Interactions of Bovine Serum Albumin with Aluminum

Interactions of aqueous solutions of aluminum polyoxocations (Al13-mers and Al30-mers) and aluminum hydroxide suspensions of varying particle sizes (2...
0 downloads 0 Views 395KB Size
Download PDF
10078

Langmuir 2006, 22, 10078-10088

Interactions of Bovine Serum Albumin with Aluminum Polyoxocations and Aluminum Hydroxide Olivier Deschaume, Kirill L. Shafran, and Carole C. Perry* School of Biomedical and Natural Sciences, Nottingham Trent UniVersity, Clifton Lane, Clifton, Nottingham, The United Kingdom ReceiVed May 8, 2006. In Final Form: August 9, 2006 Interactions of aqueous solutions of aluminum polyoxocations (Al13-mers and Al30-mers) and aluminum hydroxide suspensions of varying particle sizes (26, 55, and 82 nm) with a model protein, bovine serum albumin (BSA), have been investigated using potentiometry, conductometry, viscometry, 27Al solution NMR, UV-vis spectroscopy, dynamic light scattering, ζ-potential measurements, thermogravimetry, X-ray diffraction, and scanning electron microscopy. Increasing amounts of BSA partially convert Al13-mers and, to a larger extent, Al30-mers into amorphous Al hydroxide without gel formation. At the same time, BSA molecules can form unstable aggregates in the Al polyoxocation solutions which redisperse easily upon standing. In the case of Al hydroxide sols, BSA addition causes substantial gelation, the extent of which is proportional to the amount of BSA added and inversely related to the Al hydroxide particle size. Upon freeze-drying or centrifugation of Al species-BSA solutions, an interesting sheetlike morphology with 150-200 nm wide nanoribbons is observed for pure Al hydroxide nanoparticles and for solutions of Al polyoxocations with the highest amount of BSA studied. On the basis of the combined solution, colloidal and solid-state characterization of model Al species-BSA systems, a qualitative model of possible interactions in the Al polyoxocation-BSA and Al hydroxide-BSA systems is proposed wherein core-shell hybrid nanoparticles are formed from protein “core” and Al polyoxocation “shell” or Al hydroxide “core” and protein “shell”.

1. Introduction The biomimetic approach in materials chemistry to create novel materials ordered from the nanometer through to the mesoscopic length scale draws its inspiration from the intricate and elegant routes used by Nature.1,2 Another way to highly ordered structures is the so-called “bottom-up” or “nanotectonic” approach to formation of materials from nanosized building blocks. The latter approach is usually based on the programmed self-assembly of nanoparticles or template-directed methods such as “matrix confinement” or “substrate engineering”.3 Composites synthesized using a combined nanotectonic and biomimetic approaches may show a number of superior properties compared to analogous conventional materials. The nacre of abalone shell composed of alternating layers of aragonite (CaCO3) and biopolymers is a good example of a natural nanocomposite.4 The enhanced toughness of this nanostructured material arises both from the deflection and absorption of microcracks and from the elastic proteins holding the material together after the formation of a crack. In this study we have exploited the ability of Al ions to selfassemble into large polynuclear units in aqueous solutions and to provide molecular building blocks for a combined biomimeticnanotectonic approach to the synthesis of hybrid Al-containing materials using various biopolymers. It is well established that, in mildly acidic conditions at concentrations above the solubility limit of Al hydroxide, monomeric Al species, such as the hexaaquocation [Al(H2O)6]3+ and its hydrolysis products ([Al(H2O)5(OH)]2+, [Al(H2O)4* To whom correspondence should be addressed. E-mail: carole.perry@ ntu.ac.uk. Fax: +44 0115 8486616. Phone: +44 0115 8486695. (1) Davis, S. A.; Breulmann, M.; Rhodes, K. H.; Zhang, B.; Mann, S. Chem. Mater. 2001, 13, 3218-3226. (2) Perry, C. C. In Chemistry of AdVanced Materials: An OVerView; Interrante, L.V., Hampden-Smith, M. J., Eds.; Wiley-VCH: Weinheim, Germany, 1998; p 499. (3) Fendler, J. H. Chem. Mater. 2001, 13, 3196-3210. (4) Sellinger, A.; Weiss, P. M.; Nguyen, A.; Lu, Y.; Assink, R. A.; Gong, W.; Brinker, C. J. Nature 1998, 394, 256-260.

Figure 1. Structure of aluminum Keggin ions: Al13-mers (AlO4Al12(OH)24(H2O)127+) (I) and Al30-mers (Al30O8(OH)56(H2O)2418+) (II).

(OH)2]+), can undergo a succession of condensation reactions. These reactions lead to the formation of small oligomeric Al species, such as Al dimers and trimers5 and further transformation into large Keggin ionssAl13-mers (d ∼ 1 nm)6 and the recently characterized Al30-mer (1 × 2 nm size)7,8 (Figure 1). Basic salts containing aluminum Keggin ions have been employed in a number of applications including clay pillaring,9 preparation of Al2O3 nanoparticles,10 antiperspirant actives,11 catalysts,12 and composite materials.13 When a certain pH (pH > 4.6) and hydrolysis ratio (h > 2.6) is reached (h is the hydrolysis ratio denoted as h ) [OH-]/ [Al3+]), Al Keggin ions start to collapse, with the formation of colloidal Al hydroxide.14 Al hydroxide colloids produced by aqueous and nonaqueous sol-gel routes are among the most frequent precursors to a broad range of Al-containing inorganic and hybrid materials.15 (5) Baes, D. F., Jr.; Mesmer, R. E. In The hydrolysis of cations; WileyInterscience: New York, 1976; p 122. (6) Akitt, J. W. Prog. Nucl. Magn. Reson. Spectrosc. 1989, 21, 1-149. (7) Rowsell, J.; Nazar, L. F. J. Am. Chem. Soc. 2000, 122, 3777-3778. (8) Allouche, L.; Gerardin, C.; Loiseau, T.; Fe´rey, G.; Taulelle, F. Angew. Chem., Int. Ed. 2000, 39, 511-514. (9) Kloprogge J. T. J. Porous Mater. 1998, 5, 5-41. (10) Yao, N.; Xiong, G.; Zhang, Y.; He, M.; Yang, W. Catal. Today 2001, 68, 97-109. (11) Fitzgerald, J. J. In Antiperspirants and Deodorants; Marcel Dekker: New York, 1988; p 119. (12) Bradley, S. M.; Kydd, R. A. J. Catal. 1993, 142, 448-454. (13) Son, J.-H.; Choi, H.; Kwon, Y.-U.; Han, O. H. J. Non-Cryst. Solids 2003, 318, 186-192. (14) Bradley, S. M.; Kydd, R. A.; Howe, R. F. J. Colloid Interface Sci. 1993, 159, 405-412.

