Molecularly imprinted stimuli-responsive hydrogels for protein recognition

Polymer 53 (2012) 4359e4366

Contents lists available at SciVerse ScienceDirect

Polymer journal homepage: www.elsevier.com/locate/polymer

Molecularly imprinted stimuli-responsive hydrogels for protein recognition Nadia Adrus a, b, Mathias Ulbricht a, * a b

Lehrstuhl für Technische Chemie II, Universität Duisburg-Essen, 45117 Essen, Germany Department of Polymer Engineering, Faculty of Chemical Engineering, Universiti Teknologi Malaysia, 81310 Johor, Malaysia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 February 2012 Received in revised form 29 June 2012 Accepted 26 July 2012 Available online 1 August 2012

Temperature-responsive poly(N-isopropylacrylamide)-based (PNIPAAm) hydrogels were imprinted with lysozyme via in situ photo-initiated crosslinking polymerization. The three-dimensional network of the hydrogels was tailored by tuning the ionic content through methacrylic acid as template-binding comonomer while keeping the ratio between crosslinker (N,N0 -methylenebisacrylamide) and N-isopropylacrylamide fixed. Moderate salt concentrations (0.3 M NaCl) were found to be suited for template removal without phase separation of the hydrogel. Swelling and protein (lysozyme and cytochrome C) binding were investigated for imprinted and nonimprinted gels at temperatures below and above the lower critical solution temperature of PNIPAAm (32  C). Imprinted gels showed a much higher affinity, selectivity and binding capacity for lysozyme compared to the nonimprinted reference materials. Protein binding capacity was strongly reduced above 32  C, to zero for nonimprinted and to small values for imprinted gels. Most important, specific lysozyme binding to the imprinted gels caused a large concentration dependent deswelling. This effect was much smaller for nonimprinted gels, and the response could be modulated by the content of the comonomer methacrylic acid. Overall, this approach is interesting for the development of novel sensors or materials for controlled release applications. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Hydrogel Molecular imprinting Photopolymerization

1. Introduction Smart hydrogels are the basis of designing intelligent materials or systems. These hydrogels can respond to external stimuli and undergo a sudden change in their volume reversibly [1]. Inspired by the natural feedbackeresponse of biological systems, hydrogels are designed to function as artificial recognition elements [1]. Molecular imprinting has unveiled the path for designing tailor made hydrogels with high specificity of molecular recognition properties. Fabrication of molecularly imprinted polymers (MIP) has gained considerable attention in recent years [2e4], and MIPs are commonly prepared by means of non-covalent approach [2,5]. Due to their interesting features, molecular imprinting was investigated for various applications including antibody mimics [6,7], ultrasensitive sensors [8], biomimetic immunoassays [9] and complex separations [10]. As shown in Fig. 1, molecular imprinting proceeds through crosslinking copolymerization of functional monomers having template-binding groups in the presence of template molecules followed by their removal for creation of complementary recognition sites.

* Corresponding author. E-mail address: [email protected] (M. Ulbricht). 0032-3861/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.polymer.2012.07.062

The imprinting of low molecular weight compounds had been successfully established [12e14]. On the contrary, imprinting of relatively larger biomolecules, for instance proteins, remained a challenging task and has been investigated in less detail. Bergmann and Peppas reported that protein-imprinting requires three dimensional spatial orientation to accommodate large and complex dimensions of protein [15]. Correspondingly, designing an appropriate crosslinker to functional monomer ratio is important to ensure the accessibility of binding sites while retaining the recognition specificity of the network [16,17]. In the previous studies, various molecular imprinting protocols with stronger or weaker interactions have been developed for protein recognition, for instance via metal coordination [18] or hydrophobic interactions and hydrogen bonding, respectively [19,20]. Recent studies by Janiak et al. dealt with the effect of hydrogel charge of the MIP in terms of affinity towards bovine haemoglobin or cytochrome C [21,22]. Protein imprinting relying on moderately strong interactions between functional monomers and template assures the ease of the template removal under mild conditions as well as the reversibility of binding and repeatability of use [2]. The use of polyacrylamide (PAAm) for gel-electrophoresis has been also extended for imprinting approach. Copolymerization of PAAm with methacrylic acid (MAA) had been reported for affinity towards lysozyme [23]. The complexation of MAA and lysozyme

4360

N. Adrus, M. Ulbricht / Polymer 53 (2012) 4359e4366

Fig. 1. Principles of molecular imprinting (1 ¼ functional monomers, 2 ¼ crosslinker monomer, 3 ¼ template molecule; a ¼ self-assembly, b ¼ polymerization, c ¼ removal of template molecule). Adapted from reference [11].

