Bioadsorption of proteins on large mesocage-shaped mesoporous alumina monoliths

Colloids and Surfaces B: Biointerfaces 103 (2013) 288–297

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

Bioadsorption of proteins on large mesocage-shaped mesoporous alumina monoliths Sherif A. El-Safty a,b,∗ , M.A. Shenashen a,b,1 , M. Khairy a,b a b

National Institute for Materials Science, 1-2-1 Sengen, Tsukuba-shi, Ibaraki-ken 305-0047, Japan Graduate School for Advanced Science and Engineering, Waseda University, 3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan

a r t i c l e

i n f o

Article history: Received 15 May 2012 Received in revised form 1 October 2012 Accepted 5 October 2012 Available online 2 November 2012 Keywords: Mesocage alumina Protein Langmuir adsorption isotherms Bioadsorbents Theoretical models

a b s t r a c t With the remarkable progress in the field of gene technology, proteins have gained an important function in the field of disease diagnosis and treatment. Protein bioadsorption has drawn increasing attention partly because of the promising advances for diagnostic assays, sensors, separations, and gene technology. Mesocage alumina has a cage-type structure with high surface area and pore volume, exhibiting superior capabilities for protein adsorption. In this study, we report the size-selective adsorption/removal of virtual proteins having different shapes, sizes, functions, and properties, including insulin, HopPmaL domain, lysozyme, galectin-3, ␤-lactoglobulin, ␣-1-antitrypsin, ␣-amylase, and myosin in aqueous water using mesocage alumina. The mesoporous alumina monoliths have unique morphology and physical properties and enhanced protein adsorption characteristics in terms of sample loading capacity and quantity, thereby ensuring a higher concentration of proteins, interior pore diffusivity, and encapsulation in a short period. Thermodynamic analysis shows that protein adsorption on mesocage alumina monoliths is favorable and spontaneous. Theoretical models have been studied to investigate the major driving forces to achieve the most optimal performance of protein adsorption. The development of ultra- or micrometer-scale morphology composed of mesocage-shaped mesoporous monoliths or alumina network clusters can be effectively used to encapsulate the macromolecules into the interior cage cavities, which can greatly assist in other potentials for biomedical applications. Furthermore, the adsorption of a single protein from mixtures based on size- and shape-selective separation can open up new ways to produce micro-objects that suit a given protein encapsulation design. © 2012 Elsevier B.V. All rights reserved.

1. Introduction The great success in genomics has led to an urgent demand for biological protein adsorbents. The adsorption of biomolecules, such as proteins, on inorganic substrate materials and their determination in trace amounts, is greatly useful in various applications, including diagnostic assays, gene technology, cancer diagnosis, allergy tests, biosensors, protein separation, drug/gene delivery systems, and pharmaceutical sciences [1–11]. Proteins are not homogeneous particles based on the protein structure, indicating that not all bonded parts are equally effective in adsorption. The presence of protein at the interface is assumed to be driven solely by diffusion processes. Diffusion is thermodynamically favorable

∗ Corresponding author at: National Institute for Materials Science (NIMS), 12-1 Sengen, Tsukuba-shi, Ibaraki-ken 305-0047, Japan and Graduate School for Advanced Science and Engineering, Waseda University, 3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan. E-mail addresses: [email protected], [email protected] (S.A. El-Safty). 1 Permanent Address (M.A. Shenashen): Egyptian Petroleum Research Institute (EPRI), Nasr City, Cairo, Egypt. 0927-7765/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.colsurfb.2012.10.040

because some of the contraction and hydration energies of proteins are lost at the interface [12–14]. The discovery of insulin (INS) in 1922 marked the beginning of research and development to improve the means of delivering protein therapeutics to patients. Thus, numerous research activities have been spent in designing new drugs to counter various diseases. Protein adsorption can occur through hydrophobic, electrostatic, and hydrogen-bonding interactions [15]. Protein adsorption on inorganic surfaces has fundamental biological importance. Numerous studies have used several nanostructures for the target-separation of biomolecules from biosamples, such as nickel oxide, multifunctional magnetic nanorods, different cores/shells, and porous materials [5–9]. Furthermore, biomedical and nanotechnological applications have increasingly used interfaces between inorganic material and polypeptides in various scientific fields, such as nanobioscience, materials science, artificial implants, protein-purification strategies, biosensors, drug delivery systems, catalysis, catalysis support, and molecular biology/biotechnology [16]. The high surface area, as well as the tunable and uniform pore size, causes mesoporous materials to become versatile hosts for numerous guest molecules, such as proteins, drugs, and small

S.A. El-Safty et al. / Colloids and Surfaces B: Biointerfaces 103 (2013) 288–297

biomolecules. Thus, mesoporous alumina is one of the most important and extensively used non-siliceous mesoporous materials offering a wide range of applications [17–21]. The first successful synthesis of ordered mesoporous alumina was achieved using long-chain carboxylic acids as templates in alcoholic solvents with low molecular weight [22]. The mesostructure and porosity of alumina have a dominant function in their activity and/or affinity for a specific application [23–26]. Garcia et al. [27] indicated that the mesostructure and high surface area make ␥-alumina an efficient substrate for l-alanine adsorption. Nevertheless, the new design of mesostructure frameworks showing the ability of macromolecule immobilization and uptake into the interior mesoporosity remains challenging. For instance, mesofilters based on hexagonal cylinder or cage silica nanotubes inside the anodic alumina membrane channels for a promising size-exclusion separation of proteins have been fabricated [28,29]. However, the preparation of actively functional bioadsorbent material with uniform and large mesopores is a key requirement for the successful encapsulation of proteins and bioanalysis to control the size-selective adsorption of high molecular weight proteins without remaining attached to the adsorbent surfaces [30–32]. In this study, we report a one-pot, simple fabrication of mesocage alumina monoliths using Brij 58 and cetyl trimethyl ammonium bromide (CTAB) templates. Preparing a highly active and stable adsorbent is a key requirement for a successful application. This study further reveals a novel biomimetic strategy to design bioadsorbent materials. This strategy is based on the control of surface topographic feature construction, among others. The alumina mesoporous matrices and mesocage cavities enable highperformance selective encapsulation of biomolecules [INS, HopPmaL domain (HD), lysozyme (LYS), galectin-3 (GA3), ␤-lactoglobin (␤-LG), ␣-1-antitrypsin (␣-ATR), ␣-amylase (␣-AMY), and myosin (MYS)] with different shapes, sizes, functions, and properties. This manuscript also addresses the prominent factors affecting adsorption. The reusability of bioadsorbent monoliths for several reuse cycles of adsorption assays is of particular interest. The good immobilization and uptake assays on the mesoporous materials provide a good indication of its biological field applications, such as protein encapsulation, bioanalysis, and drug delivery systems. 2. Experiments 2.1. Chemicals All materials were used without further purification. Polyoxyethylene (20) cetyl ether (Brij 58, C16 H33 (OCH2 CH2 )20 OH, MW. av = 1124), alkyltrimethylammonium bromide CTAB, aluminum nitrate enneahydrate Al(NO3 )3 ·9H2 O, insulin (INS, 5.733 kDa, 2.4 nm), HopPmaL domain (HD, 11,085 kDa, 2.9 nm), lysozyme (LYS, 14,300 kDa, 3.2 nm), galectin-3 (GA3, 16,043 kDa, 4.0 nm), ␤-lactoglobin (␤-LG, 18.4 kDa, 4.2 nm), ␣-1-antitrypsin (␣-ATR, 47.619 kDa, 5.9 nm), ␣-amylase (␣-AMY, ∼54 kDa, 6.8 nm), and myosin (MYS, 200–500 kDa, ∼14–19 nm) were obtained from Sigma–Aldrich Company Ltd. USA. Dodecane was obtained from Wako Company Ltd. Osaka, Japan. 2.2. Synthesis of mesoporous alumina bioadsorbent The stirring-assisted approach based on direct template synthesis of the quaternary microemulsion liquid crystalline phases of Brij 58 or CTAB/C12 -alkane/Al(NO3 )3 /H2 O–HCl/ethanol composition was used to fabricate tunable mesocage alumina monoliths. Controlled mesoporous alumina bioadsorbents were synthesized with different concentrations of CTAB and Brij 58. For the direct synthesis of mesoporous alumina bioadsorbent CTAB-1, the reactant weight

