Cationic Polymer-Coated Magnetic Nanoparticles with Antibacterial Properties: Synthesis and In Vitro Characterization

Uniformly sized magnetite nanoparticles (Dn = 16 nm) were prepared by a thermal decomposition of Fe(III) oleate in octadec-1-ene and stabilized by oleic acid. The particles were coated with Sipomer PAM-200 containing both phosphate and methacrylic groups available for the attachment to the iron oxide and at the same time enabling (co)polymerization of 2-(dimethylamino)ethyl methacrylate and/or 2-tert-butylaminoethyl methacrylate at two molar ratios. The poly[2-(dimethylamino)ethyl methacrylate] (PDMAEMA) and poly[2-(dimethylamino)ethyl methacrylate-co-2-tert-butylaminoethyl methacrylate] [P(DMAEMA-TBAEMA)] polymers and the particles were characterized by 1H NMR spectroscopy, size-exclusion chromatography, transmission electron microscopy, dynamic light scattering, thermogravimetric analysis, magnetometry, and ATR FTIR and atomic absorption spectroscopy. The antimicrobial effect of cationic polymer-coated magnetite nanoparticles tested on both Escherichia coli and Staphylococcus aureus bacteria was found to be time- and dose-responsive. The P(DMAEMA-TBAEMA)-coated magnetite particles possessed superior biocidal properties compared to those of P(DMAEMA)-coated one.


Introduction
The invention of antibiotics in the 1940s saved millions of lives and many infectious diseases became far less deadly. However, bacteria have currently developed resistance to antibiotics, which is now a major threat for public health [1][2][3]. Moreover, there is a need for disinfection of drinking water in remote areas using alternative approaches to chlorination, in order to eliminate pathogens responsible for waterborne diseases. There are other physical processes to disinfect the contaminated water including filtration, thermal treatment, or UV irradiation; however, these methods suffer from limitations, such as time demand, high price, waste of resources, use of various additives and agents, etc. [4]. In a search for new, cheap, and simple solutions of these problems, a significant effort has been devoted to the development of antibacterial agents based on nanoparticles [5][6][7]. In contrast to conventional antibiotics, such materials possess a range of unique physicochemical and biological properties, e.g., high surface-to-volume ratio that increases contact area with microorganisms, high stability, possibility of easy modification with various functional groups, ligands, targeting agents and other biomolecules, enabling not only disinfection, but also its monitoring and targeting [8][9][10]. Particles possessing sterilization properties are typically based on silver, copper, various metal oxides or sulfides, or carbon nanotubes [11].

