Antibody Binding to a Tethered Vesicle Assembly Using QCM-D

Quartz Crystal Microbalance Analysis of DNA-Templated Calcium Phosphate Mineralization. Sunaro C. Ngourn , Hillary A. Butts , Alexander R. Petty , Jam...
0 downloads 0 Views 2MB Size
Download PDF
Anal. Chem. 2009, 81, 6021–6029

Antibody Binding to a Tethered Vesicle Assembly Using QCM-D Ankit R. Patel,† Kay K. Kanazawa, and Curtis W. Frank* Department of Chemical Engineering, Stanford University, 381 North-South Mall, Stauffer III, Stanford, California 94305-5025 The bilayer-tethered vesicle assembly has recently been proposed as a biomimetic model membrane platform for the analysis of integral membrane proteins. Here, we explore the binding of antibodies to membrane components of the vesicle assembly through the use of quartz crystal microbalance with dissipation monitoring (QCMD). The technique provides a quantitative, label-free avenue to study binding processes at membrane surfaces. However, converting the signal generated upon binding to the actual amount of antibody bound has been a challenge for a viscoelastic system such as the tethered vesicle assembly. In this work, we first established an empirical relationship between the amount of bound antibody and the corresponding QCM-D response. Then, the results were examined in the context of an existing model describing the QCM-D response under a variety of theoretical loading conditions. As a model system, we investigated the binding of monoclonal antidinitrophenyl (DNP) IgG1 to tethered vesicles displaying DNP hapten groups. The measured frequency and dissipation responses upon binding were compared to an independent measure of the amount of bound antibody obtained through the use of an in situ ELISA assay. At saturation, the surface mass density of bound antibody was approximately 900 ng/cm2. Further, through the application of QCM-D models that describe the response of the quartz when loaded by either a single homogeneous viscoelastic film or by a two-layered viscoelastic film, we found that a homogeneous, onelayer model accurately predicts the amount of antibody bound to the tethered vesicles near antibody surface saturation, but a two-layer model must be invoked to accurately describe the kinetic response of the dissipation factor, which suggests that the binding of the antibody results in a stiffening of the top layer of the film. Antibody and ligand binding to membrane protein receptors or antigens displayed on the surface of cell membranes are important interactions found in a variety of biological processes, including activation of various cell signaling pathways as well as initiation of the immune response. Both receptor-ligand interac* Author to whom correspondence is addressed. Fax: (650)723-9780. E-mail: [email protected]. † Present address: Late Stage Pharmaceutical and Processing Development, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080. 10.1021/ac802756v CCC: $40.75  2009 American Chemical Society Published on Web 07/06/2009

tions and antibody-antigen interactions have traditionally been studied using labeled ligands in conjunction with cell-based assays or surface-immobilized components. Some time ago, biomimetic model membranes in the format of solid-supported lipid bilayers were proposed as simple, robust systems to study these interactions.1 Such a system confers the benefits of working in vitro while maintaining the lateral fluidity of membrane components, an important requirement for the study of multivalent ligand binding and cell signaling. Lipid-linked antigens and ligands as well as membrane-associated proteins can readily be studied using the solid-supported lipid bilayer system. In spite of their attractive features, the use of solid-supported lipid bilayers poses significant challenges to the study of integral, membrane-spanning receptors and coreceptors due to the close proximity of the underlying substrate. These include the loss of lateral mobility as well as receptor activity through unfavorable interactions between the substrate and protein domains extending out from the plane of the bilayer.2-4 To address this challenge, the tethered vesicle assembly, consisting of lipid vesicles tethered to a solid supported lipid bilayer,2,5-10 has recently been proposed as a strategy to overcome the protein-substrate interactions. The surface bound vesicles are intended to house integral membrane proteins of interest without interference from the surface thereby serving as a suitable platform for studies of binding to membrane constituents. Tethering to a lipid bilayer allows tethered vesicles to move across the surface of the substrate for lab-on-a-chip applications, and the lipid bilayer serves to prevent nonspecific adsorption of receptor ligands. In addition, the assembly’s location near a surface allows the use of powerful, surface-sensitive techniques, including quartz crystal microbalance with dissipation monitoring (QCM-D), to probe binding and adsorption to the model membrane. Quartz crystal microbalance with dissipation monitoring allows real-time, label-free, quantitative analysis of surface adsorption and (1) Tamm, L. K.; McConnell, H. M. Biophys. J. 1985, 47, 105–113. (2) Boukobza, E.; Sonnenfeld, A.; Haran, G. J. Phys. Chem. B 2001, 105, 12165– 12170. (3) Tanaka, M.; Sackmann, E. Nature 2005, 437, 656–663. (4) Wagner, M. L.; Tamm, L. K. Biophys. J. 2000, 79, 1400–1414. (5) Patel, A. R.; Frank, C. W. Langmuir 2006, 22, 7587–7599. (6) Yoshina-Ishii, C.; Boxer, S. G. J. Am. Chem. Soc. 2003, 125, 3696–3697. (7) Yoshina-Ishii, C.; Boxer, S. G. Langmuir 2006, 22, 2384–2391. (8) Yoshina-Ishii, C.; Chan, Y. H. M.; Johnson, J. M.; Kung, L. A.; Lenz, P.; Boxer, S. G. Langmuir 2006, 22, 5682–5689. (9) Yoshina-Ishii, C.; Miller, G. P.; Kraft, M. L.; Kool, E. T.; Boxer, S. G. J. Am. Chem. Soc. 2005, 127, 1356–1357. (10) Yoshina-Ishii, C.; Miller, G. P.; Kraft, M. L.; Kool, E. T.; Boxer, S. G. Biophys. J. 2005, 88, 233A–233A.

