Non-agglomerated silicon–organic nanoparticles and their nanocomplexes with oligonucleotides: synthesis and properties

The development of efficient and convenient systems for the delivery of nucleic-acid-based drugs into cells is an urgent task. А promising approach is the use of various nanoparticles. Silica nanoparticles can be used as vehicles to deliver nucleic acid fragments into cells. In this work, we developed a method for the synthesis of silicon–organic (Si–NH2) non-agglomerated nanoparticles by the hydrolysis of aminopropyltriethoxysilane (APTES). The resulting product forms a clear solution containing nanoparticles in the form of low molecular weight polymer chains with [─Si(OH)(C3H6NH2)O─] monomer units. Oligonucleotides (ODN) were conjugated to the prepared Si–NH2 nanoparticles using the electrostatic interaction between positively charged amino groups of nanoparticles and negatively charged internucleotide phosphate groups in oligonucleotides. The Si–NH2 nanoparticles and Si–NH2·ODN nanocomplexes were characterized by transmission electron microscopy, atomic force microscopy and IR and electron spectroscopy. The size and zeta potential values of the prepared nanoparticles and nanocomplexes were evaluated. Oligonucleotides in Si–NH2·ODN complexes retain their ability to form complementary duplexes. The Si–NH2Flu nanoparticles and Si–NH2·ODNFlu nanocomplexes were shown by fluorescence microscopy to penetrate into human cells. The Si–NH2Flu nanoparticles predominantly accumulated in the cytoplasm whereas ODNFlu complexes were predominantly detected in the cellular nuclei. The Si–NH2·ODN nanocomplexes demonstrated a high antisense activity against the influenza A virus in a cell culture at a concentration that was lower than their 50% toxic concentration by three orders of magnitude.

The Si-NH 2 ·ODN nanocomplexes demonstrated a high antisense activity against the influenza A virus in a cell culture at a concentration that was lower than their 50% toxic concentration by three orders of magnitude.

Introduction
The development of efficient and convenient systems for the delivery of nucleic-acid-based drugs into cells is an urgent task. The solution to this problem would allow for the use of these drugs in practical medicine. Despite many efforts in this field, this problem cannot be considered as completely solved. А promising approach is the use of various nanoparticles as delivery vehicles. We have previously developed methods for immobilizing DNA fragments onto titanium dioxide nanoparticles with the formation of TiO 2 ·PL-DNA nanocomposites [1,2]. Silica nanoparticles can also be used as vehicles to deliver nucleic acid fragments into cells [3,4].
SiO 2 nanoparticles bearing amino groups on the surface were shown to bind plasmid DNA, allowing the nanoparticles to penetration into cells, and even nuclei, and to protect DNA against intracellular nucleases [5,6]. The prospect of using SiO 2 nanoparticles as nonviral nanovectors to deliver plasmid DNA and their lower toxicity compared to the widely used transfection agent lipofectamine was shown in previous work [6]. It was demonstrated that SiO 2 nanoparticles with a diameter of 20-77 nm can easily penetrate into cellular cytoplasm and nuclei [7]. Polylysine-modified nanoparticles appeared to be nontoxic upon oral administration in mice, which opens the possibility for their wide application [8]. Under physiological conditions, SiO 2 nanoparticles are known to degrade to orthosilicic acid, Si(OH) 4 , which is found in almost all human tissues and effectively excreted through the urine [9]. Organically modified silica nanoparticles are known for their low toxicity and biocompatibility and can be used for targeted imaging and therapy [10,11].
The synthesis of amino-containing silica nanoparticles has been described in many works [3,6,[12][13][14][15]. Most often, amino-containing silica nanoparticles are prepared by copolymerization during the hydrolysis of alkyl (or aryl)triethoxy(or methoxy) silanes in the presence of APTES [16,17]. A large series of works describes the synthesis of amino-containing silicon nanoparticles, so-called ORMOSIL, and their use for diagnostics and therapy [15,18]. Most often, these methods lead to the formation of nanoparticles with a diameter of 20-100 nm. Silica is particularly attractive as a functional, biocompatible, nontoxic, and inert material, which can be easily synthesized and modified [16,17].