Physicochemical Characterization
High-resolution 1 H NMR spectra were acquired in deuterated water with a Bruker Avance III 600 spectrometer operating at 600.2 MHz and processed with the Topspin 4.0.5 software (both Bruker; Billerica, MA, USA). Measurements were as follows: 90 • pulse width 10 µs, relaxation delay 10 s, spectral width 7211 Hz, acquisition time 4.54 s, and 32 scans. The integrated intensities were determined using the spectrometer integration software with an accuracy of ±1%. During the measurements, temperature was maintained at 25 ± 0.2 • C using a BVT 3000 temperature unit.
The size-exclusion chromatography (SEC) of the polymers was performed at 25 • C with a TSKgel SuperAW-L guard column (L × I.D. 4.6 mm × 3.5 cm, particle size 7 µm; Polymer Laboratories; Church Stretton, UK) and UVD 305 (Watrex; Prague, Czech Republic) and RI-101 (Shodex; Tokyo, Japan) detectors. Methanol/acetate buffer mixture (80/20 v/v) was used as a mobile phase at a flow rate of 0.5 mL/min. The molar mass was calculated using Clarity software (DataApex; Prague, Czech Republic) with poly-styrene as a calibration standard. The samples for both 1 H NMR and SEC analysis were prepared by dissolution of the polymer-coated MNP in 2 M HCl; pH of the solution was adjusted to 7 and the solution was dialyzed against water using a cellulose dialysis membrane (MWCO 3.5 kDa) and purified on a Sephadex TM LH-20 column (Merck; Kenilworth, NJ, USA).
Morphology, size, and size distribution of the particles were analyzed with a Tecnai G2 Spirit Twin 12 transmission electron microscope (TEM; FEI; Brno, Czech Republic) [10]. Numberaverage diameter (D n = ΣN i ·D i /ΣN i ), weight-average diameter (D w = ΣN i ·D i 4 /ΣN i ·D i 3 ), and dispersity (Ð = D w /D n ) were calculated from at least 500 individual particles from the TEM micrographs using Atlas software (Tescan, Brno, Czech Republic).
Hydrodynamic diameter (D h ), polydispersity (PD), and ζ-potential were obtained by dynamic light scattering (DLS) using a ZEN 3600 Zetasizer Nano instrument (Malvern Instruments; Malvern, UK) at RT and different pHs. The D h was calculated from the intensity-weighted distribution function obtained by CONTIN analysis of the correlation function embedded in Malvern software as well as electrophoretic mobility of the magnetic nanoparticles, which was converted to ζ-potential using the Smoluchowski equation.
A PerkinElmer 3110 atomic absorption spectrometer (AAS) was used to analyze the amount of iron in the particles by measuring the solution obtained after their mineralization with 68% HClO 4 /65% HNO 3 (4/1 v/v) mixture at 80 • C for 20 min. The amount of phosphorus was determined after the mineralization and mixing with sulfuric acid, ammonium molybdate, ascorbic acid, and antimony potassium tartrate using a PerkinElmer Lambda 20 UV-Vis spectrometer at 690 nm.
ATR FTIR spectra were recorded on a PerkinElmer Paragon 1000PC spectrometer equipped with a Specac MKII Golden Gate single attenuated total reflection system with a diamond crystal; the angle of incidence was 45 • .
Thermogravimetric analysis (TGA) was performed in air at 20-650 • C with a heating rate of 10 • C/min using a PerkinElmer TGA 7 analyzer and Pyris 1 software (Shelton, CT, USA).
Magnetic properties, namely coercivity and induced remanent magnetization were measured using an EV9 vibrating sample magnetometer (MicroSense; Lowell, MA, USA) at RT and maximum magnetic field of 1 T. Hysteresis loops were determined with different field steps from 500 Oe (0.05 T) in the highest field range (300 mT-1 T) to the finest step of 5 Oe (0.5 mT) in the lowest field range (−25-25 mT) at the intersections with magnetization and field axes. This allowed reliable determination of saturation remanent magnetization M rs , i.e., magnetization at zero field (the intersection with the magnetization axis), and coercive force B c that is field necessary to induce zero magnetization (the intersection with the field axis). The two parameters were determined on both the ascending and descending branches of the hysteresis loop and resulting average values were considered. Saturation induced magnetization M s was determined as the maximum induced magnetization after subtraction of the linear paramagnetic part of the loop at >0.5 T. Here, the ferrimagnetic response to external field was saturated and reflected by closed, reversible, and linear ascending and descending branches of the loop. Again, average M s values were calculated, reflecting the concentration of ferrimagnetic substance in the specimen. To check its ability to acquire remanent magnetization, the measurements were performed in zero field after application of a step-wise increasing magnetic field, with 50 Oe (5 mT) field steps and 5000 Oe (0.5 T) maximum field. Remanent magnetization in true superparamagnetic particles at RT should decay rapidly to zero. Finally, after application of the maximum field, the same process was repeated, but with the field applied in the opposite direction. The field is necessary to remove M rs named coercivity of remanence (B cr ) and, similarly like M rs and coercivity B c , it reflects magnetite particle size.

Antibacterial Activity of MNP@S-D and MNP@S-D-T
In 24-well plates (TPP; Trasadingen, Switzerland), overnight cultures of S. aureus and E. coli were resuspended in HBSS and diluted in LB (1 mL) to cultivate bacteria at 1 × 10 5 colony-forming units (CFU) per mL. The MNP@S-D and MNP@S-D-T nanoparticles (5.5, 44, and 175 µg/mL) were added to the bacteria cultures and incubated at 37 • C with shaking (125 rpm). Culture without any treatment (NT) and containing ampicillin (150 µg/mL) served as a negative and positive control, respectively. After the exposure to nanoparticles, the bacteria aliquots were taken from the wells in time intervals of 0, 1, and 4 h, diluted 100×, 1000×, or 1,000,000×, seeded on LA plates, and incubated at 37 • C for 24 h. The procedure was repeated three times in duplicates and bacterial CFU/mL were calculated. The viability was calculated as a ratio of the number of living bacteria at given time interval to the number of living bacteria in NT. Data were analyzed by the GraphPad Prism5 software and two-way ANOVA followed by Bonferroni's test.

