Biocompatible zwitterionic copolymer-stabilized magnetite nanoparticles: a simple one-pot synthesis, antifouling properties and biomagnetic separation

A simple one-pot synthesis of biocompatible and antifouling magnetite nanoparticles (Fe3O4NPs) was developed. The process involves co-precipitation and in situ coating of zwitterionic copolymer poly[(methacrylic acid)-co-(2-methacryloyloxyethyl phosphorylcholine)] (PMAMPC). The influence of one-step and two-step coating methods on the performance of modified Fe3O4NP was investigated. The PMAMPC-Fe3O4NP with a narrow particle size distribution obtained from the two-step approach were highly stable in aqueous media within a wide range of pH. The particles exhibited superparamagnetic behavior with high saturation magnetization values so that they could be easily separated from solution by a magnet. Their antifouling characteristics against 2 selected proteins, lysozyme (LYZ) and bovine serum albumin (BSA), as a function of copolymer molecular weight and composition were also evaluated. Moreover, taking advantage of having carboxyl groups in the coated copolymer, the PMAMPC-Fe3O4NP were conjugated with a model biomolecular probe, biotin. The biotin-immobilized PMAMPC-Fe3O4NP were then tested for their specific capturing of a target molecule, streptavidin. The results have demonstrated the potential of PMAMPC-Fe3O4NP prepared by the two-step in situ coating method for probe immobilization and subsequent biomagnetic separation of target molecules. The fact that the developed functionalizable magnetite nanoparticles are biocompatible and antifouling also opens up the possibility of their use in other biomedical-relevant applications.


Introduction
Superparamagnetic iron oxide nanoparticles, especially magnetite nanoparticles (Fe 3 O 4 NPs), have been widely used in a variety of biomedical applications, such as magnetic resonance imaging (MRI), 1,2 magnetic separation, 3-5 drug delivery, 6,7 and biosensors. [8][9][10] These applications require particles that exhibit superparamagnetic behaviour at room temperature and have high magnetization values. To be applied for biological and medical diagnosis, the magnetite nanoparticles must be stable in water at neutral pH and under physiological conditions and should not non-specically adsorb proteins and other biological components. Coating with polymers having antifouling characteristics has been recognized as an efficient approach that can not only offer antifouling properties to but also enhance colloidal stability of the Fe 3 O 4 NPs. [11][12][13][14][15][16][17][18][19][20] In principle, the polymer-stabilized Fe 3 O 4 NPs can be prepared by either a one-step method where in situ formation of the nanoparticles and the polymer coating takes place concurrently, or a two-step method where the nanoparticles are rstly formed and the polymer is then coated thereaer.
Poly(ethylene glycol) (PEG) and its derivatives have been recognized as typical hydrophilic polymers that are extensively used to modify the Fe 3 O 4 NPs for biomedical applications due to their protein resistant properties. [12][13][14] Besides PEG, poly(2methacryloyloxyethyl phosphorylcholine) (PMPC), a hydrophilic zwitterionic polymer which contains positively and negatively charged moieties within the same structure, has also received much attention as an alternative antifouling polymeric stabilizer for Fe 3 O 4 NPs. [15][16][17][18][19][20] It has been demonstrated that PMPC can be tethered to the surface of Fe 3 O 4 NPs by surface-initiated atom transfer radical polymerization (SI-ATRP). [16][17][18][19] This "graing from" method required a two-step approach: surface immobilization of ATRP initiator and then PMPC formation. A simple one step in situ formation of PMPCstabilized Fe 3 O 4 NPs whereby co-precipitation of ferric and ferrous salts happens simultaneously with the coating of PMPC is not possible due to a lack of anchoring group with strong affinity in the PMPC structure to bind with the surface of Fe 3 -O 4 NPs. Armes and co-workers 20 have shown that such limitation can be overcome by using a double hydrophilic block copolymer of PMPC and poly(glycerol monomethacrylate) (PGMA). PGMA would act as an absorbing block via a ve-membered cyclic chelation between the 1,2-diol moieties in the side chain of PGMA and Fe atoms on the surface of Fe 3 O 4 NPs. 12,13 A similar strategy has recently been reported by Zheng and co-workers, 15 using a random copolymer of poly[(methacrylic acid)-co- ( Previously, PMPC-based copolymers have been introduced by our group as antifouling and functional polymeric layer for biosensing probe conjugation. 