A MIP-based Impedimetric Sensor for the Detection of low-mw Molecules

Transcription

1 A MIP-based Impedimetric Sensor for the Detection of low-mw Molecules R. Thoelen *1, R. Vansweevelt 1, J. Duchateau 1, F. Horemans 1, J. D Haen 1,2, L. Lutsen 2, D. Vanderzande 1,2, M. Ameloot 3, M. vandeven 3, T.J. Cleij 1, and P. Wagner 1 1 Hasselt University, Institute for Materials Research, Wetenschapspark 1, 3590 Diepenbeek, Belgium 2 IMEC, Division IMOMEC, Wetenschapspark 1, B-3590 Diepenbeek, Belgium 3 Biomedical Research Institute, Hasselt University and transnationale Universiteit Limburg, Agoralaan, B-3590 Diepenbeek, Belgium ronald.thoelen@uhasselt.be Abstract Mimicking the selectivity and sensitivity of biological systems for sensor devices is of increasing interest in biomedical, environmental and chemical analysis. Synthetic materials with imprinted nanocavities, acting as highly selective artificial receptors, are a tailor-made solution in obtaining such a sensor. Incorporation of such molecularly imprinted polymers (MIPs) in a platform suitable for electrochemical measurements, can offer high sensitivity together with device miniaturization and an electronic read-out. As a proof of principle, a MIP-based sensor for L-nicotine has been developed. To this end, the molecular structure of L-nicotine was imprinted in a polymer matrix of PMAA. Subsequently, microparticles of the imprinted polymer were immobilized on thin films of the conjugated polymer OC 1 C 10 -PPV. These films were incorporated in an impedimetric sensing device. Using electrochemical impedance spectroscopy, the real part of the impedance was monitored for various concentrations. This setup allows for the detection of L-nicotine from 1 to 10 nm and is insensitive for the resembling molecule L-cotinine. Keywords: Molecular imprinting, L-nicotine sensor, impedance spectroscopy, conjugated polymers 1

2 1. Introduction Molecularly imprinted polymers (MIPs) are an important class of synthetic materials mimicking the molecular recognition of low-weight molecules by natural receptors (Cormack and Elorza 2004; Haupt and Mosbach 1998; Piletsky et al. 2001). The focus areas in which MIPs are being implemented are mainly materials science and separation technology (Ansell et al. 1996). However, these materials become also increasingly popular for sensor purposes. The technology provides a wide pallet of sensing devices for different kinds of target templates, such as bacteria (Harvey et al. 2006), low molecular weight molecules (Piletsky et al. 2006; Dickert et al. 2004) and proteins (Bossi et al. 2007). The detection of the binding can be achieved by electrochemical transducers with MIP membranes, piezo-electric or optical transducers (Blanco-Lopez et al. 2004; Ebarvia and Sevilla 2005; Liao et al. 2004; Merkoci and Alegret 2002; Panasyuk-Delaney et al. 2001). The imprinting technique is based on the development of either non-covalent or reversible covalent complexes formed between the target molecule and suitable functional monomers. Subsequently, a cross-linker is added to this mixture to form a matrix in which the complexes are fixed. After extraction of the target molecule from this matrix, a tailor-made, highly selective synthetic receptor is created. In this work, the aim is to develop an approach for a fast, inexpensive, MIPbased analyte-sensor. The fast readout is achieved by using impedance spectroscopy in a basic electrochemical cell configuration and monitoring the time resolved capacitance. To enhance the signal, a conjugated polymer is used to immobilize the non-conducting MIP particles onto the electrode surface. The conjugated polymer, OC 1 C 10 -PPV, which is stable in aqueous and a variety of organic solvent solutions, has also successfully been used as transducer layer in immunosensors (Cooreman et al. 2005). As proof of principle, a MIP-based sensor for L-nicotine, C 10 H 14 N 2, has been developed. L-nicotine has previously been employed as a template for molecular imprinting in chromatography. In 2001 the first successful attempt was made to implement this material in a QCM device (Tan et al. 2001), which achieved a detection limit of 25 nm. The MIP particles in Tan s work were immobilized in a PVC matrix and the analyte was dissolved in double-distilled water. The relevant medical concentration for nicotine in urine for non-smokers is 0.3 µm and for smokers 6.3 µm. 2

