Adsorption of Myoglobin and Corona Formation on Silica Nanoparticles.

The adsorption of proteins from aqueous medium leads to the formation of protein corona on nanoparticles. The formation of protein corona is governed by a complex interplay of protein-particle and protein-protein interactions, such as electrostatics, van der Waals, hydrophobic, hydrogen bonding, and solvation. The experimental parameters influencing these interactions, and thus governing the protein corona formation on nanoparticles, are currently poorly understood. This lack of understanding is due to the complexity in the surface charge distribution and anisotropic shape of the protein molecules. Here, we investigate the effect of pH and salinity on the characteristics of corona formed by myoglobin on silica nanoparticles. We experimentally measure and theoretically model the adsorption isotherms of myoglobin binding to silica nanoparticles. By combining adsorption studies with surface electrostatic mapping of myoglobin, we demonstrate that a monolayered hard corona is formed in low salinity dispersions, which transforms into a multilayered hard + soft corona upon the addition of salt. We attribute the observed changes in protein adsorption behavior with increasing pH and salinity to the change in electrostatic interactions and surface charge regulation effects. This study provides insights into the mechanism of protein adsorption and corona formation on nanoparticles, which would guide future studies on optimizing nanoparticle design for maximum functional benefits and minimum toxicity.

Introduction

The integration of nanoscience with biomedicine has resulted in numerous technological advances in molecular sensing, targeted delivery, in vivo imaging, and gene therapy.[1−4] For these applications, nanoparticles (NPs) are the fundamental units that provide the necessary biological response and functionality.[5,6] When NPs are introduced in in vivo environments, proteins from biological fluids instantaneously bind to the NP surface, leading to the formation of a shell known as the “protein corona”.[7] Thus, any further interaction of NPs with biomolecules is mediated by the protein corona.[8,9] While most of the current studies focus on the biological activity of the protein within the corona, the mechanism of its formation and local orientation of proteins within the corona remain poorly understood.[10,11] This lack of understanding of protein corona characteristics is due to anisotropic shape and nonuniform distribution of chemical functional groups on the surface of protein molecules forming the corona.[12,13] To develop the next generation of functional nanomaterials for biomedical applications, it is critical to understand the local interactions driving the formation of the protein corona and thermodynamic state of the protein on the NPs.

The corona formed on NPs can be composed of a different number of layers of protein depending on the binding affinity of protein–NP and protein–protein.[14] During the adsorption process, proteins with the highest affinity for the NP surface are irreversibly adsorbed forming the first layer known as “hard” corona.[15] Subsequently, any free protein in bulk weakly adsorbs on the hard corona leading to the formation of an outer layer called “soft” corona.[16] While protein molecules forming the hard corona are irreversibly bound to NPs, the weakly bound molecules forming the outer layer are in dynamic equilibrium with free protein molecules in bulk solvent.[17,18] Despite the knowledge on the dynamic behavior of proteins forming the corona, the assembled state of protein molecules within the corona remains unknown.[19] Recently, molecular dynamic (MD) simulations have been used to demonstrate that a globular protein adsorbing on a solid surface attains a preferred orientation, which is governed by the local electrostatic and van der Waals interactions.[13,20] These simulation studies were performed for a single protein molecule binding to a solid substrate; therefore, no protein–protein interactions were involved. In practice, adsorption of proteins on NPs is significantly influenced by protein–protein interactions. In this study, we investigate the effect of protein–NP and protein–protein interactions on the adsorption and corona formation on the surface of NPs.[21] We show that the assembled state of protein corona shell transitions from a monolayered hard corona to a multilayered “hard + soft” corona by changing the pH and salinity of the solvent.

