Specific removal of protein using protein imprinted polydopamine shells on modified amino-functionalized magnetic nanoparticles

Abstract

Thin imprinted shells over functionalized magnetic nanoparticles is an effective solution to weaken mass transfer resistance, achieve high binding capacity, and attain rapid separation. In this work, a simple, green, and effective approach was developed to imprint bovine serum albumin (BSA) on the surface of amino-modified Fe₃O₄ nanoparticles (Fe₃O₄@NH₂) using dopamine as monomer through a two-step immobilized template strategy. The results of X-ray diffraction and vibrating sample magnetometry indicated that the as-synthesized nanomaterials exhibited high crystallinity and satisfactory superparamagnetic properties. Transmission electron microscopy and Fourier transform infrared spectroscopy of the products showed that polydopamine shells successfully attached onto Fe₃O₄@NH₂. The polydopamine shells with a thickness of about 10 nm enable the template recognition sites to be accessed easily, and exhibit fast kinetics and high adsorption capacity in aqueous solution. Meanwhile, an excellent selectivity towards BSA has been presented when bovine hemoglobin (BHb), transferrin, and immunoglobulin G (IgG) were employed as competitive proteins. Good recovery after a reasonably mild elution and successful capture of the target protein from a real sample of bovine blood suggests its potential value in practical applications. In addition, the resultant polymers were stable and had no obvious deterioration after six adsorption–regeneration cycles. The versatility of the proposed method has also been verified by choosing four other proteins with different isoelectric points as the templates.

---

1. Introduction

Affinity separation and removal of high-abundant proteins from blood plasma is often used as a sample pretreatment technique before clinical and pharmaceutical proteomic studies. Several feasible methods, such as immobilized metal affinity chromatography (IMAC),^1,2 dielectrophoresis,^3,4 and immunoaffinity chromatography^5,6 have been applied to remove high-abundant proteins from biological samples. However, there are some drawbacks, including cost, time and labor consumption, poor stability, limitations of the targets, and non-reusability that prohibit the widespread applicability of these approaches. Therefore, to develop a more effective and general application method for the specific isolation and depletion of high-abundant proteins is of vital significance.

Molecularly imprinted polymers (MIPs), tailor-made receptors with specific three-dimensional recognition sites complementary in shape, size, and functionality with the templates,^7,8 are considered one of the most promising affinity matrices. MIPs have also been regarded as a satisfactory adsorbent for the separation and removal of target proteins,^9,10 as a result of good physical, chemical, and thermal stability, simple and economical production, durability, and reusability, as well as high specificity. With respect to protein imprinting, a big problem is the large molecular size which restricts the mass transfer of proteins between the cross-linked polymer network and the solution. Another problem is aqueous media where proteins prefer to exist, but the noncovalent interactions of templates and monomers can be reduced remarkably. Other problems include flexible conformation, complexity, and numbers of functional groups, all of which are unfavorable elements for imprinting proteins.

Some imprinting approaches have been developed to solve the mass transfer difficulty of proteins, for example, surface imprinting,^11,12 epitope imprinting,^13,14 and metal-coordination polymerization.^15,16 Among these, surface imprinting with the recognition cavities located at the surface of the MIPs, is regarded as the most promising method for imprinting proteins, as it is expected to weaken mass transfer resistance and make the removal of the template protein easy and complete.^17,18 Nanomaterials are ideal supports for surface imprinting, especially magnetic Fe₃O₄ nanoparticles (NPs), due to their superparamagnetism and simple preparation.^19,20 These excellent properties make imprinted Fe₃O₄ NPs useful for improving the accessibility of the recognition cavities generated and the separation efficiency of polymers from solution. Several protein molecularly imprinted works based on Fe₃O₄ NPs have been reported, in which silane or other active groups (such as –C=C, –COOH, –CHO, and –NH₂) were introduced, often involving harsh conditions and/or multiple functionalization steps.^21–23 Furthermore, the synthesis process was usually performed in organic solvents, which is a minus point for protein imprinting. Therefore, more effective and simpler strategies for protein imprinting on the surface of Fe₃O₄ NPs are still required. Our previous work²⁴ developed a much easier synthetic process for BHb imprinting employing amino group directly-functionalized Fe₃O₄ NPs as carriers. Though this work displayed some advantages in aspects of preparation technology, the hydrophilicity of the obtained products was unsatisfactory in virtue of adopting hydrophobic monomers.

Considering the solubility of protein, monomers with multifunctional groups and properties of hydrophilicity and biocompatibility are appropriate for imprinting proteins. Dopamine (DA), containing catechol and amine groups, is accepted as a popular adhesive for proteins.^25,26 The catechol groups are liable to be oxidized to generate o-quinone functionalities which make DA form thin polydopamine (PDA) films onto different kinds of materials in weak alkaline media at room temperature. To date, a certain number of successful examples using DA as the monomer and cross-linker for protein imprinting have been reported.^27–29 However, to the best of our knowledge, adopting DA as a monomer for the imprinting of proteins on the surface of amino group directly-functionalized Fe₃O₄ NPs has not been reported yet.

Taking into account the advantages of surface imprinting technology, magnetic separation, and PDA, the key idea of this work is to develop a facile approach to imprint protein on the surface of amino group directly-functionalized Fe₃O₄ NPs using dopamine as monomer through a two-step immobilized template process. The procedure of this preparation method was rather green and simple. The obtained nanomaterials could be easily separated from the solution under an external magnet within three seconds. In order to obtain the best recognition performance, the polymerization and adsorption conditions were investigated in detail. In addition, the practicability of the as-prepared imprinted polymers for biological application was further investigated by the separation and removal of a target protein from a real bovine blood sample. Furthermore, four other proteins with different isoelectric points were chosen as the templates to evaluate the versatility of the approach developed in this work.

