Periodic Reporting for period 4 - COMP-MICR-CROW-MEM (Computational Microscopy of Crowded Membranes)
Reporting period: 2020-05-01 to 2020-10-31
Cell membranes form a highly complex and heterogeneous mixture of membrane proteins and lipids. Understanding the protein-lipid interplay that gives rise to the lateral organisation principles of cell membranes is essential for life and health, as malfunctioning at the level of lipid-protein interaction is implicated in numerous diseases1, including various cancers, Alzheimer’s disease, diabetes, HIV, and heart failure. Thus, investigations of these crowded membranes is emerging as a new and exceptionally exciting frontier at the crossroads of biology, life sciences, physics, and chemistry.
However, our current understanding of the detailed organisation of cellular membranes, remains rather elusive. Characterisation of the structural heterogeneity in-vivo is very challenging, owing to the lack of experimental methods suitable for studying these fluctuating nanoscale assemblies of lipids and proteins in living cells with the required spatio-temporal resolution. Given the fundamental role of biomembranes, both within and around the cell, knowledge about the molecular level organisation is crucial. In recent years, computer simulations have become a unique investigatory tool for understanding the driving forces governing the lateral organisation of cellular membrane components and this “computational microscopy” has become indispensable as a complement to traditional microscopy methods.
In this ERC project, using advanced computational microscopy, we studied the interaction of lipids and proteins in complex, crowded, membrane patches, to enable the driving forces of membrane protein sorting and clustering to be unravelled at conditions closely mimicking real cellular membranes.
However, our current understanding of the detailed organisation of cellular membranes, remains rather elusive. Characterisation of the structural heterogeneity in-vivo is very challenging, owing to the lack of experimental methods suitable for studying these fluctuating nanoscale assemblies of lipids and proteins in living cells with the required spatio-temporal resolution. Given the fundamental role of biomembranes, both within and around the cell, knowledge about the molecular level organisation is crucial. In recent years, computer simulations have become a unique investigatory tool for understanding the driving forces governing the lateral organisation of cellular membrane components and this “computational microscopy” has become indispensable as a complement to traditional microscopy methods.
In this ERC project, using advanced computational microscopy, we studied the interaction of lipids and proteins in complex, crowded, membrane patches, to enable the driving forces of membrane protein sorting and clustering to be unravelled at conditions closely mimicking real cellular membranes.
The three major aims of the proposal were
1. To develop a novel computational microscopy framework for simulating biomolecular processes at multiple resolutions, from atomic detail to supra coarse-grained.
2. To use this new computational microscopy framework to investigate the driving forces of membrane protein sorting and clustering, focusing on the role of gangliosides, cardiolipins, and crowding conditions.
3. To use the new insights and computational microscopy to provide a molecular view of realistic biological membranes, in particular of prototype eukaryotic plasma membranes composed of hundreds of different lipids and proteins.
With respect to (1), we have made progress in coupling different resolutions together in the AdResS framework. In particular, together with our collaborator Dr. Praprotnik, we have developed a fast method to switch between so-called bundled water models and free water models. The bundled model is required to efficiently exchange with coarse-grain models, whereas the latter provides better solution properties of biomolecules. We have also
improved the overall quality of the CG Martini model, which forms the hearth of our computational microscope. This has truly been a community effort, with many people involved in the testing of the new parameter sets, both from our own group and from external collaborators. The lead of this project was in the hands of Dr. Souza.
With respect to (2), we studied the behavior of ganglioside lipids in their ability to self- assemble and to bind to membrane proteins. Extensive simulations at both the CG and all- atom level have been performed to unravel the driving forces of ganglioside-ganglioside and ganglioside-protein interactions. The role of cardiolipin has been investigated in mitochondrial membranes - we found that cardiolipin can mediate the formation of respiratory chain supercomplexes in crowded conditions.
Our increased modeling of complex membranes, aim (3), has been very successful. We now have complex models of realistic mitochondrial membranes and thylakoid membranes. In collaboration with the group of Prof. Tieleman we resolved lipid sorting in plasma membranes in models with >60 different lipid types, raising the bar in complexity. We have also setup a model for a large piece of a crowded plasma membrane model featuring tens of different plasma membrane proteins. In addition, we contributed to realistic models of bacterial membranes. In collaboration with Prof. Im, we have implemented the use of Martini models for lipopolysaccharides (a major component of the bacterial outer membrane) in CHARMM- GUI.
1. To develop a novel computational microscopy framework for simulating biomolecular processes at multiple resolutions, from atomic detail to supra coarse-grained.
2. To use this new computational microscopy framework to investigate the driving forces of membrane protein sorting and clustering, focusing on the role of gangliosides, cardiolipins, and crowding conditions.
3. To use the new insights and computational microscopy to provide a molecular view of realistic biological membranes, in particular of prototype eukaryotic plasma membranes composed of hundreds of different lipids and proteins.
With respect to (1), we have made progress in coupling different resolutions together in the AdResS framework. In particular, together with our collaborator Dr. Praprotnik, we have developed a fast method to switch between so-called bundled water models and free water models. The bundled model is required to efficiently exchange with coarse-grain models, whereas the latter provides better solution properties of biomolecules. We have also
improved the overall quality of the CG Martini model, which forms the hearth of our computational microscope. This has truly been a community effort, with many people involved in the testing of the new parameter sets, both from our own group and from external collaborators. The lead of this project was in the hands of Dr. Souza.
With respect to (2), we studied the behavior of ganglioside lipids in their ability to self- assemble and to bind to membrane proteins. Extensive simulations at both the CG and all- atom level have been performed to unravel the driving forces of ganglioside-ganglioside and ganglioside-protein interactions. The role of cardiolipin has been investigated in mitochondrial membranes - we found that cardiolipin can mediate the formation of respiratory chain supercomplexes in crowded conditions.
Our increased modeling of complex membranes, aim (3), has been very successful. We now have complex models of realistic mitochondrial membranes and thylakoid membranes. In collaboration with the group of Prof. Tieleman we resolved lipid sorting in plasma membranes in models with >60 different lipid types, raising the bar in complexity. We have also setup a model for a large piece of a crowded plasma membrane model featuring tens of different plasma membrane proteins. In addition, we contributed to realistic models of bacterial membranes. In collaboration with Prof. Im, we have implemented the use of Martini models for lipopolysaccharides (a major component of the bacterial outer membrane) in CHARMM- GUI.
In conclusion, the new Martini 3.0 model is fully operational, which will allow researchers worldwide to benefit from the improved set of parameters. The final simulations on the crowded plasma membrane modes are ready to be published, models for other realistic membrane are available from the Martini web portal. The significant increase in terms of complexity that we achieved will allow many researchers to simulate membrane related processes under more realistic conditions, and to form a bridge between the molecular and the mesoscale world.