Team:TU-Eindhoven/Modeling/Rosetta Results

• Home
• Team
  • Members
  • Advisors
  • Collaborations
  • Sponsors
  • Attributions
• Project
  • Description
  • Background
  • Scaffold Design
    • Heterodimer
    • Tetramer
  • Experiments
    • Heterodimer
    • Tetramer
  • Timeline
  • Protocols
• Modeling
  • Rosetta
    • Model
    • Results
  • PPI
    • Model
    • Results
• Results
  • Heterodimer
    • Introduction
    • Functionality
    • Orthogonality
    • Conclusion
  • Tetramer
  • NanoBiT curve
  • Proof of Concept
  • Interlab Study
  • Biobricks
    • Parts
    • Basic Part
  • Achievements
  • Future
• Policy & Practices
  • Outreach
  • Safe By Design
  • Applications
    • T-Cell Therapy
    • Transport Bacteria
    • Co-Enzym Regen.
  • Safety
    • Lab
    • Project


✉

Results

Designing proteins using a molecular structure predictor.

• Rosetta
  • Model
  • Results
• PPI Model
  • Model
  • Results

For determining the mutations necessary for the creation of orthogonal pairs our written protocol was used. Our starting Protein DataBase file (PDB) was the PDB submitted by C. Ottmann et al. containing a complex of a T14-3-3 primer, fusicoccin, and 2 CT52 (PDB ID: 2O98) in which chain A and B are T14-3-3 monomers and chain P and Q are CT52 proteins.

Computational Alanine Scanning

After subjecting the PDB to relax and backrub simulations in Rosetta several of the best scoring PDBs (according to Rosetta) were submitted to the Robetta server¹ for computational alanine scanning, together with a mutations list containing the residues that are relevant for the binding of the T14-3-3-FC-CT52 complex based on a paper by C. Ottmann et al.²

The alanine scans gave us insight in which residues were the most critical for the binding interactions and therefore the best options for mutation testing.

Figure 1: The Results of the computational alanine scanning represented as a heat map with each row containing a PDB resulting from the backrub simulation and each column a residue represented by the ID of that residue in the original PDB. Values of the heat map are ΔΔG values in Rosetta Energy Units (R.E.U.)

Figure 1 displays the Computational Alanine scan in the form of a heat map. It shows that the most important residues in the T14-3-3 are 56 63 136 181 182 189 229 233, and for CT52 932 935 936 939 943 945 947 948 950 951 952 953 954 955 956.

Point Mutant Scanning

Figure 2: A schematic representation of the definition of an orthogonal pair: a mutated CT52 only interacts with a mutated T14-3-3 and the wildtype only interacts with the wildtype.

We create an orthogonal pair (see figure 2) by first destabilizing the binding interface by introducing a mutation in the CT52, since CT52 is a more flexible protein and thus more prone to changes in secondary structure than T14-3-3.

In the next step we restabilize the binding interface by introducing compensating mutations in the T14-3-3. In the first step all relevant residues found with the Computational alanine scanning are mutated to 19 different amino acids (all except cysteine) and then evaluated on their destabilizing effect (see figure 3A). Higher destabilization means a better orthogonality between the mutated CT52 and the wildtype T14-3-3.

In the second step all residues in T14-3-3 that are within 6 Å of the mutated residue are one by one mutated into 19 mutants to test the restabilizing effects of these mutations. Some of these mutations were combined into two point mutations to increase the restabilizing effect. This step ensures that the mutated CT52 binds to the mutated T14-3-3. Energy increase/decrease of the mutated pair compared to the wildtype pair are shown in figure 3C. The third step is to determine the orthogonality, that is the energy increase, between the wildtype CT52 and the mutated T14-3-3, this is shown in figure 3B.

figure 3: The effect of mutations depicted on heat maps, displayed in the rows are the relevant residues of the CT52 and in the columns the amino acid to which the residue is mutated, values displayed are the ΔΔG values in R.E.U. A) effect of destabilizing mutations in the CT52 on the wildtype T14-3-3 mutated CT52 binding interface. B) effect of destabilizing mutations in the T14-3-3 on the mutated T14-3-3 wildtype CT52 binding interface. Displayed mutations are not the mutations in the T14-3-3, but the mutations in the CT52 for which these T14-3-3 mutations compensate. C) The energy increase/decrease of the mutated pair compared to the wildtype pair.