2. Experimental

2.1. Materials

Dopamine (DA) was purchased from Alfa Aesar Chemical Company. 1,6-Diaminohexane (DAH), anhydrous sodium acetate (NaOAc), ferric chloride hexahydrate (FeCl₃·6H₂O), ethylene glycol (EG), diethylene glycol (DEG), ethanol, trihydroxymethyl aminomethane (Tris), sodium dodecyl sulfonate (SDS), and acetic acid (HAc) were provided by Xi’an Chemicals Ltd. Bovine serum albumin (BSA; pI = 4.9, M_w = 66.0 kDa), bovine hemoglobin (BHb; pI = 6.9, M_w = 64.0 kDa), bovine pancreas ribonuclease A (RNase A; pI = 9.4, M_w = 13.7 kDa), hemoglobin (Hb; pI = 6.8, M_w = 65.0 kDa), transferrin (transferrin; pI = 5.5, M_w = 77.0 kDa), immunoglobulin G (IgG; pI = 8.0, M_w = 150 kDa), and lysozyme (Lyz; pI = 11.2, M_w = 13.4 kDa) were obtained from Sigma. The highly purified water (18.0 MΩ cm⁻¹) was prepared with a WaterPro water system (Axlwater Corporation, TY10AXLC1805-2, China) and used throughout the experiments. The bovine blood sample was kindly gifted from the local market. All reagents were of at least analytical grade and used without further treatment.

2.2. Preparation of Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs

The amino-modified Fe₃O₄ nanoparticles (denoted as Fe₃O₄@NH₂) were synthesized according to a previous method³⁰ with some modifications. FeCl₃·6H₂O (1.0 g), NaOAc (3.5 g), and DAH (6.0 g) were dissolved in a mixture of EG (10 mL) and DEG (10 mL) in a Teflon-lined stainless-steel autoclave, which was sealed and heated at 200 °C. After reaction for 8 h, the autoclave was cooled to room temperature. The resultant black products were washed with highly purified water and ethanol to remove the solvent and the unreacted DAH, and then dried in a vacuum for further use.

The optimal experimental conditions for the polymerization of the BSA surface-imprinted magnetic nanoparticles were as follows: (1) the mass of immobilized template BSA (10 mg, 20 mg, 30 mg, 40 mg, and 50 mg, respectively) was dissolved in 10 mM Tris–HCl buffer (pH = 8.5, 30 mL) in a three-neck round-bottom flask, and mixed with Fe₃O₄@NH₂ (200 mg). The mixture was stirred for 30 min to obtain the Fe₃O₄@NH₂–BSA complex. The complex was collected by an external magnetic field and left in the bottom of the flask, and the concentration of BSA in the supernatant was detected by a UV-vis spectrophotometer at about 280 nm. (2) The amount of the functional monomer DA (50 mg, 75 mg, 100 mg, 125 mg, and 150 mg, respectively) and 30 mL of Tris–HCl buffer (pH = 8.5, 10 mM) were added to the above flask, and stirred for 5 h at room temperature. The obtained products were rinsed with highly purified water until the supernatant was clear, and then washed with SDS–HAc (2%) to remove the template protein until no adsorption was detected by the UV-vis spectrophotometer. The resulting imprinted polymers (denoted as Fe₃O₄@BSA–MIPs) were collected by an external magnetic field, washed with highly purified water, and then dried under vacuum. Non-imprinted polymers (denoted as Fe₃O₄@NIPs) were prepared following the same procedure in the absence of the template protein BSA. Furthermore, the immobilization time of the BSA and the polymerization time of DA were also investigated.

Moreover, to testify the efficiency of Fe₃O₄@NH₂ as the support, bare magnetic molecularly imprinted and non-imprinted polymers (denoted as B-Fe₃O₄@BSA–MIPs and B-Fe₃O₄@NIPs) were also prepared by the same way as that of Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs except for employing bare magnetic nanoparticles as the support.

2.3. Binding experiments

To investigate the adsorption kinetics of Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs, 10 mg of adsorbent was suspended in 10 mL of BSA Tris–HCl buffer (pH = 7.0, 10 mM) solution at a concentration of 0.30 mg mL⁻¹. The mixture was incubated at regular time intervals from 5 min to 60 min at room temperature, and the polymers were magnetically separated from the solution. Then, the concentration of BSA in the supernatant was measured by UV-vis spectrophotometry.

The adsorption capacity (Q) of the template protein or the competitive protein bound to the imprinted polymers is calculated as follows:

where C₀ and C_e (mg mL⁻¹) represent the initial and equilibrium solution concentrations of the adsorbed protein, V (mL) represents the volume of the solution, and m (mg) is the weight of the imprinted polymers.

The isothermal adsorption experiment was operated through changing the concentration of BSA from 0.050 to 0.50 mg mL⁻¹ in Tris–HCl buffer (pH = 7.0, 10 mM) while employing 10 mg of Fe₃O₄@BSA–MIPs or Fe₃O₄@NIPs and shaking for 20 min at room temperature. Then, the adsorbent was isolated by a magnet and the residue BSA in the supernatant was determined by UV-vis spectrophotometry.

To evaluate the specific adsorption capability of the as-prepared polymers, 10 mg of Fe₃O₄@BSA–MIPs or Fe₃O₄@NIPs was incubated in 3 mL of Tris–HCl (10 mM, pH = 7.0) solution of BSA, BHb, transferrin, and IgG at a concentration of 0.30 mg mL⁻¹ respectively at room temperature for 20 min, then the separation and determination procedures were conducted as described earlier in the adsorption kinetics experiments.

The imprinting factor (IF) and selectivity coefficient (SC) are usually used to determine the selectivity properties of the imprinted polymers towards the template protein and the competitive protein. The IF and SC are calculated from the following equations:

where Q_MIP and Q_NIP (mg g⁻¹) represent the adsorption capacity of the protein for Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs, and IF_TEM and IF_COM are the imprinting factors of the template protein and the competitive protein.