had prevented it from free-radical attack and thus facilitated the removal of template-protein after polymerization. Hirayama et al. copolymerized either PAAm with acrylic acid or PAAm with N,Ndimethylaminopropylacrylamide on a silica bead surface for lysozyme recognition, and high sensitivity for lysozyme was observed with a quartz crystal microbalance sensor [24]. Other studies using PAAm as macromolecular matrix for fabrication of MIP materials were reported by Bergmann and Peppas [15], Xia et al. [25] and Hjertén et al. [20,26]. In the view of stimuli-responsivity, PNIPAAmbased hydrogels were used as the main component in the imprinting matrix that showed both temperature-responsive swelling and recognition. Few related examples were provided by the work of Turan et al. for recognition of myoglobin [27], by Chen et al. for recognition of lysozyme or cytochrome C [28] and by Hua et al. for recognition of bovine serum albumin [29]. Similarly, pH [30] and magnetic-responsive recognition [31] were also reported. It was also noticeable that interactions between imprint molecules and imprinted binding sites could induce signals and the amplification of these signals, because the responsivity has led to specific swelling [32] or shrinking [28,33,34]. In those studies, the preparation of MIP was typically done via chemically initiated freeradical copolymerization, since it can be performed under mild conditions and can thus prevent denaturation of protein. To the best of our knowledge, photopolymerization of molecularly imprinted stimuli-responsive hydrogels for protein recognition has not been reported yet. The main general advantages of photopolymerization as compared to thermal or redox initiation for preparation of hydrogels are: i) The polymerization reaction is normally very rapid, selective and occurs in situ [35,36]. ii) The preparation is very well applicable for integration of smart polymeric matrices within membrane formats [37] or microsystems [38]. In addition, fast photopolymerization at low temperature is beneficial for efficiency of molecular imprinting. Thus, the aim of this study was to fabricate temperature-responsive MIP hydrogels for lysozyme recognition via photopolymerization. PNIPAAm hydrogel was copolymerized with MAA as template-binding monomer. A crucial step in the preparation of MIP hydrogels is template-protein removal [39]. The efficiency of template-protein removal depends of several factors such as washing medium and conditions, composition of the hydrogels, in particular the choice of comonomer, kinetics of polymerization and many more. We believed that tuning the ionic interactions through various MAA content could facilitate both the specific binding and the effective elution of the template-protein under mild conditions, so that this could be a most efficient strategy to improve imprinting efficiency and to increase recognition binding. The significant binding capacity and selectivity of MIP over non-imprinted polymer (NIP) hydrogels were demonstrated with responsive swellingedeswelling that modulates the binding abilities. Most

importantly, specific volume shrinking was observed upon rebinding the template-protein; i.e., recognition and signal transduction can be combined within one material. Due to these interesting properties, this approach is envisioned as a step towards establishment of intelligent materials for sensor or drug delivery applications. 2. Experimental 2.1. Materials All chemicals used were at least analytical grades. N-isopropylacrylamide (NIPAAm) and tris(hydroxymethyl)aminomethane (Tris) (base) were purchased from Acros (Geel, Belgium). Photoinitiator (PI) 1-[4-(2-hydroxyethoxy)-phenyl]-2-hydroxy-2methyl-1-propane-1-one under the trade name Irgacure 2959Ò was supplied by BASF-Ciba Chemicals (Basel, Switzerland). N,N0 methylenebisacrylamide (MBAAm) and methacrylic acid (MAA) were purchased from SigmaeAldrich (Steinheim, Germany). Lysozyme (chicken egg white, MW 14.6 kDa, isoelectric point (IEP) 10.8), and cytochrome C (bovine heart, MW 12.4 kDa, IEP 9.8) were obtained from Fluka (Belgium). NIPAAm was recrystallized from nhexane. All other chemicals were used as received without further purification. Water purified with a Milli-Q system from Millipore (USA) was used for all experiments. 2.2. Synthesis of hydrogels Depending on the net charge, the diameter of lysozyme is about 3e4 nm [40]. From our previous study [35], the mesh size of hydrogel with composition of M15DC05 (cf. below) determined from rheological characterization was 5.6 nm. Thus, the hydrogel composition for imprinting procedure was developed from this previous experiment to give an appropriate crosslinker to functional monomer ratio as suggested in the literature [2]. The homogeneous precursor hydrogel solutions (Table 1 and 2) were prepared in triseHCl buffer solution (10 mM, pH 7.0) and contained NIPAAm as a main monomer (15 wt% / M15), MAA as a templatebinding monomer (1e10 wt% relative to NIPAAm / M0 01, M0 02, M0 05, M0 10), MBAAm as a crosslinker (5 wt% relative to NIPAAm / DC05) as well as PI (2 wt% relative to NIPAAm). The MIP and NIP hydrogels were prepared with and without lysozyme (10 wt% relative to NIPAAm), respectively. The solution was transferred into a glass vessel (50 mL, diameter w25 mm), purged with argon and then pre-cooled in the ice bath (w4  C) for 10 min. The solution was exposed to UV light (UVA system, Hönle AG, with l > 300 nm and intensity w30 mW/cm2) for 15 min at 4  C. After the polymerization, the hydrogel was cut into disc-like pieces (w1e2 mm in thickness), and washed in aqueous NaCl solution (0.3 or 1.0 M); the solution was refreshed every 24 h for several times. Extraction of the unreacted monomer and lysozyme were monitored using total organic carbon analyser (TOC-Vcpn system, Shimadzu, Japan), and the data were used to