289

of Al(NO3 )3 , CTAB, dodecane, HCl solution (pH 1.3), and ethanol were 5, 1, 0.5, 2.5, and 10 g, respectively. The precursor solution was stirred for 5 min to form a homogenous sol–gel solution. Different forms of mesoporous CTAB-1, CTAB-2, CTAB-3, Brij-1, Brij-2, and Brij-3 were prepared using 1 g of CTAB, 1.5 g of CTAB, 2 g of CTAB, 1 g of Brij-58, 1.5 g of Brij-58, and 2 g of Brij-58, respectively. The contents of other materials were kept constant. The liquid viscosity of the material increased as the hydrolysis/condensation reactions continued, with stirring for 2 h. The resulting optical gellike material acquired the shape and size of the reaction vessel. Adding the swelling agent C12 -alkane and co-surfactant ethanol under stirring-assisted synthesis led to the formation of large, open mesostructures. The solid alumina/surfactant monoliths were completely dried at 40 ◦ C overnight. The organic moieties were removed by calcination at 550 ◦ C in air for 5 h. 2.3. Batch-contact bioadsorption of proteins The batch adsorption of various protein types into the mesoporous alumina bioadsorbent monoliths (0.05 g) was performed in an aqueous solution (30 mL) subjected to constant stirring at different temperatures (20–35 ◦ C, ±0.1 ◦ C). The proteins used were INS, HD, LYS, GA3, ␤-LG, ␣-ATR, ␣-AMY, and MYS. The initial concentration of the proteins ranged from 5 × 10−5 mol/L to 1 × 106 mol/L. Then, the aliquot protein was collected and monitored as a function of exposure time. The concentrations of LYS and ␤-LG proteins were studied through fluorescence spectroscopy (Perkin-Elmer LS45) at emission values of 350 and 342 nm, with excitation wavelengths at 292 and 290 nm, respectively. The concentrations of INS, LYS, ␣-ATR, ␣-AMY, and MYS proteins were studied through UV–vis spectroscopy (Shimadzu 3700 model solid-state UV–vis spectrophotometer) at  values of 277, 409, 405, 252, and 267 nm, respectively. The decrease in UV–vis and fluorescence spectra of all proteins at specific wavelengths indicates the adsorbed amount of proteins inside the mesocage alumina bioadsorbent within these adsorption processes. In the experimental assays, several adsorption measurements (n ≥ 3) were conducted using a wide range of low-/high-molecular weight proteins under specific batch-contact adsorption conditions. Therefore, the standard deviation of the protein adsorption assays was 0.3–0.6%, as proven by the fitting plot of adsorption graphs. The protein/alumina was washed at least thrice with water to remove the loosely attached protein molecules to characterize the bioadsorbent material after protein adsorption assays. The composite was dried under vacuum conditions at room temperature and the sample was collected for N2 adsorption/desorption measurements. A simple chemical treatment based on acidified solution (HCl/NaCl) was used to study the reusability of mesocage alumina adsorbents and effectively remove the trapped proteins in the monolith pores without causing significant damage in the textural and physical properties of the mesostructures. Our experimental findings showed that proteins with high molecular size and weight took a longer time to be removed from the monolith interior during cleaning. 2.4. Controlled adsorption parameters 2.4.1. Adsorption capacity and isotherms To study the adsorption affinity, the adsorption amount of biomolecules (INS, HD, LYS, GA3, ␤-LG, ␣-ATR, ␣-AMY, and MYS proteins) at the equilibrium step was calculated using the following equation: qe =

(Co − Ce )V m

(1)

290

S.A. El-Safty et al. / Colloids and Surfaces B: Biointerfaces 103 (2013) 288–297

where qe is the adsorption capacity at equilibrium (mol g−1 ); Co and Ce are the initial and equilibrium protein concentrations (M), respectively; V is the volume (L) of the aqueous solution; and m is the mass (g) of the bioadsorbent used in the experiments. To investigate the adsorption isothermal assays, the adsorption characteristics of the protein molecules onto the adsorbents can be studied by the following Langmuir isotherm equation [31–33]:

1

1 Ce = + qe KL qm

qm

Ce

(2)

where qm (mmol g−1 ) is the amount of protein molecules adsorbed to form a monolayer coverage, and KL is the Langmuir adsorption equilibrium constant. From the plot of Ce /qe against Ce , qm and KL can be determined from the slope and intercept. 2.4.2. Adsorption diffusivity The intraparticle diffusion of proteins into mesocage alumina bioadsorbents can be determined by plotting the fractional attainment of equilibrium (fe ) against t1/2 of bioadsorbents at different sizes and shapes of proteins, according to Fick’s second law relationship [34–37]: √ 6 Dt fe = (3) r  where (fe ) = qt /qs is the ratio of adsorbed quantity of protein molecule at time (t) to the amount adsorbed at saturation time, r is the pore radius of monoliths,  is constant, and D is the intraparticle diffusion coefficient. 2.4.3. Surface coverage of bioadsorbents The fraction of the coverage bioadsorbents pore surfaces (fc , g/m2 ) occupied by the protein molecules were determined as following equation: fc =

Mˇ S

(4)

where M is the molecular area of the protein molecule, S (m2 /g) is the surface area of monolith MA bioadsorbents, and b is the number of molecules adsorbed per unit area of bioadsorbent. However, ˇ can be calculated as follows: ˇ = qe /S/NA , where NA is Avogadro’s number. 2.4.4. Kinetic and thermodynamic studies of adsorption assays The batch adsorption of proteins into alumina bioadsorbents was analyzed by applying Lagergren’s equation [38]: dq = kt (qe − qt ) dt

(5)

By applying the initial conditions qt = 0 at t = 0, the linear integration form of pseudo first order model can be expressed as: Ln

q − q  e t qe

= kt t

(6)

where kt is the rate constant (per gram adsorbent, pga) of the firstorder kinetics. The free energy of the activation (G*), the enthalpy of activation (H*) and the entropy of activation (S*) were calculated from Eyring’s equation: kt =

kT −G∗ /RT e h

(7)

H ∗ = E − RT ∗

∗

G = H − TS Ln

(8) ∗

H kt h S =− + RT R kT

(9) (10)

where k is Boltzmann’s constant, h is Planck’s constants and T is absolute temperature.