To evaluate if the amino groups on Fe₃O₄@NH₂ could enhance the adsorption performance of the imprinted nanoparticles, we compared the binding capabilities of Fe₃O₄@BSA–MIPs, Fe₃O₄@NIPs, B-Fe₃O₄@BSA–MIPs, and B-Fe₃O₄@NIPs. In detail, 10 mg of the above four polymers was incubated with 10 mL of the BSA solution at a concentration of 0.30 mg mL⁻¹ (pH = 7.0, 10 mM Tris–HCl) at room temperature for 20 min. Then, the extraction and detection procedures were conducted as described earlier in the adsorption experiments.

2.4. Reusability of the resultant imprinted nanomaterials

To estimate the reusability of Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs, the BSA adsorption–regeneration procedure was repeated 6 times using the same polymers. Briefly, 10 mg of the polymer was added to 10 mL of the BSA solution at a concentration of 0.30 mg mL⁻¹ (pH = 7.0, 10 mM Tris–HCl) and incubated at room temperature for 20 min. Then, Fe₃O₄@BSA–MIPs or Fe₃O₄@NIPs were removed by a magnet and the bound amount of BSA was quantified by UV-vis spectrophotometry. The reused polymers were eluted with SDS–HAc (2%) for 3 h to ensure complete removal of the adsorbed BSA. The recovered products were then reused for the adsorption of BSA for another 5 cycles, and for every cycle the supernatant was collected and determined by UV-vis spectrophotometry at about 280 nm.

2.5. Real sample analysis

10 mg of Fe₃O₄@BSA–MIPs was merged with 10 mL of the standard protein mixture (containing 0.30 mg mL⁻¹ BSA and 0.30 mg mL⁻¹ BHb) and the bovine blood sample diluted 150-fold with Tris–HCl buffer solution (pH = 7.0, 10 mM), respectively. After incubation for 20 min under gentle shaking, SDS–HAc (2%) was employed to elute the adsorbed protein for 3 h. The diluted, adsorbed, and eluted samples were used for SDS-PAGE analysis.

2.6. Versatility investigation

The versatility of this approach was evaluated by using another three magnetic imprinted polymers generated in the same way as that of Fe₃O₄@BSA–MIPs except for adopting BHb, Hb, RNase A, and Lyz as the template protein, which were denoted as Fe₃O₄@BHb–MIPs, Fe₃O₄@Hb–MIPs, Fe₃O₄@RNase A–MIPs, and Fe₃O₄@Lyz–MIPs, respectively. The full cross-reaction characterizations of the four protein-imprinted nanomaterials along with the non-imprinted Fe₃O₄@NIPs were investigated by adsorbing the corresponding template protein and the three other proteins, namely, 10 mg of the polymer was incubated with 10 mL of the Tris–HCl buffer (pH = 7.0, 10 mM) solution of BSA, BHb, Hb, RNase A, and Lyz at a concentration of 0.30 mg mL⁻¹ at room temperature for 20 min. Then, the extraction and detection procedures were conducted as described earlier in the binding experiments.

2.7. Characterization

The morphology of the obtained nanomaterials was studied using a Tecnai G2 T2 S-TWIN transmission electron microscope. Fourier transform infrared (FT-IR) spectra were obtained via a Nicolet AVATAR 330 FT-IR spectrophotometer. The identification of the crystalline phase was carried out by a Rigaku D/max/2500v/pc (Japan) X-ray diffractometer with Cu Kα radiation. The magnetic properties were analyzed with a vibrating sample magnetometer (VSM) (LDJ 9600-1, USA). The adsorption data were measured using a UV-2450 spectrophotometer (Shimadzu, Japan). Electrophoretic analysis of the protein samples was performed using regular SDS-PAGE (Bio-Rad, Hercules, CA) with 10% running and 5% stacking gels. The proteins were stained with Coomassie Brilliant Blue R-250.

3. Results and discussion

3.1. Preparation of Fe₃O₄@BSA–MIPs

The procedure for the preparation of Fe₃O₄@BSA–MIPs, combining the merits of surface imprinting technology, a two-step immobilized template strategy, functionalized magnetic nanomaterials, and the hydrophilicity of DA, is displayed in Fig. 1. First, Fe₃O₄@NH₂, the surface of which was full of numerous amino groups, was synthesized through a modified one-step solvothermal method, using FeCl₃·6H₂O as the single iron source and DAH as the ligand. Next, the template protein BSA was immobilized on the surface of Fe₃O₄@NH₂ through multi-hydrogen bonding interactions between the amino groups of Fe₃O₄@NH₂ and the amino acids on BSA to form the Fe₃O₄@NH₂–template complex. Then, a thin adherent polydopamine (PDA) shell was deposited on the surface of Fe₃O₄@NH₂ to further immobilize the template protein by using DA as the monomer and cross-linker. Notably, the dopamine units possess amino, hydroxyl, and phenyl groups which can interact with BSA through multi-hydrogen bonds and hydrophobic effects. The two immobilized template processes can orientate template proteins orderly and keep the imprinted sites uniform for improving the imprinting effect.³¹ Finally, after washing with the 2% SDS–HAc solution, Fe₃O₄@BSA–MIPs with imprinted recognition sites complementary to BSA in shape, size, and functional group orientation were obtained.

Fig. 1 The synthetic procedure of Fe₃O₄@BSA–MIPs.

3.2. Optimization of the polymerization conditions

When we carried out the optimal experimental conditions, the effect of different mass of immobilization of BSA onto Fe₃O₄@NH₂ in the range of 10–50 mg was investigated. As shown in Table S1,† the Q_IM increased with increasing Q_AD from 10 mg to 30 mg, indicating that a higher Q_AD might make Fe₃O₄@NH₂ interact adequately with BSA through forming multi-hydrogen bonds. However, the Q_IM remained almost unchanged with increasing Q_AD from 30 mg to 50 mg. A possible reason was that the amino groups on Fe₃O₄@NH₂ had been saturated with BSA. Excessive Q_AD could not increase the immobilization. Thus, the Q_AD of 30 mg was chosen in this work. Furthermore, we also evaluated the immobilization time of BSA by varying from 10 min to 50 min. The results are presented in Table S2,† from which we found that the best immobilization time was 30 min.

