Langmuir 2005, 21, 7029-7035
7029
Surface Modification of Magnetic Nanoparticles with Alkoxysilanes and Their Application in Magnetic Bioseparations Ian J. Bruce* and Tapas Sen Department of Biosciences, University of Kent, Canterbury, Kent CT2 7NJ, United Kingdom Received March 1, 2005. In Final Form: May 5, 2005 A versatile and inexpensive method for the introduction of amine groups onto the surface of silica-coated magnetite composite nanoparticles has been established based on the condensation of (aminopropyl)triethoxysilane (APTS). The process was observed to be sensitive to a range of variables, and a range of silane surface-modified nanoparticles was synthesized under various reaction conditions, that is, solvent systems [water, tetrahydrofuran (THF), ethanol, or 1:1 mixtures of them], reaction times (from 1 to 24 h), and temperatures (18, 50, and 70 °C), with water as the catalyst and silane at either 0.2% or 2% (w/v) in an attempt to optimize the process. The products of the various reactions were characterized in terms of their possession of surface -NH2 groups, morphologies, and properties with respect to DNA binding and elution before being modified with a single-stranded oligonucleotide capture sequence. It was observed that careful manipulation of temperature, time, and solvent conditions was important for optimal silanization of the nanoparticles, and in our experiments best results were obtained when silanization of the particles in suspension involved use of water as the solvent and APTS at 0.2% (w/v) and when the reaction was conducted at room temperature for 5 h and was preceded by ultrasonication of the particle suspension. The materials produced were used in experiments to selectively capture complementary nucleic acid sequences by hybridization after grafting with an oligonucleotide. The efficiency of the oligonucleotidemodified particles in the capture experiments was observed to be directly related to the original density of amine groups present at the surface of the support. The results indicate that surface engineering of the nanoparticles was possible by silanization under defined, optimized conditions. This approach could be extended to the activation of such surfaces and other materials with other functional groups.
Introduction Surfaces grafted with biomolecules are important for modern biotechnology and life sciences. For example, processes using surface-activated nanoparticles include liquid chromatography,1 magnetic separation,2-4 and multidetection systems based on biosensors in array formats (biochips).5-8 These are all examples of technologies where the controlled, optimized attachment of biomolecules to solid surfaces plays a crucial role in their ultimate utility. The success of any immobilization process relies on a careful balance of the intermolecular forces driving the interaction between the molecules to be grafted and the outermost layer of the substrate surface. In this context, surface chemistry has evolved as a powerful tool to tailor surface properties.9-11 In fact, it can even address issues * Corresponding author: phone/fax 0044 1227 824557; e-mail
[email protected]. (1) Leonard, M. J. Chromatogr. B 1997, 699, 3-27. (2) Bucak, S.; Jones, D. A.; Laibinis, P. E.; Hatton, T. A. Biotechnol. Prog. 2003, 19, 477-484. (3) Karumanchi, R.; Doddamane, S. N.; Sampangi, C.; Todd, P. W. Trends Biotechnol. 2002, 20, 72-78. (4) Saiyed, Z. M.; Telang, S. D.; Ramchand, C. N. Biomagn. Res. Technol. 2003, 1, 2 (5) Cutler, P. Proteomics 2003, 3, 3-18. (6) Lee, Y. S.; Mrksich, M. Trends Biotechnol. 2002, 20, S14-S18. (7) Lee, S. Y.; Lee, S. J.; Jung, H. T. J. Ind. Eng. Chem. 2003, 9, 9-15. (8) Phelan, M. L.; Nock, S. Proteomics 2003, 3, 2123-2134. (9) Jon, S. Y.; Seong, J. H.; Khademhosseini, A.; Tran, T. N. T.; Laibinis, P. E.; Langer, R. Langmuir 2003, 19, 9989-9993. (10) Schaeferling, M.; Schiller, S.; Paul, H.; Kruschina, M.; Pavlickova, P.; Meerkamp, M.; Giammasi, C.; Kambhampati, D. Electrophoresis 2002, 23, 3097-3105. (11) Xu, C.; Xu, K.; Gu, H.; Zheng, R.; Liu, H.; Zhang, X.; Guo, Z.; Bing, X. J. Am. Chem. Soc. 2004, 126, 9938.