The thermodynamic equilibrium constant, Kc which is dependent upon the fractional attainment of equilibrium (fe ) of the protein adsorbed molecules deduced from the following equation: Kc =

qf 1 − qf

(11)

where qf is the ratio of the amount of molecule adsorbed at a time (qt ) to that adsorbed at infinity (q∞ ), (i.e. qf = qt/ q∞ ). From the value of Kc , the Gibbs free energy change, G, can be derived. The plot of ln Kc vs. 1/T, gives the numerical values of H of the adsorption of protein adsorption; G and S using the following relations: G = −RT Ln Kc Ln Kc =

−H S + RT R

(13) (14)

where H, S, G, and T are the change in enthalpy, the change in entropy, the change in Gibbs free energy and temperature in Kelvin, respectively. R is the gas constant and Kc is equilibrium constant. 2.5. Analysis of mesoporous alumina bio-adsorbents Small angle X-ray scattering (SAXS) experiments were performed at room temperature. A two-dimensional (2D) confocal mirror (Rigaku Nanoviewer) and a pinhole collimator were used to obtain a focused high flux/high transmission. A monochromatic ˚ was also used. The 2D X-ray beam of CuK˛ radiation ( = 1.54 A) SAXS patterns were recorded using a 2D detector (Bruker Hi-Star) covering a range of momentum transfer q = (4/) sin(2/2) from 0.2 cm−1 to 10 cm−1 , where  is the wavelength of the incident Xray beam and 2 is the scattering angle. The inter-particle distance (center to center) was calculated using d = 2/qmax . Wide-angle powder X-ray diffraction (WAXRD) patterns were measured using an 18 kW diffractometer (Bruker D8 Advance) with monochromated CuK␣ radiation. The scattering reflections were recorded for 2 angles between 1◦ and 100◦ with a step size of 0.1◦ corresponding to the d-spacing between 88.2 and 1.35 nm. N2 adsorption–desorption isotherms were measured using a BELSORP MIN-II analyzer (JP BEL Co., Ltd.) at 77 K. The pore size distribution was determined from the adsorption isotherms by the nonlocal density functional theory (NLDFT). The specific surface area (SBET ) was calculated using multi-point adsorption data from the linear segment of the N2 adsorption isotherms by the Brunauer–Emmett–Teller (BET) theory. High-resolution transmission electron microscopy (HRTEM), three-dimensional transmission electron microscopy (TEM) surfaces, and scanning TEM (STEM) were performed using a JEOL JEM model 2100F microscope. HRTEM was conducted at an acceleration voltage of 200 kV to obtain a lattice resolution of 0.1 nm. The HRTEM images were recorded using a charge-coupled device camera. STEM was carried out at a camera length of 80 cm and a spot size of 1 nm. To investigate the acidic properties of aluminosilica membranes, NH3 temperature-programmed desorption (NH3 -TPD) was measured by using a BEL-Japan TPD-1S system with a quadrupole mass spectrometer. 27 Al magic-angle spinning (MAS) nuclear magnetic resonance (NMR) spectroscopy was also recorded using a Bruker AMX-500 spectrometer. 27 Al NMR spectra were measured at a frequency of 125.78 MHz with a 90◦ pulse length of 4.7 ␮s. For all samples, the repetition delay was 64 s with a rotor spinning at 4 kHz for 27 Al NMR. The chemical shift scale was externally adjusted to be zero for the 27 Al signal using an aqueous solution (1 N) of Al(NO3 )3 .

S.A. El-Safty et al. / Colloids and Surfaces B: Biointerfaces 103 (2013) 288–297

3. Results and discussion 3.1. Stirring-assisted synthesis of mesocage alumina monoliths In this work, the stirring-assisted approach of direct template synthesis was used to fabricate tunable mesocage alumina monoliths using different surfactants (CTAB and Brij-58). Controlled alumina mesostructures were synthesized by direct templating of the quaternary microemulsion liquid crystalline phases of Brij 58 or CTAB/C12 -alkane/Al(NO3 )3 /HCl solution/ethanol composition. Mesoporous alumina bioadsorbents were synthesized with different concentrations of CTAB and Brij-58. This stirringassisted synthesis approach provides several key advantages. First, the approach provides the ability to control enlargement of regular mesostructure pores using different surfactants (CTAB and Brij-58). Second, quaternary microemulsion liquid crystalline phases of CTAB or Brij-58/C12 -alkane/AlNO3 ·9H2 O/HCl solution/ethanol composition were used for control engineering pore systems, as proven by small-angle X-ray scattering (SAXS), N2 adsorption/desorption isotherm, and high-resolution transmission electron microscopy (HRTEM) profiles. This new strategy based on tunable pore sizes (2.7–20 nm) and giant mesocage alumina is highly desirable for trapping large biomolecules such as proteins (Scheme 1). Third, despite the stirring conditions, these mesocage alumina retained well-defined micrometric morphology of monoliths that can open up new ways to realize tailor-made micro-objects to a given protein encapsulation design. The development of ultra- or micrometer-sized scale morphology composed of mesocage-shaped mesoporous monoliths can be effectively used to separate large macromolecules into large pore surface matrices. The three-dimensional (3D) structure and molecular engineering model of the mesocage alumina fabricated by stirring-assisted approach of CTAB and Brij 58 surfactants are systemically shown in Scheme 1. The mesocage organization and the alumina network clusters are the key to the precise manipulation of encapsulation/immobilization of protein (Scheme 1A–C). The large cage-shaped alumina and connecting pores result in enhanced sizeselective adsorption of protein into interfaced cavities (Scheme 1D and E).