To optimize the recognition performance of Fe₃O₄@BSA–MIPs, the thickness of the PDA shell, which is controlled by the amount of DA, was adjusted from 50 mg to 150 mg. The results of the adsorption capacity (Q) and imprinting factor (IF) are presented in Fig. 2A. Both Q and IF increased with an increasing amount of DA from 50 mg to 100 mg, indicating that enough functional monomers could react adequately with template proteins and create enough recognition cavities in the network of polymers to enhance the imprinting ability of Fe₃O₄@BSA–MIPs towards BSA. However, a decrease of Q and IF was observed when the amount of DA was further increased. Too thick PDA imprinted shells may make it hard for the protein to access the recognition cavities. Therefore, 100 mg of DA was chosen for the polymerization of Fe₃O₄@BSA–MIPs in accordance with the experimental results.

Fig. 2 Effect of the amount (A) and polymerization time (B) of DA on the imprinting performance of Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs.

The polymerization time of DA was varied from 3 h to 7 h to seek the shortest experimental period and the best binding performance. The results are displayed in Fig. 2B. It is indicated that the optimal recognition sites and the most suitable shell thickness were formed when the polymerization time was 5 h. The polymerization degree was incomplete when the polymerization time was less than 5 h, because the thin PDA shell could not immobilize much template and created less imprinted cavities. Conversely, too many monomers self-polymerized and blocked the recognition sites, impacting the imprinting effect of the polymers when the polymerization time was over 5 h. Therefore, 5 h was chosen as the polymerization time in this work.

3.3. Characterization of the obtained nanomaterials

The morphological structure and size of Fe₃O₄@NH₂ and Fe₃O₄@BSA–MIPs were characterized by TEM. As can be seen from Fig. 3, the obtained nanomaterials possessed a spherical structure with a narrow particle size distribution. Through statistical analysis of the particles, the average diameter of Fe₃O₄@NH₂ (Fig. 3A) was about 60 nm. It was found that Fe₃O₄@BSA–MIPs with a distinguishable core–shell structure were successfully prepared after a two-step template immobilization process, and whose diameter increased to approximately 80 nm (Fig. 3B), indicating that the thickness of the PDA imprinting shell was just about 10 nm. The thickness was suitable for the three-dimensional structure of the BSA,³² making the mass transport between the solution and the surface of Fe₃O₄@BSA–MIPs easier.

Fig. 3 TEM images of Fe₃O₄@NH₂ (A) and Fe₃O₄@BSA–MIPs (B); the size distribution histograms of Fe₃O₄@NH₂ (C) and Fe₃O₄@BSA–MIPs (D).

The chemical groups of the pristine Fe₃O₄@NH₂ and Fe₃O₄@BSA–MIPs (Fig. 4) were characterized by FT-IR spectroscopy. The strong peak at 575 cm⁻¹ observed in both samples (curves a and b) was attributed to the stretch of the Fe–O vibration. The peaks at around 1080 cm⁻¹ and 1640 cm⁻¹, corresponding to the C–N stretch and the bending vibration of N–H²⁴ respectively, indicate that the amino functionalization was on the surface of Fe₃O₄@NH₂ (curve a). When compared with Fe₃O₄@NH₂, the new appearance of typical peaks in the spectra of Fe₃O₄@BSA–MIPs (curve b), such as phenylic C=C stretching at 1500 cm⁻¹, the enhanced adsorption peak intensity of C–N bond at 1640 cm⁻¹, and the lessened characteristic peak of the Fe–O bond at 575 cm⁻¹ should be attributed to the coating of the PDA layer on the magnetic nanoparticles.

Fig. 4 FT-IR spectra of Fe₃O₄@NH₂ (a) and Fe₃O₄@BSA–MIPs (b).

The crystalline structures of Fe₃O₄@NH₂ and Fe₃O₄@BSA–MIPs were characterized by XRD in the 2θ range of 20–70° (Fig. 5). Six characteristic peaks of Fe₃O₄ (2θ = 30.38°, 35.58°, 43.14°, 53.48°, 57.08°, and 62.66°) were observed in both samples. The peak positions at the corresponding 2θ values were indexed as (220), (311), (400), (422), (511), and (440), respectively, which matched well with the database of magnetite in the JCPDS-International Center for Diffraction Data (JCPDS card: 19-629). The distinctive diffraction peaks of Fe₃O₄@BSA–MIPs were in agreement with those of Fe₃O₄, illustrating that the obtained imprinted polymers containing Fe₃O₄ with a spinel structure and the phase of Fe₃O₄ did not change during the synthetic process.³³

Fig. 5 XRD patterns of Fe₃O₄@NH₂ (a) and Fe₃O₄@BSA–MIPs (b).

The magnetic properties of Fe₃O₄@NH₂ and core–shell Fe₃O₄@BSA–MIPs were studied using a vibrating sample magnetometer at room temperature. Fig. 6 illustrates the plots of magnetization versus magnetic field (M–H loop) of Fe₃O₄@NH₂ (curve a) and Fe₃O₄@BSA–MIPs (curve b) from which we found there was no remanence and coercivity, demonstrating that both polymers were superparamagnetic. It is noteworthy that the saturation magnetization of Fe₃O₄@BSA–MIPs (40.7 emu g⁻¹) was lower than that of Fe₃O₄@NH₂ (53.5 emu g⁻¹), which was ascribed to the shielding effect of the PDA shell on the surface of Fe₃O₄@NH₂. The experimental results show that the obtained imprinted polymers possess sufficient magnetic force to be quickly separated from the solution under an external magnet within three seconds.

Fig. 6 Magnetization curves of Fe₃O₄@NH₂ (a) and Fe₃O₄@BSA–MIPs (b).