In this study, we use myoglobin and silica as a model protein and NP, respectively. Myoglobin is a globular protein found in skeletal and cardiac muscle cells. It acts as a local oxygen reservoir to provide temporary oxygen when blood oxygen delivery is insufficient during periods of intense muscular activity.[22,23] Myoglobin contains 153 amino acids, a heme group where an iron ion is bonded to four nitrogen atoms of the porphyrin ring, a proximal histidine group (His-93), and a distal histidine group (His-64) with a net isoelectric point (IEP) of pH 7.[24] Here, we use hydrophilic silica nanospheres of diameter 30 nm as model NPs. The silica NPs have been widely used as a model to study adsorption and delivery of proteins due to their low toxicity, hydrophilicity induced by surface by the silanol groups (Si–O–H), and well-established synthetic procedures to precisely control their physical and chemical properties.[25] The neutral point of bare silica NPs is pH ∼2, i.e., the NPs are negatively charged at pH >2. Thus, myoglobin and silica are oppositely charged in the range 2 < pH < 7, and a net electrostatic attraction is expected between the protein and NPs. Here, we experimentally study the effect of pH and salt concentration on the adsorption of myoglobin molecules on silica NPs. We use the Guggenheim–Anderson–de Boer (GAB) model to represent the adsorption of the protein onto silica NPs. Additionally, we calculate the surface electrostatic maps for myoglobin, providing a better understanding of the adsorption process and corresponding protein corona formation.

Materials and Methods

Materials

Myoglobin from equine skeletal muscle (molecular weight of ∼17 kDa, purity ≥95%) was purchased from Sigma-Aldrich. Commercially available LUDOX-TMA colloidal silica (Sigma-Aldrich) was used as model spherical NPs. The silica NP dispersion was dialyzed for 7 days using deionized water, and water was changed every day to remove any undesired foreign molecules in the dispersion. Sodium chloride (NaCl) (VWR, purity 99%) was used as a 1:1 electrolyte to alter the range of electrostatic interactions. All experiments were performed with ultrapure water of resistivity 18.2 MΩ cm.

Experimental Methods

The silica NPs were characterized by dynamic light scattering (DLS), transmission electron microscopy (TEM), ζ-potential measurement, and nitrogen gas adsorption. A Litesizer 500 (Anton Paar) equipped with a 658 nm laser was used to determine the time-correlation function of the dispersion at a backscattering angle of 175°. For ζ-potential measurements, Univette cuvette (Anton Paar) was used and the electrophoretic mobility of the NPs was determined by applying a 10 V direct current across the dispersion. The TEM imaging of the NPs was performed on a JEOL JEM-2011. The specific surface area of the particles was determined using nitrogen gas adsorption performed on a Micromeritics ASAP 2020.

Calculating Net Charge, Dipole Moment, and Surface Potential of Protein

The atomic structure of myoglobin was obtained from the protein data bank (PDB). In our study, we used the crystal structure of myoglobin (PDB # 5ZZE) obtained by X-ray diffraction.[26] In step 1, the protonation state of each ionizable group on myoglobin at a given pH was assigned using the online tool PROPKA.[27,28] In step 2, the charge on an individual atom in myoglobin was obtained using webserver PDB2PQR.[29] At the end of step 2, a file containing atomic positions and protonation/charged state of each functional group was created. The surface electrostatic potential map and net dipole moment of myoglobin at the given pH were computed using visual molecular dynamics.[30]

Results and Discussion

Characterization of Silica NPs and Myoglobin

Silica NPs were characterized for their size and polydispersity using DLS and TEM. The surface charge of the NPs was quantified in terms of ζ-potential, and nitrogen adsorption was used to determine their specific surface area. The time-correlation function of the NPs as obtained by the DLS measurement is shown in Figure S1a. The time-correlation function shows a single exponent decay, which is a signature of a single particle population. The intensity distribution extracted from the time-correlation function of silica NPs is shown in Figure S1b. The hydrodynamic diameter of the silica NP was ∼30 nm, with a polydispersity index of 0.1. The diameter was further confirmed by visualizing the silica NPs using TEM, as shown in Figure a.

Figure 1

(a) TEM image of silica NPs showing the diameter (∼30 nm) with a spherical shape. (b) ζ-Potential of silica NPs as a function of pH. The silica NPs are negatively charged in the range 4 < pH < 11. (c) Net charge on myoglobin as a function of pH. Silica NPs and myoglobin are oppositely charged below IEPMGB, i.e., 4 < pH < 7. (d) X-ray crystal structure of horse myoglobin (PDB # 5ZZE). The yellow line structures indicate histidine residues in the protein. The His-93 and His-64 residues play a key role in charge regulation effects due to their physiologically relevant pKa values.

