«Kinetic Investigations of Thiolate Protected Gold Nanoparticles: Protein Interactions, Electron Transfer, and Precursor Formation By Brian N. Turner ...»
Synthesis of Au696Tiopronin265 Nanoparticles Tiopronin MPCs were synthesized as previously described. 122 500 mg HAuCl4 3H2O was dissolved in 70 mL 6:1 methanol:acetic acid and chilled in an ice bath for 30 minutes, giving a yellow solution. 110 mg tiopronin was then added to the solution, forming a ruby red intermediate. Upon fading of the solution to near colorless, 500 mg sodium borohydride in a minimum of water was immediately and rapidly added to the mixture. A violent, exothermic reaction results, leaving a black suspension. The solution is stirred overnight, yielding polydisperse nanometer sized particles. The resulting solution is centrifuged, and the solvent decanted. The particles were then resuspended in methanol, and centrifuged a second time. 1H NMR: 4.00 ppm broad singlet (tiopronin methylene on cluster), 3.76 ppm broad singlet (tiopronin CH on cluster), 1.66 ppm very broad singlet (tiopronin methyl on cluster). TEM: average particle diameter of 2.9 nm
0.8 nm. UV/Visible spectroscopy: no surface plasmon resonance band observable.
TGA: 24 % by mass organic (tiopronin). Average formula: Au 696Tiopronin265.
Place Exchange to Form Series 3 Peptide-Nanoparticle Conjugates 20 mg of Au696Tiopronin265 was dissolved in 2 mL water. Then, a 1:10 stoichiometric peptide:tiopronin mass of peptide was dissolved in 3 mL of water. The solutions were mixed together and stirred for 72 hours. The resulting nanoparticle-peptide conjugate solutions were subsequently transferred to 10,000 MWCO centrifuge filters, and centrifuged a total of 10 times. The resulting clean solutions were analyzed by 1H NMR to determine purity and exchange percentages and by UV-Vis to determine nanoparticle concentrations by Beer’s law. Drying of the solution was avoided in order to reduce particle agglomeration. Series 1 and 2 peptides (see Table 6) were place exchanged by similar methods, except that dialysis in 10,000 MWCO membranes was used in place of centrifuge filters, and the use of buffers during place exchange and purification was utilized if there was a solubility or stability problem. Series 1 peptides were exchanged with Au216Tiopronin129, and series 2 peptides were exchanged with Au118Tiopronin71.
QCM Immunosensor Construction and Detection of HRSV A stable frequency baseline was collected while PBS was pumped through the flow cell with a peristaltic pump. Protein G (200 µg/mL) was pumped through the flow cell until saturation was reached as indicated by a level frequency reading, at which point the solution was switched back to PBS. In the same manner, BSA (1 mg/mL) was then flowed, followed by Palivizumab (40 µg/mL). HRSV, diluted in PBS from growth medium to the desired concentration, was then flowed for a period of about 10 minutes.
All solutions were in PBS and flowed at a rate of 20 µL/min. After HRSV was flowed through the cell, all lines and flow cells were purged with 20% bleach solution to decontaminate.
Adaptation of the QCM Immunosensor to Detect Peptide Presenting Gold Nanoparticles The above Scheme was followed for sensor construction, except that some concentrations were changed to optimize sensitivity and/or save on material. Protein G was used at 50 µg/mL, BSA at 1mg/mL, and PZ at 10 µg/mL. Detection of the nanoparticle solution was achieved by diluting the water solution in PBS to a minimum of 4:1 to avoid drastic differences in solution conductivity and viscosity, which causes inaccuracies in converting the frequency signal to bound mass. The nanoparticle solution was then flowed through the cell as normal.
QCM Immunosensor for HRSV (Note: In concert with Dr. Yibin Zhang, a QCM based immunosensor for HRSV was developed.123 As the data is largely that of Dr. Zhang’s, only the major results that are necessary to the understanding of the peptide presenting gold nanoparticle data that follows are discussed.) A QCM immunosensor strategy for SARS-associated coronavirus124 was adapted to specifically and quantitatively detect HRSV. The platform, depicted in Figure 19, is composed of an Fc binding cell wall protein such as Protein A, G, or L from Staphylococcus Aureus, a blocking agent such as bovine serum albumin (BSA), and a polyclonal or monoclonal (mAb) detection antibody.
Figure 19: Cartoon schematic of the QCM based immunosensor for HRSV detection.
Gold electrodes wrapped about piezoelectric quartz are functionalized with Protein G, which binds PZ in an orientation amenable to HRSV detection. BSA is used to fill in bare gold spaces left after adsorption of Protein G.
