Bio-nano interactions: cellulase on iron oxide nanoparticle surfaces
- 283 Downloads
Iron oxide nanoparticles (IONs) may well represent the most promising magnetic nanostructures for a plethora of applications in health, life and environmental science. IONs are already used in medicine, catalysis and downstream processing of biotechnological products. Since most particles, utilized industrially, need expensive coatings, the application of bare nanoparticles seems economically worthwhile. In this study, three different ION species were synthesized by co-precipitation methods without stabilizing agents and were thoroughly characterized with a multi-analytical approach. We emphasize the importance of the particle characterization as transitions of the ION polymorphs into each other are possible as well as merging of distinct properties. The particle sizes, which here range from 10 to 30 nm, and the magnetic properties of IONs are crucial for the further application. The adsorption behavior of the enzyme cellulase (CEL) as a model protein is investigated on the different IONs in order to gain deeper insights into bio-nano interactions to different surface sites, charges, curvatures and morphologies, as given by the three applied adsorber materials. The protein-particle interactions are driven by electrostatic and hydrophobic forces in the case of CEL. The CEL adsorption follows a Langmuir behavior and does not exceed maximum loads of around 0.6 g g−1. IR spectroscopy gives insights into the orientation of bound CEL and indicates a stronger affinity for the β-sheet tertiary structure content while a higher load can be reached with a higher α-helix content.
KeywordsIron oxide nanoparticles Protein adsorption Enzyme immobilization Electrostatic interaction Cellulase Surface charge
Iron oxides accompany the progress of humankind since the beginning of the Iron Age when metallurgical processing was a political advantage. In a time with increasing interest in nanotechnologies, iron oxides still draw attention due to their magnetic properties. Magnetic separation emerged from the mining industry and was further developed for a huge variety of applications (Lu et al. 2007). They range from food processing and wastewater treatment (Drenkova-Tuhtan et al. 2013) to computational and medical applications such as data storage (Terris and Thomson 2005), biosensors (Graham et al. 2004), magnetic resonance imaging (Lee and Hyeon 2012), hyperthermia (Sonvico et al. 2005) and drug delivery (Mikhaylov et al. 2011). In biotechnological downstream processing, magnetic particles represent an interesting alternative for cascades of purification steps, such as liquid–liquid extraction, chromatography, filtration or centrifugation (Franzreb et al. 2006; Colombo et al. 2012; Fraga García et al. 2015). Furthermore, the applications of iron oxide nanoparticles as low-cost carrier material for catalysts with high reusability and recovery are very promising (Jesionowski et al. 2014; Rossi et al. 2014; Roth et al. 2016b). Therefore, we strive to investigate the adsorption of cellulase (CEL), an enzyme with the ability to crack the world’s most common polysaccharide cellulose to glucose (Bornscheuer et al. 2014), on different iron oxide nanoparticles. While most investigations focus on the catalytic activity, our aim is to understand protein adsorption on nanoscale surfaces and how to implement the differences in protein purification and immobilization taking advantage of tailored adsorption. Cellulase, which is industrially utilized in the sustainable degradation of biomass as feedstock for the bioethanol production, was chosen as a low-cost protein with distinct size and known electrostatic properties. The processing of iron oxide nanoparticles has been investigated extensively since the first approaches (Khalafalla and Reimers 1973, Massart 1981). Many different synthesis routes for iron oxides have been described in the last 15 years: microemulsions, sol–gel syntheses (Albornoz and Jacobo 2006), sonochemical reactions (Massart 1981), hydrothermal reactions (Wan et al. 2005), hydrolysis and thermolysis of precursors (Kimata et al. 2003), flow injection syntheses (Salazar-Alvarez et al. 2006), and electrospray syntheses. However, the alkaline co-precipitation of the iron salts ferric and ferrous chloride is among the cheapest and most commonly deployed. This route still provides a versatility for producing up to 16 distinct species of iron oxides, hydroxides and oxyhydroxides which can be formed depending on synthesis conditions (Ahn et al. 2012). In this study we want to investigate interaction with the thermodynamically most stable oxide polymorphs magnetite (Fe3O4), maghemite (γ-Fe2O3) and hematite (α-Fe2O3). The crystal systems vary from a hexagonal close packed unit cell of hematite with a trigonal crystal symmetry to maghemite and magnetite with a cubic close packed cell with a defective and inverse spinel crystal structure respectively (Cornell and Schwertmann 2003). We gained crystal information from X-ray diffraction and observed phonon modes by attenuated total reflection Fourier-transform infrared (ATR-IR) and Raman spectroscopy. The magnetic properties range from antiferromagnetically ordered hematite to a ferrimagnetic behavior in magnetite and maghemite. Differences in crystal structure can influence the surface properties of IONs as reported in the literature (Carlson and Kawatra 2013) and therefore an accurate nanoparticle characterization is essential for the interpretation of adsorption behavior. Thus, the surface charge of IONs is investigated by potentiometric titration. We discuss the relevance of the charge for the adsorption behavior of the biomolecule. According to the literature, adsorption of smaller molecules to IONs is strongly dependent on electrostatic interactions and on coordination of carboxyl groups (Schwaminger et al. 2015; Roth et al. 2016a). The focus of this investigation is to highlight the structure–property relationship of different IONs and to evaluate the role of the electrostatic interactions in the adsorption of larger biomolecules such as proteins. Therefore, we choose CEL as model molecule due to its industrial relevance. Moreover, we investigate the affinity of CEL to these three oxide surfaces at various buffer conditions to understand the impact of particle size, degree of oxidation and surface chemistry on adsorption.