3.4. Adsorption performance

3.4.1. Adsorption kinetics. Adsorption kinetics of BSA onto Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs were investigated, and the curves of binding capacity versus time are presented in Fig. 7A. It was observed that the increase of the adsorption capacity of BSA onto Fe₃O₄@BSA–MIPs was quite fast within the first 10 min, and then almost reached equilibrium after 20 min. The adsorption equilibrium time of the obtained nanomaterials was shorter than that of some other surface imprinting polymers for BSA.^34,35 The rapid adsorption rate towards BSA was due to the large surface-to-volume ratios of the carriers and the thin imprinting shells, which made BSA approach the binding sites easily with lower mass-transfer resistance. It was also noted that Fe₃O₄@BSA–MIPs had a much higher adsorption capacity than that of Fe₃O₄@NIPs, suggesting the favorable recognition performance of the resultant imprinted nanomaterials.

Fig. 7 Adsorption kinetics (A) and isotherms (B) of BSA onto Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs.

To further investigate the kinetic mechanisms of adsorption, the obtained data were calculated by the second-order rate equation that can be expressed as follows:

where Q_e and Q_t (mg g⁻¹) are the amounts of BSA adsorbed onto Fe₃O₄@BSA–MIPs or Fe₃O₄@NIPs at the equilibrium and at time t (min), respectively; K (g mg⁻¹ min⁻¹) represents the rate constant of the second-order adsorption; v₀ (mg g⁻¹ min⁻¹) is the initial adsorption rate.

As shown in Table 1, the correlation coefficients (r) of Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs are 0.9998 and 0.9964, indicating that the second-order model fitted the binding data well. The values of v₀ implied that the initial adsorption rate of Fe₃O₄@BSA–MIPs was much faster than that of Fe₃O₄@NIPs, which was in keeping with Fig. 7A. The results also illustrated that the adsorption capacity was proportional to the number of active recognition sites on the surface of the imprinted nanomaterials, and the chemical adsorption was the rate-limiting step in the recognition process of the resulting polymers.³⁶

Table 1 Equations and parameters of the adsorption kinetics and isotherms of Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs

Model	Equations and parameters	Fe3O4@BSA–MIPs	Fe3O4@NIPs
Second-order kinetics	Equation	t/Qt = 0.0082 + 0.0091t	t/Qt = 0.0989 + 0.0292t
	Qe (mg g^−1)	109.9	34.13
	K (g mg^−1 min^−1)	0.0101	0.0087
	v0 (mg g^−1 min^−1)	122.0	10.11
	r	0.9998	0.9964
Langmuir isotherm	Equation	Ce/Q = 2.1162 × 10^−4 + 0.0085Ce	Ce/Q = 0.0019 + 0.0254Ce
	Qmax (mg g^−1)	117.6	39.37
	KL (mL mg^−1)	40.18	13.37
	r	0.9969	0.9954
Freundlich isotherm	Equation	log Q = 2.2525 + 0.3381 log Ce	log Q = 1.7022 + 0.3843 log Ce
	KF (mg g^−1)	178.9	50.37
	n	0.3381	0.3843
	r	0.9811	0.9339

---

3.4.2. Adsorption isotherms. The adsorption isothermal experiments were performed at different initial concentrations of BSA, ranging from 0.050 to 0.50 mg mL⁻¹ (Fig. 7B). As the isothermal curves show, the amount of BSA bound to Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs increased quickly along with increasing the initial concentration of BSA and then reached saturation when the concentration was above 0.30 mg mL⁻¹. The experimental maximum adsorption capacity of BSA onto Fe₃O₄@BSA–MIPs (107.8 mg g⁻¹) was 3.35 times higher than that of Fe₃O₄@NIPs (32.15 mg g⁻¹). A large amount of specific recognition cavities on Fe₃O₄@BSA–MIPs results in a higher affinity towards BSA than that of non-imprinted nanomaterials. Furthermore, compared with some other works for surface imprinting BSA using Fe₃O₄ as the support, the binding amount in this work was much higher,^22,32 confirming that DA, possessing favorable hydrophilicity and abundant functional groups, could create more imprinted sites which fitted well with the template protein in terms of water-solubility, chemical effect, and steric structure.

The saturation binding data was further processed by the Freundlich and Langmuir isothermal models to estimate the adsorption properties of the resultant nanomaterials. The two models are expressed as follows:

log Q = log K_F + n log C_e (5)

where Q (mg g⁻¹) is the amount of BSA bound to Fe₃O₄@BSA–MIPs at equilibrium, Q_max (mg g⁻¹) is the apparent maximum adsorption capacity, C_e (mg mL⁻¹) is the free analytical concentration at equilibrium, K_L and K_F (mL mg⁻¹) are the Langmuir and Freundlich constants respectively, and n is the Freundlich exponent which represents the heterogeneity of the system. The values of K_L, Q_max and n, K_F can be calculated from the slope and intercept of the linear plot in C_e/Q versus C_e and log Q versus log C_e, respectively.

The Langmuir isotherm model assumes that the adsorption takes place at specific homogeneous sites as well as monolayer sorption, and each site can bind only one molecule.³⁷ While the Freundlich isotherm model is applied to multi-layer adsorption and unfavorable adsorption on heterogeneous surfaces.³⁸ Through comparison of the r values displayed in Table 1, we can conclude that the experimental data are better fitted with the Langmuir isotherm model (r > 0.99) than the Freundlich model (r < 0.94). The maximum adsorption capacities obtained from the experiment are also close to the apparent maximum adsorption capacities (117.6 mg g⁻¹ for Fe₃O₄@BSA–MIPs and 39.37 mg g⁻¹ for Fe₃O₄@NIPs) calculated using the Langmuir isotherm model. In light of the results, the binding of BSA onto the resultant nanomaterials may be all monolayer adsorption, and no further binding takes place at the binding site once a template molecule occupies this site.