Dr. Yibin Zhang was responsible for determining the ideal combination of sensing antibody and Fc binding protein (data not provided here), and determined that the combination of Protein G and Palivizumab optimized the sensitivity and specificity of the instrument towards HRSV. The sensor construction and subsequent detection of HRSV from cultured growth medium diluted in PBS is displayed in Figure 20.
Figure 20: Detection of HRSV with the QCM-PZ immunosensor. Light blue indicates that PBS is being flowed without any analyte. Prot G indicates that 200 µg/mL Protein G is flowed through the cell for the corresponding time duration in white giving a total immobilized protein mass of 99 ng on the gold electrode surface, BSA is 1 mg/mL and gave a mass of 35 ng, Palivizumab is 40 µg/mL and gave a mass of 229 ng, and RSV is 8 x 105 PFU/mL yielding a detected viral mass of 210 ng. All solutions were in PBS and flowed at a rate of 20 µL/min. Figure designed by Dr. Yibin Zhang.
Experiments to evaluate the specificity of the sensor are displayed in Figure 21. These experiments looked at replacement of the antibody and antigen to investigate non-specific binding, associated with false positives. Anti-HA and anti-PA were compared to PZ for detection of HRSV and HRSV growth medium and human metapneumovirus (HMPV), a close relative of HRSV, were compared to HRSV for detection by PZ. All detection steps were run for 10 minutes.
Figure 21: Detection and control experiments for the QCM-PZ immunonosensor.
Experiments 1 and 2 are sensors detecting HRSV with anti-HA(mouse monoclonal 12CA3 from the Vanderbilt antibody core) and anti-PA (mouse monoclonal anti-PA83 from Alpha Diagnostics) replacing PZ. Experiment 3 is the detection experiment with HRSV and PZ (labeled as Synagis in this figure). Experiments 4 and 5 use the PZ immunosensor, but RSV growth medium and HMPV are substituted for HRSV during the detection step. Figure designed by Dr. Yibin Zhang.
The results from the HRSV growth medium are taken as a negative background because it is present in the positive HRSV sample. A negative background was therefore determined as the signal of HRSV growth medium plus three standard deviations, which is 59.3 ng. Anything below this was determined to be a negative result for detection.
Neither of the control antibodies provided significant signal when compared to PZ and to the negative threshold, indicating that HRSV is not electrostatically or physically sticking to other proteins in the sensing layer. Also, none of the other antigens provided a significant signal. The lack of HRSV growth media binding would indicate that the signal is not from contaminants in the HRSV sample that originate from the viral culture.
The lack of HMPV binding suggests that the interaction between the Fab region of PZ and the antigenic surface protein of HRSV, in this case F, is of a nature dependent upon the specific primary and secondary structure of the HRSV F protein.
In conclusion, a QCM platform using the combination of protein G and PZ was found to be sufficient for the specific detection and quantitation of HRSV in buffer. The sensor was also found to be capable of detection of HRSV in infected MDCK cells and in human patient nasal wash samples. The sensor had a limit of detection of 9 x 10 3 PFU/mL.
Development of Peptide Presenting Gold Nanoparticles Peptides listed in Table 6 were place exchanged successfully onto tiopronin protected gold nanoparticles. TGA and TEM yielded the number of gold atoms in the nanoparticle core and the number of tiopronin molecules in the unexchanged material, and 1H NMR yielded the ratio of peptide to tiopronin on the surface, as demonstrated in Figure 21.
Figure 22: 1H NMR, with double water-gate solvent suppression, of a peptidenanoparticle conjugate, in this case Au696Tiopronin255Peptide(3-F)10. The extent of peptide exchange is determined by integrating the peak at 0.8 ppm, which represents leucine, isoleucine, and valine methyl protons and comparing that to the peak at 1.5 ppm, which represents tiopronin methyl protons.
Table 7 lists nanoparticle mimics that were synthesized and evaluated with the QCM immunosensor.
Table 7: Nanoparticle mimics studied with the QCM HRSV immunosensor. Peptide IDs refer to Table 3-x. nAu, ntiopronin, and npeptide refer to the number of gold atoms, tiopronin molecules, and peptide molecules per nanoparticle, respectively. d TEM is the diameter of the gold core as determined by TEM. PZ detection qualitatively describes the relative amount and nature of detection.
