Importance of Domain Closure for the Autoactivation of ERK2

• Published in: Biochemistry (Easton) 2011-09, Vol.50 (37), p.8038-8048
• Main Authors: Barr, Daniel, Oashi, Taiji, Burkhard, Kimberly, Lucius, Sarah, Samadani, Ramin, Zhang, Jun, Shapiro, Paul, MacKerell, Alexander D, van der Vaart, Arjan
• Format: Article
• Language: English
• Subjects: Enzyme Activation - genetics; Mitogen-Activated Protein Kinase 1 - chemistry; Mitogen-Activated Protein Kinase 1 - genetics; Mitogen-Activated Protein Kinase 1 - metabolism; Molecular Dynamics Simulation; Mutation - genetics; Protein Binding - genetics; Protein Folding; Protein Structure, Secondary - genetics; Protein Structure, Tertiary - genetics
• ISSN: 0006-2960; 1520-4995
• DOI: 10.1021/bi200503a

Extracellular signal-regulated kinases 1 and 2 (ERK1 and -2, respectively) play a critical role in regulating cell division and have been implicated in cancer. In addition to activation by MAPK/ERK kinases 1 and 2 (MEK1 and -2, respectively), certain mutants of ERK2 can be activated by autophosphorylation. To identify the mechanism of autoactivation, we have performed a series of molecular dynamics simulations of ERK1 and -2 in various stages of activation as well as the constitutively active Q103A, I84A, L73P, and R65S ERK2 mutants. Our simulations indicate the importance of domain closure for autoactivation and activity regulation, with that event occurring prior to folding of the activation lip and of loop L16. Results indicate that the second phosphorylation event, that of T183, disrupts hydrogen bonding involving D334, thereby allowing the kinase to lock into the active conformation. On the basis of the simulations, three predictions were made. G83A was suggested to impede activation; K162M was suggested to perturb the interface between the N- and C-domains leading to activation, and Q64C was hypothesized to stop folding of loop L16, thereby perturbing the homodimerization interface. Functional analysis of the mutants validated the predictions concerning the G83A and Q64C mutants. The K162M mutant did not autoactivate as predicted, however, which may be due to the location of the residue on the protein surface near the ED substrate docking domain.