3.4.3. Comparison of two kinds of imprinted nanoparticles. The adsorption performances of B-Fe₃O₄@BSA–MIPs and B-Fe₃O₄@NIPs were also investigated. The results of the comparison of two kinds of imprinted nanomaterials prepared adopting different supports are presented in Table 2. It is obvious that the binding amount of B-Fe₃O₄@NIPs was close to that of Fe₃O₄@NIPs. However, the adsorption capacities showed a great difference between Fe₃O₄@BSA–MIPs and B-Fe₃O₄@BSA–MIPs. Fe₃O₄@BSA–MIPs had a much higher binding amount and imprinting factor, twice as much as those of B-Fe₃O₄@BSA–MIPs, demonstrating that the amino groups played an important role in the process of protein recognition. Because Fe₃O₄@BSA–MIPs were obtained by the “template immobilization strategy”, the –NH₂ on the surface of Fe₃O₄@NH₂ could orientate the template proteins in an orderly fashion and create more homogenous binding sites in the polymer matrix.

Table 2 Comparison of two kinds of imprinted nanoparticles^a

Polymers	Q (mg g^−1)	IF
a In this experiment, 10 mg of Fe3O4@BSA–MIPs, B-Fe3O4@BSA–MIPs, Fe3O4@NIPs, and B-Fe3O4@NIPs was incubated in 10 mL of a Tris–HCl (10 mM, pH = 7.0) solution of BSA at a concentration of 0.30 mg mL^−1 at room temperature for 20 min. (n = 5).
Fe3O4@BSA–MIPs	107.8	—
Fe3O4@NIPs	32.15	3.35
B-Fe3O4@BSA–MIPs	53.19	—
B-Fe3O4@NIPs	31.84	1.67

---

3.4.4. Specificity. BHb, IgG, and transferrin were chosen as competitive proteins to illustrate the recognition specificity of the imprinted nanomaterials for template BSA due to the following reasons. (1) BHb is another high-abundant protein in bovine blood serum and has a similar molecular volume and weight to BSA; (2) IgG and transferrin are proteins which have relatively lower quantities in the blood. The adsorption of Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs to BSA, BHb, IgG, and transferrin with a concentration of 0.30 mg mL⁻¹ in 3 mL of Tris–HCl buffer were examined, respectively (Fig. 8A and Table S3†). Obviously, the imprinted polymers exhibited a much higher binding amount for template protein BSA than those of BHb, IgG, and transferrin. However, the adsorption capacity of BSA for Fe₃O₄@NIPs was quite close to those of the three other competitive proteins. The difference suggested that the specific recognition cavities complementary in shape, size, and functional groups with the template protein were formed in the thin imprinting PDA shells of Fe₃O₄@BSA–MIPs. Moreover, the imprinting factor of BSA (3.35) was higher than those of BHb (1.97), IgG (1.19), and transferrin (1.21). These results further proved the excellent imprinting efficiency of the present method.

Fig. 8 The specific adsorption capability (A) and reusability (B) of Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs.

3.5. Reusability

The property of reusability is significant for the actual application of polymers. To test the stability of the prepared magnetic nanomaterials, the adsorption–regeneration cycle was repeated six times using the same Fe₃O₄@BSA–MIPs and Fe₃O₄@NIPs. As shown in Fig. 8B, the adsorption capacity of Fe₃O₄@BSA–MIPs was reduced by only 7.1% after six cycles, which might be because some recognition cavities of the imprinted polymers were clogged after regeneration or destroyed after washing, and thus, they were no longer fit for the template molecules. Whereas, the affinity of Fe₃O₄@NIPs remained almost unchanged, because their recognition was nonspecific and the effect of washing could be ignored. After six cycles of adsorption–regeneration, the adsorption performance of the imprinted and non-imprinted magnetic nanomaterials prepared in this work were stable, which is an outstanding advantage for practical application.

3.6. Application

Serum albumin accounts for approximately 50% of total blood plasma proteins,³⁹ which is considered to act as a shield in the analysis of biomarkers that are commonly at lower concentration levels in plasma samples. Therefore, we chose fresh bovine blood diluted 150-fold with Tris–HCl (pH = 7.0, 10 mM) as the real sample to further evaluate the practicability of Fe₃O₄@BSA–MIPs. The SDS-PAGE analysis is illustrated in Fig. 9. There were two bands in lane 1, indicating the mixture of BSA and BHb. It was found from lane 2 that the band of BSA faded after treatment with Fe₃O₄@BSA–MIPs, while the band of BHb had little change. The BSA band reappeared in lane 3 after elution with 2% SDS–HAc solution, revealing that the BSA was selectively captured by Fe₃O₄@BSA–MIPs. Lane 5 displayed the supernatant of the fresh bovine blood diluted 150-fold (lane 4) after treatment with Fe₃O₄@BSA–MIPs, in which BSA almost disappeared while the other bands were retained, suggesting that Fe₃O₄@BSA–MIPs had specific recognition for BSA in the bovine blood sample and less co-adsorption with others. After elution with 2% SDS–HAc solution, only the band for BSA was observed (lane 6), indicating that the BSA was selectively isolated and recovered from the real sample after elution. The above results confirmed the specificity and practicability of Fe₃O₄@BSA–MIPs for the separation of BSA.

Fig. 9 SDS-PAGE analysis to evaluate the applicability of Fe₃O₄@BSA–MIPs towards BSA. Lane 1, 0.30 mg mL⁻¹ BSA and BHb binary standard solution; lane 2, remaining BSA and BHb solution after adsorption by Fe₃O₄@BSA–MIPs; lane 3, the BSA and BHb mixture eluted by 2% SDS–HAc; lane 4, bovine blood diluted 150-fold; lane 5, remaining bovine blood after adsorption by Fe₃O₄@BSA–MIPs; lane 6, the absorbed bovine blood eluted by 2% SDS–HAc.

