Here are several best practices for performing traditional binding experiments with the MicroCal PEAQ-ITC, VP-ITC and iTC200 systems. This is a continuation of previous blogs ‘Best Practices for Isothermal Titration Calorimetry to study binding interactions’ part 1 and part 2.

Run regular performance and validation checks on your ITC

  • Water-into-water injections
  • ITC test kit with RNase and EDTA
  • Another reference binding assay

Prior to binding experiments, perform control titrations whenever possible

  • Check for heats of dilution and other sources of heat change
  • The typical control titration is the ligand in the ITC syringe, titrated into matched buffer, using the same ITC experimental parameters as the binding experiment
  • The heat changes for the control titration should be small, reproducible, and like the heat changes at the end of the binding experiment (Figure 1, top)
  • Other control titrations are buffer-into-buffer, and buffer-into-macromolecule in the ITC cell

Figure 1: (top) Raw ITC data, with the control titration (ligand into buffer) (red), and the binding experiment (ligand into macromolecule) (black). Note that the peaks for the control titration are similar in shape and size to the final injections of the binding experiment.

(Bottom) Normalized heat changes, with the control titration (red), and the binding experiment (black). Data are fit to the one set of sites binding model.

Evaluation of ITC raw data: what does the raw date for an ideal ITC experiment for a (1:1 binding event) look like

See figures 1 and 2.

  • Titration data plot begins as a series of approximately equal, large peaks (exothermic or endothermic), representing close to 100% binding of the ligand to the macromolecule
  • Heat changes become reduced as the binding sites are saturated with further injections of ligand during the experiment
  • The final peaks, corresponding to the heats of dilution of ligand after binding is fully saturated, should be small and reproducible (Figures 1 and 2)
  • This kind of titration curve will result in a sigmoidal binding isotherm (Figure 1 bottom)
  • Need sufficient heat change above baseline noise for high-quality ITC data
  • Heats of injection (integrated area):
    • For PEAQ-ITC and ITC200: >2.5 ucals for the second (first full) peak is ideal
      • ~1 ucals for second peak is minimum heat for data analysis
    • For VP-ITC: >10 ucals for the second (first full) peak is ideal
      • ~3-5 ucals for second peak is minimum heat for data analysis
    • Below this heat detection range, data can be noisier due to lower signal/noise ratio
  • The baseline position (in ucal/sec) for the raw ITC data should be within 1 ucal/sec of the reference power setting. If the difference is greater than 1 ucal/sec, that can indicate a dirty ITC cell, and/or bubbles in the reference or sample cell
  • All the DP values (Y axis) for the raw ITC data should be above 0 ucal/sec
  • The position of the post-injection baseline should be the same as the pre-injection baseline. If not, there may not be enough time between injections, or there is another process causing a change in DP such as enzymatic hydrolysis
  • A slight drift in the baseline throughout the titration is normal. If there is a good integration baseline, the data are OK (Figure 2). You do not want large jumps in the baseline, or “steps

Figure 2. High-quality ITC data. Note the slight drift in the integration baseline (in red) – this is acceptable.

More information about the N value from ITC curve fitting

N is associated with the total macromolecule concentration and is one of the parameters that is determined during data analysis. It is important to understand that the ITC N value is NOT always the same thing as the stoichiometry (binding ratio) of ligand binding to macromolecule. N value does in fact equal the stoichiometry if the concentrations we use for the fitting are correct and 100% active.

Another way to express the N value is this equation:

N = St x [AFcell/AFsyr]

Where St is stoichiometry (number of binding sites the macromolecule has for the ligand), AF is the “active fraction” of the biomolecule (the active concentration divided by the total concentration), AFcell is the active fraction of the sample in the cell (typically the macromolecule), and AFsyr is the active fraction of the sample in the syringe (typically the ligand).

Any of the three values in this equation can change N, which means that N is a measure of the binding ratio as much as the activity of the sample(s). Two of these values must be known (or assumed) to get the third one. For example, if one is interested in the stoichiometry then the concentrations provided in the fit must be accurate and 100% active. If one is interested in the activity of the sample in the cell (often the macromolecule), then the stoichiometry and the active concentration of the sample in the syringe (ligand) must be assumed.

In the PEAQ-ITC data analysis software package, you can also fit the data by fixing N to 1 (or the desired value) and varying the cell or the syringe concentration.

If N is lower than the expected stoichiometry, your active macromolecule concentration in the cell is less than the total concentration, and/or your active ligand concentration is higher than the total concentration.

If N is 0.5, it is possible that your macromolecule (in the ITC cell) is a dimer and binds one ligand per dimer. It is also possible that your biomolecule in the syringe has 2 binding sites, and the sample in the cell has one binding site, so you should try to fit the binding curve using the “ligand in cell” option.

If N is higher than the expected stoichiometry, your active macromolecule concentration in the cell is greater than the total concentration, and/or your active ligand concentration is lower than the total concentration.

Errors in the ligand concentration in the syringe results in errors of measurements of N, KD and ΔH, and therefore the calculated free energy (ΔG) and entropic (-TΔS) contribution to binding. Errors in macromolecule concentration in the cell primarily affects the accuracy of the determined N value.

Relevant content

Previous posts