Magnetite synthesis was carried out by the co-precipitation of Fe2+ and Fe3+ aqueous salt solutions in an alkaline environment as previously reported (Roth et al. 2015). Ferric chloride (FeCl3·6H2O) and sodium hydroxide (NaOH) were purchased from AppliChem GmbH, Germany. Ferrous chloride (FeCl2·4H2O) was purchased from Bernd Kraft GmbH, Germany. Ferric perchlorate nonahydrate was purchased from Sigma-Aldrich Co, USA. Aqueous solutions of ferrous chloride (100 mL, 1 mmol L−1), ferric chloride (100 mL, 2 mmol L−1) and sodium hydroxide (500 mL, 1 mmol L−1) were prepared with degassed and deionized water. The co-precipitation of magnetite nanoparticles was performed in a stirred tank reactor under nitrogen atmosphere to prevent oxidation of the precursors and the product.
Maghemite particles were synthesized under the same conditions and with the same precursors. The only difference was the ambient atmosphere during the reaction instead of nitrogen atmosphere and the use of merely deionized water. Furthermore, the particles were stored in 1 mmol L−1 HNO3 for 24 h in order to ensure a full oxidation.
Hematite particles were obtained by hydrolysis of a 1 L solution containing ferric perchlorate (17.8 g). The solution was held at 98 °C for 10 days without stirring.
The obtained particles were washed several times with double distilled water, until a conductivity below 200 µS cm−1 was reached. For magnetite particles the washing processes were conducted under nitrogen atmosphere. A colloidal sample was stored for transmission electron microscopy (TEM) and Mössbauer spectroscopy while the rest of the precipitates were lyophilized with an ALPHA 1-2LDplus (Martin Christ Gefriertrocknungsanlagen GmbH, Germany).
Crystal structure and phase purity of the lyophilized samples were examined with powder X-ray diffraction (XRD). The measurements were performed with a Stadi-P diffractometer (STOE & Cie GmbH, Germany), equipped with a molybdenum source [Ge (111) monochromator, Kα1 radiation (λ = 0.7093 Å)] and a Mythen 1K detector (DECTRIS Ltd., Switzerland) in transmission geometry. Data was collected in the range from 2° to 50° (2θ). The software package STOE WinXPOW (STOE & Cie GmbH, Germany) was used for indexing and refinement purposes.
Mössbauer spectra of frozen particle suspensions were recorded at 4.2 K. For this, the absorber and the source of 57Co in rhodium (ca. 25 mCi) were cooled in a liquid He cryostat. The electromechanical velocity transducer (Halder Electronics, Germany) was operated with a sinusoidal velocity waveform. The spectra were fitted with a superposition of Lorentzian lines grouped into sextets using the MOS90 software version 2.2. The fitted components often show broadened lines and have to be considered as representing distributions of magnetic hyperfine fields.
Raman spectra were recorded by a SENTERRA spectrometer (Bruker Optics GmbH, Ettlingen, Germany) equipped with a 488 nm laser. Low laser powers (1 mW) were chosen in order to prevent oxidation of the samples. A baseline correction was accomplished with the software OPUS 7.2 for spectra measured by Raman spectroscopy using the concave rubber band method.
