RNA-Ligand Interactions Quantified by Surface Plasmon Resonance with Reference Subtraction

Abstract

Structured RNAs bind ligands and are attractive targets for small-molecule drugs. A wide variety of analytical methods have been used to characterize RNA-ligand interactions, but our experience is that most have significant limitations in terms of material requirements and applicability to complex RNAs. Surface plasmon resonance (SPR) potentially overcomes these limitations, but we find that the standard experimental framework measures notable nonspecific electrostatic-mediated interactions, frustrating analysis of weak RNA binders. SPR measurements are typically quantified relative to a non-target reference channel. Here, we show that referencing to a channel containing a non-binding control RNA enables subtraction of nonspecific binding contributions, allowing measurements of accurate and specific binding affinities. We validated this approach for small-molecule binders of two riboswitch RNAs with affinities ranging from nanomolar to millimolar, including low-molecular-mass fragment ligands. SPR implemented with reference subtraction reliably discriminates specific from nonspecific binding, uses RNA and ligand material efficiently, and enables rapid exploration of the ligand-binding landscape for RNA targets.

Figure 1:. Overview of reference subtraction in an SPR experiment.

(A) Referencing to a blank flow cell controls for bulk refractive index changes and allows quantification of total surface binding signal. (B) Use of a nonbinding RNA in the reference flow cell allows additional subtraction of signal due to nonspecific analyte–macromolecule binding.

Figure 2:. Small molecules bind RNA in both specific and nonspecific modes.

(A) Secondary structures of the TPP and SAM-III riboswitch RNAs and their ligands Z1 and SAM, respectively. (B) SPR sensorgrams, and (C) binding curves for each small molecule analyte. Sensorgrams and regions of the fits indicative of nonspecific binding are emphasized in red.

Figure 3:. Referencing to mutant RNA subtracts nonspecific binding.

(A) Mutations (blue) introduced in the binding sites for the TPP and SAM-III riboswitch RNAs to disrupt specific binding. (B) Representative SPR sensorgrams (left) and binding curves (right) with and without a binding site mutant in the reference channel. Arrows emphasize effect of incorporating the reference RNA. (C) Values for K_D and NS (Eqn. 1) as a function of reference channel.

Figure 4:. Impact of reference RNA choice on nonspecific binding subtraction.

(A) Secondary structures of RNAs used for reference subtraction. Note that the binding site mutant requires high levels of prior structural knowledge, whereas the ΔP4 mutant, SAM-III riboswitch, and structured RNA controls do not. (B) SPR responses for Z1 binding to the TPP riboswitch, using subtraction with different reference RNAs. (C) Values for K_D and NS (Eqn. 1) for Z1 binding the TPP riboswitch as a function of reference RNA.

Figure 5:. Correlation between dissociation constants determined by SPR with those obtained using other methods.

K_D values for ligands of the TPP (black) and SAM-III (blue) riboswitch RNAs measured by SPR, using a binding-site mutant RNA for reference subtraction (Fig. 3), plotted versus values determined by ITC and other methods (Table S1). Correlation analysis is based on values for log(K_D).