3 2. Experimental The PPV derivative, OC 1 C 10 -PPV, which was used as the base polymer for the immobilization layer was synthesized via the sulphinyl route (Louwet et al. 1995). All chemical and physical properties were in agreement with previously reported data. This polymer behaves as a p-type semiconductor with a bandgap of 2.1 ev. The imprinted polymer was synthesized using ethylene glycol dimethacrylate (EGDM) as the cross-linker, the monomer methacrylic acid (MAA) and chloroform as the porogen. Prior to polymerization, the stabilizers in the MAA and EGDM were removed. For the initiator azobisisobutyronitrile (AIBN) was used. All chemicals used in the polymerization were of analytical grade and were obtained from various commercial sources. The target molecule L-nicotine, C 10 H 14 N 2, was obtained from Acros. A similar molecule, L-cotinine, C 10 H 12 N 2 O, which was obtained from Sigma- Aldrich, was used to test the specificity. Cotinine is a byproduct of the human metabolism of nicotine. Both molecules are shown in Fig. 1. Fig. 1. molecule structure of L-nicotine (MW: Da ) and L-cotinine (MW: Da ). For the preparation of the L-nicotine molecularly imprinted polymer (MIP), a mixture of MAA (12.5 mmol), EGDM (25.2 mmol) and AIBN (0.66 mmol) was dissolved in 7 ml chloroform together with the template molecule L-nicotine (6.41 mmol). To exclude oxygen in the mixture, it was degassed for 10 min with N 2. Subsequently, for polymerization the solution was sealed and kept in a thermostatic water bath at 60 C for 72 hours. After polymerization the solid MIP was grounded with a mechanical mortar for 24 hours and sieved through a 25-µm sieve. Only particles with a size smaller than 25 μm were used. Next, the L-nicotine was removed 3

4 from the MIP powder by extensive washing with methanol (48 hours), followed by a mixture of acetic acid/acetonitrile (1/1) (48 hours) and finally again methanol (X hours) using a continuous extraction setup. A non-imprinted polymer (NIP) was synthesized in the same way but without the presence of the target molecule. The binding constant and quality of the molecularly imprinted polymer were evaluated with UV-Vis spectroscopy using the Varian Cary 500 UV-Vis-NIR spectrophotometer. The morphology of the sensor surface was investigated using optical microscopy with a Zeiss Axiovert 40 MAT, and scanning electron microscopy, with a FEI Quanta 200 FEG-SEM. Microscope images were processed using the image analysis program ImageJ 1.37v from the National Institute of Health, USA. In order to detect the L-nicotine recognition electronically, the electrochemical impedance principle was used. The activated electrodes, together with the reference electrodes, were placed in an impedimetric addition setup allowing time-resolved, differential measurements. The impedance spectra were measured with zero bias and an oscillating voltage of 50 mv in a frequency range from 100 Hz to 1 MHz with a HP4194A Gain-Phase analyzer. The time resolved measurements were performed at a fixed frequency of 213 Hz, in order to be sensitive to changes at the interface between the sensor and the electrolyte. 3. Results and discussion 3.1 Characterization of the MIPs For a refined analysis of the binding mechanisms between MIPs/NIPs and nicotine/cotinine, batch rebinding experiments were performed using UV-Vis absorption spectroscopy by measuring the nicotine absorption at 260 nm. To test the specificity of the L-nicotine imprinted polymers, the MIP particles were also exposed to several concentrations of L-cotinine. The same experiments were performed with the NIP-particles. A fixed amount of MIP particles, 20 mg, was added to different concentrations of L-nicotine C i from 0.2 to 1 mm. After stirring these solutions for one hour at room temperature, the concentration of target molecules remaining in solution (C f ) was measured. Subsequently, the amount of target molecules bound to the MIP (S b ) and the amount of bound molecules per gram MIP were calculated. Binding isotherms and corresponding Scatchard plots for the imprinted and non- 4

