- Policy & Practices
After creating our models we needed to analyze the results that it generated using mock parameters, and see whether we can explain everything that is shown. Note that this page does not actually use our acquired data, but simply the theoretical results of mock parameters and parameters found in literature. We do this here to focus on what the model predicts instead of analyzing our data.
In order to determine the mock parameters we will use in this section, we look at the available literature.
The Kd,C1 parameter is based on CT52 on its own, without any proteins fused to it. This makes it an acceptable approximation for a CT52-smallBiT fusion, since smallBiT is only 11 amino acids long. Sadly this approximation is not accurate enough for CT52-caspase9 fusion, since caspase9 is much larger than CT52 and will likely cause steric hindrance, effectively decreasing the binding affinity. For this we use the dissociation constant for CT52-caspase9 (Kd,C1 = 0.25 μM) and acooperativity
The cooperativity represents the influence of the first CT52 in the T14-3-3/FC2/CT52 complex on the binding of the second CT52. This can be a positive influence (σ > 1), negative (σ < 1) or neutral (σ = 1)
σ = 121.52 and the Kd for two CT52-caspase9 proteins (Kd,NoB = 1.64 μM)2. For the Kd for the CT52-NanoBiT proteins we assume the value found by Dixon et al.3
For all of our assays where we use NanoBiT to measure our protein activity concentrations of active complex can be determined with the use of a calibration curve. However, the purpose of our caspase-9 assays is to determine the effect of the specific activity of the caspase-9 on our homotetrameric scaffold compared to the homodimer. To calculate this activity we also need the Michaelis-Menten constant (KM = 650 μM)2, the catalytic rate (kcat = 0.2 s-1)2 and an initial substrate concentration S0 which depends on the experimental parameters.
When the model is executed, the program produces a graph that shows the concentration of the several complexes in our system as a function of the total scaffold concentration. Such a graph is shown in figure 1, containing the parameters mentioned above and some mock parameters.
The graph above was produced using 2 different values for the CT52 binding affinity to the T14-3-3/FC complex, one (C1) has the value mentioned earlier, the other (C2) has the arbitrary value of 1μM, The binding affinity of NanoBiT was used for background activity (Kd,NoB). For now we simulate with both CT52 concentrations at 50 nM and FC concentration at 70 μM. Since FC binds weakly to T14-3-3 on its own, it needs to be in excess. Note that in the legend the scaffold is abbreviated as B, fusicoccin as F, and the CT52 as C, with numbering if necessary.
The line BFFCC (or E) shows bell phased curve originating from thebiphasic effect
The biphasic effect is that when very little scaffold is present compared to the binding partners, there is little activity compared to its peak activity, which is to be expected, But the exceptional thing of scaffold proteins is that when there is very much scaffold present compared to the binding partners, activity will also decrease. This decrease happens because there are a lot of scaffolds that bind one of the binding partners, and cannot not bind a second because it is already bound to another scaffold.
that is inherent to scaffold proteins, it is slightly skewed towards higher concentrations due to the large positive cooperativity of the CT52 proteins and the peak is also at higher concentrations than the theoretical ideal ratio (2:1), instead it is at about 7:1.
The Graph also shows that at these low concentrations the background activity (CC) is negligible. The effect of the difference in binding affinities of the CT52 is also clearly visible by the red and blue (C1) and the green and cyan (C2) lines.
In order to determine the effectivity of a tetramer compared to that of a dimer, we will look at the specific activity of the CT52-proteins. For this purpose we have written a model for the formation of complexes in the homotetrameric system and for the homodimeric system. Developing a model for the heterotetrameric system has no added value compared to the heterodimeric system, as its purpose would be to determine the affinities of our mutated proteins, which our heterodimeric model can already do.
For the homodimeric and –tetrameric systems we use CT52-caspase-9 fused proteins, as these also form homodimers and thus are suitable for homoscaffolds. For this purpose we will use the alternative parameters discussed earlier.
As can be seen in figure 3, at smaller concentrations (left) higher order complexes like the 3 and 4 CT52 bound complexes are negligible, only at high concentrations (right) these become relevant. This hints that at low concentrations of binding partners the tetramer has no significant effect compared to the dimer. The graphs also show that the total enzymatic activity (the dotted line) is dominated by the background activity for low scaffold concentrations. What is remarkable is that the ratio CT52:scaffold is lower than the theoretical ideal ratio (4:1), and that at this point the background activity barely contributes to the total activity.
Figure 4 shows us that at low concentrations of CT52-caspase9 the specific activity has a higher peak for the dimer than the tetramer, showing that the tetramer is a less efficient structure at these concentrations and thus maybe less preferable for an enzymatic reaction using homodimeric proteins, however It does show lower specific background activity, indicating it is better able to suppress slight variations in concentrations.
As with low CT52 concentrations the specific activity of the CT52-caspase9 in the tetrameric system is lower than in the dimeric system. But what is very obvious, is that it suppresses the background much better than the dimer as the background activity of the dimer is far higher than the (just about visible) peak of the actual activity of the dimeric scaffold, where the tetramer stays somewhat stable for every concentration of CT52.
Most of these models predict behavior similar to previously observed systems, showing naturally occurring effects such as the biphasic effect. Although these models alone are not enough to conclude any accurate values of the properties, such as binding affinity or cooperativity, of the proteins in question, they do provide a guideline for the design of our experiments and can be used to determine our protein properties from the data received with those experiments.
-  Würtele, Martin et al. “Structural View of a Fungal Toxin Acting on a 14-3-3 Regulatory Complex.” The EMBO Journal 22.5 (2003): 987–994. PMC. Web. 29 Sept. 2016.
-  A. den Hamer, L.J.M. Lemmens, M.A.D Nijenhuis, C. Ottmann, M. Merkx, T.F.A. de Greef, L. Brunsveld, ‘manuscript submitted for publication’, 2016
-  Dixon, A. S., Schwinn, M. K., Hall, M. P., Zimmerman, K., Otto, P., Lubben, T. H., ... & Wood, M. G. (2015). NanoLuc complementation reporter optimized for accurate measurement of protein interactions in cells. ACS chemical biology,11(2), 400-408.