Probing Temperature- and pH-Dependent Binding between Quantum Dots and Bovine Serum Albumin by Fluorescence Correlation Spectroscopy

Luminescent quantum dots (QDs) with unique optical properties have potential applications in bio-imaging. The interaction between QDs and bio-molecules is important to the biological effect of QDs in vivo. In this paper, we have employed fluorescence correlation spectroscopy (FCS) to probe the temperature- and pH-dependent interactions between CdSe QDs with carboxyl (QDs-COOH) and bovine serum albumin (BSA) in buffer solutions. The results have shown that microscopic dissociation constant K′D is in the range of (1.5 ± 0.2) × 10−5 to (8.6 ± 0.1) × 10−7 M, the Hill coefficient n is from 0.4 to 2.3, and the protein corona thickness is from 3.0 to 9.4 nm. Variable-temperature measurements have shown both negative values of ∆H and ∆S for BSA adsorption on QDs-COOH, while pH has a profound effect on the adsorption. Additional, FCS measurement QDs-COOH and proteins in whole mice serum and plasma samples has also been conducted. Finally, simulation results have shown four favored QD binding sites in BSA.


Introduction
Luminescent quantum dots (QDs) have attracted tremendous interest towards applications in biomedical fields because of their outstanding optical properties, such as extreme brightness [1,2] and tunable emission wavelength, yet with a narrow emission band [3][4][5]. Meanwhile, toxicity of QDs has been extensively studied at the level of biological macro-molecules [6,7], subcellular organelles [8], cell [9][10][11], and others [12]. In biological fluids, QDs can be coated by a layer mainly of proteins, as well as other bio-macromolecules [13], which is called a protein corona. The role of the protein corona

Materials Preparation
All of the ZnS/CdSe quantum dots (QDs-COOH and QDs-PEG) were precipitated with 2-propanol to remove free Cd 2+ and 3-Mercaptopropionic acid (MPA) [33], and then diluted to 8 nM in phosphate buffer with different pH (pH = 6.0, 7.4 and 9.0) and ultrasonicated for 10 min (with the temperature below 348 K) to ensure the full dissolution of QDs, and then purified with the YM-220 ultra-filtration membrane to remove large particles.
BSA protein powder (13.28 g) was dissolved in 2 mL phosphate buffer with different pH (pH = 6.0, 7.4 and 9.0) and then diluted to different concentrations (50 nM to 100 µM). The as-prepared BSA solutions were ultrasonicated for 10 min to ensure the full dissolution of BSA and purified with the YM-220 ultra-filtration membrane to remove large particles. The BSA solutions were freshly prepared for every experiment, or otherwise stored at −4 • C to reduce protein aggregation and sedimentation.
Mouse blood was sampled by periorbital puncture [34]. To obtain the plasma, fresh blood was collected in anticoagulant tubes and then centrifuged at room temperature at 3000 rpm for 10 min. The plasma was light yellow in the upper layer of the tube. To obtain the serum, fresh blood was collected in normal tubes without any anticoagulant and kept at room temperature for 2-4 h. Then the blood was centrifuged at room temperature at 3000 rpm for 10 min. The serum could be obtained from the supernatant. All of the blood samples, including plasma and serum, were freshly prepared for every experiment.

Physicochemical Characterization of QDs
The morphology and sizes of QDs were imaged by transmission electron microscopy (TEM, Tecnai F20, FEI, Hillsboro, Oregon, OR, USA) at an operation voltage of 200 kV. QD suspension (at a concentration of 8 nM) was dropped onto a copper grid coated by a carbon membrane and then dried at room temperature for TEM imaging. Ultraviolet-visible spectroscopy (UV-VIS) spectra of QDs were acquired on a Lambda 950 spectrophotometer (Perkin-Elmer Corporation, Waltham, MA, USA). The hydrodynamic size and zeta potential of the QDs were measured on a Malvern Zetasizer Nano ZS (Malvern Instruments, Worcestershire, UK).