The specific surface area of silica NPs (as) was determined using nitrogen gas adsorption and corresponding analysis on the basis of the Brunauer–Emmett–Teller (BET) model, which was as = ∼114 m2 g–1, seen in Figure S2. The pH dependence of the surface charge of the silica NPs was quantified using ζ-potential titration, as shown in Figure b. The silica NPs were negatively charged in the complete experimental pH range of 4–10, with a ζ-potential value of −45 mV at pH 7. The highly negative ζ-potential of the silica NPs at this pH highlights their colloidal stability in the aqueous dispersion via electrostatic double-layer repulsion.[31]

Myoglobin is a globular protein with an ellipsoidal shape of dimensions 4.5 × 3.5 × 2.5 nm3.[32] The net charge of protein in aqueous environments is strongly dependent on the pH of the dispersion.[33] The charge on a myoglobin molecule at a given pH is determined from pKa values calculated using PROPKA. The calculations show that the net charge on protein changes from +18e to −5e (e is the elementary charge) upon increasing the pH from 4 to 10, which is in good agreement with the previously reported net charge of myoglobin (Figure c).[34] It can also be observed that the isoelectric point of myoglobin (IEPMGB) is pH ∼7, and the magnitude of the charge on myoglobin remains <2e in the pH range 7–9. The iron atom in myoglobin is bound by His-93 (proximal histidine), and His-64 (distal histidine) forms a hydrogen bond with the bound oxygen (Figure d).[35] For protein adsorption, these histidine residues play an important role due to their ampholytic nature at a physiologically relevant pH (discussed later).[36] Based on the ζ-potential measurements for the silica NPs and protein charge determination, it can be inferred that in the pH range 4–7, the silica NPs and myoglobin molecules are oppositely charged and a net electrostatic attraction between a silica NP and myoglobin can be expected. Note that myoglobin retains its folded state in the pH range 5–12 and denatures outside this pH regime.[37,38]

Protein Adsorption Isotherm

Adsorption isotherms of myoglobin on silica NPs were determined in the pH range 4–10 by the solvent depletion method.[39−41] The first step for isotherm measurements is determining the calibration curve for the protein concentration in the absence of NPs. In a typical calibration curve determination, known concentrations of myoglobin in the range 0.1–5 mg mL–1 are prepared in DI water at a desired pH. The UV/vis absorbance of these control samples was measured at 280 nm using a NanoDrop 2000 spectrophotometer, which showed a linear increase in agreement with the Beer–Lambert law (Figure S3). A fixed amount of silica NP dispersion was added to a myoglobin solution of known concentration such that the final concentration of NPs was 1 wt %. The pH was readjusted to a desired value using a minimum amount of 5 N sodium hydroxide or hydrochloric acid. The myoglobin–silica dispersions were equilibrated at 25 °C for 24 h and centrifuged at 18 210g for 2 h to separate silica NPs with adsorbed protein (Figure a,b). The unadsorbed myoglobin concentration in the supernatant was determined using the UV/vis absorbance values at 280 nm (Figure c) and the calibration curve (Figure S3). The amount of myoglobin adsorbed on the silica NP surface (Γ) is calculated as[42]where V is the volume of protein solution, cο is the initial myoglobin concentration, ceq is the equilibrium concentration of protein in solvent bulk, ms and as, are the mass and specific surface area of the silica NPs, respectively.

Figure 2

(a) Schematic of the experimental procedure to determine the adsorption isotherms of myoglobin on silica NPs. Myoglobin and silica NPs are mixed at a desired pH and equilibrated for 24 h at room temperature. After separating silica NPs from the aqueous phase, the amount of unadsorbed protein was determined by measuring the absorbance of the supernatant at 280 nm. (b) Schematic representation of hard corona (solid-line circle) and soft corona (dashed circle) around a silica NP. (c) Absorption spectrum of myoglobin (conc. = 2.0 mg mL–1) before adding silica NPs (red line) and after adding silica NPs and centrifugation at equilibrium (dashed line). (d, e) Sample cell of myoglobin in water showing the change of color before adding silica NPs (d) and after adding silica NPs and separating the NPs with adsorbed myoglobin using centrifugation (e).