N. Adrus, M. Ulbricht / Polymer 53 (2012) 4359e4366

4361

Table 1 Reaction mixture compositions for synthesis of NIP and MIP hydrogels with ionic comonomer. Hydrogel NIP

MIP

M15M0 01DC05 M15M0 02DC05 M15M0 05DC05 M15M0 10DC05 M15M0 01DC05L10 M15M0 02DC05L10 M15M0 05DC05L10 M15M0 10DC05L10

Mass of NIPAAm [g]

Mass of PI [g]

Mass of MAA [g]

Mass of MBAAm [g]

Mass of lysozyme [g]

Volume of buffer [ml]

3.75 3.75 3.75 3.75 3.75 3.75 3.75 3.75

0.075 0.075 0.075 0.075 0.075 0.075 0.075 0.075

0.0375 0.0750 0.1875 0.3750 0.0375 0.0750 0.1875 0.3750

0.1875 0.1875 0.1875 0.1875 0.1875 0.1875 0.1875 0.1875

0 0 0 0 0.375 0.375 0.375 0.375

25 25 25 25 25 25 25 25

a protein mixture at 20  C under shaking condition. This protein solution consisted of equimass amounts of lysozyme and cytochrome C (total concentration 0.5 g/L). After the equilibrium sorption was reached, the remaining protein in the solution was measured using UVeVIS spectrometry at 280 nm and 408 nm, corresponding to the absorbance of total protein (considering their individual absorbance coefficients) and cytochrome C, respectively.

calculate the monomer to hydrogel conversion from the mass balance. The efficiency of the template-protein removal was evaluated by comparing the conversions obtained for MIP (assuming that the template-protein will be eluted completely) and NIP hydrogels. All pieces of hydrogels were freeze-dried at 18  C for two days prior to further characterization. 2.3. Characterization of hydrogels

3. Results and discussion 2.3.1. Swelling test About 10 mg of dried gels pieces were placed in triseHCl buffer (10 mM, pH ¼ 7.0) within a temperature range from 20 to 45  C for 24 h. After this equilibration, the mass of these samples was determined. The degree of swelling was calculated from the ratio of the mass of the hydrogel in the swollen and dried states based on the average from triplicate measurements.

3.1. Template-protein removal and physical properties of the hydrogels Template-protein removal is based on the fact that the interactions between the template and the template-binding groups in the polymer can be broken by a washing medium. In literature, different NaCl concentrations, usually from 0.1 to 1.0 M, were often used for washing. The ionic strength in the washing solution increases the efficiency of protein removal because ionic interactions between the protein and functional groups (here carboxylate) can be broken, but this condition might not be optimal. Fig. 2 shows the effect of washing MIP and NIP hydrogels with different salt concentration. As can be seen, washing with higher salt concentration (1.0 M) lead to shrinkage of the gels, and this was found to be irreversible. This can be associated to lowering the osmotic pressure because of reduced difference between the concentrations of counterions in the network and in the surrounding solution. The resulting aggregation of the hydrophobic network segments may not be fully reversible under aqueous conditions. Thereby, the protein may become entrapped inside the polymeric network. This may also cause negative effects on rebinding process. The hydrogels were also washed using a lower salt concentration (0.3 M, Fig. 2). This condition was chosen based on extended preliminary experiments. Hydrogels treated under those conditions maintained their integrity (Fig. 2). Another indication that elution of template-protein took place during the washing process was that the slightly yellowish colour of MIP hydrogel obtained directly after the synthesis decreased significantly to a transparent appearance. However, this was not the case for washing with 1.0 M NaCl. In the literature [23] it had been stated that shielding lysozyme from being copolymerized with monomers has facilitated a relatively easier template-protein removal under mild conditions. It was hypothesized that this was due to complexation of MAA which in turn prevented lysozyme from the attack of free radicals. It was

2.3.2. Rebinding test with template-protein All rebinding tests were performed for fully swollen hydrogels (mass of dried gels w 10 mg). For obtaining an equilibrated sorption, the experiment was done for at least 3 days. 2.3.2.1. Effect of template-binding monomer content. The rebinding test was performed at constant lysozyme concentration (10 mL, 0.5 g/L in triseHCl buffer). The MIP and NIP hydrogels with various MAA content were incubated in the template-protein solution at 20  C and placed on a shaker. The binding was monitored from the decrease of lysozyme UV absorbance signal at 280 nm using a UVeVIS spectrometer (Varian Cary 50 Probe, Santa Clara, USA). 2.3.2.2. Sorption isotherm. For sorption isotherm test, hydrogels with 2 wt% MAA content (M0 02) were equilibrated in various lysozyme concentrations (initial concentration range 0e5 g/L) on a shaker at 20  C. Analogous protocol as above to determine the amount of lysozyme bound to the samples was employed. 2.3.2.3. Effect of temperature. The hydrogel samples (M0 02) were equilibrated in a solution of 0.5 g/L lysozyme in triseHCl buffer. The experiments were performed on a shaker at temperatures below or above or in the vicinity of the LCST of PNIPAAm. Analysis of the amount of bound protein was done as described above. 2.3.3. Competitive protein binding tests To determine the selectivity of MIP hydrogel to lysozyme, experiments were carried out for the samples (M0 02) in 10 mL of