10.1021/la061285h CCC: $33.50 © 2006 American Chemical Society Published on Web 10/25/2006

Interactions of BoVine Serum Albumin

Our group has previously concentrated on the synthesis of high-purity soluble Al polynuclear species (Al13 and Al30-mers) and insoluble Al species (Al hydroxide) having a variety of dimensions.16-19 Synthesis of these species using a common method of neutralization with base16-18 and a more advanced anion exchange route19 has been followed by various characterization techniques to obtain a better understanding of Al ion speciation in aqueous solutions. Previous studies on protein-aluminum interactions have largely concentrated on elucidating bioavailability of this element and absorption/elimination pathways in living organisms.20 A major goal has been to search for a potential link of aluminum to Alzheimer’s disease,21 as well as to investigate the negative effect of aluminum on plant growth22-24 and other adverse effects caused by aluminum toxicity.25 In this respect, interactions of soluble aluminum species at low concentrations with various proteins such as transferrin, albumin26 calmodulin,27 and mucin28 as well as with many other biomolecules and naturally occurring ligands29-32 have been studied. Fewer studies have investigated interactions of Al species with proteins and other biopolymers with regard to materials chemistry applications. Examples of such studies include interactions of Al ions with elastin, actin,33 and poly-L-glutamic acid.34 Hem and co-workers35-37 have studied interactions of commercial Al hydroxide adjuvants with BSA and other proteins as model antigens. A significant effect of the surface charge of Al hydroxide on predominantly electrostatic interactions with BSA has been demonstrated in these studies. Another study in a related field has demonstrated the substantial impact of aluminum hydroxide as the adjuvant on antigen structure and stability.38 Large number of studies has been conducted using albumins and lysozyme as model proteins. The interactions of these convenient model biopolymers with a variety of inorganic materials have been studied including Al oxide/hydroxide,35-41 titania,42 and silica.43 (15) Brinker, C. J.; Scherer, G. W. In Sol-gel science: the physics and chemistry of sol-gel processing; Academic Press: New York, 1990; p 590. (16) Perry, C. C.; Shafran, K. L. J. Inorg. Biochem. 2001, 87, 115-124. (17) Shafran, K.; Deschaume, O.; Perry, C. C. AdV. Eng. Mater. 2004, 6, 836-839. (18) Shafran, K. L.; Perry, C. C. Dalton Trans. 2005, 2098-2105. (19) Shafran, K. L.; Deschaume, O.; Perry, C. C. J. Mater. Chem. 2005, 15, 3415-3423. (20) Berthon, G. Coord. Chem. ReV. 2002, 228, 319-341. (21) Exley, C., Ed. Aluminum and Alzheimer’s disease; Elsevier: Amsterdam, 2001. (22) Igual, J. M.; Dawson, J. O. Can. J. Bot. 1999, 77, 1321-1326. (23) Copeland, L.; De Lima, M. L. J. Plant. Physiol. 1992, 140, 641. (24) Rufty, T. W., Jr.; McKown, C. T.; Lazof, D. B.; Carter, T. E. Plant Cell EnViron. 1995, 18, 1325. (25) Corain, B.; Bombi, G. G.; Tapparo, A.; Perazzolo, M.; Zatta, P. Coord. Chem. ReV. 1996, 149, 11. (26) Harris, W. R. Coord. Chem. ReV. 1996, 149, 347. (27) Haug, A.; Vitorello, V. Coord. Chem. ReV. 1996, 149, 113. (28) Exley, C. J. Inorg. Biochem. 1998, 70, 195. (29) Powell, J. J.; Whitehead, M. W.; Ainley, C. C.; Kendall, M. D.; Nicholson, J. K.; Thompson, R. P. H. J. Inorg. Biochem. 1999, 75, 167. (30) Kiss, T.; Hollo´si, M. In Aluminum and Alzheimer’s disease; Exley, C., Ed.; Elsevier: Amsterdam, 2001; p 361. (31) Harris, W. R.; Berthon, G.; Day, J. P.; Exley, C.; Flaten, T. P.; Forbes, W. F.; Kiss, T.; Orvig, C.; Zatta, P. J. Toxicol. EnViron. Health 1996, 48, 543. (32) Arnoys, E. J.; Schindler, M. Anal. Biochem. 2000, 277, 1. (33) Krejpcio, Z.; Wojciak, R. W. Pol. J. EnViron. Stud. 2002, 11, 251. (34) Jan, J. S.; Shantz, D. F. Chem. Commun. 2005, 2137. (35) Wittayanukulluk, A.; Jiang, D.; Regnier, F. E.; Hem, S. L. Vaccine 2004, 22, 1172-1176. (36) Al-Shakhshir, R. H., Regnier, F. E.; White, J. L.; Hem, S. L. Vaccine 1995, 13, 41-44. (37) Al-Shakhshir, R. H.; Regnier, F.; White, J. L.; Hem, S. L. Vaccine 1994, 12, 472-474. (38) Jones, L. S.; Peek, L. J.; Power, J.; Markham, A.; Yazzie, B.; Middaugh, C. R. J. Biol. Chem. 2005, 280, 13406-13414.

Langmuir, Vol. 22, No. 24, 2006 10079

To our mind a better understanding of the interactions of aluminum species with biomolecules is a necessary prerequisite for successful application of a combined biomimeticnanotectonic approach to advanced Al-containing materials. The use of pure and monodisperse systems of nanosized Al precursors has a clear advantage in the controlled synthesis of Al-containing materials due to an improved understanding and better control of aluminum speciation. The present contribution describes the effect of a model protein, bovine serum albumin (BSA), on the generation and some solidstate properties of hybrid Al-protein composite materials formed from various high-purity Al-containing aqueous nanosized precursors. 2. Materials and Methods 2.1. Preparation of Initial Solutions. A stock aqueous solution of aluminum chloride (1.0 M) was prepared by dissolving crystalline AlCl3‚6H2O (99%, Fisher Scientific) in distilled, deionized water. A model Al13-mer solution was prepared from the stock Al chloride solution by a soft hydrolysis techniquesstatic anion exchange described elsewhere.19 The hydrolysis ratio of the solution was adjusted to h ) 2.45, the theoretical hydrolysis ratio of Al13-mers.17 For the preparation of a model Al30-mer solution, the Al13-mer solution prepared at room temperature was thermally aged at 85 °C for 48 h as described in ref 19. Three different sols of aluminum hydroxide were prepared by adjusting the hydrolysis ratio of the stock solution of AlCl3 to h ∼ 3.0 by using the anion exchange method19 and by aging the resulting sols for 1, 2, and 6 months, which led to mean particle sizes 26 ( 3, 55 ( 6, and 82 ( 8 nm, respectively, as measured by dynamic light scattering (DLS). The final aluminum concentration in all model solutions of the Al13-mer, the Al30-mer, and Al hydroxide was 0.4 mol/L. The percent fraction of the Al13-mer was measured to be 92% of the total Al content, and the fraction of Al30-mer was 90% in the corresponding model solutions. No insoluble matter was detected in the Al13-mer and Al30-mer solutions as measured by quantitative 27Al solution NMR and a Ferron kinetic assay.19,44 The fractions of Al hydroxide in all three model suspensions exceeded 95% of the total Al content. An aqueous stock solution of BSA (50 mg/mL) was prepared by dissolving the protein (minimum 96%, from Sigma) in distilled deionized water. The solution as prepared had a pH of 7.05. 2.2. Preparation of Al-BSA Solutions. A series of model AlBSA solutions were prepared at room temperature by the addition of different amounts of fresh BSA stock solution to each of the five model Al-containing systems. All Al-BSA solutions were vigorously stirred during and after mixing of Al-containing and BSA solutions (for 60 s) using a vortex stirrer and left aging at room temperature (25 ( 0.2 °C) for 24 h. The time of aging was chosen on the basis of preliminary kinetic experiments which showed that most of the parameters of the Al-BSA mixtures (e.g. pH, viscosity, turbidity, and particle sizes) had stabilized after 24 h of monitoring. The final aluminum concentration in the Al-BSA solutions was 0.2 M, and the BSA concentration was varied from 0 to 25 mg/mL in steps of 2.5 mg/mL. 2.3. 27Al Solution NMR Spectroscopy. 27Al solution NMR spectra were acquired using a JEOL JNM-EX270 FT NMR-spectrometer operating at field strength of 6.35 T with a D2O lock using single pulse method with an x-pulse of 10 µs, 1024 scans, relaxation delay 0.1 s, and temperature 20 °C. A 5 mm glass NMR probe was used (39) Rezwan, K.; Meier, L. P.; Gauckler, L. J. Biomaterials 2005, 26, 43514357. (40) Rezwan, K.; Meier, L. P.; Gauckler, L. J. Langmuir 2005, 21, 34933497. (41) Rezwan, K.; Studart, A. R.; Vo¨ro¨s, J.; Gauckler, L. J. J. Phys. Chem. B 2005, 109, 14469-14474. (42) Oliva, F. Y.; Avalle, L. B.; Camara, O. R.; De Pauli, C. P. J. Colloid Interface Sci. 2003, 261, 299-311. (43) Coradin, T.; Coupe, A.; Livage, J. Colloids Surf., B 2003, 29, 189-196. (44) Jardine, P. M.; Zelazni, L. W. Soil Sci. Soc. Am. J. 1986, 50, 895-900.