such as preservation of the native conformation of the immobilized molecule or its orientation with respect to the surface so as to ensure proper function. These are particularly important considerations in a biological context. The great diversity demonstrated by biomolecules ideally requires flexible but generic methods for materials surface modification that permit any particular molecular species’ immobilization by a standard approach. Surface modification with organosilanes is an attractive approach in this context as it is compatible with many of the materials used in a biological context, i.e., silica gel, glass slides, or silicon wafers.12 Organosilanes are bifunctional molecules with the general formula X-(CH2)n-SiRn(OR′)3-n, where X represents the headgroup functionality, (CH2)n a flexible spacer, and Si(OR)n the anchor groups by which (after hydrolysis of the alkoxy group) they can attach to free Si-OH surface groups. Alkoxysilanes with a variety of X functionalities are commercially available, amino groups being one of the most frequently used for biological applications. These can be employed in silanization of surfaces via a range of procedures including organic phase, aqueous phase, and chemical vapor deposition of the silane. Silanization in solution is the most common approach since it does not restrict the use of volatile compounds and is not required to be performed under vacuum. Typical silane concentrations lie between 0.1 and 2% (w/v) and the presence of a small amount of water is necessary for the reaction to occur. During the surface activation reaction, two reactions take place simultaneously with respect to the silane: the (12) Coradin, T.; Lopez, P. J. Chembiochem. 2003, 4, 251-259.
10.1021/la050553t CCC: $30.25 © 2005 American Chemical Society Published on Web 06/21/2005
7030
Langmuir, Vol. 21, No. 15, 2005
hydrolysis of the n silane alkoxy groups to the highly reactive silanol species, and the condensation of the resultant silanols with the free OH groups of the surface to render stable Si-O-Si bonds. Oligomerization in solution also occurs as a competing reaction with covalent binding to the surface and constitutes an important consideration specially when n ) 2 or 3. The procession of these reactions and consequently the characteristics of the final surface layer depend on reaction variables such as solvent type, temperature, or time, as well as on the catalyst and organosilane concentrations used. The ways in which these parameters influence the surface modification process are complicated and various. Frequently a factor can lead to opposing effects by positively and negatively influencing different stages of the overall process. As an example, hydrophilic and protic solvents (like alcohols) usually accelerate hydrolysis and condensation kinetics, thus promoting the surface modification process. However, they can also compete with the silane for surface silanol groups by H-bonding. Solvent molecules can also invert the silane hydrolysis reaction or even form stable complexes with the hydrolyzed species that can lower their reactivity. Additionally, the reaction temperature can affect the outcome of the process. For example, raising the reaction temperature accelerates the silane condensation kinetics, both onto the surface and in solution, and can lead to the formation of silane aggregates. Small silane aggregates that result from condensation in solution may still be able to react with the surface (through unreacted silanol groups), but if solution-phase oligomerization goes too far, large aggregates will be inhibited sterically from reacting with surface -OH groups. This situation is even more complicated in the case of surface modification of nanoparticles. Particle aggregation phenomena and interparticle cross-linking through siloxane bridges may appear as a consequence of uncontrolled reaction conditions. Moreover, issues such as the extent of the modification (the thickness of the surface layer and the density of functional groups), and the homogeneity of the coating may be crucial for the final performance of the surface-activated material. As an example, the sensitivity of a diagnostic test will be limited by the number (density) and distribution of surface groups that are able to interact with the complementary/target molecular species. This work reports studies to determine the influence of reaction conditions (solvent, reaction temperature, reaction time, and silane and catalyst concentration) on the silanization process of silica-coated magnetite nanoparticles with (aminopropyl)triethoxysilane (APTS). The effect of reaction variables on the final product were quantified in terms of its organic content (by C and N elemental analysis), the presence and intensity of APTS characteristic bands in the IR spectrum of dried product samples, and the density of surface-active amine groups determined by colorimetric titration with nitrobenzaldehyde.13 We also studied the efficiency of such materials in a biological/diagnostic application, hybrid capture, of single-stranded nucleic acids after grafting with short single-stranded oligonucleotide. Experimental Procedures Materials. All reagents used were available commercially and were of the highest purity grade. Oligonucleotides (Guaranteed Oligos, HPLC-purified) were purchased from Proligo GmbH (Hamburg, Germany). APTS and hexamethyldisilazane (13) del Campo, A.; Sen, T.; Lellouche, J.; Bruce, I. J. J. Magn. Magn. Mater. 2005, 293, 33-40.