In this work, we describe the synthesis of silicon-based nanoparticles by the hydrolysis of APTES alone, the preparation of oligonucleotide-containing nanocomplexes on their basis, and characterization of both nanoparticles and nanocomplexes. We also demonstrated the biological activity of the latter with an example of inhibition of influenza A virus replication in cell culture.

Results and Discussion
Si-NH 2 nanoparticles were synthesized by the hydrolysis of aminopropyltriethoxysilane (APTES). The resulting product is a clear aqueous solution containing nanoparticles in the form of low molecular weight polymer chains consisting of the [-Si(OH)(C 3 H 6 NH 2 )O-] monomer units as shown below.
Due to the positively charged amino groups in Si-NH 2 at the neutral pH values, the nanoparticles can be electrostatically conjugated with oligodeoxyribonucleotides (ODN) through the binding to the negatively charged internucleotide phosphate groups to form Si-NH 2 ·ODN complexes. Like Si-NH 2 , the Si-NH 2 ·ODN nanocomplexes are dissolved in aqueous solutions. The addition of NaCl did not lead to precipitation. Si-NH 2 and Si-NH 2 ·ODN can be precipitated by acetone. The unbound oligonucleotide after precipitation of Si-NH 2 ·ODN remains in solution and can be detected spectrophotometrically. This "acetone precipitation" technique allowed us to evaluate the optimal molar ratio of the amino groups in the nanoparticles to the phosphate groups in an oligonucleotide (NH 2 /p ≈ 10) when preparing the Si-NH 2 ·ODN nanocomplexes ( Figure 1a).
A similar result was obtained when the mixture of ODN and Si-NH 2 was analyzed by electrophoresis in agarose gel ( Figure 1b). The higher the NH 2 /p ratio, the higher complexation occurred; at the ratio of 11, the band corresponding to the unbound oligonucleotide almost disappeared and the band corresponding to the Si-NH 2 ·ODN complex could be observed at the top of the electrophoregram. The results indicate the binding of several Si-NH 2 nanoparticles to one oligonucleotide molecule.
The binding of the Si-NH 2 molecules to the internucleotide phosphate groups of an oligonucleotide leads to the formation  of an ODN salt. In this work, this salt complex is conventionally named the Si-NH 2 ·ODN nanocomplex.
The Si-NH 2 nanoparticles and Si-NH 2 ·ODN nanocomplexes were characterized by physico-chemical methods.
The IR spectrum of Si-NH 2 shows the local environment of the silicon atom ( Figure 2a). The presence of the absorption bands that correspond to valence and deformation vibrations of the CH 2 , NH 2 , NH 3 + , Si-CH 2 , Si-OH, (Si-O) n , C-N groups [19] confirm the proposed structure of the Si-NH 2 molecule. The splitting of the absorption band at 1085 cm −1 , which corre-sponds to asymmetric valence vibrations of the Si-O bond, into two bands at 1100 cm −1 and 1000 cm −1 is typical for polymer structures with Si-O-Si fragments and indicates the presence of at least dimers in the Si-NH 2 nanoparticles. The bands at 3058 cm −1 and 1627 cm −1 show the presence of the protonated NH 3 + groups.
The electron absorption spectrum of the Si-NH 2 solution has two bands at 49000 cm −1 and 36500 cm −1 (205 nm and 275 nm, respectively) (Figure 2b), the first of which corresponds to the (n-π*) transfer and belongs to the NH 2 group [19]. The band at 36500 cm −1 can be attributed to the Si-O-Si structures because  the absorption in the range of 35000-37000 cm −1 is characteristic for SiO 2 nanoparticles and can be caused by electronictype paramagnetic defects or n-π* transitions of the adsorbed organic ligand [19].
