Stirring speed should not be very high, to avoid bubble formation (cavitation) and high noise level. At the same time, it should not be too low, otherwise homogenization of the cell content after injection would be inefficient and heat transfer would be very slow (no convection, only conduction).

Having said that, you may select a stirring speed from 0 to ~500 rpm. If you select 0 rpm (no stirring) the peaks will be broad and small. I would try ~100 rpm to see what happen.

By the way, do you incubate the protein+ligand in the waterbath, or only the protein? Do you incubate for the same time as in an ITC experiment?

First of all thank you.

I will try to do the experiment at low speed.

I have done assay with the protein at 370C. with substrate and have not seen any precipitation but the volume and protein concentration are low in the assay condition (20ul and 10uM resp.) than the the ITC experiment (200ul, 40uM).

Even if I titrate only buffer in ITC experiment (keeping protein (40uM) in the sample cell) I see precipitate. So in water bath I had incubated protein in buffer (without nucleotide) for the same time (around 55 min.) as the ITC experiment.

You can also change pH or ionic strength. Your protein at a pH close to the isoelectric point is prone to aggregation. Stirring may just accelerate the process.

The proposal of Armando is very good. Consider that stirring is necessary for both to maintain homogeneous temperature but most of all to generate rapid mixing for fast peak return and equilibration of base line before next ligand injection. Each system is different regarding aggregation, and protein concentration is an important variable for protein-protein interactions. Stirring allows molecules to encounter others and the shear effect also perturbs their solvation shell exposing some hydrophobic groups. So the model proposed by Armando might also benefit from repeating the experiment at different concentrations. ITC requires a lot of protein so a single experiment before saturating conditions changing stirring speed at 50, 100 and 200 might be enough to determine the optimal value, looking at the peak return to base line. Also, you might change the time interval between injections to deal with broader peaks due to low stirring speed. Of course, try to use the lowest concentration possible. Good luck!

I was thinking again in your problem Pratik. If the ligand has ionizable groups, the ligand may cause a change in the isoelectric point of the protein-ligand complex respect to the isolated receptor. Also if the ligand induces a conformational change in the protein this could make it prone to aggregation? These could be causes of precipitation different to stirring. In the first case a pH change may help, In the second case lowering the concentration is the trivial recommendation or using a lower ionic force to reduce hydrophobic effect. Also the use of an additive such as sugars or detergents at different concentrations and extrapolating the results to zero could be a solution. Greetings.

Hi,I assume that you perform ITC experiment to get information regarding stoichiometry of binding. Because you may not know the stoichiometry, the concentration of ligand that you add at the end of ITC titration experiments might be much higher than the stoichiometry of binding which may cause aggregation. I suggest that if possible you do a titration experiment using UV-visible spectroscopy to follow aggregation of your protein upon addition of ligand. If you did not observe aggregation, then you can ajust concentration of protein and ligand in ITC and you can also play with ITC settings. Because your protein might be valuable and it might be expensive to use it for finding the lowest stirring speed that you can use, i suggest that you use dilution of salt in buffer. You can titrate a buffer that contains salt into a buffer that does not have salt.

best of luck

Stirring is necessary for homogeneization ( solution and heat). The procedures suggested in the answers given above are very good. 

Hello Everyone,

I would like to share this with you all that the precipitation problem is now solved by supplementing buffer with detergent 0.05% triton-x100. The stirring speed  I am using is 500RPM. There is no visible precipitate anymore.

The N value I am getting is near 1 (N= 0.6). So I think I should crosscheck the accuracy of protein and ligand concentration in order to improve N value. Should the  N value be exactly 1? how much error is acceptable? 

Any other suggestions are welcome. 

Thank you.

Dear Mr Pratik Patil,

I am also facing similar problem. As mentioned by you addition of Triton X100 (0.01 %) solved the problem with precipitation of protein in ITC cell but it is taking too long for the initial equilibriation so I had to force start my titration. I am having water in my reference cell. is that a reason? What did you fill in the reference cell in your experiment?