21,22 Thiol-terminated poly [(methacrylic acid)-co-(2-methacryloyloxyethyl phosphorylcholine)] (PMAMPC-SH) was tethered to gold-coated substrate prior to conjugation with biotin employing carboxyl groups from the MA units. By using surface plasmon resonance (SPR), highly efficient detection of a specic target analyte, avidin, in diluted blood plasma was successful owning to the ability to resist nonspecic protein adsorption of the PMPC units. 21 Although the success in preparation of magnetic iron oxide nanoparticles coated with PMPC-based polymer has been continuously reported, [15][16][17][18][19][20] there are several problems associated with the complex synthesis methodologies and the inferior magnetic properties of the synthesized nanoparticles that limit their biomedical-relevant applications. Therefore, much attention is now focused on not only the development of simple methods to incorporate the functional polymeric layer on the surface of the magnetite nanoparticles but also providing the desirable magnetic properties. Herein, we reported a simple coating method, called a two-step in situ coating, where the Fe 3 O 4 NPs were rstly grown by co-precipitation and then PMAMPC was directly added to anchor onto the Fe 3 O 4 NPs, of which the formation of the nanoparticles and the polymer coating can be done in a "one-pot" system. Unlike the PMAMPC-Fe 3 O 4 NPs prepared via the one-step in situ coating of PMAMPC as previously reported by others, 15 the PMAMPC-Fe 3 O 4 NPs prepared by the two-step in situ coating in this work possessed a higher saturation magnetization value so that they can be easily separated under an external magnetic eld without the requirement for centrifugation. To the best of our knowledge, this is the rst report on comparative studies between the onestep and two-step in situ coating methods for the preparation of PMAMPC-Fe 3 O 4 NPs and their properties. Additionally, the application of PMAMPC-Fe 3 O 4 NPs for biospecic separation was also demonstrated. Biotin conjugated to the carboxyl groups from MA units in the copolymer coated on PMAMPC-Fe 3 O 4 NPs were specically capable of capturing the target molecules, streptavidin, as evaluated in comparison with the non-target molecule, bovine serum albumin (BSA).

Synthesis of PMAMPC-Fe 3 O 4 NPs
Fe 3 O 4 NPs were prepared and coated with PMAMPC using two in situ methods: one-step and two-step methods. In the case of the two-step method, the Fe 3 O 4 NPs were rstly formed by chemical co-precipitation of ferrous/ferric ions in alkaline medium and then PMAMPC solution was directly added to Fe 3 O 4 NPs suspension providing a coating of PMAMPC on Fe 3 O 4 NPs through chelating bond. Typically, FeCl 3 $6H 2 O (298 mg, 1.10 mmol) and FeCl 2 $4H 2 O (109 mg, 0.55 mmol) were dissolved in 20 mL DI water at room temperature. Aer the salt was completely dissolved, the mixture was added to a ask under nitrogen atmosphere at 60 C and mechanically stirred at 750 rpm for 30 min. 6 mL of aqueous ammonia solution (28% w/v) was slowly added and the solution color changed from orange to black. The colloidal mixture was continuously stirred for another hour. Subsequently, 40 mg of PMAMPC (PMA 37 -MPC 63 , 54.5 kDa) corresponding to 68.8 mmol of carboxyl groups (COOH) was directly added to the black colloidal mixture. The mixture was stirred for 1 h and cooled to room temperature. The PMAMPC-Fe 3 O 4 NPs obtained were separated by a magnet and washed with DI water several times to remove unreacted chemicals. The PMAMPC-Fe 3 O 4 NPs were dispersed in distilled water and stored as aqueous colloidal suspensions. The amounts of the PMA 37 MPC 63 in the feed solution was 10, 20 and 40 mg. Copolymers with varied composition (PMA 21 MPC 79 , PMA 37 MPC 63 and PMA 66 MPC 34 ) and molecular weight (12.0, 25.9 and 54.5 kDa) were also used for the preparation of PMAMPC-Fe 3 O 4 NPs in order to investigate the effect of copolymer composition and molecular weight on antifouling properties.