Synthesis of MNP and Their Physicochemical Characterization
OA-stabilized magnetite cores were synthesized by a thermal decomposition method in a non-polar high-boiling organic solvent (OD; Figure 1). The method allowed us to prepare magnetic nanoparticles with a very narrow size distribution, high crystallinity, and controlled diameter [39]. The synthesis proceeded in three steps: (i) formation of poly(iron oxo) clusters from an iron oleate serving as building blocks for the particle growth, (ii) short burst nucleation with formation of the particle seeds, and (iii) nanoparticle growth with increasing reaction temperature [36,40]. Due to the combination of short nucleation time and the presence of a stabilizer that prevents the nuclei aggregation and ensures the same conditions for each particle growth the resulting particles had controlled size (Dn = 16 nm) with low dispersity (Ð < 1.05) according to TEM (Figure 2a). Additionally, polydispersity (PD = 0.13) from DLS documented a very narrow particle size distribution ( Table 1). The hydrodynamic diameter of nanoparticles measured in toluene was larger (Dh = 36 nm) than Dn due to the OA adsorbed on the particle surface. The FTIR spectra of MNP exhibited three peaks at 1460, 2844, and 2919 cm −1 attributed to νd(CH) deformation vibration, and νs(CH2) symmetric and νas(CH3) asymmetric stretching vibration, respectively ( Figure 3a). The bands at 1559 and 1703 cm −1 corresponded to νa(COO − ) asymmetric and ν(COOH) stretching vibrations, respectively. Therefore, FTIR spectroscopy confirmed the presence of OA coating on the Fe3O4 nanoparticle surface. According to TGA, the amount of organic compounds on the MNP surface equaled to 84 wt.%. Due to the combination of short nucleation time and the presence of a stabilizer that prevents the nuclei aggregation and ensures the same conditions for each particle growth the resulting particles had controlled size (D n = 16 nm) with low dispersity (Ð < 1.05) according to TEM (Figure 2a). Additionally, polydispersity (PD = 0.13) from DLS documented a very narrow particle size distribution ( Table 1). The hydrodynamic diameter of nanoparticles measured in toluene was larger (D h = 36 nm) than D n due to the OA adsorbed on the particle surface. The FTIR spectra of MNP exhibited three peaks at 1460, 2844, and 2919 cm −1 attributed to ν d (CH) deformation vibration, and ν s (CH 2 ) symmetric and ν as (CH 3  respectively. Therefore, FTIR spectroscopy confirmed the presence of OA coating on the Fe 3 O 4 nanoparticle surface. According to TGA, the amount of organic compounds on the MNP surface equaled to 84 wt.%. the resulting particles had controlled size (Dn = 16 nm) with low dispersity (Ð <1.05) according to TEM (Figure 2a). Additionally, polydispersity (PD = 0.13) from DLS documented a very narrow particle size distribution ( Table 1). The hydrodynamic diameter of nanoparticles measured in toluene was larger (Dh = 36 nm) than Dn due to the OA adsorbed on the particle surface. The FTIR spectra of MNP exhibited three peaks at 1460, 2844, and 2919 cm −1 attributed to νd(CH) deformation vibration, and νs(CH2) symmetric and νas(CH3) asymmetric stretching vibration, respectively (Figure 3a). The bands at 1559 and 1703 cm −1 corresponded to νa(COO − ) asymmetric and ν(COOH) stretching vibrations, respectively. Therefore, FTIR spectroscopy confirmed the presence of OA coating on the Fe3O4 nanoparticle surface. According to TGA, the amount of organic compounds on the MNP surface equaled to 84 wt.%.

