Engineering the Interface between Glucose Oxidase and

Article pubs.acs.org/Langmuir

Engineering the Interface between Glucose Oxidase and Nanoparticles Edurne Tellechea,‡,§,∥ Kenneth J. Wilson,⊥ Ernesto Bravo,§ and Kimberly Hamad-Schifferli*,†,‡ †

Department of Mechanical Engineering and ‡Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States § FideNa (Foundation for the R + D of Nanotechnology), Pamplona, Navarra, Spain ∥ Agrobiotecnology Institute, Public University of Navarra, Pamplona, Navarra, Spain ⊥ Department of Chemistry, Bloomfield College, Bloomfield, New Jersey, United States S Supporting Information *

ABSTRACT: The behavior of glucose oxidase (GOx) on gold nanoparticles (NPs) was investigated as a function of (1) NP surface chemistry, (2) stabilizing protein additives, and (3) protein microenvironment. GOx secondary structure and unfolding was probed by circular dichroism (CD) spectroscopy and fluorescence, and GOx enzymatic activity was measured by a colorimetric assay. We also examined the activity and structure of GOx after displacement from the NP surface. Generally, GOx behavior was negatively impacted by conjugation to the NP, and conjugation conditions could vary the influence of the NP. Surface chemistry and protein microenvironment could improve behavior, but addition of stabilizing proteins negatively influenced activity. After displacement from the NPs, GOx tended to remain unfolded, indicating that the interactions with the NP were irreversible.

■

INTRODUCTION Glucose oxidase is an important enzyme for many applications.1 Because of its high turnover, specificity, and stability, it has been widely used in a diverse range of commercial applications. Due to its ability to convert glucose into gluconic acid, it has been used to detect glucose in sensors for diabetics.2−4 GOx is present in honey, where as it acts as a natural food preservative because it converts O2 to H2O2, which is antibacterial. Thus, it has been adopted by the food industry for food preservation. Furthermore, because GOx activity involves electron transfer, it has been attractive for biofuel cells,5 which can potentially utilize glucose as a fuel.6−9 For many of these applications, the protein needs to be immobilized onto a substrate, either for direct electrical connection in biosensing and biofuel cells10 or for increasing surface area, which would enable efficient scaleup. These surfaces have been chemically and physically diverse, including carbon electrodes, flat or structured metal surfaces, or polymer and glass matrices. In particular, nanoparticles (NPs) have been attractive due to their high surface to volume ratios. Unfortunately, making functional interfaces between proteins and surfaces presents significant challenges.11,12 Protein structure and function are both strongly influenced by the NP surface chemistry, which can be highly variable.13 How surface properties affect immobilized enzymes is incredibly complex,14,15 and often the numerous noncovalent interactions between the surface and protein results in irreversible denaturation and loss of function. This is particularly crucial © 2012 American Chemical Society

for electrochemical applications, as electron transfer is highly dependent on the distance between the protein and the substrate, and even protein orientation can influence electron transfer efficiencies.7,16 Generally, these surface and interface effects worsen for NP and nanostructured surfaces, where curvature, the inhomogeneous nature of the surface, and the chemical passivation all contribute significantly to enzyme destabilization, unfolding, and loss of function.17,18 Making functional and optimized interfaces to proteins has been hindered by the fact that there are no good direct experimental probes, so understanding of the NP−protein interface is still limited. Some have been able to improve electron transfer rates using matrices and different nanomaterials.19 Occasionally, the surface can unexpectedly improve enzyme folding and activity,20 as has been observed for GOx confined to microscale channels.21 However, a consistently successful approach has yet to be discovered. This has been due to the fact that we still have little control over the interface. There are few studies that try to correlate how surface properties affect GOx structure and subsequently how this impacts activity. Often, the robustness of the enzyme is gauged by cyclic voltammetry, which yields only indirect information about activity. Consequently, there are no good general strategies or design rules for surface parameters that optimize Received: December 23, 2011 Revised: February 23, 2012 Published: February 23, 2012 5190

dx.doi.org/10.1021/la2050866 | Langmuir 2012, 28, 5190−5200

Langmuir

Article

irreversible, as GOx displaced from the NP did not refold or regain its activity. These results enable us to devise strategies for conjugating NP and GOx in a way that minimizes the impact of the NP. This information could improve GOx utility and efficiency in its broad range of applications, potentially enhancing the capabilities of glucose sensors, food preservation, biofuel cells, and all the other uses of this versatile enzyme.

GOx structure and activity. Therefore, there is a need to understand how the interface affects protein structure and function of GOx. Such information would enable strategic design of the interface to optimize both structure and function, thus improving GOx activity for biosensors, food preservation, and the numerous other applications where GOx is used. Here we investigate the behavior of GOx on Au NPs, which were chosen as model nanoscale surfaces because they can be synthesized with a high degree of control and have been utilized extensively in biological solutions and environments. Moreover, Au NPs have highly versatile surface chemistry, where the chemistry can be modified systematically. In addition, how proteins and DNA couple to Au NPs both covalently and noncovalently has been studied.22−25 We characterized both the enzymatic activity and secondary structure of GOx on the NPs to allow correlation to surface properties. We systematically changed NP surface chemistry, protein microenvironment, and protein additives to help with folding (Figure 1a). In addition, we examined the properties of GOx after it was displaced from the NP. The NP strongly perturbs both the structure and the activity of the GOx, and the effect of the NP could be either exacerbated or diminished by varying the above parameters, showing that the interface between GOx and the NP is complex. The loss in GOx structure and function was

