
9037 Tromsø, Norway. Fax: +47 776 44765; Tel: +47 776 45505; E-mail:
johan.svenson@uit.no
3148 | Org. Biomol. Chem., 2009, 7, 3148–3155
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Fig. 2 Schematic illustration of the various binding cavities in the different polymers studied. (P1) 4-acetamidophenyl-(4-vinylphenyl) carbonate
polymer; (P2) phenyl-(4-vinylphenyl) carbonate polymer; (P3) methyl-(4-vinylphenyl) carbonate polymer; (P4) styrene polymer; (P5 and P7)
4-acetamidophenyl-(4-vinylphenyl) carbonate and MAA polymer; (P6) acetanilide imprinted styrene and MAA polymer; (P8 and P10) non-covalently
4-acetamidophenol imprinted MAA polymers (1:3 and 1:4); (P9 and P11) MAA reference polymers.
We postulated that another approach to improving the recog-
nition characteristics of 4-acetamidophenol could be to further
strengthen the interaction between the phenolic hydroxyl and
the polymer during the polymerisation process. We envisaged
that this could be achieved using a so called semi-covalent
molecular imprinting strategy, where the template is linked to
the functional monomer through a reversible covalent bond.
Following polymerisation, the template is cleaved then later
rebound through non-covalent interactions. The use of covalent
template–monomer interactions during the polymerisation step
should result in a more homogeneous recognition site distribution.
A further development of this, termed the sacrificial spacer
approach, was presented by Whitcombe et al. for the imprinting
of another poorly functionalised template, cholesterol.16 This
approach even takes into account the steric requirements of
the non-covalent rebinding interaction. In that study, cholesteryl
(4-vinylphenyl) carbonate was used as a covalently bound template
monomer during the polymerisation process. Hydrolytic cleavage
of the carbonate esters released cholesterol and CO2 to reveal
a binding site spacious enough to allow hydrogen bonding to
take place between the residual phenolic group in the polymer
and the hydroxyl on the template during the rebinding process.
The polymers obtained using this method bound cholesterol with
a single dissociation constant, thus resembling a true biological
receptor in this respect. This imprinting strategy has also been
applied to the imprinting of other hydroxyl containing structures
such as the anaesthetic propofol (2,6-diisopropyl phenol),17,18
nandrolone,19 and bisphenol A.20
Based on this strategy, a series of 4-acetamidophenol im-
printed EGDMA co-polymers were prepared using the semi-
covalent sacrificial spacer approach based on 4-acetamidophenyl-
(4-vinylphenyl) carbonate (P1), the non-covalently interacting
monomer MAA (P8, P10), or a combination of both strategies
(P5, P7). A series of reference systems was prepared from either
styrene (P4) or other carbonate ester based monomers (P2–
P3), from non-covalent imprinting of acetanilide (P6) or in
the total absence of template (P9, P11), Fig. 2. The affinity
for 4-acetamidophenol was evaluated by radioligand equilibrium
binding studies whereas HPLC studies were employed in order
to determine template selectivity and imprinting effect. The
results provide insights into the lower limit of template size and
functionality amenable for use in conjunction with semi-covalent
imprinting protocols, and through the comparison of covalent
and non-covalent polymerisation strategies, furnish information
regarding the design of suitable reference systems.
Experimental
Chemicals
All solvents used were of analytical or HPLC grade. Tetrahydro-
furan (THF) was dried by distillation from Na/benzophenone.
MAA was obtained from Merck and was purified by vacuum
distillation. 2,2¢-azobis(2-isobutyronitrile) (AIBN) was obtained
from Janssen Chimica and was recrystallised from methanol
before use. EGDMA was obtained from Aldrich and was
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3148–3155 | 3149
©

View Article Online
purified by extraction with 0.1 M NaOH (3 ¥ 75 ml to
100 ml EGDMA), washed with 75 ml brine, dried over
MgSO4, filtered and passed through basic Al2O3 before use.
Acetanilide (99%) was obtained from Merck and was recrys-
tallised from water. 4-aminophenol (99%) was from Riedel-
de-Hae¨n. 4-acetoxystyrene (96%), 2-acetamidophenol (97%),
3-acetamidophenol (97%), 4-acetamidophenol (98%), benzanilide
(98%), benzoyl chloride (99%), N,N-diisopropylethylamine
(DIEA, 99.5%), triethylamine (99%), triphosgene (98%),
2,6-di(tert-butyl)-4-methylphenol (≥99%), CDCl3 (99.8%), methyl
chloroformate (99%), phenyl chloroformate (99%) and styrene
(99%) were all obtained from Aldrich. [3H]-4-acetamidophenol
(50 Ci mmol-1) was from American Radiolabeled Chemicals Inc.
(USA).
which was used without further purification for the preparation
of 4.
Synthesis of 4-acetamidophenyl-(4-vinylphenyl) carbonate (4).