We use the Guggenheim–Anderson–de Boer (GAB) model to represent the experimental adsorption isotherms and determine maximum surface excess and adsorption constants for myoglobin binding onto silica NPs. The GAB model is a three-parameter model that can be used to describe multilayer adsorption onto surfaces, as shown in Figure b. According to this model, the adsorption state of molecules above the monolayer is different from the first layer (hard corona), which introduces an additional adsorption constant for the outer layer (soft corona).[43] In the GAB model, the amount of the surface adsorbed protein (Γ) is expressed as[44]where Γm is the maximum surface excess of protein forming a hard corona, ceq is an equilibrium concentration of protein in bulk solvent, Khard is the adsorption constant for the hard corona, i.e., first layer of protein directly adsorbed onto silica NPs, and Ksoft is the adsorption constant for soft corona, i.e., outer layers of myoglobin bound to the protein-decorated silica NPs. For Ksoft = 0, the GAB model reduces to the two-parameter Langmuir model, which represents the monolayer adsorption process. Note that Γm provides the maximum amount of protein forming the hard corona and does not include the protein molecules forming the soft corona.

In the regime where pH < IEPMGB, myoglobin molecules and silica are oppositely charged (Figure b,c) and the attractive electrostatic interaction exists between the protein and NP surface. The isotherms show a rapid increase in the amount of myoglobin adsorbed on silica NPs at ceq < 0.5 mg mL–1. The observed increase in surface adsorbed myoglobin highlights a strong attraction between the protein and silica NPs. This increase is followed by a plateau on the isotherm, which corresponds to the maximum surface coverage of NPs by protein molecules. The experiments show that the maximum surface excess of myoglobin on silica NPs increases with increasing pH from 4 to 6 (Figure a). The addition of NaCl results in an increase in the amount of protein adsorbed at pH < IEPMGB but a decrease in the adsorbed amount at pH > IEPMGB. Here, we determine the maximum geometric monolayer coverage of NPs by protein (Figure , dashed line) by dividing the surface area of the NPs with the projection area of the protein, i.e., 4.5 × 3.5–15 nm2. The geometric monolayer coverage is the highest possible NP surface coverage by the noninteracting protein in its side-on configuration. We find that in the presence of salt, the adsorption approaches this monolayer limit and exceeds at pH 4 (discussed later). We acknowledge that a fraction of myoglobin may have denatured at pH 4 or during the fine adjustment of the pH, which would impact the adsorption behavior.[37,38] However, our experiments show that despite the possible change in secondary structure, the positively charged myoglobin has a strong affinity for the silica surface. In this study, we specifically focus on the adsorption of myoglobin and its corona formation on silica NPs at different pH values. A separate detailed study would be necessary to investigate the effect of pH on the secondary structure of myoglobin adsorbed within the hard and soft coronas. For pH > IEPMGB, despite both myoglobin and silica NPs being negatively charged, significant adsorption of the protein on silica NPs can be observed (Figure b). However, a gradual (instead of rapid) increase in the amount of protein adsorbed in the region ceq < 0.5 mg mL–1 highlights a weaker affinity adsorption at pH > IEPMGB in comparison to pH < IEPMGB. We find that the maximum surface excess of myoglobin on silica NPs decreases with increasing pH from 8 to 10. The observed difference in the adsorption behavior at pH > IEPMGB and pH < IEPMGB can be attributed to the change in protein–NP and protein–protein electrostatic interactions, which alters the adsorption affinity and packing of molecules on the surface of silica NPs.[45]

Figure 3

Amount of protein adsorbed (Γ) as a function of the equilibrium concentration of protein (ceq) at different pH and salinity. (a, b) Adsorption isotherms for myoglobin binding to silica NPs with no added NaCl at (a) pH < IEPMGB and (b) pH > IEPMGB. The scattered points are the experimental data, and the lines represent the best fit using the GAB model given in eq . For pH < IEPMGB, the amount of myoglobin adsorbed rapidly reaches the maximum value due to the strong electrostatic attraction between the protein and NP. For pH > IEPMGB, the amount of myoglobin adsorbed on NPs shows a gradual increase. (c, d) Experimental isotherms showing the adsorption of myoglobin on silica NPs with 100 mM NaCl added to solution at (c) pH < IEPMGB and (d) pH > IEPMGB.

