Molecularly imprinted polymers as antibody mimics in automated on-line fluorescent competitive assays

Article abstract of DOI:10.1021/ac070277i

An automated molecularly imprinted sorbent based assay (MIA) for the rapid and sensitive analysis of penicillin-type β-lactam antibiotics (BLAs) has been developed and optimized. The polymers were prepared using penicillin G procaine salt as template (PEN

Full text of DOI:10.1021/ac070277i

Anal. Chem. 2007, 79, 4915-4923
Molecularly Imprinted Polymers as Antibody
Mimics in Automated On-Line Fluorescent
Competitive Assays
†
,†
‡
,§
Javier L. Urraca, Mar ı´ a C. Moreno-Bondi,* Guillermo Orellana, B o1 rje Sellergren,* and
|
Andrew J. Hall
Departments of Analytical and Organic Chemistry, Faculty of Chemistry, Universidad Complutense de Madrid,
E-28040 Madrid, Spain, Institut f u¨ r Umweltforschung, Universit a¨ t Dortmund, Otto Hahn Strasse 6,
D-44221 Dortmund, Germany, and Institute of Pharmacy, Chemistry and Biomedical Sciences, University of Sunderland,
Wharncliffe Street, Sunderland, SR1 3SD, United Kingdom
An automated molecularly imprinted sorbent based assay
MIA) for the rapid and sensitive analysis of penicillin-
to the direct analysis of penicillin G in spiked urine
samples with excellent recoveries (mean value 92%).
Results displayed by comparative analysis of the optimized
MIA with a chromatographic procedure for penicillin G
showed excellent agreement between both methods.
(
type â-lactam antibiotics (BLAs) has been developed and
optimized. The polymers were prepared using penicillin
G procaine salt as template (PENGp) and a stoichiometric
quantity of a urea-based functional monomer to target the
single oxyanionic species in the template molecule. Highly
fluorescent competitors (emission quantum yields of 0.4-
Immunological methods are commonly applied in clinical,
environmental, agricultural, food, and forensic laboratories, as they
provide a very powerful analytical tool for a wide range of analytes,
such as proteins, hormones, toxins, antibiotics, etc. Immunoassays
are very sensitive and selective, do not require skilled workers
or sophisticated instrumentation, can run many analyses simul-
taneously, and are generally cost-effective for large sample loads.
However, reagent stability, the high cost, and difficulties associated
with antibody production, together with the need to use laboratory
animals, are often cited as problems. In addition, the production
of antibodies for toxic compounds or immunosuppressants is
particularly difficult because of their adverse action on the
metabolism and the immune system.
0
.95), molecularly engineered to contain pyrene labels
while keeping intact the 6-aminopenicillanic acid moiety
for efficient recognition by the cross-linked polymers, have
been tested as analyte analogues in the competitive assay.
Pyrenemethylacetamido penicillanic acid (PAAP) was the
tagged antibiotic providing for the highest selectivity when
competing with PenG for the specific binding sites in the
molecularly imprinted polymer (MIP). Upon desorption
from the MIP, the emission signal generated by the PAAP
was related to the antibiotic concentration in the sample.
The 50% binding inhibition concentration of penicillin G
-
6
standard curves was at 1.81 × 10 M PENG, and the
-
7
detection limit was 1.97 × 10 M. The sensor showed a
Molecularly imprinted polymers (MIPs) have unique properties
that make them suitable for use in immuno-type assays. They can
be engineered to exhibit good sensitivity and specificity for various
analytes of medical, environmental, and industrial interest. They
are also highly robust, showing excellent operational stability
dynamic range (normalized signal in the 20 to 80% range)
-7
-6
from 6.80 × 10 to 7.21 × 10 M (20-80% binding
inhibition) PENG in acetonitrile:HEPES buffer 0.1 M at
pH 7.5 (40:60, v/v) solutions. Competitive binding stud-
ies demonstrated various degrees of cross-reactivity with
penicillin-type â-lactam antibiotics such as ampicillin
1
under a wide variety of conditions. However, in comparison to
biological assays using antibodies as selective recognition ele-
ments, molecularly imprinted sorbent-based assays (MIAs) have
several shortcomings, some of which have been addressed in
recent years, such as the limited water compatibility, slow kinetics,
the dominance of radioactive tracers as analyte probes, and the
need to separate the MIP with the bound probe from the unbound
probe in solution after incubation.2
(
(
71%), oxacillin (66%), penicillin V (56%), amoxicillin
13%), and nafcillin (46%) and a lower response to other
isoxazolyl penicillins such as cloxacillin (27%) and di-
cloxacillin (16%). The total analysis time was 14 min per
determination, and the MIP reactor could be reused for
more than 150 cycles without significant loss of recogni-
tion. The automatic MIA has been successfully applied
,3
*
Author to whom correspondence should be addressed. E-mail: mcmbondi@
(1) Alexander, C.; Andersson, H. S.; Andersson, L. I.; Ansell, R. J.; Kirsch, N.;
Nicholls, I. A.; Mahony, J. O.; Whitcombe, M. J. Mol. Recognit. 2006, 19,
106-180.