3.2. Characteristics of mesocage alumina network monoliths The SAXS profiles (Fig. 1A) of calcined mesocage alumina exhibited a significant broad reflective peak at q = 1.75 and 1.79 nm−1 , corresponding to d-space = 3.58 nm and 3.51 nm of the alumina mesocage synthesized in the presence of soft-template agents Brij 58 and CTAB, respectively. Although the disordered mesostructure (Fig. 1A) characterizes the fabricated alumina monoliths, the scattering patterns indicate the formation of mesoporous alumina structures (Fig. 1) [39,40]. Finely resolved Bragg diffraction peaks were clearly obtained using mesoporous alumina bioadsorbents despite the high loading adsorption level of protein molecules on the pore surface. The wide-angle X-ray diffraction (WAXD) patterns of the mesoporous alumina monoliths show broad Bragg diffraction peaks. This finding suggests that this mesoporous alumina is readily crystallized within the confined nanospace (Fig. 1B) [41]. The alumina cluster orientation resulted in open nanoscale alumina network pores. Scheme 1D–E show the stability of the orientational mesocage alumina pores after protein immobilization. Our finding indicates that a highly intense scattering peak is clearly evident in the mesoporous alumina bioadsorbents despite the high loading level of protein molecule adsorption on the pore surface. The stability of the mesocage alumina matrices increases the possibility of high flux and uptake of proteins during adsorption (Fig. 1A).

291

The HRTEM images of mesocage alumina monoliths reveal that the worm-like mesopores are connected to large regions of alumina monolith domains fabricated using Brij 58 and CTAB surfactants, which agree with the SAXS profiles (Fig. 2A and B). The electron diffraction (ED) pattern (Fig. 2D, inset) clearly shows pat√ √five-ring √ √ terns, which √ √ indicate the d-spacing ratios of 8, 11, 8, 11, 12, and 19 for the cubic Fd3m space of the ␥-Al2 O3 phase. This result revealed the formation of crystalline alumina pore surfaces around the worm-like mesoscopic cavities, which lead to the simple binding of protein during bioadsorption. The results of TEM and ED analyses indicate that the mesocage alumina show mesoscopic pores and cubic matrices in the crystal structures, which agree with the SAXS profiles (Fig. 1A) and geometrical analysis (Scheme 1) [39,42–45]. Dark field-scanning transmission electron microscopy (DF-STEM) and STEM-energy dispersive spectroscopy (EDS) mappings were carried out to characterize the surface composition and atomic distribution of mesoporous alumina. STEM–EDS mapping indicates that Al and O were uniformly distributed on the platelet surface, wherein the [Al]/[O] ratio was 1.7 (Fig. S1). Specific surface area and pore size have key functions in protein adsorption on the mesomaterial. Thus, the specific surface area, pore volume, and pore size distribution of mesoporous alumina were characterized using N2 adsorption and desorption isotherms (Fig. S2). The N2 adsorption isotherms feature the H2 -type hysteresis loop [43–49]. Moreover, typical cage mesopores with uniform unimodal or bimodal entry point were observed in the alumina bioadsorbents [40]. The lower closure point of the hysteresis loop was observed within a very narrow range of relative pressure (i.e., approximately 0.42 or 0.45 for N2 at 77 K), indicating the cavitationinduced stepwise desorption in the ink-bottle pores [41]. These results can be attributed to capillary condensation, which is significantly dependent on the surfactant content found within a narrow range of tubular pores. The lower closure point of the hysteresis loop was significantly dependent on the surfactant content, indicating the columnar-induced stepwise desorption in open tubular pores. This phenomenon results in a simpler diffusion of large macromolecules, such as open cage-like pore scaffolds, than that of ink-bottle pore materials with narrow entrances [50,51]. Based on the N2 isothermal results, the increased surfactant content resulted in large increments in the spherical mesopore cavities, which vary from 8 nm to 20 nm for Brij 58 and from 2.7 nm to 9 nm for CTAB. In addition, the decrease in the textural parameters of alumina after protein adsorption assays indicates the interior inclusion of proteins into mesopore cavities (Fig. S2E) [39,48–50]. 27 Al nuclear magnetic resonance spectra and NH -temperature 3 programmed desorption (NH3 -TPD) patterns were obtained to confirm the presence of surface acidity of the guest-free samples (Fig. S3). The coordination and location of aluminum sites in the frameworks are key determinants in the generation of surface acidity in alumina monoliths (Fig. S3A). The acidity of mesoporous alumina was determined through NH3 -TPD using mass spectroscopy to measure pure NH3 desorption without any impurity. The NH3 -TPD results show two main peaks of NH3 desorption at around 200 ◦ C and a small, broad peak at 200 ◦ C to 500 ◦ C, indicating the formation of weak “Lewis” and mildly strong “Brønsted” acid sites of the OH-groups of alumina bioadsorbents (Fig. S3B). 3.3. Applicability of the geometrical mesocage alumina monoliths on protein bioadsorption Geometrical studies on a mesoporous alumina worm-like model were designed using the Gaussian program to identify key factors, such as surface energy, charge distribution, hydrophobicity, and electrostatic interaction. These factors feasibly control protein encapsulation into the interior bioadsorbent cavity. Scheme 1 shows the theoretical models of the alumina pore structures built

292

S.A. El-Safty et al. / Colloids and Surfaces B: Biointerfaces 103 (2013) 288–297

Scheme 1. Geometrical model of mesoporous cage alumina (A) with molecular pore structures (B) and networks (C). The immobilized proteins in the inner or outer pore surfaces of alumina with the relative formation energy (D and E, respectively) (The E unit is kcal/mol).

Fig. 1. SAXS (A) and WAXD patterns (B). SAXS of mesoporous alumina bioadsorbent synthesized using Brij 58-2 (a) and CTAB-2 (b) surfactants. (a, b)* SAXS of mesocage alumina-loaded ␣-AMY protein after adsorption assays.

Fig. 2. Representative HRTEM (A, B) and ED (B-insets) micrographs of mesoporous alumina bioadsorbents fabricated using Brij 58 (A) and CTAB (B).