Bruce and Sen (HMDS) were purchased from Gelest-ABCR GmbH (Karlsruhe, Germany). The magnetic separator used was a Magix Magnetic Rack (Nuclyx Ltd., Hoddeston, U.K.) Ultrasonication was performed in a VWR ultrasonicator at a frequency of 45 Hz at room temperature. Characterization Methods. Elemental analysis (C, H, and N) was performed by Redox Spa (Milan, Italy). Fourier transform infrared (FT-IR) spectra were recorded on a Shimadzu 8300 (Shimadzu Corp., Kyoto, Japan) and samples were dried at 95 °C in vacuo for at least 5 h prior to fabrication of the KBr pellet. In this context 5 mg of each sample was thoroughly mixed and crushed with 500 mg of KBr, and 80 mg of that mixture was used for pellet fabrication (pressure applied was 10 tons for 10 min). Fifty scans of the region between 400 and 4000 cm-1 were collected for each FT-IR spectrum recorded. UV absorbance was measured on a Cary 100 Scan UV-vis spectrophotometer (Varian Ltd., Oxford, U.K.). The absorbance of 4-nitrobenzaldehyde was measured at 282 nm and that of oligonucleotides at 260 nm. Fluorescence was measured with a Cary Eclipse fluorescence spectrophotometer (Varian Ltd., Oxford, U.K.) at excitation and emission wavelengths of 460 and 515 nm, respectively. Transmission electron micrographs (TEM) were recorded on a Phillips 300 instrument. A drop of suspension of silica-coated magnetite in water before and after modification was placed onto a carbon-coated copper grid (Lacey carbon 400 mesh Cu-50 grids from Agar Scientific, Cambridge, U.K.) and dried at 35 °C. Synthesis of Magnetite and Silica-Coated Magnetite Nanoparticles. Magnetite nanoparticles were synthesized by precipitation of Fe3O4 from a solution of iron sulfate, potassium nitrate, and potassium hydroxide.14,15 Silica coating was performed by deposition of silica from a supersaturated solution of silicic acid by titration with hydrochloric acid as previously reported by our group.14 Passivation of Glassware. Glass reaction vessels were passivated by leaving them overnight in a desiccator under vacuum in the presence of HMDS. Determination of Solid Content of Nanoparticle Suspensions. Nanoparticle suspension densities were determined by drying 200 µL (×3) of each sample at 95 °C in a vacuum overnight to obtain a dry mass estimation. Synthesis of Amine-Modified Magnetite and SilicaCoated Magnetite Nanoparticles. Silica-coated magnetite nanoparticles were suspended in either water or organic solvent. In the case of nanoparticles suspended in organic solvent, this was preceded by multiple washings (×5) with the corresponding solvent to ensure that all water etc had been removed prior to reaction. For surface activation reactions, 150 mg of nanoparticles was added to a freshly prepared solution of APTS (0.2 or 2% w/v) in the desired solvent (or solvent mixture), and the final volume of the suspension was adjusted to 15 mL. In cases where the reaction was performed in organic solvents, 150 µL of water was also added as catalyst. The mixture was stirred vigorously on a magnetic stirrer at the desired temperature while reacting, and 2 mL aliquots were removed at fixed time intervals and washed five times (2 mL each) with fresh solvent (or solvent mixture). Silanization reactions were also performed in 1/1 mixtures of ethanol/water or THF/water. In cases where the reaction had been performed in organic solvents, two additional washing steps with water were performed. Colorimetric Assay of Amine Density. Nanoparticles (5 mg) were placed in a 1.5 mL Eppendorf tube and washed (×4) with 1 mL of coupling solution [0.8% (v/v) glacial acetic acid in dry methanol]. Subsequently, 1 mL of 4-nitrobenzaldehyde solution (7 mg in 10 mL of coupling solution) was added to the particles and the suspension was allowed to react for 3 h with gentle end-over-end rotation. After removal of the supernatant and washing (×4 in 1 mL of coupling solution), 1 mL of hydrolysis solution (75 mL of H2O, 75 mL of MeOH, and 0.2 mL of glacial (14) Bruce, I. J.; Taylor, J.; Todd, M.; Davies, M. J.; Borioni, E.; Sangregorio, C.; Sen, T. J. Magn. Magn. Mater. 2004, 284, 145-160. (15) Philipse, A. P.; Vanbruggen, M. P. B.; Pathmamanoharan, C. Langmuir 1994, 10, 92-99.
