Christina Scharnagl and Dieter Langosch
Technical University of Munich (Germany)
Local Project ID:
HPC Platform used:
SuperMUC of LRZ
Integral membrane proteins exhibit conformational flexibility at different structural levels and time scales. Our work focusses on the biophysical basis of the interdependence of transmembrane helix dynamics, helix-helix recognition, and helix-lipid interactions. In this context, we try to understand the impacts of these phenomena on biological processes, such as membrane fusion, lipid translocation, and intramembrane proteolysis. Our approach closely connects experimental work and established computational analysis in order to interpret and guide the experiments and to validate the simulations.
Integral membrane proteins facilitate communication between the inside of the cell and its exterior. Their transmembrane domains (TMDs) perform structural and functional tasks, i.e. they drive protein/protein interactions, exhibit sequence-specific conformational dynamics on multiple size and time scales, and couple tightly to the hydrated lipid bilayer. These properties are interconnected and support a diversity of biological functions. Membrane proteins are notoriously difficult to study by experimental methods. In silico techniques like molecular dynamics (MD) simulations provide powerful tools of high spatial and temporal resolution that can effectively complement experimental methods. However, this level of detail is connected to a high demand on computational resources, only offered by high performance clusters like SuperMUC.
The scientists of the Technical University of Munich (TUM) applied MD simulations to study the conformational dynamics of various TMDs in lipid bilayers and a membrane mimetic solvent. Central questions and results are summarized in Fig. 1.
In project A the researchers focus on the conformational dynamics of the TMD of the amyloid precursor protein (APP). APP is enzymatically cleaved within its TMD by γ-secretase (GSEC), forming toxic peptides regarded as molecular cause of Alzheimer's disease (AD). Finding the link between the molecular architecture of the APP TMD and cleavage is therefore of utmost importance. It seems plausible that the TMD itself is optimized for unfolding of the scissile bond. However, this expectation was challenged by experiments and MD simulations [2-4]. The achieved results suggest an entirely new model of intramembrane proteolysis  where reaching a cleavage-competent state requires searching a complex energy landscape by the substrate/enzyme complex (Fig. 2).
This view initiated formulation of a number of computational and experimental approaches in a collaborative research program (DFG grant FOR2290, https://www.i-proteolysis.de/). Here, the TMD dynamics of ~100 known substrates of GSEC as well as non-substrates (Gauss Collaboration project pr48ko) are compared. Knowing the key dynamical motifs for processing will help to identify new substrates and to elucidate the physiological functions of these proteases in the brain and other organs.
Project B investigated TMDs de novo designed to exhibit different conformational dynamics. The researchers want to clarify how dynamics and peptide/lipid interactions affect functional properties of the helices, such as their ability to fuse lipid membranes and to flip lipid molecules. While a detailed analysis is still ongoing, preliminary results indicate that (i) the membrane constrains helix dynamics, (ii) lipids preferentially contact peptides via electrostatic interactions of their head groups (Fig. 1 B), and (iii) peptide/lipid interaction is a two-way-process (unpublished results). The work is supported by a DFG research grant (Mechanisms of Membrane Fusion and Lipid Flip).
In project C, the TMD-TMD interaction of human quiescin sulfhydryl oxidase II, a membrane protein that aids folding of other proteins, was investigated. Atomistic simulations in a bilayer system mimicking a cellular membrane revealed how the membrane can regulate interactions between membrane helices (Fig. 1C) by providing interfacial water and/or by binding of lipid molecules to the helices .
Sufficient exploration of the conformational space of protein and lipids poses a challenge in all projects. Hence, the researchers ran multiple independent copies of the simulation systems in parallel to cover aggregate times of 10-50 microseconds. This technique benefited greatly from SuperMUC's architecture and the Redisexec framework developed by the LRZ.
 Pester, O., Barret, P., Hornburg, D., Hornburg, P., Pröbstle, R., Widmaier, S., Kutzner, C., Dürrbaum, M., Kapurniotu, A., Sanders, C. R., Scharnagl, C. & Langosch, D. (2013). The Backbone Dynamics of the Amyloid Precursor Protein Transmembrane Helix Provides a Rationale for the Sequential Cleavage Mechanism of γ-Secretase. J. Am. Chem. Soc. 135, 1317-1329.
 Pester, O., Götz, A., Multhaup, G., Scharnagl, C. & Langosch, D. (2013). The Cleavage Domain of the Amyloid Precursor Protein Transmembrane Helix does not Exhibit Above-Average Backbone Dynamics. ChemBioChem 14, 1943-1948.
 Scharnagl, C., Pester, O., Hornburg, P., Hornburg, D., Götz, A. & Langosch, D. (2014). Side-Chain to Main-Chain Hydrogen Bonding Controls the Intrinsic Backbone Dynamics of the Amyloid Precursor Protein Transmembrane Helix. Biophys. J. 106, 1318-1326.
 Langosch, D., Scharnagl, C., Steiner, H. & Lemberg, M. K. (2015). Understanding Intramembrane Proteolysis: From Protein Dynamics to Reaction Kinetics. Trends Biochem Sci 40, 318-327.
 Ried, C., Scharnagl, C. & Langosch, D. (2016). Entrapment of Water at the Transmembrane Helix–Helix Interface of Quiescin Sulfhydryl Oxidase 2. Biochemistry 55, 1287–1290
Christina Scharnagl1*, Alexander Götz1, Dieter Langosch2*
1 Technical University Munich, Department of Physics E.14
2 Technical University Munich, Chair Chemistry of Biopolymers and Center for Integrated Protein Science (Munich)
* Principal Investigators
Dr. Christina Scharnagl
Technische Universität München
Fakultät für Physik
Maximus-von-Imhof-Forum 4, Raum P051
D-85350 Freising (Germany)
e-mail: christina.scharnagl [at] tum.de