Freeze-dried particles were also characterized by ATR-IR spectroscopy. Proteins adsorbed on the particles were investigated as water suspensions (mass ratio CEL:ION, 2:1). The Bruker Optics Tensor 27 spectrometer (Bruker Optics GmbH, Ettlingen, Germany) equipped with a Bruker Platinum ATR accessory (diamond crystal with one reflection) and a room temperature deuterated tri glycinesulfate (RT-DTGS) detector was used in all experiments. Each sample was recorded twice against air background (from 4000 to 400 cm−1 with a resolution of 4 cm−1), changing the position between two measurements on the ATR crystal, and averaged 64 times. For all spectra, an atmospheric compensation and an ATR correction (experimental ATR correction factor for magnetite 1.91, maghemite 1.88 and hematite 1.87) and a concave baseline correction was applied with the software OPUS 7.2. Also, a max/min normalization was accomplished without consideration of the CO2 region.
Magnetic properties of the precipitates were characterized with a superconducting quantum interference device (SQUID) magnetometer MPMS (Quantum Design Inc., USA) at a temperature of 300 K. Magnetic field was varied between −50 and +50 kOe.
The particle dimensions were assessed by transmission electron microscopy (TEM) using a JEM 100-CX (JEOL GmbH, Germany). For the TEM measurements the colloidal samples were diluted in degassed and deionized water, ultrasonicated to disperse any agglomerates and precipitated on carbon coated copper grids (Quantifoil Micro Tools GmbH, Germany). The pictures were manually processed in ImageJ. For each sample, 100 particles were measured in random order.
The specific surface area was determined by the BET method with a Gemini VII 2390 Surface Area Analyzer measuring nitrogen adsorption isotherms at 77 K.
The nanoparticles’ point of zero charge (PZC) was determined by potentiometric titrations which were accomplished in an OptiMax™ reactor (Mettler-Toledo GmbH, Germany) from pH 4 to 10. The degassed particle suspensions were adjusted to a concentration of 2 g L−1 and equilibrated at a pH of 4 overnight. The whole titration was conducted under nitrogen atmosphere at 298.5 K with HCl and NaOH as titrands and a NaCl concentration of 100 mmol L−1.
Prior to further analysis and activity measurements, the mixtures were washed in the respective buffer.
3 Results and discussion
Crystal structure, particle size as well as oxidation state can influence bio-nano interactions. Therefore, a thorough particle characterization is required in order to correctly interpret structure–property relationships. In this context, we synthesized IONs with similar particle shapes and sizes for the presented study. The average particle diameter of the synthesized hematite is larger than particle diameters of magnetite and maghemite due to the low thermodynamical stability of hematite in the lower nanoscale range. However, we wanted to ensure similar particle properties by employing wet-chemical syntheses and waiving the use of stabilizers. XRD is used to determine the different crystal structures and represents an indicator for particle sizes. The X-ray diffractograms of IONs (Fig. 1) show well defined diffraction patterns without impurities for the three iron oxide particles. While the cubic structures of maghemite and magnetite yield very similar diffraction patterns, the trigonal crystal system of hematite yields a clearly different diffractogram (Cornell and Schwertmann 2003). The Scherrer equation yields the primary particle size from reflection broadening of the IONs which corresponds to 25 nm for hematite, 9 nm for maghemite and 10 nm for magnetite. The reflections corresponding to the maghemite crystallites are slightly shifted to higher diffraction angles compared to magnetite crystallites. This behavior can be explained by the defective spinel structure of maghemite connected with a shift of the reflections and a corresponding change of the lattice constant from 8.40 to 8.36 Å (Goss 1988; Kim et al. 2012). Furthermore, the planes 210 and 211 which are scarcely visible indicate the existence of maghemite. The polymorphs maghemite and hematite are not visible in the respective other diffractogram. The small particle sizes impede the differentiation between structure and size owed defects in the crystalline structure (Baaziz et al. 2014).