(2) Ansell, R. J. J. Chromatogr. B 2004, 804, 151-165.
(3) Hunt, C. E.; Pasetto, P.; Ansell, R.J.; Haupt, K. Chem. Commun. 2006, 1754-
1756.
quim.ucm.es; b.sellergren@infu.uni-dortmund.de.
†
Department of Analytical Chemistry.
Department of Organic Chemistry.
Universit a¨ t Dortmund.
University of Sunderland.
‡
§
|
1
0.1021/ac070277i CCC: $37.00 © 2007 American Chemical Society
Analytical Chemistry, Vol. 79, No. 13, July 1, 2007 4915
Published on Web 06/06/2007

Heterogeneous phase immunosensors, in combination with
flow techniques, can be automated easily and combine the
sensitivity and selectivity of immunoassays with the accuracy and
Scheme 1
4
,5
simplicity of the flow methods. These formats have advantages
compared with other sensing formats: (i) it is possible to use
the same reagents employed in microtiter plate formats, avoiding
the synthesis of special reagents, (ii) the desorption of the
captured immunocomplexes is very easy and effective, and (iii)
the active working life of the immuno-support reaches a reusability
of hundred of cycles. They can be applied in different fields,
allowing automation and reducing sample manipulation. The
principles behind these methods could, in principle, be applied
in MIAs, provided that the appropriate labeled analyte derivatives
are available and that the binding and regeneration kinetics are
favorable.
them have been tested only in organic solvents, again to promote
efficient competition between the labeled analogue and the analyte
for the polymeric binding sites, a great limitation for the analysis
of aqueous samples.1
Here, we wish to report the first example of an automated MIA
compatible with aqueous samples. Targeting BLAs, a flow-through
solid-phase competitive assay has been developed which exhibits
excellent robustness and performance when applied to biological
samples. This performance was enabled through the use of a
stoichiometrically imprinted polymer showing good target binding
in aqueous media (see Scheme 1). This polymer was prepared
using penicillin G procaine salt as the template molecule and a
stoichiometric quantity of a recently developed urea-based func-
Most of the labeled analytes described in the literature for
2,13,16,17
6
MIAs are radioactive derivatives, but fluorescent tags are also
7
,8
known. The principal approaches that have found a greater
application in the development of fluorescent based MIAs are
based on the imprinting of the analyte and the use of either a
9
-12
related analyte-labeled derivative
competitive assays.
or an unrelated probe for the
13-16
As we have shown previously,17 in the first
case, the spatial arrangement, nature, and size of the fluorescent
tag must be tailored to the template structure to achieve efficient
competition for the MIP binding sites. We have also demonstrated
that the labeled conjugate showing the best performance in a MIP-
based assay was also the one providing the highest sensitivity in
tional monomer to target the single oxyanionic species in the
template molecule.1
8-20
Ethyleneglycol dimethacrylate served as
5
an immunoassay based on the same type of measurements. The
crosslinker and methacrylamide as an additional hydrogen-
bonding comonomer.
In the assay, the analyte and a constant amount of labeled
fluorescent analogue, [2S,5R,6R]-3,3-dimethyl-7-oxo-6-[(pyren-
second approach relies on the fact that the probe shows some
similar functionality and size to the analyte and binds, although
weakly, to the best imprinted binding sites. With these consid-
erations, it may become a difficult task to find an unrelated probe
with the above-mentioned characteristic. In fact, if the labeled
conjugates are not selected properly, there is no guarantee that
the sites interrogated with the probe are those with the best
1ylacetyl)amino]-4-thia-1-azabicyclo[3.2.0]heptane-2-carboxylic acid
(PAAP), are allowed to compete for the binding sites of the MIP,
which was packed into a reactor. After application of a desorbing
solution, the fluorescence of the labeled derivative eluted from
the sorbent is measured and related to the analyte concentration
in the sample. The application of the desorbing solution allows
the regeneration of the support without affecting its binding
characteristics, thus allowing long-term application. The system
has been fully automated, and several parameters affecting the
sensor performance have been optimized, such as the binding
solvent composition, the amount of polymer, tracer concentration,
assay flow rates for reagent binding and elution, and the amount
and nature of the desorbing solution. The method has been
applied to the analysis of penicillin G in urine samples, and the
results have been validated by HPLC with diode array detection
2
selectivity for the analyte. In most cases, the MIAs developed to
date usually require long incubation times in order to achieve
efficient competition between the analyte and the fluorescent
analogue, thus limiting the applicability of the polymers in
combination with automatic flow techniques. Further, many of
(
(
(
(
4) Gonz a´ lez-Mart ´ı nez, M. A.; Puchades, R.; Maquieira, A. TrAC Trends Anal.
Chem. 1999, 18, 204-218.
5) Benito-Pe n˜ a, E.; Moreno-Bondi, M. C.; Orellana, G.; Maquieira, A.;
Ameronguen, A. V. J. Agric. Food Chem. 2005, 53, 6635-6642.
