After too many hours, I’m giving up on fluorescence polarization (FP) as a generic binding assay. Clearly, it works in some cases with some labels. I got a weak response from one of the aptamers I tried, but another two failed to show anything. And I’m screaming for signal even in the “positive” case. Published results suggest several other aptamers have a FP shift upon binding. But, clearly, not all of them.
For a dye like fluorescein, the local environment around the molecule is probably more important than the size of the complex. So, unless binding changes the conformation in a way that is near to the dye, there’s no FP change. It would be interesting to try a survey of published DNA aptamers and their targets to see which show an FP change and which don’t. The results could be related to structural elements. That could inform design. One group did do some design work to make a FP-aptasensor for small molecules. The binding site and the fluorophore modification were strategically placed close together so that binding was more likely to affect FP. More work in this direction could help such designs. But, ultimately, FP is not as generically useful as I would like.
Today I tried thermofluorimetric analysis based on work by the Easley lab at Auburn University. I got immediate results from two published aptamers comparable to the results in their Analytical Methods paper.
The basic idea is to melt the aptamer and then look for the changes in the melt curve after addition of the target. The target should stabilize the aptamer, so there should be a peak in the melt curve at higher temperatures. Indeed, when I add protein to my aptamer, I see such a peak. I did it at multiple concentrations and it looks like a binding curve. But when I fit the data I get a weaker Kd than the original papers suggested.
It seems to me that deriving the dissociation constant using this method will inherently report the equilibrium constant at elevated temperatures. The methods paper showed strong binding, but they chose a particularly strong sub-nanomolar aptamer as proof of concept. They didn’t show a Kd calculation. So maybe at 62 °C the 0.1 nM aptamer is acting like a 1 nM aptamer. That would be a lot weaker than the original but would still give a clear binding curve.
I need to try more conditions to prove that I’m able to derive a Kd from thermofluorimetric assay. Or not.
I also need to do a protein-only control. My DNA dye should be pretty specific, but it’s important to check that it’s not interacting with the protein and giving spurious signals.