Analytical Chemistry, Vol. 81, No. 15, August 1, 2009

6021

Figure 1. An overview of the tethered vesicle system. An ELISA assay was used to independently quantify the amount of primary antibody bound to the surface of the tethered vesicles and was calibrated against results obtained from the binding of antibody to supported lipid bilayers of known antigen surface densities.

binding with sensitivity as high as 5 ng/cm2. Since the pioneering work of Kasemo and co-workers,11-13 who showed the technique could be used to study the mechanism of supported lipid bilayer formation, a number of studies14-20 have used the technique to understand binding of DNA, antibodies, and other biological macromolecules to the surface of a supported lipid bilayer. In previous work,5 we have used QCM-D and fluorescence microscopy to quantitatively characterize the formation of a bilayertethered vesicle assembly in which biotinylated vesicles were tethered to a supported lipid bilayer on silicon oxide via streptavidin. In addition, we have characterized the effects of different vesicle sizes and compositions on the nature of the assembly. In this study, we focus on the binding of antibodies in solution to components of the tethered vesicle membrane. Specifically, our model system consists of tethered vesicles composed of a small fraction of lipids displaying the dinitrophenyl (DNP) hapten group. QCM-D and enzyme linked immunosorbent assays (ELISAs) are used to quantify the dynamics of an anti-DNP IgG1 monoclonal antibody binding to the tethered vesicle system, enabling the use of QCM-D measurements for quantitative, label-free binding analysis. Figure 1 presents a schematic overview of the tethered vesicle assembly and the ELISA assay, which is used as an independent measure of the amount of bound antibody for comparison with the QCM-D results. In addition to demonstrating the ability to detect and quantify the amount of bound antibody via QCM-D, this study serves as a proof-of-concept for quantitative, label-free, receptor-ligand binding using tethered vesicles. In the scientific literature, relatively little work has appeared in which antibody-antigen interactions are examined using QCM(11) Keller, C. A.; Kasemo, B. Biophys. J. 1998, 75, 1397–1402. (12) Keller, C. A.; Glasmastar, K.; Zhdanov, V. P.; Kasemo, B. Phys. Rev. Lett. 2000, 84, 5443–5446. (13) Reimhult, E.; Hook, F.; Kasemo, B. Langmuir 2003, 19, 1681–1691. (14) Larsson, C.; Rodahl, M.; Hook, F. Anal. Chem. 2003, 75, 5080–5087. (15) Graneli, A.; Edvardsson, M.; Hook, F. ChemPhysChem 2004, 5, 729–733. (16) Graneli, A.; Rydstrom, J.; Kasemo, B.; Hook, F. Langmuir 2003, 19, 842– 850. (17) Glasmastar, K.; Larsson, C.; Hook, F.; Kasemo, B. J. Colloid Interface Sci. 2002, 246, 40–47. (18) Hook, F.; Ray, A.; Norden, B.; Kasemo, B. Langmuir 2001, 17, 8305–8312. (19) Reimhult, E.; Larsson, C.; Kasemo, B.; Hook, F. Anal. Chem. 2004, 76, 7211–7220. (20) Svedhem, S.; Dahlborg, D.; Ekeroth, J.; Kelly, J.; Hook, F.; Gold, J. Langmuir 2003, 19, 6730–6736.

6022

Analytical Chemistry, Vol. 81, No. 15, August 1, 2009

D. One previous QCM-D study examining the binding of antibody to constituents of a supported lipid bilayer20 has shown a significant effect of hydrodynamically entrained water on the measured changes in quartz resonant frequency and dissipation factor. Unbound water confined between closely spaced surface species or within the adsorbed film can dramatically affect the QCM-D results leading to significant deviations from the linear Sauerbrey relationship21 between the amount of adsorbed mass and the measured change in quartz resonant frequency shown in eq 1

∆f ) -

n · f0 · ∆m F0 · h0

(1)

In eq 1, ∆f is the measured change in resonant frequency of a quartz crystal with unloaded fundamental resonant frequency, f0, upon a change in mass, ∆m. The density and thickness of the unloaded crystal are represented as F0 and h0, respectively, and the harmonic number is represented by n. Another study employing QCM22 has ignored the effect of water entrainment in the surface-bound antibody layer leading to possible overestimation of bound antibody. Deviations from the Sauerbrey relation can result from water entrained within the adsorbed layer acting as an inertial mass as well as through dissipative viscous losses in viscoelastic materials, such as the tethered vesicle system, which exhibits significantly greater dissipative losses than planar lipid bilayers.5 Thus, approximating the binding of any species to surfaces of the tethered vesicles as a Sauerbrey film leads to a significant error in the estimate of bound mass. Further, the analysis is complicated by the unknown membrane surface area, which depends on vesicle size, vesicle shape, and surface tethering density. The objective of this study was 2-fold: (1) to demonstrate the ability to detect antibody binding to a viscoelastic layer of tethered vesicles using QCM-D and (2) to demonstrate the effectiveness of viscoelastic modeling of the measured frequency and dissipation factor shifts in quantifying the amount of antibody bound. To accomplish this, we developed and used a flow-based ELISA assay (21) Sauerbrey, G. Z. Physik 1959, 155, 206–222. (22) Senaratne, W.; Takada, K.; Das, R.; Cohen, J.; Baird, B.; Abruna, H. D.; Ober, C. K. Biosens. Bioelectron. 2006, 22, 63–70.

that allowed independent quantification of the amount of bound antibody to serve as a point of comparison. Due to uncertainty in the number of accessible antigens on the surface of the tethered vesicles, we also investigated the binding of antibody to a solid supported lipid bilayer, which provides a well-defined geometry that was used to relate the ELISA signal to the accessible hapten surface density. The study revealed a number of important findings in interpreting QCM-D responses for binding to viscoelastic films and for the binding of antibodies to antigens in fluid membranes. In addition, some perspective is provided for interpreting the binding of other ligands to receptors hosted by tethered vesicles. METHODOLOGY Materials. Egg-yolk-derived L-R-phosphatidylcholine (egg PC); 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(Cap Biotinyl) (Biotin-X-PE); 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamineN-[6-[(2,4-dinitrophenyl)amino]caproyl] (DNP-X-PE); and 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-l-serine] (POPS) were purchased dissolved in chloroform from Avanti Polar Lipids (Alabaster, AL). For the ELISA, a monoclonal rat anti-DNP IgG1 (clone LODNP-2) was used as the primary antibody, and a horseradish peroxidase (hrp)-labeled mouse antirat IgG1 monoclonal (clone MARG1-2) was used as the secondary antibody; both were purchased from Zymed Laboratories/Invitrogen Life Technologies (Carlsbad, CA). One-step ABTS (Pierce Biotechnology, Rockford, IL) was used as the substrate for the hrp-catalyzed reaction. Phosphate buffered saline (PBS) solution was prepared by diluting a 10× concentrate at pH 7.4 (Gibco/Invitrogen Life Technologies, Carlsbad, CA) with deionized water obtained from a Milli-Q RO system (Millipore, Billerica, MA). Blocking buffer was prepared by dissolving lyophilized bovine serum albumin (BSA) (Sigma, St. Louis, MO) in PBS to a concentration of 3.00 g/L. Lyophilized streptavidin was also purchased from Sigma and dissolved in PBS. Tris buffer at pH 8.0 was used in the preparation of the supported lipid bilayers and was prepared by dissolving Tris (Invitrogen Life Technologies) and sodium chloride (Mallinckrodt Laboratory Chemicals, Phillipsburg, NJ) in Milli-Q water to final concentrations of 10 mM and 100 mM, respectively. 7x detergent concentrate (MP Biomedicals, Aurora, OH) and 0.1 M sodium dodecylsulfate (SDS) (VWR Scientific, West Chester, PA) were used for various cleaning steps. Vesicle Preparation. Vesicles of different lipid compositions were made as described previously.5 Briefly, a thin lipid film of desired composition was dried on the walls of a round-bottom flask from a chloroform solution under a stream of nitrogen gas. After removing any residual chloroform in the film through vacuum desiccation over a minimum period of two hours, the film was hydrated with the desired aqueous solution (i.e., buffer or Milli-Q water) and then extruded through a track-etch membrane of appropriate pore diameter using an Avanti MiniExtruder. Vesicles used for bilayer formation were hydrated with Milli-Q water and extruded 16 times through an 80-nm pore diameter membrane prior to extrusion 11 times through a 30-nm pore diameter membrane. Vesicles prepared for tethering were hydrated with PBS containing 2 wt % BSA and extruded 41 times through a 50-nm pore diameter membrane to obtain a very narrow, unimodal size distribution centered around 80-85 nm. Note that

