Thermometric Sensing of Nitrofurantoin by Noncovalently Imprinted

ARTICLE pubs.acs.org/ac

Thermometric Sensing of Nitrofurantoin by Noncovalently Imprinted Polymers Containing Two Complementary Functional Monomers Umporn Athikomrattanakul,* Nenad Gajovic-Eichelmann, and Frieder W. Scheller Fraunhofer Institute for Biomedical Engineering, Am Muehlenberg 13, Potsdam 14476, Germany

bS Supporting Information ABSTRACT: Molecularly imprinted polymers (MIPs) for nitrofurantoin (NFT) recognition addressing in parallel of two complementary functional groups were created using a noncovalent imprinting approach. Specific tailor-made functional monomers were synthesized: a diaminopyridine derivative as the receptor for the imide residue and three (thio)urea derivatives for the interaction with the nitro group of NFT. A significantly improved binding of NFT to the new MIPs was revealed from the imprinting factor, efficiency of binding, affinity constants and maximum binding number as compared to previously reported MIPs, which addressed either the imide or the nitro residue. Substances possessing only one functionality (either the imide group or nitro group) showed significantly weaker binding to the new imprinted polymers than NFT. However, the compounds lacking both functionalities binds extremely weak to all imprinted polymers. The new imprinted polymers were applied in a flow-through thermistor in organic solvent for the first time. The MIP-thermistor allows the detection of NFT down to a concentration of 5 μM in acetonitrile + 0.2% dimethyl sulfoxide (DMSO). The imprinting factor of 3.91 at 0.1 mM of NFT as obtained by thermistor measurements is well comparable to the value obtained by batch binding experiments.

F

or a number of years, considerable efforts have been made to develop biomimetic receptors with similar affinity as antibodies but with superior physical and chemical resistance and long shelf life. Synthetic polymers, called molecularly imprinted polymers (MIPs), were prepared using covalent binding of the target analyte to the polymer as introduced by Wulff et al.1 Later the noncovalent approach was reported by Mosbach’s group,2 which is most frequentely used in molecular imprinting.3 The MIPs are in a two-step procedure: the prearrangement of the template and the monomers, followed by polymerization. Similar as in many biological systems, the noncovalent binding between the target molecule and the functional monomers can be enhanced by multipoint interactions. Therefore, combinations of two or more functional monomers carrying binding functions which are complementary to different regions (or epitopes) of the target have been applied by several authors. The first application of two different functional monomers in the same MIP was reported by Ramstr€om et al.4 They revealed that the MIP prepared from a methacrylic acid and 2-vinylpyridine rebinds the template, an amino acid derivative, better than the MIP with only one monomer. Takeuchi et al.5 prepared MIPs for the triazine herbicides ametryn and atrazine by combinatorial molecular imprinting using methacrylic acid (MAA) and 2-(trifluoromethyl)acrylic acid (TFMAA) as functional monomers. Rebinding experiments showed that a higher amount of MAA preferred the binding of atrazine while TFMAA favored a stronger binding of ametryn. The preparation of 5-fluorouracil-imprinted polymers using the functional monomer, 2,6-bis(acrylamido)pyridine (BAP) together with the commercial monomer TFMAA was studied by Kugimiya et al.6 r 2011 American Chemical Society

TFMAA was used for the interaction with fluoride whereas BAP bound the imide group via triple hydrogen bonds. The molar ratio of template (5-fluoroureacil, 5-FU)/monomer1 (BAP)/monomer2 (TFMAA)/cross-linker (ethyleneglycol dimethacrylate, EDMA) was 1:1:1:20. The imprinting factor of this MIP was higher than that of the MIP with only BAP and the selectivity for 5-FU was also improved. More examples of MIPs prepared from combinations of commercial functional monomers have been demonstrated by other research groups.7
10 Besides the frequently used batch rebinding and high-pressure liquid chromatography (HPLC) characterization, label-free measurements based on fluorescence,11,12 quartz crystal microbalance (QCM),13
15 surface plasmon resonance (SPR),16
18 and calorimetry19
21 have also been applied for the evaluation of MIPs. Lettau et al.22 introduced a thermistor device for measuring the substrate conversion and binding with a catalytically active imprinted polymer. The polymer was prepared using a covalent imprinting approach. The same group coupled a catalytically active MIP with an enzyme within one column 2 years later.23 The thermometric measurement sums the heat generated by the hydrolysis of phenyl acetate and the subsequent oxidation of phenol. At the same time, the MIP- thermistor for the detection of fructosyl valine was applied by Rajkumar et al.24 The interaction of the phenyl boronic acid containing MIP with the target analyte fructosyl valine generated an exothermic peak signal. On the basis of the Received: April 29, 2011 Accepted: September 17, 2011 Published: September 29, 2011 7704