FCS Measurements
FCS measurements were performed on a home-built system, which was based on an inverted Olympus IX73 microscope. A NKT supercontinuum white-light pulse laser was used as the excitation laser. The repeat frequency, excitation wavelength, and laser power were 3.123 MHz, 592 nm, and 60 µW, respectively. The excitation light was focused by a 60× water immersion objective. The fluorescence emitted by the sample was collected by the same objective and detected by an Excelis Detector (SPCM-AQRH-16). Timeharp 200 (PicoQuant, Berlin, Germany) was used for photon-counting. The coverslip used in the cuvette was coated with a layer of PEG (2 k) to avoid adsorption of fluorescent molecule. The temperature was controlled by an mK1000 series temperature controller (INSTEC, Inc., USA).
QD solutions (200 µL, 8 nM) were mixed with the equal volume BSA solutions with different concentrations (50 nM to 100 µM) diluted in different phosphate buffers (pH = 6.0, 7.4, and 9.0). The solution was incubated at different temperatures (298 K, 300 K, 303 K, and 308 K) for 1 h before FCS measurements. Each sample was measured at least three times and then the average was taken to reduce the error.
For the case of a Gaussian excitation profile, Aragon and Pecora [35] derived the three-dimensional autocorrelation function in terms of the average number of species crossing the focal volume N , τ i is the residence time of a molecule in the focal volume, and ω is a structure parameter equal to the ratio of the longitudinal to transverse size of the focal volume. The residence time, τ i , is related to the corresponding diffusion coefficient, D i , by D i = w xy 2 /(4τ i ) for the transverse direction, where W xy represents the radius in the x-y plane direction of the point spread function of the confocal system. The hydrodynamic radius is calculated according to the Stokes-Einstein equation: where k B is the Boltzmann constant, T is the absolute temperature, and η 0 is the solvent viscosity, R is the hydrodynamic radius.
To quantify the interactions between BSA and QDs, a model was adopted [36]: where d z ([Protein]) and d z (0) are the hydrodynamic diameters of QDs with and without protein, respectively, and c is a scaling constant. Since the number of bound protein molecules should be much less than the number of free protein molecules, the total protein concentration was used as free protein concentration. Fitting the dependence of the hydrodynamic radius on protein concentration directly gives the values of the microscopic dissociation constant K D and the Hill coefficient n for BSA-QD binding. For positively/negatively cooperative absorption, n is larger/less than 1, respectively. For serum protein adsorption on nanoparticles, n is close to unity [37,38].

Circula Dichroism (CD) Spectroscopy
CD spectra of QDs-COOH and BSA were recorded at room temperature on a Jasco J-500C spectropolarimeter. The solution of QDs-COOH and BSA was well dispersed in water and buffer with concentration of 0.2 mg/mL in different pH (6.0, 7.4, and 9.0).

Simulation
The initial structure of BSA was derived from the crystal structure of bovine serum albumin with PDB ID 4F5S (released on 13 May 2012 with 2.47 Å resolution) [39]. The ZnS/CdSe core region in QDs was treated as a spherical structure, and other parts of QDs remain. Both the BSA initial structure and QD models were then parameterized by CHARMM c36m force field parameters [40] and solvated in a rectangular simulation box with explicit TIP3P waters. After salvation, the simulation system was subjected to a steepest descent energy minimization for about 5000 steps, followed by a conjugate gradient for the next 5000 steps and equilibrated by a 500-ps molecular dynamics simulation to reduce the van der Waals conflicts. Finally, 1-ns molecular dynamics simulations were performed using the NAMD package [41] with a temperature of 298 K, periodic boundary, NPT ensembles, and different pH conditions (6.0, 7.4, and 9.0). The SHAKE algorithm was applied to constrain all of the chemical bonds, and atom velocities for start-up runs were obtained based on the Maxwell distribution at 298 K. For our simulation system, 10 independent simulations were carried out with different start-up velocities. The isothermal compressibility was set to 4.5 × 10 −5 /bar for the solvent simulations. The electrostatic interactions were treated by the particle mesh Ewald (PME) algorithm with and interpolation order of 4.0 and a grid spacing of 0.12 nm. The van der Waals interactions were calculated by using a cut-off of 12 Å. All of the molecular dynamics simulations were performed with a time step of 2 fs, and the coordinates for all the simulation systems were saved every 1 ps.

