Team:EPF-Lausanne/Results

 

 

 

 

Results of Modeling



{|class="wikitable" border="0" cellpadding="10" cellspacing="1" style="padding: 1px; background-color:#007CBC; text-align:center" !width="20%" align="left" valign="top" style="background:#ffffff; color:black"| = Summary of the main results =

 Wild type simulations
First of all, the equilibration (stabilization of temperature, pressure and density) was accurate for both states of LOV2 domain.
 * For dark state, see here


 * For light state, see here

Secondly, we analyzed many important characteristics of the system:


 * 1. RMSD was analyzed of all residues' alpha carbon, which shows us that the protein was stable and that our simulation was apparently trustable.


 * For dark state, see here


 * For light state, see here


 * 2. RMSF was analyzed for residues' side chains. We were able to localize, helped by some differential analysis some residues that move much more than others, which would mean that these moving residues were possibly implicated in the movement transmission that induces the general conformational change of the protein upon light activation. The movement of these residues were not correlated. And further analysis demonstrate that we were not able to see the conformational change (see below).


 * For dark state, see here


 * For light state, see here


 * For the differential analysis, see here


 * 3. The angle between the main beta sheet and the J-alpha helix of LOV2 domain was computed, and we could not see any periodic main movement neither in the dark nor the light state.


 * For the differential analysis of the angle, see here


 * 4. Side chain dihedral angle of the reactive cystein (residue 450) was computed for both state.


 * Interestingly, in the dark state we were able to find that the sulfur atom of this cystein point 30% of the time toward the cofactor, FMN (molecule that reacts with the protein upon light activation) and 70% of the time toward the opposite side of the FMN.


 * Important graphs and explanations, see here


 * These results were expected, because they confirm the in vitro results obtain by [Halavaty et al.], and this confirm once again that our simulation seems to be accurate.


 * In the light state, the cystein is covalently bonded with the cofactor, FMN, so the side chain dihedral angle was far more stable than in the dark state. This result was completely expected because after a covalent bonding with a big set of atom such as the FMN, the cystein's side chain is less free to move.


 * Important graphs and explanations, see here

<font color="#007CBC"> Simulations of non-light activated LOV2 domain with specific mutations
Based on the cystein's side chain movement analysis, residue mutations near the active site were designed. The goal was to "push" the side chain of the cystein more often in the direction of the FMN with steric interactions.

For a better view of the active site, see here

<font color="#007CBC"> I427F
Here the isoleucine 427 was replaced by a phenylalanine.

This mutation gave very interesting results:
 *  It changed the amount of time the cystein's side chain point toward the FMN from 30% in the wild type to ~57% in the I427F mutant. 


 * Important result, see here

So, an important hypothesis appears:
 * If the cystein's side chain points more often toward its reacting carbon in the FMN, there is more chances that upon light activation a covalent bond will be made. Moreover, if there is more steric obstruction toward the unbonding this newly formed covalent bond, this covalent bond will be stabilized, and it will finally leads to a general stabilization of the light activated state of the protein.

<font color="#007CBC"> L453G
Here the leucine 453 was replaced by a glycine in order to let empty space for glycine's side chain to move toward the FMN.

This mutation gave less interesting results than the first mutation:
 * It changed the amount of time the cystein's side chain point toward the FMN from 30% in the wild type to ~31% in the I427F mutant.


 * Important result, see here

So, an important hypothesis appears:
 * In that case, it seems that the cystein's side chain dosen't move a lot if empty space is available. This hypothesis sounds rational because the cystein in the wild type protein move in a more or less stable way in its available space. So, in increasing only the space available will make move the side chain slightly more but not significantly.


 * }

Click on each title below to access the detailled results.

=color="#007CBC"> Fusion of the LOV domain and the trpR DNA-binding domain=

The first step in our computational study of the LOV domain was to fuse the 2 domains of interest in VMD, namely the LOV domain and the TrpR DNA-binding domain. It allowed to visualize the different proteins tried by Sosnick. The working protein, that we call LovTAP is the result of the fusion at PHE22 of TrpR.

=<font color="#007CBC">Dark State simulation=

This is where we run a long simulation on the dark state system and analyze the output.

In the analysis, we tried to achieve the following goals:
 * find a structural change in the Jα helix based on the simulation
 * find residues showing different comportment in dark and light state

=<font color="#007CBC">Light state simulation=

The light state corresponds to the photoactivated state of the LOV domain, and here are gathered results concerning the light state from a 60ns simulation starting after previous equilibration.

We mainly focused on an analysis of dihedral angles to understand the movement of useful residues.

=<font color="#007CBC">Differential Analysis=

Now that the two states are well-characterized, we want to confront the two visions of the protein. This part is thus devoted to the comparison of the two states.

After a detailed analysis based on both previous simulation, we were able to determine that the stability of the Cystein 450 is highly correlated with the creation of the covalent bound to the FMN.

=<font color="#007CBC">Mutations=

Our final goal is to find a way to make the protein more stable, or to increase its affinity: that's why we imagined some ponctual mutations on some particular residues to do so. We picked the more mobile residues in the beta sheet, closest to the CYS450, and see if they can improve the overall stability. Here is a list of the mutations planned: These were partly based on studies made by : see here for more information.
 * ILE427 mutated in PHE
 * LEU453 mutated in GLY
 * 1) Zoltowski: Mechanism-based tuning of a LOV domain photoreceptor
 * 2) Christie, Steric Interactions Stabilize the Signaling State of the LOV2 Domain of Phototropin 1

We ran two other simulations after mutating the LOV domain at these residues and we discovered a much better stability of the cystein due to I427F. In this configuration, the cystein points toward the FMN in 57,2% of the cases, which is almost twice better as in the wild type protein!

=<font color="#007CBC">Validation of the equilibration =

This part brings together results validating our equilibration. This one is composed of 3 different steps:
 * first a minimization, where we try to find a minimum of energy. In fact, it is essential to find a stable point on the potential energy surface in order to begin dynamics. At a minimum on the potential energy surface, the net force on each atom vanishes.

Constraints are imposed during minimization. To minimize we need a function (provided by the forcefield) and a starting set of coordinates. The magnitude of the first derivative can be used to determine the direction and magnitude of a step (i.e. change in the coordinates) required to approach a minimum configuration. To reach the minimum the structure must be successively updated by changing the coordinates (taking a step) and checking for convergence. Each complete cycle of differentiation and stepping is known as a minimization iteration. The equilibration can itself be divided into 3 phases: - an NPT - an NVT - again an NPT The aim of doing a minimization followed by an equilibration simulation is to generate a trajectory for the system, which will be analysed further.
 * a second step composed with a heating of our protein allows to increase the temperature from 5 to 300K.
 * finally, we do an equilibration. This equilibration stage is required because the input structure is typically not within the equilibrium phase space of the simulation conditions, particularly in systems as complex as proteins, which can lead to false trajectories in protein dynamics.

This part gathers together plots confirming that our minimization-heating-equilibration were correct, and that we followed with a good file of trajectories.