The current biophysics toolbox is composed of a number of well-established methods that are continuously being enriched by novel techniques that are higher throughput to further our mechanistic understanding of low molecular weight drug candidates (Ottl J. Biophysics/label-free assays in hit discovery and verification. In M Cooper, LM Mayr (eds), Label-free Technologies for Drug Discovery. Wiley; 2011, pp. 155–169). Among the available biophysical methods, a number of techniques are based on biosensors where ligand interactions with immobilized biological molecules are monitored. Biosensor-based techniques vary in type and instrument design within each type, and they are gaining traction due to their label-free nature and the ability to provide real-time data (Daghestani et al., Sensors 2010;10:9630–9646).
A long-standing and perhaps one of the most widely used biosensor-based techniques is surface plasmon resonance (SPR). First demonstrated in the late 1960s and commercialized in the 1990s, SPR utilizes single transverse magnetic (TM) polarization for detection of binding events. In contrast, a more recently introduced evanescent technique (commercialized in 2000), dual polarization interferometry (DPI), allows measurements in both the TM and transverse electric (TE) modes, thus providing information not only on changes in refractive index but also in thickness. As a result, DPI offers the opportunity to probe structural changes that occur during binding in addition to the kinetics of the interactions.
Potentially, DPI could meet the challenging need to monitor the real-time interplay between binding mode of action (MoA) and structural change. In the past, DPI has been successfully shown to detect protein structural changes upon ion binding, such as Zn2+ (Fresquet et al., J Biol Chem 2007;282:34634). Herein, using calmodulin (CaM) as a model system, Coan et al.* demonstrated that DPI was capable of detecting protein structural change induced by small molecule ligand binding.
The authors first examined the immobilized CaM system by performing CaCl2 injections with its final concentration ranging from 3.9 to 500 μM (see Figure 1). By using MgCl2 injection in the same range to account for bulk effect and to eliminate nonspecific binding effect, the authors deduced a KD value for CaM-Ca2+ in the range of 14–34 μM. Given that CaM contains four high-affinity Ca2+ binding sites, the authors further attributed the relatively low affinity (literature value: 1–10 μM) to the simplicity of the one-site binding model that was used to fit the data. The authors monitored signal changes by injecting a known CaM ligand, trifluoperazine (TFP) (see Figure 2) (along with a known nonbinder 5-hydroxytryptophan, which failed to exhibit any binding response).
Despite the limited solubility, a single digit micromolar KD value, comparable to literature reports, was derived from the concentration range tested (0.625–10 μM). In addition, an approximate 2:1 stoichiometry was estimated based on the max mass response. The discrepancies in stoichiometry between the one obtained in this study and some previously reported values could be linked to the manner the protein was immobilized (undirected in this case vs. directed strategies).
Overall, this case study illustrated that DPI could report and distinguish protein conformational change upon binding to small molecule ligands or ions that acted through different MoA. The detection of binding event possessing a high size ratio (protein:ion > 400) once again highlights the sensitivity of the technique, while the capability to reliably determine KD for small molecule binders further extends the technique to its less frequently tapped arena. Since CaM is a relatively well-characterized protein (see for example Bornhop et al., Science 2007;317:1732–1736) and is known for its substantial structural rearrangement upon binding, it would be interesting to apply DPI to a broad range of targets, especially those that may undergo small or even subtle ligand-induced structural changes.
*Abstract from Analytical Chemistry 2012, Vol. 84: 1586–1591
In early drug discovery, knowledge about ligand-induced conformational changes and their influence on protein activity greatly aids the identification of lead candidates for medicinal chemistry efforts. Efficiently acquiring such information remains a challenge in the initial stages of lead finding. Here we investigated the application of dual polarization interferometry (DPI) as a method for the real-time characterization of low molecular weight (LMW) ligands that induce conformational changes. As a model system we chose calmodulin (CaM), which undergoes large and distinct structural rearrangements in response to calcium ion and small molecule inhibitors such as trifluoperazine (TFP). We measured concentration dependent mass, thickness, and density responses of an immobilized CaM protein layer, which correlated directly with binding and conformational events.
Calcium ion binding to CaM induced an increase in thickness (≤0.05 nm) and decrease in density (≤−0.03 g/cm3) whereas TFP induced an increase in both thickness (≤0.05 nm) and density (≤0.01 g/cm3). The layer measurements reported here show how DPI can be used to assess and differentiate ligands with distinct structural modes of action.