MicroCal VP-ITC User Manual page 49

Section 3: Running an ITC Experiment
that you plan to use. For c values larger than ca. 10, the final concentration of ligand in the cell
after all injections are completed should be ca. 1.5 times the total concentration of macromolecule
binding sites in the cell at the beginning of the experiment,
i.e., X tot x ∆v/V = n x M tot x 1.5,
Where ∆v is the total volume of injectant to be used, V is the cell volume (ca. 1.3 ml), and n is the
ligand/macromolecule stoichiometry. For cases where n=1, the ligand concentration, X tot , ,
should be ca. 8.5 times M tot using a 250 µl syringe. If the c value for your system is lower than
10, you may wish to increase the final ligand/macromolecule ratio from 1.5 up to 2.0 or even 2.5
as is evident by referring back to the previous figure showing binding isotherms as they depend
on c. It should be realized however that accurate curve fitting is possible even when saturation of
sites is not achieved.
There may be other factors, specific to your system, that are important considerations in
experiment design, such as the total amount of macromolecule or ligand that is available for the
experiment and/or solubility restrictions on the macromolecule or the ligand.
Several other experimental design problems should be mentioned. First, the buffer in which the
ligand is dissolved should be an exact match (i.e., pH, buffer concentration, salt concentration,
etc.) to the buffer in which the macromolecule is dissolved, or else large spurious heat effects
from buffer mixing will result. For example, if the ligand is dissolved directly into the buffer
which was dialyzed against the macromolecule solution, the exact pH of the ligand solution may
change due to titration of ionizable groups on the ligand. If this happens, then the ligand solution
should be back-titrated carefully until the pH is identical to that of the macromolecule solution
before doing the experiment.
Second, control experiments (i.e., ligand solution added to buffer in cell without the presence of
macromolecule) to determine the heat of dilution of ligand should be carried out in the same way
as the experiment with macromolecule present, and these heats of dilution should be subtracted
from the corresponding injection into the macromolecule solution. You will usually find these
heats of dilution to be small and frequently negligible (unless the ligand dimerizes or aggregates
with itself!) but they should be checked as a precaution.
Finally, you may find occasionally that the first injection in a series of injections shows a smaller
heat effect than it should. This can result from bending the syringe needle a little when seating the
injector into the barrel, or leakage resulting from having the syringe in the cell a long time before
the first injection is made (particularly if it is stirring all the while). It you find this to be a
persistent problem with certain systems, even when care is taken to avoid the aforementioned
factors, you may wish to make a small first injection (e.g. one 1 µl injection followed by ten 10 µl
injections) and then delete the first data point before doing curve-fitting in Origin.
Because of release or uptake of protons during many biological binding reactions, the observed
heats of binding may be strongly dependent on which buffer is used. In fact, certain binding
reactions which have extremely small ∆H and produce virtually no signal in buffers with small
∆H ion (e.g., phosphate) can sometimes be studied nicely in buffers with a large ∆H ion (e.g., tris)
where the signal will be much larger.
39