Table 2 NIP and MIP versions of P(NIPAAm-co-MAA) hydrogels with varied MAA content synthesized via photopolymerization (2 wt% PI and 10 wt% lysozyme relative to NIPAAm; 15 min UV time). Gel

NIP MIP

Content of MAA relative to NIPAAm (M0 ) 00

01

02

05

10

M15M0 00DC05 e

M15M0 01DC05 M15M0 01DC05L10

M15M0 02DC05 M15M0 02DC05L10

M15M0 05DC05 M15M0 05DC05L10

M15M0 10DC05 M15M0 10DC05L10

4362

N. Adrus, M. Ulbricht / Polymer 53 (2012) 4359e4366 Table 4 Molar ratios between template-binding monomer MAA and lysozyme from protein binding to NIP hydrogels.

Fig. 2. Photograph showing the effect of sodium chloride concentration used for washing of MIP and NIP hydrogels prepared with 5 wt% MAA.

also reported that the fraction of extractable lysozyme increased with increasing MAA content copolymerized to PAAm [23]. However this was not the case for P(NIPAAm-co-MAA) in this study. Presumably, by increasing content of MAA, more electrostatic binding sites may decrease the efficiency of protein removal. A higher strength of interaction can be explained based on coordinated multivalent binding of more than one carboxylic group of the polymer to one protein molecule as seen for the NIP hydrogels (cf. Section 3.3; Table 4). The incomplete template-protein removal will have a negative influence for recognition binding; it will mask binding sites. It may also impose a resistance towards water uptake or mass transport; that can reduce swelling or binding capacity. In general, an increasing fraction of electrostatic interactions by comonomer segments in the hydrogel makes the material more similar to an ion exchanger matrix. As shown in Table 3, NIP and MIP hydrogels synthesized with 2 or 5 wt% MAA content relative to NIPAAm achieved a conversion of around 90%. High degree of monomer conversion is very important due to several factors. The accuracy of network quantification based on statistical models depends on the high monomer conversion. Secondly, high conversion will ensure the mechanical stability of hydrogels for certain applications (e.g., separation media). In

Table 3 Monomer to hydrogel conversion for NIP and MIP with various MAA contents after washing in 0.3 M NaCl solution (M0 02 and M0 05 represent 2 and 5 wt% MAA relative to NIPAAm, respectively). Gel

Composition

NIP

M15 M15 M15 M15

MIP

M0 02 M0 05 M0 02 M0 05

DC05 DC05 DC05 DC05

Conversion [%] Lys0 Lys0 Lys10 Lys10

94 97 87 92

M0 [%]

Binding capacity [mg/g]

NM0 /Nlysozyme

1 2 5

39 199 313

0.6:1 2:1 5:1

addition, proportionally more ionic sites with increasing MAA content can only be assured at high monomer conversion. Assuming that lysozyme is only reversibly bound in the hydrogel, the elution of template-protein (as well as of unreacted monomers) in the washing medium of the NIP and MIP gels can be investigated by monitoring the TOC values. If monomer conversion is the same for MIP and NIP and the template-protein is removed from MIP hydrogels, a correspondingly higher TOC value for MIP than NIP will be measured. However, the source of TOC cannot be distinguished. Overall lower apparent monomer conversion for MIP than for NIP gels was observed (Table 3), what could indicate that template-protein was entirely removed. Hydrogels synthesized with various MAA content could be distinguished visually. As shown in Fig. 3, hydrogels prepared with 5 wt% MAA were transparent (homogeneous gel). In contrast, hydrogels prepared with higher MAA content (i.e., 10 wt%) were opaque. This indicated macroscopic heterogeneities of the network. NIPAAm with isopropyl side group is relatively more hydrophobic than MAA. Thus, with increasing MAA content, phase separated hydrogels were obtained, most likely because more ionic groups lead to more water what is bound to MAA-based network segments in the network, competing with the solvation of NIPAAm-based segments via hydrogen bonds. In addition, the phase separation could also have taken place as a result of chemical incompatibility of the comonomers during the polymerization [41]. 3.2. Temperature-responsivity Fig. 4 shows the temperature-dependent swelling behaviour of NIP and MIP hydrogels prepared with 2% template-binding monomer when the temperature of the buffer increased from 15 to 45  C. Below the LCST, all the hydrogels possessed high degree of swelling; a strong decrease of swelling was observed between 30 and 35  C. This observation is consistent with the classical temperature-responsive behaviour of PNIPAAm homo- or copolymers reported in the literature [35,36,42,43]. Furthermore, significant differences between swelling degree below and above LCST of PNIPAAm were observed regardless the content of template-binding monomer or the presence of protein during synthesis (Fig. 5). These data confirmed the temperatureresponsivity of all NIP and MIP gels synthesized in this work. On the one hand, the degree of swelling below LCST increased with increasing MAA content for both NIP and MIP gels. By increasing template-monomer content, increasing ionic group content is the reason for high water uptake. On the other hand, mainly below LCST and also at lower MAA content; i.e., up to 2 wt%, the equilibrium swelling ratio of the MIP hydrogels was higher than that of their NIP counterparts. This is presumably because the molecular imprinting technique creates nanoscale cavities, with additional sites for hydration within the hydrogel matrix (cf. ref. [39]). However, this was not the case for the MIP hydrogels prepared with the highest template-monomer content; and that lower swelling compared to the NIP gels could be due to incomplete protein removal (cf. 3.1). This would be analogous to the effect of protein binding to MIP gels, i.e., a deswelling of the gel (this will be shown and discussed below; cf. 3.4).