The yellow oil containing 3 was dissolved in THF (40 ml) and
added dropwise to a stirred ice cold solution of 4-vinylphenol (2)
(2.0 g, 16.6 mmol) and DIEA (5.0 ml, 28.7 mmol) in THF (60 ml)
containing a trace amount of 2,6-di(tert-butyl)-4-methylphenol
under an inert atmosphere (N2). The solution was stirred at RT for
15 h whereafter the resulting amine salt was removed by filtration
and the filtrate was evaporated. The resultant yellow crystals were
dissolved in CH2Cl2 (100 ml), extracted with water (2 ¥ 50 ml),
NaHCO3 (50 ml, 5% w/v), and brine (50 ml). The organic phase
was dried over MgSO4, filtered, evaporated and purified by column
chromatography (heptane:ethyl acetate, 2:8). The last traces of 1
were removed by recrystallisation from ethanol (95%) to afford 4
as colourless crystals (57%); mp 161–163 ◦C; Found: C, 68.25; H,
5.15; N, 4.6. Calc. for C17H15NO4: C, 68.7; H, 5.1; N, 4.7%; lmax
(CHCl3)/nm 251 (log e 4.48); nmax/cm-1 3264 (NH), 3067 and 2818
Apparatus
NMR spectra were acquired on a Bruker AC 250 MHz instrument
or on a Varian Unity InovaTM 500 MHz instrument. Homonuclear
1H connectivities were determined using COSY experiments.
Chemical shifts (d) are presented in ppm and J values are reported
in Hz. FT-IR spectra were recorded using samples dispersed
in KBr on a Nicolet Avatar 320 FT-IR instrument by diffuse
reflectance IR spectroscopy. Mass spectra of positive ions obtained
by electron impact (EI, 70 eV) were measured using an Agilent
6890 GC-system with an Agilent 5973 MS detector. Liquid
scintillation counting was performed on a Tri-Carb 2100TR
Liquid Scintillator Analyzer from Packard using Beckman Ready
SafeTM scintillation cocktail. Chromatographic evaluations were
performed using a Merck-Hitachi LaChrom HPLC system com-
prised of a series L-7100 pump, L-7200 autosampler, L-7455 diode
array detector and a D-7000 interface. BET surface area analysis
was performed by N2 adsorption on a Micromeritics ASAP 2400
instrument. High resolution mass spectrometry was performed at
the Chemical Center (Lund, Sweden). Elemental analyses were
performed by Mikrokemi AB (Uppsala, Sweden). Data analyses
were conducted with the software package GraphPad Prism 4.03
(GraphPad Software Inc., San Diego, USA).
=
(CH), 1767 and 1658 (C C), 1610 (NH), 1552 and 1500 (Ar); dH
(500 MHz; CDCl3; Me4Si; 25 ◦C) 7.54–7.52 (2 H, d, J 9.0, 2 ¥
ArH), 7.45–7.43 (2 H, d, J 8.6, 2 ¥ ArH), 7.23 (2 H, d, J 8.7, 2 ¥
ArH), 7.22 (2 H, d, J 9.0, 2 ¥ ArH), 6.74–6.68 (1 H, dd, J 10.9
and 17.6, CH2=CH), 5.74–5.71 (1 H, dd, J 0.7 and 17.6, trans-
CH2=CH), 5.28–5.26 (1 H, dd, J 0.6 and 10.9, cis-CH2=CH) and
2.16 (3 H, s, CH3); dC (63 MHz; CDCl3; Me4Si; 25 ◦C) 168.2, 152.1,
150.4, 147.1, 136.0, 135.9, 135.7, 127.3, 121.4, 120.9, 120.8, 114.5
and 24.5; m/z (FAB) 297.1007, C17H15NO4 requires 297.1001.
Synthesis of phenyl-(4-vinylphenyl) carbonate (5). Phenyl-(4-
vinylphenyl) carbonate (5) was prepared from phenyl chlorofor-
mate and 4-vinylphenol (2) according to the method of Whitcombe
et al.16 The product was purified by column chromatography
(heptane:ethyl acetate, 8:2) and was obtained as colourless crystals
(47%); mp 50–52 ◦C; lmax (CHCl3)/nm 251 (log e 4.24); nmax/cm-1
=
=
3062 and 2994 (CH), 1769 (C O), 1627 (C C) and 1591 (Ar);
dH (250 MHz; CDCl3; Me4Si; 25 ◦C) 7.47–7.39 (4 H, m, 4 ¥
ArH), 7.30–7.22 (5 H, m, 5 ¥ ArH), 6.77–6.66 (1 H, dd, J 10.9
and 17.6, CH2=CH), 5.76–5.69 (1 H, dd, J 0.7 and 17.6, trans-
CH2=CH), 5.29–5.25 (1 H, dd, J 0.6 and 10.9, cis-CH2=CH);
dC (63 MHz; CDCl3; Me4Si; 25 ◦C) 152.0, 151.0, 150.4, 135.8,
135.7, 129.6, 127.3, 126.3, 120.94, 120.88 and 114.4; m/z (FAB)
240.0791, C15H12O3 requires 240.0786.
