The experiment was run at 6 μcal/s and the raw data show a
very small decrease in this baseline level. It is recommended that the
raw data is presented in this way so that any drift or acute changes
(steps) in the course of the experiment can be seen. In some
software, the generation of the baselines required for the integration of the peaks gives the opportunity to set all the baseline at a
“zero level” making the whole data collection look perfectly flat at
zero differential power. This type of representation should be
avoided.
These data have been fit using Malvern’s PEAQ software. This
program automatically integrates the excess heat effect from each
injection and normalizes this to the number of moles of ligand
added during each injection giving molar heats in kcal/mol. The
integration was checked manually for each peak and baselines
adjusted if there was an obvious inconsistency. The integrals are
plotted against the molar ratio of [ligand]/[protein] at the end of
each injection and the resulting binding isotherm can be fit to a K d
value of 6 μM, an enthalpy of À13 kcal/mol with stoichiometry of
0.92. The small heats at the end of the titration (high molar ratio)
represent the background control heat arising at each injection
when the protein is fully saturated with ligand and can be included
as a fitted parameter. Preferably, the end point can be confirmed by
measuring separately in a control experiment where the identical
trisaccharide solution used is injected into buffer alone (see Note
12). The heats determined in this control experiment can be subtracted from the original binding integrals, as has been done here,
and the corrected data fit with a zero end point (see Fig. 2 lower).
Where the experimental data do not reach saturation of all binding
sites on the protein, it becomes more important to perform such a
control since the fitted end point is being extrapolated to values
that are not measured directly in the binding experiment.
3.3 Are
the Concentrations
of Protein and Ligand
Optimal?
The sigmoidal shape of the ITC binding titration can be defined by
a unit-less parameter, the “c-value,” which expresses how far the
protein concentration in the cell is above or below the K d value
(c ¼ [protein]/K d , in Fig. 2 where [lysozyme] was 37 μM c ~ 6). It
is possible to simulate data with the same K d over a variety of
experimental c-values as shown in Fig. 3.
Experiments performed at high c-value >250 (cell concentration far above K d ) essentially generate a step function plot. All the
added ligand becomes bound until all sites on the protein are filled
at a molar ratio of 1, and then no binding occurs and the control
end point heat is observed. In this case, the enthalpy and stoichiometry of the interaction are very well defined but the K d cannot be
quantified other than concluding it is significantly below the cell
concentration. Working in this c-value regime is known as a stoichiometric titration. At intermediate c-values between 5 and
250, the plots are more sigmoidal and K d can be determined with
Isothermal Titration Calorimetry
143
very small decrease in this baseline level. It is recommended that the
raw data is presented in this way so that any drift or acute changes
(steps) in the course of the experiment can be seen. In some
software, the generation of the baselines required for the integration of the peaks gives the opportunity to set all the baseline at a
“zero level” making the whole data collection look perfectly flat at
zero differential power. This type of representation should be
avoided.
These data have been fit using Malvern’s PEAQ software. This
program automatically integrates the excess heat effect from each
injection and normalizes this to the number of moles of ligand
added during each injection giving molar heats in kcal/mol. The
integration was checked manually for each peak and baselines
adjusted if there was an obvious inconsistency. The integrals are
plotted against the molar ratio of [ligand]/[protein] at the end of
each injection and the resulting binding isotherm can be fit to a K d
value of 6 μM, an enthalpy of À13 kcal/mol with stoichiometry of
0.92. The small heats at the end of the titration (high molar ratio)
represent the background control heat arising at each injection
when the protein is fully saturated with ligand and can be included
as a fitted parameter. Preferably, the end point can be confirmed by
measuring separately in a control experiment where the identical
trisaccharide solution used is injected into buffer alone (see Note
12). The heats determined in this control experiment can be subtracted from the original binding integrals, as has been done here,
and the corrected data fit with a zero end point (see Fig. 2 lower).
Where the experimental data do not reach saturation of all binding
sites on the protein, it becomes more important to perform such a
control since the fitted end point is being extrapolated to values
that are not measured directly in the binding experiment.
3.3 Are
the Concentrations
of Protein and Ligand
Optimal?
The sigmoidal shape of the ITC binding titration can be defined by
a unit-less parameter, the “c-value,” which expresses how far the
protein concentration in the cell is above or below the K d value
(c ¼ [protein]/K d , in Fig. 2 where [lysozyme] was 37 μM c ~ 6). It
is possible to simulate data with the same K d over a variety of
experimental c-values as shown in Fig. 3.
Experiments performed at high c-value >250 (cell concentration far above K d ) essentially generate a step function plot. All the
added ligand becomes bound until all sites on the protein are filled
at a molar ratio of 1, and then no binding occurs and the control
end point heat is observed. In this case, the enthalpy and stoichiometry of the interaction are very well defined but the K d cannot be
quantified other than concluding it is significantly below the cell
concentration. Working in this c-value regime is known as a stoichiometric titration. At intermediate c-values between 5 and
250, the plots are more sigmoidal and K d can be determined with
Isothermal Titration Calorimetry
143