I think SPR might be an ideal technique for this. You can immobilize the DNA on the chip and then measure an affinity for your TF in the presence and absence of your second protein. Check out the RedCaT technology described in this paper by one of my former colleagues.

http://nar.oxfordjournals.org/content/41/14/7009

SPR is a good choice but if you have never done SPR then it might take a while to optimize it. If you have done EMSA then it might be worth trying it. In EMSA you will be detecting the DNA (free or bound). The free proteins most likely will not enter the gel. It would depend on their pKa. Normally DNA binding proteins are positively charged.

If the TF affinity for the promoter site is weak (in microM range), then ITC might be another good method to try.

I'd try EMSA doing titrations, then you could ideally obtain the Kds for each of the complexes. Of course you'll be comparing the Kd of your TF with that of the TF/protein complex for each DNA, but there's no way to separate the effect of the other protein in your TF in isolation to the interaction with the DNA. Anyway, that's what you'd be interested in, in the first place. 

Probably I'd pick a constant DNA concentration and then titrate increasing concentrations of TF in one gel and increasing concentrations of a 1:1 mixture of TF/protein in the second gel. (do you know the stoichiometry of your TF/protein complex?). You could also run a third gel titrating increasing concentrations of your protein in the presence of constant DNA and constant protein partner, then you'd basically compare with the first gel the affinity of your TF in both situations... 

I would recommend MicroScale Thermophoresis (MST). It allows you to do straight forward assay setup, and the experiment is fast and easy to perform.

In your case, I would first use fluorescently labeled DNA, then bind your TF to it. In a second experiment, I would analyze your interacting protein binding to the DNA. Now, in the third experiment, you could pre-incubate the DNA with the interacting protein and then analyze TF binding to that complex. Alternatively, you could pre-incubate the DNA with TF and then add the interacting protein. This way you will be able to determine a set of affinities and see if the affinity of TF is enhanced in the presence of the interacting protein or not!

Read more about MST applications (and especially for interactions with more than just 2 interactants):

http://www.nanotemper-technologies.com/applications/

Hi all,

Thanks for your kind responses.

Ellis, I agree with Soumya that SPR is challenging and I'm not experienced. Is your lab at UCSD experienced? I'm also in San Diego, so we could have a meeting (in person) if you would be interested in a collaboration...

As Soumya and M.Eugenio suggested EMSA (titrating TF and/or binding partner) might be the best solution. I come to realize that titration is the key (and what make everything more tedious and time consuming -I'm in a rush to finish these experiments and submit-).

Heide, thanks for you response. I've been checking MST and your paper review. Can you tell me what's the difference between MST and ITC (or Thermofluor)? 

I think ITC and Thermofluor would not work because I have 3 molecules and hence 2 binding constants (assuming the protein that binds the TF does not bind DNA).  And differentiating between protein-protein binding constant and protein-DNA binding constant won't be easy.

Let me know,

Thanks again!

Sure, I'll try to point out the differences:

In MST, a temperature gradient is induced and molecules will move in this temperature gradient (it's typically about 2-3K increase, but on a very small scale of only few µm). The movement of a molecule depends on the size, charge and hydration shell, that means that a molecule will move differently if in the unbound state or bound to a ligand.

The Monolith instruments are using a fluorescence based approach, so one ligand is fluorescent (intrinsic tryptophan fluorescence or attached dyes) and used in constant concentrations whereas the non-fluorescent ligand is prepared in a dilution series around the expected Kd. The direct output is the binding affinity, but you could also calculate thermodynamic parameters via the Van't Hoff Plot.

In contrast, ITC measures the heat release or heat uptake when molecules bind to each other. One molecule is present in a measurment cell, the binding partner is present in a syringe and titrated in. Here, you also obtain information of binding affinities and doing some maths, you can deduce thermodynamic parameters.

Both ITC and MST measure binding affinities in free solution. ITC has some restrictions towards higher affinity (very low concentrated samples will not induce that much heat change) and lower affinity (solubility and limited amount of sample). MST in contrast only uses very little sample (4µl per titration point), but of course is also limited by solubility of the sample. For high affinity interactions, the Monolith NT.115Pico provides higher sensitivity in fluorescence detection and thus is able to detect pM affinites. For an ITC experiment, the buffers of the target in the cell and the ligand in the syringe need to be absolutely identical to avoid any dilution artefacts. The MST experiment can be performed in any buffer (even cell lysate and serum), the only restriction is that the composition needs to be constant for all titration points.

Thermofluor is something different. In thermofluor experiments, you denature your protein by heating. To monitor the denaturation process, dyes such as Sypro Orange are added to the protein. The dye does not fluoresce per se, but during denaturation, hydrophobic parts of the protein become surface exposed which the dye attaches to and starts to fluoresce. In this assay, you can test different buffer conditions/additives/ligands and see which condition is able stabilize your target protein.

If you would like to learn more about MST and if you are interested to test your own sample, you could join one of the MST Meetings:

http://www.nanotemper-technologies.com/news-events/events/article/3rd-north-american-mst-meeting/

Let's see if I got it straight: 

With MST you are measuring differences in fluorescence (deltaF in your Jerabek-Willemsen  et al 2014 paper, where you review this technique) at different temperatures.  That difference in fluorescence is due to the bound-unbound state of the molecules that you are introducing. 

Then, I could measure the fluorescence of my promoter (labelled with Cy5) bound to the TF and with and without the interacting partner (and using as negative control only the interacting partner and promoter -hopefully not binding at all, otherwise everything will be messed up because there will be an extra Kd). 

Is that right?

I think you got it right. I just want to point out, where the deltaF comes from:

Fluorescence is used to monitor the movement of the molecules. In the spot of heating, the fluorescence signal is detected. If the  fluorescence signal decreases during heating, molecules move away from the heated spot (this is what we mostly see for biomolecules, it's called positive thermophoresis). The movement of the molecule is different if in the unbound or bound state, so one species shows a stronger thermophoresis, the other one a weaker thermophoresis. Therefore, the difference in fluorescence which is plotted in the graph corresponds to the different movement of molecules in the temperature gradient.

Right, you could use Cy5 labeled promotor and check the affinities for TF with and without interacting partner. The negative control is also straight forward. With MST, you could also investigate the binding affinity of the TF towards the interacting partner.

Let me know if you have any further questions!

Well, my next question would be about the amount of protein and DNA I would need.

And whether you would be willing to establish a collaboration if I send the proteins and DNA ... my lab unfortunately won't buy the necessary equipment since this would be a time point experiment most likely not going to be repeated with different proteins ...

Thanks!

Cy5 is a very bright fluorscent dye, you can easily go down to 10nM. The protein concentration depends on the affinity. For MST, you will prepare a dilution series of the protein starting with concentrations around 20fold above the Kd to obtain suffient data points for a good binding curve, both in the saturation and baseline.

You can quantify the interaction between any kind of biomolecules with MST! So I'm sure you will find a huge bunge of applications which you could use MST for :-) 

Let's further discuss via email. Thanks for the discussion and for your interest!!