N. Adrus, M. Ulbricht / Polymer 53 (2012) 4359e4366

4363

12.0

20°C

45°C

Swelling degree

10.0 8.0 6.0 4.0 2.0 0.0 M'01

M'02

NIP

M'05

M'01

M'02

M'05

MIP

Fig. 5. Degree of swelling for MIP and NIP hydrogels at two temperatures as a function of MAA content.

3.3.1. Binding capacity and imprinting factor The influence of template-binding monomer content onto binding behaviour of hydrogels with the template-protein at pH 7 and RT was investigated (Fig. 6). A very clear correlation between binding capacity and MAA content had been observed; binding capacity increased with increasing template-monomer content. This trend is analogous to the influence of MAA content on swelling observed in Fig. 5. At neutral pH, lysozyme is positively charged while MAA has a negative fixed charge due to deprotonation of carboxyl groups. Therefore the template-monomer imposed electrostatic interactions either with water or protein due to its ionic character. Almost no protein binding (w1 mg/g) was observed for the hydrogels prepared without template-monomer (cf. Fig. 6). This experiment implied that ionic interaction is necessary for binding.

Binding observed for NIP hydrogels was associated only to the non-specific electrostatic interactions. The MIP hydrogels on the other hand bound significantly more protein than NIP hydrogels (Fig. 6). Imprinting factor is the ratio between binding capacities of MIP and NIP gels. First, the imprinting factor increased with increasing template-binding monomer content. However, a further increase of template-monomer content (5%) reduced the imprinting factor (cf. Fig. 6) as a result of increasing non-specific binding within NIP hydrogels. The effect of template-binding comonomer content onto the binding capacity of the gel was therefore analysed. For this purpose, the number of template-monomer molecules (NM’) to bind one lysozyme molecule (Nlysozyme) was estimated based on the binding data of NIP hydrogels. Table 4 shows that more than one MAA molecule in the M’02 and M’05 hydrogels binds one lysozyme molecule. This indicates stronger electrostatic interactions with increasing template-binding comonomer. Thus, it is more likely that at higher MAA content, ion exchanger characteristic of the gel may have stronger influence than just only shielding lysozyme during polymerization (cf. Section 3.1). Even more critical would be the non-specific binding besides recognition in imprinted sites in MIP hydrogels. The non-specific binding can hence be reduced using a lower content of the template-monomer, but this reduced the overall binding capacity. However, the reduction was less pronounced for the MIP hydrogels.

Fig. 4. Degree of swelling for MIP and NIP hydrogels prepared with 2 wt% MAA as function of temperature.

Fig. 6. Binding capacities and resulting imprinting factor obtained with 0.5 g/L lysozyme in TriseHCl buffer (10 mM, pH 7.0) at RT.

Fig. 3. Photograph showing the influence of MAA content on properties of MIP and NIP gels (washing: 0.3 M NaCl).