Synthesis of 4-vinylphenol (2). 4-vinylphenol (2) was prepared
by hydrolysis of 4-acetoxystyrene with aqueous potassium hydrox-
ide according to the method of Corson et al..21 Recrystallisation
from n-pentane afforded shiny colourless plates of 2 (46%); lmax
(EtOH)/nm 261 (log e 4.04); nmax/cm-1 3400–3000 br (OH), 2911
Synthesis of methyl-(4-vinylphenyl) carbonate (6). A solution
of methyl chloroformate (2.6 ml, 33.3 mmol) in THF (35 ml) was
added dropwise to a stirred ice cold solution of 4-vinylphenol (2)
(2.0 g, 16.7 mmol) and DIEA (7.3 ml, 41.6 mmol) in THF (55 ml)
containing a trace amount of 2,6-di(tert-butyl)-4-methylphenol
under an inert atmosphere (N2). The solution was stirred at
RT for 20 h until the disappearance of 2 was evident on TLC
(heptane:ethyl acetate, 8:2). The formed amine salt was removed by
filtration and the remaining solution was evaporated. The resultant
oil was dissolved in CH2Cl2 (100 ml), extracted with water (2 ¥
50 ml), NaHCO3 (50 ml, 5% w/v), and brine (50 ml). The organic
phase was dried over MgSO4, filtered, evaporated and purified by
column chromatography (heptane:ethyl acetate, 8:2) to afford a
colourless oil (77%). Found: C, 67.4; H, 5.8. Calc. for C10H10O3:
C, 67.4; H, 5.7%; lmax (EtOH)/nm 249 (log e 4.74); nmax/cm-1 3046
(CH), 1661 (C C) and 1505 (Ar); dH (250 MHz; CDCl3; 25 ◦C),
=
7.32–7.29 (2 H, d, J 8.5, 2 ¥ ArH), 6.81–6.78 (2 H, d, J 8.5, 2 ¥
ArH), 6.71–6.60 (1 H, dd, J 10.9 and 17.6, CH2=CH), 5.64–5.57
(1 H, d, J 17.6, trans-CH2=CH), 5.15–5.10 (1 H, d, J 10.9, cis-
◦
CH2=CH) and 4.71 (1 H, s, OH); dC (63 MHz; CDCl3; 25 C),
155.2, 136.1, 130.7, 127.6, 115.4 and 111.7; m/z (EI) 120 (M+,
100%), 91 (44), 65 (13) and 39 (6).
Synthesis of 4-acetamidophenyl chloroformate (3). A solution
of 4-acetamidophenol (1) (3.0 g, 19.8 mmol) and DIEA (5.8 ml,
33.7 mmol) in THF (75 ml) was added dropwise to a stirred ice
cold solution of triphosgene (7.0 g, 23.8 mmol) in THF (30 ml)
under an inert atmosphere (N2). The resultant solution was stirred
at RT for 5 h, until the disappearance of 1 was evident on TLC
(heptane:ethyl acetate, 2:8). The resultant amine salt was removed
by filtration and the filtrate was evaporated to yield a yellow oil
=
=
and 2953 (CH), 1757 (C O), 1627 and 1508 (C C), 1436 (CH)
3150 | Org. Biomol. Chem., 2009, 7, 3148–3155
This journal is The Royal Society of Chemistry 2009
©

View Article Online
◦
and 1225 (C–O); dH (250 MHz; CDCl3; Me4Si; 25 C) 7.42–7.39
monolith was manually ground and sieved through a 63 mm
sieve. Fine particles were removed by repeated sedimentation from
400 ml acetone (5 ¥ 40 min) and air-dried. The carbonate esters
were hydrolysed by stirring the suspended polymers (1.5 g) in 0.1 M
NaOH in methanol (100 ml) at RT. Template removal from P1 was
monitored by HPLC, which showed that complete hydrolysis was
obtained after 19 h. The polymers were collected by filtration and
washed with 0.1 M NaOH in methanol (2 ¥ 20 ml), methanol (2 ¥
20 ml), water (100 ml), and acetone (100 ml), then air dried.
(2 H, d, J 8.7, 2 ¥ ArH), 7.15–7.11 (2 H, d, J 8.8, 2 ¥ ArH), 6.75–
6.63 (1 H, dd, J 10.9 and 17.6, CH2=CH), 5.73–5.66 (1 H, d, J
17.6, trans-CH2=CH), 5.26–5.22 (1 H, d, J 10.9, cis-CH2=CH),
3.89 (3 H, s, CH3); dC (63 MHz; CDCl3; 25 ◦C) 154.1, 150.6, 135.7,
135.5, 127.2, 121.0, 114.2 and 55.3; m/z (FAB) 178.0634, C10H10O3
requires 178.0630.
Synthesis of (4-hydroxyphenyl)benzamide (7). A mixture of
4-aminophenol (2.7 g, 25.0 mmol) and triethylamine (3.5 ml,
25.1 mmol) in 1,4-dioxane (35 ml) under an inert atmosphere (N2)
was treated with dropwise addition of benzoyl chloride (4.5 ml,
38.8 mmol). The mixture was stirred at RT for 3h, then poured
into water (100 ml). The product was collected by filtration, the
residue dissolved in ethyl acetate and purified by extraction with
10% NaOH (3 ¥ 75 ml). The aqueous phase was treated with 3 M
HCl to pH 1 and extracted with ethyl acetate (3 ¥ 75 ml). The
organic phase was washed with water (75 ml), saturated NaHCO3
(75 ml), and again with water (75 ml) before it was dried over
MgSO4, filtered and evaporated to furnish white crystals of 7
(32%); mp 215–216 ◦C; nmax/cm-1 3382 (NH), 3324 (OH), 3052
Polymer titrations
Polymer particles were suspended in toluene (25 mg/ml), and
from this stock solution appropriate volumes were distributed
to Eppendorf tubes. Each incubation mixture contained [3H]-4-
acetamidophenol (0.27 pmol) and the desired amounts of polymer
(0.5 to 20 mg) in a total volume of 1 ml toluene. The samples
were incubated on a rocking table at 23 ◦C until equilibrium
was reached (2 h). After centrifugation at 20,800 ¥ g for 5 min,
500 ml of the supernatant was removed from each incubation
tube and mixed with 2 ml scintillation cocktail then measured
by liquid scintillation counting. All experiments were performed
in triplicate.
