Team:Costa Rica/Computational Model
We aimed to model the 3D structure of our chimeric protein AgrA to try to predict its topology
and functionality. We also made a docking of the protein to understand how it dimerizes. As mentioned in the Description
we pretend to make a chimeric AgrA protein which consists in a combination of the same protein
from two different organisms of the same phylum. AgrA is a transcription factor composed of two
domains which are connected by a flexible linker. The first domain, known as response domain, is
the part of the protein that is phosphorylated when the quorum sensing is activated. The second
domain, named DNA binding domain, binds to the promoter and promotes the transcription of
virulence genes. We combined the response domain of C. difficile with the DNA-binding
domain of S. aureus and joined them using three different linkers: the 2 natives linkers
(one from C. difficile and the other from S. aureus) and a flexible synthetic
linker, generally known as (G4S)3. In order to comprehend and analyze the stability and functionality of our chimeric AgrA, we modeled
the three-dimensional structures of AgrA with the DNA binding domain of S. aureus and the
response domain of C. difficile joined with different linkers. This was done with the
program i-Tasser which uses a
variation of the Monte Carlo simulation. The structures predicted are shown in the next figure.
Figure 1: Predicted structure of the protein with the
linker: A) synthetic (G4S)3 , B) native C. difficile linker, C) native S.
aureus linker. In the three structures predicted, the DNA binding domain of the AgrA protein is very similar to the crystallized 3D structure
reported, with some little differences in a few loops. Even though the response domain have not
been crystallized, the structure obtained is similar to the one predicted before by
(Srivastava et al., 2014) in a different study.
The 3D structures have a relatively good C- score which
oscillates between -0.53 and -0.51.
Figure 2: Example of phosphorylated protein. Then, we tried to make a docking of the homodimerization that the proteins undergo to be able to
bind to DNA. We got various probabilities of dimerization. Here we show the 3 main
probabilities according to the energy and the RMSD calculated. Figure 3: Main probabilities of dimerization according
to the energy and the RMSD calculated. Later on, we tried to simulate how the homodimer binds to DNA. Nevertheless, the
results that we accomplished showed a different outcome from the one obtained previously in the
crystal structure (PDB: 3BS1). Because of this we were not able to conclude anything about this
docking. More research needs to be done to simulate this
interaction and to completely characterize the 3D structure of the protein. Sandeep K. Srivastava, Kalagiri Rajasree (2014), Influence of the AgrC-AgrA Complex
on the Response Time of Staphylococcus aureus Quorum Sensing, Enviromental
Microbiology. Doi: 10.1128/JB.01530-14. Yang J. Zhang (2015), I-TASSER server: new development for protein structure and
function predictions. Nucleic Acids Research, 43: W174-W181.
Prediction Model
The following step was to phosphorylate the protein. However, we encountered that the
regular programs used to predict post-translational modifications, including Rosetta and Coot, are not able to
phosphorylate an Asp residue. Also, when phosphorylated manually
using chimera, other programs like Rosetta or EXPasy were not able to understand this
modification. Because of that, we made the phosphorylation using a MD program, which is an extension of
the Grommacs program for macromolecules. This online tool is called Vienna-PTM, using it we were
able to phosphorylate the structure and make a relax, to achieve its native
structure. An example of the phosphorylated protein is shown in figure 2.
References