According to the transmission electron microscopy (TEM) images, there is a difference in morphology between Si-NH 2 nanoparticles and their complexes with ODN. Discrete nanometer-sized nanoparticles (Si-NH 2 ) are quite homogeneous ( Figure 3a). Si-NH 2 ·ODN(1) forms chain-like structures (Figure 3b), which can be explained by the formation of aggregates containing more Si-NH 2 particles. Probably, the Si-NH 2 counter ions in the Si-NH 2 ·ODN complex are polymerized during the sample preparation (drying) due to their close proximity to one other.
The Si/P weight ratio for the analyzed Si-NH 2 ·ODN(1) sample was evaluated by inductively coupled plasma mass spectrome-try (ICP-MS). The experimental Si/P value (10.9) showed a reasonable correlation with the calculated data (10.0).
The atomic force microscopy (AFM) image of the Si-NH 2 nanoparticles was not obtained in a satisfactory quality because of the very small size of the particles. Fortunately, several slightly larger Si-NH 2 ·ODN(1) nanocomplexes could be visualized ( Figure 4). According to the AFM data, the size (height) of these nanocomplexes is 1.2-1.5 nm. As in the case of the TEM image (Figure 3b), the particles in the sample can form chainlike structures, which can increase the particle size detected by AFM.
Reasonable results on the diameter of the studied particles were obtained by dynamic light scattering (Table 1 and Table 2), which allows measurement of the particle size in solution (without effects of the drying process, which most likely leads to polymerization and agglomeration of the particles). The zeta potential of the studied samples was evaluated by phase analysis light scattering. The value of zeta potential of Si-NH 2 is positive due to positively charged amino groups in neutral medium. When the nanoparticles were bound to an oligonucleotide, the zeta potential value, as expected, became negative (Table 1).  Table 1).
The very small size of the Si-NH 2 particles and Si-NH 2 ·ODN complexes (1-5 nm) allows for the formation of true solutions. The prepared nanoparticles and nanocomplexes, Si-NH 2 and Si-NH 2 ·ODN, can be stored for a long time (at least for a month at 25 °C and three months at 4 °C) without changes in size (Table 2), which indicates no aggregation of these preparations occurred in these time frames.
Oligonucleotides in the Si-NH 2 ·ODN nanocomplexes were shown to retain their ability to form complementary duplexes. The Si-NH 2 ·ODN(2) and Si-NH 2 ·ODN(3) nanocomplexes containing the complementary oligonucleotides have the same size (d ≈ 3.6 nm, Table 1); however, we observed an increase in the diameter to 4.8 nm for the particles in the mixture of these nanocomplexes. This fact indicates the formation of the com-plementary duplex. A similar increase in the size was observed when ODN(2) was added to the preformed Si-NH 2 ·ODN(3) nanocomplex (Table 1) The concentration of Si-NH 2 and Si-NH 2 ·ODN resulting in 50% MDCK cell death (TC 50 ) was found to be 10-20 mM (for Si). The nontoxic concentration of the samples (0.014 mM for silicon) was used to study their ability to interact with the RNA target in cells with an example using the inhibition of influenza A virus (IAV) reproduction. The antiviral activity of the Si-NH 2 ·ODN(4) nanocomplex was not inferior to that of the TiO 2 ·PL-ODN nanocomplexes reported in our previous papers [20][21][22]. The high inhibition efficiency of the virus reproduction with the use of the Si-NH 2 ·ODN nanocomplex is provided by the fact that the nanocomplex delivers ODN into the cytoplasm of the cell, and ODN can penetrate into the cell nuclei ( Figure 5). The delivered ODN can interact with virus RNA via the antisense mechanism.

Conclusion
We developed a very convenient and simple method for the preparation of non-agglomerated Si-NH 2 nanoparticles and Si-NH 2 ·ODN nanocomplexes. The Si-NH 2 nanoparticles were obtained by a one-step synthesis from APTES, and the resulting product could be used without further purification to form Si-NH 2 ·ODN nanocomplexes. The Si-NH 2 nanoparticles and Si-NH 2 ·ODN nanocomplexes were characterized by physicochemical methods. The proposed nanocomplexes appeared to form very small nanoparticles (which are dissolved in aqueous solutions), then penetrate into the cellular cytoplasm to deliver oligonucleotides to the nucleus. The main advantage of the proposed Si-NH 2 nanoparticles and Si-NH 2 ·ODN nanocomplexes is that they are not prone to aggregation and form true aqueous solutions. The Si-NH 2 ·ODN nanocomplexes exhibit biological activity, which has been shown with an example of the inhibition of influenza A virus replication. The proposed nanocom-plexes can effectively and selectively interact with RNA targets in the cell culture.
