BioExcel partners at the Institute for Research in Biomedicine IRB in Barcelona unravel the mechanisms for enzyme regulation by using advanced techniques for enhanced sampling and free energy calculations.
When one is managing a large-scale operation in which any misstep can lead to a catastrophe, it is only natural that one evolves into a control freak. Our cells are no exception and in case of cellular processes, they run a tight ship through many regulators, such as protein kinases. This superfamily of enzymes transfers a phosphate group from a high-energy molecule (like ATP) to other proteins, thereby modifying their activity. With over 500 different kinases in the human genome, these proteins regulate a wide range of processes: from signal transduction and cell cycle progression to metabolism, transcription, cell movement, apoptosis, and differentiation. Any atypical change in their activity can lead to a variety of pathological conditions, which is why it is important to understand what makes them (in)active in the first place.
p38α belongs to a group of protein kinases that are primarily involved in cellular stress response and that have been implicated in chronic inflammatory diseases, cancer, and heart and neurodegenerative diseases, which is why we wanted to learn more about its activation. The most typical mechanism involves a dual phosphorylation of a large flexible loop. This event allows p38α to first bind ATP and then a substrate molecule which the protein phosphorylates. Zhang and colleagues solved the X-ray structure of the phosphorylated protein which showed that the dual phosphorylation triggers a large conformational change. However, Tokunaga and colleagues contradicted these results in their NMR study and showed that large changes occur only upon ATP or substrate binding.
To figure out where these discrepancies in the proposed models come from, we employed numerous molecular dynamics simulations combined with an advanced sampling technique called metadynamics. We performed the calculations using the state-of-the-art molecular simulation package GROMACS (with metadynamics plugin PLUMED). This combination gave us an advantage over standard molecular dynamics simulations, as it allowed us to observe large conformational changes in a reasonable amount of computational time and in great detail. Furthermore, we could add statistical significance to the observed conformations.
We were able to prove that the observed X-ray structure is a direct consequence of the crystal environment and that the conformational change indeed occurs upon ATP binding, as suggested by NMR. We built on this finding further and explained why the protein binds ATP only in its phosphorylated form and how the substrates with a docking peptide enhance the catalytic reaction. We were also able to highlight important electrostatic interactions which we believe are specific to this kinase family and which also clarify the results of several mutational studies. Even though p38α has been extensively targeted in case of inflammatory diseases and some cancers, none of the drugs have made it yet to the market. We believe that the novel conformations we found could be used as a starting point in virtual screening studies aimed at uncovering new inhibitors. In addition, the described electrostatic interactions will hopefully allow us to explore alternative activation pathways with increased specificity.
Antonija Kuzmanic, Ludovico Sutto, Giorgio Saladino, Angel R. Nebreda, Francesco L. Gervasio and Modesto Orozco.
Changes in the free-energy landscape of p38α MAP kinase through its canonical activation and binding events as studied by enhanced molecular dynamics simulations
eLife (2017) DOI: 10.7554/eLife.22175