■

MATERIALS AND METHODS

Materials. Glucose oxidase type X-S (GOx) was purchased from Sigma Aldrich in powder form and stored at −20 °C. The buffer used with the protein contained 10 mM sodium phosphate, adjusted to pH 7.4. The buffer used with salt contained an additional 25 mM NaCl, 100 mM NaCl, 150 mM NaCl, or 100 mM K2SO4. Polyethylene glycol (PEG) in solution was added to the buffer used at concentrations of 0.1% or 1%. Water used throughout this study was purified by a Milli-Q purification system (Millipore). All chemicals were purchased from Sigma Aldrich or Strem Chemicals Inc. Methods. Au NP Synthesis. Citrate-stabilized Au NPs (d = 9.6 ± 2.2 nm) were synthesized according to literature methods.26 Briefly, 80 mL of 1% HAuCl4 was reduced by adding 20 mL of 1% sodium citrate, 1% tannic acid, and 25 mM of sodium carbonate at 60 °C under vigorous stirring. After 10 min the colloidal solution was cooled at room temperature. The average size of the particles was obtained by analysis of TEM images (JEOL 2011, Figure 1) with ImageJ software.27 Concentration of the NP solution was calculated from the peak of absorption spectra at 520 nm and known extinction coefficients. Surface Modification of Au NPs. The surfaces of the NPs were modified with different ligands (negative, neutral, and zwitterionic). Citrate NPs (NP−citrate) were divided in two batches. Small amounts of the negatively charged ligand BPS [bis(p-sulfonatophenyl)phenylphosphine dihydrate dipotassium salt] was added to one batch, and the reaction was mildly stirred overnight. Each batch was divided into three aliquots. The first one was retained. The second was incubated with the neutral polymer mPEG-SH (methoxypolyethylene glycol thiol, MW = 356.5) and the third with L-cysteine, a zwitterion, in a 1:300 (NP:ligand) stoichiometry. mPEG-SH and L-cysteine both contain a thiol group that facilitates a covalent bond to Au on the NP, while the P in BPS forms a dipolar bond with Au. Six different samples were utilized: NP−citrate, NP−citrate−PEG, NP−citrate−Cys, NP− BPS, NP−BPS−PEG, and NP−BPS−Cys. Gel electrophoresis was used to assay the surface chemistry. Agarose gel electrophoresis (2%) was performed in 0.5 × TBE at E = 80 mA/ cm for 60 min. Conjugation of GOx to Au NPs. NP−GOx conjugation was achieved by incubating the GOx with the NPs in solution at room temperature for 1 h. Before incubation, GOx was kept in 10 mM NaP pH 7.4 for 1 h at 4 °C. The incubation ratio was 1:500 NP:GOx, which was chosen because it resulted in significant mobility shifts in gel electrophoresis (Supporting Information, Figure S6). This ratio most likely represents a large excess of protein, so it helps ensure conjugation. NP−GOx was purified by centrifugation at 4 °C. Displacement of GOx from Au NPs. The GOx was displaced from the NPs using overnight incubation in mercaptohexanol (1 mM MCH) at 4 °C, a strategy that has been utilized for quantification of thiolated DNA on NP surfaces.28 MCH has greater affinity to Au NPs than the protein and thus displaces GOx from the surface upon binding via Au−S bonding. Changes in color and aggregation of the NPs occurred, which indicated the exchange of GOx for MCH. After incubation, the sample was centrifuged and the supernatant was collected for further measurements of protein concentration, activity, and secondary structure. Fluorescence Spectroscopy. The fluorescence of GOx tryptophan (Trp) and flavine adenine dinucleotide (FAD) was recorded on a Fluoromax spectrofluorometer. The fluorescence spectra were

Figure 1. Engineering the nanoparticle surface. (a) schematic of the strategies explored. (b) NP surface chemistries utilized. (c) TEM (JEOL 2011, 120 keV) image (size bar of 100 nm) of NPs and histogram (d = 9.6 ± 2.2 nm) by ImageJ.27 5191


Langmuir

Article

Figure 2. Effect of surface chemistry on GOx behavior. (a) Gel electrophoresis of NP and NP−GOx. Lanes: (1) NP−citrate, (2) NP−citrate modified with PEG (NP−citrate−PEG), (3) NP−citrate modified with Cys (NP−citrate−Cys), (4) NP−BPS, (5) NP−BPS modified with PEG, (6) NP−BPS modified with Cys, (7) NP−citrate−GOx, (8) NP−citrate−PEG−GOx, (9) NP−citrate−Cys−GOx, (10) NP−BPS−GOx, (11) NP− BPS−PEG−GOx, (12) NP−BPS−Cys−GOx. GOx fluorescence: (b) Trp (λ340 nm) and FAD (λ467 nm) intensity vs T, (c) Trp λmax vs T, (d) NP coverage, (e) Trp λmax, and (f) FAD/Trp ratio as a function of NP surface chemistry. measurements were carried out at 25 °C on an Aviv model 202 using a 1 mm path length quartz cuvette. The CD signal was recorded between 260 and 190 nm, with a digital integration time of 3−5 s/data point, and reported data were averaged over 15−20 accumulations. [GOx] = (2−7.5) × 10−7 M.

measured with 10 mm path length cell, and the excitation and emission slits were set at 8 nm each. Intrinsic fluorescence spectra were obtained by exciting at 290 nm and the emission spectra was recorded from 310 to 440 nm. For FAD fluorescence, excitation wavelength was 467 nm and the emission was recorded from 490 to 750 nm. Protein coverage and folding on the NPs were quantified by fluorescence. GOx was displaced by MCH from the NPs. The number of GOx per NP was obtained from the ratio between protein Trp fluorescence on the supernatant solution and the initial NP concentration. Protein folding was probed by measuring the shift of the maximum emission peak of Trp fluorescence (λmax) and the changes in the FAD intensity fluorescence. Activity of GOx. The enzymatic activity of GOx was quantified by a standard assay, which is the conversion of 0.05 mM Amplex Red to resorufin by the enzyme peroxidase (0.1 Unit/mL) in the presence of H2O2 (glucose:H2O2 ratio of 1:1). When GOx is active, it converts glucose in the presence of O2 to make gluconic acid and H2O2. The kinetics measurements of wild type, Au NP−conjugated, and MCHdisplaced GOx were carried out at 25 °C in a Thermo-Electron plate reader; data were collected over 30 min in triplicate experiments each time. The assay was performed spectrophotometrically (54 000 cm−1 M−1) and the production of the colorimetric product resorufin was monitored by absorption at 571 nm. [D-glucose] = 0−50 mM and [GOx] = (5 × 10−9)−(5 × 10−8) M. KM and kcat/KM were calculated using Lineweaver−Burk and Eadie−Hofstee diagrams, and only the data that matched were used. Data from different syntheses and experimental runs were collected to calculate the average. All statistical analysis were performed with Statistical Product and Service Solutions (SPSS; v.19.0), using unifactorial analyses of variance. The Levene test was used, and LSD statistics was applied for variables with homogeneity of variance and the Dunnett T3 test for cases of nonhomoscedasticity. CD Spectroscopy. The secondary structure of the GOx both on and off the NP was probed by circular dichroism (CD) spectroscopy. CD

■

RESULTS AND DISCUSSION Engineering the NP Surface Chemistry. GOx is a homodimeric glycoprotein with one tight noncovalently bound FAD cofactor per monomer. The GOx monomer has 583 amino acid residues, 17 α-helices, and 30 β-strand structures and consists of two separate structural domains. Its overall size is ∼6.0 × 5.2 × 3.7 nm. We utilized GOx from Aspergillus niger (UnitProtKB accession number P13006).29 NPs 9.6 nm diameter were synthesized. NP surfaces were modified with different surface coating ligands so that the effect of NP surface chemistry on GOx behavior could be studied (Figure 1b). We started with either NP−citrate as a starting material and then modified it with PEG-SH or cysteine (Cys), which binds to the Au NP by the SH group. Because it presents both NH3+ and COO− groups, it is a zwitterion. Zwitterionic ligands are known to minimize nonspecific adsorption better than PEG and thus have been of interest for antifouling.30 Alternatively, the citrate ligand was replaced with BPS, which is known to better stabilize the NPs in buffer solutions and improve DNA and protein conjugation.28,31 NP−BPS was then modified with PEG-SH and Cys as described above. NPs with different surface chemistries were then conjugated to GOx either during or after the surface modification. Conjugation was achieved by simple incubation for 1 h at a NP:GOx ratio of 1:500. Free GOx was purified by the NP− GOx conjugates by spin centrifugation. 5192