Mössbauer spectroscopy at 4.2 K is an efficient tool to distinguish between magnetite and maghemite as ferric and ferrous ions demonstrate different hyperfine fields on the octaeder gaps of the spinel structure. Figure 2 indicates a Fe2+ share of more than 30% for magnetite while no Fe2+ can be observed for the other iron oxide species. No divalent iron ions with hyperfine fields between 47 and 49 T are visible in the second spectrum. Maghemite demonstrates similar hyperfine fields around 52 T as reported for the Fe3+ species of magnetite at 4.2 K (Berry et al. 1998; Oh et al. 1998). The larger magnetic hyperfine field component of hematite (54 T) compared to maghemite reveals an antiferromagnetic state of hematite which is well known for bulk hematite below the Morin transition temperature (Oh et al. 1998; Bødker and Mørup 2000). Since the Morin transition is often suppressed in nano-crystalline hematite, at 4.2 K a weakly ferromagnetic component (30%) represented by a slightly weaker field (53 T) is still visible. Mössbauer spectroscopy reveals an insight into the magnetic splitting of atom cores and indicates the magnetic field generated by surrounding electrons. Hence, this technique allows a prediction of the magnetic properties of materials.
For separation processes, the magnetic properties of IONs can play a crucial role and therefore should be considered for adsorption studies. Here, the ferrimagnetic magnetite and maghemite nanoparticles have quite high saturation magnetizations of 80 and 67 emu g−1, respectively (Fig. 3). These values are in good agreement with the saturation magnetizations of particles produced by co-precipitation syntheses reported in literature (Fang et al. 2012; Roth et al. 2016b) and quite close to bulk properties (Cornell and Schwertmann 2003). The magnetization curves of both iron oxides demonstrate superparamagnetic behavior, showing almost no (≤1 emu g−1) remanence (Lu et al. 2007). Here, maghemite shows less remanence than magnetite which sometimes is reported otherwise in literature and might result from the slightly smaller primary crystallite size. The superparamagnetic behavior is reported for iron oxide nanoparticles with diameters below 20 nm (Kolhatkar et al. 2013). No hysteresis can be observed for the weakly ferromagnetic hematite nanoparticles which as well differ from the bulk material (Bødker et al. 2000). The saturation magnetization was not reached with an external magnetic field of 50 kOe. The maximum magnetization we were able to determine was 1.5 emu g−1, which prevents the application of these hematite particles for magnetic separations but indicates partial ferromagnetic behavior.
Raman scattering is another interesting technique to differentiate between different IONs due to their phonon modes. Among the investigated IONs, hematite demonstrates the most distinct scattering as its hexagonal crystal structure features an inversion center. Here, seven peaks can be observed and assigned to the different phonon modes (Kim et al. 2005) of hematite (Fig. 4a). While the peaks observed at 226 and 494 cm−1 can be assigned to the A1g modes (Chamritski and Burns 2005) four peaks at 245, 293, 410, and 612 cm−1 can be assigned to the Eg modes (Chamritski and Burns 2005). The peak at 663 cm−1 can be either assigned to a Raman forbidden infrared active vibration (Bersani et al. 1999) which might be active for disordered nano-sized systems or to magnetite/maghemite impurities. Compared to other reported bands for hematite, we observed a broad full width at half maximum (FWHM) and a blue shift which can be assigned to the phonon confinement of nanoparticles.
The polymorph maghemite demonstrates weaker scattering behavior than hematite due to the lack of an inversion center and a nano-sized defect spinel structure. Maghemite shows strong phonon confinement with large peak broadening of the bands at 688, 507 and 360 cm−1, which can be accounted to the Raman active modes A1g, Eg and Tg, respectively (Jubb and Allen 2010). The peak broadening is an indicator for the dimensions of the IONs and the defective structure of maghemite. The inverse spinel structure of magnetite is shown by the strong peak at 669 cm−1 which can be assigned to an A1g mode. This mode is much smaller compared to maghemite. Therefore, a surface oxidation can hardly be noticed but a broadening is still on hand compared to literature values for bulk magnetite. The Tg modes can scarcely be observed at 534 and 303 cm−1.