3.7. Method validation

The generality of the proposed imprinting approach was further investigated through the full cross-selectivity test of four protein-imprinted nanomaterials along with non-imprinted Fe₃O₄@NIPs, and the results are summarized in Table 3. It was clearly observed that the adsorption capacity, imprinting factor, and selectivity coefficient of different imprinted polymers to corresponding template proteins were all higher than those of the other three proteins, demonstrating that the developed method was valid to imprint different kinds of proteins with different isoelectric points. It is worth mentioning that the BHb-imprinted nanomaterials used in this work have a larger adsorption capacity (181.1 mg g⁻¹) towards BHb and a higher imprinting factor (4.96) than those of our previous work (37.58 mg g⁻¹, 3.60).²⁴ The greatest difference lies in that the latter adopted two types of silane coupling agent as monomers while the former employed DA as monomer. It is also worth noting that the Hb-imprinted nanoparticles used in this work have a considerably larger adsorption capacity (187.9 mg g⁻¹) towards Hb and a relatively higher imprinting factor (5.48) than those of Zhou’s work (17.50 mg g⁻¹, 5.01).²⁰ The main contrast was that the latter employed bare Fe₃O₄ nanoparticles as the support while the former used Fe₃O₄@NH₂ instead. Therefore, NH₂-modification on the surface of Fe₃O₄ nanoparticles enhances the performance of the imprinted nanomaterials.

Table 3 Cross-selectivities of BSA, BHb, Hb, RNase A, and Lyz adsorbing by Fe₃O₄@BSA–MIPs, Fe₃O₄@BHb–MIPs, Fe₃O₄@Hb–MIPs, Fe₃O₄@RNase A–MIPs, Fe₃O₄@Lyz–MIPs, and Fe₃O₄@NIPs^a

Polymers	Proteins	Q (mg g^−1)	IF	SC
a In these experiments, 10 mg of Fe3O4@BSA–MIPs, Fe3O4@BHb–MIPs, Fe3O4@Hb–MIPs, Fe3O4@RNase A–MIPs, Fe3O4@Lyz–MIPs, and Fe3O4@NIPs was incubated in 10 mL of a Tris–HCl (10 mM, pH = 7.0) solution of BSA, BHb, Hb, RNase A, and Lyz at a concentration of 0.30 mg mL^−1 respectively at room temperature for 20 min (n = 5).
Fe3O4@BSA–MIPs	BSA	107.8 ± 2.7	3.35 ± 0.02	—
	BHb	72.13 ± 1.8	1.97 ± 0.08	1.70 ± 0.07
	Hb	70.15 ± 1.9	2.05 ± 0.03	1.63 ± 0.04
	RNase A	82.61 ± 2.2	2.01 ± 0.07	1.67 ± 0.11
	Lyz	99.1 ± 3.4	2.10 ± 0.04	1.59 ± 0.09
Fe3O4@BHb–MIPs	BSA	73.89 ± 1.7	2.30 ± 0.09	2.16 ± 0.08
	BHb	181.1 ± 3.6	4.96 ± 0.23	—
	Hb	146.3 ± 3.8	4.27 ± 0.26	1.16 ± 0.02
	RNase A	93.82 ± 2.9	2.28 ± 0.07	2.18 ± 0.25
	Lyz	105.7 ± 3.1	2.24 ± 0.05	2.21 ± 0.32
Fe3O4@Hb–MIPs	BSA	71.41 ± 2.0	2.22 ± 0.04	2.47 ± 0.16
	BHb	159.2 ± 3.7	4.36 ± 0.06	1.26 ± 0.35
	Hb	187.9 ± 3.5	5.48 ± 0.03	—
	RNase A	91.85 ± 2.4	2.24 ± 0.02	2.45 ± 0.26
	Lyz	109.3 ± 2.9	2.31 ± 0.09	2.37 ± 0.12
Fe3O4@RNase A–MIPs	BSA	47.61 ± 2.2	1.48 ± 0.06	4.06 ± 0.12
	BHb	66.22 ± 3.6	1.81 ± 0.03	3.32 ± 0.29
	Hb	63.01 ± 3.9	1.84 ± 0.05	3.27 ± 0.37
	RNase A	246.8 ± 8.3	6.01 ± 0.26	—
	Lyz	136.5 ± 2.7	2.89 ± 0.04	2.08 ± 0.08
Fe3O4@Lyz–MIPs	BSA	43.94 ± 1.4	1.37 ± 0.08	4.50 ± 0.13
	BHb	62.73 ± 2.5	1.72 ± 0.03	3.58 ± 0.46
	Hb	59.42 ± 1.9	1.73 ± 0.02	3.56 ± 0.39
	RNase A	119.6 ± 3.6	2.91 ± 0.05	2.12 ± 0.25
	Lyz	291.1 ± 11.2	6.16 ± 0.31	—
Fe3O4@NIP	BSA	32.15 ± 0.94	—	—
	BHb	36.54 ± 1.2	—	—
	Hb	34.26 ± 1.4	—	—
	RNase A	41.06 ± 2.9	—	—
	Lyz	47.26 ± 1.3	—	—

---

4. Conclusion

In this work, a facile, general, and economical approach was successfully developed to imprint proteins on the surface of Fe₃O₄@NH₂ using dopamine as monomer. Because of combining the advantages of surface imprinting technology, immobilized template strategy, magnetic separation, and PDA, these resulting imprinted nanomaterials exhibited fast kinetics, a large adsorption capacity, high selectivity, and satisfactory reusability. Furthermore, the favorable versatility for proteins with different isoelectric points, and the successful application in the specific capture of the target protein from a real sample of bovine blood and good recovery after a reasonably mild elution, indicate that the proposed method could be an alternative solution for the selective removal of high-abundant proteins from complex biological samples.

Acknowledgements

The authors are grateful for financial support from the National Natural Science Foundation of China (no. 21305107), the Fundamental Research Funds for the Central Universities (no. 08143081, 08142034), and China Postdoctoral Science Foundation (no. 2014M562388).