Langmuir

Article

it changes as the protein unfolds, which exposes the Trp to water, a different environment from the hydrophobic pocket it normally occupies. λmax of Trp fluorescence for free GOx increased with temperature in almost a linear fashion from 342 to 355 nm over T = 25−75 °C (Figure 2c), indicating that Trp λmax can be used to monitor GOx unfolding. Different NP surface chemistries resulted in similar λmax values, with the highest being for the NP−PEG, NP−Cys, and NP−BPS−Cys (Figure 2e). Compared to λmax with thermal denaturation for free GOx, the amount of unfolding corresponded to a temperature of ∼58 °C. Unfolding was the greatest for GOx on NP−Cys and NP−PEG and the least for NP−BPS and NP−BPS−PEG. FAD fluorescence intensity (λexc = 467 nm, emission λmax = 525 nm) is also temperature sensitive. FAD fluorescence is quenched by the aromatic residues Tyr515, Trp426, Tyr68, and Trp111 and increases when the protein is unfolded. For extreme denaturation, FAD can become dissociated from the protein and exhibit even higher fluorescence. Thus, the ratio of FAD/Trp fluorescence can be used to evaluate GOx denaturation. The temperature dependence of FAD fluorescence was confirmed (open squares, Figure 2b), which increased sigmoidally in intensity with increasing temperature and had a midpoint of Tm ∼ 50 °C. We compared FAD/Trp ratios of GOx on the different NPs (Figure 2f). For all the NP surface chemistries, the FAD/Trp was higher than for free GOx (normalized to FAD/Trp = 1.0), indicating again that the GOx was unfolded on the NPs. This can be due to FAD intensity increasing and/or Trp fluorescence decreasing with unfolding. PEG and Cys samples had higher ratios (4.41 and 7.03, respectively) than NP−citrate (2.85) and NP−BPS (2.3). Generally, surface chemistries originating from NP−citrate had higher ratios than those originating from NP−BPS. NP−Cys generally had the worst values and thus greatest denaturation, and NP−BPS and NP−BPS−PEG were more similar. Circular dichroism (CD) spectroscopy of the GOx on the NPs also confirmed GOx unfolding. Free GOx has a strongly αhelical structure (28% helix, 18% sheet), which has been observed by others.34−37 The GOx far-UV CD curve for the native state has two minima at 208 and 222 nm, supporting a αhelical structure (black line, Figure 3 a,b).34 However, for NP− GOx, the CD spectra changed dramatically to resemble that of a random coil structure, suggesting that the GOx was strongly denatured on the NP surfaces. Secondary structure differed slightly depending on NP surface chemistry. GOx was the least denatured on NP−BPS. The ligands PEG-SH and Cys did not improve GOx structure but instead contributed to GOx unfolding. NPs by themselves did not exhibit any negative CD signals in the same spectral range and were very similar to the buffer spectra (Supporting Information, Figure S3). Control experiments where free GOx was incubated with MCH and BPS with no NPs present exhibited no changes in GOx secondary structure (Supporting Information, Figure S5). Additionally, it is important to note that CD spectra for NP− protein conjugates are difficult to obtain without significant spectral noise, even with extensive signal averaging. This can be attributed to the fact that the CD signal is due to an absorption difference in the UV, where the NPs absorb strongly. Furthermore, NP−protein conjugates are generally at a much lower concentration than what is typically used for free proteins, which are on the order of 1 mg/mL or greater. This can lead to challenges in obtaining CD spectra of NP−protein conjugates.

Gel electrophoresis showed mobility shifts that varied with surface chemistry. NP−citrate was not stable enough in the buffer to run in the gel (lane 1), which has been observed previously.32 PEG-SH functionalization increased the NP mobility (lane 2). The NP retained a negative charge and was able to migrate, which could be due retained charges from residual citrate or surface charges of the NP or incomplete PEGylation.33 Functionalization with Cys resulted in NPs that did not move from the well, most likely due to decreased stability and charge (lane 3). NP−BPS ran in the positive direction, indicating successful NP stabilization and a net negative charge (lane 4). In addition, the narrow band indicated a uniform charge distribution and thus uniform BPS functionalization. NP−BPS functionalized with PEG-SH resulted in a slight mobility shift to lower mobility, which could be due to an increase in size because of the larger PEG ligand, and/or a decrease in charge, as PEG-SH is neutral (lane 5). Cys functionalization of NP−BPS resulted in no noticeable shift. Most of the shifts in the gel were small, consistent with previous observations.28,33 Gel mobilities of the NPs shifted upon conjugation to GOx. NP−citrate−GOx had a higher mobility relative to NP−citrate, indicating that the negatively charged GOx was bound to the NP, stabilizing it (lane 7). For the other NPs, conjugation to GOx resulted in shifts to lower mobility, most likely due to the increase in size of the NP upon conjugation. NP−PEG decreased in mobility upon GOx conjugation (lane 8), while NP−Cys increased in mobility (lane 9). Conjugation to NP− BPS (lane 10) decreased mobility the most compared to NP− citrate−PEG−GOx (lane 11) and NP−citrate−GOx−Cys (lane 12). The different mobilities could be explained for citrate gold NPs and BPS by the different coverages. For both NP−citrate and NP−BPS, the mobility of the GOx conjugates exhibited the trend of PEG > Cys > NP. Fluorescence spectroscopy was used to quantify both the amount of GOx on the NP surface as well as the GOx denaturation. GOx has two fluorescing species: the aromatic residues (Trp, Tyr, Phe) and the cofactor FAD. Free protein fluorescence showed that Trp intensity at 343 nm was relatively flat until about 60 °C, when GOx melts (Tm = 62 °C), indicating that Trp intensity was suitable for determining coverage for moderately unfolded proteins (Figure 2b, circles). By quantifying Trp fluorescence of GOx displaced from the NP by MCH, the coverage was determined to be 10 GOx/NP for NP−citrate, or ∼1 GOx monomer/25 nm2 (Figure 2d). For NP−citrate modified with PEG-SH, the resulting coverage was lower, 3 GOx/NP, indicating that the PEG-SH inhibited conjugation. Cys modification of NP−citrate also inhibited conjugation, resulting in only 6.5 GOx/NP. NP−BPS resulted in coverages similar to NP−citrate (9 GOx/NP). Likewise, modification with PEG-SH and Cys resulted in lower coverages of ∼3 GOx/NP. Evidence for successful displacement of the GOx from the NP by MCH was the color change of the solution from red to blue, due to aggregation of the NPs because the GOx is no longer on the surface to help stabilize it. However, it is important to note that MCH displacement may not necessarily completely displace all the GOx, so these numbers are lower limits for coverages. Thus, NP surface chemistry was found to influence the efficiency of protein conjugation. On the basis of the Trp fluorescence shifts, the secondary structure of GOx was found to be highly perturbed. Trp fluorescence has been used to assay protein unfolding because 5193