Detection of Peptide Presenting Gold Nanoparticles with the QCM Immunosensor for HRSV As previously summarized in Table 7, many of the nanoparticle mimics evaluated did not interact significantly with PZ. The two notable exceptions were Au696Tiopronin261Peptide(3-F)4 (mimic A) and Au118Tiopronin71Peptide(2-F)3 (mimic B).
As Mimic B was found to non-specifically bind to the non-sensing proteins on the gold surface (data available upon request), Mimic A will be the focus of the remainder of this chapter. Of importance, one should recall that mimic A was functionalized with the same peptide that gave the strongest signal in a peptide ELISA assay, as discussed in Chapter II (referred to there as peptide A, and here as peptide 3-F). The primary sequence of this peptide and of peptide 2-F contains the same sequence as peptides studied by MedImmune79 and, more recently, by McLellan and coworkers. 90, 125 The authors were not aware of the latter two works at the time that the following studies were conducted.
Mimic A is of the most interest as it bound strongly and specifically to the QCM HRSV immunosensor. A proof of concept experiment was performed in which a sensor was constructed in the usual way (see methods), except that PZ was left out entirely. In this instance, mimic A was not detected by the sensor. Upon addition of PZ, however, the mimic was detected by its strong binding to the sensor, as shown in Figure 23.
Figure 23: Detection of Au696Tiopronin255Peptide(3-F)10 by the QCM immunosensor with and without PZ. Blue sections represent the flow of PBS without analyte. White sections indicate the binding of protein G (50 µg/mL, 192 ng bound), BSA (1 mg/mL, 53 ng bound), nanoparticle mimic (818 µg/mL, no detectable amount of material bound), PZ (10 µg/mL, 92 ng bound), and nanoparticle mimic again (409 µg/mL, 150 ng bound after 15 minutes of flow), respectively.
In order to further characterize the affinity of mimic A for the PZ sensor, a concentration study was undertaken. Six concentrations of mimic A in PBS were evaluated for their bound mass and binding on rate in order to determine kf, kr, and Ka for Scheme 3-1.
Scheme 3-1: Langmuirian binding model for an antigen and antibody. The first case is appropriate for large antigens, such as virions, and the second case better describes smaller antigens (peptides, small proteins, nanoparticles).
The data for mimic detection, along with data for the construction of the sensor layer for each experiment is summarized in Table 8.
Table 8: Summary of mimic A detection at various concentrations, along with sensor characterization data for each experiment. n is the number of bound molecules of the corresponding protein. θ is the percent coverage of Protein G by PZ. NP is the mass of mimic A bound after the 3 minutes in the middle of the binding curve. mmax is the
theoretical maximum mass of nanoparticle that could bind in either a 1:1, 2:1, or 4:1
model (see Scheme 3-1). None of the sample concentrations analyzed provided an empirical mmax in the time durations allowed, which were limited by the amount of sample available for analysis. * indicates that the data could not be determined due to an instrumental error. Sensor construction data for 818 µg/mL was lost due to a crash of the data logging program before the detection step.
Normally, it is straightforward to determine equilibrium association (Ka) and dissociation (Kd) from m vs. C data, but in our case, the inability to determine saturation data renders this method unfeasible. Binding coefficients were instead determined from the maximum on rates at each concentration. Figure 24 illustrates the concentration dependence of the on rate.
Figure 24: Representative QCM mass versus time curves at various concentrations of mimic A during the detection step.
For any binding model in Scheme 3-1, the kinetics are characterized by:
( ) ( ) (3-1) where mt is the bound mass at time t, mmax is the maximum mass that can bind to the sensor, the total quantity in the natural logarithm is the uncovered surface percentage and the fractional quantity is the surface coverage percentage at a given time, and τ is the time constant associated with a given concentration of antigen.126, 127 In order to determine mmax in the absence of data at saturation, a calculated value is used based on the average number of PZ molecules bound to the sensor. For a 2:1 binding model, which is expected for this system, mmax = 210 ng. This value of mmax provides better fits of data to theory than if individually calculated values for each sensor are used, probably because bound mass is not expected to track linearly with immobilized antibody due to the “hook effect.”128, 129 The negative inverse time constant can be determined from the slope of the linear portion of a linear plot of equation 3-1, in the form of the logarithmic quantity versus t, as illustrated in Figure 25 for one of the 409 µg/mL experiments.
Figure 25: Plot to determine the time constant for mimic A binding to the Palivizumab sensor, τ, for a single experiment. For this 409 µg/mL experiment, -τ-1 = -6.50 x 10-4 s-1.
The slope is only determined for the linear portion (black), where the binding rate has reached a maximum.
The binding coefficients are determined from the equation: