High-affinity immobilization of proteins using biotin- and GST-based coupling strategies

Abstract

Surface plasmon resonance (SPR) is a highly sensitive method for the detection of molecular interactions. One interacting partner is immobilized on the sensor chip surface while the other is injected across the sensor surface. This chapter focuses on high-affinity immobilization of protein substrates for affinity and kinetic analyses using biotin/streptavidin interaction and GST/anti-GST-antibody interaction.

Figure 1. A typical single flow cell sensorgram generated during the loading of anti-GST antibody

The flow cell surface was activated for direct amine coupling of the anti-GST antibody (αGST-Ab) by injecting the coupling solution (40 μl NHS/EDC at 5 μl/min flow rate with EXTRACLEAN command). Anti-GST antibody was then diluted to 30 μg/ml in immobilization buffer and 45 μl were injected at 5 μl/min flow rate, omitting the EXTRACLEAN command. To deactivate esters on the sensor chip surface, ethanolamine was injected (35 μl at 5 μl/min flow rate with EXTRACLEAN command). To test the binding surface, supplied recombinant GST protein was diluted to 5 μg/ml and 100 μl were injected over the sensor chip at 20 μl/min flow rate. To remove GST from the binding surface, 40 μl of glycine-HCl pH 2.2 was injected at 20 μl/min flow rate.

Figure 2. Affinity and kinetic measurements using an anti-GST antibody CM5 sensor chip

A. Flow cells 1 and 2 were loaded with 1000 RUs of GST and GST-RGS12GoLoco proteins, respectively. Increasing concentrations of Gα_i1 were injected using the KINJECT command (300 μl injections with a 200 second dissociation phase at 20 μl/min flow rate). Specific binding was determined by subtracting non-specific binding to a GST control flow cell from the GST-RGS12GoLoco response curve. B. Binding affinities were determined by plotting the maximum response attained at each concentration of Gα_i1 versus the concentration of the Gα_i1, then by fitting to a Langmuir binding isotherm using GraphPad Prism 5.0 (GraphPad Software, La Jolla, CA) to determine dissociation constant (K_D). Kinetic analyses of the association (k_a) and dissociation (k_d) rates were also used to determine K_D using the simultaneous k_a/k_d (1:1 Langmuir) model (BIAevaluation). Differences in K_D determinations between kinetic and equilibrium analyses are likely to reflect mass-transport and/or rebinding limitations to binding assay execution, as indicated by large k_a values, nearly linear initial association seen on sensorgrams, and/or greatly slowed dissociation (see Schuck, P. et al. in this Volume for further information).

Figure 3. Diverse applications of an anti-GST CM5 sensor chip

A. A GST-RGS fusion protein was immobilized on the surface of flow cell 1 and GST was immobilized on the surface of flow cell 2 (the latter surface to use to subtract sensorgram changes in resonance due to buffer shifts). Gα_i1 pre-incubated with either GDP or GDP+Mg²⁺+AlF₄⁻ (AMF), a transition-state mimetic form known to bind avidly to RGS proteins, was injected over the sensor surface. GST-RGS bound to the Gα_i1(AMF) but not Gα_i1(GDP) as expected. B. GST-H-Ras loaded with a non-hydrolyzable GTP analogue, GTPγS, was loaded in flow cell 1, GST-H-Ras loaded with GDP was loaded in flow cell 2, and GST control was loaded in flow cell 3. An N-terminal construct of c-Raf-1, a known H-Ras binding partner, was injected across the sensor chip surface. The data show the nucleotide state selectivity of the H-Ras/c-Raf-1 interaction. C. GST-RGS14 was loaded in flow cell 1 and GST was loaded in flow cell 2. Gα_i1 was injected in GDP+Mg²⁺+AlF₄⁻ (AMF) Running Buffer over the sensor chip surface and a Gα_i1/RGS14 specific interaction was detected. Specific binding for all injections was determined by subtracting the non-specific binding observed from a GST control flow cell from the experiment-containing flow cells.

Figure 4. Confirmation of biotinylated protein production

A. SDS-PAGE gel stained with Coomassie Blue dye reveals the migration of biotinylated Gα_i1, designated b-Gα_i1, and unlabeled Gα_i1. B. Immunoblot of SDS-PAGE gel transferred to nitrocellulose using streptavidin-horseradish peroxidase antibody to detect biotinylated protein.

Figure 5. Loading Streptavidin (SA) Sensor Chip

A typical sensorgram generated during the loading of biotinylated protein to flow cells 1 and 2. The flow cell surface was stabilized with 3 × 20 μl injections of Regeneration Buffer before loading of biotinylated-Gα_i1 (260 nM final concentration, as diluted in Biotin-conjugate Immobilization Buffer), one flow cell at a time, until saturation is observed (note plateau in sensorgram curves).

Figure 6. Affinity measurements using a SA sensor chip

A. Flow cell 1 was loaded with 450 RUs of Biotin-KB-801 peptide, and flow cell 2 was loaded with 350 RUs mNotch peptide as a control. Increasing concentrations of Gα_i1 in GDP+Mg²⁺+AlF₄⁻ (AMF) Buffer, the transition state mimetic form, were injected over the sensor surface as indicated, using the KINJECT command (300 μl injections with a 200 second dissociation phase at 20 μl/min flow rate). Binding curves were obtained by subtracting non-specific binding to a non-interacting biotinylated peptide from all experiment-containing flow cells. B. Binding affinities were determined by plotting the maximum response attained at each concentration of Gα_i1 versus the concentration of the Gα_i1, then by fitting to a rectangular hyperbola to determine K_D using GraphPad Prism 5.0 (Graphpad Software, La Jolla, CA). C. Flow cell 1 was loaded with 450 RUs of Biotin-Gα_i1, and flow cell 2 was loaded with an equivalent RU signal of denatured Biotin-Gα_i1. Varying concentrations of RGS10 in GDP+Mg²⁺+AlF₄⁻ (AMF) Buffer were injected over the sensor surface as indicated, using the KINJECT command (300 μl injections with a 200 second dissociation phase at 20 μl/min flow rate). Specific binding was determined by subtracting non-specific binding from a flow cell containing denatured biotin-Gα_i1. D. Binding affinities were determined by plotting the maximum response attained at each concentration of RGS10 versus the concentration of the RGS10, then by fitting to a rectangular hyperbola to determine K_D using GraphPad Prism 5.0 (GraphPad Software, La Jolla, CA). These final panels were reproduced from Soundararajan et al. 2008 (ref. (7)). Copyright (c) 2008 National Academy of Sciences, U.S.A.