dx.doi.org/10.1021/ac201099h | Anal. Chem. 2011, 83, 7704–7711

Analytical Chemistry thermometric response, an extraordinary imprinting factor of 41 was found, 29 times higher than the value obtained by batch binding characterization of the same MIP. These reports showed that flowthrough calorimetry and thermistor measurements can reveal significant information about the transient binding behavior (and catalysis) of MIPs. Nitrofurantoin (NFT) is an antibiotic drug which has been banned since 1997 by the European Union as an animal food additive.25,26 Because of its illegal use in many countries and the ever growing demand for improved food safety, there is great interest in affordable methods for NFT detection in foodstuffs at the micromolar range, which could eventually be used for monitoring for abuse on the farm or markets and substitute the standard liquid chroamtography
tandem mass spectrometry (LC
MS/MS) method.27 We have developed different MIPs for NFT using the noncovalent approach. Carboxyphenyl aminohydantoin (CPAH) was successfully applied as a “dummy” template for the chemically unstable NFT. Either 2,6-bis(methaacrylamido) pyridine (BMP) for the imide residue of NFT or 1-(4-vinylphenyl)-3-(3,5-bis(trifluoromethyl) phenyl thiourea (VTU) or 1-(4 vinylphenyl)3-(pentafluorophenyl) thiourea (PTU) for the interaction with the nitro group were applied in our previous studies.28,29 The NFT binding was analyzed in batch binding studies, and it was found that all these MIPs bind NFT better than other structurally related compounds with the only exception being the template used for imprinting, CPAH, which was bound with a slightly higher affinity. However, the affinity of these MIPs is not yet sufficient for the application in a sensor configuration. Therefore, the aim of this work is to present the development of MIPs with an improved affinity for NFT using two tailor-made functional monomers, which are complementary to the imide and the nitro moiety of the target NFT and to apply these MIPs in the thermistor for the label-free measurement of NFT. Moreover, the equilibrium binding properties of the imprinted polymers were characterized using UV
vis spectrophotometry and compared with MIP-based thermistor results. To the best of our knowledge, the analytical application of MIPs with two complementary binding sites prepared by a noncovalent approach in combination with the thermistor has not been reported so far.

’ EXPERIMENTAL SECTION Materials. Unless otherwise specified, all chemicals were purchased from Sigma-Aldrich, Germany. 1-Hydroxycyclohexyl phenyl ketone (Irgacure 127) was purchased from Wako, Japan. Acetonitrile, dimethyl sulfoxide (DMSO), and methanol were obtained from Carl Roth, Germany. All chemicals and solvents were of analytical or HPLC grade. Sep-Pack NH2 cartridges (6 mL, 1 g) were purchased from Waters, United Kingdom. Synthesis of an Analogue Template and Functional Monomers. The analogue template, carboxyphenyl aminohydantoin (CPAH), and the functional monomers, 2,6-bis(methaacrylamido) pyridine (BMP), 1-(4-vinylphenyl)-3-(3,5-bis(trifluoromethyl)phenyl thiourea (VTU), and 1-(4-vinylphenyl)-3-(pentafluorophenyl) thiourea (PTU), Figure 1, were synthesized as already described in our previous report.28,30 1-(4-Vinylphenyl)-3-(3,5-bis(trifluoromethyl) phenyl urea (VFU) was synthesized according to the protocol published earlier by Sellergren’s group.31 Preparation of MIPs Based on Two Functional Monomers. Three different MIPs were prepared, all containing the diaminopyridine derivative (BMP) as the first monomer and either one of the three (thio)urea-based derivatives (VFU, VTU, and PTU)

ARTICLE

as the second monomer. According to their composition, they were called MIP-BMP-VFU, MIP-BMP-VTU, and MIP-BMPPTU. The polymer formulation of these MIPs followed our previous reports28,29 using CPAH as the template, pentaerythritol triacrylate (PETRA) as the cross-linker, DMSO/acetonitrile (67:33) as the porogen but with the molar ratio of template/monomer1/ monomer2/cross-linker as 1:1:1:12. Briefly, 0.5 mmol of CPAH, 0.5 mmol of monomer 1 (BMP), and 0.5 mmol of monomer 2 (VFU, VTU or PTU) were mixed together in 6 mL of DMSO/ acetonitrile (67:33) in 10 mL glass vials and incubated for 4 h at 25 °C to allow for self-assembly of the host
guest complexes. Then 6 mmol of PETRA and 0.5 wt % of photoinitiator (Irgacure 127) were added and each vial was purged with argon for 5 min. After 2 h preincubation at 4 °C, polymerization was initiated under ultraviolet light at 366 nm for 6 h exposure. After the bulky polymers were formed, they were ground in a ball mill (Retsch, type S 100), sieved, and then washed with hot methanol in a Soxhlet apparatus overnight for template removal. The course of template removal was monitored with a UV
vis spectrophotometer at a wavelength of 300 nm. The template stripped polymer was dried for 24 h at 30 °C. The nonimprinted control polymers NIP-BMP-VFU, NIP-BMP-VTU, and NIP-BMP-PTU were prepared in the same manner as MIP but without the template. Morphologies of the polymers were recorded as scanning electron microscopy (SEM) images with a Gemini Leo 1550 instrument (Carl Zeiss, Oberkochen, Germany) at an operating voltage of 3 keV. The pore parameters and surface areas of the imprinted polymers were measured using a Quantachrome instrument (automated surface area and pore size analyzer, Quadrasorb SI). In total, 30 mg of dry imprinted polymers were degassed at 60 °C for 24 h under nitrogen flow to remove adsorbed gases and moisture. The nitrogen adsorption/desorption analysis were performed at 77 K. The surface areas from multipoint N2 adsorption isotherms were calculated using the Brunauer, Emmett, and Teller (BET) equation. Characterization of MIPs by Batch Rebinding Studies. The imprinted polymers for both NFT and CPAH binding were characterized by batch rebinding studies as in our previous study.28,29 Briefly, CPAH or NFT solution was prepared in acetonitrile + 2% DMSO. The DMSO addition was necessary because CPAH and NFT are not soluble in pure acetonitrile. The mixture was incubated at 25 °C under continuous stirring with varied incubation time for adsorption kinetics. The adsorption isotherm was determined from the concentration dependence of NFT and CPAH binding in batch binding studies. The analyte (5 mL) with varied initial concentrations (0.001—2.5 mM) was added to 10 mg of imprinted polymer in each tube. The incubation time for the target binding to all polymers was 24 h at 25 °C under continuous stirring to ensure that the equilibrium was reached. The cross-reactivity of the imprinted polymers was evaluated for selectivity using the structurally related compounds nitrofurazone (NFZ), furazolidone (FZD), 5-nitroanthranilic acid (NAA), 4-nitrobenzyl bromide (NBB), p-nitrophenol (NP), 2-nitrofuran (NF), 5,5-diphenylhydantoin (DPH), and the nonrelated compound p-aminophenol (AP), Figure 1, comparing to the template CPAH and the target NFT. All analytes were prepared at a concentration of 0.1 mM in acetonitrile + 2% DMSO, and binding was evaluated by a 24 h incubation at 25 °C. After the incubation time, the sample solutions were taken and transferred to a centrifuge tube for centrifugation at 10 000 rpm in 10 min. The supernatant of nonbound compounds in each tube was analyzed by UV
vis spectrophotometry at the maximum wavelength 7705