3.3. Protein binding

4364

N. Adrus, M. Ulbricht / Polymer 53 (2012) 4359e4366

3.3.2. Protein sorption isotherm Binding capacity was further investigated for hydrogels prepared with 2% template-monomer content as a function of lysozyme concentration (Fig. 7). An increase in binding capacity was observed until a saturation level. Overall, MIP exhibited larger binding capacities than NIP hydrogels, especially at low lysozyme concentration. Besides ionic interactions, the higher binding capacity obtained for MIP hydrogels was mainly due to the recognition behaviour towards lysozyme as its template-protein. Thus, recognition favoured a binding process through the fitting of multiple binding sites within the hydrogel matrix created by the imprinting procedure and summation of non-covalent interactions such as electrostatic interactions and hydrogen bonding. Considering the shape of the curves in Fig. 7, the data have been fitted to the Langmuir isotherm model in order to estimate the binding constant. The binding constants for MIP and NIP hydrogels obtained in this study (3.9  105 M1 and 7.7  104 M1, respectively) were in a good agreement with those reported in the literature for similar materials prepared via redox-initiated polymerization (4.2  105 M1 and 9.9  104 M1, respectively; cf. [28]). 3.3.3. Protein binding selectivity In order to determine the selectivity of MIP hydrogel towards its template-protein, a competition sorption test was performed. A protein mixture of 0.5 g/L lysozyme and cytochrome C was used. Cytochrome C was used due to its similar properties compared to lysozyme in terms of size and IEP (cf. Section 2.1). The result of this experiment for hydrogels prepared with 2% template-binding monomer is shown Fig. 8. Firstly, NIP and MIP hydrogels demonstrated no significant difference in binding toward the reference protein cytochrome C. This competition experiment clearly indicates that cytochrome C binding was caused by non-specific interactions. For the NIP hydrogels, more cytochrome C than lysozyme was bound. It should be noted that the cytochrome C molecule is slightly smaller than the lysozyme molecule (cf. Section 2.1), and, thus, the cross linked NIP hydrogels could bind more cytochrome C because of better accessibility of the binding sites. For a simple grafted PAAm-based adsorber, somewhat more cyctochrome C than lysozyme was bound, what had been related to larger sterical hindrance for the larger protein [44]. However, since the difference in size was small, this could not fully justify the relatively large difference in binding for NIP hydrogels. Also, on standard chromatography materials, either cation-exchange or

Binding capacity [mg/g]

800.0

600.0

400.0

200.0 MIP

Fig. 8. Protein binding selectivity for MIP and NIP hydrogels determined with total initial protein concentration of 0.5 g/L in triseHCl buffer (10 mM, pH 7.0) at RT; hydrogel composition: M15M0 02DC05.

reversed phase, a slightly lower retention of cytochrome C, i.e., slightly weaker binding, than lysozyme is often observed. The two proteins differ with respect to the structure of the basic amino acids: Lysozyme is mainly arginine-based, whereas cytochrome C is mainly lysin-based. This had been used for creating lysozyme selectivity based on polymers with arginine-selective side groups [44]. In the same study, the capacity of a grafted cation-exchanger material for the two proteins had been the same [44]. However, most importantly, the data in Fig. 8 reveal that MIP hydrogels exhibited high selectivity towards lysozyme compared to cytochrome C; i.e., just the opposite compared to the NIPs with the same composition. Even though lysozyme and cytochrome C have similar properties at their molecular level, MIP hydrogels preferentially bind template-protein; i.e., lysozyme to the larger extent than cytochrome C. This is a strong indication that MIP gels can distinguish proteins not simply on the basis of size or charge, but to a large extent due to recognition and binding fitting complementarily to the shape of the template-protein. 3.3.4. Temperature induced binding response The effect of temperature on protein binding of NIP and MIP hydrogels is shown in Fig. 9. It can be seen that the binding capacity of both hydrogels decreased as the temperature increased above LCST of PNIPAAm and the lowest binding was observed at highest temperature; i.e., 45  C. The temperature-induced binding response can be explained as follows. The MIP hydrogels serve as the selective recognition element provided that their recognition sites are preserved analogous to the imprinting state. Due to the temperature-responsivity, PNIPAAm based hydrogels collapsed into a compact 3D network and this was not accessible for the protein binding. In addition, deformation of imprinted sites may also have occurred. Nevertheless, MIP hydrogels had overall higher binding capacity than NIP hydrogels also above LCST. This would indicate that a small fraction of imprinted sites was probably on the outer surface of the gel, also at higher temperature with distinct affinity for lysozyme binding.

NIP 3.4. Specific volume change induced by protein-responsivity and its proposed mechanism

0.0 0

1

2

3

4

5

Lysozyme concentration [g/L] Fig. 7. Binding isotherms for MIP and NIP hydrogels with lysozyme in triseHCl buffer (10 mM, pH 7.0) at RT; hydrogel composition: M15M0 02DC05. Binding constants for MIP and NIP hydrogels were 3.9  105 M1 and 7.7  104 M1, respectively.

The binding of lysozyme in the MIP hydrogels at a temperature below LCST resulted in a strong decrease in hydrogel volume. This effect was first studied for various protein concentrations as shown in Fig. 10. It can be clearly seen that increasing lysozyme concentration continuously induced shrinking of the MIP hydrogels up to

N. Adrus, M. Ulbricht / Polymer 53 (2012) 4359e4366

4365

1 NIP

MIP

Vads / Vsw

0.95 0.9 0.85

0.8 0.75

0.7 M'00 Fig. 9. Influence of temperature on lysozyme binding of M15M0 02DC05 MIP/NIP hydrogels. Initial protein concentration 0.5 g/L; hydrogel composition: M15M0 02DC05.