Choosing Mutations

Based on the results represented in

figure 3A-C

For example, we first looked at heatmap A, the darker the colour, the stronger the destabilizing effect and thus the better the orthogonality between the mutated CT52 and the wildtype T14-3-3: if you look at the residue 947 on the y-axis you can see several dark colors, one of which is at mutation F on the x-axis. Next, we looked at the same coordinates on heatmap B (so residue 947 mutation F) to see whether those same coordinates also have a dark color in this heatmap. It did, which means the destabilizing effect of the mutated T14-3-3 and wildtype CT-52 is also quite strong. Lastly we looked at heatmap C, for the binding to still be strong enough to actually bind, the free Gibbs energy should preferably be lower than the wildtype-wildtype binding, this means the color in the heatmap should be blue for the chosen coordinates. For the coordinates residue 947 mutation F, the color is indeed slightly blue. For our application we need to be able to regulate the interaction with fusicoccin, so the CT52 shouldn’t bind strongly to the scaffold without fusicoccin present, this means that the Gibbs free energy should not decrease too much, which is the case when the color is dark blue, that this happens. This is the reason that, for example, the mutation I947W wasn’t chosen.

, several potential mutations were chosen to test in wet lab experiments, see

table 1.

The chosen mutations have high ΔΔG changes in figure 3A and 3B, and are blue (meaning it has a decrease in ΔΔG compared to the wildtype). A lower energy in our mutated pair is preferred over a higher energy, not only because a lower energy means a more stable binding, but also because Rosetta’s point mutant scan application tends to be a bit off in its predictions and thus a high energy could be even more unfavorable than predicted. For our application, however, the decrease in energy should not be too large, as this could decrease the regulability of our complex.

In figure 4 these mutations are presented in a crystal structure of the T14-3-3/CT52 complex.

figure 4: Our chosen mutations represented in a crystal structure in pyMol (original PDB ID: 2O98). A-D) in image A to D the visible purple or blue amino acids show the newly designed mutation pairs, the transparent green amino acids represent the original amino acids before Rosetta modeling was applied. The PDBs with mutations have been relaxed, backrubbed, mutated to the desired amino acids and their sidechains have been repacked. A) The CT52-I947F mutation with the T14-3-3-S71L and T14-3-3-I72V mutation, here the interaction is entirely apolar thanks to the replacement of Serine with Leucine. B) The CT52-I947H mutation with the T14-3-3-S71L mutation, this interaction consists of the apolar side of Histidine interacting with the now apolar Isoleucine. C) This is the CT52-K943D mutation with the E19R mutation, this interaction has an inverse charge compared to the wildtype. D) This is the CT52-S953K mutation with the T14-3-3-W237R mutation, here apolar interactions have been turned into polar interactions. E) The entire T14-3-3/CT52 complex showing all our designed mutations and their location in the protein. blue: W237R and S953K, cyan: E19R and K943D, pink: I947F (interacts with purple and brown), brown: I72V (interacts with pink), purple: S71L (interacts with pink and light-brown), light-brown: I947H (interacts with purple).

Determining the properties of the mutated proteins

In order to determine the properties of our mutated complexes, most importantly the binding affinity/dissociation constant and the cooperativity (which is defined as the increase/decrease in binding affinity of the second CT52 because of the first bound CT52), a model based on mass-action and Michaelis-Menten kinetics was developed to convert the activity measured in the readout to these properties.

back to top