we used a smaller tethered vesicle size than in our previous work,5 as the smaller size resulted in lower dissipation values and less nonspecific adsorption. Vesicle solutions were prepared at 10 mg/ mL lipid concentration and were diluted to ∼0.6 mg/mL just prior to use. QCM-D Measurements. QCM-D measurements were performed using a Q-Sense E4 instrument operating at 25.1 °C under a continuous flow rate of 0.1 mL/min. Silicon oxide-coated 4.95 MHz sensor crystals were purchased from Q-Sense and were stored in 0.1 M sodium dodecylsulfate (SDS). Just prior to use, the crystals were cleaned as described before5 with water and ethanol rinses prior to exposure to an oxygen plasma for 2 min. The crystals were mounted into the flow chambers after being dried following a final rinse with ethanol. Frequency and dissipation shifts at the third, fifth, seventh, ninth, 11th, and 13th overtones were measured using QSoft 401 v 1.4.2.110 supplied by Q-Sense. In a compromise between noise and time resolution, data points were collected approximately once every 3 s for each resonance. In general, each time a new solution was introduced into the measurement chamber, the solution was initially allowed to flow from a reservoir through the flow module housing the sensor crystal and out to waste for 10 min at 0.1 mL/min; the volume of this flow path was ∼0.5 mL. After the flow path had been flushed (10 min), the exit stream was recycled to maintain continuous flow during long adsorption steps. Net changes in frequency and dissipation for streptavidin binding, vesicle tethering, and antibody binding were determined by taking the difference between the average of one minute’s worth of data collected starting three minutes after the start of the blocking buffer rinse and the average of one minute’s worth of data collected starting two minutes prior to injection. After each experiment, the entire flow path was incubated with 0.1 M SDS for a minimum of one hour and then rinsed extensively with water and ethanol. Formation of the Tethered Vesicle Assembly. A tethered vesicle assembly was prepared in a sequential manner similar to that previously described.5 In this approach, a lipid bilayer was first formed by exposing osmotically shocked biotinylated vesicles to a clean SiO2-coated sensor crystal equilibrated in water. To obtain osmotically shocked vesicles, vesicles prepared in water were diluted with Tris buffer (10 mM Tris, 100 mM NaCl, pH 8.0) immediately prior to injection. In a departure from the general methodology described in the previous section, a flow rate of 1 mL/min instead of 0.1 mL/min was used to introduce the bilayer-forming vesicle solution in order to prevent the osmotically shocked vesicles from reaching a stable state before being exposed to the substrate. The assembly process was modified from previous work to include a blocking step to prevent nonspecific adsorption of antibodies to tubing walls, etc. For this purpose, a blocking buffer consisting of 0.3 wt % BSA in PBS was allowed to flow through the system for a minimum period of 45 min following bilayer formation. All solutions subsequently introduced into the flow module contained the same amount of BSA. After bilayer formation and the introduction of blocking buffer, a 0.05 mg/mL solution of streptavidin in blocking buffer was introduced, and streptavidin was allowed to bind to the biotinylated Analytical Chemistry, Vol. 81, No. 15, August 1, 2009