10080 Langmuir, Vol. 22, No. 24, 2006 with a coaxial insert (both from Wilmad) filled with a secondary standard containing a known concentration of NaAlO2 (99.9%, Fisher Scientific) and 0.1 M NaOH in D2O (99.8%, BDH). An aqueous solution of Al(NO3)3 (0.016 mol/L) in HNO3 (0.1 mol/L) was used as a primary standard. Raw 27Al NMR spectra were processed using a Fast Fourier Transformation algorithm (Galactic GRAMS/32, V. 5.1); spectral apodization was carried out using a Gaussian function and line broadening of 25 Hz. Quantification of the 27Al NMR spectra was carried out using an algorithm described previously.18,19 Confirmation of the presence of Al30-mer in samples in the presence of BSA was ascertained by performing a series of variabletemperature NMR experiments from 20 to 90 °C for a sample containing 25 mg/mL BSA. The spectral results are presented in the Supporting Information. 2.4. Determination of the Residual Concentration of Al Ions after Mixing with BSA. The samples containing Al species and BSA were centrifuged for 30 min at 3000 rpm using a standard laboratory centrifuge. The centrifugation speed was chosen to prevent free BSA or Al hydroxide sedimentation. The supernatant of the solutions was then retained for residual aluminum and BSA concentration measurements and the solid material used for SEM characterization. In the case of Al13- or Al30-based samples, phase separation was not observed probably due to the very small size of the particulates, and the residual protein and aluminum concentrations where therefore not measured. The concentration of free aluminum species in supernatant solutions after centrifugation was determined using a colorimetric Ferron assay44 with preliminary acid digestion of the supernatant solutions to break down any polynuclear species and Al hydroxide that was present. For this purpose, an aliquot of 0.1 mL of the supernatant solution was mixed with 2.0 mL of 1.0 M HCl and 7.9 mL of distilled, deionized water and digested for 48 h at 60 °C in a sealed tube. A 0.2 mL aliquot of the resulting solution was then mixed with 2.3 mL of deionized water and 2.5 mL of Ferron reagent (pH 5.2), containing 2 × 10-3 M Ferron (8-hydroxy-7-iodoquinoline5-sulfonic acid, Fluka), 0.2 M acetic acid, and 0.1 M hydroxylamine hydrochloride.44 The absorbance of the solutions was measured at 370 nm, 1 h after mixing, and the overall concentration of Al ions in the supernatant was calculated from the calibration curve obtained under the same conditions using an atomic absorption standard solution of Al ions (0.986 mg/L, Aldrich) in place of the Al-BSA samples. 2.5. Measurement of the BSA Concentration. The free BSA concentration was measured by a Bradford total protein concentration assay. An aliquot of 0.1 mL of the sample supernatant was diluted in 9.9 mL of distilled deionized water. A volume of 1.0 mL of the resulting solution was mixed with 5.0 mL of the Bradford reagent containing 0.1 g/L Coomassie Brilliant Blue G-250, 5.0% ethanol, and 8.5% phosphoric acid (filtered before use) according to the method described in ref 45. The absorbance of the solutions was integrated for 15 s at a wavelength of 590 nm 5 min after mixing, and the protein concentration was calculated from the calibration curve obtained with BSA solutions of known concentration. 2.6. Measurements of pH, Conductivity, and Viscosity. The pH and the conductivity of the Al species-BSA samples were measured after aging using a PHM-250 pH-meter with Red Rod glass electrode and a temperature sensor (Radiometer). A CDM-230 conductivity meter with two-plate conductivity probe and a temperature sensor (Radiometer) was used for conductivity measurements. The pH probe was calibrated using four IUPAC buffers (pH 1.689, 4.005, 7.001, and 10.008) from Radiometer. The absolute error of pH measurements was (0.1. Viscosity measurements were performed at 25 ( 2 °C using an AND SV-10 vibro-viscometer (A&D Ltd., Japan) with gold-coated transducer and temperature sensor. The readings of viscosity were acquired in 10 mL plastic cells with fixed size and position. 2.7. Dynamic Light Scattering and ζ-Potential Measurements. DLS measurements of Al species-BSA solutions were acquired using a Coulter N5Plus particle size meter after 24 h of aging at a (45) Stoscheck, C. M. Methods Enzymol. 1990, 182, 50.