=
(CH), 1647 (C O), 1542 (NH) and 1220 (CO); dH (500 MHz;
DMSO-d6; 25 ◦C) 10.01 (1 H, s, NH), 9.24 (1 H, OH), 7.93–7.92
(2 H, d, 2 ¥ ArH), 7.58–7.49 (5 H, m, 3 ¥ ArH) and 6.75–6.73 (2 H,
d, 2 ¥ ArH); dC (125 MHz; DMSO-d6; 25 ◦C) 164.8, 153.6, 135.1,
131.1, 130.6, 128.2, 127.4, 122.2 and 114.9.
Equilibrium binding studies
Binding studies in different solvents were performed essentially as
described above. Polymers (2.5 or 5 mg) were incubated with [3H]-
4-acetamidophenol (0.27 pmol) in a total volume of 1 ml solvent
in Eppendorf tubes.
NMR studies
1H-NMR spectra were recorded at 25 ◦C using a 29 mM solution
of acetanilide in the presence of various concentrations of styrene
(0 to 1.75 M) in CDCl3.
Competitive experiments
Competitive binding studies were performed essentially as de-
scribed above in toluene containing 0.1% dimethyl formamide.
Polymer samples (2.5 mg) were incubated together with [3H]-
4-acetamidophenol (0.27 pmol) and competing, unlabelled
4-acetamidophenol (10 or 100 mM).
Polymer synthesis
In a typical polymer synthesis, monomer (5 mol%), crosslinker
(95 mol%) and template (for P6, P8 and P10) were dissolved in
chloroform or acetonitrile in a glass test tube (see Table 1 for
details). Initiator (AIBN, 1 mol% with respect to polymerisable
double bonds) was added and the tube was connected to a vacuum
line where the mixture was degassed by three repeated freeze–thaw
cycles, and sealed at reduc◦ed pressure. Polymerisation was carried
out in a water bath at 60 C for 24 h, and the resultant polymer
Chromatographic evaluation
Polymer particles were suspended in chloroform/acetonitrile
(85:15, v/v) in a slurry reservoir and packed into stainless steel
Table 1 Polymer compositions and physical data
P1
P2
P3
P4
P5
P6
P7
P8
P9
P10c
P11c
Acetanilide (mmol)
4-acetamidophenol (mmol)
4 (mmol)
–
–
1.23
–
–
–
–
23.36
10
–
1:19
423
40.5
–
–
–
1.25
–
–
–
23.71
10
–
1:19
411
40.9
–
–
–
–
–
–
–
–
–
–
0.43
–
–
0.43
–
–
–
–
0.43
0.85
24.21
10
–
–
0.43
–
–
–
0.43
–
–
–
–
–
–
–
–
–
–
1.00
–
–
–
–
–
–
–
–
–
5 (mmol)
6 (mmol)
1.27
–
–
24.07
10
–
1:19
436
48.9
–
1.29
–
24.54
10
–
1:19
466
50.6
Styrene (mmol)
MAA (mmol)
EGDMA (mmol)
Chloroform (ml)
Acetonitrile (ml)
Ratio M:Ca
–
–
–
–
0.85
24.21
10
0.85
24.21
–
1.30
24.66
–
1.30
24.66
–
10
1:19
477
99.3
4.00
20.00
–
4.00
20.00
–
10
1:5
364
76.2
–
–
10
10
10
1:2:57
506
63.3
1:2:57
531
64.7
1:2:57
510
93.6
1:19
493
84.5
1:5
350
89.1
BET surface areab (m2/g)
4
6
5
5
7
7
5
5
5
4
4
b
˚
Pore diameter (A)
a Denotes the ratio between monomer (M) and cross-linker (C). b After hydrolysis. c Polymer synthesis according to Tan et al.14
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3148–3155 | 3151
©

View Article Online
HPLC columns (100 ¥ 4.6 mm) at 300 bar using a single action
reciprocating plunger pump (Haskel Engineering Supply Co.,
USA) with 400 ml acetone as the packing solvent.
The template containing monomer 4-acetamidophenyl-(4-
vinylphenyl) carbonate (4) was synthesised in two steps from
4-acetamidophenol (1), triphosgene and 4-vinylphenol (2). For
the synthesis of reference polymers, two similar carbonate ester
analogs were prepared: from 4-vinylphenol and phenyl chloro-
formate or methyl chloroformate to afford phenyl-(4-vinylphenyl)
carbonate (5) and methyl-(4-vinylphenyl) carbonate (6), respec-
tively.
Chromatographic analyses were performed in triplicate using
1 mM of the analytes: acetanilide, benzanilide, (4-hydroxyphenyl)
benzamide, 2-acetamidophenol, 3-acetamidophenol or 4-
acetamidophenol dissolved in chloroform/methanol (94:6, v/v).
Chloroform/methanol (99:1, v/v) was used as the mobile phase
◦
(1 ml/min), and elution was monitored at 246 nm at 24 C. The
The semi-covalent 4-acetamidophenol imprinted polymer P1
was synthesised by co-polymerisation of 4-acetamidophenyl-(4-
vinylphenyl) carbonate (4, 5 mol%) and EGDMA (95 mol%)
in chloroform by thermally-initiated free radical polymerisation
using AIBN (1 mol%). Suitable reference polymers (P2–P3)
were prepared similarly from phenyl-(4-vinylphenyl) carbonate
(5) or methyl-(4-vinylphenyl) carbonate (6) and EGDMA. These
reference polymers were expected to contain smaller binding sites
than P1 since the monomers were based on either phenyl- (P2) or
methyl carbonate (P3). To evaluate the influence of the phenolic
moieties incorporated in the polymer cavities, a styrene based
polymer (P4), lacking the ability to engage in hydrogen bonding
interactions with paracetamol, was also prepared in chloroform.