5 imprinted polymers are shown in Fig. 2. Due to the high affinity of the specific binding sites in the MIP, the amount of L-nicotine bound to the MIP is higher than in the NIP. The amounts of L-cotinine bound to MIP and NIP are equal and significantly lower as compared to the L-nicotine binding characteristics. The binding of L-cotinine is non-specific and may involve the free acid groups at the surface of the polymer. Fig. 2. Graph of binding isotherms (a) and corresponding Scatchard plots (b) for imprinted and non-imprinted polymers exposed to L-nicotine and L-cotinine. The binding sites in imprinted polymers have a large heterogeneous distribution due to different organization of the complementary functional groups and different shape-selective cavities (Umpleby et al. 2000; Spivak 2005). For a single class of binding sites, the Langmuir isotherm would be a straight line (Langmuir 1918). The Scatchard plots, Fig. 2b, indicate however the existence of a non-linear relationship, thus implying a hetergenous distribution of the binding sites, each with different affinity. For this type of distributions the binding isotherm can be better analyzed using the Freundlich model (Freundlich 1926), see Equation 1. This model describes an adsorption system with emphasis on two factors, namely the lateral interaction between the adsorbed molecules and the energetic surface heterogeneity. S = A (1) ν b C f Where S b is the amount of target molecule bound per gram of MIP/NIP, A is the Freundlich constant and ν the Freundlich heterogeneity parameter. The binding 5

6 isotherms in Fig. 2a were fitted using Equation 1. The parameters are presented in Table 1 and can be used to plot the affinity distribution according to Equation 2, where N(K) is the amount of binding sites available for each association constant K. This can be seen in Fig. 3. sin( πν ) ν N( K i ) = A K i (2) π Using the parameters the total number of binding sites, N tot, fore the average association constant, K av, within the range from mm -1 can be calculated using Equation 3 and 4 (Spivak 2005). These values are also listed in Table 1. sin( πν ) ν ν N tot = A ( Kmin Kmax ) (3) π 1 ν 1 ν ν ( K max K min ) K av = (4) ν ν ν ( K K ) 1 min max Table 1. Freundlich parameters from fit in Fig. 2a together with the number-average association constant and the total number of binding sites for the range of binding constants from mm -1. A ν (mm -1 ) N tot (µmole/g) MIP L-nicotine NIP L-nicotine MIP L-cotinine NIP L-cotinine K av 6

7 Fig. 3. The affinity distribution for imprinted and non-imprinted polymers based on the Freundlich model, the graphs were calculated using Eq. 2 with the parameters from Table 1, error bars are smaller than symbol size. The results in Table 1 confirm the higher affinity of the MIP for L-nicotine than for L- cotinine. In addition, it can be seen that the MIP particles have twice as many binding sites for L-nicotine as the NIP, while the number of binding sites for L-cotinine is even lower. For the region between 1 and 100 mm -1 the amount of binding sites for L- cotinine is only 0.46 µmole/g for the MIP and 0.11 µmole/g for the NIP. 3.2 Sensing concept and results The sensing devices consisted of a glass cover slip substrate, 20 mm by 12 mm, on which eight platina electrodes were sputtered, resulting in a device consisting of four sensitive regions. A 200 nm thick film of the conjugated polymer OC 1 C 10 -PPV, was spincoated on top of the electrodes from a chlorbenzene solution. Subsequently, the MIP particles were transferred to the OC 1 C 10 -PPV covered electrode using a poly(dimethylsiloxane) (PDMS) stamp, which was pressed on the electrode with a small force of about 100 Pascal during a few seconds. After removal of the stamp, a thin film of micro particles of MIP remained on the surface. Subsequently, the polymer surface was heated to 110 C, which is above glass transition temperature OC 1 C 10 -PPV, during ten minutes causing the powder particles to sink into the 7