Deschaume et al. measuring angle 90.0° and acquisition time 120 s, temperature 25 °C, and equilibration time 20 min, in triplicate. Particle size distribution analysis was performed using the unimodal intensitybased model from the Coulter CONTIN program. Prior to DLS measurements the samples were diluted with distilled, deionized, and nanofiltered water (Whatman filter, 0.1 µm) and treated for 1 min in an ultrasonic bath to break up any loose aggregates. This preparation technique was used to eliminate the effects of multiple scattering (observed in concentrated sols and leading to an underestimation of particle size), viscosity variation, and dust, leading to errors in particle size measurements. ζ-potential measurements of samples containing Al species and BSA were performed using a Zetasizer Nano ZS from Malvern Instruments (Malvern, Worcestershire, U.K.) using a disposable capillary cell. Typically, 1 mL aliquot of each sample was injected into the capillary cell and measurements were made at 25 °C using field strengths of approximately 10 V/cm. The electrophoretic mobilities of the particles were determined using phase analysis light scattering (PALS) and converted into ζ potentials using the Smoluchowski approximation.46,47 To enable this conversion and to reduce the effect of viscosity, the samples were diluted 100 times with deionized water immediately before measurement. 2.8. Preparation of Solid Materials and Scanning Electron Microscopy. In the case of phase separation after centrifugation of the samples (Al hydroxide-BSA) the resulting gellike materials were rinsed with water and centrifuged again. The solid content of the samples at the bottom of the centrifuge tubes was subsequently freeze-dried for 24 h using a Virtis freeze-drier before being mounted on SEM stubs and coated with carbon using a standard procedure. The SEM images were acquired using a JEOL-JSM-840A SEM, acceleration voltage 25 keV, and working distance 8 mm. 2.9. Thermogravimetric Analysis. The weight loss, temperature, and time data arrays were collected between 20 and 900 °C using a continuous heating rate of 10 °C/min on a Stanton Redcroft TG760 balance connected to a computer. The weight loss curve was normalized to the weight of the sample at 900 °C. The dm/dT derivatives were then calculated, and the resulting DTG curves fitted using Gaussian functions in the peak-fitting module of Thermo Galactic Grams/32 software. 2.10. Powder X-ray Diffraction Measurements. X-ray powder diffraction data were acquired at room temperature using a Panalytical X′Pert Pro powder diffractometer operating with Cu KR1 radiation, continuous scans from 2θ ) 1 to 80°, step size 0.002° in 2θ, scan step time 30.9 s, and continuous spinning of the sample during the run.

3. Results and Discussion The series of Al-BSA samples prepared from five model solutions of aluminum species (Al13, Al30, and three Al hydroxide sols with different particle sizes) containing various amounts of BSA were analyzed using a range of solution and colloidal techniques. After careful centrifugation of the insoluble part of the samples, the precipitated solid was freeze-dried and analyzed using several solid-state techniques as well as solution techniques after redissolution of the obtained materials. 3.1. Physical Properties and pH of the Al-BSA Solutions. Conductivity, pH, and viscosity measurements of the samples prepared from Al species and BSA were carried out simultaneously, after aging the samples for 24 h. The data are presented in Figure 2. As follows from Figure 2A, the initial conductivity of the pure aluminum species solutions varies in the following order: Al30mer > Al13-mer > Al hydroxide (26 nm) > Al hydroxide (55 nm) > Al hydroxide (82 nm). With increasing BSA concentration, solution conductivity decreases for Al polycations (∼5% decrease (46) McNeil-Watson, F.; Tscharneuter, W.; Miller, J. F. Colloids Surf., A 1988, 140, 53. (47) Miller, J. F.; Schatzel, K.; Vincent, B. J. Colloid Interface Sci. 1991, 143, 532.

Interactions of BoVine Serum Albumin

Langmuir, Vol. 22, No. 24, 2006 10081

Figure 2. (A) Conductivity, (B) pH, and (C) viscosity (C) of Al species-BSA samples prepared at various BSA concentrations.

from the initial value for the Al13 and ∼10% for the Al30), while remaining practically unchanged for the samples prepared from Al hydroxide (Figure 2A). The initial pH of the three Al hydroxide mixtures with BSA (Figure 2B) shows an inverse trend to that of conductivity, the Al30-BSA system showing the lowest initial pH of 3.9. The pH of Al13-BSA solutions increases with increasing BSA concentration, reaching the upper pH limit of Al13-mer stability (pH ∼ 4.6).14,17 The pH of the Al30-BSA samples decreases slightly when BSA concentration increases from 0.0 to 7.5 mg/mL and then tends to increase for higher BSA concentrations. However, the overall pH change in the case of the Al30-BSA systems does not exceed 0.1 units of pH, and the pH value never exceeds pH ∼ 4.0. The initial pH of the Al hydroxide-BSA systems is systematically higher than that of the Al polycations and BSA and continues to increase with increasing BSA concentration. There are small differences between the values of pH for the Al hydroxide-BSA solutions prepared from the hydroxide sols with varying particle sizes. According to the data of Figure 2C, the initial viscosity of Al hydroxide sols without BSA is approximately the same as that of pure water (0.98 ( 0.02 cP at 25 °C). The viscosity of the Al hydroxide-BSA mixtures increases exponentially with increasing BSA concentration up to 17.5 mg/mL of BSA in solution, reflecting the increasing gelation processes. The growth of viscosity is much more apparent for the samples prepared from Al hydroxide particles with 26 and 55 nm diameter. For these samples, the strong gel obtained inside the reaction vessels above a certain BSA concentration could not be transferred into the viscometer cell without damaging the gel structure and biasing viscosity readings. Therefore, no measurements of viscosity were made for the samples of Al hydroxide (26 nm particle diameter)

containing more than 12.5 mg/mL of BSA and for the samples with 55 nm diametersabove 17.5 mg/mL BSA concentration. Both Al polyoxocation-containing systems (Al13-BSA and Al30-BSA) showed no significant increase of viscosity whatever the BSA concentration (viscosity values remained close to that of pure water at 1.00 ( 0.02 cP), in drastic contrast to hydroxidecontaining systems (Figure 2C), indicating little or no gelation for the systems in question. 3.2. Dynamic Light Scattering and ζ-Potential Measurements. DLS measurements were performed to follow particle size evolution in the model Al-BSA systems. The results of these measurements are presented in Figure 3 as a function of BSA concentration. The relative standard deviation of the particle size data presented in Figure 3 did not exceed ∼5% for the Al13-based systems, ∼3% for the Al30-based systems, and ∼2% for all Al hydroxide-based systems (measured in triplicate). In the case of samples prepared from the three different Al hydroxide sols, the measured particle sizes increase with growing BSA concentration (Figure 3A). The relative slope of the inclining particle size curve as well as the largest particle size measured appears to be inversely related to the particle size of the initial Al hydroxide suspension. For the 26 nm Al hydroxide sol, the particle size growth is most rapid and the largest particle size observed by DLS (∼1200 nm) is the highest. In contrast, the Al hydroxide system initially containing 82 nm particles shows the slowest growth of particle size with increasing BSA concentration, and the largest particle size is between 200 and 400 nm at BSA concentrations of 22.5-25 mg/mL (Figure 3A). Therefore, the propensity of Al hydroxide particles to aggregate in the presence of increasing concentrations of BSA, possibly acting as a coagulation agent similarly to the work,50 arises from the overall surface area of the suspension, which in turn depends on particle size.

10082 Langmuir, Vol. 22, No. 24, 2006

Deschaume et al.

Figure 4. Evolution of ζ-potential of Al species-BSA samples as a function of BSA concentration: (A) Al13-mer-BSA and Al30mer-BSA samples; (B) Al hydroxide-BSA samples.