Electrostatically driven adsorption processes are strongly dependent on the charges or potentials of the interacting adsorbate–adsorbent pair, as well as the presence of electrolytes in the surrounding medium.[46] Here, we investigate the effect of the presence of a 1:1 electrolyte (NaCl) on the adsorption of myoglobin onto silica NPs. We added 100 mM NaCl to the dispersion of myoglobin and silica NPs to reduce the Debye length to ∼1 nm.[47] At pH < IEPMGB with added NaCl, the adsorbed protein sharply increases at ceq < 0.2 mg mL–1 (Figure c), which is similar to the absorption behavior with no added NaCl (Figure a). However, the amount of protein adsorbed with added salt continually increases beyond the high-affinity regime (ceq < 0.5 mg mL–1). This continuous increase of Γ at ceq > 0.5 mg mL–1 is attributed to protein–protein attraction and the formation of a soft myoglobin corona on silica NPs, which is not observed in the absence of added NaCl. At pH > IEPMGB with 100 mM NaCl, the adsorption isotherms show a rapid increase of protein adsorption in the range ceq < 0.5 mg mL–1 and shows gradual increase at ceq > 0.5 mg mL–1. This behavior can be attributed to the adsorption of protein molecules via the amino acid charge regulation mechanism, as discussed in Sections and 3.4.

Effect of pH on Adsorption with No Added NaCl

To understand the adsorption mechanism of myoglobin onto silica NPs, the parameters of the GAB model, namely, Γm, Khard, and Ksoft, were extracted for each pH and salinity by fitting the experimental data using eq (Figure b–d). The Γm of adsorption of myoglobin of silica NPs increases upon increasing the pH from 4 to 8. Then, Γm monotonically decreases at pH > IEPMGB. The increase in Γm from pH 4 to 8 is attributed to the decrease in repulsion between adsorbed myoglobin molecules on the NP and free molecules in solvent bulk. At pH 4, myoglobin molecules are strongly positively charged, which adsorb onto the negatively charged silica surface by electrostatic attraction. Due to a high net charge of myoglobin (Figure c), electrostatic repulsion exists between myoglobin molecules adsorbed on the NP and unadsorbed molecules in bulk. The electrostatic repulsion between the adsorbed and unadsorbed molecules hinders the further adsorption and limits the Γm at pH 4.

Figure 4

(a) Schematic representation of adsorption constants Khard for hard corona formation driven by protein–NP interaction and Ksoft for soft corona formation driven by protein–protein interaction. The values of Khard and Ksoft are estimated by fitting the experimental adsorption isotherms using the GAB model, as given in eq . (b) Change in the maximum amount of protein adsorbed in hard corona (Γm) as a function of dispersion pH with and without adding NaCl. In the absence of added NaCl, the Γm is maximum at pH 8 due to the preferred orientation of the adsorbed protein on silica surface. (c, d) Adsorption constants for hard corona and soft corona, respectively, as a function of pH and salinity.

Upon increasing the pH to 8, the charge on the protein significantly decreases (IEP of myoglobin of ∼7), thus reducing the repulsion between adsorbed and unadsorbed protein molecules. This decrease in the repulsion drives the observed increase in the maximum surface excess. It should be emphasized that the apparent large adsorption of myoglobin onto silica NPs at pH 8 is not an artifact of self-aggregation of protein molecules in bulk. This was verified experimentally where no aggregate separation, i.e., no significant protein depletion from the solvent upon centrifugation (18 210g, 2 h), was observed for dispersion containing only myoglobin at pH 4, 8, and 10 (Figure S4).[48]

At pH > IEPMGB, where the charge of myoglobin molecules and silica surface are both negative, the adsorption of protein is driven by the charge regulation of amino acids.[36,42] In protein adsorption, histidine plays a significant role due to its highly ampholytic nature at physiologically relevant pH.[36] For silica NPs dispersed in water with no added NaCl, the concentration of hydrogen ions (counterions) is higher near the negatively charged surface than that of the bulk, which results in a pH gradient in the diffused electrical double layer.[47,49] Near the surface of silica NPs, the pH is lower than the bulk solution due to higher concentrations of hydrogen ions that act as counterions. When the pH is above the IEPMGB, the neutral tautomers in histidine near the surface of silica NPs are protonated into a cationic state, which results in positively charged residues of myoglobin molecules adsorbing onto the negatively charged silica surface.[13,50,51] Therefore, the myoglobin molecules carrying a weak negative net charge are adsorbed by the charge regulation effect at pH 8, and the weak protein–protein repulsion allows the additional adsorption of myoglobin leading to a maximum Γm. Upon increasing pH from 8 to 10, a large negative net charge induces a protein–NP repulsion, which leads to a decrease in Γm. In the absence of added NaCl, i.e., Ksoft = 0; thus, the protein only forms hard corona. An important factor influencing the amount of protein in hard corona is the orientation of individual proteins on the NP surface.[52] Here, we compute the surface electrostatic potential map and net dipole moment of myoglobin to better understand the relation among pH, protein orientation, and amount of adsorbed protein in the hard corona.