8 conjugated polymer matrix. The sample was packaged into an addition setup with four sensing sites, shown in Fig. 4, and the impedance was measured time resolved at a fixed frequency of 213 Hz. The addition setup allowed addition of the target molecule by adding of few droplets in the buffer solution to reduce artifacts from replacing the buffer during the measurement. The four sensing sites allowed simultaneous monitoring of a bare Pt electrode, a Pt electrode covered with OC 1 C 10 - PPV and a Pt electrode covered with OC 1 C 10 -PPV-MIP layer and OC 1 C 10 -PPV-NIP layer. Before and after each measurement a complete impedance spectrum was measured. Fig. 4. Sensing device with four sensing channels with 500 µm electrode spacing, Pt / Pt-OC 1 C 10 PPV / Pt- OC 1 C 10 PPV-MIP / Pt- OC 1 C 10 PPV-NIP. After immobilizing the MIP particles on top of the electrode surface, the density was investigated using optical microscopy. It could be seen that the macroscopic covering percentage of the particles was about 23.5 %. The average size of the clusters of particles was 49.3 µm². Fig. 5 shows the optical microscope image. Fig. 5. Optical microscope images of MIP particles on the transducer layer (a) and the size distribution graph (b), the solid line is a guide to the eye. 8

9 To study the coverage at the sub-micron scale, scanning electron microscopy (SEM) was used. In between the large clusters, a coverage of 6.1 % of small fragments was found with an average size of µm². With the electron microscope we can distinguish the separate MIP particles on top of the electrodes as seen in Fig. 6a. To investigate the sinking of the MIP particles into the polymer layer, a crosssection was taken of the same device, seen in Fig. 6b. The sunken splinter has a diameter of 300 nm. Fig. 6. SEM image of MIP particles seen from above (a) and the cross-section (b), using an accelerating voltage of 15.0 kev and a large-field detector (LFD). Electrical measurements were performed by using electrochemical impedance spectroscopy. Simultaneous a bare Pt electrode and OC 1 C 10 -PPV coated electrodes without and with MIP and NIP particles were measured in phosphate buffered saline (PBS) solution before addition of nicotine. PBS is used to mimic the ionic strength of body liquids and to maintain a constant ph. The latter may be important since L- nicotine is known to have different protonations depending on the acidity of the solution. (Troje et al. 1997). To investigate the recognition event, the real part of the impedance is tracked in time at the lower frequency regime (213 Hz). At this regime the signal represents events that take place at the interface between the electrolyte and the sensor surface. Binding of nicotine effects the formation of the double layer at the electrode surface, which is reflected in the capacitive changes at the lower frequency range. In addition, it influences the charge transfer from the electrolyte to the electrode. 9

10 Fig. 7a shows the dose-response curves of the MIP- and NIP-based sensor electrode after addition of L-nicotine to the PBS solution. The black curve, which corresponds to the signal of the MIP coated electrode, shows an increase of 45% when 10 nm of L-nicotine is introduced in the setup, whereas the signal of the NIP coated electrode only increased with 30%. The change of the bare OC 1 C 10 -PPV is in same order of magnitude as the NIP sample. Also the bare OC 1 C 10 -PPV and NIP electrodes react on addition of L-nicotine, which is probably due to adsorption at the interface. In the case of the NIP, this adsorption may involve the free acid groups, whereas in the case of OC 1 C 10 -PPV this may be the result of π-π interactions. The impedance change is measured 20 minutes after each addition of L-nicotine. Subsequently, the influence of L-cotinine is investigated in the same way as was done for L-nicotine. L-cotinine differs only by one oxygen atom from L-nicotine (cf. Fig. 1). Fig. 7b shows the response of the four different channels. A signal increase is observed due to adsorption of L-cotinine to the polymer surface. The response is not specific while both responses are in the same order and markedly lower than the reaction of L-nicotine of the sensor. It can be expected that this adsorption has a similar origin as discussed above for the non-specific adsorption of L-nicotine. Fig. 7. Impedance change due to L-nicotine exposure to the sensor surface in the concentration range of 1 10 nm (a) and due to L-cotinine exposure in the concentration range of 1 12 nm (b). 10