S.A. El-Safty et al. / Colloids and Surfaces B: Biointerfaces 103 (2013) 288–297

based on spinel cubic Fd3m crystal lattices, as proven by WAXD. In this case, the pore wall surface consists of an octahedral Al2 O3 arrangement (Scheme 1B) where the oxygen atoms are closely packed and the Al atoms fill the Fd3m positions. Thus, the Al atoms are octahedrally coordinated by oxygen (Scheme 1C). The energy relaxation of the unit cell structure shows the localized distortion of the octahedral symmetries with minor appearance of tetrahedral symmetries. This distortion clearly appears in the longer Al O ˚ Calculations also bond length from the average 1.81 A˚ to 1.92 A. reveal that the Al charge is +0.91 ± 0.05 and the oxygen charge is −0.53 ± 0.07. The development of alumina network clusters and ultraor micrometer-sized scale morphology composed of mesocageshaped mesoporous monoliths, which can be effectively used to encapsulate large macromolecules into the interior cage cavities, would greatly assist in other potentials for size- and shape-selective separation (see below). LYS protein was used for the adsorption calculation model imported from the crystal structure protein data bank (ID: 1LSM) to clarify the interior immobilization of proteins during the adsorption assays of alumina bioadsorbents. The adsorption model was created using a soft docking algorithm implemented on the SYBYL-X suite program [51]. The activity of the inner pore wall surface was determined from the formation energy calculation of the protein/surface docked models. Various chemical properties were used to calculate the most stable conformations of protein adsorption (e.g., electrostatic interaction and charges). The results show that the formation energy of the inner immobilization of protein is less than that of the surface location by 5 kcal/mol (Scheme 1D and E). This finding indicates that protein adsorption is mainly preferred in the inner rather than in the outer pore surface of the mesoporous alumina. 3.4. Key components in the bioadsorption assays of macromolecules Several key factors were studied using batch-adsorption assays to investigate the protein adsorption performance of mesoporous alumina monoliths. These factors include the molecular shape and size, concentration, and composition of proteins, such as INS (2.4 nm), HD (2.9 nm), LYS (3.2 nm), GA3 (4.0 nm), ␤-LG (4.2 nm), ␣-ATR (5.9 nm), ␣-AMY (6.8 nm), and MYS (14 nm to 19 nm), in addition to temperature, pH solution, and active aluminum sites (Fig. S4A–D). All Lewis acid sites can transform into Brønsted acid sites of the alumina bioadsorbents, and are the key factors for enhanced adsorption uptake. These natural surfaces of acid sites strongly induce H-bonding and dispersive interactions with biomolecules. A series of different protein (1 × 10−6 M) adsorption assays were performed using Brij 3 as bioadsorbent to study the effect of molecular shape and size of proteins on their adsorption into the internal pores of the mesoporous alumina (Fig. S4A, B). Figure S4 (A and B) indicates the significant effect of the molecular size of proteins in the adsorption efficiency of mesoporous alumina in terms of the large quantity and high-speed adsorption of proteins or any biomolecules. Figure S4 (A–C) shows that a fast reaction occurred after short time intervals depending on each protein, and then the adsorption rate increased slowly with time until an equilibrium state was attained. Figure S4 (C and D) shows that the equilibrium adsorption capacity increases with increased LYS concentration and temperature. The adsorption rate and capacity depend on the molecular weight and size of proteins, concentration, and temperature. Proteins typically carry electrostatic charge, often both positive and negative on the same protein. The protein net charge, charge distribution, and decay length of electrostatic interactions with the surface can all affect the adsorption uptake of proteins on the solid

293

Fig. 3. Adsorption capacity of proteins (INS, pH 5.3; MYS, pH 5.2) on CTAB-2 and Brij 1 mesocage alumina in 0.2 M acetate buffer solution and water. The pH solution was prepared using 0.2 mol of sodium acetate (CH3 COONa) in water and adjusted with acetic acid (CH3 COOH) to the desirable pH values.

surface. An experimental setup was designed in this study to investigate the effect of pH on the adsorption affinity of proteins into the mesocage alumina to study the effect of electrostatic force between proteins and alumina surfaces. The adsorption assays for both lowINS and high-MYS molecular weight proteins were carried out in acetate buffer (at their isoelectric points, i.e., pH 5–5.3) and in aqueous solutions (pH 6.5) (Fig. 3). The adsorption of INS is one order of magnitude higher than that of MYS in both water and buffer solution due to the significant effect of the molecular weight and size of proteins on protein uptake (Fig. 3). The experimental data exhibited a slight change in the adsorption capacity of proteins into the mesocage alumina. Our findings further revealed that protein adsorption slightly increased with increase in protein sizes at their isoelectric points. Thus, the adsorption affinity for light proteins (e.g., INS) was more enhanced in the aqueous solution than that in the acetate buffer at isoelectric pH 5.3. However, the adsorption of MYS was enhanced in the buffer compared with that in the aqueous solution (Fig. 6), which is mainly due to the aggregation or multimerization of INS at this specific pH. Giger et al. reported that INS has relatively unstable monomers and dimers that tend to form large aggregates (i.e., high hydrophobic form) in a manner highly dependent upon pH and ionic strength [52]. Therefore, INS tends to change its conformation near the isoelectric point, which is inconsistent with MYS. This finding indicates that the charged structure and protein size have significant effects on adsorption. 3.5. Protein size-selective bioadsorbent-based mesocage alumina The size and shape of each protein molecule have important functions in the adsorption of protein molecules [30,31]. A series of adsorption experiments were carried out using lowmolecular weight and small-molecular size proteins, such as INS (size = 2.4 nm) and ␤-LG (size = 4.2 nm), using CTAB-alumina bioadsorbents (pore size = 4.9 nm to 9 nm; Fig. 4A) to study protein size-selective adsorption in the internal pore cavities. However, Brij-alumina monoliths (pore size = 9 nm to 20 nm) were effectively used as size-selective bioadsorbents of high molecular weight and size of protein, such as ␣-AMY (size = 6.8 nm) and MYS (size = 14–19 nm; Fig. 4B). The results indicate that the mesocage pore sizes of alumina bioadsorbent, as well as the molecular weight and size of proteins, are the key components in enhancing adsorption efficiency in terms of large-quantity uptake and high-speed capacity of proteins or any biomolecule.

294

S.A. El-Safty et al. / Colloids and Surfaces B: Biointerfaces 103 (2013) 288–297

Fig. 5. Linear form of the Langmuir plot (A and B) for the adsorption of proteins into (1.6 g/L) mesoporous alumina bioadsorbent at 20 ◦ C.

Fig. 4. Time dependence of the adsorption of 1 × 10−5 M proteins on adsorption assays using 1.6 g/L mesoporous alumina synthesized with CTAB (A) and Brij (B).