dx.doi.org/10.1021/ac201099h |Anal. Chem. 2011, 83, 7704–7711

Analytical Chemistry

ARTICLE

Figure 1. Chemical structures of compounds used in this study (NFT, nitrofurantoin; CPAH, carboxyphenyl aminohydantoin; BMP, 2,6bis(methaacrylamido) pyridine; VFU, 1-(4-vinylphenyl)-3-(3,5-bis(trifluoromethyl)phenyl urea; VTU, 1-(4-vinylphenyl)-3-(3,5-bis(trifluoromethyl)phenyl thiourea; PTU, 1-(4-vinylphenyl)-3-(pentafluorophenyl) thiourea; NFZ, nitrofurazone; FZD, furazolidone; NAA, 5-nitroanthranilic acid; NBB, 4-nitrobenzyl bromide; NP, p-nitrophenol; NF, 2-nitrofuran) and other compounds (DPH, 5,5-diphenylhydantoin; AP, paminophenol).

of each analyte at 300 nm (CPAH, NF), 370 nm (NFT, NFZ, FZD), 350 nm (NAA), 275 nm (NBB), 405 nm (NP), 240 nm (DPH), and 310 nm (AP), respectively. The concentrations of nonbound compounds were calculated using an individual calibration curve obtained from UV-vis measurement at maximum wavelength for each compound. The amount of the analyte bound (B, μmol/g) to the imprinted polymer was evaluated by subtracting the concentration of the nonbound target molecule from the initial concentration. MIP-Based Thermometric Study. The MIP-based thermistor used in this study is based on a flow-through thermometric calorimeter (“enzyme thermistor”) as invented by Mosbach and Danielsson.32 It was operated in the setup introduced by Rajkumar et al.,24 but in this study noncovalently imprinted polymer was studied and organic solvents like acetonitrile was used as the mobile phase. Measurements were conducted using a commercial enzyme thermistor (model ET-2007, Omic Bioscience, China) with a flow-through calorimetry approach. In total, 100 mg of the noncovalently imprinted polymer or the respective control polymer were packed in a 500 μL microreactor connected with the thermistor probe. The thermistor was perfused with solutions of CPAH or NFT with concentrations from 0 to 100 μM in acetonitrile + 0.2% DMSO at a flow rate of 0.75 mL/min and operated at a constant temperature of 30 °C. Because of the low solubility of CPAH and NFT in pure acetonitrile, concentrations between 250 μM and 2.5 mM were prepared in acetonitrile + 2% DMSO. The thermistor was first perfused with the analyte solution, followed by the mobile phase for the next 15 min. The temperature changes in flow-through mode by the thermistor were measured for each concentration with three different batch preparations. NFT Detection in Feed Samples by MIP-Based Thermistor. A nitrofurantoin stock solution (200 μg/mL) was prepared by dissolving in acetonitrile + 0.2% DMSO. Poultry feed was spiked

with NFT and used as a preliminary real sample. Sample preparation and extraction followed the method from Barbosa et al.33 with some adjustments. Briefly, 5.0 g of minced feed was weighed in a conical tube and spiked with the appropriate volumes of nitrofurantoin stock solutions to give final concentrations of 1.2, 3.6, and 9.5 μg/g, respectively. Then 20 mL of 79 mM ammonium acetate solution (pH 4.6) was added, and the pH was adjusted to pH 8 with a 30% (v/v) ammonium hydroxide solution. After 15 min of incubation, 30 mL of ethyl acetate was added and stirred for 20 min in a rotary shaker. Afterward the mixture was centrifuged for 10 min at 3000 rpm. The organic phase was collected and filtered through a 0.45 μm PVDF miniuniprep vial before the solution was evaporated to dryness. The residue was reconstituted in 2 mL of acetone and put onto a Seppack NH2 cartridge, which was conditioned with 5 mL of acetone. The eluate containing nitrofurantoin 5 mL was collected and evaporated to dryness. The residue was dissolved with 25 mL of acetonitrile + 0.2% DMSO and pumped into the thermistor with the same conditions mentioned above.