6) Vlatakis, G.; Andersson, L. I.; Muller, R.; Mosbach, K. Nature 1993, 361,
6
45-647.
7) Piletsky, S. A.; Turner, A. P. F. New materials based on imprinted polymers
and their application in optical sensors. In Optical biosensors: Present and
future; Ligler, F. S., Rowe Taitt, C. A., Eds.; Elsevier: Amsterdam, 2002.
8) Andersson, L. I. J. Chromatogr. B 2000, 739, 163-173.
9) Piletsky, S. A.; Piletskaya, E. V.; El’skaya, A.; Levi, R.; Yano, K.; Karube, I.
Anal. Lett. 1997, 30, 445.
(DAD).
(
(
EXPERIMENTAL SECTION
(
(
(
10) Navarro Villoslada, F.; Urraca, J. L.; Moreno-Bondi, M. C.; Orellana, G. Sens.
Actuators, B: Chem. 2007, 121, 67-73.
11) Surugiu, I.; Danielsson, B.; Ye, L.; Mosbach, K.; Haupt, K. Anal. Chem. 2001,
Chemicals. The urea-based functional monomer, N-[3,5-bis-
(trifluoromethyl)phenyl]-N′-(4-vinylphenyl)urea (MPVU), was pre-
7
3, 487-491.
12) Levi, R.; McNiven, S.; Piletsky, S. A.; Cheong, S. H.; Yano, K.; Karube, I.
Anal. Chem. 1997, 69, 2017-2021.
(17) Benito-Pe n˜ a, E.; Moreno-Bondi, M. C.; Aparicio, S.; Orellana, G.; Cederfur,
J.; Kempe, M. Anal. Chem. 2006, 78, 2019-2027.
(
(
13) Haupt, K.; Mayes, A. G.; Mosbach, K. Anal. Chem. 1998, 70, 3936-3939.
14) Kroeger, S.; Turner, A. P. F.; Mosbach, K.; Haupt, K. Anal. Chem. 1999,
(18) Hall, A. J.; Manesiotis, P.; Emgenbroich, M.; Quaglia, M.; Lorenzi, E.;
Sellergren, B. J. Org. Chem. 2005, 70, 1732-1736.
7
1, 3698-3702.
(19) Urraca, J. L.; Hall, A. J.; Moreno-Bondi, M. C.; Sellergren, B. Angew. Chem.
2006, 45, 5158.
(20) Urraca, J. L.; Moreno-Bondi, M. C.; Hall, A. J.; Sellergren, B. Anal. Chem.
(
15) Piletsky, S. A.; Terpetschnig, E.; Anderson, H. S.; Nichols, I. A.; Wolfbeis,
O. S. Fresenius J. Anal. Chem. 1999, 364, 512-516.
(
16) Greene, N.; Shimizu, K. D. J. Am. Chem. Soc. 2005, 127, 5695-5700.
2007, 79, 695-701.
4916 Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

Figure 1. Chemical structures of the â-lactam antibiotics included in the study.
pared as described previously.18 Ethyleneglycol dimethacrylate
EDMA) was purchased from Aldrich and purified prior to use
as follows: EDMA was washed sequentially with 10% NaOH
aqueous), water, and then brine. After being dried over MgSO
steps: MeOH (100 mL), MeOH/ 0.1 M HCl (aq) (9:1, v/v, 100
(
mL), and finally MeOH (100 mL). The wash solutions were
combined and evaporated to dryness under reduced pressure, and
1
(
4
,
the solid residue was then weighed and examined by H NMR
it was distilled under reduced pressure to give inhibitor-free
monomer. Methacrylamide was purchased from Sigma-Aldrich
spectroscopy, showing that template removal was near quantitative
(extract residue weight ) 300 mg). Thereafter, the MIP was
crushed and sieved, and particles in the size range 25-50 µm
were collected for use in the SPE experiments. Prior to use, they
were sedimented using MeOH/water (80:20, v/ v) in order to
remove fine particles. A control, nonimprinted polymer was
prepared in the same manner but with omission of the template
molecule.
(
St. Louis, MO). ABDV was obtained from Wako (Neuss,
Germany) and used as received. The antibiotics (Figure 1)
penicillin G potassium (PENG) and procaine (PENGp) salts,
penicillin V potassium salt (PENV), amoxicillin (AMOX), nafcillin
sodium salt (NAFCI), cloxacillin sodium salt (CLOX), dicloxacillin
sodium salt (DICLOX), oxacillin (OXA), cephapirin (CEPHA), and
ampicillin (AMPI) were supplied by from Bayer AG (Leverkusen,
Germany).
HPLC grade acetonitrile and methanol were purchased from
SDS (Peypin, France), and water for HPLC was purified with a
Milli-Q system (Millipore, Bedford, MA). All solutions prepared
for HPLC were passed through a 0.45 µm nylon filter before use.
HEPES was supplied by Aldrich (Steinheim, Germany), and a
buffer solution, pH 7.5, was prepared by dissolving 23.830 g in 1
L of purified water (0.1 M). Trifluoroacetic acid (TFA, +99%) was
from Fluka (Buchs, Switzerland).