Magnetic Nanoparticles with Surface Alkoxysilanes acetic acid) was added to the particles and the tube was shaken for a further hour. The supernatant was then removed from the particles with a magnetic separator and its absorbance was measured at 282 nm. The amount of 4-nitrobenzaldehyde in the hydrolysis solution was calculated by interpolation, by use of a calibration curve constructed from a range of standard solutions of 4-nitrobenzaldehyde prepared separately. Salmon Sperm DNA Adsorption and Elution Assays. Sheared salmon sperm DNA (produced by repeated passage through a hypodermic syringe needle; 50 µg) was diluted in 400 µL of TEN buffer [100 mM Tris-HCl, pH 8.0, 50 mM ethylenediaminetetraacetic acid (EDTA), pH 8.0, and 500 mM NaCl]. Poly(ethylene glycol) (PEG, Mr 8000; 400 µL) in 4 M NaCl was added and the solution was mixed with 2 mg of nanoparticles (previously washed in sterile water) in a sterile Eppendorf tube. The mixture was incubated with gentle agitation for 5 min at 25 °C, after which the support material was immobilized and the supernatant was retained for assay of its absorbance at 260 nm. DNA was eluted by washing the support sequentially, twice, in sterile deionized water (200 µL each) for 5 min each at room temperature. Each 200 µL aliquot was assayed independently for its absorbance at 260 nm to estimate its DNA content. A standard curve was constructed for comparison purposes from a series of salmon sperm DNA solutions of known concentration. Covalent Coupling of Single-Stranded Oligonucleotides to Nanoparticles. SSC buffers (1× and 13×) were prepared by diluting a stock solution of SSC buffer (20×) (175.3 g of NaCl, 88.2 g of sodium citrate, and 1 L of H2O, pH 7.4) with distilled, deionized water, adjusted to pH 7.4, and autoclaved before use. Glutaraldehyde solutions were prepared immediately before use. Modified nanoparticles (2 mg) were washed (×3) with 1 mL of coupling buffer (1× SSC buffer, pH 7.3) for 2 min at 18 °C. After removal of the supernatant, 0.5 mL of a 5% (w/v) glutaraldehyde solution in coupling buffer was added and the suspension was incubated for 3 h with end-over-end rotation at 18 °C. The material was subsequently washed (×3) with 1 mL of coupling buffer to remove excess glutaraldehyde. A 3.3 µM solution (0.5 mL) of 5′-amine-modified dT25 was added and the mixture was left incubating overnight with shaking. The oligomodified nanoparticles were then washed once with coupling buffer and placed in 0.8 mL of NaBH3CN solution [0.03% (w/v) in coupling buffer] for 30 min at 18 °C. The material was then washed (×3) with 0.8 mL of coupling buffer and finally resuspended in 200 µL of the same. DNA Capture Experiments. dT25-modified particles (1 mg) were washed twice with 0.5 mL of water and heated to 80 °C for 4 min. A solution (200 µL) of either 5′-fluorescein-modified oligodA25(specific/complementary sequence capture) or dTTCTTGGCCGCACATTTGTCA (non-specific/noncomplementary adsorption) at the required concentration (see Table 2) in 13× SSC/ 0.05% BSA was added, and the suspension was incubated with gentle shaking for 30 min at 18 °C. The supernatant was removed and kept for analysis. After the nanoparticles were washed (×3) with 1 mL of 13× SSC, 200 µL of water was added and the nanoparticle suspension was heated to 85 °C for 4 min to disassociate the annealed complimentary/captured sequences. The supernatant was removed and kept for fluorescence analysis. Two control experiments were also performed in order to evaluate the nonspecific binding: the adsorption of the fluorescently labeled sequence onto aldehyde-modified nanoparticles and the hybrid capture of a noncomplementary fluorescently labeled sequence.
Results and Discussion Surface Modification with (Aminopropyl)triethoxysilane: (A) Influence of the Solvent. Surface modification reactions were performed at room temperature on silica-coated magnetite nanoparticles in three different types of solvents: two polar organic solvents, one protic (ethanol) and one aprotic (THF), and water. A small amount of water was deliberately added to the reaction mixture in the case of organic solvent-based reactions as it is needed to catalyze the reaction and as a reagent for
Langmuir, Vol. 21, No. 15, 2005 7031
Figure 1. IR spectra of APTS-modified silica-coated magnetite nanoparticles synthesized in different solvents before and after washing with water. The IR spectrum of pure APTS is included for comparison. Surface modification was performed at room temperature for 20 h.
the hydrolysis and condensation of the alkoxysilane groups. APTS and its hydrolysis and condensation products have excellent solubility in all these solvents, and no sign of insoluble aggregate formation was observed in any of the experiments performed. After reaction, the nanoparticles were extensively washed with the same solvent used in the reaction and a small amount was taken and dried for IR and elemental analysis. Figure 1 shows the IR spectra of these materials. The strong IR band at 580 cm-1 is characteristic of the Fe-O vibrations related to the magnetite core, and the weak bands at 880 and 1030 cm-1 correspond to Si-OH and Si-O-Si or Si-O-Fe stretching vibrations of the silica shell.14,16 The IR spectrum of the APTS-modified material shows an increase in the band at 1035 cm-1 (Si-O stretching), and a new band at 1131 cm-1, which may be attributed to C-N stretching modes. The band at 790 cm-1 corresponds to NH2 bending modes of the APTS molecule. The intensity of the APTS bands for the material modified in ethanol or water was relatively weak. These observations are in accord with the fact that the surface silane layer represents a small amount of the material’s mass in comparison with its bulk of magnetite. Surprisingly, intense bands were obtained when the reaction was performed in THF, suggesting that much more APTS had been deposited to the nanoparticles’ surface. Elemental analysis of the samples confirmed these trends (see Table 1). The change in relative intensity of -NH2 bending to -CH stretching resulting from APTS modification of the material in THF suggests that under those conditions -NH2 bending could be hindered in the modified material. This could be due to hydrogen bonding of NH2 group with Si-O-Si or unreacted Si-OH surface groups. The observation of such strong intensities for surfacelayer-related IR bands in the case of materials arising from THF-mediated reactions is unlikely to be compatible with the formation of a silane monolayer or even a multilayer at the particle surface but rather suggests an (16) Ma, M.; Zhang, Y.; Yu, W.; Shen, H. Y.; Zhang, H. Q.; Gu, N. Colloid Surface A 2003, 212, 219-226.