11 To eliminate the contribution of non-specific adsorption of L-nicotine to the surface of the MIP, it is better to plot the signal in a differential way. This is achieved by subtracting the normalized impedance signal of the NIP from the normalized MIP signal. Fig. 8 shows a clear difference between the reaction of the sensor to L-nicotine and L-cotinine. A linear response is found from 2 to 5 nm with a response of 4.3% increase of the differential signal per nanomolar. For concentrations higher than 5 nm the impedance change is similar for MIP and NIP. This is due to full occupation of the specific binding places in the MIP and in this way the MIP reacts in the same way as the NIP. Fig. 8. Differential signal, NIP subtracted from the MIP. 4. Conclusion The microscopy results show a good immobilization of the powdered MIP to the OC 1 C 10 -PPV covered electrode surface using the PDMS stamping technique. UV- Vis Spectroscopy has been used to characterize the MIPs before implementing them in the sensing device. Binding properties have beene investigated using the Freundlich model, which indicates the occurrence of specific recognition of L-nicotine and insensitivity towards L-cotinine. The measured values of the binding parameters are comparable to literature values. It is demonstrated that by using impedimetric detection, it is possible to detect low-mw molecules using molecularly imprinted polymers, which are immobilized on a conjugated polymer thin film. This technique 11

12 offers the possibility to detect 10 nm with a response increase of 45 percent, which is much lower than relevant medical concentrations. The reference measurements with L-cotinine, which only differs one oxygen atom from L-nicotine, indicates a high specificity. As a fast, easy-to-use sensor with electric readout, the MIP-based impedimetric sensor could be useful for the detection of small molecules in the pharmaceutical, environmental, diagnostic and biotechnological sector. Acknowledgements This work is supported by the School for Life Sciences of transnational University Limburg, by Hasselt University via the BOF-project Development and Characterisation of Fluorescent Molecularly Imprinted Polymers for Nanotechnology and Applications in Advanced Electronic Devices, Chemo- and Bio-sensors and the Fund for Scientific Research Flanders via the Scienctific Research Community FWO- WOG (W N) Hybrid systems at nanometer scale. Technical assistance by J. Sogen and J. Baccus (Hasselt University) is gratefully acknowledged. References Ansell, R. J., Kriz, D., and Mosbach, K., Current Opinion in Biotechnology 7, Blanco-Lopez, M. C., Lobo-Castanon, M. J., Miranda-Ordieres, A. J., and Tunon- Blanco, P., Trends in Analytical Chemistry 23, Bossi, A., Bonini, F., Turner, A. P. F., and Piletsky, S. A., Biosensors and Bioelectronics 22, Cooreman, P., Thoelen, R., Manca, J., vandeven, M., Vermeeren, V., Michiels, L., Ameloot, M., and Wagner, P., Biosensors and Bioelectronics 20, Cormack, P. A. G. and Elorza, A. Z., Journal of Chromatography B-Analytical Technologies in the Biomedical and Life Sciences 804, Dickert, F. L., Lieberzeit, P., Miarecka, S. G., Mann, K. J., Hayden, O., and Palfinger, C., Biosensors & Bioelectronics 20, Ebarvia, B. S. and Sevilla, F., Sensors and Actuators B-Chemical 107, Freundlich, H., Colloid and Capillary Chemistry Methuen, London. 12

13 Harvey, S. D., Mong, G. M., Ozanich, R. M., Mclean, J. S., Goodwin, S. M., Valentine, N. B., and Fredrickson, J. K., Analytical and Bioanalytical Chemistry 386, Haupt, K. and Mosbach, K., Trends in Biotechnology 16, Langmuir, I., J. Am. Chem. Soc. 40, Liao, H. P., Zhang, Z. H., Li, H., Nie, L. H., and Yao, S. Z., Electrochimica Acta 49, Louwet, F., Vanderzande, D., and Gelan, J., Synthetic Metals 69, Merkoci, A. and Alegret, S., Trac-Trends in Analytical Chemistry 21, Panasyuk-Delaney, T., Mirsky, V. M., Ulbricht, M., and Wolfbeis, O. S., Analytica Chimica Acta 435, Piletsky, S. A., Alcock, S., and Turner, A. P. F., Trends in Biotechnology 19, Piletsky, S. A., Turner, N. W., and Laitenberger, P., Medical Engineering & Physics 28, Spivak, D. A., Advanced Drug Delivery Reviews 57, Tan, Y. G., Yin, J., Liang, C. D., Peng, H., Nie, L. H., and Yao, S. Z., Bioelectrochemistry 53, Troje, Z. S., Frobe, Z., and Perovic, D., Journal of Chromatography a 775, Umpleby, R. J., Bode, M., and Shimizu, K. D., Analyst 125,