The mesopore tuning of the cage window and cavity is highly significant in the controlled mass transportation in the porous system, as well as in the size-selective enrichment and separation of biomolecules with different sizes from a complex system. The mesoporous alumina was used with two mixtures of differentsized proteins (i.e., ␣-AMY and MYS; INS and ␣-AMY mixtures). These proteins served as model guest molecules for the adsorption experiments on Brij-1 (pore size = 20 nm) and CTAB-2 (pore size = 20 nm) alumina (Fig. S5A, B) to investigate the size selectivity of mesoporous alumina in bioadsorption. The results indicate that mesoporous alumina can be created for size-selective applications, such as enrichment and separation, as well as drug delivery, by applying engineering techniques in the pore connection and multi-modal pore system.

temperature is represented by the Langmuir isotherm to verify such adsorption behavior (Fig. 5). The straight line of the Ce /qe vs. Ce plot of this adsorption assay proves the formation of monolayer coverage of proteins on the interior pore surfaces of monoliths. The linear adsorption curves indicate that a wide range of biomolecule concentrations can be removed in a one-step treatment. A linear graph (Fig. 5A and B) with a correlation coefficient range of 0.98–0.99 clearly shows that the Langmuir adsorption isotherms characterize the adsorption assays for all proteins. The monolayer adsorption capacity qm and the Langmuir coverage constant KL are obtained from the slope and intercept of the linear plot. The qm data indicate that the practical removal of 1.0 g of INS, for example, requires 16.6 g of giant alumina. Moreover, the KL values are consistent with the adsorption/desorption rates, indicating that the protein adsorption assays are fully reversible (see below). The qm and KL values decrease in the order MYS < ␣-AMY < ␣-ATR < ␤LG < GA3 < LYS < HD < INS. Thus, the size and shape of each protein molecule have important functions in molecular adsorption. 3.7. Biomolecule diffusion and mass transport

3.6. Bioadsorption isotherms of biomolecules on mesocage alumina networks Langmuir adsorption, which is the monolayer adsorption, depends on the assumption that the intermolecular forces decrease rapidly with distance, and consequently predict the existence of monolayer coverage of the protein adsorbate at the outer pore surface of the alumina adsorbent. The homogeneous adsorbent structures have key functions based on the Langmuir equation, which proposes that empirical symmetry exists in all adsorbent–adsorbate sites and energy. The monolayer coverage of proteins on alumina bioadsorbent pore surfaces at constant

In adsorption experiments, protein mass transport is most probably transformed from a bulk solution to hierarchical alumina monoliths through intraparticle diffusion (D), which is often the rate-limiting step in numerous adsorption processes. The fractional attainment of equilibrium (fe ) against t1/2 of mesoporous alumina bioadsorbents was plotted at different sizes and shapes of proteins according to Fick’s second law to understand the factors affecting biomolecule mass transport from aqueous phase to the binding sites of mesoporous alumina monoliths [39,40]. Figure S6 shows that S-curves can be classified into three portions. Our results show that the diffusion of biomolecules across the

S.A. El-Safty et al. / Colloids and Surfaces B: Biointerfaces 103 (2013) 288–297

295

Table 1 Kinetic and thermodynamic parameters for protein bioadsorption on the pore surface of mesocage alumina monoliths. Protein

M.Wt (kDa)

T (K)

Kinetic parameters kt

Ea (kJ/Mol)

Thermodynamic parameters H* (kJ/Mol)

S* (J/(K Mol))

G* (kJ/Mol)

Kc

49.55

−460.45

134.9 137.3 139.6 141.9

0.09 0.02 0.01 0.003

S (J/(K Mol))

G (kJ/Mol)

0.176

−617.1

2.27 2.23 2.21 2.18

123.1 125.2 127.3 129.4

0.28 0.07 0.03 0.01

0.177

−613

2.26 2.22 2.20 2.16

111.9 113.8 115.8 117.7

0.70 0.17 0.08 0.02

0.180

−614.1

2.27 2.22 2.21 2.17

−324.59

95.1 96.8 98.4 100

1.20 0.32 0.13 0.04

0.182

−616.7

2.28 2.24 2.22 2.18

−274.63

80.5 81.9 83.2 84.6

1.98 0.50 0.18 0.06

0.19

−638.2

2.36 2.31 2.29 2.26

−224.62

65.8 66.9 68.1 69.2

3.67 0.70 0.27 0.09

−665.6

2.46 2.42 2.39 2.36

18.57

−172.52

50.6 51.4 52.3 53.2

6.68 1.11 0.41 0.12

0.22

−721.3

2.67 2.62 2.60 2.56

15.46

−143.76

42.1 42.9 43.6 44.3

5.44 5.02 1.57 0.42

0.26

−852.9

3.16 3.10 3.07 3.03

200–500

293 298 303 308

0.019 0.029 0.039 0.058

␣-AMY

54

293 298 303 308

0.025 0.037 0.046 0.065

49.72

45.23

−420.04

␣-ATR

47.6

293 298 303 308

0.027 0.039 0.048 0.065

43.40

41.07

−381.88

␤-LG

18.4

293 298 303 308

0.031 0.044 0.051 0.068

38.49

34.92

GA3

16

293 298 303 308

0.038 0.051 0.059 0.075

33.63

29.52

14.3

293 298 303 308

0.043 0.053 0.062 0.079

HD

11.1

293 298 303 308

0.049 0.058 0.069 0.083

25.97

INS

5.7

293 298 303 308

0.057 0.065 0.075 0.087

21.71

MYS

LYS

63.02

37.16

24.17

aqueous diffusive boundary layer at the adsorbent-water interface can also affect the molecular transfer to the receiving adsorbent phase [39,53]. The fraction of the coverage alumina bioadsorbent surfaces (fc , g/m2 ) occupied by protein molecules can be determined in Fig. 6. The fc and D parameters of the bioadsorbents decrease in the following order: MYS < ␣-AMY < ␣-ATR < ␤LG < GA3 < LYS < HD < INS. Based on these adsorption parameters, the higher adsorption effectiveness of these proteins is clearly consistent with the low molecular weight and concentrations of proteins. 3.8. Kinetic and thermodynamic studies of protein adsorption Kinetic and thermodynamic studies feature the stability of protein immobilization in interior pore cavities. The stability of immobilized proteins is one of the important characteristics in defining the adsorption efficiency in terms of the adsorption rate and capacity. We performed thermodynamic and kinetic studies of protein adsorption at different temperatures ranging from 20 ◦ C to 35 ◦ C to investigate the adsorption feature and diffusion of proteins into mesocage alumina cavities. A pseudo-first-order equation was used to investigate the adsorption mechanism and to describe protein adsorption in solid–liquid systems based on mesocage alumina bioadsorbent. Figure S7 shows that the first-order kinetic equation best describes the data on protein adsorption into the mesocage alumina bioadsorbent. The values of the rate constant kt (per gram of bioadsorbent) of protein adsorption were determined from the slope of the linear first-order kinetic equation (Table S1). The kinetic studies showed that kt decreases in the following INS > HD > LYS > GA3 > ␤-LG > ␣-ATR > ␣-AMY > MYS sequence:

H (kJ/Mol)