7032
Langmuir, Vol. 21, No. 15, 2005
Bruce and Sen
Table 1. Percentage C, H, and N Contenta of Silica-Coated Magnetite Nanoparticles after Surface Modification with APTS at RT in Different Solvents elemental analysis (wt %) reaction solvent
C
H
N
H2O EtOH THF
0.60 1.04 6.02
0.26 0.30 1.56
0.14 0.31 1.86
a
Determined by combustion elemental analysis.
extensive nonspecific adsorption process by which the silane molecules physisorb at the particle surface by means of reversible noncovalent interactions. This was confirmed by experimentation. Some of the material generated in the presence of THF was washed in water. When the wash supernatant was analyzed by FT-IR, typical APTS bands were observed. Furthermore, the materials were also reanalyzed post-washing by FT-IR and it was observed that the strength of the APTS-related bands was significantly lower than those of the parent material (see Figure 1). The main difference between ethanol and THF as solvents used in silanization of the nanoparticles is the protic character of the former and therefore its capability to form H-bonds with the Si-O- surface. The same principle also applies for water. In these cases a favorable solvent-surface interaction could create a strongly bound solvent layer around the particles, which might have hindered the relatively weaker silane particle from causing nonspecific adsorption. This would not have been the case for THF-mediated reactions. APTS autocatalyzes its own hydrolysis and condensation due to the basic character of its amino group (an aqueous solution of APTS exhibits a pH of 11).17,18 For this reason, APTS hydrolyzes (and polymerizes) completely within a few minutes when dissolved in water, but reactions in organic solvents proceed less quickly. Additionally, polar organic solvents accelerate hydrolysis when compared to nonpolar solvents, as has been observed from 1H NMR experiments of hydrolyzing solutions of silanes.19 With these considerations in mind, the rates of hydrolysis and condensation of APTS solutions should follow the order water . ethanol > THF. We are currently unable to determine exactly how these differences relate to the results observed. To avoid the occurrence of nonspecifically bound silane to the particle surfaces, all subsequent silanization reactions were performed in water or in 1/1 mixtures of ethanol/water or THF/water and extensively washed with water before analysis or further use. (B) Influence of Reaction Time and Temperature. The rate and extent to which alkoxysilane hydrolysis and condensation proceeds in solution and onto the nanoparticle surface depends on the duration and temperature of the reaction. Generally speaking, an increase in these parameters should lead to a greater degree of polymerization and subsequently a higher density of active surface groups. However, steric crowding and particle aggregation may also be promoted, which could lead to an opposite effect. A detailed study was undertaken to determine the (17) Beari, F.; Brand, M.; Jenkner, P.; Lehnert, R.; Metternich, H. J.; Monkiewicz, J.; Siesler, H. W. J. Organomet. Chem. 2001, 625, 208216. (18) Kang, H. J.; Meesiri, W.; Blum, F. D. Mater. Sci. Eng. A-Struct. 1990, 126, 265-270. (19) Osterholtz, F. D.; Pohl, E. R. J. Adhes. Sci. Technol. 1992, 6, 127-149.
Figure 2. Density of surface amine groups (in nanomoles per milligram of nanoparticles) of silica-coated magnetite nanoparticles modified with APTS under different reaction conditions. Each plot’s legend indicates the solvent [water (A) or 1:1 mixtures of water with tetrahydrofuran, THF (B), or ethanol, EtOH (C)], the reaction temperature (room temperature or 50 °C), and the silane concentration (0.2 or 2% w/v) used in each experiment. The nanoparticle suspension in plot D was ultrasonicated (US) for 5 min prior to reaction with APTS.