Langmuir

Article

are more labile, forming transient redox-active disulfide bonds that are alternately reduced and oxidized in the course of an enzymatic reaction. Clearly, GOx interaction with the NPs compromises secondary structure and possibly could be exposing internal amino acids to enable binding to the NP.6 To understand how these structural changes impact function, enzymatic activity of GOx on the different NPs was measured by a colorimetric assay. NP−GOx were incubated with its substrate glucose, Amplex Red, and horseradish peroxidase (HRP). The products of GOx are H2O2 and gluconic acid. In the presence of H2O2, HRP converts Amplex Red into resorufin, which is a chromophore that can be detected optically and thus can serve as a measure of GOx activity. Resorufin concentration was monitored by absorption at 571 nm as a function of time for different glucose concentrations in order to obtain Lineweaver−Burk plots, which enabled quantification of KM and kcat/KM (Figure 3c). Kinetic curves were obtained for glucose concentrations ranging from 0.1 × 10−3 to 50 × 10−3 M (Supporting Information, Figure S4). KM (black squares) generally decreased when GOx was conjugated to the NPs. NP−citrate−GOx had KM = 0.019 M while free GOx had KM = 0.029 M, which agreed with previous reports.38,39 This suggested that the binding affinity for the substrate glucose was improved upon immobilization. This could be either due to the binding pocket being sterically more accessible because of interactions of the Cys521 and/or the disulfide with the NP and/or partial unfolding of the GOx. Modifying the NP−citrate with PEG-SH resulted in similar KM values to NP−citrate (KM = 0.020 M), and Cys resulted in a higher value of KM (0.027 M), indicating that this surface chemistry did not affect substrate affinity. Apparently the Cys ligand either hinders the substrate from reaching the binding site40 or affects GOx structure and/or orientation. For NP− BPS−GOx, KM was lower (0.017 M), demonstrating stronger substrate binding. Modifying NP−BPS surface with PEG-SH and Cys increased the KM, indicating that these ligands weaken substrate binding, although KM was still lower than for free GOx. Changing NP surface chemistry varied kcat/KM. Generally, kcat/KM for GOx on all the NPs was in the range of 100−200 M−1 s−1 (red circles, Figure 3c), more than an order of magnitude less than that of free GOx (kcat/KM = 2450 M−1 s−1, after y-axis break), indicating that GOx activity was compromised when bound to a NP. NP−citrate−GOx had kcat/KM = 155 M−1 s−1. Modifying NP−citrate with PEG-SH and Cys decreased kcat/KM. NP−BPS−GOx had approximately the same turnover rate (160 M−1 s−1), and modifying the surface chemistry to PEG-SH increased kcat/KM (186 M−1 s−1). Cys modification increased kcat/KM even more (198 M−1 s−1), showing that surface modification strategies can have some positive impact for NP−BPS. The activity measurements show that the structural changes in the GOx are accompanied by an increase in substrate binding strength but loss in activity upon NP conjugation. These results show that the NP and NP surface chemistry can strongly influence GOx structure and consequently enzymatic activity. Stabilization of GOx by Other Proteins. We probed the effect of proteins additives that could potentially stabilize GOx on the NP surface. This strategy has been used by others successfully to improve folding of immobilized proteins.41,42 We used human serum albumin (HSA), as it is a negatively charged carrier protein found in the blood in great abundance and has been employed to stabilize biomolecules and small

Figure 3. CD spectra of (a) free GOx (black line) GOx on NP−citrate (red solid line), NP−citrate−PEG (red dots/line), and NP−citrate− Cys (red dashed line); (b) free GOx (black line), GOx on NP−BPS (blue solid line), NP−BPS−PEG (blue dots/line), and NP−BPS−Cys (blue dashed line). Effect of surface chemistry on (c) KM (black squares) and kcat/KM (red circles). kcat/KM of free GOx after y-axis break. Data are represented as average value ± SE (n = 4−8). The * represents significant differences (P ≤ 0.05).

On the basis of the GOx structure, and the fact that the denaturation is strongest for NP−Cys, we tried to explain the protein behavior based on the amino acids that could be interacting with the NP and NP ligand. GOx has three Cys, two of which form a disulfide bond (Cys164 and Cys206). The remaining Cys, Cys521, is not a disulfide but on a loop. It is not on the surface of the protein but relatively close to it, so it could potentially bind to the NP, especially if it is accompanied by protein denaturation. In addition, Cys521 is close to the active site His516. Thus, it could be that the thiol links to the NP, forcing the GOx to unfold, similar to what has been observed for yeast cytochrome c (Saccharomyces cerevisiae).33 In these previous studies, when NPs were conjugated to cytochrome c from horse heart (Equus caballus) which lacks a surface cysteine, the protein stayed folded, showing that Cys binding to a NP can be detrimental to folding. The disulfide bond in GOx (Cys164 and Cys206) is quite stable, but it is near the GOx surface and in an area that is basic. The intrachain disulfide bond can contribute to protein folding and stability, but depending on the protein environment, some disulfide bonds 5194


Langmuir

Article

NP−protein conjugates for covalently linked cytochrome c has been shown to improve folding44 due to increased screening between the charged amino acid side chains on the protein and the negatively charged NP ligands, which are thought to perturb protein structure due to electrostatic repulsion/ attraction. Only NP−BPS were tested with NaCl due to their high stability. Coverage of GOx on NP−BPS was found to be dependent on NaCl concentration (Figure 5a, blue triangles).

molecules. We also used lysozyme, which is often chosen as a model for a small, positively charged protein. Lysozyme has also been found to significantly increase (3-fold) the thermal stability of soluble GOx, where apparently the ionic interactions between two proteins minimize the unfolding and nonspecific aggregation.43 Two surface chemistries were chosen for this approach, NP− citrate and NP−BPS, because both their GOx conjugates exhibited lower KM values (Figure 4c). The strategy consisted

Figure 4. Effect of stabilization by other proteins on (a) coverage (blue triangles) and (b) on KM (black squares) and kcat/KM (red circles). kcat/KM of free GOx after y-axis break. Data are represented as average value ± SE (n = 4−8). The * represents significant differences (P ≤ 0.05).

of incubating NPs first with GOx, followed by incubation with HSA or lysozyme. Free protein was purified from the NPs by centrifugation. The incubation ratio 1:10 (NP:lysozyme) was selected because at higher concentration the conjugate precipitated. HSA was incubated at a final concentration of 0.1% HSA. NP−GOx activity measurement (Figure 4b, black squares) determined that KM was higher with both HSA and lysozyme incubation, indicating that these other proteins inhibited substrate binding. In addition, changing NP surface chemistry (citrate vs BPS) did not significantly affect substrate binding, as KM values were approximately the same (KM ∼ 0.025 M). Also, addition of HSA and lysozyme tended to lower kcat/KM, indicating that activity was also compromised in the presence of these two proteins (red circles). However, for NP−citrate− GOx, lysozyme had little effect on kcat/KM. These results show that protein additives negatively affected GOx activity on the NP surface. Manipulating the GOx Microenvironment. We altered the protein microenvironment using salts and small molecules that are known to affect GOx folding and stability. NP−citrate and NP−BPS were chosen for this approach. We explored the use of NaCl, which has been known to improve the tertiary structure and stability of free GOx.39 Cations can induce compaction of negatively charged proteins, significantly reducing their hydrodynamic radius. Additionally, they can affect the surface water structure or interact with charged groups on the amino acid side chains. Also, addition of NaCl to

Figure 5. Effect of protein microenvironment stabilization on activity. Inset: schematic of protein microenvironment modifications used. (a) NaCl for NP−BPS−GOx, (b) PEG in solution on NP−citrate−GOx, (c) PEG in solution on NP−BPS−GOx, (d) glucose on NP−citrate− GOx and NP−BPS−GOx. Coverage (blue tiangles), KM (black squares), and kcat/KM (red circles). Data are represented as an average value ± SE (n = 3−8). The * represents significant differences (P ≤ 0.05).