In the case of the one-step method, the PMAMPC-Fe 3 O 4 NPs were synthesized by a chemical co-precipitation of ferrous/ ferric ions in the presence of PMAMPC (PMA 37 MPC 63 , 54.5 kDa) in alkaline medium. FeCl 3 $6H 2 O (298 mg, 1.10 mmol), FeCl 2 $4H 2 O (109 mg, 0.55 mmol) and PMAMPC (40 mg, 68.8 mmol of COOH) were dissolved in 20 mL DI water at room temperature. The Fe 3 O 4 NPs were formed and simultaneously chelated with the carboxylate ions in PMAMPC aer the addition of 6 mL of aqueous ammonia solution (28% w/v) under nitrogen atmosphere at 60 C. The mixture was stirred at 750 rpm for 2 h and cooled to room temperature. The PMAMPC-Fe 3 O 4 NPs obtained were separated by a magnet and washed with DI water several times. The PMAMPC-Fe 3 O 4 NP were dispersed in distilled water and stored as aqueous colloidal suspensions. Uncoated Fe 3 O 4 NPs were also synthesized by the same method in the absence of PMAMPC.

Characterization
The morphology, size and size distribution of the Fe 3 O 4 NPs were evaluated using transmission electron microscopy (TEM, Philips TECNAI 20). Fourier transform infrared (FT-IR) spectra were recorded by PERKIN ELMER using a KBr disk method. The amount of polymer content was determined by thermogravimetric analysis (TGA, MettlerToledo TGA/SDTA 851 e ). The weight loss of the dried samples was monitored under nitrogen atmosphere at temperatures ranging from 25 to 800 C with a heating rate of 10 C min À1 . The crystalline phase was studied by X-ray diffraction (XRD, Bruker AXS Model D8 Discover). The hydrodynamic diameter and zeta potential were measured by dynamic light scattering (DLS) using a Malvern Nano ZSP instrument (UK) at room temperature at a scattering angle of 90

Colloidal stability study
Colloidal stability of the unmodied Fe 3 O 4 NPs and PMAMPC-Fe 3 O 4 NPs suspensions (0.1 mg mL À1 ) were visually monitored in buffer solutions having pH in the range 1-11 in the absence of external magnetic eld. The quantitative measure of colloidal stability by DLS has been performed.

Protein adsorption study
To test the antifouling properties of the PMAMPC-Fe 3 O 4 NPs, LYZ and BSA were used as models for positively charged and negatively charged proteins, respectively. The protein was dissolved in PBS (10 mM, pH 7.4). 500 mL of 1 mg mL À1 protein solution was added to a 500 mL colloidal solution of PMAMPC-Fe 3 O 4 NPs (2 mg mL À1 ) and incubated under gentle shaking at room temperature for 30 min. Then, the particles were magnetically separated. Supernatant was collected and analyzed for the remaining protein by the Bradford protein assay. The amount of adsorbed protein on the PMAMPC-Fe 3 -O 4 NPs (q e , mg g À1 ) was calculated using eqn (1).