References

1. F. R. Wang, C. Chmil, F. Pierce, K. Ganapathy, B. B. Gump, J. A. MacKenzie, Y. Mechref and K. Bendinskas, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2013, 934, 26–33.
2. J. L. Cao, X. H. Zhang, X. W. He, L. X. Chen and Y. K. Zhang, J. Mater. Chem. B, 2013, 1, 3625–3632.
3. M. Javanmard, S. Emaminejad, C. Gupta, J. Provine, R. W. Davis and R. T. Howe, Sens. Actuators, B, 2014, 193, 918–924.
4. M. Javanmard, S. Emaminejad, R. W. Dutton and R. W. Davis, Anal. Chem., 2012, 84, 1432–1438.
5. D. N. Gunasena and Z. El Rassi, J. Sep. Sci., 2011, 34, 2097–2105.
6. R. L. Gundry, M. Y. White, J. Nogee, I. Tchernyshyov and J. E. Van Eyk, Proteomics, 2009, 9, 2021–2028.
7. B. Li, J. J. Xu, A. J. Hall, K. Haupt and B. Tse Sum Bui, J. Mol. Recognit., 2014, 27, 559–565.
8. A. Mirmohseni, M. Shojaei and R. Pourata, RSC Adv., 2014, 4, 20177–20184.
9. E. Yildirim, E. Turan and T. Caykara, J. Mater. Chem., 2012, 22, 636–642.
10. K. El Kirat, M. Bartkowski and K. Haupt, Biosens. Bioelectron., 2009, 24, 2618–2624.
11. J. Bognár, J. Szücs, Z. Dorkó, V. Horváth and R. E. Gyurcsányi, Adv. Funct. Mater., 2013, 23, 4703–4709.
12. T. Wangchareansak, C. Sangma, K. Choowongkomon, F. Dickert and P. Lieberzeit, Anal. Bioanal. Chem., 2011, 400, 2499–2506.
13. M. Bosserdt, N. Gajovic-Eichelman and F. W. Scheller, Anal. Bioanal. Chem., 2013, 405, 6437–6444.
14. A. M. Bossi, P. S. Sharma, L. Montana, G. Zoccatelli, O. Laub and R. Levi, Anal. Chem., 2012, 84, 4036–4041.
15. V. Saumya, K. P. Prathish and T. P. Rao, Talanta, 2011, 85, 1056–1062.
16. L. Uzun, R. Uzek, S. Şenel, R. Say and A. Denizli, Mater. Sci. Eng., C, 2013, 33, 3432–3439.
17. F. Kartal and A. Denizli, J. Sep. Sci., 2014, 37, 2077–2086.
18. E. Moczko, A. Guerreiro, E. Piletska and S. Piletsky, Langmuir, 2013, 29, 9891–9896.
19. R. X. Gao, X. R. Mu, J. J. Zhang and Y. H. Tang, J. Mater. Chem. B, 2014, 2, 783–792.
20. W. H. Zhou, C. H. Lu, X. C. Guo, F. R. Chen, H. H. Yang and X. R. Wang, J. Mater. Chem., 2010, 20, 880–883.
21. W. W. Huang, X. Yang, S. Zhao, M. Zhang, X. L. Hu, J. Wang and H. T. Zhao, Analyst, 2013, 138, 6653–6661.
22. X. J. Li, B. L. Zhang, L. Tian, W. Li, T.J. Xin, H. P. Zhang and Q. Y. Zhang, Sens. Actuators, B, 2014, 196, 265–271.
23. T. Jing, H. R. Du, Q. Dai, H. Xia, J. W. Niu, Q. L. Hao, S. R. Mei and Y. K. Zhou, Biosens. Bioelectron., 2010, 26, 301–306.
24. R. X. Gao, X. R. Mu, Y. Hao, L. L. Zhang, J. J. Zhang and Y. H. Tang, J. Mater. Chem. B, 2014, 2, 1733–1741.
25. Y. Zhang, K. Panneerselvam, R. Ogaki, L. Hosta-Rigau, R. V. D. Westen, B. E. B. Jensen, B. M. Teo, M. F. Zhu and B. Städler, Langmuir, 2013, 29, 10213–10222.
26. A. Nematollahzadeh, A. Shojaei, M. J. Abdekhodaie and B. Sellergren, J. Colloid Interface Sci., 2013, 404, 117–126.
27. A. Tretjakov, V. Syritski, J. Reut, R. Boroznjak, O. Volobujeva and A. Öpik, Microchim. Acta, 2013, 180, 1433–1442.
28. Z. W. Xia, Z. A. Lin, Y. Xiao, L. Wang, J. N. Zheng, H. H. Yang and G. N. Chen, Biosens. Bioelectron., 2013, 47, 120–126.
29. R. Liu, M. Sha, S. S. Jiang, J. Luo and X. Y. Liu, Talanta, 2014, 120, 76–83.
30. L. Y. Wang, J. Bao, L. Wang, F. Zhang and Y. D. Li, Chem.–Eur. J., 2006, 12, 6341–6347.
31. L. J. Liu, J. J. Zheng, G. J. Fang and W. H. Xie, Anal. Chim. Acta, 2012, 726, 85–92.
32. X. J. Li, B. L. Zhang, W. Li, X. F. Lei, X. L. Fan, L. Tian, H. P. Zhang and Q. Y. Zhang, Biosens. Bioelectron., 2014, 51, 261–267.
33. F. X. Chen, S. L. Xie, J. H. Zhang and R. Liu, Mater. Lett., 2013, 112, 177–179.
34. D. R. Kryscio and N. A. Peppas, Anal. Chim. Acta, 2012, 718, 109–115.
35. M. S. Zhang, J. R. Huang, P. Yu and X. Chen, Talanta, 2010, 81, 162–166.
36. I. Koç, G. Baydemir, E. Bayram, H. Yavuz and A. Denizli, J. Hazard. Mater., 2011, 192, 1819–1826.
37. S. Y. Tao, C. Wang, W. Ma, S. Wu and C. G. Meng, Microporous Mesoporous Mater., 2012, 147, 295–301.
38. R. Dan, Y. Z. Wang, L. Du, S. H. Du, M. D. Huang, S. Yang and M. Zhang, Analyst, 2013, 138, 3433–3443.
39. M. A. O. da Silva and M. A. Z. Arruda, Talanta, 2009, 77, 985–990.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra07965e

---