Physicochemical Characterization of Carboxyl QDs (QDs-COOH)
The core size of QDs-COOH revealed by TEM (Figure 1a,b) is 4.5 ± 0.5 nm. The hydrodynamic radius of QDs-COOH is 8.5 ± 0.1 nm measured by DLS (Figure 1c). The slightly larger hydrodynamic radius can be due to the electrical double layer on QDs-COOH surfaces. The absorption and emission peaks of QDs-COOH are around 580 and 620 nm, respectively ( Figure 1d). The radius size (R) and zeta potential (ζ) of QDs in buffer solutions with different pH are listed in Table 1. All QDs-COOH showed negative surface charge in buffer solutions.  Table 1. All QDs-COOH showed negative surface charge in buffer solutions.

BSA-QD Interaction at Different Temperatures
Typical FCS correlation curves for QDs-COOH with different BSA concentrations (pH = 7.4, T = 298 K) is shown in Figure 2b. The characteristic diffusion time shifts to a longer time at higher BSA concentrations, indicating BSA binding on QDs-COOH. Figure 2c-f show hydrodynamic radii of QDs-COOH derived from the FCS curves at different BSA concentrations and different temperature (300, 303, and 308 K, respectively). Further fitting with Equation (3) give the microscopic dissociation constant K' D and Hill coefficient n. As the temperature increases, the microscopic dissociation constant (K' D ) gradually increases (weaker interaction) and the protein corona thickness (∆R) gradually decreases (Figure 2g). Similar results were observed at pH 6.0 and pH 9.0 ( Figures S1 and S2). This indicated a decreased affinity between QDs-COOH and BSA at higher temperatures. Further, lnK is plotted against 1/T to extract thermodynamic parameters for BSA-QD interactions according to: where K is the binding constant (inverse to K D ), k B N A is the gas constant. The free energy change (∆G) is then calculated by: Nanomaterials 2017, 7, 93 6 of 12 where K is the binding constant (inverse to K′D), kBNA is the gas constant. The free energy change (ΔG) is then calculated by: All of the results listed in Table 2 show that ΔH < 0, ΔS < 0, and ΔG < 0.This indicates that the binding of BSA to the QDs-COOH was enthalpy-driven, in which the affinity is lower at higher temperatures. The affinity may be contributed by the hydrogen-bond interactions [42] electrostatic interaction, cavities, or hydrogen-bond interactions [43][44][45][46] between QDs-COOH and BSA.     Table 2 show that ∆H < 0, ∆S < 0, and ∆G < 0.This indicates that the binding of BSA to the QDs-COOH was enthalpy-driven, in which the affinity is lower at higher temperatures. The affinity may be contributed by the hydrogen-bond interactions [42] electrostatic interaction, cavities, or hydrogen-bond interactions [43][44][45][46] between QDs-COOH and BSA.