Synthesis of Pyrene-Labeled â-Lactam Antibiotics (BLAs).
Pyrene-labeled BLAs and their analogues (Figure 2) were pre-
pared from 6-APA or the corresponding â-lactam antibiotic and
the succinimidyl esters (obtained from the commercial acids and
N-hydroxysuccinimide) of pyrenebutyric or pyreneacetic acids
2
1
(Aldrich; St. Louis, MO) in acetone/water/NaHCO3. In brief
outline, after stirring the mixture overnight, the acetone was
removed under reduced pressure and the aqueous mixture was
subjected to repetitive extraction with diethyl ether in order to
remove any unreacted succinimidyl ester. Then the aqueous phase
was brought to pH 2.5 with 10% phosphoric acid and extracted
several times with diethyl ether. The combined extracts were dried
over anhydrous magnesium sulfate. After removal of the drying
agent, the organic solvent was eliminated by rotavaporation to
yield the target compound. The purity and structure of the labeled
Synthesis of the Imprinted and the Nonimprinted Poly-
mers. The polymers were prepared as described elsewhere.1
Briefly, the template PENGp (286 mg, 0.5 mmol), functional
monomer MPVU (186 mg, 0.5 mmol), methacrylamide (84 mg, 1
mmol), EDMA (3.8 mL, 20 mmol), and the free radical initiator
ABDV (44 mg, 1% w/w total monomers) were dissolved in MeCN
9,20
1
substrates was confirmed by elemental analysis and FT-IR, H
(
5.6 mL). After dissolution, the solution was transferred to a glass
NMR, and 13C NMR spectroscopy, as well as mass spectrometry
with electrospray ionization (ESI-MS) and will be reported
elsewhere.22 The following pyrene derivatives were synthesized
in that way: [2S,5R,6R]-3,3-dimethyl-7-oxo-6-[(pyren-1-ylacetyl)-
tube, cooled to 0 °C, and then purged with N for 10 min. After
2
purging, the glass tube was sealed and polymerization initiated
thermally by placing the tube in a water bath set at 40 °C. for 48
h to allow polymerization. The MIP monolith was removed from
the tube and broken into smaller fragments. The template
molecule was removed through the following sequential washing
(
21) Orellana, G.; Aparicio Lara, S.; Moreno-Bondi, M.C.; Benito Pe n˜ a, E.
Synthesis of fluorescent derivatives of â-lactam antibiotics. Spanish Patent
2,197,811.
Analytical Chemistry, Vol. 79, No. 13, July 1, 2007 4917

Figure 2. Chemical structures of the fluorescent derivatives of 6-aminopenicillanic acid.
amino]-4-thia-1-azabicyclo-[3.2.0]heptane-2-carboxylic acid (PAAP);
2S,5R,6R]-3,3-dimethyl-7-oxo-6-[(4-pyren-1-ylbutanoyl]amino]-4-
thia-1-azabicyclo[3.2.0]-heptane-2-carboxylic acid (PBAP); [2S,5R,6R]-
detector wavelengths were set at 341 nm (excitation) and 396 nm
(emission).
To evaluate the optimum concentration of PAAP for the
[
6
-{[(2R)-2-amino-2-(4-hydroxyphenyl)ethanoyl]amino}-3,3-dimethyl-
competitive assays, MIP and NIP samples (19 mg) were incubated
with 1 mL of increasing concentrations (1 nM to 250 nM) of PAAP
in MeCN/aq HEPES buffer (pH 7.5) (40:60, v/v) in the presence
and in the absence of 400 µM PENG. The supernatants were
analyzed by HPLC-FLD as described above.
7
-oxo-4-thia-1-azabicyclo[3.2.0]heptane-2-carboxylic acid (PAAX);
[
2S,5R,6R]-3,3-dimethyl-7-oxo-6-({(2R)-2-phenyl-2-[(pyren-1-yl-
acetyl)amino]-ethanoyl}amino)-4-thia-1-azabicyclo[3.2.0]heptane-
-carboxylic acid (PAAM); [2S,5R,6R]-3,3-dimethyl-7-oxo-6-
{(2R)-2-phenyl-2-[(pyren-1-ylbutanoyl)amino]ethanoyl}amino)-4-
2
(
Measuring System. The automated flow injection manifold
5
thia-1-azabicyclo[3.2.0]-heptane-2-carboxylic acid (PBAM).
HPLC Evaluation of the Polymers. The MIPs were slurry-
packed into stainless steel HPLC columns (150 mm × 4.6 mm
i.d.) using methanol/water (80:20, v/v) as the solvent. HPLC
evaluation of the polymers for recognition of the labeled BLAs
was performed using an HPLC 1100 instrument (Agilent) equipped
with a quaternary pump, an autoinjector, and a fluorescent
detector. For these experiments, the following conditions were
used: 1 mL/min flow rate, 20 µL sample volume, 400 µM analyte
concentration, excitation at 341 nm, and detection at 396 nm.
