Multi-scale modeling of molecular basis for odor and taste


ac_mimic.tarThe advent of exascale computation will impact tremendously on cell biology and pharmacology. Multi-scale simulations and enhanced sampling calculations on exascale platforms on one hand will allow describing accurately realistic biological systems in biologically relevant timescale and, on the other hand, will enable high-throughput of prediction of molecular data, e.g. enzymatic reactions constants, association/dissociation kinetic constants in protein/ and ligand/protein complexes. These data, often not accessible by experimental means, allow fixing constraints on the parameters required for mathematical modelling of signalling pathways, as for example the neuronal cascade. Hence, the combination of molecular simulation and system biology will offer an unprecedented predictive tool to investigate the effects of mutations and of drugs for subcellular events, testable by experiment. This will open a new avenue in understanding human biology and in counteracting cell derangement associated to diseases. The latter, indeed, most often, originate from misregulated signalling pathways, due to altered protein expression levels or protein mutations.
As for odorant receptors, we are currently particularly interested in enzymatic reactions involved in the cascade triggered by odorant molecules binding to their target receptors. Hybrid quantum mechanical/molecular mechanical (QM/MM) approach, in particular when coupled to molecular dynamics (MD) as a sampling method, allows mechanistic studies of chemical reactions in complex environments, and is well established for the study of enzymatic reactions. The availability of a QM/MM MD code with exceptional scaling performance on massively parallel supercomputers would drastically reduce the time to-solution. Thus, it is expected that the availability of high performance computing resources will allow undertaking systematic mechanistic studies of the enzymes involved in an intracellular signalling pathway.

What are we doing in BioExcel?

The CPMD code is among the best performing codes for ab initio molecular dynamics simulations. BioExcel is contributing to develop a new QM/MM interface for the CPMD code that promises to be more HPC oriented, more flexible and totally free. However, the new QM/MM interface needs to be tested and validated on a real state-of-the-art problem in order to prove its correctness and performance.
For this reason, in collaboration with the Human Brain Project (HBP, www.humanbrainproject.eu) we are interested to investigate the enzymatic mechanism of adenylyl cyclase (AC) in complex with its acti
vator protein Gsa by employing the QM/MM approach of CPMD. ACs are key component of many signaling cascades, including the odorant and taste perception ones, and their role is to catalyze the conversion of adenosine triphosphate (ATP) to cyclic adenosine monophosphate (cAMP). cAMP is a second messenger involved in several signaling cascade pathways and this conversion involves the hydrolysis of the Pα atom of ATP. These reactions are ubiquitous in biology, yet they are still not fully understood and highly debated in literature, especialy because of the role played by metal ions and water molecules.

A bit more about our collaborators and the context of the Use Case

The proposed Use Case project is performed in collaboration with the HBP. The research group involved in these simulations is one of the five groups dedicated to molecular simulation in the Brain Simulation Platform (i.e. sub-project 6 or SP6) of HBP. While other partners in the molecular simulation team perform protein diffusion simulations in crowded environments or coarse-grain simulations of protein-protein complexes, the task of this research team is to provide atomistic information of enzyme-catalyzed chemical reactions and of protein-ligand binding and unbinding processes that govern neuronal cascades.

Why is this Use Case particularly interesting for BioExcel?

The planned improvement of the QM/MM interface in its scaling performance by a team of BioExcel, combined with the replacement of the old Gromos96 MD engine by the more modern MD engine from the GROMACS package will provide an unprecedented powerful massively parallel and freely available tool to unravel the unknown mechanisms involved in the several molecular pathways, including the odorant and taste perceptions.

What is the current status of the Use Case?

A first working version of the new QM/MM interface has been already developed and tested. Currently, the developers are working to optimise the performance and make the interface user friendly. At the same time, the investigation of the Use Case through the new interface has started: comparisons of results and performance with the old QM/MM interface will provide a real-case benchmark for the accuracy and efficacy of the new QM/MM interface.
At the beginning of the project, several QM/MM models were built and tested with the CPMD old interface, in order to find the correct setup and partitioning for this complex system. For pedagogical purposes, we provide here below for the system with the deprotonated ribose oxygen O3’ of the ATP moiety both the models (MODEL1-10) that resulted unstable, and the one that has been later used for production runs. The most instabilities were apparent not during the preliminary equilibration phase but only during the initial production runs (MODEL4-10). The instabilities mainly involve the coordination shell of the magnesium ions inside the enzyme.

Models for the old QM/MM interface  Model for the new QM/MM interface
  FINAL MODEL