4-acetamidophenol imprinted polymers were also synthesised
using a combination of one equivalent 4-acetamidophenyl-(4-
vinylphenyl) carbonate and two equivalents of the non-covalently
interacting functional monomer MAA in either chloroform (P5)
or acetonitrile (P7). As a suitable reference for these systems,
acetanilide was imprinted in a styrene MAA co-polymer using
chloroform as a porogen (P6). Non-covalently imprinted 4-
acetamidophenol polymers were also prepared in acetonitrile,
using either three or four equivalents of MAA (P8 and P10),
with corresponding reference polymers prepared in the absence of
template (P9 and P11).
Post-polymerisation treatment of P1 with methanolic sodium
hydroxide was performed to remove the template molecule by hy-
drolysis of the carbonate ester. The release of 4-acetamidophenol
from P1 was monitored by HPLC analysis, which after 19 h no
longer showed any product, indicating that complete hydrolysis
had occurred. All polymers (P2–P11) were subjected to the
same treatment as described above to assure template removal.
Hydrolysis of P1, P2, P5 and P7 was also confirmed by IR
measurements where the aromatic carbonate shoulder at 1779 cm-1
disappeared after treatment. In the case of P3 the carbonate
shoulder was not observed, either before or after hydrolysis.
Polymers were then subjected to a series of radio-ligand binding
studies in a range of different solvents and solvent mixtures using
[3H]-4-acetamidophenol. All polymers demonstrated high binding
affinity for 4-acetamidophenol in toluene (data not shown), which
was expected as recognition of this template relies mostly upon
electrostatic interactions. Accordingly, lower binding was observed
in more hydrophilic media such as acetonitrile and buffer systems,
due to the ability of these solvents to compete for electrostatic
interactions.32 In an attempt to minimise non-specific ligand
binding in toluene, additional equilibrium binding studies were
performed in toluene containing 0.1% dimethyl formamide (Fig. 3,
black bars). In these studies, the highest affinity for the template
was obtained for the non-covalently imprinted polymer P10,
void volume (V0) was determined by injections of acetone, and
the retention volumes (VR) for analytes were assigned to the point
of the peak corresponding to 50% of the peak area by manual
integration using gravimetric analysis.
Results and discussion
Ideally, a method for the rapid and accurate determination of
paracetamol (and its metabolites) in acute situations should be
robust, sensitive and selective. Although MIPs offer the potential
to help fulfil these requirements, significant success with this
target using various molecular imprinting systems has thus far
been elusive. We suggest that the limited size and functionality
of the target are issues that challenge the molecular imprinting
technique. The reported5 significance of both the phenolic and
amide moieties of paracetamol for recognition by antibodies
provides some support for this argument.
The paracetamol molecular imprinting studies thus far
reported13–15 have employed non-covalent strategies, an advantage
of which is the relative ease of polymer preparation and the
vast array of functional monomers available. Nonetheless, this
approach is inherently complicated by the presence of competing
interactions (e.g. template–template, solvent–template, functional
monomer–functional monomer), which contribute to the site
heterogeneity.22–24 Tan et al.’s use of 4-VP to provide stronger
interactions with the mildly acidic phenolic hydroxyl, in con-
junction with MAA for interaction with the amide, provided an
improvement in the recognition performance of the polymer.14
Another general strategy for improving the quality of the
template–functional monomer interactions is the use of multi-
dentate functional monomers capable of stoichiometric non-
covalent interaction, in particular as developed and exploited by
Wulff et al.25,26 and Sellergren et al.,27,28 and more recently by
Takeuchi et al.29 and ourselves.30
Strong template–monomer interactions and more homogenous
recognition site distributions are best exemplified through the
use of reversible covalent bonds for molecular imprinting.31
Semi-covalent imprinting strategies, in particular when using
the sacrificial spacer approach, have proven useful for poorly
functionalised structures, e.g. cholesterol.16 The possibility of
applying a sacrificial spacer-based semi-covalent approach to
the imprinting of paracetamol seemed achievable through the
use of the phenolic hydroxyl as a basis for attachment of
the polymerisable spacer, a carbonate ester-containing template
monomer that upon hydrolysis releases 4-acetamidophenol and
CO2 while exposing the binding site. Moreover, we envisaged that
the sacrificial spacer approach could be further augmented by
additional non-covalent interactions (hydrogen bonding between
methacrylic acid and the amide functionality of paracetamol).
where a binding of 61
1% was determined. Only a minor
difference in binding (4%) between the imprinted (P10) and non-
imprinted (P11) polymers was observed. We attribute this result
3152 | Org. Biomol. Chem., 2009, 7, 3148–3155
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Fig. 3 Binding of [3H]-4-acetamidophenol to P1–P11 in toluene containing 0.1% dimethyl formamide in the presence of various concentrations of cold
4-acetamidophenol. No cold ligand present (black bar), 10 mM cold ligand present (grey bar), 100 mM cold ligand present (white bar). B/T denotes the
amount of bound radioligand (B) relative to the total amount added (T). Error bars reflect SEM.
to the high abundance of carboxylic acid residues present in
these polymers relative to the other polymer systems studied,
and the resultant high incidence of non-specific electrostatic
interactions.
