Binding, brightness, or noise? Extracting temperature-dependent properties of dye bound to DNA

We present a method for extracting temperature-dependent thermodynamic and photophysical properties of SYTO-13 dye bound to DNA from fluorescence measurements. Together, mathematical modeling, control experiments, and numerical optimization enable dye binding strength, dye brightness, and experimental noise (or error) to be discriminated from one another. By focusing on the low-dye-coverage regime, the model avoids bias and can simplify quantification. Utilizing the temperature-cycling capabilities and multi-reaction chambers of a real-time PCR machine increases throughput. Significant well-to-well and plate-to-plate variation is quantified by using total least squares to account for error in both fluorescence and nominal dye concentration. Properties computed independently for single-stranded DNA and double-stranded DNA by numerical optimization are consistent with intuition and explain the advantageous performance of SYTO-13 in high-resolution melting and real-time PCR assays. Distinguishing between binding, brightness, and noise also clarifies the mechanism for the increased fluorescence of dye in a solution of double-stranded DNA compared to single-stranded DNA; in fact, the explanation changes with temperature.