NP−BPS with no NaCl resulted in ∼9 GOx/NP−BPS. As NaCl was increased, coverage decreased. For 150 mM NaCl, the resulting coverage was only 2 GOx/NP. This trend is opposite to the expected result of higher coverages with higher NaCl due to increased screening of the electrostatic repulsion between the negatively charged NP−BPS and the negatively charged GOx. It could be possible that both electrostatic and 5195


Langmuir

Article

hydrophobic interactions are involved in conjugation.45,46 If hydrophobic interactions play an important role, addition of salt could increase the surface ionic charge, making the NP surface less hydrophobic due to the interactions with the negatively charged BPS, consequently diminishing GOx coverage. NaCl also affected the conjugated GOx activity. KM exhibited a nonmonotonic dependence on NaCl concentration (Figure 5a, black squares). At low NaCl KM increased to ∼0.030 M. For [NaCl] > 25 mM, KM decreased back to the original value for NP−BPS−GOx (KM = 0.017 M). However kcat/KM increased with NaCl, reaching 550 M−1 s−1 for [NaCl] = 150 mM, showing that activity could be improved. This improvement in activity was not observed for the free GOx. In previous reports, addition of NaCl to free GOx was found to actually reduce enzymatic activity, even though it stabilized tertiary and quaternary structure.39 In our case, addition of 150 mM NaCl to GOx did not significantly change KM (0.029 M) or kcat/KM (2410 M−1 s−1, above y-axis break). Addition of 100 mM K2SO4 was also explored, as it has been used previously to stabilize GOx due to increased hydrophobic interactions in the protein.47 Substrate binding was not affected by K2SO4, as KM was approximately the same with or without K2SO4, but kcat/KM increased to ∼520 M−1 s−1 (Figure 5a). Addition of K2SO4 to free GOx causes the KM to decrease from 0.029 to 0.020 M, but kcat/KM does not change. The KM on the NP and free protein in K2SO4 was found to be the same (data not shown). The effect of incubating NP−GOx conjugates (NP−citrate and NP−BPS) with free PEG in solution was explored, as it has been used to stabilize proteins. PEG (MW = 3000 g/mol) and no thiol was introduced to NP−GOx in two different ways (Figure 5, sketch). First, premade NP−citrate−GOx or NP− BPS−GOx conjugates were incubated with free PEG (“after”). The second approach was to incubate NP−citrate or NP−BPS with free PEG and GOx simultaneously (“sim”). For NP− citrate, addition of free PEG increased KM, especially for simultaneous incubations (black squares, Figure 5b). This phenomenon was also observed for NP−BPS but resulted in smaller changes in KM. Presence of free PEG tended to decrease kcat/KM from its value of 150 M−1 s−1, and the lowest kcat /K M values were observed for NP−BPS incubated simultaneously (red circles, Figure 5c). Generally, free PEG is detrimental for NP−GOx activity. Furthermore, the order of incubation matters, where simultaneous incubation results tended to compromise the GOx function more. Finally, we investigated the effect of conjugating GOx to NPs in the presence of glucose, which could potentially stabilize the protein by sitting in the binding pocket. Similar strategies have been used with cofactors for orienting GOx on solid surfaces.16,48 First, GOx was incubated with a high concentration of glucose (500 × excess) and then conjugated to the NP. Following conjugation, free glucose was separated from the NP−GOx conjugates by centrifugation. Incubation with glucose influenced the resulting GOx coverage on the NP. Coverage decreased for NP−citrate−GOx from 10 to 7 GOx/ NP (blue triangles, Figure 5d). For NP−BPS−GOx, coverage increased from 9 to 15 GOx/NP, and for BPS−PEG−GOx, coverage increased slightly from 3 to 4 GOx/NP. Addition of glucose improved GOx activity. KM decreased with glucose incubation, from 0.019 to 0.017 M for NP−citrate and from 0.017 to 0.013 M for NP−BPS (black squares, Figure 5d). This shows that, for these surface chemistries, conjugation in the

presence of glucose could help maintain the substrate binding site. On the other hand, kcat/KM for NP−citrate and NP−BPS− PEG when they are incubated with glucose and NP−BPS experienced a slight decrease (red circles, Figure 5d). This behavior of BPS−PEG−GOx with glucose differed from the other surface chemistries. NP−PEG−BPS−GOx incubated with glucose exhibited an increase in KM (0.025−0.030 M), so the glucose did not help substrate binding. However, kcat/KM increased slightly with glucose incubation, from 186 to 235 M−1 s−1. For this surface chemistry, substrate incubation can help enzymatic activity. Protein after Displacement from the NP Surface. We probed the structure and activity of GOx after it was displaced from the NP surface, as it could yield information about how GOx interacts with the NP and the NP ligands. GOx displacement from the NP was achieved by high concentrations of mercaptohexanol (MCH), which binds to the Au NP via its thiol (Figure 6). Displaced GOx was separated from the NPs by

Figure 6. Activity of GOx before and after displacement from the NP surface: (a) KM on the NP (black dashed) and after displacement by MCH (solid black), (b) kcat/KM on the NP (red dashed) and after displacement (solid red), (c) Trp λmax of GOx on NPs with and without glucose. Data are represented as the average value ± SE (n = 4−8). The * represents significant differences (P ≤ 0.05).

spin centrifugation and then quantified in solution for structure and activity. Trp λmax of GOx after displacement from NPs was 355 nm on the NP−citrate and 352 nm on the NP−BPS, which indicated that the GOx was unfolded on the NP (Figure 2e). For all the NP surface treatments, KM decreased on the order of 40% once GOx was displaced from NP−citrate (Figure 6a). 5196