Competitive binding studies were performed with [3H]-4-
acetamidophenol in toluene containing 0.1% DMF using unla-
belled 4-acetamidophenol (10 or 100 mM) as the competing ligand
(Fig. 3, grey and white bars). In general, the polymers demon-
strated only a minor reduction in binding (≤10%) with increasing
concentrations of unlabelled ligand. This result indicates that
the majority of the observed binding under these conditions is
due to non-specific interactions with the EGDMA backbone and
randomly incorporated hydroxyl groups. However, both P6 and
P7 display sites selective for 4-acetamidophenol. The fact that P6
(an acetanilide imprinted reference polymer containing styrene
and MAA) also rebinds 4-acetamidophenol selectively suggests
that p–p interactions between the aromatic residues present in
4-acetamidophenol and the polymer also contribute to the recog-
nition. Attempts to study p–p interactions between paracetamol
and styrene using NMR titrations failed due to poor solubility.
However, the presence of such weak aromatic interactions in
chloroform was confirmed by NMR titration of acetanilide and
styrene (see ESI†). Unfortunately, again due to solubility issues, it
was not possible to further increase the concentration of unlabelled
ligand in the competitive experiments to quantify the extent of
interaction. Instead we turned to chromatographic studies in order
to gain more knowledge regarding the ligand specificities of the
polymers.
Chromatographic evaluation of the recognition characteris-
tics of the polymers was performed in chloroform/methanol
(99:1, v/v) where all polymers demonstrated selectivity for 4-aceta-
midophenol over the structurally similar ligands acetanilide, ben-
zanilide, (4-hydroxyphenyl)benzamide, 2-acetamidophenol and
3-acetamidophenol, Fig. 4. A substantial increase in capacity
factor was observed in all the polymers for all ligands containing
a phenolic residue, and selectivity was even seen for 4-acetamido-
phenol over both 3-acetamidophenol and 2-acetamidophenol,
Table 2. Moving the hydroxyl group from the para to meta
position in the ligand results in a weakening of the interaction
point to the polymer cavity followed by a reduction in retention.
2-acetamidophenol shows a very low capacity factor in all the
polymers which is explained by the inability of this ligand to
Interestingly, polymer P7, which was prepared in acetonitrile
using a combination of semi- and non-covalent imprinting, also
showed high binding (45 1%). When chloroform was employed
as the porogen (P5) the binding and surface areas were lower.
Binding to P6 was similar to P5, despite the lack of the hydroxyl
functionality provided by the residual phenol. The contribution of
the MAA-derived non-covalent interactions to the total binding
is apparent, in particular upon comparison of P7 and P5 with the
purely semi-covalently based systems that showed lower binding
(35% for P1 and P2, 28% for P3), and even after consideration is
taken of the gas-accessible surface areas (BET-data). In the case
of P3, the lower binding may be explained by the formation of a
somewhat smaller cavity in this polymer, leading to exclusion of
the ligand from the binding site. The styrene-containing reference
polymer P4 bound only 14 1% paracetamol which indicates
that incorporation of a phenolic residue in the polymer cavity can
contribute to ligand recognition. This is further supported by the
fact that P4 has a larger BET surface area (466 m2/g) than P1
(423 m2/g) and still displays lower binding of 4-acetamidophenol.
The results obtained for P4 also indicate that non-specific binding
to the EGDMA backbone is substantial. This kind of weak
interaction between template and crosslinker has previously been
shown by molecular dynamics studies to be common in EGDMA
systems, due to the vast number of cross-linker molecules present
in the polymerisation mixture.33 The omni-MIPs developed by
Spivak and co-workers provide a good example of the role cross-
linking functional monomers can play on the induction of ligand
selective binding sites.34,35
The difference in BET surface areas between P1 (423 m2/g)
and P7 (510 m2/g) was substantial and may provide a likely
explanation for the greater binding capacity of P7. However, the
binding observed using P10, with a considerably smaller surface
area (350 m2/g), indicates that specific binding may be involved in
this case.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3148–3155 | 3153
©

View Article Online
Table 2 Capacity factors (k¢)a obtained from HPLC analysis in chloroform/methanol (99:1) with ligand concentration 1 mM
Analyte
P1
P2
P3
P4
P5
P6
P7
P8
P9
P10
0.40
0.10
3.73
1.04
9.11
13.53
P11
0.39
0.11
3.84
1.09
8.88
12.52
Acetanilide
Benzanilide
(4-hydroxyphenyl)benzamide
2-acetamidophenol
3-acetamidophenol
4-acetamidophenol
0.39
0.19
4.59
1.09
6.20
7.74
0.45
0.23
5.47
1.25
7.05
8.87
0.45
0.24
5.41
1.27
7.06
9.01
0.35
0.16
3.65
0.97
4.97
5.70
0.42
0.21
5.24
1.26
7.66
9.70
0.39
0.20
4.06
1.07
5.96
7.07
0.30
0.13
3.86
0.99
6.90
8.58
0.25
0.10
3.32
0.88
6.73
8.47
0.26
0.10
3.26
0.90
6.30
7.76
a k¢ = (VR - V0)/V0. The SEM of the capacity factors presented are all ≤0.02.
Table 3 Normalised selectivity factors (R)a between imprinted polymers and their respective reference systems as a measure of imprinting effect
P1 vs
P3
P5 vs
P6
P7 vs
P6
P8 vs
P9
P10 vs
P11
Analyte
P2
P4
Acetanilide
Benzanilide
(4-hydroxyphenyl)benzamide
2-acetamidophenol
3-acetamidophenol
4-acetamidophenol
1.00
0.97
0.96
1.00
1.01
1.00
1.00
0.93
0.99
1.00
1.02
1.00
0.83
0.85
0.93
0.83
0.92
1.00
0.79
0.75
0.94
0.86
0.94
1.00
0.65
0.53
0.78
0.76
0.95
1.00
0.88
0.86
0.93
0.90
0.98
1.00
0.94
0.90
0.90
0.89
0.95
1.00
a R = (k¢analyte MIP/k¢analyte REF) ¥ (k¢template REF/k¢template MIP).
point is necessary for obtaining recognition, even for a template
as small as paracetamol. Incorporation of a second complexation
point resulted in a more pronounced imprinting effect. This is
illustrated by results obtained with the mixed system P7, where
MAA had the ability to interact with the amide functionality in
4-acetamidophenol. When comparing P5 and P7 it was seen that a
more pronounced imprinting effect is obtained in the acetonitrile
based polymer system than in that synthesised in chloroform.