Methanol was used as the void marker and the retention factors
is similar to one described previously. An eight-way distribution
valve (Kloehn, Las Vegas, NV) equipped with a 2.5 mL syringe
pump is connected to another eight-way distribution valve. The
whole system is controlled using the Winpump software provided
by Kloehn (Las Vegas, NV). The output of the pump is connected
to a stainless steel column (Agilent, Germany) (20 × 2.1 mm),
thermostated at 20 °C and packed with the MIP. The output flow
is driven to a flow-through cell (100 µL, Starna, Germany) placed
in the fluorometer sample holder. Fluorescence intensity measure-
ments at 341 nm (exc.) and 396 nm (em.) were carried out in a
Fluoromax 2 (Horiba-Jobin Yvon, Longjumeau Cedex, France),
with the instrumental parameters and data processing controlled
with the original software (Datamax). All solutions were thermo-
stated at 20 °C using a Precisterm JP water bath (Selecta,
Barcelona, Spain).
(
k) were calculated as k ) (t
time of the analyte and t is the retention time of the void marker.
Imprinting factors (IF) were calculated as IF ) kMIP/kNIP
R 0 0 R
- t )/t , where t is the retention
0
.
Binding of PAAP to the MIP/NIP. To calculate the optimum
amount of polymer to be used in the competitive assay, a constant
amount of PAAP (250 nM) was incubated with increasing
concentrations of MIP or NIP (1-25 mg mL ) and allowed to
equilibrate for 24 h at room temperature. The experiment was
repeated under the same conditions but in the presence of PENG
Fluorescent Competitive Assay. The measuring protocol is
based on the principles of a competitive fluoroimmunoassay. All
solutions were prepared in MeCN:aq. HEPES buffer (0.1 M, pH
-
1
7
.5) (40:60, v/v). Initially, the sample (0.95 mL) was mixed in the
syringe with a constant amount of PAAP (120 nM, 0.25 mL), and
(400 µM). After equilibration, the amount of PAAP remaining in
1
0
mL of the solution was injected into the reactor at a flow rate of
.75 mL min to allow retention of the analytes on the MIP. The
the supernatant was measured by HPLC-FLD analysis. The
analytical column was a LUNA C18 (2) (150 mm × 4.6 mm, 5
µm) column protected by a RP18 guard column (4.0 mm × 3.0
mm, 5 µm), both from Phenomenex (Torrance, CA). The mobile
phase was 45:55 water/MeCN containing 0.08% phosphoric acid.
-1
reactor was washed three times with 1 mL of HEPES buffer (0.1
M, pH 7.5) to remove all the unbound complexes. The analytical
signal was generated upon dissociation of the PAAP molecules
retained on the polymer after competition using 2.5 mL of a
-
1
Analyses were performed at a flow rate of 1.0 mL min at room
temperature. The injection volume was 20 µL, and the fluorescence
-1
methanol solution at a flow rate of 1 mL min
.
Before a new measurement, the reactor was washed with 2.5
-
1
mL of HEPES buffer (0.1 M, pH 7.5) at a flow rate of 1 mL min
.
(
22) Orellana, G.; Aparicio Lara, S.; Moreno-Bondi, M. C.; Benito Pe n˜ a, E.
Manuscript in preparation).
(
A complete cycle for the whole automated assay procedure
4918 Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

required approximately 14 min, including regeneration. The
reactor showed great stability, and the system could be used for
more than 150 measurements. Occasionally, the reactor had to
be back-flushed with washing buffer to maintain a constant flow
rate.
gradient program was used for the mobile phase, combining
solvent A (water with 0.01% TFA) and solvent B (MeCN with 0.01%
TFA) as follows: 0% B (3 min), 0-37% B (5 min), 37% B (11 min),
37-67% B (5 min), 67% solvent B (5 min). Analyses were
performed at a flow rate of 1.5 mL min , and the column
temperature was kept at 35 °C. The injection volume was 200 µL,
and the UV detector wavelength was set at 220 nm. For the
recovery studies, matrix-matched calibration standards were
prepared for each sample. All the analyses were carried out in
triplicate.
-
1
Experimental signals were normalized using the following
expression:
normalized response ) (B - B )/(B - B )
(1)
∞
0
∞
where B is the signal (fluorescence intensity) measured in the
presence of the increasing analyte concentrations, B is the
background fluorescence obtained in the presence of an excess
of PENG, and B is the signal in absence of antibiotic. The
normalized response was plotted as a function of the analyte
concentration (in logarithmic scale), and the experimental data
were fitted to a four-parameter logistic equation (sigmoidal):
RESULTS AND DISCUSSION
∞
1
. Chromatographic Evaluation. The ability of the molecu-
larly imprinted polymers to retain the different â-lactam antibiotics
BLAs) as well as the template molecule (PENG) was investigated
0
(
by comparing their retentions on the NIP and MIP. Figure 3
shows the retention factors (k) and the imprinting factors (IF)
for the five labeled fluorescent BLAs and for the template molecule
on the MIP and NIP using mobile phases containing acetonitrile:
aqueous HEPES (0.1 M, pH 7.5) in the 50:50 to 0:100 (v/v) range.
