The Road to Calorimeter Master Vol.4
Let’s Develop the Ability to Optimize ITC Experiments!
The characters and story in this post are fictional.
For technical content, we have received advice from Professor Fukada, a visiting researcher at Osaka Prefecture University.
※ At the end of this article, you can download the material.
【Previous Episode】
Mr. Nakamura from Maruban Pharmaceuticals was instructed by his boss, Mr. Tanaka, to set up a calorimeter. Under Dr. Fukada’s guidance, he was able to master the basics of measuring and analyzing with the iTC200.
【This Episode】
Having acquired basic skills, Mr. Nakamura decided to measure the antibody-antigen that he had on hand. The antigen was a protein with an approximate molecular weight of 50 kDa that he purchased, and the antibody was one he prepared. The affinity is predicted to be about a few nM from other measurement results.
According to papers, it seems that most cases use antibodies in the cell and antigens in the syringe. Calculating based on the C value (*1), if KD is a few nM, then the concentration is too low, so at least 10 μM should be ensured. With a 1:1 binding, 100 μM of the antigen is required, but for an antibody which is divalent, 200 μM of the antigen is needed. Hmm? Would it be okay to set the antibody to 5 μM and the antigen to 100 μM to save samples? Let’s try measuring at a low concentration. First, let’s start with a control experiment (*2). Referring to the conditions on P.9 of the Japanese simplified manual regarding the amount and number of drops, each should be 2 µl for 18 drops. However, the first drop should be 0.4µl.
※1 The C value is explained in The Road to Calorimeter Master Vol.2.
※2 For the control experiment, see The Road to Calorimeter Master Vol.3.
The dilution heat seems large, doesn’t it? Or is it normal? Anyway, let’s measure the sample.
There is almost no change from the control experiment?! The heat in the latter part of the measurement hasn’t converged to zero, either… We should be using buffer with the same composition, so a buffer mismatch is unlikely… Is the syringe dirty? No way, I checked the system before the measurement. Dr. Fukada~.
Mr. Nakamura, what’s the matter?
Actually, it’s my sample, antigen and antibody, but when measured, there’s hardly any difference between the two data.
Oh my, indeed. This doesn’t reveal what we’re measuring, does it? Have you considered the cause?
The DP value is fine with the set Reference Power of 5 μcal/sec, so it’s definitely not contamination of the cell. There’s an issue with the titration response, so I thought the titration syringe might be dirty, but we conducted a system check before sample measurement and there were no issues. So, I believe it to be a buffer mismatch, but I’m using the same buffer composition. So, it’s unlikely.
You just mentioned, “I’m using the same buffer composition,” didn’t you? Didn’t you dialyze the sample?
Yes, this time, the antibody was prepared by me, but I used a Carrier Free commercial product for the antigen. I prepared the antigen concentration using a buffer of the same composition as the one used to purify the antibody by gel filtration.
So, you didn’t dialyze after gel filtration purification. Did you prepare the antigen with the gel filtration buffer?
Strictly speaking, I used something prepared separately for antigen concentration adjustment. But, the composition should be the same.
Mr. Nakamura, it seems likely that the buffer composition has differed. No matter how similar the composition, there is always a slight deviation due to reagent measurement and the like. Even preparations made on the same day can undergo slight changes if the date changes. Also, you mentioned you used Carrier Free products, which were freeze-dried. Did you confirm the conditions under which they were formulated? Some may use TFA or acetonitrile, and those could remain as residue or affect pH, etc.
What!? Really!?
To match the buffer composition of both the cell sample and titration sample as closely as possible, dialysis is crucial. Especially for purchased products or those not prepared completely by yourself, it is better to dialyze them with the antibody to eliminate more uncertain factors.
Understood. I’ll dialyze and remeasure. ITC is quite delicate…
To add, “Strictly matching the buffer composition between the cell and syringe” is extremely important when conducting ITC measurements. If significant thermal reactions are observed in the control experiment or thermal reactions that change as titration proceeds, it may indicate a mismatch between the ligand solution and buffer. Naturally, one should also be cautious of inadequate methanol drying from syringe drying.
Before proceeding with experiments, be sure to check buffer matching. If you cannot eliminate trends observed in control experiment results despite careful buffer matching, it may be due to aggregation or self-association of the ligand in the syringe. In such cases, reviewing measurement conditions is necessary.