Interestingly, the choice of acetonitrile over chloroform as a
porogen has previously been reported to increase the selectivity
of MAA-EGDMA imprinted co-polymers.37–39
Collectively, the results demonstrate that polymers offering only
a single interaction point, as in the case for the semi-covalent
systems (P1–P3), while capable of engendering the polymers with
selectivity for 4-acetamidophenol, did not elicit a clear imprinting
effect. We conclude that this is due to the small size and low degree
of functionalisation of the polymerisable paracetamol template
(4). We propose that these factors severely limit the quantity
of favourable contributions from van der Waals interactions
between polymer backbone and template structure, which seem
to be necessary in purely semi-covalent systems involving poorly
functionalised templates in order to obtain template selectivity.17,38
In contrast, the non-covalent systems (P8, P10) showed high
template rebinding, but only a weak imprinting effect, as randomly
distributed carboxyl moieties exerted a dominating influence.
From the results obtained with the polymers investigated in
this study, the best strategy for creating MIPs selective for the
small and poorly functionalized 4-acetamidophenol, was to rely
upon the simultaneous use of semi-covalent and non-covalent
imprinting. Here, the combination of the control of template–
polymer interaction at the phenolic hydroxyl of paracetamol
afforded by the sacrificial spacer approach in conjunction with
the amide–carboxyl interactions furnished by MAA, and even
cross-linker, yielded the most promising result, as based upon the
good imprinting effects obtained for 4-acetamidophenol using P5
Fig. 4 Structures of 4-acetamidophenol and related analogues.
create favourable recognition interactions in the polymer func-
tionalities. This is most likely due to the relative inaccessibility of
the phenolic hydroxyl due to the proximity of the ortho-acetamido
substituent. This problem is exacerbated by the presence of a stabil-
ising and competing internal hydrogen bond within the ligand. The
lower capacity factor obtained for (4-hydroxyphenyl)benzamide
over that for 4-acetamidophenol is noteworthy as they both display
a phenolic residue in the para position. This difference can to some
extent be explained by the fact that (4-hydroxyphenol)benzamide
possesses a two and a half times larger rotational volume36 than 4-
acetamidophenol, leading to exclusion of the former ligand from
some of the binding sites. When comparing the series of polymers
studied, a general trend observed is that the capacity factors
were greatest for the mixed and purely non-covalent systems
and more modest for the semi-covalent polymers. The highest
capacity factors were obtained for the non-covalent polymer P10,
although only a moderate imprinting effect was seen for this
system, Table 3. Interestingly, the effect of imprinting was not
evident at all for the purely semi-covalent system (P1) which
indicates that more than one additional electrostatic interaction
3154 | Org. Biomol. Chem., 2009, 7, 3148–3155
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
and P7. This reflects the results obtained with antibodies5 where
the significance of both functionalities for binding was established.
4 R. E. Coxon, G. Gallacher, J. Landon and C. Rae, Ann. Clin. Biochem.,
1988, 25, 49–52.
5 D. W. Potter, N. R. Pumford, J. A. Hinson, R. W. Benson and D. W.
Roberts, J. Pharmacol. Exp. Ther., 1989, 248, 182–189.
6 X. M. Liu, J. S Zhang and X. G. Chen, Anal. Sci., 2008, 24, 1031–
1037.
Conclusions
7 H. Zou, S. H. Gao, W. S. Chen, Y. Q. Zhong, X. T. Jiang and Y. Y. Pei,
Chromatographia, 2008, 68, 251–257.
8 B. Dejaegher, M. S. Bloomfield, J. Smeyers-Verbeke and Y. V. Heyden,
Talanta, 2008, 75, 258–265.
9 B. Sellergren, ed., Molecularly Imprinted Polymers. Man-made Mimics
of Antibodies and Their Applications in Analytical Chemistry, Elsevier,
Amsterdam, The Netherlands, 2001.
10 M. Yan and O. Ramstro¨m, eds, Molecularly Imprinted Materials.
Science and Technology, Marcel Dekker, New York, 2005.
11 C. Alexander, H. S. Andersson, L. I. Andersson, R. J. Ansell, N. Kirsch,
I. A. Nicholls, J. O’Mahony and M. J. Whitcombe, J. Mol. Recognit.,
2006, 19, 106–180.
12 J. Svenson and I. A. Nicholls, Anal. Chim. Acta, 2001, 435, 19–24.
13 M. L. Yang and Y. Z. Li, Anal. Lett., 2004, 37, 2043–2052.
14 Y. G. Tan, Z. L. Zhou, P. Wang, L. H. Nie and S. Z. Yao, Talanta, 2001,
55, 337–347.
15 A. Gomez-Caballero, M. A. Goicolea and R. J. Barrio, Analyst, 2005,
130, 1012–1018.
16 M. J. Whitcombe, M. E. Rodriguez, P. Villar and E. N. Vulfson, J. Am.
Chem. Soc., 1995, 117, 7105–7111.