’ RESULTS AND DISCUSSION Binding Capacity and Affinity of the Imprinted Polymers. In order to improve the affinity and selectivity of the MIPs for NFT, the (thio)urea-based functional monomers (VFU, VTU, or PTU) and the diaminopyridine derivative (BMP) were combined in the same polymer. In this way, the interactions of the polymer with both the nitro group and imide moiety of NFT are summed up leading to “multi-point interactions”. The equilibrium binding of NFT to all MIPs and NIPs was evaluated at different concentrations in acetonitrile + 2% DMSO (Figure 2). The result revealed that at all concentrations the binding for CPAH (Figure 2a) and for NFT (Figure 2b) were decreasing in the order MIP-BMP7706

dx.doi.org/10.1021/ac201099h |Anal. Chem. 2011, 83, 7704–7711

Analytical Chemistry

Figure 2. Batch binding study: concentration dependence of three different imprinted polymers (MIPs) and the related control polymers (NIPs) for (a) CPAH binding and (b) NFT binding. (Data represents mean ( s.d., n = 3).

VFU > MIP-BMP-VTU > MIP-BMP-PTU. The same tendency was found for the association constants of the prepolymerization complexes of the respective functional monomers with the target in our previous report.30 The equilibrium binding curves and the respective Scatchard analysis were used to calculate the imprinting factor, efficiency of binding, affinity constants (K), and maximum number of binding sites (Bmax) of these polymers (Table 1). In this work, the K and Bmax were obtained using the software LMMpro version 1.06 by C. Schulthess. The imprinting factor was determined from the ratio of the quantity bound by the MIPs to the quantity bound by the NIPs. The efficiency of binding was calculated as percentage of the analyte quantity bound by the MIP at saturation divided by the theoretical maximum binding capacity. The latter was calculated as the amount of the target (CPAH, 500 μmol) used for the polymerization divided by the mass of polymer (∼7.09 g) of MIP-BMP-VFU after template removal. The imprinting factors at saturation concentrations (1 mM) of CPAH and NFT were 4.37 and 3.06, respectively, for the MIP-BMP-VFU. The Scatchard analysis showed that the analyte binding to the imprinted polymers can be described by two kinds of binding sites, one with higher affinity (K1) and lower amount (Bmax1) and

ARTICLE

another with a lower affinity (K2) but with higher capacity (Bmax2). An example of a Scatchard plot is represented in Figure S1 in the Supporting Information. The affinity constants K1 of the MIPs were 175
189 M
1 and 140
159 M
1 for CPAH and NFT, respectively. They were at least 50% higher than for MIPs prepared with only one functional monomer (either BMP or one (thio)urea derivative) from the previous reports.28,29 For instance, with our best one-monomer MIP (MIP-VFU) K1 was 129 and 121 M
1 for CPAH and NFT binding, respectively. The same trend was seen when comparing binding capacities: Bmax1 (i.e., the capacity of high-affinity binding sites obtained by Scatchard analysis) with the two-monomer polymer MIPBMP-VFU was 2.8 and 1.4 μmol/g for CPAH and NFT, respectively, as compared to 1 and 0.92 μmol/g for CPAH and NFT, respectively, with the best one-monomer polymer MIPVFU.29 According to our previous report,28 an increase of the ratio between the template CPAH and functional monomer BMP from 1:1 to 1:4 resulted in an increase of the slope of the linear measuring range by 2% and of the binding to the MIPs at saturation by only 5% while the imprinting factors of both MIPs were almost identical with 2.47 and 2.49, respectively. These findings demonstrate that increasing the amount of functional monomer leads only to a small increase of the analyte binding. However, the simultaneous use of two complementary functional monomers results in a marked improvement of the affinity of the MIPs for nitrofurantoin (via multipoint interactions). This interpretation is also supported by the cross-reactivity experiments, which show the highest binding for substances carrying two “epitopes” complementary to the two functional monomers. The morphology and textures of the polymer particles were analyzed applying SEM as shown in Figure 3. It can be seen that the highly granular particles are irregular in size, ranging between 25 and 50 μm, which is typical for bulk polymers. At the same resolution, the SEM images clearly show that the pore morphology was substantially different between the MIP- and NIP-BMPVFU. The MIP-BMP-VFU had a rough structure and a higher porosity with bigger cavities than the NIP-BMP-VFU, indicative of the surface increase often observed with molecularly imprinted polymers. From nitrogen sorption data, the surface area and total pore volume of the imprinted polymer were obtained and correlated with SEM images. For example, the surface area and total pore volume of MIP-BMP-VFU were 157.4 m2/g and 0.172 mL/g, while these values of NIP-BMP-VFU were 126.9 m2/g and 0.157 mL/g, respectively. These results are also consistent with the previous study results34,35 that the MIPs tend to have higher surface area than that of corresponding NIPs. Furthermore, relatively larger pores or a large surface area in the imprinted polymers may contribute to the improvement of recognition properties of MIPs and the analyte diffusion to the binding sites is facilitated. Cross-Reactivity of MIPs. The cross-reactivity as depicted in Figure 4 shows that all imprinted polymers bind the target molecules CPAH and NFT stronger than all other compounds. The MIP-BMP-VFU has the optimal performance as compared with both MIP-BMP and MIP-VFU. On the basis of the two complementary binding sites, the interaction with CPAH and NFT is stronger than with structurally related compounds, which possess only one binding moiety. NFZ and FZD are similar to NFT in both size and shape, and they contain a nitro group and either an amide or a ketone group. Therefore, both compounds could bind to (thio)urea derivatives via the nitro group similar to NFT whereas they bind to BMP via 7707