effect of these parameters on the surface activation of the nanoparticles. The relative insensitivity of IR analysis with respect to surface modification of the nanoparticles with APTS was likely to render it incapable of detecting what could have been subtle changes with respect to surface composition resulting from these experiments. Elemental analysis also suffered from the same limitation.13 Therefore, a colorimetric assay was adopted that permitted the accurate, sensitive quantification of surface amine density of APTS-modified materials.13 This assay was based upon the reaction of UV-sensitive 4-nitrobenzaldehyde with nanoparticle surface amine groups under anhydrous conditions to form an imine.20 The reaction can be reverted by the addition of water, and the 4-nitrobenzaldehyde liberated can be quantified by measuring the UV absorbance of the hydrolysis solution at 282 nm. The number of hydrolyzed 4-nitrobenzaldehyde molecules per gram of support represents the density of active amine sites on the particle surface. Figure 2 shows the results of colorimetric analyses performed on silica-coated magnetite nanoparticles that were modified with APTS for different periods at various temperatures. Identical experiments were performed in the various solvents employed at two different silane concentrations. It was observed that increasing the reaction time (from 1 to 24 h) led to an increase in amine density at the particle surface and that this effect was more pronounced at higher reaction temperatures (see Figure 2). This observation is similar to other results obtained recently with pure magnetite.13 A 4-fold increase of surface -NH2 density was (20) Moon, J. H.; Kim, J. H.; Kim, K.; Kang, T. H.; Kim, B.; Kim, C. H.; Hahn, J. H.; Park, J. W. Langmuir 1997, 13, 4305-4310.
Magnetic Nanoparticles with Surface Alkoxysilanes
Langmuir, Vol. 21, No. 15, 2005 7033
Figure 3. TEM images of unmodified (left) and NH2-modified (right) silica-coated magnetite nanoparticles.
observed when the reaction temperature was increased from RT to 70 °C.13 Increasing the reaction temperature should promote silanol condensation in solution and onto the nanoparticle surface. In addition, thermal energy may destabilize possible silanol-protic solvent complexes and therefore promote more favorable silanol-surface interactions.21 It was observed that sonication of the particle suspension before commencing the silanization reaction led to a slight increase in the final amine density values of the materials. Sonication prior to reaction may disrupt the formation of particle aggregates and/or cause the disaggregation of reversible aggregates already formed in the original suspension. (C) Influence of Alkoxysilane Concentration. APTS at 0.2% and 2% (w/v) concentration was used in silanization experiments and the results are shown in Figure 2. Silane concentration did not appear to significantly influence amine density possessed by the final material when reactions were performed in EtOH/water or THF/water mixtures. However, a different trend was observed for silanizations performed in water at high temperature. In this case, reactions at 0.2% alkoxysilane concentration approached a maximum surface amine density of approximately 12 nmol mg-1 after 8 h, whereas reactions in 2% APTS water solution reached higher amine densities (up to 23 nmol mg-1) by increasing reaction times up to 24 h. The surface amine density saturation point (15 nmol mg-1) was reached just after 5 h in water-based reactions involving 0.2% APTS when ultrasonication was employed. Additionally, the results indicate that increased APTS concentration (2%) did not dramatically improve the surface -NH2 density of the nanoparticles for any of the solvent systems tested. Particle Morphology. The morphology of the nanoparticles before and after surface modification was studied by TEM. Nanoparticles were rhombic in morphology (Figure 3) and their sizes ranged from 30 to 150 nm diameter. Little or no difference was observed between the original and APTS-modified material (Figure 3), indicating that silanization does not appear to have a deleterious effect on the support material. Surface Properties: Salmon Sperm DNA Adsorption Experiments. The change in surface properties of the nanoparticles after silanization was explored and (21) Dubitsky, Y.; Zaopo, A.; Zannoni, G.; Zetta, L. Mater. Chem. Phys. 2000, 64, 45-53.
Figure 4. Adsorption and elution of salmon sperm DNA onto and from silica-coated magnetite nanoparticles and their aminoand aldehyde-modified analogues.