Inductively coupled plasma mass spectrometry (ICP-MS) was performed using an Agilent 7700 ICP MS (USA) spectrometer.
The IR spectrum of the Si-NH 2 solution was recorded on a Cary 660 FTIR spectrometer (Agilent Technologies, USA) within the range of 4000-500 cm −1 at 4 cm −1 resolution, and 100 scans were accumulated in the attenuated total reflectance (ATR) mode using a GladiATR unit (Pike Technologies). The spectrum of the substance is presented as absorbance after computer processing, in particular, subtraction of the water spectrum (solvent) from the spectrum of the solution was performed (Figure 2a).
The electron spectrum of the Si-NH 2 solution was taken on a Shimadzu UV-2501PC spectrophotometer (Shimadzu, Japan) in the absorbance mode. The spectrum was recorded with absorption compensation relative to water in the range 11000-54000 cm −1 in a quartz cuvette with an optical path length of 1 mm (Figure 2b).
The size and zeta potential values of Si-NH 2 and Si-NH 2 ·ODN were measured at 5-50 mM concentration (for Si) in physiological solutions or water on a Zetasizer Nano ZS Plus instrument (Malvern, UK) with the ratio of NH 2 /p = 10 in the Si-NH 2 ·ODN nanocomplexes. The zeta potential values of the samples were obtained using a DTS1070 cuvette. Each sample was prepared at least in triplicate and measured three times at room temperature. The values of particle size and zeta potential were averaged over those experiments.
Transmission electron microscopy (TEM) images were taken on a JEM 1400 (Jeol, Japan) electron microscope at an accelerating voltage of 80 kV.
We analyzed a 35 mM solution of Si-NH 2 nanoparticles and Si-NH 2 ·ODN nanocomplexes with the ratio of NH 2 /p ≈ 10 in the latter case. The studied sample (10 μL) was mixed with a drop of 2% aqueous uranyl acetate; the mixture was covered with a standard carbon-coated copper mesh and after drying for one minute was fixed in a holder and introduced in the chamber of the electron microscope. The results are presented in Figure 3.
Atomic force microscopy (AFM) was performed on a Solver P47 Bio atomic force microscope (NT-МDT, Russia) in a tapping mode. The aqueous solution of the Si-NH 2 ·ODN(1) sample (10 µL, 0.16 µM, NH 2 /p = 10) was applied to a freshly cleaved mica area of 25-30 mm 2 . The adsorption was held for one minute at room temperature; the remaining fluid was blown by air. The results are shown in Figure 4.
Preparation of silicon-organic nanoparticles and Si-NH 2 ·ODN nanocomplexes APTES (8 mL, 34.16 mmol) was added dropwise to 100 mL of hot (70 °C) water. The mixture was stirred at this temperature for 15 h and then cooled to the room temperature. The concentration of the product (0.46 M) was evaluated by inductively coupled plasma mass spectrometry (ICP-MS) on an Agilent 7700 ICP-MS spectrometer or by titration of the amino groups with hydrochloric acid. The yield of the product was 95-97%.