6023

lipids in the bilayer for 10 min. This was followed by a 10-min rinse with blocking buffer. The final step of the assembly process was the introduction of biotinylated vesicles containing various amounts of DNP-X-PE into the flow module. One hour was allowed for vesicle tethering after which there was a final 10-min rinse with blocking buffer. A lipid composition of 95% Egg PC, 2% Biotin-X-PE, and 3% POPS was used for the bilayer with the Biotin-X-PE content being high enough to ensure the bilayer had reached its streptavidin binding capacity.5 Vesicles used for tethering purposes consisted of approximately 95% Egg PC, 0.0125% Biotin-X-PE, and various amounts of DNP-X-PE, with the balance made up by POPS. This was done to ensure that the vesicles and the bilayer maintained a constant surface charge. Antibody Binding Experiments. Antibody binding experiments were conducted on both tethered vesicle assemblies formed as described above as well as on lipid bilayers composed of 95% Egg PC and various amounts of DNP-X-PE, with the balance made up by POPS. A 0.00625 mg/mL monoclonal anti-DNP solution (1: 160 dilution of 1 mg/mL) in blocking buffer was allowed to bind to the DNP-X-PE containing lipid bilayer or tethered vesicles for 1 h under continuous flow at 0.1 mL/min. After binding, the system was rinsed with blocking buffer for 10 min. Flow-Based ELISA. For ELISA measurements, a solution of monoclonal secondary hrp-labeled antibody was introduced into the QCM-D measurement chamber after the primary antibody (anti-DNP) binding step. After a 10-min blocking buffer rinse, 1-step ABTS solution was allowed to flow over the entire assembly at a fixed flow rate of 21 µL/min. This allowed the ABTS solution sufficient residence time above the crystal to produce a measurable change in the absorbance of the effluent stream over the range of DNP-X-PE concentrations investigated. After steady state was established (15 min), effluent was collected for 5 min. One hundred microliters of this effluent solution was transferred to a standard 96-well plate, and its absorbance was generally measured within 5 min of collection using a microplate reader (Molecular Devices SpectraMax 190). Because the extent of reaction was directly influenced by the flow rate, syringe pumps were used for more precise control of flow instead of a peristaltic pump for this part of the protocol. This practice reduced the uncertainty of the absorbance measurement by avoiding inconsistencies in the residence time of the ABTS solution in the flow module. General Viscoelastic Modeling Methodology. The net changes in the normalized frequency and dissipation at all the harmonics except the fundamental were fit to one- and two-layer mechanical viscoelastic models proposed by Voinova et al.,23 using a Simplex-based unconstrained minimization routine in Matlab’s optimization toolbox. The quality of the fit was evaluated by the value of the following fitness function

Q(n, µ, η, F, t) )

∑ n

(

)

∆f nmeasured - ∆f ncalculated 2 + σ∆fn

(

∆Dnmeasured

- ∆Dncalculated σ∆Dn

)

2

(2)

(23) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Phys. Scr. 1999, 59, 391–396.