Infrared spectroscopy can be utilized to investigate surface defects of nanoparticles. Here, the broadening of peaks corresponding to hematite’s Eu (384 and 554 cm−1) and A2u (474 cm−1) modes can be compared to literature and can indicate an inhomogeneous structure of particles due to their nanometer range (Fig. 4b). The modes are also slightly shifted compared to bulk values of hematite which is a further indicator for an inhomogeneous crystal structure. A weak shoulder at about 690 cm−1 is typical for poorly crystalline iron hydroxides and oxyhydroxides (Chernyshova et al. 2010). The maghemite spectrum shows different T1u modes (440, 564, 582, 637, 688, 730 cm−1) (Santoyo Salazar et al. 2011) which indicate the defective and distorted spinel structure. In contast to magnetite the bands which can be attributed to surface and hydroxyl modes (688, 730 cm−1) are more pronounced compared to the respective main Fe–O vibration for spinel structures [~580 (575) cm−1]. Moreover, magnetite shows a smaller band at 444 cm−1 and less peak broadening at 580 cm−1 than maghemite. Surface modes around 690–730 cm−1 are observable for magnetite which can be related to surface oxidation, surface defects and surface hydroxylation (Gotic et al. 2009).
The Point of Zero Charge (PZC) varies for the three iron oxides (Fig. 5). Magnetite and maghemite demonstrate PZCs between 7.8 and 8, which is in good agreement with literature (Cornell and Schwertmann 2003). Zeta potential measurements for similar maghemite particles indicate connatural isoelectric points around pH 7.5 (Gdula et al. 2016). Hematite has its PZC at 5.4 which is often described for hematite soils. The charging behavior with different ionic strengths is typical for nanoparticles in this range (Illés and Tombácz 2003). However, the electrokinetic surface properties can be changed by different ions in the Stern layer (Vereda et al. 2015). As the surface charge corresponds to adsorbed ions, the amount of binding sites can be calculated. The binding sites of magnetite and maghemite were determined by the proton adsorption capacity at pH 4 to be around 4.5 and 3.2 sites nm−2, respectively. The here synthesizes hematite nanoparticles exhibit 4.8 sites nm−2 calculated at pH 11. These values for iron oxide nanoparticles are in good agreement with literature values. (Sun et al. 1998; Tombácz et al. 2007).
TEM micrographs of nanoparticles slightly differ from the XRD diameters but evidence the same size trend (Fig. 6). Hematite demonstrates the largest particle distribution with an average diameter of 34 ± 5 nm. Maghemite particles are 11 ± 3 nm in average while magnetite particles reach a size of 12 ± 3 nm in average. This size distribution fits well the observed BET surfaces for the particles 44, 153 and 99 m2 g−1 for hematite, maghemite and magnetite, respectively. These values were used for the normalization of adsorption isotherms to an adsorbed amount of CEL per surface area of particles.
The precisely characterized IONs were dispersed in different buffer solutions and the effect of surface properties of the three polymorphs was investigated toward the adsorption of CEL at 25 °C. At a pH of 5 in 50 mmol L−1 sodium acetate buffer, CEL adsorbes differently to the three iron oxide surfaces (Fig. 7 a). The highest load qmax (0.73 g g−1) can be observed for maghemite while the maximum load of CEL on magnetite is 0.57 g g−1. This load is significantly higher than the load observed for the adsorption of CEL on IONs at 50 °C (Roth et al. 2016b). The discrepancy between magnetite and maghemite can be explained by the higher specific surface area of maghemite while both IONs demonstrate similar surface charges. Presumably, the slightly higher PZC for magnetite is responsible for a higher propensity of CEL to adsorb on the surface indicated by a lower equilibrium constant Kd corresponding to 0.26 g L−1 for magnetite and 0.35 g L−1 for maghemite, respectively. The Kd for magnetite at 25 °C is higher than the Kd at 50 °C indicating a less favorable adsorption equilibrium at lower temperatures (Roth et al. 2016b). Hematite demonstrates significantly lower loads of cellulase (0.10 g g−1) while the half maximum is reached earlier at 0.12 g L−1. Magnetite and maghemite surfaces are charged in the sodium acetate buffer at pH 5 and therefore electrostatic interactions might be possible with the CEL near its isoelectric point at 4.9 (Kudina et al. 2014).
