Integrative Molecular Dynamics Simulations Untangle Cross-Linking Data to Unveil Mitochondrial Protein Distributions

Cross-linking mass spectrometry (XL-MS) enables the mapping of protein-protein interactions on the cellular level. When applied to all compartments of mitochondria, the sheer number of cross-links and connections can be overwhelming, rendering simple cluster analyses convoluted and uninformative. To address this limitation, we integrate the XL-MS data, 3D electron microscopy data, and localization annotations with a supra coarse-grained molecular dynamics simulation to sort all data, making clusters more accessible and interpretable. In the context of mitochondria, this method, through a total of 6.9 milliseconds of simulations, successfully identifies known, suggests unknown protein clusters, and reveals the distribution of inner mitochondrial membrane proteins allowing a more precise localization within compartments. Our integrative approach suggests, that two so-far ambigiously placed proteins FAM162A and TMEM126A are localized in the cristae, which is validated through super resolution microscopy. Together, this demonstrates the strong potential of the presented approach.