when C 0 and C e are the protein concentrations before and aer adsorption (mg mL À1 ), v is the volume of the aqueous phase (mL), and m is the weight of the nanoparticles (g). ) were added to wells for 24 h. The nanoparticles were removed from the cells and then the morphology of cells before and aer culturing with the nanoparticles were observed using a Olympus inverse phase contrast microscope (OLYMPUS IX70, Japan) equipped with an objective (LCAch 20X/ 0.04phC, Olympus, Japan) of 20Â magnication. MTT solution (10 mL, 5 mg mL À1 in PBS) was added into each well for 2 h before aspiration of the solution. Dimethyl sulfoxide (500 mL) was added into each well to solubilize the blue formazan crystal product. Subsequently, the absorbance was measured at the wavelength of 550 nm using a microplate reader (Molecular Device, USA). The amount of formazan was proportional to the number of functional mitochondria in viable cells. All experiments were performed in triplicate and cells in culture media without the nanoparticles were used as a control. Percentage of cell viability was expressed as absorbance of treated well/ absorbance of control well Â 100. 23,24 Data were expressed as means AE S.D. (n ¼ 3).

Immobilization of biotin onto PMAMPC-Fe 3 O 4 NPs
The carboxyl groups of the coated PMAMPC were rst activated by an aqueous solution of EDC and NHS. A 100 mL of fresh mixed solution of EDC (50 mg mL À1 ) and NHS (50 mg mL À1 ) was added to 500 mL of aqueous suspension of PMAMPC-Fe 3 -O 4 NPs (10 mg mL À1 ) for 30 min. The activated PMAMPC-Fe 3 -O 4 NPs were separated from the solution by a magnet and washed with DI water. Then, the activated PMAMPC-Fe 3 O 4 NPs were re-suspended in 500 mL of NH 2 -biotin solution (10 mg mL À1 ) in DI water and incubated at room temperature overnight. Aer the incubation step, the nanoparticles were separated with a magnet, washed and re-suspended in DI water.
Biospecic separation of biotinylated PMAMPC-Fe 3 O 4 NP 100 mL of biotinylated PMAMPC-Fe 3 O 4 NPs (1 mg mL À1 ) was added to 100 mL of streptavidin (1 mg mL À1 ). The mixture was mixed by vortex and incubated by shaking vigorously at room temperature for 45 min. By a magnetic separation, the supernatant solution was collected and used to determine the content of un-captured streptavidin by UV/Vis spectrophotometer. The capture efficiency (%) was calculated using eqn (2).
when C 0 and C e are the streptavidin concentration in solution before and aer magnetic separation expressed as the absorbance value at 282 nm, respectively. A similar procedure was also performed for BSA separation.  Fig. S2 in ESI, † the uncoated Fe 3 O 4 NPs showed a major weight loss of 4% in a temperature range of 25-100 C which should be due to loss of water. A similar magnitude of water loss from both types of PMAMPC-Fe 3 O 4 NPs was also found in the same temperature range. PMAMPC-Fe 3 O 4 NPs showed the additional weight loss between 200 and 400 C which could be attributed to the thermal decomposition of the PMAMPC copolymer. The weight losses of 14.8 and 13.5% corresponding to 11.1% and 9.7% of PMPC content were detected for the PMAMPC-Fe 3 O 4 NPs prepared by the one-step and twostep methods, respectively. The results demonstrated that using the same amount of PMAMPC copolymer in the feed, a slightly greater PMAMPC content was incorporated into the PMAMPC-Fe 3 O 4 NPs prepared by the one-step method than that in the two-step method. This may be due to the less steric hindrance in one-step method where the incorporation of the polymer chain into the nanoparticles and the growth of the nanoparticles take place concurrently. The two-step method based on "graing to" approach where the nanoparticles are rstly formed and the polymer is then coated on the nanoparticles surface therefore gives a low coating density since the  attached polymer chains possess a steric barrier to the approaching polymer chains. So, the PMAMPC-Fe 3 O 4 NPs obtained from the two-step method should have a lower PMAMPC content.