At a pH of 8.5 all particles should be negatively charged. Here, hematite demonstrates the highest qmax (0.47 g g−1) while less CEL adsorbes on magnetite and maghemite (0.39 and 0.32 g g−1) (Fig. 7b). This behavior might be explained by the stronger electrostatic interactions between hematite and CEL at these conditions though a mass transport limitation due to agglomerating particles cannot be excluded (Fraga García et al. 2014). Moreover, CEL binding on magnetite and maghemite demonstrate the least favorable adsorption equilibrium in TRIS buffer (Kd 0.38 and 0.37 g L−1) while the Kd of hematite is lower in TRIS (0.08 g L−1) than in sodium acetate buffer. At pH of 6 with a salt concentration of 100 mmol L−1 sodium chloride, the three iron oxide nanoparticles behave similarly. Here, the Kd values range from 0.02 to 0.04 to 0.06 g L−1 for hematite, magnetite and maghemite respectively, demonstrating the highest affinity towards CEL adsorption on all IONs (Fig. 7c). The maximum protein load is similar for CEL on all IONs in relation to the particles’ specific surface area (0.11 g g−1 hematite; 0.32 g g−1 magnetite and 0.50 g g−1 maghemite).
In order to verify the influence of electrostatic interactions on the adsorption of CEL on IONs, we investigated the influence of the NaCl concentration from 0.01 to 1 M (Fig. S1). While the thermodynamic adsorption equilibrium is favored with lower salt concentrations, the maximum load increases with higher concentrations on magnetite (Table S2). As high ion concentrations in the double layer decrease electrostatic interactions at the bio-nano interface, lower protein loading values would be expected. The observed behavior indicates an electrostatic adsorption which is coupled with other adsorption mechanisms such as hydrophobic interactions. Moreover, the adsorption kinetic is very fast as we observed the adsorption equilibrium of CEL on magnetite to be reached after 1 min of incubation (Fig S3). The adsorption behavior of CEL on ION can be compared to the adsorption of the protein bovine serum albumin (BSA) to inorganic nanoparticles which has a similar isoelectric point (4.7) and has been described in literature (Rezwan et al. 2004; Yu et al. 2016). For BSA, the nanoparticles’ surface charge influences the adsorption equilibrium and the amount of adsorbed protein can be directly correlated to the pH (Peng et al. 2004; Li et al. 2006). From the maximum CEL load, the size and density of nanoparticles and the molecular weight of CEL (Fig. S3), the number of CEL molecules per nanoparticle (Table 1) can be calculated (SI).
The number of molecules per nanoparticle is comparable to other investigations on protein adsorption on nanoparticle surfaces.(Yu et al. 2016) Furthermore, we were able to verify the activity of adsorbed CEL (SI) which was significantly higher on maghemite than on magnetite nanoparticles. The specific activity of CEL on magnetite is in good agreement with literature values for magnetite nanoparticles while the specific activity of CEL on maghemite is higher and similar to CEL on silica support (Roth et al. 2016b). This behavior indicates an inhibition of cellulase activity by divalent iron ions (Tejirian and Xu 2010). CEL adsorbed in TRIS buffer still demonstrates an enzyme activity after being rebuffered in sodium acetate for the pNP assay which emphasizes the advantages of adsorption and the respective increasing resistence of enzymes against pH shifts (Serefoglou et al. 2008; Jordan et al. 2011).
Overview on adsorption parameters and maximum loading as well as specific cellulase activity at different buffer conditions and nanoparticles
Kd (g L−1)
qmax (g g−1)
Activity (U g−1)
Kd (g L−1)
qmax (g g−1)
Activity (U g−1)
Kd (g L−1)
qmax (g g−1)
The infrared spectra of CEL on the different IONs confirm the protein adsorption and give deeper insights towards the binding mechanism. For all spectra the amide band I corresponding to the the C=O stretch vibration and the amide II band corresponding to the N–H bend and C–N stretch modes can be observed around 1640 and 1540 cm−1, respectively (Morhardt et al. 2014; Roth et al. 2016b). The band around 1411 cm−1 represents a COO− stretching mode corresponding to carboxylic groups in the protein or acetate buffer (Fig. 8a). Hence, this band is larger for the CEL adsorbed on the iron oxide in sodium acetate buffer (Morhardt et al. 2014; Schwaminger et al. 2015). The large band around 1550 cm−1 can be explained by an overlap of the vibrations corresponding to the amide II band and the COO− vibration of the buffer bound to the surface. The complexation of acetate molecules on the iron oxide surfaces can furthermore lead to a shift and broadening of this band comparable to citrate adsorption which is widely investigated (Situm et al. 2016). The existence of acetate ions on the surface indicates the possibility of electrostatic interactions with CEL even at its IEP. The TRIS buffer can be observed to be adsorbed on the IONs’ surfaces as well (Fig. 8b). Here, the C–H bending vibration around 1480 cm−1 and the C–OH bending vibrations around 1040 cm−1 indicate TRIS in the IR spectrum. Furthermore, the amide band I shifts to a lower wave number indicating an interaction between the TRIS buffer and the protein. As the C–OH vibration is observable on the CEL adsorbed on IONs in sodium chloride solution as well, a contribution of the CEL glycosilylations to the adsorption should be accounted (Khajehpour et al. 2006; Roth et al. 2016b).