For low-molecular-weight ligands that can’t undergo dialysis, you can prepare them by dissolving the ligand directly in the dialysate after polymer dialysis. However, be cautious with samples containing residual components such as additives or salts (e.g., synthetic peptides or nucleotides). After dissolving, check the solution’s pH with a pH meter. If the ligand solution’s pH deviates more than 0.05 from the buffer pH, adjust it using a very small amount of HCl or NaOH solution. Additionally, thermal changes due to small differences in salt concentration can be subtracted through control experiments.
Also, for low-molecular-weight ligands with protons dissociating groups, even if at concentrations of a few mM, if the buffer components are around 10 mM and thus not very different in concentration, sufficient buffering might not be achieved, affecting the pH, so please be careful.
The Next Day….
Dialysis complete. Concentration adjusted! No pH deviation. Set the sample and start measurement!!
Control experiment’s dilution heat is constant. The sample measurement also converged to dilution heat in the latter half! Dialysis is crucial!! Next, let’s analyze it.
Fitting is good! KD is 6nM, and the binding ratio is 1.8.
Mr. Nakamura, you’ve got quite good data there.
Yes, I’ve re-recognized the importance of dialysis!
That’s right. By the way, how did you subtract the control data this time?
I subtracted the raw data as per the manual.
That’s right.
To supplement, the manual suggests using the raw data for the subtraction of control experiments. If dilution heat is constant, and if the control experiment has a high noise level, subtracting the average value is also acceptable. Select based on the situation.
Now, let’s have a question here. This figure shows the data from which the control raw data was subtracted, but there is a problem regarding the analysis. Where do you think it is?
Is it the bad fitting?
That’s right. Then, why do you think the fitting is poor?
Is it because the model is incorrect, or something?
No, I don’t think it’s a model error. Where does the fitting seem particularly bad to you?
Is it where it’s plateauing towards the end, perhaps?
Yes. And why do you think that?
…I’m not sure.
When performing fitting, analysis software calculates it to converge to zero in the latter part of the measurements. In viewing this data, the plot of measurement data appears above zero, doesn’t it? Meanwhile, you’ll notice the red fitting line converges to zero at the end.
Oh, really! That’s right!
So, Mr. Nakamura, why do you think the measurement result isn’t zero even though you’ve subtracted the control?
Could it be that there is a difference between the dilution heat of the control experiment and that of the sample measurement, and that the heat generated by the control experiment resulted in a positive value upon subtraction?
That’s right. Then, what do you think should be done in such a case?
Umm, I’m not sure.
When you really can’t manage to subtract the dilution heat properly and get such data, there’s a technique where you can edit the data so that the values after the sigmoidal curve converge to zero and then analyze.
A sneaky trick!? How does it work?
Using something called the Y Translate feature is convenient.
Y Translate?
It’s a tool that, as the name suggests, shifts values on the Y-axis. Please refer to the materials for how to operate it.
To download the material on how to operate the Y Translate, please fill in the required information in the form at the bottom of the article.
Yes. Thank you.
I need to get used to the analysis software too… Is there a more detailed manual on analysis?
Yes, there is. You can download it from Malvern’s homepage (English version) as a PDF file. (Registration on the Malvern website is required.)
We provide specific guidance on analysis examples. All data is saved in the delivered PC.
We also introduce the fitting formula used in the analysis, so please refer to it.
Y Translate done. Will the fitting be good…
The mismatch between the plot in the latter half of titration and fitting data has reduced!
Mr. Nakamura, you’ve deepened your understanding again, haven’t you?
Yes!
By the way, Mr. Nakamura. What buffer did you use for this measurement?
I used phosphate buffer during antibody preparation, so I used phosphate buffer for ITC measurement as well.
Is that so? Then Mr. Nakamura, here’s a question. If the pH is the same, but the buffer components are different, do you think you’ll get the same data?
What!?
For example, this time you’re using phosphate buffer, but what would happen if you use HEPES or Tris of the same pH?
Hmm. Generally, experimental results can change with buffer changes, so I think there will be some differences, but…
That’s right. Actually, there’s a hint in the Japanese simplified manual.
Really?
Please have a look at the buffer composition on P.5.
Oh, really. There is a difference in the protonation enthalpy among buffers. But in reality, how different are they?
Now that your measurements have settled down, shall we chat over tea about it? (For details, see Professor Fukada’s Column 3)
Yes! Professor Fukada! Thank you!!