Langmuir

Article

helicity remained low (6%), and β-strand and random coil proportions remained approximately the same, suggesting that once the GOx is displaced from the NP−citrate, it cannot refold. Similar behavior was observed for NP−BPS−GOx (Figure 7b) and NP−citrate−PEG−GOx (Figure 7c). From the deconvolution (Table 1), the proportions of α-helicity, βstrand, and random coil did not change significantly after MCH displacement. This shows that the NP can have irreversible effects on the secondary structure. Control experiments of free GOx incubated only with BPS and MCH resulted in no change in CD spectra compared to that of the free GOx (Supporting Information, Figure S5). However, for NP−BPS−PEG−GOx, the GOx secondary structure improved after displacement from the NP (Figure 7d). GOx on the NP was strongly denatured, with a random coil spectra signature (gray dots/lines). However, after displacement (black), the CD spectrum exhibited two minima at 208 and 222 nm, indicative of the α-helical structure. Spectral deconvolution showed that α-helicity increased from 8% to 13% and β-strand proportion decreased from 46% to 32%. Still, even though folding improves, it does not fully recover. This also shows that functionalization conditions can make a difference in the residual effect on the protein structure. Both NP surface chemistries result in denaturation of the GOx. However, when released from the NP, apparently the refolding of the protein differs. Others have observed a residual denaturation for bovine serum albumin (BSA) on Au NPs. Brewer et al.50 suggested that the protein spreads on the NP− ligand surface, resulting in denaturation, and when displaced from the NP surface, some of the NP ligand comes off with the protein. This ligand can remain complexed with the protein and maintain its denatured state. We also detected the presence of BPS in the solution after displacement by fluorescence (λem = 585 nm, data not shown). Due to the nature of PEG-SH and citrate, their presence could not be probed by fluorescence. Presumably, either PEG is not released or when it is released from the NP it does not bind to the GOx in a way that maintains its denatured state. Glucose incubation also made a marked difference in the GOx secondary structure after displacement. NP−GOx conjugates that were created with glucose present also showed recovery of secondary structure after displacement from the NPs. For NP−citrate−GOx incubated with glucose (Figure 7a, black dots/lines), the displaced GOx exhibited minima at 208 and 222 nm, a signature of α-helicity. Compared to incubations without glucose (black line), GOx was able to retain much of its secondary structure. Deconvolution (Table 1) showed that the α-helicity was 28% with glucose, close to that of the free GOx (32%) and significantly higher than for NP−citrate−GOx without glucose (6%). Similar behavior was observed for NP− BPS−GOx incubated with glucose (Figure 7b, black dots/ lines), exhibiting 28% α-helicity with glucose and 9% without. Evidently, glucose allows GOx to maintain its structure during conjugation, which is retained after displacement from the NP. It is possible that when the GOx is bound to glucose, it is more stable and does not permit the disulfide and/or Cys521 to interact with the NP. In addition, glucose positively impacted activity both before and after displacement from the NPs. KM of GOx on NP−citrate and NP−BPS with glucose incubations was lower relative to KM without glucose both before and after displacement (Figure 6a). Also, kcat/KM of the displaced protein was higher than for GOx that did not have glucose incubation during the conjugation (Figure 6b). However, glucose

This indicates that the binding site was blocked when the GOx was on the NP surface. Chemical displacement by MCH most likely allows the active site to be exposed again, thus increasing substrate binding strength. This suggests that the GOx most likely binds with the substrate binding pocket facing the NP, which is possibly aided by interactions between the Cys in the disulfide bond and Cys521. However, kcat/KM also decreased after displacement for all the NP surface chemistries (Figure 6b). Values went from 150 M−1 s−1 for GOx on NP−citrate (hatched red) to 100 M−1 s−1 when displaced from the NP by MCH (solid red). NP−BPS with glucose incubation also resulted in similar decreases in kcat/KM (∼50%). This suggests that even though the NP causes a great amount of denaturation and loss of activity, removing GOx from the NP makes it worse. In addition, because the KM and kcat/KM values are not restored to their original values for the free GOx, this indicates that the NP has strongly and irreversibly destabilized the protein. Removing GOx from the NP surface results in even more denaturation, indicating that when the GOx interacts with the NP−citrate surface, there is some stabilization involved. Similar decreases in KM and kcat/KM were observed for NP−BPS−GOx (Figure 6b). In addition, this denaturation is not a consequence of the MCH. Control experiments probing the effect of MCH on the activity of free GOx showed that the activity was not significantly different from that of free GOx with no MCH, where kcat/KM decreased from 2450 to 2220 M−1 s−1 and KM increased from 0.029 to 0.037 M (Supporting Information, Table S4). The secondary structure of GOx after MCH displacement from the NP was probed. Comparing GOx on NP−citrate (gray dots/lines) vs after displacement (black line), the minimum in the CD spectrum shifted to shorter wavelengths (Figure 7a). In addition, the signal at 222 nm became more positive. Deconvolution49 (Table 1) showed that the α-helical contribution decreased from 32% to 8% upon conjugation to NP−citrate, suggesting unfolding. After MCH displacement, α-

Figure 7. CD of GOx before and after displacement from the NP surface: (a) on NP−citrate (gray dot/line) after displacement (solid black line) and glucose incubated GOx after displacement (black dot/ line), (b) on NP−BPS (gray dot/line) after displacement (solid black line) and glucose incubated GOx after displacement (black dots/line), (c) on NP−citrate−PEG (gray dots/line) and after displacement (solid black line), and (d) on NP−BPS−PEG (gray dot/line) and after displacement (solid black line). 5197


Langmuir

Article

Table 1. CD Deconvolution Results Using K2d49 GOx helix (%) strand (%) random coil (%) Re distance MAX error helix (%) strand (%) random coil (%) Re distance MAX error

32 18 50 124 >0.227 GOx 32 18 50 124 >0.227

NP− citrate

NP−citrate MCH

NP−citrate−glucose MCH

GOx

8 6 41 36 50 58 78 74 0.18 0.18 NP−citrate−PEG

28 28 44 107 >0.227 NP−citrate−PEG MCH

helix (%) strand (%) random coil (%) Re distance MAX error

7 50 43 49 0.10

9 35 56 31 0.08

helix (%) strand (%) random coil (%) Re distance MAX error

incubation did not blue shift λmax, indicating that denaturation was still present (Figure 6c). Nevertheless, glucose incubation had net a positive effect on the protein after displacement. These results show that for certain surface chemistries, the NP irreversibly affects the protein denaturation and activity. In addition, the degree of recovery of protein structure/activity depends on NP surface chemistry and protein microenvironment.

32 18 50 124 >0.227 GOx

NP− BPS

NP−BPS MCH

NP−BPS−glucose MCH

9 9 36 35 55 56 35 94 0.08 0.18 NP−BPS−PEG

28 28 44 96 0.18 NP−BPS−PEG MCH

8 46 46 74 0.18

13 32 54 50 0.10

32 18 50 124 >0.227

stabilize folding but did not result in large improvements in activity. Also, charge stabilization (NaCl) improved activity but did not greatly improve secondary structure. This means that when strategizing immobilization, one needs to probe both structure and activity of the enzyme, as they may not always be correlated. These strategies can be useful for applications that require tethering the GOx and other enzymes to a surface.