followed by developing and using a model system involving the adsorption and desorption of salmon sperm DNA from the nanoparticles’ surface. This effectively served as a quick and sensitive test to corroborate the presence of new groups on the nanoparticle surface after silanization etc. Experimental conditions were as previously reported by our group,14 and Figure 4 shows the results of these studies with nonmodified, APTS-modified, and glutaraldehyde-modified silica-coated magnetite nanoparticles. Adsorption was performed under chaotropic conditions with elution in water. A high degree of adsorption of DNA onto silica- and -NH2-modified surfaces was observed under chaotropic conditions. In considering these observations, it is important to remember that the silica surface is negatively charged at pH 7 and that DNA adsorption would most likely be mediated by charge screening effects of surface Si-O- groups by the salt cation under chaotropic conditions. Also, the APTS surface-activated nanoparticles would be positively charged at pH 7, and in that case DNA adsorption would most likely occur by attractive electrostatic interactions between the surface and the negatively charged DNA phosphate groups.22 Adsorption of DNA onto the glutaraldehyde surface was observed to
7034
Langmuir, Vol. 21, No. 15, 2005
Bruce and Sen
Table 2. Results from Hybrid Capture and Nonspecific Adsorption of Complementary (dA25) and Noncomplementary (dTTCTTGGCCGCACATTTGTCA) Fluorescently Labeled Oligonucleotides to Nanoparticlesa Functionalized with the Capture Sequence dT25 NaCl (M)
blocking agent
surface funct
none none none none none none SSC, pH 7.3 SSC, pH 7.3 SSC, pH 7.3 SSC, pH 7.3 SSC, pH 7.3 SSC, pH 7.3 13× SSC, pH 7.3 13× SSC, pH 7.3 13× SSC, pH 7.3
2 2 2 2 2 2 0.15 0.15 0.15 0.15 0.15 0.15 2 2 2
none none 0.05% BSA 0.05% BSA 0.05% PEG 0.05% PEG 0.05% PEG 0.05% PEG 0.05% BSA 0.05% BSA 0.005% BSA 0.005% BSA 0.05% BSA 0.05% BSA 0.05% BSA
-CHO dT25 -CHO dT25 -CHO dT25 -CHO dT25 -CHO dT25 -CHO dT25 -CHO dT25
13× SSC, pH 7.3 13× SSC, pH 7.3
2 2
0.05% BSA 0.05% BSA
coupling buffer, pH
total oligo in soln (nmol) Complementaryb 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 0.24
Noncomplementaryb -CHO 0.3 dT25 0.3
adsorbed (nmol)
hybrid captured (nmol)
calcd nonspecific adsorption (nmol)
1.5 1.5 0.24 0.58 0.30 0.41 0.4 0.47 0.01 0.1 0.05 0.15 0.22 0.11 0.12
0.08 0.11 0 0.24 0 0.09 0 0.12 0 0.07 0 0.07 0 0.11 0.12
1.42 1.39 0.24 0.34 0.30 0.32 0.4 0.35 0.01 0.03 0.05 0.08 0.22 0 0
0.01 0.01
0 0
0.01 0.01
a APTS-coated silica-coated magnetite nanoparticles (1 mg) possessing an amine surface density of 15 nmol mg-1 and oligo-dT surface 25 density of 0.4 nmol mg-1 were used for these experiments. The experimental conditions of the adsorption/capture are specified in the first five columns from the left. The remaining columns show the initial amount of dA25 in solution, the total amount of dA25 adsorbed to the surface under the specific conditions, the amount of dA25 that elutes from the surface in water at 85 °C (hybrid captured), and the calculated amount of dA25 that remains on the nanoparticle surface after the elution (nonspecifically adsorbed). b Complementary probe was dA25; noncomplementary probe was dTTCTTGGCCGCACATTTGTCA.
be lower, probably as a consequence of the more neutral nature of the aldehyde-modified surface. The differences in surface chemistries of the three materials were also reflected in the way in which the DNA could be desorbed from the material surfaces. DNA adsorbed on silica surfaces can be eluted in water as a consequence of the electrostatic repulsion between ionized silanols and the phosphate backbone at low ionic strength. Elution from the amino-modified surface in water did not occur as a consequence of the electrostatic attraction between ammonium and phosphate groups being unaffected by the ionic strength of the environment (under the conditions used). Elution of the DNA from the glutaraldehyde surface occurred easily. Figure 4 illustrates the results obtained. Surface Coupling of Oligonucleotides: Bioactive Nanoparticles. The surface amino groups of APTSmodified nanoparticles can be used for coupling aminomodified oligonucleotides, which can be employed in the hybrid capture of complimentary sequences from a solution. This is a common approach in molecular diagnostic assays. Immobilization of 5′-NH2-modified oligonucleotides to the APTS-modified nanoparticles requires the use of a homofunctional coupling agent to convert the primary amine groups resident on the surface to -CHO, and glutaraldehyde has been widely used for this purpose. The conversion takes place in SSC buffer at approximately pH 7. Coupling of the oligonucleotide involves the formation of an imine bond that is subsequently reduced to the stable secondary amine by use of a reducing agent (NaBH3CN). The amount of immobilized oligonucleotide can be determined by UV spectroscopy (absorbance 260 nm) of the supernatant after the reaction has been completed. By this approach, 5′-amino-modified dT25 was immobilized onto APTS-modified nanoparticles (amine den(22) Carre, A.; Lacarrriere, V.; Birch, W. J. Colloid Interface Sci. 2003, 260, 49-55.