As has been described previously, pH 7.5 is required to ensure
deprotonation of the carboxylic acid groups of the antibiotic and
the fluorescent tracers, which is necessary for binding to occur
to the polymeric urea moieties.20 The evaluated fluorescent BLAs
show strong retention on both MIP and NIP when a mobile phase
Amax - Amin
normalized signal )
+ Amin
(2)
[
(
analyte] b
1
+
)
EC50
where Amax is the asymptotic maximum (maximum emission in
absence of analyte), b represents the slope of the curve at the
inflection point, EC50 is the analyte concentration at the inflection
point (concentration giving 50% inhibition of Amax), and Amin is
the asymptotic minimum. The detection limit (LOD) was calcu-
lated as the analyte concentration for which the tracer binding to
the antibody was inhibited by 10%, and the dynamic range (DR)
of the method was evaluated as the analyte concentrations that
produced a normalized signal in the 20-80% range.
of 100% aqueous HEPES (0.1 M, pH 7.5) is used (kMIP and kNIP
>
10). This behavior can be attributed to the nonspecific hydropho-
bic interactions between the pyrene moiety of the labeled
antibiotics and the bulk polymer matrices, which must be avoided
in order to develop the competitive assay. The addition of up to
5
0% acetonitrile in the mobile phase reduces drastically the
retention times of all the labeled analytes, as well as PENG, on
both the MIP and NIP. However, the retention is always higher
on the MIP. This behavior can be explained by considering that
the increment of the acetonitrile content in the mobile phase leads
to a decrease in hydrophobic interactions between the aromatic
moiety of the dyes and the sorbents, favoring the specific
recognition of the â-lactam part of the molecules by the PENG-
imprinted sites of the MIP. Obviously, the interaction depends
also on the size, shape, and geometry of the labeled BLA tested,
and the best imprinting factor was obtained for the PAAP, which
shows a three-dimensional structure quite similar to that of
PENG.17 The retention factor for PAAP using acetonitrile:aqueous
HEPES 40:60 (0.1 M, pH 7.5) as mobile phase was very similar
to that of PENG (kPAAP ) 1.91, kPENG ) 1.25). Moreover, in both
cases the retention factors on the non-imprinted polymer (NIP)
are very low (kPAAP ) 0.42, kPENG ) 0.25) and the corresponding
imprinting factors were similar (IFPAAP ) 4.57 and IFPENG ) 4.93).
Thus, we decided to use this solvent composition for the
development of the MIA to ensure the efficient competition of
the labeled BLA and PENG for the polymeric binding sites and
to minimize the nonspecific interactions with the selective recogni-
tion polymer.
Selectivity Studies. Cross-reactivity studies were carried out
by measuring the competitive curves for other chemically related
and nonrelated antibiotics under the optimized conditions in the
-
8
-3
range of 2 × 10 to 9 × 10 M. Cross-reactivity (CR) was
calculated as the percentage between the EC50 value for PENG
and the EC50 for the interfering compound,
EC50 (PENG)
%
CR )
× 100% (3)
EC50 (cross-reacting compound)
Sample Analysis and Validation. Urine samples were col-
lected from a healthy individual not medicated with antibiotics
for more than 6 months. Portions of 10 mL were spiked with an
-
6
aqueous solution of PENG (final concentrations 4 × 10 and 8
-
6
×
10 M), and the pH was adjusted to 7.5 and made up to a final
volume of 25 mL with HEPES (pH 7.5, final buffer concentration
0
4
.1 M). For sample analysis, 6 mL of the solution was mixed with
mL of acetonitrile. The solution was filtered through a 0.45 µm
nylon filter before use and injected in the measuring system. At
least three replicate samples were analyzed to validate the results
obtained.
For validation purposes, the samples were also analyzed by
HPLC-DAD, 1100 instrument (Agilent) equipped with a quaternary
pump, an auto-injector. The analytical column was a LUNA C18
2. MIA Optimization and Analytical Characterization. To
evaluate the optimum amount of MIP for the PenG assay, a fixed
concentration of PAAP (250 nM) was incubated in batch for 24 h
with increasing amounts of MIP in the absence and in the
presence of a constant concentration of PenG (400 µM) in 1 mL
of acetonitrile:aqueous HEPES (40:60, v/v) as the solvent. A
(
3
150 × 4.6 mm, 5 µm) protected by a RP18 guard column (4.0 ×
.0 mm, 5 µm), both from Phenomenex (Torrance, CA). A
Analytical Chemistry, Vol. 79, No. 13, July 1, 2007 4919

Figure 3. Retention factors (k) in (a) the MIP, (b) NIP, and (c) imprinting factors (IF) for the five fluorescent â-lactam antibiotics and the
template molecule using mobile phases consisting of acetonitrile:aqueous HEPES (0.1 M, pH 7.5) in the 50:50 to 0:100 (v/v) range.