Results and discussion
Black colloidal suspensions of PMAMPC-Fe 3 O 4 NPs obtained from both the one-step and two-step methods showed good colloidal stability in aqueous media for up to 30 days of storage at room temperature (Fig. 2a). However, only the PMAMPC-Fe 3 O 4 NPs prepared by the two-step method could be rapidly separated from the dispersion using an external magnetic eld for 5 min (Fig. 2b), showing they had an excellent magnetic response. The magnetization curves measured at room temperature for the PMAMPC-Fe 3 O 4 NPs prepared by the onestep and two-step methods are compared in Fig. S3, ESI. † There was no residual magnetism or hysteresis loop in the magnetization of both samples, suggesting the Fe 3 O 4 NPs produced are superparamagnetic. High saturation magnetization values of 48 and 47 emu g À1 were detected for the uncoated Fe 3 O 4 NPs and PMAMPC-Fe 3 O 4 NPs prepared by the two-step method. The saturation magnetization value of the PMAMPC-Fe 3 O 4 NPs prepared by the one-step method (38 emu g À1 ) was lower than the other samples. This may be ascribable to the lower weight fraction of the magnetic material within the magnetic nanoparticles-polymer composite as previously reported others. 15,18,30 The magnetite content of PMAMPC-Fe 3 O 4 -NPs prepared by the one-step method was given as a weight fraction of 88.9% which was lower than that of PMAMPC-Fe 3 -O 4 NPs prepared by the two-step method (90.3%). The reduction of magnetization value from 47 to 38 emu g À1 may not seem large. However, the difference is signicant enough to possess an impact on the ability to be separated by a magnet. In fact a similar observation was reported by Majeed and co-workers 31 on the pentaerythritol tetrakis 3-mercaptopropionatepolymethacrylic acid functionalized magnetic iron oxide (PTMP-PMA-Fe 3 O 4 NP s ). They have found that the particles cannot be separated by a magnet once the magnetization value was decreased from 58 to 45 emu g À1 . The smaller mean particle size (7.9 AE 2.2 nm) corresponded to the PMAMPC-Fe 3 O 4 NPs prepared by the two-step method (11.7 AE 2.8 nm) (See Fig. 2c-f) may also account for the reduced magnetism. 28,29 These results support the reasons why the PMAMPC-Fe 3 O 4 NPs prepared by the one-step method could not be separated out of the solution by a magnet. Fig. S4 in ESI † shows the crystalline structure of uncoated Fe 3 O 4 NPs in comparison with PMAMPC-Fe 3 O 4 NPs prepared by both methods. All samples showed the standard XRD pattern of magnetite nanoparticles corresponding well with previous reports. 15,28,30 However, the area of the diffraction peaks of the PMAMPC-Fe 3 O 4 NPs prepared by the one-step method (Fig. S4c †) were lower than those of the uncoated Fe 3 O 4 NPs and the PMAMPC-Fe 3 O 4 NPs prepared by the two-step method. The peak area data are summarized in Table S2, ESI. † It is possible that this as a consequence of the polymer coating during nanoparticle formation interfering with the crystallization process of magnetite nanoparticles. The particle sizes of the nanoparticles were also calculated using the Scherrer formula. 32 The detail of calculation is provided in Table S3, ESI. † The sizes of PMAMPC-Fe 3 O 4 NPs prepared by the one-step and two-step method calculated from the XRD data are 9.8 and 11.4 nm, respectively which corresponded quite well with sizes evaluated by the TEM analysis (Fig. 2).