The resulting Si-NH 2 nanoparticles were attached to oligodeoxyribonucleotides using the electrostatic interaction between the negatively charged internucleotide phosphate groups and positively charged amino groups in Si-NH 2 . The amount of Si-NH 2 nanoparticles necessary for the complete binding of an oligonucleotide was evaluated in the following experiments. In experiment (a), different amounts of Si-NH 2 (0, 1, 3, 5, 10, 16, and 22 µL of 0.046 M solution) were added to ODN(1) (2.7 µL of 10 −3 M; ≈40 nmol of phosphate groups) in water in seven vials. The volume of the mixtures was adjusted to 100 µL with water in each vial. After vigorous stirring for 1 min, the mixture was left at room temperature for 30 min to facilitate complexation. The formed Si-NH 2 ·ODN(1) complex was precipitated by acetone (precipitate 1). Supernatants from each vial were separated, and the unbound ODN in these supernatants were precipitated by the addition of 1 µL of 3 M LiClO 4 (precipitate 2). The precipitates 1 and 2 were dried at 60 °C to remove the acetone and were dissolved in 1 M NaCl. The amount of the bound (precipitate 1) and unbound (precipitate 2) oligonucleotide was evaluated spectrophotometrically at 260 nm. The results are presented in Figure 1a. In experiment (b), ODN (1) and Si-NH 2 were mixed (as in (a)) and the samples (1 µL) were loaded onto 0.7% agarose gel prepared by heating of agarose in Tris-glycine buffer. Electrophoresis was performed at 5 V·cm −1 for 30 min. Oligonucleotide spots were visualized by StainsAll (Figure 1b).
In our experiments, the Si-NH 2 ·ODN nanocomplexes were obtained by mixing 0.46 M Si-NH 2 (2 µL) and ODN in 0.16 M NaCl or water (98 µL) while keeping the ratio between the amino and phosphate groups at NH 2 /p ≥ 10, unless otherwise specified.

Thermal denaturation of ODN duplexes
Thermal denaturation of ODN duplexes was studied using a UV-1800 spectrophotometer equipped with a TMSPC-8 instrument (Shimadzu, Japan) for analyzing the melting temperature of the samples.  Figure 5.

Cell viability assays
The experiments were carried out similarly as described in [20]. In short: Two days after the formation of a continuous monolayer of MDCK cells at 37 °C and 5% CO 2 , the cells were washed with nutrient medium without serum. The Si-NH 2 or Si-NH 2 ·ODN samples were diluted with RPMI-1640 medium to the needed concentration (2-35 mM for Si) and incubated with MDCK cells at 37 °C and 5% CO 2 for two days. Destructive changes in the cells were evaluated by an inverted microscope. MDCK cells without studied samples were used as a control and taken for 100%. The cells were stained with trypan blue [23], and the number of viable cells were counted in a Goryaev chamber.

Antiviral activity of nanocomplexes
The antiviral activity experiments were performed analogously as described in [22]. The influenza A virus (IAV) strains and MDCK cells were prepared as in [22]. The cells at ≈80% confluence were initially infected with A/chicken/ Kurgan/05/2005 virus (H5N1), which was added in each well in RPMI-1640 medium (100 μL) containing trypsin (2 μg/mL) at a MOI of 0.1. The control sample was RPMI-1640 medium (100 μL) containing trypsin (2 μg/mL). After 1 h of virus adsorption at room temperature, the virus-containing medium was removed, and the cells were rinsed with RPMI-1640 medium without trypsin. The Si-NH 2 , ODN(4), Si-NH 2 ·ODN(4), and Si-NH 2 ·ODN(5) studied samples taken in RPMI-1640 medium without trypsin (100 μL/well) at a concentration of 0.014 mM for Si and 0.1 μM for ODN were applied to the infected MDCK cells, followed by incubation for 4 h at 37 °C, 5% CO 2 , and 100% humidity. After incubation for 4 h at room temperature, the medium containing the sample was removed, the cells were rinsed with RPMI-1640 medium without trypsin, and the same medium containing trypsin was added in each well (100 μL). After incubation for 48 h, serial ten-fold dilutions (from 10 -1 to 10 -8 ) of the cultural virus-containing liquid from each well were applied to MDCK cells for 48 h to evaluate the virus titer. The presence of the virus was visually determined under a microscope by the cytopathic action and in the hemagglutination reaction with a 1% suspension of chicken erythrocytes. The virus titer was expressed in terms of log TCID 50 /mL ( Figure 6). The titer was evaluated by noting the highest dilution of the virus, which caused the hemagglutination reaction.