BSA-QDs-COOH Interaction at Different pH
Typically, BSA is presented in an N configuration at pH 7.4. But BSA can undergo N-F (normal-fast) transition when pH < 5.0 and N-B (normal-basic) transition when pH > 9.0 [47,48] (Figure 3a). FCS measurements have been done at different pH of 6.0, 7.4, and 9.0 to investigate the interaction of QDs-COOH and BSA (Figure 3b-e). Figure 3c shows the autocorrelation curves for QDs-COOH with BSA at different concentrations at pH 6.0 and Figure 3c-e show hydrodynamic radii at different BSA concentrations at different pH (pH 7.4 and 9.0). K D is larger at pH 7.4 than that at pH 6.0 and 9.0 (Figure 3f), indicating a weaker interaction at pH 7.4 and an overall thinner protein corona at pH 7.4 ( Figure 3g).
The isoelectric point (i.e., pI) of BSA is 4.7 and pK a for carboxyl (-COOH) group is around 5.5 [49,50]. Both BSA and QD-COOH will have more negative surface potentials at higher pH, so weaker interaction is observed at pH 7.4 than that for pH 6.0, which is expected because of the increased repulsion from the more negative charge at the higher pH for both QDs-COOH and BSA. However, stronger interaction is observed at pH 9.0 than that at pH 7.4. BSA might undergo a conformational change [43][44][45] after combining with QDs-COOH or part of the N conformation of BSA shifts to the basic (B) conformation spontaneously at pH 9.0, and then the secondary and tertiary structures of BSA may change.
The peptide chains of the BSA protein are made up of units, such as α-helicity, β-sheet, and random curliness [51]. Contrasting to the N conformation, the B conformation is characterized by a significant loss in helical content and a phase of a decrease in the secondary structure [52], which results in partial unfolding of the secondary structure. Thus, BSA turns to be a loosening of the molecule with a loss of rigidity [53][54][55]. In addition, the disruption of tertiary contacts of BSA may be beneficial to the hydrogen bonding reaction between BSA and QDs-COOH. Thus, the more flexible structure of BSA at pH 9.0 can have higher affinity to QDs-COOH. CD spectroscopy has also been conducted to investigate the secondary structure of BSA before and after binding to QDs-COOH ( Figure 4). First, significant loss of the helical structure is clearly seen for BSA at pH 9.0 (Figure 4a). In the CD measurements of BSA on QDs-COOH, to minimize contributions of unadsorbed BSA, excess QDs (20 nM) and a low concentration of BSA (5 nM) were used. After binding to QDs-COOH, a slight decrease of the helical structure was also observed (Figure 4b-d). This indicates only a small secondary structure change for BSA adsorbed on the QDs-COOH surface.  The isoelectric point (i.e., pI) of BSA is 4.7 and pKa for carboxyl (-COOH) group is around 5.5 [49,50]. Both BSA and QD-COOH will have more negative surface potentials at higher pH, so weaker interaction is observed at pH 7.4 than that for pH 6.0, which is expected because of the increased repulsion from the more negative charge at the higher pH for both QDs-COOH and BSA. However, stronger interaction is observed at pH 9.0 than that at pH 7.4. BSA might undergo a conformational change [43][44][45] after combining with QDs-COOH or part of the N conformation of BSA shifts to the basic (B) conformation spontaneously at pH 9.0, and then the secondary and tertiary structures of BSA may change.
The peptide chains of the BSA protein are made up of units, such as α-helicity, β-sheet, and random curliness [51]. Contrasting to the N conformation, the B conformation is characterized by a significant loss in helical content and a phase of a decrease in the secondary structure [52], which results in partial unfolding of the secondary structure. Thus, BSA turns to be a loosening of the molecule with a loss of rigidity [53][54][55]. In addition, the disruption of tertiary contacts of BSA may be beneficial to the hydrogen bonding reaction between BSA and QDs-COOH. Thus, the more flexible structure of BSA at pH 9.0 can have higher affinity to QDs-COOH.
CD spectroscopy has also been conducted to investigate the secondary structure of BSA before and after binding to QDs-COOH ( Figure 4). First, significant loss of the helical structure is clearly seen for BSA at pH 9.0 (Figure 4a). In the CD measurements of BSA on QDs-COOH, to minimize  As a contrast, BSA adsorption on QDs-PEG at different pH in buffer solution was also investigated by FCS as a contrast ( Figure S3). Due to the presence of PEG, no BSA adsorption was observed at room temperature, confirming that PEG can reduce the binding of proteins to nanoparticles [56].

Protein Binding in Whole Mice Serum and Plasma Samples
FCS measurements have also been used to measure the protein binding on QDs-COOH in serum and plasma samples of mice. Due to the absence of blood cells (especially red blood cells), no significant interference to FCS measurements was observed. The thickness of the protein corona (∆R, corrected by the viscosity of serum and plasma) in plasma was found to be higher than that in serum and BSA solution ( Figure 5). This may be due to more proteins and other chemicals with different  As a contrast, BSA adsorption on QDs-PEG at different pH in buffer solution was also investigated by FCS as a contrast ( Figure S3). Due to the presence of PEG, no BSA adsorption was observed at room temperature, confirming that PEG can reduce the binding of proteins to nanoparticles [56].