a protein concentration of about 1.2 g/L; then swelling became constant. In case of NIP hydrogels, only slight volume decrease was observed and this occurred only at high lysozyme concentration which is most probably due to non-specific effect of ion pairing. Specific shrinking for MIP gels was observed due to specific interactions between template-protein and imprinting sites and subsequent conformational change within the hydrogel network. Presumably, the best proposed mechanism for this significant specific volume shrinking is that binding to imprinted sites causes local conformation and solvation changes which will ultimately lead to different solvation also in the wider environment of the imprinted sites. Specifically, formation of ion pairs between carboxylate and cationic protein groups will reduce the charge density in the gel and hence lead to lower amount of bound water. The release of bound water could be a strong entropic driving force. This process is started by a recognition sequence of MIP hydrogels followed by adaptation sequence; i.e., conformational adaptation within the network towards the template-protein that ultimately leads to shrinking. Similarly, the protein-responsivity was also studied for variation of template-binding monomer content as shown in Fig. 11. Clearly, proportional tendency of volume shrinking and swelling was found. In the absence of ionic groups, no change of volume was observed. Shrinking became more pronounced with increasing 1 NIP

0.95

Vads / Vsw

MIP

0.9 0.85 0.8

0.75 0.7 0

1

2

3

4

5

Lysozyme concentration [g/L] Fig. 10. Volume shrinking for MIP and NIP hydrogels as function of lysozyme concentrations in triseHCl buffer (10 mM, pH 7.0) at RT; hydrogel composition: M15M0 02DC05.

M'01

M'02

M'05

Fig. 11. Volume shrinking upon protein binding for MIP and NIP hydrogels with increasing template-monomer content. Initial protein concentration of 0.5 g/L lysozyme in TriseHCl buffer (10 mM, pH 7.0) at RT.

template-monomer content. Several explanations can be offered. It is most likely, that more bound water can be released from these hydrogels due to an increasing formation of ion pairs between carboxylate and cationic protein groups. Higher templatemonomer content introduced more imprinted sites with higher water sorption (cf. Section 3.2) and therefore induced larger degree of shrinkage upon protein binding. Because the gels were macroscopically homogeneous (cf. Fig. 3) and relatively large samples (1e2 mm thick; cf. Sections 2.2 and 2.3) have been used, the times for complete deswelling and swelling upon changes of conditions were relatively long. Therefore, only equilibrium data (after 24 h) have been recorded. However, it is known that the rate can largely be increased by either introducing macropores into the bulk hydrogels (cf. [42]) or by reducing the materials thickness [45]; for dimensions in the lower mm range, response times in the range of seconds are to be expected. 4. Conclusions PNIPAAm-based NIP and lysozyme MIP hydrogels were synthesized via photo-initiated free radical copolymerization. To our knowledge, photopolymerized MIP hydrogels for protein recognition had not been reported yet. Overall, this study has made four important contributions: i) The advantages of photopolymerization techniques e especially for in situ fabrication or integration of hydrogels within microsystems e can also be used for the preparation of protein-imprinted hydrogels. ii) Template-protein removal without irreversible hydrogel collaps is possible using a moderate NaCl concentration e through optimization of ionic content of MIP hydrogels at sufficient degree of monomer conversion. iii) MIP hydrogels show higher protein binding capacity and selectivity towards lysozyme than NIP hydrogels, and the temperature-responsive swelling-deswelling of the PNIPAAm-based materials modulated the binding ability. iv) Protein recognition and signal transduction can be combined within one material because selective protein to the MIP hydrogels leads to a pronounced deswelling. Achieving such large effects is possible by the combination of imprinting with a hydrogel matrix showing a critical phase transition; when using these materials at a temperature close to LCST, the

4366

N. Adrus, M. Ulbricht / Polymer 53 (2012) 4359e4366

ability to undergo the phase transition will amplify the response to protein recognition/binding. This will enable additional functions of such hydrogels, for instance the development of protein (micro)sensors based on measuring the swelling pressure (cf. [46]). Due to these interesting features, this approach is envisioned to facilitate establishment of novel sensor or other biomimetic materials. As such, preparation of MIP hydrogels via photopolymerization can also be adapted to a macroporous support, for instance via pore-filling functionalization in order to obtain MIP hydrogels as recognition and transducer element in a proteinselective membrane (cf. [37]). Acknowledgements The Ministry of Higher Education Malaysia is gratefully acknowledged for supporting the Ph.D. scholarship for N.A. Valuable preliminary data on hydrogels washing by Tobias Hennecke and experimental support by Kevin Bojaryn are also greatly acknowledged. References [1] Cohen Stuart MA, Huck WTS, Genzer J, Müller M, Ober C, Stamm M, et al. Nat Mat 2010;9:101e13. [2] Byrne ME, Park K, Peppas NA. Adv Drug Deliv Rev 2002;54:149e61. [3] Komiyama M, Takeuchi T, Mukawa T, Asanuma H. Molecular imprinting: from fundamentals to applications. Weinheim: Wiley-VCH; 2003. [4] Alexander C, Andersson HS, Andersson LI, Ansell RJ, Kirsch N, Nicholls IA, et al. J Mol Recog 2006;19:106e80. [5] Sellergren B. Trends Anal Chem 1997;16:310e20. [6] Wulff G, Gross T, Schönfeld R, Schrader T, Kirsten C. Molecular imprinting for the preparation of enzyme-analogous polymers. In: Molecular and ionic recognition with imprinted polymers, vol. 703. ACS Symposium Series; American Chemical Society; 1998. p. 10e28. [7] Andersson LI, Nicholls IA, Mosbach K. Antibody mimics obtained by noncovalent molecular imprinting. In: Immunoanalysis of agrochemicals, vol. 586. ACS Symposium Series; American Chemical Society; 1995. p. 89e96. [8] Cai D, Ren L, Zhao H, Xu C, Zhang L, Yu Y, et al. Nat Nanotech 2010;5:597e601. [9] Xu ZX, Gao HJ, Zhang LM, Chen XQ, Qiao XG. J Food Sci 2011;76:R69e75.