Surface Modification of MNP with Sipomer PAM-200
For the application of antibacterial MNP in disinfection of aqueous media, waterdispersible nanoparticles are required. However, the original magnetic particles contained hydrophobic OA coating and therefore they have to be modified with a hydrophilic polymer such as PDMAEMA to provide both dispersibility and colloidal stability of particles in water. To ensure efficient attachment of both PDMAEMA and its copolymers to the iron oxide surface, Sipomer PAM-200 terminated with phosphate groups was selected as a mediator. While the phosphate group enabled attachment of Sipomer PAM-200 to the iron oxide, which was superior to a mere physical adsorption, its reactive methacrylic group with a vinyl bond separated by a ten-carbon spacer from phosphate anchoring group allowed the free-radical copolymerization of DMAEMA monomer [16]. Analogously to OA, Sipomer PAM-200 due to its low contrast was not visible on the TEM micrograph ( Figure 2b). The hydrodynamic diameter of MNP@S only slightly decreased to 28 nm compared to the initial nanoparticles, probably due to poorer solvation of Sipomer PAM-200 shell in toluene than that of OA (Table 1).

Modification of MNP@S with PDMAEMA and P(DMAEMA-TBAEMA)
From various antimicrobial polymers intended as a coating for the MNP, cationic PDMAEMA and PTBAEMA were selected due to their mucoadhesive, antibacterial, and stimuli-sensitive properties [25,41]. As the PTBAEMA has limited solubility in water, TBAEMA was copolymerized with a highly hydrophilic DMAEMA monomer to obtain water-dispersible antibacterial magnetic agents. At the beginning, different DMAEMA/TBAEMA molar ratios (1/0, 0.75/0.25, 0.5/0.5, 0.25/0.75) were studied in the preparation of polymers. While the MNP coated with copolymers prepared at the DMAEMA/TBAEMA ratio of 0.25/0.75 mol/mol were too hydrophobic and thus nondispersible in water, the MNP@S-D, MNP@S-D-T1, and MNP@S-D-T2 prepared with DMAEMA/TBAEMA ratios equaling to 1/0, 0.75/0.25, and 0.5/0.5 mol/mol, respectively, were water-dispersible and therefore tested in further experiments ( Table 1). The compositions of PDMAEMA and P(DMAEMA-TBAEMA) copolymers with different DMAEMA/TBAEMA molar ratios were investigated by 1 H NMR spectroscopy by comparing methyl signals of side chains from DMAEMA and TBAEMA (Figure 4). The "1" resonance found at 2.1 ppm corresponded to (CH 3 ) 2 from DMAEMA chains, while signal "1" situated at~1 ppm was related to (CH 3 ) 3 protons from TBAEMA. Calculated monomer ratios in the (co)polymers were in good agreement with those added in the polymeri-zation feed. The weight-(M w ), number-average molar mass (M n ), and polydispersity (M w /M n ) of the polymers determined by SEC analysis showed that with increasing DMAEMA/TBAEMA ratio M w of S-D, S-D-T1, and S-D-T2 polymers increased from 145 to 195 and 240 kDa, respectively; polydispersity remained rather low (M w /M n = < 1.2) due to the purification of polymers by dialysis and/or chromatography.  (Table 1). At this pH, amino groups of PDMAEMA were partially protonated and, as a result, the ξ-potential was quite high. It was typical that the hydrodynamic diameter Dh of particles measured by DLS in water was larger than the number-average diameter according to TEM due to several reasons. In DLS, the hydrodynamic diameter is determined by the autocorrelation function that compares the fluctuation of intensity of scattered light and provides z-average diameter (more sensitive to bigger particles). In contrast, TEM provides the number-average diameter that is in principle smaller than the Dh. Moreover, the DLS measures the particles in water, where the polymer coatings can be swollen, while the TEM analyzes dry specimens. Finally, last but not least, the particles form doublets, triplets, and other small aggregates in water, whereas individual particles are calculated from TEM micrographs. The presence of phosphate groups on the particles was confirmed by the UV-Vis spectroscopy that revealed 0.88, 0.69, and 0.42 wt.% of phosphorus in the MNP@S-D, MNP@S-D-T1, and MNP@S-D-T2, respectively. These values agreed with the literature data, where Sipomer PAM-200-coated γ-Fe2O3 nanoparticles contained 1.49 wt.% of phosphorus [16]. The content of PDMAEMA-based coatings on the particles, w(polymer), was indirectly calculated not only from the amount of phosphorus, but also from the quantity of iron determined by AAS. Since the PDMAEMA or PTBAEMA did not contain any Fe or P, the content of PDMAEMA-based polymers on the particles could be calculated from the amount of the elements w(Fe) and w(P) according to Equation (1) (Table 1). At this pH, amino groups of PDMAEMA were partially protonated and, as a result, the ξ-potential was quite high. It was typical that the hydrodynamic diameter D h of particles measured by DLS in water was larger than the number-average diameter according to TEM due to several reasons. In DLS, the hydrodynamic diameter is determined by the autocorrelation function that compares the fluctuation of intensity of scattered light and provides z-average diameter (more sensitive to bigger particles). In contrast, TEM provides the number-average diameter that is in principle smaller than the D h . Moreover, the DLS measures the particles in water, where the polymer coatings can be swollen, while the TEM analyzes dry specimens. Finally, last but not least, the particles form doublets, triplets, and other small aggregates in water, whereas individual particles are calculated from TEM micrographs. The presence of phosphate groups on the particles was confirmed by the UV-Vis spectroscopy that revealed 0.  [16]. The content of PDMAEMA-based coatings on the particles, w(polymer), was indirectly calculated not only from the amount of phosphorus, but also from the quantity of iron determined by AAS. Since the PDMAEMA or PTBAEMA did not contain any Fe or P, the content of PDMAEMA-based polymers on the particles could be calculated from the amount of the elements w(Fe) and w(P) according to Equation (1) (Table 2). This was in agreement with the amount of coating determined by TGA (79 wt.%; Table 2). Similarly, MNP@S-D-T1 and MNP@S-D-T2 contained 65 and 82 wt.% of the P(DMAEMA-TBAEMA) copolymer according to AAS, while TGA revealed 66 and 78 wt.% of the polymer, which was again in a reasonable agreement. The surface composition of MNP@S-D, MNP@S-D-T1, and MNP@S-D-T2 was further analyzed by FTIR spectroscopy (Figure 3a). The bands at 1148, 1454, and 1720 cm −1 were attributed to ν as (C-O-C) asymmetric stretching, δ(CH 2 ) bending, and ν(C=O) stretching vibration, respectively. The two peaks located at 2861 and 2941 cm −1 were ascribed to ν s (CH 2 ) symmetric and ν as (CH 3 ) asymmetric stretching vibrations respectively. The broad band at 3380 cm −1 was assigned to ν(OH) stretching vibration of adsorbed water. The content of polymer on the particle surface was then analyzed by TGA (Figure 3b). In the MNP@S-D-T2, the initial weight loss observed at <100 • C was assigned to dehydration of particles. In the MNP@S-D, MNP@S-D-T1, and MNP@S-D-T2, the weight loss occurred at the temperatures ranges 219-584, 210-580, and 184-594 • C, respectively, that were associated with the decomposition of the polymer shell. After considering the Sipomer PAM-200 contribution (8 wt.%; Figure 3b), the content of PDMAEMA or P(DMAEMA-TBAEMA) was 77, 63, and 76 wt.% for MNP@S-D, MNP@S-D-T1, and MNP@S-D-T2, respectively. This amount of polymer coating was enough to ensure good colloidal stability of particles even after one month of storage. In addition, with increasing proportion of TBAEMA in the coating from 0 to 50 mol.%, the thermal stability of shell decreased, i.e., the decomposition of polymer started already from 219 to 184 • C, respectively. Consequently, it can be concluded that all three types of nanoparticles were stable at temperatures < 180 • C, which is quite sufficient in terms of prospective heat sterilization needed for biological experiments.
To prove the superparamagnetic character of the particles, their magnetic properties were examined. The measurements of remanent magnetization in zero field, i.e., acquisition of remanent magnetization as well as back-field remagnetization, yielded noisy curves close to zero. The inability of material to acquire stable remanent magnetization is typical for truly superparamagnetic particles, where the collective behavior due to interparticle interactions is absent. This was supported by the shape of hysteresis loops passing virtually through the origin and yielding coercive force B c < 1 Oe (0.1 mT). Additional-ly, determination of M rs from the induced magnetization measurements was obstructed by the very narrow loops (exemplified on Figure 3c; Table 1). Moreover, the data suggested that there were no interparticle magnetic interactions which might cause collective behavior and unwanted particle clustering in the magnetic field. Despite high degree of uncertainty, both parameters (M rs and B c ) suggested an increasing trend ( Table 1). The only reliably determined magnetic parameter was saturation induced magnetization M s , reflecting the concentration of ferrimagnetic particles and confirming that the particles could be removed from aqueous suspensions using a magnet [42]. Stable single-domain and coarse multidomain magnetite and maghemite have saturation magnetization of~92 and~80 A·m 2 /kg, respectively; however, magnetic parameters of synthetic nanosized iron oxides are known to be size-dependent [43,44]. Coercive force and saturation remanence of our synthesized particles were well below the minimum values for nanosized magnetite, indicating that the particle size was~10 nm [44]. Consequently, saturation magnetization of~55 A·m 2 /kg was used to estimate the upper limit of iron oxide concentration (Table 1). Hence, the content of Fe 3 O 4 in MNP@S-D, MNP@S-D-T1, and MNP@S-D-T2 could be calculated, amounting to 13, 20, and 13 wt.%, respectively, that corresponded to 80 and 87 wt.% of coating (Table 2). These results approximately agreed with those from analysis of phosphorus and TGA.

