Computational Study Of The Full-length 3d Structures Of MEKI, 2 And Activation Mechanism Of MEK1 By Phosphorylation

It is widely reported that various signals can be transducted by identical or similar signal molecules, and lead to distinct biology effect. The mechanism still remains to be elucidated about how the common signal pathway sort and transmit different signals.As a pivotal intracellular common pathway, ERK/MAPK signal pathway can be activated by a variety of cell surface receptors after binding to their corresponding extracellular stimuli, and then induce consequent activation between Raf, MEK and ERK by phosphorylation. The phosphorylated ERK is able regulate gene expression through interacting with hundreds of downstream protein substrate (such as transcriptional regulatory factor). It is reported that diverse signals evoke different or even opposite cytobiology effect after been transducted by ERK/MAPK signal pathway. Notably, there are only two MEKs (MEK1 and MEK2) which share high sequence similarity, but they are capable of tranducting multiple signals. The capacity is obviously considered to be closely connected with their subtle structural transformation; however, the detailed mechanism is still unclear. Therefore, the investigation about the dynamic characters of MEKs under various functional state will provide an admirable paradigm on the explanation of signal sorting mechanism by common signal pathway.Several MEK structures have been determined by crystallization, however, they all lack of three functionally important regions: the N-terminal region, the proline-rich region and the C-terminal region, all of which are hard to be crystallized due to their high flexiblity. The N-terminal region is of great importance not only because it is essential for the interaction with ERK, but also for the down-regulatory role it played in the activation of itself. Furthermore, there is an nuclear export sequence in this region involving MEK subcellular localization. Three phosphorylation sites in proline-rich region can up- or down-regulate MEK activity and impose effect on the ability of association between MEK1and signal molecules of other tiers. The function of C-terminal region is related to the subcellular localization and activity regulation. Apparently, all these missing parts greatly impede deeper research about the functional mechanism of MEK, it is therefore regard to be extraordinary urgent to obtain the full-length three dimension structure of MEK1 and MEK2.At present, the computational structural biology is able to perform considerable accurate three-dimension structure construction by homology modeling and ab initio folding, and to implement molecular dynamic simulation on their structural features. Based on this advantage, we try to construct the full-length three dimension structures of MEK1 and MEK2, and investigate their dynamic characteristics and the relation between phosphorylation and activation to approach the signal transduction mechanism of ERK/MAPK common signal pathway.Since the first 38 and 59 residues are truncated to facilitate the crystallization in MEK1 and MEK2, respectively, the structure information of N-terminal region is not available in PDB databank. Thus we adopted the ab initio folding method to construct these two structures. Secondary structure prediction indicate that the 26-62 residues in MEK1 and 30-66 residues in MEK2 are prone to fold into a helix, we therefore chose MEK1 1-64 residues and MEK2 1-68 residues to conduct ab initio folding with multiple trajectories and long simulation time. The result shows that a helix is formed with more than 30 residues, and the helix is able to keep structural intact in major coordinate clusters, which is consistent with secondary structure prediction. So we chose the structure with the lowest potential energy as typical structure and consider it as a template to construct the full-length three dimension structure of MEK1. Compared to MEK1, there is also a long helix formed in MEK2, with obvious lower stability and occasional unwinding, twisting and turning into pieces in its C terminal region. Besides, a structure distinct with MEK1 appears that several helix fragments huddle together, the low potential energy of which is considerably low. However, this novel structure is unstable and not frequently appears in the simulation progress. Thus we chose two structures as its typical structure: the structure with the lowest potential energy in the whole simulation progress and the structure with the lowest potential energy in the major coordinate clusters.We used MODELLER software to construct the full-length three dimension structure of MEK1. First, we took the typical structure of MEK1 N-terminal region (1-59 residues) we obtained above and the crystal structure 3EQI (45-381 residues) as templates for the construction. The structure of proline-rich region and C-terminal region are constructed by software without any template. The primary analysis indicate that the main body of protein may exhibit intense electrostatic repulsion to the N-terminal region, and lead to the possibility that the N-terminal region may stay in a position far away from the main body of protein. Therefore, we construct another structure using the structure of full-length ab initio N-terminal region (1-64 residues) and crystal structure 3EQI (61-381 residues) as templates, the N-terminal region of which is apart from the main body of protein. We performed molecular dynamic simulation on both structures with long time and multiple trajectories. The result indicates that the potential energy of the first structure is obviously lower than that of the second structure with a higher structural stability. The 1-6 residues which function is to interact with ERK is buried between N terminal helix and the main body of protein, where MEK is unable to interplay with ERK, and only in a small quantity of structures that 1-6 residues are exposed on the surface and enable the interaction with ERK. This structural characteristic can well explain the fact reported before that MEK and ERK exhibit weak binding capacity. In the second structure, N-terminal helix swings apart from the main body of protein. The analysis indicate that an immense structural transformation will occur in N terminal region after the interaction with ERK to position ERK to the appropriate place, and the ability of N terminal helix to swing may facilitate the structural transformation. Finally, N terminal helix displays considerable flexiblity in both structures, which may be the reason for the failure of the crystallization.Similarly, we constructed MEK2 structure with two ab initio folding structures of MEK2 N-terminal region (1-68 residues) and crystal structure 1S9I as templates, and obtained two full-length three dimension structures of MEK2. We performed molecular dynamic simulation on both structures with long time and multiple trajectories. The result showed that it is more reasonable for the N-terminal region to adopt a long helix conformation. Besides, compared with the N-terminal region of MEK1, that of MEK2 exhibit higher flexibility, which may be related to their sequence discrepancy or the different position where the N-terminal helix was put in the first place. Another structure should be constructed with MEK1 full-length three dimension structure as template to get adequate spatial sampling, however, the work has not been done due to the limit of computational resources.The activation of MEK1 is achieved by the dual phosphorylation of Ser218 and Ser222. To probe the character and mechanism of conformational changes during the phosphorylation, we further constructed the structure of MEK1-ATP complex, the structure of MEK1 with Ser218 and Ser222 monophosphorylated, respectively, and the structure of MEK1-ATP complex with both Ser218 and Ser222 phosphorylated, and performed molecular dynamic simulations on each structure. The result shows that Ser218 and Ser222 are exposed more after the bind of ATP to MEK1, and are therefore more vulnerable to be phosphorylated. Monophosphorylation can induce the changes of distance and angle between N-terminal helix and the main body of protein, and dualphosphorylation is able to induce similar but much more intense changes, which eventually lead to the exposure of the ERK-binding residues of N terminal region. As it was reported before, MEK are capable of binding with ATP no matter being phosphorylated or not, and the dualphosphorylation is required for the regular activation. Our research not only well explained these reported phenomena, but also clearly displayed the features and mechanism of conformation change during activation, and also laid the foundation for the further investigation of the progress of interaction between of activated MEK1 and ERK.To sum up, we constructed reasonable MEK1 full-length three dimension structure and performed research on its dynamic characters, and our work revealed the features and mechanism of conformation change during activation. We also preliminary constructed the full-length three dimension structure of MEK2. Our work laid the foundation for the further research of signal tranduction mechanism through ERK/MAPK common signal pathway.