Protein Binding in Whole Mice Serum and Plasma Samples
FCS measurements have also been used to measure the protein binding on QDs-COOH in serum and plasma samples of mice. Due to the absence of blood cells (especially red blood cells), no significant interference to FCS measurements was observed. The thickness of the protein corona (∆R, corrected by the viscosity of serum and plasma) in plasma was found to be higher than that in serum and BSA solution ( Figure 5). This may be due to more proteins and other chemicals with different sizes and affinities in plasma [57]. Additionally, the corona thickness is thinner at higher temperatures. As a contrast, BSA adsorption on QDs-PEG at different pH in buffer solution was also investigated by FCS as a contrast ( Figure S3). Due to the presence of PEG, no BSA adsorption was observed at room temperature, confirming that PEG can reduce the binding of proteins to nanoparticles [56].

Protein Binding in Whole Mice Serum and Plasma Samples
FCS measurements have also been used to measure the protein binding on QDs-COOH in serum and plasma samples of mice. Due to the absence of blood cells (especially red blood cells), no significant interference to FCS measurements was observed. The thickness of the protein corona (∆R, corrected by the viscosity of serum and plasma) in plasma was found to be higher than that in serum and BSA solution ( Figure 5). This may be due to more proteins and other chemicals with different sizes and affinities in plasma [57]. Additionally, the corona thickness is thinner at higher temperatures.

Simulation of BSA-QDs Interaction
Based on the structural analysis, four QD binding sites were identified in the BSA protein structure (as shown in Scheme 1). Binding site 1 resides in the central cavity of the BSA which is beneficial to the binding. Sites 2-4 are situated at the interface of two subdomains in the neighborhood (Site 2 resides at the interface between subdomains IA and IB. Site 3 resides at the interface between subdomains IIA and IIB, and site 4 resides at the interface between subdomains IIIA and IIIB) which contain substitutions of amino acids that create interactions across inter-domain boundaries within the BSA. There was a coordination of the carboxyl groups on the QDs with specific positively-charged amino acid residues which attributed to the high affinity by both electrostatic reaction and hydrogen-bonding interactions [42,43]. At pH 6.0, QDs were inclined to bind at binding sites 2-4 due to the electrostatic interactions formed by the residues with specific positive charges. At pH 7.4, QDs could bind to BSA in both the central cavity and the interface between neighborhood subdomains. However, at pH 9.0 QDs would prefer to bind with BSA in its central cavity (binding site 1) due to the favorable hydrogen bonding and hydrophobic interactions which lead to the higher binding capacity. reaction and hydrogen-bonding interactions [42,43]. At pH 6.0, QDs were inclined to bind at binding sites 2-4 due to the electrostatic interactions formed by the residues with specific positive charges. At pH 7.4, QDs could bind to BSA in both the central cavity and the interface between neighborhood subdomains. However, at pH 9.0 QDs would prefer to bind with BSA in its central cavity (binding site 1) due to the favorable hydrogen bonding and hydrophobic interactions which lead to the higher binding capacity.

Conclusions
In summary, we have investigated the interaction between CdSe QDs and BSA to probe their temperature-and pH-dependent binding by FCS via measuring the microscopic dissociation constant K′D, Hill coefficient n, and the thickness of the protein coronas in buffer solution, serum, and plasma. The binding of BSA to the QDs-COOH was found to be enthalpy-driven. The results also showed that the incensement of temperature was a disadvantage to the binding of protein and QDs, while pH had a profound and complex effect on the interaction by changing the structure of the BSA. This work also proved the effectiveness and accuracy of the FCS method to evaluate the kinetics and thermodynamic parameters of biomolecule interaction and the practicability of the FCS method in real serum and plasma samples. Finally, simulation results have showed four favored QD binding sites and different combination modes in BSA.

Conclusions
In summary, we have investigated the interaction between CdSe QDs and BSA to probe their temperature-and pH-dependent binding by FCS via measuring the microscopic dissociation constant K D , Hill coefficient n, and the thickness of the protein coronas in buffer solution, serum, and plasma. The binding of BSA to the QDs-COOH was found to be enthalpy-driven. The results also showed that the incensement of temperature was a disadvantage to the binding of protein and QDs, while pH had a profound and complex effect on the interaction by changing the structure of the BSA. This work also proved the effectiveness and accuracy of the FCS method to evaluate the kinetics and thermodynamic parameters of biomolecule interaction and the practicability of the FCS method in real serum and plasma samples. Finally, simulation results have showed four favored QD binding sites and different combination modes in BSA.