dx.doi.org/10.1021/ac201099h |Anal. Chem. 2011, 83, 7704–7711

Analytical Chemistry

6.00 ( 0.25 2.75 ( 0.25 1.16 ( 0.18 153.67 ( 31 6.93 ( 0.35 2.80 6.77 ( 0.15 2.85 ( 0.11 1.61 ( 0.32 187.62 ( 40 8.20 ( 0.41 3.72 MIP-BMP-PTU

Binding capabilities of the imprinted polymers are characterized from binding isotherms determined by batch rebinding study. (Data represent mean ( s.d., n = 3). b K and Bmax evaluated from linear line (R2 = 0.97
0.99) using Langmuir analysis are the affinity constant and the apparent maximum number, respectively. K1 and Bmax1 are the parameters for the higher affinity binding sites. (Data represent mean ( s.d., n = 5) c K2 and Bmax2 are the parameters for the lower affinity binding sites. (Data represent mean ( s.d., n = 5) a

6.19 ( 0.24

6.53 ( 0.19 3.77 ( 0.32

3.10 ( 0.21 1.21 ( 0.25

1.43 ( 0.33 140.30 ( 18

158.83 ( 23 7.31 ( 0.37

8.11 ( 0.41 3.06

2.90 7.05 ( 0.28

7.97 ( 0.22 4.25 ( 0.15

3.42 ( 0.16

2.80 ( 0.40

1.95 ( 0.38

175.41 ( 25

188.78 ( 37

10.13 ( 0.50 4.37

3.94

MIP-BMP-VFU

MIP-BMP-VTU

8.75 ( 0.44

K2c

Bmax1b (μmol/g) (M
1)

K1b efficiency of imprinting

binding (%)a Bmax1b (μmol/g)

efficiency

of binding (%)a factora polymer

imprinting

K1b (M
1)

K2c (M
1)

Bmax2c (μmol/g)

factora

NFT CPAH

Table 1. Summary of Binding Parameters Obtained by Batch Rebinding and Scatchad Analysis of All Imprinted Polymers with CPAH or NFT

(M
1)

Bmax2c (μmol/g)

ARTICLE

hydrogen bonds but weaker than the imide group of CPAH and NFT. The substances possessing only one functionality, either an imide group or a nitro group, such as DPH, NAA, NBB, NP, and NF bind to MIP-BMP-VFU with pronouncedly lower amounts than NFT. AP, which has no complementary binding group to either one of the functional monomers, shows extremely weak binding to all imprinted polymers (Figure 1). Its signal reflects the nonspecific interaction with the polymer backbone. A similar selectivity pattern for all tested nitrocompounds was obtained with the nitro-group addressing MIP-VFU, although the overall binding capacity was markedly lower than with MIP-BMP-VFU. As expected, DPH, an imidocompound without the nitro-group, was not bound at all by this nitro-group addressing MIP. Conversely, only CPAH, NFT, and DPH, possessing an imido group, were strongly bound by the MIP-BMP, while the imino-containing compounds NFZ and FZD showed much weaker binding and AP and NP were not bound at all. For MIP-BMP-VTU and MIP-BMP-PTU, the cross-reactivity data showed the same tendency but with a lower binding capacity both for CPAH and NFT (data not shown). All in all, the cross-reactivity pattern of the two-monomer MIPs may be understood by a surprisingly simple observation: the binding capacity for each of the tested compounds can be estimated by the superposition of the respective capacities measured for the two one-monomer MIPs. Only CPAH, the actual template monomer, shows a higher than expected binding capacity, possibly due to the perfect fit to the imprinted cavities. However, this observation should be considered as preliminary. These results confirm previous reports that MIPs prepared with two functional monomers show improved binding capacity, affinity, and selectivity for the recognition of the target molecule. Thermometric Sensing Based on Noncovalently Imprinted Polymers. Because the polymer MIP-BMP-VFU showed the optimal binding capacity, affinity and selectivity to both CPAH and NFT in batch rebinding studies, it was selected for the measurements in the thermistor. The interaction with NFT at concentrations between 0 and 100 μM in acetonitrile + 0.2% DMSO generated temperature signals as shown in Figure 5. The phenomenon of temperature change (ΔT) recorded by the thermometric measurement is corresponding to the enthalpy change (ΔH), total number of product molecules (np), and heat capacity of the system (Cp) as described in eq 1 by Danielsson’s group.36 ΔT ¼


