As you correctly stated, you need to perform a binding experiment. There are many ways to do it that give a value for association constant. The most informative in terms of thermodynamics is isothermal titration calorimetry, which gives deltaH and deltaS values.

As you correctly stated, you need to perform a binding experiment. There are many ways to do it that give a value for association constant. The most informative in terms of thermodynamics is isothermal titration calorimetry, which gives deltaH and deltaS values.

Try Microscale Thermophoresis?

http://www.sciencedirect.com/science/article/pii/S0022286014002750

Binding experiment with an active enzyme and substrate gives always the Km value, which is not equal to Kd.. 

Vello Tougu makes a good point. If you have a single-substrate enzyme, you can't study the equilibrium binding interaction between the substrate and the enzyme directly because the reaction will occur and the substrate will be consumed. Such experiments can be useful, however, for some multiple-substrate enzymes to measure the affinity of one of the substrates when that substrate can bind with no reaction occurring. Whether this is feasible depends on the kinetic mechanism of the enzyme.

For some single-substrate enzymes, for which the substrate is in rapid equilibrium with the enzyme (in other word, the binding rate is much faster than the reaction rate, Km=Kd. In those cases, you can measure Kd by doing a simple Michaelis-Menten plot. The problem is to be sure that the rapid equilibrium condition is met.

For 2-substrate enzymes with an Ordered kinetic mechanism, the Kd of the first substrate to bind can be determined by Michaelis-Menten kinetics. For 2-substrate enzymes with a Rapid Equilibrium Random kinetic mechanism, the Kds of both substrates can be determined this way. As above, the problem is to know the kinetic mechanism.

Dear Sergii

If I were you, I would prefer to use the computational biological tools (or you may call these as bioinformatics tools) to find the accurate binding energy. To do the esame experimentally, you need to determine the equilibrium and god bless you if if it a Rapid equilibrium condition. It is very difficult to get reproducible results on repetition.

Personally, I am very skeptical that an accurate measurement of binding energy of a ligand with a protein can be computed.

I completely agree with Dr. Shapiro. The best option is ITC. Other experiments involving Km and Kd estimation are not really accurate to quantify binding energy.

Other options (nor really accurate) are those where yo can check the activity at different temperatures in presence of different stabilization agents such us salt or glycerol (you get the activation energy involved in the reaction catalysed by your enzyme).

The previous answers point out the unreliability of using Km values as denoting dissociation constant but if you have a good competitive inhibitor, estimates of Ki values (likely to be dissn constants) for the the wt and mutants, the relative deltaG values calculated from the Ki values may provide useful comparative data for the substrate analog binding energies of the various forms of the enzyme.