Using fluorophore-labeled oligonucleotides to measure affinities of protein-DNA interactions

Abstract

Changes in fluorescence emission intensity and anisotropy can reflect changes in the environment and molecular motion of a fluorophore. Researchers can capitalize on these characteristics to assess the affinity and specificity of DNA-binding proteins using fluorophore-labeled oligonucleotides. While there are many advantages to measuring binding using fluorescent oligonucleotides, there are also some distinct disadvantages. Here we describe some of the relevant issues for the novice, illustrating key points using data collected with a variety of labeled oligonucleotides and the relaxase domain of F plasmid TraI. Topics include selection of a fluorophore, experimental design using a fluorometer equipped with an automatic titrating unit, and analysis of direct binding and competition assays.

Figure 1

Binding of TraI36 to a fluorophore-labeled single-stranded oligonucleotide has different effects on different fluorophores. The changes in fluorescence intensity (top) and anisotropy (bottom) are shown for a 17-base oligonucleotide (5′-TTTGCGTGGGGTGTGTG-3′) 3′-labeled with TAMRA (open circles), 6-FAM (filled triangles), or Cy3 (filled squares). The solid lines depict simultaneous fits to intensity and anisotropy data using SPECTRABIND (Toptygin and Brand, 1995a; Toptygin and Brand, 1995b) as described (Stern and Schildbach, 2001).

Figure 2

The linkage between fluorophore and DNA can affect fluorescent properties. TraI36 titrations into solutions of 22-base (5′-TTTGCGTGGGGTGTGTGCTTTT-3′) single-stranded oligonucleotide 3′-labeled with TAMRA using either a CPG (open circles) or NHS (filled diamonds) linkage are shown. The solid lines depict simultaneous fits to intensity and anisotropy data using SPECTRABIND (Toptygin and Brand, 1995a; Toptygin and Brand, 1995b) as described (Stern and Schildbach, 2001).

Figure 3

The proximity of fluorophore to bound protein can affect fluorescent properties. Single-stranded 22-base (5′-TTTGCGTGGGGTGTGTGCTTTT-3′; open squares) or 17-base (5′-TTTGCGTGGGGTGTGTG-3′; filled squares) oligonucleotides were synthesized with a 3′-6-FAM probe. Intensity (top) or anisotropy (bottom) changes occurring upon titration of TraI36 are shown. The solid lines depict simultaneous fits to intensity and anisotropy data using SPECTRABIND (Toptygin and Brand, 1995a; Toptygin and Brand, 1995b) as described (Stern and Schildbach, 2001).

Figure 4

Changing solution conditions can alter fluorescent behavior of probes. TraI36 binding to 22-base oligonucleotides 3′-labeled with TAMRA (open circles) or Cy3 (filled squares) at pH 10.5 causes less significant changes to intensity (top) and decreases instead of increases in anisotropy (bottom) (compare to Figure 1). The solid lines depict simultaneous fits to intensity and anisotropy data using SPECTRABIND (Toptygin and Brand, 1995a; Toptygin and Brand, 1995b) as described (Stern and Schildbach, 2001).

Figure 5

Amino acid substitutions in TraI36 can alter fluorophore behavior. Shown are the increases in fluorescent anisotropy of a 3′-TAMRA-labeled 22-base oligonucleotide upon binding of wild type TraI36 (open circles) or the TraI36 variant proteins E187A (filled squares) or R150A (filled diamonds). The solid line depicts simultaneous fits to intensity and anisotropy data using SPECTRABIND (Toptygin and Brand, 1995a; Toptygin and Brand, 1995b) as described (Stern and Schildbach, 2001).

Figure 6

Titration scheme using an automatic titrator. Shown is a curve for the binding of an F TraI36 variant protein to an oligonucleotide containing a sequence from the R100 plasmid. For an interaction with K_D = 7 nM and fluorescent oligonucleotide concentration of 4 nM, three step sizes (0.08 nM step size to 0.96 nM, 3 nM step size to 30 nM, 65 nM step size to 400 nM) were used to generate this binding curve. Note the breaks at the point at which the step size changes. The titration scheme has to be adjusted to prevent these from occurring at points in the curve where they could reduce the quality of the fit. The solid line depicts a fit to the data using KaleidaGraph as described (Larkin et al., 2007).