ΔHnp Cp

ð1Þ

The binding of the target to the MIP is reflected by an exothermic peak signal after injection of the analyte, followed by an endothermic desorption peak after washing with the solvent (Figure 5a.) The exothermic signal of the adsorption may reflect enthalpy changes by the noncovalent complex formation and changes of the solvation of the analytes at the imprinted polymer. On the other hand, the endothermic peak represents the dissociation of the analyte from the polymer cavities and may also include local movements of the polymer chains close to binding sites.19,37 Both signals reflect the sum of the interactions in the binding cavities and at the backbone of the polymer. The height of the peaks in the thermograms of the MIP increases linearly with the NFT concentration up to a concentration of 50 μM and levels off sharply at 100 μM, indicating that the solubility limit of NFT was reached. Higher NFT 7708

dx.doi.org/10.1021/ac201099h |Anal. Chem. 2011, 83, 7704–7711

Analytical Chemistry

ARTICLE

Figure 3. Scanning electron micrographs of imprinted polymers prepared. Magnification at 300 (a) MIP-BMP-VFU; 50 000 (b) MIP-BMP-VFU; and 50 000 (c) NIP-BMP-VFU. The left-hand bar corresponds to 20 μm for part a and 300 nm for parts b and c.

Figure 4. Cross-reactivity of three different MIPs determined by batch binding (MIP-BMP; only imide-binding monomer; MIP-VFU, only nitro-binding monomer; MIP-BMP-VFU, both monomers). A total of 10 mg of imprinted polymer was incubated with 5 mL of a 0.1 mM solution of each respective compound in acetonitrile + 2% DMSO. (Data represents mean ( s.d., n = 3).

concentration standards, requiring a higher DMSO percentage of 2% (v/v) in the mobile phase as in the batch binding study were causing leakages at different sites in the thermistor and could not be measured. The thermogram of the control polymer NIPBMP-VFU showed that NFT binds to the NIP as well but with significantly lower heat changes (Figure 5b). This result emphasized that a non-negligible binding of NFT to the nonimprinted polymer is expected because it contains the same specific receptor functions as the imprinted one, although its surface area and pore volume are smaller. The concentration dependent thermogramms for NFT measured with the MIP- and NIP-BMP-VFU absorbents are presented in Figure 6. The “MIP thermistor” allows the detection of NFT down to a concentration of 5 μM in acetonitrile + 0.2% DMSO. The imprinting factor (IF) evaluated from these curves is 3.91 (0.1 mM), whereas from previous batch binding studies a value of 4.36 was obtained at this concentration. Not only binding of the analyte to the imprinted polymer prepared by noncovalent imprinting but also the nonspecific adsorption are weakly exothermic processes of comparable magnitude. Therefore, the thermistor measurements of analyte binding to the MIP and respective NIP mirror the results of batch binding. Hence, the IF values of both methods are almost identical. On the other hand, a fructose binding MIP prepared by covalent imprinting

Figure 5. Thermograms for the interaction of the imprinted polymers (a) MIP-BMP-VFU and (b) the control polymer NIP-BMP-VFU with different concentrations of NFT in acetonitrile + 0.2% DMSO.

showed a 29-time higher apparent IF for the MIP-thermistor than the respective equilibrium binding value by batch as reported by Rajkumar et al.24 Their unexpected finding can be explained by a negligible contribution of the nonspecific heat of adsorption to the heat signal when compared to the heat accompanying covalent bond formation that occurs during specific binding. In our studies, equilibrium binding experiments were performed at 25 °C (or 298 K) whereas the thermistor 7709

dx.doi.org/10.1021/ac201099h |Anal. Chem. 2011, 83, 7704–7711

Analytical Chemistry

ARTICLE

other reports24 could measure the target at the millimolar range only. This measuring range (LOQ ∼ 1 μg/g) compares favorably with liquid chromatography with ultraviolet detection (LC
UV) for a single nitrofuran screening and confirmation method as reported by others.33,38,39 Therefore, these results show the potential of a MIP-based thermistor biosensor as an alternative screening method for controlling the illegal use of nitrofurantoin on animal feed in farms in the near future.

Figure 6. Concentration dependence of NFT binding to the imprinted polymer based on two functional monomers MIP-BMP-VFU and the control polymer NIP-BMP-VFU investigated by thermometric sensing in acetonitrile + 0.2% DMSO. Temperature changes at the exothermic peak (t = 445 s) were plotted. (Data represent mean ( s.d., n = 3).