sity 15 nmol mg-1, Figure 2: 0.2% APTS, RT, water medium, ultrasonication, 5 h reaction product). It was observed that the resulting amount of covalently immobilized dT25 (0.4 nmol mg-1) was 37 times smaller than the density of free -NH2 groups detected by colorimetric analysis. This could possibly have resulted from the fact that the complementary oligonucleotide is a large molecule compared to the nitrobenzaldehyde used in the colorimetric analysis. It may have encountered greater steric hindrance effects when approaching the support surface, which may have lowered the yield of the coupling reaction. Hybrid Capture of Complementary Capture Sequences. A model system was developed and used in the performance testing of the oligonucleotide-modified particles in hybrid capture and nonspecific adsorption of complementary and noncomplementary fluorescently labeled oligonucleotide. Capture/adsorption efficiency was measured by analysis of the difference in total amount of fluorescently labeled oligonucleotide present in a solution before and after its incubation with the nanoparticles. Subsequent desorption and measurement of bound material was used in confirmation. Various parameters determine the efficiency of hybrid capture, and a systematic attempt was made to define the variables involved so as to optimize the system and its activity before the assay was applied experimentally to the various materials arising from the project. For example, the ionic strength and the pH of the hybridization medium together with the presence of blocking agents such as BSA and PEG can have a large effect on the efficiency of the system. At pH 7 (the pH at which these types of materials are normally used), the oligonucleotide phosphate groups are ionized and the molecule behaves as a negatively charged polyelectrolyte. Base pairing for duplex formation is promoted by high salt concentrations. However, high salt concentrations could promote H bonding between the unmodified aldehyde groups (still present on the surface of the support after grafting with the capture oligonucle-
Magnetic Nanoparticles with Surface Alkoxysilanes
Langmuir, Vol. 21, No. 15, 2005 7035
Table 3. Comparison of Surface Amine Density and Hybrid Capture Efficiency of APTS-Modified and Silica-Coated Magnetite Nanoparticles silanization conditions time T (°C) (h)
solvent
capture silane -NH2 density efficiencya concn (%) (nmol/mg) (nmol/mg) Pure Magnetite 2 2 2 2 2
50 50 50 50 50
1 2 5 8 23
H2O H2O H2O H2O H2O
10.4 14.1 17.4 18.6 24
0.044 0.061 0.086 0.081 0.146
50 18 50 18 50 50
23 23 23 23 23 23
Silica-Coated Magnetite H2O 0.2 13 THF/H2O 2 7.2 THF/H2O 2 19.6 EtOH/H2O 2 8.4 EtOH/H2O 2 24 H2O, US 5 min 2 27
0.064 0.034 0.103 0.03 0.126 0.161
a Efficiency in hybrid capture of a complementary sequence after having been grafted to a capture probe.
otide) and therefore also promote nonspecific adsorption of oligonucleotide species from solution. For this reason the addition of a competitor nonspecifically binding molecule at relatively high concentration is desirable (and usually necessary) to effectively block the nonspecific binding of the nucleic acid species to surface-unmodified aldehyde groups. PEG or BSA are the agents usually added for this purpose, with BSA often yielding better results. Experimental results are illustrated in Table 2 and were generated by use of nanoparticles with amine and dT25 densities of 15 and 0.4 nmol mg-1, respectively. The maximum hybrid capture efficiency of fluorescein-labeled dA25 was observed to be 0.24 nmol when 0.05% (w/v) BSA was used as blocking agent in 2 M NaCl. However, these conditions also resulted in high levels of nonspecific binding (0.34 nmol). When 2 M NaCl and BSA at 0.05% (w/v) were used in 13×SSC buffer at pH 7.3, no nonspecific binding was observed. The capture efficiency, however, was still relatively high, approximately 0.1 nmol. Capture efficiency under these conditions remained high even when a very low concentration of dA25 (0.24 nmol) was used in experiments. The lower yield may have been a consequence of steric hindrance caused by the BSA molecule competing effects. Control experiments performed with noncomplementary sequences showed no hybrid capture, as expected (see last row of Table 2). Under the optimized reaction conditions identified above, experiments were performed on various oligo-dTmodified nanoparticles produced from amino-modified materials (Figure 2) to investigate what effect surface
Figure 5. Comparison of the change in hybrid capture efficiency with surface NH2 density of silica-coated magnetite and magnetite nanoparticles.
amine density had on the process, and results are presented in Table 3. The correlation between the efficiency of hybrid capture and surface amine density is also presented in Figure 5, permitting comparison with our previously reported data for magnetite.13 It can be observed that the mole ratio of captured oligonucleotides is more than 2 orders of magnitude smaller than the number of available surface amine groups detected colorimetrically (Table 3) and this ratio seems to be independent of the amine density. Conclusions An optimized oligonucleotide-modified derivative of silanized silica-coated magnetite nanoparticles has been produced and defined with potential for use in diagnostic and research applications. The influence of silanization reaction parameters on the surface density of activating groups on the nanoparticles has been established, and DNA-DNA hybrid capture by the particles’ oligonucleotide-modified forms has been optimized. The results of this work not only establish the value of our materials and methods in a molecular diagnostic context but also will help other researchers to significantly improve their understanding of surface modification processes with alkoxysilanes. In turn, this will allow others to improve the performance of their nanomaterials for the same and other biological applications. Acknowledgment. We thank the European Union FP5 (CHEMAG project G5RD-CT-2001-00534) for financial support. LA050553T