Increasing pH changes the protonation state of the amino acids of myoglobin and results in an increase in the negatively charged residues on the surface of the protein. This change in the surface charge distribution can be theoretically quantified using electrostatic potential mapping and estimating the dipole moment of the protein (Figure a–c). The dipole moment of a protein is the vector joining the center of the negative charges to the center of the positive charges on the protein molecule. The magnitude of the dipole moment vector is the measure of anisotropy in the distribution of charges on the surface of a protein, and the direction of the dipole moment vector points toward the center of the positive charges of the protein.

Figure 5

(a–c) Electrostatic surface maps for myoglobin (PDB # 5ZZE) at pH 4 (a), 8 (b), and 10 (c). The negative and positive residues on the protein are represented in red and blue, respectively. The inset in (a) is the color scale in the unit of elementary charge, e. The electric dipole moment vector of the protein is indicated by the arrow. Here, the view of the protein is fixed and the changes in the dipole moment with pH are represented by the change in the length and orientation of the vector shown by the arrow. (d) Magnitude of the dipole moment of myoglobin as a function of pH. The value of the total dipole moment shows a maximum near IEPMGB due to asymmetric charge distribution on the protein surface.

The magnitude of the dipole moment of myoglobin increases from pH 4 to 6, remains nearly constant from pH 6 to 9, and decreases at pH 10. The dipole moment calculations indicate that the charge distribution on the surface of myoglobin changes from more symmetric at pH 4 to less symmetric at pH 8 and then again more symmetric at pH 10 (Figure d). We believe that a more asymmetric charge distribution on the surface of the protein drives an oriented attachment of the protein molecule on the silica surface, thus favoring a closer packing. The preferential orientation of the protein combined with a reduced protein–protein repulsion on the NP surface due to the small net charge of myoglobin at pH 8 (Figure c) may lead to its improved packing and highest Γm.

We find that Khard shows a rapid decrease with increasing pH from 4 to 10 (Figure c). The large value of Khard at pH < IEPMGB highlights the strong electrostatic attraction between myoglobin molecules and the silica surface. As pH increases from 4 to 6, the net charge of myoglobin significantly decreases from +15e to +5e (Figure c). As a result, Khard dramatically decreases due to weakened electrostatic attraction at pH 6. At pH > IEPMGB, the adsorption of myoglobin on the silica surface is driven by the charge regulation effect. In this regime, proteins are adsorbed with a preferred orientation due to asymmetric surface charge distribution (large dipole moment), and Khard decreases from pH 6 to 8.

A further decrease in Khard from pH 8 to 10 is attributed to the increasing electrostatic repulsion between the silica surface and free molecules in solvent bulk as the net charge of myoglobin becomes more negative. In the complete pH range, the value of Ksoft remains very small (<0.025 mL mg–1) (Figure d), i.e., myoglobin only forms a hard corona on NPs in the absence of added salt.

Effect of Added NaCl on Adsorption

Addition of salt drives the formation of a soft corona on top of the hard corona structure formed by myoglobin on silica NPs. For the negatively charged surface in a solution with a high NaCl concentration, oppositely charged counterions bind to negative sites.[47] Therefore, adding NaCl lowers the surface potential of silica NPs and screens all electrostatic pair interactions, including NP–NP, protein–protein, and protein–NP.[53] In our adsorption studies, we find that the addition of NaCl into myoglobin–silica NP dispersions leads to a decrease in the amount of protein adsorbed in hard corona, except at pH 4 (Figure b).