Table 2. Recovery of Nitrofurantoin from Spiked Poultry Feed detection of nitrofurantoin (μM)a spiked concentration in μM

mean ( s.d.,

(estimated in μg/g of feed)

n=3

% recovery

40 μM (∼9.5 μg/g)

39.47 ( 1.33

98.6 ( 3.3

15 μM (∼3.6 μg/g) 5 μM (∼1.2 μg/g)

14.13 ( 0.71 4.89 ( 0.43

94.2 ( 4.7 97.8 ( 18.9

a

Data were calculated from the calibration curve (Figure S2 in the Supporting Information) at ranges of 0
50 μM nitrofurantoin standard solution with R2 = 0.9971.

measurement was operated at 30 °C (or 303 K) based on the instrument setup. However, the low enthalpy based on the weak interactions (noncovalent interactions) should have only a small effect of temperature as compared with the covalent binding reactions. The interaction of CPAH, which was used as the template in the polymerization, with the MIP- and NIP-BMP-VFU was also investigated (data not shown). The thermometric signals for CPAH were slightly higher than for NFT with measurements down to 2 μM CPAH and an IF of 7.91 for 0.1 mM of CPAH were achieved. The analytical characteristics for three different concentrations of nitrofurantoin spiked in feed samples were determined using the MIP-based thermistor for the first time, and the results were shown in Table 2. The thermogram of NFT was demonstated in Figure S3 in the Supporting Information. Recovery of nitrofurantoin from the spiked feed was between 94.2 to 98.6% from three different concentrations (5, 15, and 40 μM) with the limit of detection (LOD) and the limit of quantitation (LOQ) being 2.99 μM (0.71 μg/g) and 4.3 μM (1.02 μg/g), respectively. Besides the batch binding protocol being used for sample preparation techniques, the thermistor protocol for MIPs applied in this work should be considered as preliminary results and can be used with established sample measurement down to the micromolar range. On the other hand, covalently MIPs in the thermistor from

’ CONCLUSIONS MIPs addressing in parallel two different functional groups of the analyte nitrofurantoin (NFT) are presented in this paper. The synergistic combination of two functional monomers results in an enhanced affinity for the analyte by the “multi-point interactions”. This added contribution was clearly reflected by a significantly higher binding affinity, capacity, and selectivity for nitrofurantoin when compared to imprints prepared with either one of the functional monomers. The binding enhancement by “multi-point interactions” was even more visible with the actual template molecule, carboxyphenyl-aminohydantoin (CPAH), a mimic of the chemically unstable NFT. The nonimprinted (control) polymer contains the same amount of the two functional monomers as the respective MIP. However, during the polymerization no spatial arrangements are created for the simultaneous interaction of the target with both functionalities. Therefore the binding of the target should be dominated by the interaction with only one of the (randomly arranged) functional monomers. The increased binding of the targets NFT and CPAH by the imprinted polymer obviously results from the shape of the cavities created by the template molecule during the polymer formation. The dominating effect of hydrogen bonding between the functional monomers and the target is exemplified by the very low binding of compounds missing both main functional groups, the imide and the nitro group. Imprinted polymers synthesized by a noncovalent approach were applied for the first time in a flow-through thermistor in organic solvent. The binding process of the analyte to the MIP is reflected by an exothermic signal, and the desorption generates a symmetric endothermic peak, illustrating the reversibility of the binding process. The imprinting factors for the target molecule as calculated from thermometric signals are well comparable to the values obtained by equilibrium binding experiments. The combination of the flow calorimeter with the analyte recognition by noncovalently imprinted MIPs, the “MIP-thermistor”, allows for label-free detection of NFT in animal feed samples at the parts per million concentration range. ’ ASSOCIATED CONTENT

bS

Supporting Information. Additional information including figures. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Phone: +49 (0)331 58187 315. Fax: +49 (0)331 58187 119. E-mail: [email protected].

’ ACKNOWLEDGMENT The authors thank the German Ministry for Education & Research BMBF for financial support under the Grant BioHySys 0311993. Further acknowledgements go to Dr. Martin Katterle 7710

dx.doi.org/10.1021/ac201099h |Anal. Chem. 2011, 83, 7704–7711

Analytical Chemistry for helpful discussions and Dr. Marte Alejandro Ramírez Ortegon for graphic assistance on table of content.