■

■

ASSOCIATED CONTENT

* Supporting Information

CONCLUSIONS In summary, interactions with NPs can significantly alter the structure and function of GOx. The NP is generally a strongly perturbing agent for the GOx, and interactions between the protein and the NP and NP ligand result in irreversible denaturation. In addition, the NP can improve substrate binding while reducing catalytic turnover. This indicates that steric interactions are not as much of an issue as is denaturation. This is based on the fact that the KM values for NP−GOx were lower than for free GOx, and that secondary structure in CD is often greatly compromised by the NP. It is likely that the three Cys in GOx have an affinity for the NP surface, and formation of the thiol−Au bonds either cause or accompany GOx unfolding. Choice of surface chemistry can influence GOx coverage, structure, and activity. Neutral and zwitterionic ligands are generally not helpful, where they act more as a steric barrier for the active site and simultaneously reduce the coverage of the GOx on the NP. The order of chemical modifications also influences behavior, though to a lesser extent. Modifying the protein microenvironment by addition of salt and the substrate glucose can help improve activity and folding. Salt screening was found to be successful for improving activity. Incubation with its substrate glucose helps stabilize secondary structure, indicating that preserving the active site is likely to be occurring. Protein additives did not help as expected, despite their success for 2D surfaces. Apparently, steric inhibition becomes significant, and protein activity is not improved. Displacement of GOx from the NP yielded information about the interface between the GOx and NP. Because the KM decreased after displacement from the NP, the interface most likely involves interactions with the GOx substrate binding pocket. Because in some cases the GOx was irreversibly denatured and unable to recover its structure and activity after displacement, these interactions at the NP−GOx interface significantly perturb GOx structure and probably the active site. However, stabilization of secondary structure does not necessarily mean preservation of activity. Glucose was able to

S

Temperature-dependent fluorescence of free GOx, CD spectra of NPs and free GOx, kinetic data used to obtain KM and kcat/ KM, and statistical analyses. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Barbara Imperiali and Deborah Pheasant at the MIT CSBi Bioinstrumentation Facility (BIF). We thank Narenda Maheshri for instrumental use. E.T. was supported by MISTI, the project SABioD (Department of Innovation, Government of Navarra), and the grant Jeronimo de Ayanz (Department of Education, Government of Navarra). K.J.W. was supported by the MIT Summer Research Program.

■

REFERENCES

(1) Wilson, R.; Turner, A. P. F. Glucose oxidase: An ideal enzyme. Biosens. Bioelectron. 1992, 7 (3), 165−185. (2) Newman, J. D.; Turner, A. P. F. Home blood glucose biosensors: A commercial perspective. Biosens. Bioelectron. 2005, 20 (12), 2435− 2453. (3) Miranda, O. R.; Creran, B.; Rotello, V. M. Array-based sensing with nanoparticles: “Chemical noses” for sensing biomolecules and cell surfaces. Curr. Opin. Chem. BIol. 2010, 14 (6), 728−736. (4) Miranda, O. R.; Chen, H.-T.; You, C.-C.; Mortenson, D. E.; Yang, X.-C.; Bunz, U. H. F.; Rotello, V. M. Enzyme-amplified array sensing of proteins in solution and in biofluids. J. Am. Chem. Soc. 2010, 132 (14), 5285−5289. (5) Cooney, M. J.; Svoboda, V.; Lau, C.; Martin, G.; Minteer, S. D. Enzyme catalysed biofuel cells. Energy Environ. Sci. 2008, 1 (3), 320− 337. (6) Otsuka, I.; Yaoita, M.; Higano, M.; Nagashima, S.; Kataoka, R. Tapping mode AFM study on the surface dynamics of a single glucose

5198


Langmuir

Article

oxidase molecule on a Au(111) surface in water with implication for a surface-induced unfolding pathway. Appl. Surf. Sci. 2004, 235 (1−2), 188−196. (7) Ivnitski, D.; Branch, B.; Atanassov, P.; Apblett, C. Glucose oxidase anode for biofuel cell based on direct electron transfer. Electrochem. Commun. 2006, 8 (8), 1204−1210. (8) Rauf, S.; Ihsan, A.; Akhtar, K.; Ghauri, M. A.; Rahman, M.; Anwar, M. A.; Khalid, A. M. Glucose oxidase immobilization on a novel cellulose acetate−polymethylmethacrylate membrane. J. Biotechnol. 2006, 121 (3), 351−360. (9) Bankar, S. B.; Bule, M. V.; Singhal, R. S.; Ananthanarayan, L. Glucose oxidaseAn overview. Biotechnol. Adv. 2009, 27 (4), 489− 501. (10) Ren, X.; Meng, X.; Chen, D.; Tang, F.; Jiao, J. Using silver nanoparticle to enhance current response of biosensor. Biosens. Bioelectron. 2005, 21 (3), 433−437. (11) Lynch, I.; Cedervall, T.; Lundqvist, M.; Cabaleiro-Lago, C.; Linse, S.; Dawson, K. A. The nanoparticle−protein complex as a biological entity; a complex fluids and surface science challenge for the 21st century. Adv. Colloid Interface Sci. 2007, 134−135, 167−174. (12) Nel, A. E.; Madler, L.; Velegol, D.; Xia, T.; Hoek, E. M. V.; Somasundaran, P.; Klaessig, F.; Castranova, V.; Thompson, M. Understanding biophysicochemical interactions at the nano−bio interface. Nat. Mater. 2009, 8, 543−557. (13) Gagner, J. E.; Lopez, M. D.; Dordick, J. S.; Siegel, R. W. Effect of gold nanoparticle morphology on adsorbed protein structure and function. Biomater. 2011, 32 (29), 7241−7252. (14) Bickerstaff, G. F. Characterization of Enzyme Activity, Protein Content, and Thiol Groups in Immobilized Enzymes. In Immobilization of Enzymes and Cells; Humana Press: Totowa, NJ, 1996; Vol. 1, pp 1−11. (15) Keighron, J. D.; Keating, C. D. Enzyme:nanoparticle bioconjugates with two sequential enzymes: Stoichiometry and activity of malate dehydrogenase and citrate synthase on Au nanoparticles. Langmuir 2010, 26 (24), 18992−19000. (16) Xiao, Y.; Patolsky, F.; Katz, E.; F.Hainfeld, J.; Willner, I. Plugging into enzymes: Nanowiring of redox enzymes by a gold nanoparticle. Science 2003, 299, 1877−1881. (17) Vertegel, A. A.; Siegel, R. W.; Dordick, J. S. Silica nanoparticle size influences the structure and enzymatic activity of adsorbed lysozyme. Langmuir 2004, 20, 6800−6807. (18) Aubin-Tam, M.-E.; Hamad-Schifferli, K. Structure and function of nanoparticle−protein conjugates. Biomed. Mater. 2008, 3, 034001. (19) Shan, C.; Yang, H.; Song, J.; Han, D.; Ivaska, A.; Niu, L. Direct electrochemistry of glucose oxidase and biosensing for glucose based on graphene. Anal. Chem. 2009, 81 (6), 2378−2382. (20) Asuri, P.; Karajanagi, S. S.; Vertegel, A. A.; Dordick, J. S.; Kane, R. S. Enhanced stability of enzymes adsorbed onto nanoparticles. J. Nanosci. Nanotechnol. 2007, 7 (4−5), 1675−1678. (21) Wang, G.; Yau, S.-T. Spatial confinement induced enzyme stability for bioelectronic applications. J. Phys. Chem. C 2007, 111 (32), 11921−11926. (22) Nelson, E. M.; Rothberg, L. J. Kinetics and mechanism of singlestranded DNA adsorption onto citrate-stabilized gold nanoparticles in colloidal solution. Langmuir 2011, 27 (5), 1770−1777. (23) Tsai, D.-H.; DelRio, F. W.; Keene, A. M.; Tyner, K. M.; MacCuspie, R. I.; Cho, T. J.; Zachariah, M. R.; Hackley, V. A. Adsorption and conformation of serum albumin protein on gold nanoparticles investigated using dimensional measurements and in situ spectroscopic methods. Langmuir 2011, 27 (6), 2464−2477. (24) Chirra, H. D.; Sexton, T.; Biswal, D.; Hersh, L. B.; Hilt, J. Z. Catalase-coupled gold nanoparticles: Comparison between the carbodiimide and biotin−streptavidin methods. Acta Biomater. 2011, 7 (7), 2865−2872. (25) de Puig, H.; Federici, S.; Baxamusa, S. H.; Bergese, P.; HamadSchifferli, K. Quantifying the nanomachinery of the nanoparticle− biomolecule interface. Small 2011, 7 (17), 2477−2484. (26) Beesley, J. E. Colloidal Gold: A New Perspective for Cytochemical Marking; Oxford University Press: Oxford, 1989.