Figure 3. Evolution of average particle sizes of Al species-BSA samples as a function of BSA concentration measured by DLS at a scattering angle of 90°: (A) Al hydroxide-BSA samples with different Al hydroxide particle sizes; (B) Al hydroxide-BSA samples prepared from 82 nm Al hydroxide with adjusted pH; (C) Al13mer-BSA and Al30-mer-BSA samples.

In the case of the 82 nm Al hydroxide sols, the evolution of particle size with BSA concentration has been followed for three initial values of pH of the initial Al hydroxide sol (Figure 3B). At an initial pH of 3.8, the particle size varies over a relatively narrow range (100-150 nm) above a BSA concentration of 2.5 mg/mL and tends to decrease with increasing amounts of BSA. When the Al hydroxide sol was adjusted to a higher pH (4.2 or 4.7), the particle size increases with BSA concentration according to a preset value of the pH, especially at high BSA concentrations (>15 mg/mL). (48) Novak, J. T.; Park, C. Water Sci. Technol. 2004, 49, 73. (49) Anzai, J. I.; Guo, B.; Osa, T. Nucleic Acids Res. 1996, 40, 35-40. (50) Rezwan, K.; Meier, L. P.; Rezwan, M.; Vo¨ro¨s, J.; Textor, M.; Gauckler L. J. Langmuir 2004, 20, 10055-10061.

The data of Figure 3B indicates that pH is an important parameter governing the aggregation-gelation process in the Al hydroxide-BSA systems. At pH < 4.0 these systems show only limited aggregation, while at higher pH the system is allowed to aggregate and thence to gel, especially in the presence of high BSA concentrations. As follows from Figure 3C, the particle size of the pure Al13mer and Al30-mer solutions was found to be close to the actual sizes of these species (1 ( 0.3 and 2 ( 0.7 nm, respectively18). Once BSA is added, the average particle size in all cases immediately increases to ∼10.75 ( 0.5 nm. For the Al13-merBSA samples the mean size of the suspension does not change profoundly with further increase of BSA concentration, while in the case of the Al30-containing solutions the average particle size tends to increase at BSA concentrations above 17.5 mg/mL (Figure 3C). The mean diameter of the BSA molecule alone that was estimated to be ∼8-9 nm from preliminary DLS measurements (the actual size of the BSA molecule is 4 nm × 4 nm × 14 nm49,50). The larger-than-expected mean particle size of Al polycation-BSA samples could arise, for instance, from the absorption of Al13-mer clusters on the “surface” of BSA molecules that are negatively charged under mildly acidic conditions (pH < 5.0). These hypothetical clusters would have a slightly larger size, e.g. 10-11 nm in the case of one layer of Al13 around 8 nm-sized BSA molecule and, possibly, more than 12 nm in the case of an Al30 layer than might have been anticipated. At constant Al polycation concentration and increasing BSA amounts the average particle size observed with DLS would decrease and eventually get closer to the original BSA size. The hypothesis above is supported by ζ-potential measurements of the model systems in question (Figure 4).

Interactions of BoVine Serum Albumin

Figure 5. (A) Free BSA concentration and (B) aluminum ion concentration in the supernatant solutions of the Al hydroxideBSA samples after centrifugation.

The ζ-potential value for the Al polyoxocation-BSA system is positive at all BSA concentrations measured, with only small differences between Al13- and Al30-containing samples being observed (Figure 4A). Solutions containing Al13 and Al30 without BSA could not be measured as the size of these species were too small to be detected. Thus, the acquired positive values of ζ-potential result from the presence of BSA molecules which in pure solutions of similar concentration showed negative potentials (∼-8.6 ( 0.08 mV). The observed charge reversal of BSA can be explained by the adsorption of Al polyoxocations on the surface of protein. With increasing concentration of BSA the ratio of Al polyoxocations to protein falls constantly resulting in decrease of ζ-potential (Figure 4A). For Al hydroxide, the value of ζ-potential is also positive but more than two times higher than for the Al polyoxocation-BSA solutions at similar concentrations (Figure 4B). The ζ-potential of the Al hydroxide-BSA system also tends to drop with increasing amounts of BSA. This behavior can be explained by the adsorption of negatively charged BSA molecules on highly charged particles of Al hydroxide with a resulting cancellation of charge. 3.3. Residual Concentrations of BSA and Aluminum Ions. To clarify the extent of Al hydroxide-BSA composite formation after aging of the samples, free supernatant concentrations of BSA and aluminum were measured as a function of BSA concentration after removal of the insoluble matter by centrifugation as described in the Experimental Section. The results are presented in Figure 5 for the samples prepared from Al hydroxide sols with varying particle sizes. The average measurement error did not exceed 0.02 mg/mL for BSA and 0.005 mol/L for Al ions.

Langmuir, Vol. 22, No. 24, 2006 10083

As follows from Figure 5A, the free BSA concentration in the supernatant of the Al hydroxide-BSA samples increases with increase of total protein concentration for all three Al hydroxide suspensions. For samples containing 26 nm diameter Al hydroxide particles, “free” BSA concentration reaches the maximum at a total BSA level of 7.5 mg/mL. For the system containing 82 nm diameter Al hydroxide particles, the maximum of “free” BSA is at ∼15 mg/mL of a total BSA content. The system containing 55 nm diameter Al hydroxide particles appears to follow an intermediate trend. At higher BSA contents, the amount of “free” BSA decreases significantly for all three Al hydroxide systems. Perhaps, at this point BSA-Al hydroxide aggregates (or gels) become large enough to be separated from solution by lowspeed centrifugation. “Free” total aluminum concentration in the centrifugate solution continuously decreased with increasing total BSA concentration in the systems containing Al hydroxide (Figure 5B). This decrease was equally substantial for all three suspensions tested (from ∼100% to ∼35% of the total Al concentration which was 0.2 mol/L initially). The curves generally followed a sigmoidal trend indicating that association of BSA and Al hydroxide particles occurs in a “titration-like” fashion. The trends discussed above probably arise from the formation of colloidal, self-assembled composites in BSA-Al hydroxide mixtures. The extent of the hybrid composite formation is indicated by the sedimentation of these particles by centrifugation and, consequently, lower “free” concentrations of Al species and BSA in the supernatant. One can hypothesize that, for low BSA concentrations, the particle size is probably below the sedimentation threshold set by the gentle centrifugation speed of 3000 rpm, whereas, for higher BSA amounts, the Al hydroxide-BSA “conjugates” are quickly separated and the increase in total BSA produces a decrease in free aluminum concentration. This hypothesis is supported indirectly by the DLS data (Figure 3A) and viscosity measurements (Figure 2C) which show the particle size and the viscosity in the Al hydroxide-BSA systems increasing with increasing total BSA concentration. 3.4. Redissolution Experiments and 27Al Solution NMR. The results described above were concerned with the solution characterization of the obtained Al-BSA model systems. In the case of the Al polycation-BSA systems, a fraction of these solutions was freeze-dried to obtain solid products. Freeze-drying was chosen as a suitable method for quenching all reactions in solution and for a gentle, nondestructive transformation of solution species into solid state. For Al hydroxide-BSA systems, the part sedimented by centrifugation was subjected to freeze-drying. As has been shown by measurement of the residual Al and BSA concentrations (Figure 5), substantial amounts of the protein and the Al hydroxide were in the insoluble phase after centrifugation, especially at high BSA concentrations. In the case of Al polycation-BSA mixtures, there was no significant phase separation after centrifugation. Therefore, a part of the native solutions was freezedried and analyzed in the solid state. For the Al hydroxide-BSA systems, no measurable amounts of soluble Al species were detected by 27Al NMR spectroscopy (data not presented). In contrast, all samples prepared from Al polycations and BSA were soluble. The first step of analysis of the freeze-dried samples was to redisperse them in pure water, to identify and measure amounts of individual Al species in redissolved samples. Such analysis has been carried out using quantitative 27Al solution NMR spectroscopy according to the procedure developed earlier.17,18

10084 Langmuir, Vol. 22, No. 24, 2006

Figure 6. 27Al solution NMR spectra of the (A) Al13-mer-BSA and (B) Al30-mer-BSA samples as a function of total BSA concentration. T ) 25 °C. The peak at 80 ppm arises from the internal reference solution of aluminate ions in D2O.

The NMR spectra of all of the Al polycation-BSA systems before and after the freeze-drying/redissolution procedure were acquired. The spectra of the initial Al polyoxocation-BSA samples (before freeze-drying) are presented in Figure 6, as an illustrative example. Redissolution was carried out with accurately weighed amounts of solid materials in a fixed volume of distilled, deionized water (normally, 1 or 2 mL). 27Al NMR spectra acquired from the Al -mer-BSA solutions 13 show peaks at ∼63 ppm (tetrahedral core of the Al13-mer) and ∼0 ppm (octahedral signal of Al monomers) along with a signal at 80 ppm arising from aluminate ions of the internal reference solution6 (Figure 6A). In the spectra of the Al30-mer-BSA systems (Figure 6B), along with the mentioned signals at ∼63 and ∼0 ppm, a broader signal at 70 ppm is observed, which corresponds to the tetrahedral “core” Al atoms of the Al30-mer.8 Unambiguous assignment of the signal at 70 ppm has been made on the basis of variable-temperature NMR experiments that are provided as Supporting Information. On the basis of the 27Al NMR spectra of the Al polyoxocationBSA samples before and after freeze-drying/redissolution, Al speciation diagrams were calculated using an algorithm described previously.17,18 The diagrams, which include Al monomers, Al13mers, Al30-mers, and Al hydroxide, are presented in Figure 7. The amount of Al hydroxide, normally undetectable by means of 27Al solution NMR, was estimated indirectly as a difference of the total Al content (either in solution or in solid state, depending on the experiment) and the sum of the fractions of all soluble species (Al monomers, Al13-mers, and Al30-mers). The concentration of small Al oligomers (e.g. Al dimers or trimers) was not taken into account in this study due to the very low concentration of these species, as indicated by the absence of the associated

Deschaume et al.

Figure 7. 27Al solution NMR spectra of the (A) Al13-mer-BSA and (B) Al30-mer-BSA samples as a function of total BSA concentration. T ) 25 °C.

broad peak at ∼2-4 ppm or its strong overlap with larger peaks at ∼0 and ∼10 ppm (cf. Figure 6). The immediate conclusion following from the data of Figure 7 is that generally there is a good correlation between all measured species in the samples before and after freeze-drying/redissolution procedure. Therefore, such a procedure does not significantly disturb the distribution of soluble species in the Al polycationBSA samples. The concentration of the major species, the Al13mer in the initial Al13-mer-BSA samples before and after freezedrying/redissolution, decreases from ∼98-99% to ∼90% upon increasing BSA from 0 to 25 mg/mL. This decrease is correlated with a decline of Al monomer concentration (from ∼2% to ∼0.2%) and anticorrelated with the increase of insoluble matter (from ∼0% solution NMR, probably due to an increase of solution pH (Figure 2B)). However, these hydroxide particles do not aggregate strongly due to the presence of BSA in solution, as supported indirectly by the constant viscosity of the samples with various BSA concentrations (Figure 2C). The data of Figure 7B indicates that, in the case of the Al30mer-BSA systems, there is a strong effect of increasing BSA concentration on the initial and freeze-dried samples. The concentration of the major species (Al30-mers) decreases from ∼80-86% to ∼62-65% upon increase of BSA concentration. A similar trend, but to a lesser extent, is observed for the minor quantities of the Al13-mer and Al monomers. The general trends observed for the fractions of soluble species detected in the Al30mer-BSA (Figure 7B) samples do not correlate with the corresponding pH trends in Figure 2B. The apparent contradiction is probably due to the stronger acidity of the Al30-mer or higher concentration of Al monomers in the Al30-mer-containing

Interactions of BoVine Serum Albumin

Langmuir, Vol. 22, No. 24, 2006 10085

Figure 8. Results of thermogravimetric analysis of the system Al hydroxide (26 nm)-BSA: (A) raw “weight loss” curves; (B) the first derivative thermogravimetric (DTG) curves; (C) example of the peak fitting used for semiquantitative analysis of the DTG curves; (D) DTG peak areas as a function of BSA concentration.

solution, compared to the Al13-mer samples. Higher acidity makes the Al30-mer solution more sensitive to BSA additions, and it would require higher BSA concentrations or a base stronger than BSA to increase the pH substantially. 3.5. TGA Measurements. As mentioned above, the model systems containing pure Al hydroxide and BSA could not be investigated directly using redissolution experiments and 27Al solution NMR. Therefore, solid-state techniques such as X-ray diffraction, thermogravimetry, and SEM were employed to shed light on the speciation of these systems. The result of thermogravimetric analysis of the solid samples of Al hydroxide (26 nm)-BSA model systems with various BSA content prepared by freeze-drying of the respective solutions are presented in Figure 8. The original TGA data normalized to the final “dry” weight of the samples, Figure 8A, varied in a systematic fashion with increasing BSA concentration. The data were smoothed using an FFT algorithm (Origin 6.1 software), and the first derivative (DTG) curves were calculated (Figure 8B). The DTG curves revealed several peaks with the maxima indicated in the legend, which were fitted using Gaussian functions in ThermoGalactic Grams/32 software as shown in the illustrative example in Figure 8C. The position of the maxima of the fitted DTG peaks at 49 ( 4, 157 ( 20, and 239 ( 12 and 354 ( 33, 459 ( 63, and 604 ( 87 °C remained relatively constant, i.e., within 10-15%, with increasing BSA in the initial samples. The DTG curve of pure freeze-dried BSA contained a peak at 300 °C which was not observed for the Al hydroxide (26 nm)-BSA samples. This modification of the DTG pattern of the protein can be explained by a destabilization of BSA upon binding to the Al hydroxide surface leading to conformational modifications.38 Following the treatment of the TGA data described above, a semiquantitative analysis of the peak areas of the DTG profiles

was performed. Areas of the DTG peaks at 49, 354, 459, and 604 °C tended to increase with increasing BSA content of the samples, which suggests their association with the thermal degradation of BSA. The other two peaks at 157 and 239 °C showed an opposite trend and, therefore, can be preliminarily attributed to the inorganic component of the analyzed composites (Al hydroxide). By summing the areas of the DTG peaks proposed to have arisen from BSA and from Al hydroxide, it was possible to reconstruct the trends observed for the bioorganic and inorganic components of the composites, as shown in Figure 8D. The Alrelated DTG trend associated with Al hydroxide correlates quite well with the mass fraction of the dry residue that is most probably pure Al2O3. The sum of BSA-related DTG peaks shows a trend opposite to that of the inorganic component (Figure 8D), which, in turn, is in good agreement with the solution data on the concentration of “free” BSA and aluminum in solutions after centrifugation (Figure 5). 3.6. Results of Powder X-ray Diffraction. Selected samples of the Al-BSA systems containing various Al species and different amounts of BSA were analyzed using X-ray diffraction (Figure 9). XRD patterns of the Al13-mer-BSA system (25 mg/mL of BSA) shown in Figure 9 contains an intense peak composite peak at ∼9.6° in 2θ and several other very broad features which possibly indicate the presence of a crystalline phase with extremely small crystallite size (roughly estimated to be 1-3 nm from Sherrer’s formula). We have compared the diffractograms (a and b) in Figure 9 with the PXRD data available for the Al13-mer sulfate salt51 and our PXRD measurements of the Al13-mer chloride salt.19 However, there was no complete match with any (51) Johansson, G. Ark. Kemi 1962, 20, 321.

10086 Langmuir, Vol. 22, No. 24, 2006

Figure 9. XRD patterns obtained from the Al-BSA materials prepared: (a) Al13-BSA 0 mg/mL; (b) Al13-BSA 25 mg/mL; (c) Al30-BSA 0 mg/mL; (d) Al30-BSA 25 mg/mL; (e) Al hydroxide (d ∼ 26 nm)-BSA 0 mg/mL; (f) Al hydroxide (d ∼ 26 nm)-BSA 25 mg/mL.

Deschaume et al.

For the samples obtained from 82 nm Al hydroxide and BSA, a similar observation can be made (Figure 10). However, the sheets obtained from pure 82 nm Al hydroxide are more disorganized in comparison with the ones prepared from smaller particles, and the spacing between the nanoribbons is larger. For the lowest BSA concentrations, a stacking of the sheets is observed, the spacing of the ribbons becoming more chaotic. Finally, at the largest BSA concentration used, the sheets are much smaller and completely disorganized. SEM observations of the morphology of Al hydroxide-BSA systems indicate that the effect of hydroxide particle size is important. With increasing particle size of Al hydroxide, the nanoribbon morphology gets sharper and more ordered (at least in the two dimensions of the sheets). In the case of pure Al13-mer, the freeze-dried material is largely disorganized on the micrometer level, although relatively thick (3-6 µm wide) ribbons can be observed by SEM (Figure 10). Upon BSA addition, the samples containing Al13-mers change progressively until finally they exhibit the same morphology as for pure hydroxide-based materials. A similar picture is observed for the Al30-mer-based materials. Freeze-drying of pure Al30-mer solutions leads to a mixed fibrous and sheetlike structure which progressively disappears with addition of higher amounts of BSA. An increasing number of large stacked sheets comprising 3-5 µm diameter holes is then observed in this case. As BSA concentration increases, sheets with nanoribbons similar to those observed for pure Al hydroxide samples prevail. The reason for the “nanoribbon-sheet” morphology of Al hydroxide-containing samples is not clear. This morphology could arise either from the freeze-drying procedure itself or from the preferential packing of Al hydroxide particles into symmetrically located two-dimensional structures. Further morphological and spatial chemical analysis is required to explain this phenomenon. Nevertheless, whatever the origin of the nanoribbon morphology, it remains a good “fingerprint” of the presence of significant amounts of Al hydroxide.

4. Conclusions

Figure 10. SEM pictures of Al species-BSA hybrid materials prepared by freeze-drying.

of these patterns. Therefore, the exact identification of the crystalline phase was not possible in this case. Perhaps the presence of high amounts of BSA does not completely prevent crystallization of the Al13-mer chloride, although the crystallite size in this case is likely to be much lower than the bulk freezedried Al13-mer solution.19 For the Al30-mer containing samples, there is no crystallization of the Al30-mer in the chloride form, whatever the concentration of BSA, as follows from the PXRD patterns of pure and BSAcontaining freeze-dried samples (Figure 9c,d). The same conclusion follows from the PXRD patterns acquired for the systems containing Al hydroxide (∼26 nm) (Figure 9e,f). 3.7. SEM Observations. The solid products obtained by freezedrying of various Al species-BSA samples were subjected to SEM analysis to observe the morphology of the prepared materials. The pure 26 nm hydroxide-based samples without BSA exhibited a morphology characterized by sheets covered with equally spaced parallel ribbons (Figure 10). At increasing BSA concentration, the ribbons disappear and the sheets become increasingly disorganized. For the highest BSA concentrations, the sample consists only of disorganized flakes and small sheets.

Aluminum-BSA nanohybrid materials have been obtained by mixing aluminum hydroxide nanosols, as well as Al13-mer or Al30-mer solutions with varying amounts of BSA. The composition and formation mechanisms of these materials have been investigated using various solution- and solid-state characterization techniques. The chosen time scale of aging of AlBSA mixtures (24 h) allowed us to concentrate on the long-term interactions of Al species with BSA at conditions close to equilibrium. All five Al-containing precursors used in this study possess significant positive surface charge which correlates well with the initial pH of the corresponding solutions (suspensions) and decreases in the following order: Al30-mer (pH ) 3.85) > Al13mer (pH ) 4.26) > Al hydroxide 1 (26 nm, pH ) 4.61) > Al hydroxide 2 (55 nm, pH ) 4.73) > Al hydroxide 3 (82 nm, pH ) 4.75). The formal charges on the Al30-mer and Al13-mer are +18 and +7 correspondingly, while the high charge on the Al hydroxide nanoparticles has been indicated by high values of ζ-potential measurements (Figure 4). The observed ζ potentials (51.6 and 54.5 mV) of the as-prepared Al hydroxide suspensions are in agreement with their small mean particle sizes (82 and 55 nm) as well as with relatively low pH values of the suspensions (Figure 2B). The BSA molecule with isoelectric point at pH ∼ 4.67-5.0 represents a very weak polyacid showing a range of acidities

Interactions of BoVine Serum Albumin

Langmuir, Vol. 22, No. 24, 2006 10087

Figure 11. Schematic representation of the interactions of BSA with various Al species.

and, therefore, an inhomogeneous distribution of negative charges.49 Thus, for most cases studied, the interaction between BSA and positively charged Al polycations or Al hydroxide nanoparticles will be predominantly electrostatic, as has been shown previously.35-41 Such interactions presumably lead to coverage of BSA molecules by smaller Al polycations. The resulting particles are slightly larger than the original BSA molecules according to the DLS data (Figure 3C) and could be termed hybrid “core-shell” nanoparticles. The overall charge of these agglomerates depends on the BSA concentration (Figure 4A). There could be a certain point, when the concentration of BSA is high enough to partially cancel the charge of Al polycations and some properties of the system could become similar to those of Al hydroxide, as evidenced indirectly by the identical morphology of the freeze-dried samples (Figure 10). However, unlike for Al hydroxide-BSA systems, the hybrid particles of Al polycations with the protein remain soluble. The increasing concentration of BSA appears to force Al13mers, and especially Al30-mers, to convert partially into Al hydroxide, apparently, as a result of an acid-base reaction. This is supported by 27Al solution NMR data showing a progressive decrease of soluble Al species (Figure 7A,B). We assume that Al polycations could act as acids due to their still high charge/Al atom ratio,52 while BSA could act as a base, which is possible, taking into account the heterogeneity of the functional groups on the surface of this protein.49,50 Substantial decrease of conductivity, more profound for the Al30-mer which is a stronger acid than for the Al13-mer, is observed indicating binding of Al polycations to BSA or partial conversion into Al hydroxide. The extent of Al hydroxide formation and its particle size is very limited, as follows from the absence of gelation and large aggregates, as measured by viscometry and DLS (Figures 2C and 3C). Another explanation for the observed decrease in concentration of Al polycations is the formation of insoluble BSA-“Al” aggregates that are not detected by solution NMR. The ζ-potential measurements provide supporting evidence for the predominantly electrostatic interactions between the “Al”containing species and BSA (Figure 4). Correlation of increasing BSA concentration with decreasing ζ-potential indicates that the negative charge of BSA is being canceled by either Al polyoxocations or Al hydroxide depending on the experimental system under investigation. When solutions containing Al polyoxocations and BSA are forced into the solid state by freeze-drying, there is no significant change in Al speciation, as shown by 27Al NMR spectroscopy after redissolution of freeze-dried materials in water (Figure 7). Powder XRD provides a hint that crystallization of the Al13-mer (52) Jolivet, J.-P. Metal Oxide Chemistry and Synthesis. From Solution to Solid State; Wiley & Sons: Chichester, U.K., 2000.

and Al30-mer salts is suppressed in the presence of high BSA concentrations (Figure 9). The overall result of interactions between BSA and Al hydroxide is most probably the formation of a hybrid material in which BSA molecules are adsorbed on the surface of larger Al hydroxide particles. These “core-shell” hybrid nanoparticles are the reverse of those found from the Al polycations as now the hydroxide particles form the core that is decorated by BSA molecules. This process is promoted by increasing BSA concentration, as supported by the dramatically growing average size of the mixtures (Figure 3C), higher values of viscosity and pH (Figure 2C,B), and positive values of the measured ζ-potentials (Figure 4B). The conductivity of these model systems does not change in this case (Figure 2A), as there are practically no soluble Al species involved in the process. Measurements of free BSA indicate (Figure 5) that a large part of the protein is involved in the formation of a composite gellike structure that was amorphous to X-rays after centrifugation and freeze-drying (Figure 9). Free Al concentration varies as a function of BSA content in a titration-like manner for all three Al hydroxide systems studied. Therefore, increasing amounts of BSA cover the surface of Al hydroxide particles gradually, thereby neutralizing positive surface charge and creating favorable conditions for the aggregation process as also supported by ζ-potential measurements (Figure 4). Semiquantitative thermogravimetric analysis (Figure 8) shows a systematic increase of BSA content and a corresponding decrease of Al hydroxide in the freeze-dried samples with increasing BSA amounts (Figure 8D). However, this trend is not linear and reaches a “steady state” above BSA ∼15 mg/mL. This fact could indicate a “saturation limit” of the surface coverage of Al hydroxide by BSA molecules. Moreover, a modification of the BSA molecule can be evidenced from the difference between the DTG trace of the pure protein and of the protein associated with Al hydroxide. The fact that the largest particle size increase observed by DLS is for the system with smallest initial hydroxide particles (26 nm, Figure 3A) indicates that the surface area of the Al hydroxide particles (which is higher for the 26 nm suspension) plays an important role in the aggregation-gelation process induced by BSA. The value of pH also strongly affects the size of the resulting Al hydroxide-BSA aggregates (Figure 3B), formed as a result of predominantly electrostatic interactions. The scheme in Figure 11 serves to summarize the suggested mechanisms of interaction of various Al species and BSA and to relate them to the morphology of the freeze-dried materials. As follows from the scheme, the inorganic “core” of Al hydroxide is gradually covered by organic “shell” of BSA molecules and aggregation probably proceeds via protein molecules having their effective overall charge reduced to zero. In contrast, Al polycations

10088 Langmuir, Vol. 22, No. 24, 2006

cover the surface of BSA molecules leading to a charged hybrid “core-shell” particles with organic “core” and inorganic “shell”. The positively charged “shell” protects the hybrid particles from aggregation. However, the mechanisms of interaction described above are in a complex interplay, as evidenced by the change in morphology of the resulting Al species-BSA materials. For Al polycationBSA materials the morphology is highly disorganized at low BSA content and becomes structured in two dimensions as very thin sheets with “nanoribbons” at higher BSA levels (Figures 10 and 11). The Al hydroxide-BSA system shows an opposite trend. As Al hydroxide particles are being progressively surrounded by increasing amounts of adsorbed BSA, at some point all of them become separated from one another. The interaction between particles occurs through BSA layers. This process results in the destruction of the “nanoribbon” morphology pertinent of pure Al hydroxide and complete “randomization” of the solid materials at high BSA content. The occurrence of the nanoribbon morphology which was reproducibly observed in pure Al hydroxide and Al polycationBSA systems could be in part a result of the freeze-drying process used for the preparation of the solid materials. Nevertheless, the reproducibility of the pattern suggests it is a “fingerprint” of the systems from which this morphology was obtained.

Deschaume et al.

This study has shed light mainly on aqueous aluminum speciation in the presence of a model protein and the various interactions occurring between the bioorganic and inorganic parts of the system in question, without an in-depth study of the changes occurring in the protein molecule itself. This aspect of our model systems could be investigated further, as well as fine structure and surface properties of the “core-shell” composites. The use of aqueous processing for the generation of composite materials by the self-assembly of highly pure Al species and biopolymers has the potential for being an environmentally friendly, extremely soft, and easy route toward advanced materials. We hope that nanohybrid composites similar to those obtained in this study will find interesting applications due to their intricate structure and morphology, as well as becoming precursors of choice for the synthesis of other advanced Alcontaining materials with unusual properties. Acknowledgment. We gratefully acknowledge the generous help of Dr. Michael Kaszuba of Malvern Instruments (Malvern, Worcestershire, U.K.) with the measurement of ζ-potentials. Supporting Information Available: Variable-temperature 27Al solution NMR data. This material is available free of charge via the Internet at http://pubs.acs.org. LA061285H