6024

Analytical Chemistry, Vol. 81, No. 15, August 1, 2009

where n is the harmonic number of each resonance monitored (n ) 3, 5, 7, ... 13), and µ, η, F, and t are the film parameters: shear modulus, viscosity, density, and thickness, respectively. The value of this function is zero for a perfect fit and increases as the quality of the fit degrades. Note that each term is weighted by the inverse of the experimentally measured standard deviation of the shifts from multiple runs, σ. This is analogous to a χ2 fit for a single variable problem. Matlab’s fminsearch unconstrained minimization routine employing a Simplex-based algorithm was used to determine the best fitting film parameters. It was found that the specific values of the fitted film parameters varied as a function of the initial guess value, indicating the presence of local minima in the Q function for a given set of experimental data. To overcome this problem, the global minimum in the function was found by initiating the minimization at multiple sets of starting film parameters. Sets of film parameters that produced very low values for the fitness function, but which were nonphysical (i.e., parameter values less than zero), were ignored in determining the global minimum. The density for the one-layer fits was fixed and assumed to be that of water (1 g/cm3). RESULTS Experimental Overview. A typical antibody binding experiment including the formation of the tethered vesicle system is shown in Figure 2, which presents both the frequency and dissipation changes at all the observed overtones for tethered vesicles composed of 3 mol % DNP-X-PE. Bilayer formation showed a net normalized frequency change of approximately -25 Hz and a net dissipation shift of less than 0.1 × 10-6 at all of the overtones, indicating the formation of a high quality lipid bilayer.12 Streptavidin binding resulted in a normalized frequency shift of approximately -30 Hz at all of the resonances and exhibited net dissipation shifts less than 1 × 10-6 consistent with prior work indicating streptavidin surface saturation.5 The introduction of biotinylated vesicles resulted in large, immediate changes in both the recorded frequency and dissipation shifts at all of the measured resonances. Similarly, primary antibody binding resulted in large, immediate shifts in both frequency and dissipation factor. Dissipation Kinetics upon Antibody Binding to the Tethered Vesicles. The kinetics of the changes in the resonant frequencies upon antibody binding to the tethered vesicles were unremarkable, but the behavior of the dissipation factor and its dependence on harmonic number was particularly noteworthy. Figure 3 presents the dissipation changes upon antibody binding to the tethered vesicles as a function of time for a variety of tethered vesicle DNP-X-PE concentrations. At the lower resonances, the dissipation factor exhibits a local maximum and minimum before reaching its final values, while higher order resonances only show a point of inflection. ELISA Assays for Antibody Binding to the Tethered Vesicle Assembly. The ELISA was performed in situ after tethered vesicle assembly construction. In Figure 4, the average measured absorbance of the effluent stream and the normalized frequency shift upon antibody binding to the tethered vesicles are both plotted against the vesicle DNP-X-PE concentration. The absorbance data suggest antibody surface saturation occurs between 1.5 and 2 mol % DNP-X-PE in the tethered vesicles. The magnitude of the net frequency shift decreases linearly until 1.5-2

Figure 2. Typical antibody binding experiment. Normalized changes in (a) resonant frequency and (b) dissipation factor at each of the measured overtones are plotted against time for the formation of the tethered vesicle assembly as well as subsequent antibody binding to tethered vesicles composed of 3 mol % DNP-X-PE. The introduction of (1) osmotically shocked vesicles, (2) blocking buffer, (3) streptavidin, (4) tethering vesicles, and (5) anti-DNP are marked on both plots. Appropriate rinse steps (unmarked) were typically performed 10 min prior to the introduction of each new species. The baseline is taken to be that of a bare SiO2-coated quartz crystal in water.

mol % DNP-X-PE. For DNP-X-PE concentrations greater than this range, the frequency shifts asymptotically approach a fixed value. For both sets of data, each data point in the figure represents the average of 3-4 runs, and the error bars represent one standard deviation around the mean. Correlation of QCM-D and ELISA Results. As mentioned previously, the ELISA measurements served as an independent measure of the amount of antibody bound to the surface of the tethered vesicles to which the mass of bound antibody obtained from modeling the QCM-D frequency and dissipation shifts could be correlated. A solid supported lipid bilayer was used to obtain known surface densities of DNP to calibrate the ELISA responses against DNP surface densities. See the Supporting Information for additional information regarding the binding of antibody to the lipid bilayer. Figure 5 plots the measured net normalized frequency shift upon antibody binding to tethered vesicles against the antibody surface mass density obtained from ELISAs as it approaches saturation (i.e., for vesicle DNP-X-PE content e2 mol %). Each data point arises from the mean value of the frequency shift and absorbance measurement for between two and ten different runs. There are two clearly defined regions in the plot. Below an

Figure 3. Changes in dissipation factor at each resonance upon antibody binding to tethered vesicles composed of various percentages of DNP-X-PE. Note that the baseline for each resonance represents the respective measured value of the dissipation factor prior to the introduction of antibody, and time ) 0 is taken to be 5 min prior to the introduction of antibody solution.

Figure 4. The net changes in resonant frequency at the 3rd harmonic upon antibody binding to the tethered vesicle assembly (black circles) and the corresponding ELISA absorbance measurements (green triangles). The error bars represent the standard deviations of at least three measurements.

antibody surface mass density of approximately 800 ng/cm2 (corresponding to 0.75 mol % DNP-X-PE in the tethered Analytical Chemistry, Vol. 81, No. 15, August 1, 2009

6025

Figure 5. The normalized net frequency shifts upon antibody binding to the tethered vesicle assembly for vesicle DNP-X-PE concentrations