I am currently doing enzymatic inhibition assays and in my current setup I observe that a commercial inhibitor is showing inhibition as expected. Still, the inhibition is not as strong as the literature claims. Ic50 is ~20x higher. I am wondering if the glycerol, ca. 8 percent from my protein stock may slow down the kinetics that determine the binding of the inhibitor to the protein? Is this a known problem? If so, I could overcome this by longer pre incubation time? In the future I want do a complete buffer exchange from storage conditions to assay conditions anyway to eliminate glycerin and all other components that are not supposed to be in the final buffer by a spin column, to eliminate this possible source of deviation. Still, at this stage my interest was, if the assay will work in my lab at all as a proof of concept (which it does).
-
$\begingroup$ Is it 8% glycerol by volume or mass? Can you provide more detail about the volumes mixed, and perhaps about what's in the sample. $\endgroup$– Buck Thorn ♦Commented Feb 12 at 8:39
-
1$\begingroup$ It is by volume. The sample is quite complex and contains buffer salts, a surfactant (DDM), DTT, two proteins, two substrates and one peptide. The two proteins are injected into the matrix without substrate and they contain 52 Percent glycerol each, their volume about 1 and 9 microliters working in 60 microliters in total on 384 well plate. $\endgroup$– raptorlaneCommented Feb 12 at 9:15
2 Answers
I am wondering if the glycerol, ca. 8 percent from my protein stock may slow down the kinetics that determine the binding of the inhibitor to the protein?
When you change the solvent, a lot of things change, including the kinetics of a reaction and the equilibrium constants of ligand binding. What the exact reason is for the discrepancy of your observations and the literature value is impossible to say without a full set of experiments. However, it is much easier to remove the glycerol and try to reproduce the conditions used in the literature.
Is this a known problem?
The viscosity of 8% glycerol is very different from that of water. You would expect everything to slow down (both inhibitor and substrate binding/dissociation). I don't think there is a general rule of thumb that inhibitors are less potent in the presence of glycerol. There is no obvious theoretical reason why pre-incubation would change the situation.
-
$\begingroup$ I did the same experiment without and with additional 10 min preincubation time. Before I almost did not see any inhibition at all but the increased time had a great effect. So, I would assume this is a plausible working hypothesis at the moment. $\endgroup$ Commented Feb 11 at 18:13
-
1$\begingroup$ The result might be about macroscopic mixing. If your inhibitor is sitting in one corner of the vial and the enzyme and substrate in another, the inhibition will be minimal. Also, even with pre-incubation, you don't reproduce the literature values, so it is not the entire explanation. With the scarcity of data, there are multiple plausible working hypotheses. I'm assuming you don't want to run your assays in high glycerol in any case, so removing it is annoying but probably necessary. $\endgroup$– Karsten ♦Commented Feb 11 at 19:52
-
$\begingroup$ I would think so, and I can visually see the smearing of the glycerol when the protein is injected to the wells. I mix every well with the pipette tip until the smearing disappears, but obviously I can't tell by visual inspection if all density and viscosity gradients are homogenized. Preincubation takes place on a plate shaker. I know when I mix the protein after concentration of the äkta fractions with glycerol it takes some minutes after the mixture is homogenized. Also, more importantly without inverting of the falcon tube there are two layers that keep separated. So, definitely a factor. $\endgroup$ Commented Feb 11 at 21:43
Reverse engineering the composition of your sample: here 8% is ambiguous, but assume it means the volume ratio of the original substances is 8:92.
Glycerol has density ~1.261 g/cm3 ($\pu{20 ^\circ C}$), and water ~1.00 g/cm3, so the implied glycerol content by mass is 9.9%, and the dynamic viscosity of the mixture can be computed with an online calculator as 1.29 cP ($\pu{20 ^\circ C}$). The viscosity of water is 1.00 cP ($\pu{20 ^\circ C}).$
A ~30% increase in viscosity will not have an extravagant effect on the diffusion coefficient. Assuming a Stokes model ($D = k_\mathrm{B}T/6\pi\eta a$) $D$ will be reduced by a factor of ~23%. The diffusion coefficient is related to the mean square displacement as $\left <r^2 \right > \propto \tau \times D$ so that the time required to diffuse some rms distance is $ \propto \eta$ and will increase by ~30%. This is not going to be significant, or at least would not account for an order of magnitude or more difference in kinetics.
If the supplier ships the proteins in glycerol it is safe to assume that this is not going to have a negative effect on stability. If the explanation for the difference in the assay is due to the solvent and not, say, to faulty proteins, it might be due to specific interactions.
References
The online viscosity calculator references the following:
- Cheng (2008) Ind. Eng. Chem. Res. 47 3285-3288.
- Volk and Kähler (2018) Experiments in Fluids 59 75.
-
$\begingroup$ I expressed the proteins by myself. Glycerol is usually used to protect the protein from damaging during freezing and to stabilize specific conformations, so freezing without glycerol is often the problem, not having it. $\endgroup$ Commented Feb 12 at 9:53
-
$\begingroup$ If the proteins are supposed to be folded and you have enough you might check stability in various ways. If you already did this, next thought is what else are you doing to check the proteins are ok (functional)? $\endgroup$– Buck Thorn ♦Commented Feb 12 at 10:02
-
1$\begingroup$ I do a flouresence activity assay. The product will react with a probe that will become fluorescently active. $\endgroup$ Commented Feb 12 at 12:44
-
1$\begingroup$ I did nonlinear analysis now and the value is not that terrible far away from the literature, but the data points show some spreading. Actually, I can very well think of glycerol might be the problem. Not that it completely "destroys" diffusion, but the mixing will properly differ quite significant from well to well I would assume. $\endgroup$ Commented Feb 12 at 12:46
-
1$\begingroup$ The scattering per well is ok. But triplets deviation is too high. But the general trend works, with increasing inhibitor concentration, the activity drops as expected. I account this as mentioned to mixing and pipetting error at the moment. I will redo the experiments after glycerin removal. $\endgroup$ Commented Feb 14 at 9:42