0.20

(Table 1). These kt values are consistent with the same sequence of adsorption amounts, as well as with the fc and D parameters of these proteins. The temperature dependence of kt values for LYS was studied as shown in Fig. S1. The results indicate that kt increased with temperature. The activation energy E of protein adsorption was deduced from an Arrhenius plot (Fig. S8). Other activation parameters of the free activation energy G*, activation enthalpy H*, and activation entropy S* were calculated from the Eyring equation (Fig. S9) and are listed in Table 1. The lower E value with higher kt value was in agreement with the sequence of the adsorption affinity of the proteins in the mesoporous alumina bioadsorbent. The values of the activation parameters (Table 1) are almost consistent with those reported for the diffusion of small molecules through the interior particle pores of solid adsorbents, reflecting the ease of diffusion of protein, particularly with size <7 nm, into the interior mesoporous alumina bioadsorbent monoliths. The plot of ln Kc vs. 1 = T (Fig. S10) gives the numerical values of H of protein adsorption. G and S are calculated and presented in Table 1. Our findings reveal that Kc increases with temperature, whereas the absolute value of G increases with decrease in temperature. Thus, protein adsorption is spontaneous and more favorable at relatively high temperatures without denaturation, which confirms an endothermic adsorption process with positive H. The large S value indicates the greater stability of thermodynamic adsorption of proteins into the interior pores of mesocage alumina. The S value increases in this sequence: INS > HD > LYS > GA3 > ␤LG > ␣-ATR > ␣-AMY > MYS, contributing to a greater value of Kc and greater stability of thermodynamic adsorption. This result reflects the attraction between proteins and the active acid sites into the

296

S.A. El-Safty et al. / Colloids and Surfaces B: Biointerfaces 103 (2013) 288–297

Fig. 6. Fraction coverage of 1.6 g/L mesoporous alumina bioadsorbents with different sizes and shapes of proteins at 20 ◦ C.

micrometer-sized particle of monoliths, which draws the proteins to diffuse toward the particle center. 3.9. Reusability of bioadsorbent monoliths The limited applications of porous bioadsorbents in medicine are due to the damage of pore density and uniformity within reuse cycles. A major advantage of alumina bioadsorbent monoliths is their reducibility and reversibility after multiple cycles of protein adsorption. Due to the difficulty of clearing pores clogged with proteins, a self-cleaning process was reported to overcome the clogging of membrane pores by macromolecules [28]. Reversible adsorption was carried out using acidified solution (HCl/NaCl) to effectively remove the trapped proteins in the giant alumina pores. The desorption rate of proteins is shown in Fig. S11. The adsorbed proteins desorb slowly when exposed to a buffer solution (Fig. S11A). The desorption rates appear to be influenced by the protein-alumina pore surface interactions and the molecular weight of proteins (Fig. S11B). The absence of the desorption rate indicates that desorption occurred mainly from the interior cavities of trapped proteins (Fig. S11). Our studies indicate that the Langmuir equilibrium constant KL values are consistent with the ratio of adsorption/desorption rates of proteins (k1 /k−1 ), indicating that the protein adsorption assays are fully reversible (Fig. S11B, inset). Fig. 7 shows that the monoliths retain a high-protein adsorption efficiency after numerous reuse cycles (i.e., dead-end adsorption) [53–56]. The efficiency of monoliths during the reuse cycles of protein adsorption assays decreases in the order of MYS < ␣AMY < ␣-ATR < ␤-LG < G3A < LYS < HD < INS. This finding indicates

Fig. 7. Reusability study of 1 × 10−6 M protein adsorption onto 1.6 g/L Brij-2 bioadsorbent at 20 ◦ C. The efficiency (E%) of the bioadsorbent was calculated from the % recovery ratio of the adsorbed amount of protein at equilibrium (qr ) per reuse cycle (No.) and the initially adsorbed amount (qo ) obtained from the initial use of bioadsorbent monoliths.

the effectiveness of protein size, shape, and weight in continuous dead-end adsorption. The adsorption efficiency drastically decreased with cycling number. However, the reused mesocage alumina bioadsorbents remained effective to some extent for protein adsorption after six recycles. 4. Conclusion The stirring-assisted approach of direct template synthesis was used to fabricate tunable mesocage alumina monoliths using different surfactants (CTAB and Brij 58). The alumina mesopore matrices and mesocage cavities resulted in the high performance of the selective encapsulation of biomolecules with different shapes, sizes, functions, and properties (e.g., INS, HD, LYS, GA3, ␤-LG, ␣ATR, ␣-AMY, and MYS). Several key factors, including the molecular shape and size, concentration, and composition of proteins, in addition to temperature, pH solution, and active aluminum sites, can control their potential immobilization during the adsorption assays in terms of interior diffusivity and adsorption capacity. Our findings revealed that the pore sizes of mesoporous alumina bioadsorbents have key functions in the efficiency of size-selective adsorption assays, particularly during protein separation in the mixture. For the first time, geometrical models of alumina bioadsorbents were studied to investigate the major driving forces to achieve the most optimal performance of encapsulation/adsorption in terms of diffusivity and quantity, particularly with large molecular weight,

S.A. El-Safty et al. / Colloids and Surfaces B: Biointerfaces 103 (2013) 288–297

size, and concentrations of proteins. Furthermore, the occurrence of single protein adsorption from the mixtures based on sizeand shape-selective separation can open up new ways to fabricate micro-objects that suit a given protein encapsulation design. Thermodynamic studies showed that the stability of protein immobilization and protein adsorption on mesocage are favorable and spontaneous. Moreover, the reducibility and reversibility of alumina adsorbents pose a unique and interesting challenge.

[22] [23] [24] [25] [26] [27] [28] [29] [30]

Appendix A. Supplementary data

[32]

data associated with Supplementary cle can be found, in the online http://dx.doi.org/10.1016/j.colsurfb.2012.10.040.