17 M. Petcu, J. Cooney, C. Cook, D. Lauren, P. Schaare and P. Holland,
Anal. Chim. Acta, 2001, 435, 49–55.
18 M. Petcu, J. G. Karlsson, M. J. Whitcombe and I. A. Nicholls, J. Mol.
Recognit., 2009, 22, 182–189.
19 C. J. Percival, S. Stanley, A. Braithwaite, M. I. Newton and G. McHale,
Analyst, 2002, 127, 1024–1026.
20 V. P. Joshi, M. G. Kulkarni and R. A. Mashelkar, J. Chromatogr. A,
1999, 849, 319–330.
21 B. B. Corson, W. J. Heintzelman, L. H. Schwartzman, H. E. Tiefenthal,
R. J. Lokken, J. E. Nickels, G. R. Atwood and F. J. Pavlik, J. Org.
Chem., 1958, 23, 544–549.
22 I. A. Nicholls, J. Mol. Recognit., 1998, 11, 79–82.
23 M. J. Whitcombe, L. Martin and E. N. Vulfson, Chromatographia,
1998, 47, 457–464.
24 J. Svenson, J. G. Karlsson and I. A. Nicholls, J. Chromatogr. A, 2004,
1024, 39–44.
25 G. Wulff and K. Knorr, Bioseparation, 2001, 10, 257–276.
26 J. Q. Liu and G. Wulff, J. Am. Chem. Soc., 2008, 130, 8044–8054.
27 A. J. Hall, P. Manesiotis, M. Emgenbroich, M. Quaglia, E. De Lorenzi
and B. Sellergren, J. Org. Chem., 2005, 70, 1732–1736.
28 J. L. Urraca, A. J. Hall, M. C. Moreno-Bondi and B. Sellergren, Angew.
Chem. Int. Ed., 2006, 45, 5158–5161.
Polymers with selectivity for 4-acetamidophenol (paracetamol)
have been prepared using three different imprinting approaches. A
series of batch binding and chromatographic studies demonstrated
that a pure non-covalent imprinting system with MAA bound
paracetamol most efficiently although displayed only a weak
imprinting effect, suggesting that binding was due mainly to non-
specific interactions. Polymers prepared using solely the semi-
covalent approach displayed low binding and no imprinting effect
under the conditions studied, highlighting the necessity for more
than one electrostatic interaction point for creating polymers capa-
ble of the selective molecular recognition of 4-acetamidophenol.
The use of MAA in combination with a semi-covalent strategy
based upon a sacrificial carbonate ester (4), resulted in an improved
imprinting effect. It is therefore evident that in order to obtain
successful imprinting of small and poorly functionalised template
structures, such as 4-acetamidophenol, neither the non-covalent
nor the semi-covalent imprinting method alone is powerful enough
to create selective recognition. At least in this case, the best
solution to this problem requires the use of a combination of
these methods, thereby taking advantage of the template fixation
obtained from the semi-covalent approach and the complexing
power obtained from the non-covalent approach to introduce
additional interaction points to the template. This study not only
helps to define the lower limit of template size and functionality
amenable for use with semi-covalent imprinting protocols, but also
highlights the potential benefits that can be obtained by combining
the strengths of different imprinting strategies. We believe that the
improvements, albeit modest, obtained with this strategy warrant
the use of this approach in other systems with similar challenges.
The basis for this argument is strengthened through comparison
with conclusions drawn from studies with antibodies selective for
paracetamol (4-acetamidophenol) and its metabolites.
29 H. Kubo, H. Navial and T. Takeuchi, Chem. Commun., 2003, 22, 2792–
2793.
30 J. P. Rosengren, J. G. Karlsson and I. A. Nicholls, Org. Biomol. Chem.,
2004, 2, 3374–3378.
31 G. Wulff, Chem. Rev., 2002, 102, 1–27.
32 J. G. Karlsson, L. I. Andersson and I. A. Nicholls, Anal. Chim. Acta,
2001, 435, 57–64.
Acknowledgements
We are grateful to the National Research School in Pharmaceutical
Sciences (Fla¨k), the Swedish Research Council (VR), the Swedish
Knowledge Foundation (KKS) and the University of Kalmar for
financial support. We also thank Dr Susanne Wikman (University
of Kalmar, Sweden) and Dr Michael J. Whitcombe (Cranfield
University, UK) for helpful discussions.
33 J. O’Mahony, B. C. G. Karlsson, B. Mizaikoff and I. A. Nicholls,
Analyst, 2007, 132, 1161–1168.
34 M. Sibrian-Vazquez and D. A. Spivak, J. Am. Chem. Soc., 2004, 126,
7827–7833.
35 J. LeJeune and D. A. Spivak, Anal. Bioanal. Chem., 2007, 389, 433–440.
36 Based on molecular diameter measurement following energy minimi-
sation using MM2 in Chem3D Ultra v. 8.0.
37 B. Sellergren and K. J. Shea, J. Chromatogr., 1993, 635, 31–49.
38 R. J. Ansell and K. L. Kuah, Analyst, 2005, 130, 179–187.
39 R. J. Ansell, D. Y. Wang and J. K. L. Kuah, Analyst, 2008, 133, 1673–
1683.
Notes and references
1 A. Bertolini, A. Ferrari, A. Ottani, S. Guerzoni, R. Tacchi and S. Leone,
CNS Drug Rev., 2006, 12, 250–275.
2 J. A. Vale and A. T. Proudfoot, Lancet, 1995, 346, 547–552.
3 J. Chesham and G. E. Davies, Clin. Exp. Immunol., 1985, 61, 224–231.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3148–3155 | 3155
©