[10] Sellergren B. Separation of enantiomers using molecularly imprinted polymers. In: Subramanian G, editor. Chiral separation techniques. Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA; 2007. p. 399e431. [11] Haupt K. Anal Chem 2003;75:376 Ae83 A. [12] Kugimiya A, Kuwada Y, Takeuchi T. J Chromatogr A 2001;938:131e5. [13] Ikegami T, Lee WS, Nariai H, Takeuchi T. Anal Bioanal Chem 2004;378: 1898e902. [14] Vlatakis G, Andersson LI, Muller R, Mosbach K. Nature 1993;361:645e7. [15] Bergmann NM, Peppas NA. Ind Eng Chem Res 2008;47:9099e107. [16] Scott RA, Peppas NA. Macromolecules 1999;32:8674e7. [17] Stancil KA, Feld MS, Kardar M. J Phys Chem B 2005;109:6636e9. [18] Qin L, He XW, Zhang W, Li WY, Zhang YK. Anal Chem 2009;81:7206e16. [19] Hjertén S. J Chromatogr A 1973;87:325e31. [20] Hjertén S, Liao J, Nakazato K, Wang Y, Zamaratskaia G, Zhang H. Chromatographia 1997;44:227e34. [21] Janiak DS, Ayyub OB, Kofinas P. Macromolecules 2009;42:1703e9. [22] Janiak DS, Ayyub OB, Kofinas P. Polymer 2010;51:665e70. [23] Ou SH, Wu MC, Chou TC, Liu CC. Anal Chim Acta 2004;504:163e6. [24] Hirayama K, Sakai Y, Kameoka K. J Appl Polym Sci 2001;81:3378e87. [25] Xia YQ, Guo TY, Song MD, Zhang BH, Zhang B-L. Biomacromolecules 2005;6: 2601e6. [26] Liao JL, Wang Y, Hjertén S. Chromatographia 1996;42:259e62. [27] Turan E, Özçetin G, Caykara T. Macromol Biosci 2009;9:421e8. [28] Chen Z, Hua Z, Xu L, Huang Y, Zhao M, Li Y. J Mol Recog 2008;21:71e7. [29] Hua Z, Chen Z, Li Y, Zhao M. Langmuir 2008;24:5773e80. [30] Uysal A, Demirel G, Turan E, Çaykara T. Anal Chim Acta 2008;625:110e5. [31] Kan X, Zhao Q, Shao D, Geng Z, Wang Z, Zhu JJ. J Phys Chem B 2010;114: 3999e4004. [32] Watanabe M, Akahoshi T, Tabata Y, Nakayama D. J Am Chem Soc 1998;120: 5577e8. [33] Miyata T, Jige M, Nakaminami T, Uragami T. Proc Natl Acad Sci U S A 2006; 103:1190e3. [34] Sarhan, Wulff G. Makromol Chem 1982;183:85e92. [35] Adrus N, Ulbricht M. React Func Polym, under revision. [36] Singh D, Kuckling D, Choudhary V, Adler HJ, Koul V. Polym Adv Technol 2006; 17:186e92. [37] Adrus N, Ulbricht M. J. Mater Chem 2012;22:3088e98. [38] Yu C, Svec F, Fréchet JMJ. Electrophoresis 2000;21:120e7. [39] Bereli N, Andaç M, Baydemir G, Say R, Galaev IY, Denizli A. J Chromatogr A 2008;1190:18e26. [40] Kisler JM, Steven GW, O’Connor AJ. Mater Phys Mech 2001;4:89e93. [41] Turan E, Demirci S, Caykara T. J Polym Sci Polym Phys 2008;46:1713e24. [42] Fänger C, Wack H, Ulbricht M. Macromol Biosci 2006;6:393e402. [43] Zhang XZ, Wu DQ, Chu CC. J Polym Sci Part B Polym Phys 2003;41:582e93. [44] He D, Sun W, Schrader T, Ulbricht M. J Mater Chem 2009;19:253e60. [45] Naini CA, Franzka S, Frost S, Ulbricht M, Hartmann N. Angew Chem Int Ed 2011;50:4513. [46] Wack H, Ulbricht M. Polymer 2009;50:2075e80.