The Path to Calorimeter Mastery Vol.4: Let’s Optimize ITC Experiments! Continuation

Next, let’s try measuring proteins and compounds. Since the affinity is initially unknown, let’s measure proteins at 10 μM and compounds at 100 μM in the usual way.

It’s not forming a sigmoidal curve. Did I set the concentrations wrong… The dilution heat of controls is almost the same for each titration. The DP value is 0.015 μcal/sec, which is small, isn’t it? Moreover, it’s smaller than the heat in the last titration of the sample, indicating that the heat generated by the sample is very small. Can I just subtract this directly…?

Mr. Nakamura, you seem to be in trouble.

Professor Fukada, indeed. I measured the proteins and compounds, but it seems like I’m not getting a sigmoidal curve, and the heat obtained is very low.

Is that so? Mr. Nakamura, did you try using the simulation software before measuring?

Oh, the one included with iTC200 control software you mentioned earlier. Sorry, I haven’t used it…

It’s mentioned in the manual. Why don’t you check page 39?

Alright, I’ll check it out.

Also, the sample measured this time seems to have relatively weak affinity.

Right. I think it’s likely around a few μM.

How did you set the concentrations?

I tried with proteins at 10 μM and compounds at 100 μM.

Then, what happens if you simulate with a presumed K_D of 5 μM?

Ah, it’s drawing a sigmoidal curve. However, the concentration is set higher than I used.

Yes, this is the concentration recommended by the simulation software. What if you simulate using the concentrations you used, Mr. Nakamura?

Protein at 10 μM, compound at 100 μM…

Oh! It’s not drawing a sigmoidal curve!!

That’s right. If you are aware of the expected K_D as in this case, it might be good to simulate and verify before attempting the actual measurement.

I see. The simulation software is convenient!

Here’s a question: This is data measured by ITC for Bovine Carbonic Anhydrase interacting with small molecules Sulfanilimide and AMBSA. What do you think when you see this data?

The diagram for Sulfanilimide somewhat forms a sigmoidal curve, but the diagram for AMBSA appears linear, not forming a sigmoidal curve.

Exactly. So, how do you think you could make it form a sigmoidal curve?

Hmm… Earlier, I couldn’t get a sigmoidal curve because of lower concentrations, so should I increase the concentrations?

Exactly. Incidentally, K_D for Sulfanilimide is about 8 μM and for AMBSA, 10 μM is said.

I should try simulating.

In the concentrations in the previous diagram, it was not possible to draw a sufficient sigmoidal curve, but in the simulation results, the cell concentration is set relatively high. What should be done if this concentration cannot be prepared?

Mr. Nakamura, do you remember the setting range for the C-value?

Setting range?

Please check the manual. What range is desirable to fall within?

Ah! It was ideally 5-250, and 1-1000 as acceptable range (please refer to pages 7-8 of the Japanese simple manual).

Yes, that’s right. You can also change the C-value in the simulation software. Furthermore, you can simulate how it will be when actually measured with the prepared sample concentrations.

I see! It’s about experimental design.

That’s correct. Then Mr. Nakamura, since you might be measuring compounds with even lower affinity in the future, let me tell you the tips for those measurements and analyses.

Yes! Please, go ahead!!

What do you think when you see this data, Mr. Nakamura? It’s the interaction between protein and fragment…

Since it’s not forming a sigmoidal curve, should I increase the concentration and try measuring?

Supposing this experiment was carried out with a cell protein concentration of 30 μM and fragments at 3 mM, what happens?

What? Is the fragment concentration that high? Compared to that, the cell protein concentration isn’t that high…

Let’s go back to the manual. Please see page 9. It’s about designing an experiment assuming an affinity weaker than 100 μM.

It’s not forming a sigmoidal curve.

Yes, for very weak affinities, there could often be a difficulty in setting concentrations to form a sigmoidal curve. Could you simulate it?

Wow! It’s impossible to have 1 mM of protein!!

That’s right. Therefore, in such cases, as shown in the manual, it’s necessary to set protein concentrations between 1/5 to 1/100 of the expected K_D, while ensuring at least 10 μM, and titration samples between 10 to 50 times the K_D at a few mM to ensure the sample is fully saturated in the later stages of titration. Look at the x-axis of the normalized plot data of the molecular ratio from earlier. You see there’s a big value, right?

Yes, but it doesn’t form a sigmoidal curve…

That’s true. With the current state, neither binding ratio nor enthalpy change can be determined. Here, a little ingenuity is required in the analysis.

A little ingenuity?

Yes, indeed. As written in the manual, analyzing with a hypothesized binding ratio is the way to go. If the sample you’re measuring shows a 1:1 binding, theoretically, you can assume the binding ratio is “1”. Do you understand?

Well, theoretically…

The binding ratio corresponds to the midpoint of the sigmoidal curve, so by inputting “1” as a constant in the initial value when fitting, you can perform fitting. When the Fitting Session window appears, input “1” for the N value and uncheck “Vary?” to set it as a constant. As a result, the software can use a part of the received curve’s sigmoidal pattern to simulate the sigmoidal curve, eventually determining K_D and ΔH.

I see!

However, it assumes the sample concentration retains 100% activity, so be careful.

Yes!

Thanks to this, Mr. Nakamura could perform the actual sample measurement, learning about the sample preparation, measurement conditions, and analysis methods that suit the nature of each sample. When irregular reactions happen, instead of forcing the analysis, think about what it should normally be like, and consider why it didn’t turn out that way.
Next time, we plan to introduce how to interpret and address irregular data.

[Blog Postscript]
What do you think is the most frequently asked question about ITC? The answer is: “solution doesn’t flow during cleaning of the titration syringe.” Actually, we address this issue in Volume 2 of the Path to Calorimeter Mastery.

Damaged glass parts of the titration syringe, poor connection of fill port adapter, and deterioration of O-rings are mentioned, but recently, another reason for malfunction has been observed.

That is, clogged filter on the back of the Washing Module.

Replacing this filter can sometimes solve the issue. If following the instructions in “When cleaning water doesn’t flow into MicroCal iTC200 titration syringe,” doesn’t alleviate the issue, try replacing this filter. Note, the filter is included with the system upon delivery.

However, depending on when the system was delivered, this filter might be inside the Washing Nodule, not visible as in the picture. In such a case, please contact our technical representatives if you find it difficult to replace on your own.