Colloidal stability of PMAMPC-Fe 3 O 4 NPs
The stability of PMAMPC-Fe 3 O 4 NPs dispersion was determined by monitoring the settling of the nanoparticles in aqueous media. As shown in Fig. 3A, the uncoated Fe 3 O 4 NPs started to aggregate and settle to the bottom of vial within 15 min (as indicated by red arrows). It could be veried with an increase in diameter of the uncoated Fe 3 O 4 NPs from nanometer range to micrometer range within 15 min as evaluated by DLS (Fig. 3D). On the other hand, the PMAMPC-Fe 3 O 4 NPs synthesized by the two in situ methods exhibited good colloidal stability with unchanged particle size, suggesting that the hydrophilic PMAMPC coating signicantly improves the stability of Fe 3 O 4 -NPs suspension. In addition, the increase of copolymer coating on Fe 3 O 4 NPs surface by increasing the PMAMPC amount in the feed solution possessed a strong impact on colloidal stability (Fig. 3A, column 3-5). The quantity of coated PMAMPC estimated from the weight loss in the temperature range 200-400 C by TGA analysis are shown in Fig. S5, ESI. † Because of the dissociation of COOH to COO-groups, at pH $ 5, the surface of PMAMPC-Fe 3 O 4 NPs would become positively charged resulting in electrostatic repulsion between neighboring nanoparticles preventing aggregation. On the other hand, the COOH groups remain protonated at pH < 5, inducing the agglomeration of nanoparticles as can be seen in Fig. 3B. The zeta potential values shown in Fig. 3D support the assumption that colloidal stability of PMAMPC-Fe 3 O 4 NPs is due to COOH dissociation.

Protein adsorption study
The antifouling property of PMAMPC-Fe 3 O 4 NPs synthesized by the two-step method was determined by monitoring the adsorption of two proteins (LYS and BSA) in PBS solution (pH 7.4). As expected, the presence of PMAMPC on the surface of PMAMPC-Fe 3 O 4 NPs resulted in low adsorption of both LYS and BSA as compared with the uncoated Fe 3 O 4 NPs (Fig. 4). Under the test conditions in the PBS solution (pH 7.4, 10 mM), BSA (pI ¼ 4.8) would be negatively charged whereas lysozyme (pI ¼ 12) would be positively charged. The ability to resist protein adsorption is concentration-dependent. In other words, the amount of adsorbed protein decreased with increasing PMAMPC amount introduced to the feed solution during the one-pot synthesis (Fig. 4a). This coincides with the fact that the high amount of PMAMPC contained a greater number of hydrophilic MPC units on the surface of Fe 3 O 4 NPs, resulting in more effective suppression of protein adsorption. A similar result was found in both of the increasing polymer chain length and number of MPC units in the copolymer (Fig. 4b and c) in which a greater number of MPC units were obtained. The results strongly suggested that MPC units in the PMAMPC copolymer chain played an important role in preventing protein adsorption. This is in excellent agreement with many reports previously published on the fact that the introduction of MPC units in the copolymer exhibits an outstanding resistance to non-specic interactions with proteins due to its cellmembrane mimic structure. [15][16][17][18][19][20][21][22] Evaluation of cytotoxicity In this research, cytotoxicity of the Fe 3 O 4 NPs was evaluated against human umbilical vein endothelial cells (EA.Hy926). As can be seen in Fig. 5, the morphology and growth of cells were not affected by the presence of PMAMPC-Fe 3 O 4 NPs, implying that the PMAMPC-Fe 3 O 4 NPs are biocompatible and not toxic to the cells. The cell-membrane mimic structure of PMPC provides a better environment for cells so that they can maintain their activity and stability. In addition, the PMAMPC-Fe 3 O 4 NPs did not seem to attach well to the cell surface as opposed to the uncoated Fe 3 O 4 NPs which may be described as a result of the   anti-fouling characteristic of PMPC-based polymer. On the other hand, cells were obviously damaged in the presence of uncoated Fe 3 O 4 NPs, indicating that they are toxic to the cells. The effect of Fe 3 O 4 NPs on the cell viability was also evaluated. The results shown in Fig. 6 indicated that PMAMPC-Fe 3 O 4 NPs slightly reduced the cell viability (80-84%) when compared with the control aer culturing for 24 h. Interestingly, the cell viability remained almost unchanged when the Fe 3 O 4 NPs concentration increased, even at relatively high concentration (up to 0.25 mg mL À1 ). The uncoated Fe 3 O 4 NPs caused a significant reduction in cell viability (44-60%) in a dose-dependent manner. Apparently, the coating of zwitterionic PMAMPC signicantly improves the biocompatibility of the Fe 3 O 4 NPs as evidenced by the results from cell morphology and the cell viability. This outcome is in good agreement with the work previously reported by Zheng and co-workers 15 who have demonstrated that the PMAMPC-Fe 3 O 4 NPs prepared via the one-step in situ coating of PMAMPC were biocompatible and showed low cell uptake efficiency which correlates well with the high percentage of cell viability.

