Article pubs.acs.org/Langmuir
pH Effect on Protein G Orientation on Gold Surfaces and Characterization of Adsorption Thermodynamics Blake N. Johnson and Raj Mutharasan* Department of Chemical and Biological Engineering, Drexel University, Philadelphia, Pennsylvania 19104, United States S Supporting Information *
ABSTRACT: The pH effect on adsorbed antibody-binding protein (protein G) orientation on gold (Au) and its adsorption thermodynamic characteristics were investigated using quartz crystal microbalance (QCM) and X-ray photoelectron spectroscopy (XPS). The adsorbed protein G orientation was measured by binding response of two antibody−antigen systems: the model bovine serum albumin (BSA) and the foodborne pathogen E. coli O157:H7. Surface coverage was not significantly affected by pH, but its orientation was. The most properly oriented protein G for antibody binding was achieved at near-neutral pH. Adsorption was verified by XPS measurements using nitrogen (N) 1s, oxygen (O) 1s, and Au 4p peak heights. Adsorption energetics were determined by van’t Hoff and Langmuir kinetic analyses of adsorption data obtained at 296, 303, and 308 K. Large characteristic entropy change of protein adsorption was observed (ΔS° = 0.52 ± 0.01 kcal/mol·K). The adsorption process was not classical physisorption but exhibited chemisorption characteristics based on significant enthalpy change (ΔH° = −25 ± 6 kcal/mol).
■
INTRODUCTION Orienting immobilized proteins is critical in various scientific and measurement applications.1,2 Although many protein− substrate combinations have been investigated, the antibody binding protein (protein G or protein A)−gold (Au) system is especially important in antibody-based immunoassays and protein−protein interaction studies.2−5 However, limited fundamental information is available on protein G immobilization on Au surfaces and on methodology that enhances a proper orientation. Further, the thermodynamic characteristics of protein G adsorption on Au have not yet been reported. To shed further light on these important issues, we measured thermodynamic properties of the model protein G−Au protein−substrate system. Also, in an attempt to optimize immobilized protein G orientation for maximizing antibodybinding capacity, we find that solution pH is a critical parameter, as it affects orientation. In general, proteins are immobilized on surfaces by various techniques, such as adsorption and covalent bonding.1,6−8 When surfaces have accessible functional groups, a protein may be directly immobilized via reactions to form covalent linkage.9−12 Thus, various techniques for grafting surface functional groups have been investigated.11,13−18 Although immobilization via covalent bonds is desirable, adsorption is extensively used in practical applications,19−22 especially on Au and polystyrene surfaces, due to simplicity in implementation and reasonable binding affinity. In such cases, however, the orientation may vary. Therefore, several approaches have been investigated for facilitating controlled orientation, such as © 2012 American Chemical Society
incorporating non-native amino acids and functional groups that have high surface affinity.5,21,22 Such approaches, however, modify the recognition molecule which can potentially affect selectivity. Previous investigations have shown that ionic strength, pH, cosolvents, and surface topography affect protein adsorption,18,20,23−25 and thus, it is logical to examine these parameters for controlling orientation and surface coverage density. In this study, we examine the effects of pH on protein G orientation and surface coverage as well as characterize the thermodynamics of its adsorption to Au.