The increase in Γm at pH 4 upon adding NaCl highlights a screening of protein–protein repulsions between strongly charged myoglobin molecules on the surface of the NP. This screening leads to a small increase in the amount of protein adsorbed on the NP.[54] The addition of NaCl also screens the attraction between the silica surface and myoglobin in solvent. As a result, the value of Γm decreases with added NaCl and remains almost constant (Γm = 0.6 ± 0.1 mg m–2) from pH 6 to 10 (Figure b). We find that a decrease in the value of Khard upon the addition of NaCl indicates a weakening of attraction between silica NPs and the myoglobin. This further ascertains that the formation of hard corona is driven by electrostatic attraction, which are screened with the addition of NaCl.

Addition of salt also screens the electrostatic repulsion between adsorbed myoglobin molecules on the silica surface and myoglobin in solvent bulk.[55] This is evident from the increase in the value of Ksoft at all pH upon the addition of NaCl (except pH 10), as shown in Figure d, as well as the significant difference in the values of adsorption free energy (ΔGads0) for two distinct binding sites. We calculate ΔGads0 from the relation[41,56]

where NA is Avogadro’s number, kB is the Boltzmann constant, T is the temperature, and Kads is the adsorption constant. We find that the values of ΔGads0 at pH 6 in 100 mM NaCl aqueous solution with Khard and Ksoft are −16 and −4.5 kJ mol–1, respectively. The significant difference in the values of ΔGads0 indicates that one population of the protein is strongly bound to the silica surface while the other is weakly bound, presumably onto adsorbed proteins, thus forming corona as hard and soft. This interpretation of soft corona formation based on binding energy is in distinction from previous reports where the protein separable during centrifugation was associated with the presence of soft corona.[9]

To investigate the effect of added NaCl on the amount of adsorbed protein in soft corona, we quantify the fraction of myoglobin present in soft corona (ϕ), by using the relation ϕ = (Γ – Γm)/Γ. Note that Γm is the maximum surface excess of myoglobin in hard corona only. We find that in the absence of NaCl at all tested pH ϕ ∼ 0, i.e., myoglobin only forms hard corona on silica NPs without added NaCl. Upon the addition of NaCl, ϕ increases significantly in nearly all tested pH, highlighting that myoglobin molecules form a soft corona upon the addition of NaCl (Figure a). As pH increases from pH 4 to 10, the ϕ decreases from 0.6 to 0, indicating a gradual loss of soft corona on NPs. This gradual loss of soft corona with increasing pH can be attributed to the difference in electrostatic screening mechanisms of NaCl stability at low and high pH (Figure b). A previous study showed that chloride ions are more effective in screening charges than sodium ions because of their larger polarizability.[57] As a result, short-range repulsion between positively charged myoglobin molecules is more effectively screened at pH < IEPMGB than for a negatively charged myoglobin at high pH > IEPMGB. This change in screening mechanism leads to the observed decrease in Ksoft and ϕ with increasing pH. Therefore, myoglobin forms a soft corona at pH < IEPMGB in the presence of 100 mM NaCl, which is lost upon increasing the pH > IEPMGB.

Figure 6

(a) Fraction of myoglobin in the soft corona as a function of pH and concentration of NaCl. (b, c) Schematic representation of the structural change of the protein corona around a silica NP upon adding NaCl to a myoglobin–silica NP dispersion at pH < IEPMGB (b) and pH > IEPMGB (c).

Conclusions

We studied the influence of pH and added NaCl on the binding of myoglobin onto silica NPs. In the absence of added NaCl at pH < IEPMGB, the myoglobin molecules are adsorbed by electrostatic attraction with a presumably random orientation at the silica surface due to the symmetric charge distribution of myoglobin. At pH > IEPMGB, both silica NPs and myoglobin are negatively charged, and the adsorption is driven by the charge regulation mechanism of amino acid groups in myoglobin. In this pH range, a high dipole moment of myoglobin may drive its adsorption in a preferred orientation on the silica surface. In the absence of NaCl, myoglobin forms a hard corona on silica NPs. The addition of NaCl screens electrostatic repulsion between free protein and protein molecules adsorbed on silica NP surface, thus driving the formation of an additional soft corona at pH < IEPMGB. This soft corona is lost upon increasing the pH > IEPMGB due to electrostatic repulsion between protein–protein and protein–NP surface. The approach presented in the study could allow progressive access to the “hidden” binding sites in the hard corona, which currently is a challenge,[58,59] thus providing a platform to design nanomaterials with controlled uptake and binding of proteins with a preferred orientation.