(27) Abramoff, M. D.; Magelhaes, P. J.; Ram, S. J. Image processing with ImageJ. Biophotonics Int. 2004, 11 (7), 36−42. (28) Park, S.; Brown, K. A.; Hamad-Schifferli, K. Changes in oligonucleotide conformation on nanoparticle surfaces by modification with mercaptohexanol. Nano Lett. 2004, 4 (10), 1925−1929. (29) Hecht, H. J.; Kalisz, H. M.; Hendle, J.; Schmid, R. D.; Schomburg, D. Crystal structure of glucose oxidase from Aspergillus niger refined at 2.3 Å resolution. J. Mol. Biol. 1993, 229 (1), 153−172. (30) Estephan, Z. G.; Schlenoff, P. S.; Schlenoff, J. B. Zwitteration as an alternative to PEGylation. Langmuir 2011, 27 (11), 6794−6800. (31) Zanchet, D.; Micheel, C. M.; Parak, W. J.; Gerion, D.; Alivisatos, A. P. Electrophoretic isolation of discrete Au nanocrystal/DNA conjugates. Nano Lett. 2001, 1 (1), 32−35. (32) Reed, A. M. W.; Metallo, S. J. Oriented protein adsorption to gold nanoparticles through a genetically encodable binding motif. Langmuir 2010, 26 (24), 18945−18950. (33) Aubin-Tam, M.-E.; Hamad-Schifferli, K. Gold nanoparticle− cytochrome c complexes: The effect of nanoparticle ligand charge on protein structure. Langmuir 2005, 21 (26), 12080−12084. (34) Hecht, H. J.; Schomburg, D.; Kalisz, H.; Schmid, R. D. The 3D structure of glucose oxidase from Aspergillus niger. Implications for the use of GOD as a biosensor enzyme. Biosens. Bioelectron. 1993, 8 (3−4), 197−203. (35) Khatun Haq, S.; Faiz Ahmad, M.; Hasan Khan, R. The acidinduced state of glucose oxidase exists as a compact folded intermediate. Biochem. Biophys. Res. Commun. 2003, 303 (2), 685−692. (36) Zhao, H.-Z.; Sun, J.-J.; Song, J.; Yang, Q.-Z. Direct electron transfer and conformational change of glucose oxidase on carbon nanotube-based electrodes. Carbon 2010, 48 (5), 1508−1514. (37) Caves, M. S.; Derham, B. K.; Jezek, J.; Freedman, R. B. The mechanism of inactivation of glucose oxidase from Penicillium amagasakiense under ambient storage conditions. Enzyme Microb. Technol. 2011, 49 (1), 79−87. (38) Swoboda, B. E. P.; Massey, V. Purification and properties of the glucose oxidase from Aspergillus niger. J. Biol. Chem. 1965, 240 (5), 2209−2215. (39) Ahmad, A.; Akhtar, M. S.; Bhakuni, V. Monovalent cationinduced conformational change in glucose oxidase leading to stabilization of the enzyme. Biochemistry 2001, 40 (7), 1945−1955. (40) Estephan, Z. G.; Jaber, J. A.; Schlenoff, J. B. Zwitterion-stabilized silica nanoparticles: Toward nonstick nano. Langmuir 2010, 26 (22), 16884−16889. (41) Gouda, M. D.; Kumar, M. A.; Thakur, M. S.; Karanth, N. G. Enhancement of operational stability of an enzyme biosensor for glucose and sucrose using protein based stabilizing agents. Biosens. Bioelectron. 2002, 17 (6−7), 503−507. (42) Tyhach, R.; Rupchock, P.; Pendergrass, J.; Skjold, A.; Smith, P.; Johnson, R.; Albarella, J.; Profitt, J. Adaptation of prostatic-group-label homogeneous immunoassay to reagent-strip format. Clin. Chem. 1981, 27 (9), 1499−1504. (43) Gouda, M. D.; Thakur, M. S.; Karanth, N. G. stability studies on immobilized glucose oxidase using an amperometric biosensorEffect of protein based stabilizing agents. Electroanalysis 2001, 13 (10), 849− 855. (44) Aubin-Tam, M.-E.; Hwang, W.; Hamad-Schifferli, K. Sitedirected nanoparticle labeling of cytochrome c. Proc. Natl. Acad. Sci. U. S. A. 2009, 106 (11), 4095−4100. (45) Lindman, S.; Lynch, I.; Thulin, E.; Nilsson, H.; Dawson, K. A.; Linse, S. Systematic investigation of the thermodynamics of HSA adsorption to N-iso-propylacrylamide/N-tert-butylacrylamide copolymer nanoparticles. Effects of particle size and hydrophobicity. Nano Lett. 2007, 7 (4), 914−920. (46) Monopoli, M. P.; Walczyk, D.; Campbell, A.; Elia, G.; Lynch, I.; Baldelli Bombelli, F.; Dawson, K. A. Physical−chemical aspects of protein corona: Relevance to in vitro and in vivo biological impacts of nanoparticles. J. Am. Chem. Soc. 2011, 133 (8), 2525−2534. (47) Gouda, M. D.; Singh, S. A.; Rao, A. G. A.; Thakur, M. S.; Karanth, N. G. Thermal inactivation of glucose oxidase: Mechanism 5199


Langmuir

Article

and stabilization using additives. J. Biol. Chem. 2003, 278 (27), 24324− 24333. (48) Srere, P. A.; Uyeda, K.; Klaus, M. Functional groups on enzymes suitable for binding to matrices. In Method in Enzymology; Academic Press: New York, 1976; Vol. 44, pp 11−19. (49) Andrade, M. A.; Chacón, P.; Merelo, J. J.; Morán, F. Evaluation of secondary structure of proteins from UV circular dichroism spectra using an unsupervised learning neural network. Protein Eng. 1993, 6 (4), 383−390. (50) Brewer, S. H.; Glomm, W. R.; Johnson, M. C.; Knag, M. K.; Franzen, S. Probing BSA binding to citrate-coated gold nanoparticles and surfaces. Langmuir 2005, 21 (20), 9303−9307.

5200


Engineering the Interface between Glucose Oxidase and

Recommend Documents