Especially in Fig. 8c the amide I band is very pronounced which allows to investigate the secondary structure of the peptide backbone (Pavlidis et al. 2012). The deconvolution is described in the SI and renders different shares of β-sheet and α-helices with different buffer systems. Interestingly, CEL adsorbed in NaCl demonstrates the highest β-sheet content while the maximum load is lower or equally low than in other buffer systems (Table S6). This might indicate a lower loading of IONs due to different CEL folding as β-sheets usually exhibit larger surfaces than α-helices. Furthermore, the larger CEL surface might lead to higher affinities with IONs and therefore leads to lower Kd values. Hence CEL folding is an important factor for the adsorption beside electrostatic interactions. Investigations towards BSA adsorption demonstrate large β-sheet content for hydrophobic adsorption (Roach et al. 2005). Accordingly, CEL in TRIS and sodiumacetate buffer show a similar lower β-sheet content and an increased α-helices content and demonstrate comparable adsorption constants and indicate polar and electrostatic interactions (Roach et al. 2005). Maghemite nanoparticles show a pecularity of higher β-turn contents than magnetite, which might explain the higher enzymatic activity (Table 1).
Pure iron oxide nanoparticles were synthesized and thoroughly characterized towards their physical properties. The particle sizes range between 10 and 30 nm and the crystal structure was determined to be cubic for magnetite and maghemite and trigonal for hematite. A systematic characterization of employed materials was necessary to assign surface properties to distinct crystal properties. Crystal distortion can be detected for all IONs by ATR-IR and Raman spectroscopy. From these surface defects of the IONs different local and overall surface charges derive that lead to diverse adsorption behavior of CEL. While a different crystal structure was evidenced for magnetite and maghemite, the adsorption properties as well as the surface charge are quite similar. By changing the pH and the surrounding buffer, different amounts and different affinities for the adsorption to the IONs can be observed. The surface charge of iron oxide particles can easily be tuned by buffer ions. We demonstrate the possibility of tailoring CEL adsorption behavior on structurally different iron oxides with different surface charges. The attraction between an oppositely charged surface and CEL is stronger than for lesser or similarly charged particles and therefore a higher enzyme load can be achieved (Schultz et al. 2008). Beside electrostatic interactions other adsorption mechanisms occur for CEL on nanoparticles, as verified by the higher loads for higher sodium chloride concentrations.
The protein secondary structure as well plays an important role on the adsorption of CEL on IONs and influences the amount of adsorbed molecules. The maximum CEL load is limited to around 0.6 g g−1 CEL per ION which is similar to the literature values of BSA and often explained by a monolayer arrangement of the protein on the surface (Peng et al. 2004; Rezwan et al. 2004; Yu et al. 2016).
Further experiments are necessary for a better understanding of protein adsorption of on IONs. Especially the protein coordination in the adsorbed state and identification of binding domains are fundamental to gain deeper insights into the mechanisms at the bio-nano interface. Iron oxide nanoparticles represent an interesting tool for bioseparation and adsorption processes due to their magnetic and surface properties. The enzyme immobilization on support materials is of great scientific and industrial interest with incresing demands for biotechnological processes. CEL adsorption on iron oxide nanocarriers depicts a low-cost recyclable assembly which can be employed in the enzymatic degradation of cellulose in biomaterials such as straw lysate and bagasse.
The authors would like to express their gratitude to Prof. Dr. Tom Nilges for his support with powder XRD (TU München) and Dr. Peter Weidler for valuable discussions (Karlsruhe Institute of Technology, Institute of Functional Interfaces, Germany). Furthermore, we would like to express our very great appreciation to Dr. Marianne Hanzlik for help with TEM measurements and Stefan Darchinger for the performance of gel electrophoresis. Moreover, we are particularly grateful for the financial support of this work by the Federal Ministry of Education and Research (Grant number 031A173A) and the Bavarian Ministry of Economic Affairs and Media, Energy and Technology (Grant number 1340/68351/3/11).
- Baaziz, W., Pichon, B.P., Fleutot, S., Liu, Y., Lefevre, C., Greneche, J.-M., Toumi, M., Mhiri, T., Begin-Colin, S.: Magnetic iron oxide nanoparticles: reproducible tuning of the size and nanosized-dependent composition, defects, and spin canting. J. Phys. Chem. C. 118(7), 3795–3810 (2014)CrossRefGoogle Scholar
- Berry, F.J., Skinner, S., Thomas, M.F.: Mössbauer spectroscopic examination of a single crystal of Fe3O4. J. Phys. 10(1), 215–220 (1998)Google Scholar
- Drenkova-Tuhtan, A., Mandel, K., Paulus, A., Meyer, C., Hutter, F., Gellermann, C., Sextl, G., Franzreb, M., Steinmetz, H.: Phosphate recovery from wastewater using engineered superparamagnetic particles modified with layered double hydroxide ion exchangers. Water Res. 47(15), 5670–5677 (2013)CrossRefGoogle Scholar
- Goss, C.J.: Saturation magnetisation, coercivity and lattice parameter changes in the system Fe3O4-γFe2O3, and their relationship to structure. Phys. Chem. Miner. 16(2) (1988)Google Scholar
- Gotić, M., Koščec, G., Musić, S.: Study of the reduction and reoxidation of substoichiometric magnetite. J. Mol. Struct. 924–926, 347–354 (2009)Google Scholar
- Khalafalla, S., Reimers, G.W.: Magnetofluids and their manufacture. USA US3764540 A, 09.10.1973Google Scholar
- Mikhaylov, G., Mikac, U., Magaeva, A.A., Itin, V.I., Naiden, E.P., Psakhye, I., Babes, L., Reinheckel, T., Peters, C., Zeiser, R., Bogyo, M., Turk, V., Psakhye, S.G., Turk, B., Vasiljeva, O.: Ferri-liposomes as an MRI-visible drug-delivery system for targeting tumours and their microenvironment. Nat. Nanotechnol. 6(9), 594–602 (2011)CrossRefGoogle Scholar
- Roth, H.-C., Schwaminger, S., Fraga García, P., Ritscher, J., Berensmeier, S.: Oleate coating of iron oxide nanoparticles in aqueous systems: the role of temperature and surfactant concentration. J. Nanopart. Res. 18(4) (2016a)Google Scholar
- Santoyo Salazar, J., Perez, L., de Abril, O., Truong Phuoc, L., Ihiawakrim, D., Vazquez, M., Greneche, J.-M., Begin-Colin, S., Pourroy, G.: Magnetic iron oxide nanoparticles in 10–40 nm range: composition in terms of magnetite/maghemite ratio and effect on the magnetic properties. Chem. Mater. 23(6), 1379–1386 (2011)CrossRefGoogle Scholar
- Serefoglou, E., Litina, K., Gournis, D., Kalogeris, E., Tzialla, A.A., Pavlidis, I.V., Stamatis, H., Maccallini, E., Lubomska, M., Rudolf, P.: Smectite clays as solid supports for immobilization of β-glucosidase: synthesis, characterization, and biochemical properties. Chem. Mater. 20(12), 4106–4115 (2008)CrossRefGoogle Scholar
- Situm, A., Rahman, M.A., Goldberg, S., Al-Abadleh, H.A.: Spectral characterization and surface complexation modeling of low molecular weight organics on hematite nanoparticles: Role of electrolytes in the binding mechanism. Environ. Sci. 3(4), 910–926 (2016)Google Scholar
- Sonvico, F., Mornet, S., Vasseur, S., Dubernet, C., Jaillard, D., Degrouard, J., Hoebeke, J., Duguet, E., Colombo, P., Couvreur, P.: Folate-conjugated iron oxide nanoparticles for solid tumor targeting as potential specific magnetic hyperthermia mediators: synthesis, physicochemical characterization, and in vitro experiments. Bioconjugate Chem. 16(5), 1181–1188 (2005)CrossRefGoogle Scholar
- Tombácz, E., Illés, E., Majzik, A., Hajdú, A., Rideg, N., Szekeres, M.: Ageing in the inorganic nanoworld: example of magnetite nanoparticles in aqueous medium Croat. Chem. Acta 80(3–4), 503–515 (2007)Google Scholar