Figure 4. (a) Relative fraction of PAAP (%) bound to the MIP in the absence and in the presence of a constant amount of PenG (400 µM) as
-
1
a function of the polymer concentration (1-25 mg mL ). PAAP (250 nM) in acetonitrile:aq. HEPES buffer (pH 7.5, 40:60 (v/v)) (n ) 3; 24 h
incubation). (b) Fraction of PAAP bound (%) to a constant amount of polymer (19 mg in 1 mL of acetonitrile:aq. HEPES buffer (pH 7.5, 40:60
(v/v)), as a function of the fluorescent probe concentration (1 nM to 250 nM), after 24 h incubation: MIP (b) and NIP (9) in the absence of
PENG and MIP in the presence of 400 µM PENG (4).
control experiment was carried out using the NIP in place of the
MIP. As shown in Figure 4a, PAAP competes with PENG for the
polymeric binding sites, as the labeled BLA is displaced from the
MIP in the presence of the template molecule. The curve obtained
in the presence of PENG can be superimposed on that obtained
for the NIP in the absence of the template, demonstrating that
the labeled BLA is displaced by PENG from the specific binding
sites. The largest difference in the fraction of PAAP bound to the
4920 Analytical Chemistry, Vol. 79, No. 13, July 1, 2007

Figure 5. (a) Dose-response curve measured with the MIA over long measuring times. (b) Calibration curves (n ) 3) for PENG (from 0.0013
to 890 µM) in acetonitrile:aqueous HEPES (0.1 M, pH 7) (40/60, v/v) for the MIP (1), the NIP (0), and in urine samples for the MIP(4). PAAP
tracer 25 nM, column packed with 19 mg of polymer in all cases. The experimental points have been fitted to eq 2.
polymer, in the presence and in the absence of the analyte,
corresponds to a MIP amount of 19 mg, which was thus selected
for further experiments.
The imprinted polymer (19 mg) was packed in a stainless steel
reactor connected to the automatic flow-through system for the
subsequent measurements (see scheme in Supporting Informa-
tion). As has been reported for immunoassays, the contact time
between the labeled compound and the MIP may have a direct
effect on the sensitivity of competitive MIAs. A PAAP solution
(20 nM, 1 mL) was injected into the system at flow rates ranging
The amount of tracer (Ag*) used in the assay should be kept
low enough to achieve good sensitivity but high enough to provide
an acceptable signal. Figure 4b shows the change in the fraction
of PAAP bound (%) to a constant amount of MIP/ NIP (19 mg)
as function of the fluorescent probe concentration after 24 h
incubation. Saturation is achieved sooner for the MIP than the
NIP, and there are significant differences between the amounts
of PAAP bound to the polymers, especially for concentrations
lower than 50 nM, where the nonspecific retention is minimized.
As described in the previous experiment, when the assay was
performed in the presence of a constant amount of 1 mL of PENG
-
1
from 0.125 to 1.5 mL min . The amount of fluorescent antibiotic
bound to the sorbent and the sensitivity of the assay increased
as the sample flow rate decreased, but the measuring times were
-
1
also longer. Thus, a flow rate of 0.25 mL min was selected as a
compromise between optimal response and minimum analysis
time.
Methanol was chosen as eluting solvent, as it allowed both
quantitative elution of the labeled BLA and reuse of the reactor
for more than 150 cycles. The flow rate of the eluting solvent also
has an important effect on the sensitivity of the assay. A strong
increase in the analytical signal was observed with decreasing
(400 µM), PAAP was displaced from the MIP binding sites and
the fraction bound to the polymer was equivalent to that obtained
for the NIP. The largest differences between the MIP and the
NIP were obtained when using 25 nM PAAP.
Analytical Chemistry, Vol. 79, No. 13, July 1, 2007 4921

-
8
Figure 6. Cross-reactivities related to PenG when the MIP assay is conducted for an antibiotic concentration ranging from 2 × 10 a 9 ×
-
3
10
M in acetonitrile:aqueous HEPES (0.1 M, pH 7) (60/40, v/v). Fluorescent competitor: 25 nM PAAP (n ) 3).
methanol flow rates (from 1.5 to 0.125 mL min-1), as the
interaction with the polymer is favored and allows the disruption
of binding between the PAAP, PENG, and the polymer. Finally, a
flow rate of 0.75 mL min was selected for the elution step.
Figure 5a shows a typical dose-response curve measured with
the immunosensor over long measuring times. The normalized
competition curves obtained for PENG standards, in acetonitrile:
aqueous HEPES (0.1 M, pH 7) at concentrations ranging from
Table 1. Recovery Results for the MIA and HPLC Urine
Samples Analysis, Spiked with PENG at Several
Concentration Levels (n ) 3)
-
1
spiked,
MIA-FLD,
HPLC-DAD,
10-6 M
10-6 Ma
10-6 Ma
analyte
PENG
PENG
8
4
7 ((2)
3.7 ((0.8)
8 ((1)
3.9 ((0.6)
-
8
-3
a (ts/
x
n (t95%, n ) 3: 4.303).