this artiversion, at

References [1] H.B. Lu, J.I. Homola, C.T. Campbell, G.G. Nenninger, S.S. Yee, B.D. Ratner, Sens. Actuators B 74 (2001) 91. [2] L.B. Carneiro, J. Ferreira, M.J.L. Santos, J.P. Monteiro, E.M. Girotto, Appl. Surf. Sci. 257 (2011) 10514. [3] S. Bertazzo, K. Rezwan, Langmuir 26 (2010) 3364. [4] K.A. DeFriend, M.R. Wiesner, A.R. Barron, J. Membr. Sci. 224 (2003) 11. [5] B. Kalska-Szostko, E. Orzechowska, Mater. Chem. Phys. 129 (2011) 256. [6] C.J. Xu, K.M. Xu, H.W. Gu, X.F. Zhong, Z.H. Guo, R.K. Zheng, J. Am. Chem. Soc. 126 (2004) 3392. [7] Z. Liu, M. Li, X. Yang, M. Yin, J. Ren, X. Qu, Biomaterials 32 (2011) 4683. [8] H.M. Chen, C.H. Deng, X.M. Zhang, Angew. Chem. Int. Ed. 49 (2010) 607. [9] R. Dringen, Y. Koehler, L. Derr, G. Tomba, M.M. Schmidt, L. Treccani, L.C. Ciacchi, K. Rezwan, Langmuir 27 (2011) 9449. [10] M. Heule, K. Rezwan, L. Cavalli, L.J. Gauckler, Adv. Mater. 15 (2003) 1191. [11] A. Lopez Valdivieso, J.L. Reyes Bahena, S. Song, R. Herrera Urbina, J. Colloid Interface Sci. 298 (2006) 1–5. [12] F. Bellezza, A. Cipiciani, L. Latterini, T. Posati, P. Sassi, Langmuir 2518 (2009) 10918. [13] J. Deere, E. Magner, J.G. Wall, B.K. Hodnett, Chem. Commun. 5 (2001) 465. [14] J. Deere, E. Magner, J.G. Wall, B.K. Hodnett, Catal. Lett. 85 (2003) 19. [15] J.D. Andrade, Surface Interfacial Aspects of Biomedical Polymers: Protein Adsorption, vol. 2, Plenum, New York, 1985, pp. 1–80. [16] M. Mahato, P. Pal, T. Kamilya, R. Sarkar, G.B. Talapatra, J. Phys. Chem. B 114 (2010) 495–502. [17] K. Niesz, P. Yang, G.A. Somorjai, Chem. Commun. (2005) 1986–2198. [18] A.M. Karim, V. Prasad, G.M. Pourmpakis, W.W. Lonergan, A.I. Frenkel, J.G.G. Chen, D.G. Vlachos, J. Am. Chem. Soc. 131 (2009) 12230. [19] L.B. Sun, J. Yang, J.H. Kou, F.N. Gu, Y. Chun, Y. Wang, J.H. Zhu, Z.G. Zou, Angew. Chem. Int. Ed. 47 (2008) 3418. [20] D. Grosso, G.J. de, A.A. Soler-Illia, E.L. Crepaldi, B. Charleux, C. Sanchez, Adv. Funct. Mater. 13 (2003) 37. [21] J. Cejka, Appl. Catal. A 254 (2003) 327.

[31]

[33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56]

297

F. Vaudry, S. Khodabandeh, M.E. Davis, Chem. Mater. 8 (1996) 1451. M. Trueba, S.P. Trasatti, Eur. J. Inorg. Chem. 17 (2005) 3393. D.L. Trimm, A. Stanislaus, Appl. Catal. 21 (1986) 215. Z.-Y. Yuan, T.-Z. Ren, A. Vantomme, B.-L. Su, Chem. Mater. 16 (2004) 5096. W.Q. Cai, J.G. Yu, B. Cheng, M. Jaroniec, J. Phys. Chem. C 113 (2009) 14739. A.R. Garcia, R.B. de Barros, A. Fidalgo, L.M. Ilharco, Langmuir 23 (2007) 10164. S.A. El-Safty, Trends Anal. Chem. 30 (3) (2011) 447. S.A. El-Safty, A.Md. Shahat, R. Awual, J. Colloid Interface Sci. 359 (2011) 9–18. N.G. Balabushevitch, G.B. Sukhorukov, N.A. Moroz, D.V. Volodkin, N.I. Larionova, E. Donath, H. Mohwald, Biotechnol. Bioeng. 76 (3) (2001) 207–213. Q. Yu, Y. Zhang, H. Chen, Z. Wu, H. Huang, C. Cheng, Colloid Surf. B: Biointerfaces 76 (2010) 468–474. G.M.L. Messina, C. Satriano, G. Marletta, Colloid Surf. B: Biointerfaces 70 (2009) 76–83. X.-J. Hu, J.-S. Wang, Y.-G. Liu, X. Li, G.-M. Zeng, Z.-L. Bao, X.-X. Zeng, A.-W. Chen, F. Long, J. Hazard. Mater. 185 (2011) 306. S.A. El-Safty, J. Colloid Interface Sci. 260 (2003) 184. S.A. El-Safty, Y. Kiyozumi, T. Hanaoka, F. Mizukami, Appl. Catal. A 337 (2008) 21. S.A. El-Safty, A. Shahat, W. Warkocki, M. Ohnuma, Small 7 (2011) 62–65. P.A. Mangrulkar, S.P. Kamble, J. Meshram, S.S. Rayalu, J. Hazard. Mater. 160 (2008) 414. J.X. Lin, S.L. Zhan, M.H. Fang, X.Q. Qian, J. Porous Mater. 14 (2007) 449. D. Tang, R. Niessner, D. Knopp, Biosens. Bioelectron. 24 (2009) 2125. S. Ajitha, S. Sugunan, J. Porous Mater. 17 (2010) 341. S.A. El-Safty, T. Hanaoka, F. Mizukami, Chem. Mater. 17 (2005) 3137. R. Tian, H. Zhang, M. Ye, X. Jiang, L. Hu, X. Li, X. Bao, H. Zou, Angew. Chem. Int. Ed. 46 (2007) 962. C.J. Lucio-Ortiz, J.R. De la Rosa, A.H. Ramirez, J.A. De los, R. Heredia, P. Angel, S. Munoz-Aguirre, L.M. De Leon-Covian, J. Sol–Gel Sci. Technol. 58 (2011) 374. M. Moran-Pineda, S. Castillo, T. Lopez, R. Gomez, A.E. Cordero-Borboa, O. Novaro, Appl. Catal. B 21 (1999) 79. Y. Zhao, W.N. Martens, T.E. Bostrom, H.Y. Zhu, R.L. Frost, Langmuir 23 (2007) 2110. J.R. Matos, M. Kruk, L.P. Mercuri, M. Jaroniec, L. Zhao, T. Kamiyama, O. Terasaki, T.J. Pinnavaia, Y. Liu, J. Am. Chem. Soc. 125 (2003) 821. S.A. El-Safty, Y. Kiyozumi, T. Hanaoka, F. Mizukami, J. Phys. Chem. C 112 (14) (2008) 5476. S.A. El-Safty, A.Md. Shahat, R. Awual, M. Mekawya, J. Mater. Chem. 21 (2011) 593. P. Roy, T. Dey, K. Lee, D. Kim, B. Fabry, P. Schmuki, J. Am. Chem. Soc. 132 (2010) 7893. S.A. El-Safty, M.A. Shenashen, Anal. Chim. Acta 694 (2011) 151. SYBYL-X 1.2, Tripos International, 1699 South Hanley Rd., St. Louis, Missouri, 63144, USA. K. Giger, R.P. Vanam, E. Seyrek, P.L. Dubin, Biomacromolecules 9 (2008) 2338–2344. M. Colilla, M. Manzano, M. Vallet-Regí, Int. J. Nanomed. 3 (4) (2008) 403. S.A. El-Safty, Ahmed Shahat, Md. Rabiul Awual, J. Colloid Interface Sci. 359 (2011) 9. D. Qi, J. Wu, K. Wei, H. Ding, Y. Wang, X. Zhang, Y. Qian, Guan, Anal. Bioanal. Chem. 398 (2010) 1715. T. Tsukagoshi, Y. Kondo, N. Yoshino, Colloid Surf. B: Biointerfaces 54 (2007) 101–107.