■
MATERIALS AND METHODS
Reagents. All QCM studies were carried out in 10 mM phosphate buffered saline (pH 7.2 PBS, Sigma Aldrich, ionic strength 141 mM) except where noted. Acetate buffers (pH 5.0 and 3.0) were prepared from acetic acid and sodium acetate (Fisher). Sodium chloride (NaCl, Sigma-Aldrich) was used for adjusting acetate buffer ionic strength. Deionized (DI) water (18 MΩ, Milli-Q, Millipore, Billerica, MA) was used in all sample preparation and rinsing procedures. Lyophilized recombinant protein G (proG, Pierce) from group G streptococci (molecular weight (MW) 30 kDa,26 average diameter 4 nm) was used in all experiments. Goat polyclonal anti-E. coli O157:H7 IgG and samples of inactivated E. coli O157:H7 (KPL) were used in all E. coli sensing experiments. Rabbit polyclonal anti-bovine serum albumin IgG (anti-BSA, Fisher) and BSA (Sigma-Aldrich) were used for all BSA sensing experiments. Biological reagents were reconstituted in 10 mM PBS and 0.01% w/w sodium azide (Sigma-Aldrich), stored at 251 K, Received: March 2, 2012 Revised: April 5, 2012 Published: April 12, 2012 6928
dx.doi.org/10.1021/la3009128 | Langmuir 2012, 28, 6928−6934
Langmuir
Article
and removed only immediately prior to use. DI water, 95% ethanol (EtOH, Fisher), and 0.1% v/v Tween 80 (polyoxyethylene (20) sorbitan mono-oleate, Fisher) solution in DI water were used in all rinsing procedures. Quartz Crystal Microbalance (QCM). A 5 MHz QCM installed in a flow cell (Stanford Research Systems-SRS, model QCM200QCM25) with a 1-in.-diameter quartz crystal was used (1.2-cmdiameter circular gold sensing area (DAu)). The vendor-provided crystal sensitivity factor (Cf) was 56.5 Hz·cm2/μg at room temperature. Prior to all immobilization experiments, the QCM’s gold surface, flow cell, and reagent reservoirs were cleaned. The Au surface was cleaned in oxygen plasma for five minutes (Harrick Plasma, Ithica, NY, Plasma Cleaner model PDC-001, 200 W, RF level set to high position during plasma period), followed by a two minute rinse with fresh piranha solution, and a final rinse with copious amounts of ethanol, Tween surfactant solution, and DI water. (CAUTION: Piranha solution bubbles vigorously upon preparation and should be handled with care.) The cleaning protocol gave concurrent starting values in resonant frequency ( f) and repeatability in measured binding responses. Flow rate (Q) (100−150 μL/min) was maintained throughout all binding experiments using a calibrated syringe pump. A typical binding experiment was done by switching between buffer and analyte reservoirs maintained at constant temperature and concentration throughout the entire experiment. Stock proG, anti-E. coli O157:H7, E. coli O157:H7 antigen, anti-BSA, and BSA binding concentrations used were 50 μg/mL, 10 μg/mL, 7E07 cells/mL, 10 μg/mL, and 50 μg/mL, respectively. Stock solution sample volumes of 800 μL of proG, 500 μL of antibody, and 50 μL of antigen were all diluted in 5.4 mL buffer to achieve the desired final concentration. X-ray Photoelectron Spectroscopy (XPS). A Phi VersaProbe 5000 XPS system (Physical Electronics, Inc., Chanhassen, MN, USA) with a monochromatic aluminum Kα radiation source (26W) at pass energy of 117.5 eV (45° takeoff angle) was used for characterizing surfaces. The binding energy was calibrated using the Au 4f peak as a reference at 84.0 eV. A scanning beam diameter of 100 μm was used in all measurements. Gold substrates were immersed in proG for the same time and concentration as was used in QCM binding experiments. XPS spectra were obtained immediately after a thorough rinsing step followed by blow drying with a gentle dry nitrogen stream.
■
RESULTS AND DISCUSSION Typical Binding Experiment. A typical binding experiment involved multiple sequential steps in which the flow to the QCM was changed among buffer, proG, antibody, and antigen solutions in a continuous fashion without stopping the flow. As shown in Figure 1A, the first step is stabilization of the freshly mounted and clean QCM in buffer (PBS or acetate) followed by a change to a proG solution prepared in the same buffer. After two minutes (time required for the analyte to reach the QCM), the QCM resonant frequency decreased exponentially over 30−45 min period due to crystal masschange caused by proG adsorption. After equilibrium was reached, the flow was returned to PBS for 10 min for removal of loosely bound species and to measure possible density or viscosity effects. Introduction of antibody solution caused a similar exponential decrease response due to binding between the adsorbed proG and the antibody Fc-domain, an interaction that has been extensively studied.27−29 After the resonant frequency stabilized due to saturation, the flow was switched to the antigen sample. Again, the resonant frequency decreased due to increased added-mass caused by antigen binding. Finally, after the binding equilibrium was reached, the flow was returned to PBS. Adsorbed Protein G Saturates the Au Surface. Before investigating the pH effects, we first examined the fractional coverage achieved using the 6.45 μg/mL proG solution. As shown in Figure 1B, proG adsorption in PBS caused a resonant
Figure 1. (A) Typical binding experiment shows proG binding response to a bare Au surface (1), a PBS rinse step (2), the antibody binding response to the adsorbed proG (3), the antigen binding response to the immobilized antibody (4), and a final PBS rinse (5). Included is a negative control in which no binding analyte was used, which shows that the resonant frequency remains constant (standard deviation