1
0
to 10 M, are depicted in Figure 5b. The EC50 value
-
6
corresponds to 1.81 × 10 M PENG, and the detection limit was
-
7
1
.97 × 10 M. The sensor showed a dynamic range (normalized
-7 -6
the template (1.35) displays the highest cross-reactivity. Neverthe-
less, shape selectivity also plays a role in the competition with
PAAP. For instance NAFCI that has a log P 3.37 but still shows
a cross-reactivity of 46%. Cephapirin, a cephalosporin-type â-lactam
antibiotic, shows very low cross-reactivity (8%). Hence the size,
geometry, and polarity of the molecules all play an important role
in the recognition properties of the polymer and in the competition
for the binding sites with PAAP.
signal in the 20-80% range) from 6.80 × 10 to 7.21 × 10
M
PENG. The LOD for the analysis of PENG is much lower than
that obtained using the same labels and a methacrylic acid-based
1
7
molecularly imprinted polymer, and only 30 times higher than
that obtained using anti-PENG antibodies as the recognition
element. Nevertheless, the MIA provides shorter analysis times,
5
and the cost per analysis is drastically reduced in comparison to
immunoassays.
4
. Sample Analysis. As a proof of concept, the optimized
The intraday relative standard deviation (n ) 3) was below
fluorescent MIA has been applied to the analysis of PENG in
human urine samples. With this aim, a calibration curve was
prepared by spiking the urine samples with increasing concentra-
11% in the calibration range. Interday reproducibility was evaluated
by measuring the calibration standards on different days. The IC50
-
6
-6
-1
values ranged from 1.52 × 10 to 2.02 × 10 ng mL (RSD
0%, n ) 3). Typical RSD values reported in the literature for
-
8
-3
tions of PENG potassium salt in the range of 10 to 10 M. As
shown in Figure 5b, there is a slight matrix effect in the analyses
of urine, which interestingly and for unknown reasons consists
of an increase in the dynamic range of the assay. Hence, matrix-
matched calibration curves were used for the analysis of these
samples. The analytical characteristics of the assay were: dynamic
1
2
3
immunoassays are between 10-25%, and the MIA results
reported here are within this range.
3. Cross-Reactivity. The specificity of the developed fluores-
cence MIA was evaluated in the presence of other â-lactam
antibiotics. Cross-reactivity was calculated as the percentage
between the EC50 value for PENG and the EC50 for the interfering
compound, and the results are depicted in Figure 6. The polymer
showed group specificity, and the relative cross-reactivity was most
pronounced for other penicillins, such and AMPI (71%) and OXA
-
7
-5
-7
range 7.87 × 10 to 1.71 × 10 M, detection limit 2.97 × 10
-
6
M and EC50 4.00 × 10 M.
Urine samples were fortified with increasing concentrations
-
6
-6
(
4 × 10 and 8 × 10 M) of PenG and analyzed using the
optimized method and HPLC-DAD as an alternative technique for
validation purposes. As shown in Table 1, no significant differences
at a 95% confidence limit were observed between the results
obtained by both methods.
(
(
66%). Medium values were obtained for PENV (56%) and NAFCI
46%), and lower cross-reactivities were obtained for CLOX (27%),
DICLOX (16%), and AMOX (13%). The observed cross-reactivity
of penicillins to compete with PAAP for the polymer binding sites
seems to be related predominantly to the hydrophobicity of the
antibiotic. For instance, DICLOX is more hydrophobic than PENG
CONCLUSIONS
In conclusion an automated MIA compatible with aqueous
2
4
(
log P values 2.89 and 1.21, respectively ) while AMOX is much
samples has been developed. Targeting BLAs, the flow-through
2
4
more hydrophilic (log P ) -0.12 ) so that both penicillins show
little cross-reactivity. However, AMPI with a log P value closer to
(24) Whishart, D. S.; Knox, C.; Guo, A. C.; Shrivastava, S.; Hassanali, M.;
Stodthard, P.; Chang, Z.; Woolsey, J. Nucleic Acid Res. 2006, 34, D668-
D672.
(
23) Parker, G. A. J. Assoc. Off. Anal. Chem. 1991, 74, 868-871.
922 Analytical Chemistry, Vol. 79, No. 13, July 1, 2007
4

solid-phase competitive assay exhibited excellent robustness and
performance when applied to biological samples. This unique
performance was enabled through the use of a recently reported
stoichiometrically imprinted polymer showing good target binding
in aqueous media and should pave the way for the use of MIPs
in routine analytical applications.
J.L.U. thanks the Spanish Ministry of Education and Science for
a predoctoral grant.
SUPPORTING INFORMATION AVAILABLE
Additional information as noted in text. This material is
available free of charge via the Internet at http://pubs.acs.org.
ACKNOWLEDGMENT
Received for review February 9, 2007. Accepted April 4,
This work has been funded by the Madrid Community
2007.
Government (ref S-0505/ AMB/0374), the ESF, the ERDF, and
the Ministry of Science and Education (ref CTQ2006-15610-C02).
AC070277I
Analytical Chemistry, Vol. 79, No. 13, July 1, 2007 4923