’ REFERENCES (1) Wulff, G.; Sarhan, A. Angew. Chem., Int. Ed. 1972, 11, 341–343. (2) Arshady, R.; Mosbach, K. Macromol. Chem. Phys. 1981, 182, 687–692. (3) Alexander, C.; Andersson, H. S.; Andersson, L. I.; Ansell, R. J.; Kirsch, N.; Nicholls, I. A.; O’Mahony, J.; Whitcombe, M. J. J. Mol. Recognit. 2006, 19, 106–180. (4) Ramstr€om, O.; Andersson, L. I.; Mosbach, K. J. Org. Chem. 1993, 58, 7562–7564. (5) Takeuchi, T.; Fukuma, D.; Matsui, J. Anal. Chem. 1999, 71, 285–295. (6) Kugimiya, A.; Mukawa, T.; Takeuchi, T. Analyst 2001, 126, 772–774. (7) Alizadeh, T. Anal. Chim. Acta 2010, 669, 94–101. (8) Baggiani, C.; Baravalle, P.; Giovannoli, C.; Tozzi, C. J. Chromatogr., A 2006, 1117, 74–80. (9) Hoai, N. T.; Yoo, D.-K.; Kim, D. J. Hazard. Mater. 2010, 173, 462–467. (10) Zhanga, W.; Qina, L.; Hea, X.-W.; Lia, W.-Y.; Zhanga, Y.-K. J. Chromatogr., A 2009, 1216, 4560–4567. (11) Ge, S.; Yan, M.; Cheng, X.; Zhang, C.; Yu, J.; Zhao, P.; Gao, W. J. Pharm. Biomed. Anal. 2010, 52, 615–619. (12) Piletsky, S. A.; Terpetschnig, E.; Andersson, H. S.; Nicholls, I. A.; Wolfbeis, O. S. Fresenius J. Anal. Chem. 1999, 364, 512–516. (13) Dickert, F. L.; Forth, P.; Lieberzeit, P. A.; Voigt, G. Fresenius J. Anal. Chem. 2000, 366, 802–806. € (14) Say, R.; G€ultekin, A.; Ozcan, A. A.; Denizli, A.; Ers€oz, A. Anal. Chim. Acta 2009, 640, 82–86. (15) Suedee, R.; Intakong, W.; Dickert, F. L. Talanta 2006, 70, 194–201. (16) Lai, E. P. C.; Fafara, A.; VanderNoot, V. A.; Kono, M.; Polsky, B. Can. J. Chem. 1998, 76, 265–273. (17) Matsunaga, T.; Hishiya, T.; Takeuchi, T. Anal. Chim. Acta 2007, 591, 63–67. (18) Roche, P. J. R.; Ng, S. M.; Narayanaswamy, R.; Goddard, N.; Page, K. M. Sens. Actuators, B 2009, 139, 22–29. (19) Kirchner, R.; Seidel, J.; Wolf, G.; Wulff, G. J. Inclusion Phenom. Macrocyclic Chem. 2002, 43, 279–283. (20) Manesiotis, P.; Hall, A. J.; Courtois, J.; Irgum, K.; Sellergren, B. Angew. Chem., Int. Ed. 2005, 44, 3902–3906. (21) Weber, A.; Dettling, M.; Brunner, H.; Tovar, G. E. M. Macromol. Rapid Commun. 2002, 23, 824–828. (22) Lettau, K.; Warsinke, A.; Katterle, M.; Danielsson, B.; Scheller, F. W. Angew. Chem., Int. Ed. 2006, 45, 6986–6990. (23) Lettau, K.; Katterle, M.; Warsinke, A.; Scheller, F. W. Biosens. Bioelectron. 2008, 23, 1216–1219. (24) Rajkumar, R.; Katterle, M.; Warsinke, A.; M€ohwald, H.; Scheller, F. W. Biosens. Bioelectron. 2008, 23, 1195–1199. (25) Hiraku, Y.; Sekine, A.; Nabeshi, H.; Midorikawa, K.; Murata, M.; Kumagai, Y.; Kawanishi, S. Cancer Lett. 2004, 215, 141–150. (26) Hoogenboom, L. A. P.; van Kammen, M.; Berghmans, M. C. J.; Koeman, J. H.; Kuiper, H. A. Food Chem. Toxicol. 1991, 29, 321–328. (27) Pereira, A. S.; Pampana, L. C.; Donato, J. L.; Nucci, G. D. Anal. Chim. Acta 2004, 514, 9–13. (28) Athikomrattanakul, U.; Katterle, M.; Gajovic-Eichelmann, N.; Scheller, F. W. Biosens. Bioelectron. 2009, 25, 82–87. (29) Athikomrattanakul, U.; Katterle, M.; Gajovic-Eichelmann, N.; Scheller, F. W. Talanta 2011, 84, 274–279. (30) Athikomrattanakul, U.; Promptmas, C.; Katterle, M. Tetrahedron Lett. 2009, 50, 359–362. (31) Hall, A. J.; Manesiotis, P.; Emgenbroich, M.; Quagilia, M.; de Lorenzi, E.; Sellergren, B. J. J. Org. Chem. 2005, 70, 1732–1736.

ARTICLE

(32) Mosbach, K.; Danielsson, B. Biochim. Biophys. Acta 1974, 364, 140–145. (33) Barbosa, J.; Moura, S.; Barbosa, R.; Ramos, F.; Noronha da Silveira, M. R. Anal. Chim. Acta 2007, 586, 359–365. (34) Br€uggemann, O. Biomol. Eng. 2001, 18, 1–7. (35) Shi, X.; Wu, A.; Qu, G.; Li, R.; Zhang, D. Biomaterials 2007, 28, 3741–3749. (36) Ramanathan, K.; Danielsson, B. Biosens. Bioelectron. 2001, 16, 417–423. (37) Haupt, K. Ultrathin Electrochemical Chemo- and Biosensors; Mirsky, V. M., Ed.; Springer: Berlin, Germany, 2004; Vol. 2, pp 23
27. (38) Draisci, R.; Giannetti, L.; Lucentini, L.; Palleschi, L.; Brambilla, G.; Serpe, L.; Gallo, P. J. Chromatogr., A 1997, 777, 201–211. (39) McCracken, R. J.; Kennedy, D. G. J. Chromatogr., A 1997, 771, 349–354.

7711

dx.doi.org/10.1021/ac201099h |Anal. Chem. 2011, 83, 7704–7711