Biomagnetic separation by PMAMPC-Fe 3 O 4 NPs
Besides being the anchoring points to the surface of Fe 3 O 4 NPs, the carboxyl groups of the coated PMAMPC are employed as active sites for attaching amino-functionalized biotin (NH 2biotin) via EDC/NHS coupling chemistry. The success of biotin immobilization can be veried by the decrease of the C]O stretching peak of the carboxyl groups at 1710 cm À1 in the FT-IR spectrum shown in Fig. 1d. The biotinylated PMAMPC-Fe 3 O 4 NPs were then further tested for specic conjugation with its target molecule, streptavidin. Aer the conjugation with streptavidin, separation and washing of the biotinylated PMAMPC-Fe 3 O 4 NPs can be easily performed with the aid of a magnet without the requirement for centrifugation. In order to conrm the specicity of biotinylated PMAMPC-Fe 3 O 4 NPs, the separation of streptavidin was compared with BSA, the non-target molecule (Fig. 7). In the case of streptavidin, the peak at 282 nm which corresponds to the absorbance of protein chromophores disappeared from the UV-vis spectrum aer magnetic separation with biotinylated PMAMPC-Fe 3 O 4 NPs implying that the streptavidin molecules were captured and entirely removed from the solution. On the other hand, only a slight decrease in intensity of the same peak from the UV-vis spectrum of BSA was observed aer magnetic separation with biotinylated PMAMPC-Fe 3 O 4 NPs. The capture efficiency can be calculated by comparing the concentration of streptavidin or BSA in solution before and aer magnetic separation which were found to be 91.5% and 28.2%, respectively. The results indicate the specicity of the biotinylated PMAMPC-Fe 3 O 4 NPs for streptavidin, over BSA. The nonspe-cic adsorption of the biotinylated PMAMPC-Fe 3 O 4 NPs towards BSA corresponds quite well with the data displayed in Fig. 4 in which a nonspecic adsorption of BSA by the PMAMPC-Fe 3 O 4 NPs prepared from 40 mg of PMA 37 MPC 63 , 54.5 kDa was observed. This nonspecic adsorption may be suppressed if blocking of unreacted carboxyl groups is applied aer biotin conjugation.

Conclusions
A simple method for the synthesis of PMAMPC-coated Fe 3 O 4 -NPs by a two-step in situ reaction has been demonstrated. The anchoring of PMAMPC onto the surface of Fe 3 O 4 NPs was proven to be through carboxylate chelating interactions. The PMAMPC-Fe 3 O 4 NPs showed excellent long-term colloidal stability in aqueous media at neutral pH and their stability was pHdependent. Despite being coated by the zwitterionic copolymer, the PMAMPC-Fe 3 O 4 NPs exhibited a high saturation magnetization value so that they can be easily separated under an external magnetic eld without the requirement for centrifugation. The antifouling property of PMAMPC-Fe 3 O 4 NPs was found to be dependent on the copolymer content, molecular weight and composition. The coating of zwitterionic PMAMPC can denitely improve biocompatibility of the Fe 3 O 4 NPs, suggesting that the PMAMPC-Fe 3 O 4 NPs may also be further used in contact with cells, for example, as drug carriers. The carboxyl groups in the copolymer coated on the PMAMPC-Fe 3 O 4 NPs are readily available for biotin conjugation. The biotinylated PMAMPC-Fe 3 O 4 NPs can bind specically with the target, streptavidin as opposed to the non-specic target, BSA.