Hi, Prem!

Are You supposing that You see only free NADH fluorescence, and NADH binded to protein has no fluorescence? (or NADH binded to protein has some fluorescence level lower than free NADH?)

"there is decrease in the flourescence maxima (intensity) after addition of the ligand and then I reached to the saturation" - it's on the curve versus time or versus NADH concentration?

If you were measuring the effect of NADH binding on the intrinsic tryptophan fluorescence of the protein, then it is essential that you correct for the inner filter effect caused by NADH at every NADH concentration (as long as the NADH concentrations were high enough to cause significant absorbance), because NADH absorbs maximally at 340 nm, about the same wavelength that tryptophan emits.

In order to make this correction, repeat the titration, but replace the protein with N-acetyltryptophanamide (NATA) at a concentration sufficient to get a good fluorescence signal. Use the same excitation and emission wavelengths and the same experiment geometry as for the protein. The NATA fluorescence intensity will decrease as the NADH concentration increases due to the inner filter effect, but not due to binding, as with the protein. Calculate a correction factor at each NADH concentration as the ratio of the control (no NADH) fluorescence to the NATA fluorescence in the presence of NADH. Multiply the measurements in the protein titration by these correction factors.

An added complication could be NADH fluorescence. Check for that by titrating just the NADH alone. Subtract the NADH fluorescence, if any, from all measurements.

Once all the corrections are done, fit the binding data to a binding isotherm equation using nonlinear regression. If the Kd is much higher than the protein concentration, you can use the simple equation

y = yo + a[L]/{Kd + [L]}

where y is the fluorescence intensity measured at ligand concentration [L], yo is the fluorescence intensity when [L]=0, and yo+a is the fluorescence intensity at saturation.

However, if the Kd is not far above the protein concentration, then you should use the  Morrison equation, which is a version of the quadratic formula.

If there are multiple interacting sites for NADH, then the possibility of binding cooperativity exists and a more complex analysis is needed.

Dear Vladimir

"it's on the curve versus time or versus NADH concentration ? " 

It is actually flourescence intensity vs NADH conc curve. 

Dear Prem,

If you are monitoring NADH fluorescence emission, you should not be adding it over the protein because you will measure two effects, the enhanced emission resulting from the concentration increase, and the variations in emission resulting from the binding to the protein. 

You want to keep the concentration of your emissive species constant, and add the other molecule over it. In that case, NADH in the cuvette and add the protein over it. This is a good review with practical tips; although its oriented to supramolecular chemistry, it is applicable to any other type of interaction.

Thordarson, P. (2011) Determining association constants from titration experiments in supramolecular chemistry. Chem. Soc. Rev. 40, 1305–1323