Antibacterial Activity of the Particles
Cationic polymers, exemplified by thermo-and pH-responsive PDMAEMA, or hydrophobic TBAEMA are known to be promising antimicrobial agents [25,41]. Here, antimicrobial properties of MNP@S-D, MNP@S-D-T1, and MNP@S-D-T2 were investigated against two commonly used species of pathogen bacteria, namely, S. aureus and E. coli. These bacteria are Gram-negative (E. coli) and Gram-positive (S. aureus) containing structurally different cell walls. Biocidal processes involving the nanoparticle uptake by microorganisms consist of several steps, such as adsorption of the particles on bacterial cell surface, penetration through the cell wall, binding to cytoplasmic membrane and its disruption, release of cytoplasmic contents, and, finally, cell death [26,45]. The antibacterial mechanism of cationic DMAEMA and hydrophobic TBAEMA polymers might be based on a similar principle like the electrostatic interaction between a cationic compound and negatively charged constituents on the cell surface and/or hydrophobic interactions [25,26,34,46]; however, these effects can be strain-specific due to the distinct cell wall structures. In contrast to the Gram-positive bacteria, the Gram-negative ones possess an extra outer membrane composed of phospholipids, proteins, and lipopolysaccharides, which generally enable stronger protection against agents migrating to the cytoplasm [46]. In addition, Gram-negative bacteria require less charged and hydrophobic polymers to permeabilize the outer membrane [26]. Antimicrobial

Conclusions
Highly efficient antibacterial magnetic nanoparticles were prepared via a thermal decomposition that was followed by their modification with Sipomer PAM-200 and coating with PDMAEMA or P(DMAEMA-TBAEMA) obtained by radical (co)polymerization of 2-

Conclusions
Highly efficient antibacterial magnetic nanoparticles were prepared via a thermal decomposition that was followed by their modification with Sipomer PAM-200 and coating with PDMAEMA or P(DMAEMA-TBAEMA) obtained by radical (co)polymerization of 2-(dimethylamino)ethyl methacrylate and/or 2-tert-butylaminoethyl methacrylate. It is a great advantage of Sipomer PAM-200 that it consists of both phosphate group strongly interacting with iron oxide and methacrylate containing vinyl group prone to radical polymerization. The particles were uniform in size with diameters reaching 16 nm. Magnetic properties of the cationic polymer-coated nanoparticles clearly reflected their superparamagnetic character, without signs of clustering. As a result, the particles were easily attracted by even low external magnetic field. The antimicrobial effects of MNP@S-D, MNP@S-D-T1, and MNP@S-D-T2 against two different kinds of bacteria, S. aureus and E. coli, were found to be concentration-and time-dependent. Antibacterial properties were enhanced with increasing concentration of the nanoparticles and time of incubation. In particular, the MNP@S-D-T1 even at low concentrations (44 µg/mL) exhibited a strong antibacterial effect on both microbial cultures. This makes these particles suitable as an efficient and reusable antibacterial agent.

Conflicts of Interest:
The authors declare no conflict of interest.