[Protein-analysis] Re: Drug binding: When fluorescence polarization fails

[Marvin]

Dominic-Luc Webb wrote:
> I have worked for some time with fluroescence polarization (FP)
> in studies of drug-protein binding. My drug, which is
> hydrophobic, typically has a fluorescent adduct and the
> solution has 0.01% Tween 20 to keep down non-specific binding
> to test tubes, multiwell plates, etc. Assay usually works well.
> 
> I usually do a check with a series of glycerol concentrations
> to ensure that FP is working OK. I expect at high glycerol
> concentration, the higher viscosity lowers molecular rotation
> which is detected as an increase in the amount of polarized
> light in the static (parallel) channel. This usually works
> fine, fluorescein for instance. I recently have found some
> labelled drugs that show no increase in polarization, and
> even a small decrease (about 20 mP) even when adding glycerol
> to 50%. Again, Tween 20 is 0.01% throughout. Fluorescence
> intensity remains relatively constant with only a slight
> parallel increase in both channels. The underlying physics of
> FP always seemed pretty straightforward, but now I am
> confronted with such data that has been reproduced several
> times and it is not clear how this came about.
> 
> Strangely, when a protein that is known to be a drug target is
> present, there is a strong increase in polarization (100-300 mP).

That is what I'd expect.  Binding to a large molecule (i.e., 
the protein) reduces the rotation rate of the labeled drug, 
thus increasing the polarization.

> The glycerol data has created an uncertainty about negative
> data for test proteins that I am screening for drug binding.
> 
> Is there anyone in here that might have some hints how this
> lack of change in polarization in glycerol might have come
> about, maybe some suggestions how to troubleshoot this?

It could be that the change of solvent to 50% glycerol 
reduces the fluorescence lifetime, thus increasing the 
measured polarization.

> 
> Dominic
>