Final Report Summary - MOLECULAR ASSEMBLY (A multidisciplinary approach for the computational assembly of large molecular machines from electron density maps) Proteins are the clockwork of the complex machinery that underlies human life. Many diseases, including cancer, Alzheimer and many infectious diseases, can be attributed directly to mechanisms operating at the protein level. Many of the most important functions in the cell are carried out by proteins organized in large, dynamic molecular machines, such as the ribosome, the 26S proteasome and the nuclear pore complex. A mechanistic, atomic-resolution understanding of molecular machines is needed for rational drug design against these diseases and is therefore of high biomedical relevance. Cryo-electron microscopy (cryo-EM) is an immensely powerful technique to study macromolecular machines. Cryo-EM observes individual molecules in various states, giving important biological insight into the dynamics of a complex, complementary to a high-resolution crystal structure. One of the major challenges in structural biology is to obtain atomic-level insight into molecular machines for which these atomic structures are unknown. Cryo-EM is particularly well suited to this type of molecular machines, providing biological insights unattainable by any other method. The complete 26S proteasome, for example, has been characterized by the Baumeister group using cryo-EM, showing how the core particle is capped at two sides by a regulatory unit. Unfortunately, cryo-EM maps have a much lower resolution than crystallographic maps. Therefore, cryo-EM does not by itself provide atomic-level insight into molecular machines. To obtain this insight, the molecular machine must be assembled computationally from atomic structures of the subunits.The main objective of the current project has been the development of methodology for the computational assembly of large molecular machines from cryo-EM electron density maps. This objective was achieved already in the first year of the project, with the development of a novel assembly protocol, ATTRACT-EM. ATTRACT-EM builds upon the docking program ATTRACT, previously developed in the Zacharias group, combined with novel methodology for fitting and scoring candidate assemblies. ATTRACT-EM has demonstrated its ability for the rigid assembly of molecular machines and showing improvement over existing methods. Using ATTRACT-EM, the large GroES-GroEL complex was for the first time successfully assembled from a low-resolution experimental cryo-EM density map. ATTRACT-EM was first presented in a talk at the Ringberg retreat of the Baumeister group, and has now been published in the open-access scientific journal PLOS ONE.Towards the end of the project, further improvements in ATTRACT-EM were achieved and demonstrated on concrete test cases. Now, the protocol works very well also for smaller systems, as well as for cases where the components are flexible and have to be broken into smaller parts. These improvements have been presented at the International Caesar Conference on Electron Microscopy, and a manuscript for publication is currently being prepared. These results were made possible due to the methodological advances that were achieved within the scope of this project. In addition, these advances also proved fruitful for generic atomic force-field-driven assembly with ATTRACT, leading to improvements in docking accuracy and speed. These improvements were presented at the International CAPRI conference, and one manuscript for publication has been submitted and a second is under preparation. Together, these results and activities have established ATTRACT-EM in the cryo-EM community as a highly successful protocol for simultaneous assembly of rigid units into cryo-EM density maps. Therefore, the project has had a considerable impact on the cryo-EM community, empowering scientists both inside and outside Europe. More specifically, ATTRACT-EM gives experimentalists larger control over the assembly process, allowing them to achieve better results in the computational assembly of existing protein components into their density maps. This, in turn, yields to improved accuracy at the atomic level, facilitating subsequent applications such as drug design that require high-resolution atomic details. Future developments will focus on increasing the impact and the benefits of ATTRACT-EM to users in the cryo-EM community. The expected final results of the project will involve the finalization of the manuscripts mentioned above, and the maturation of tools and graphical user interfaces to facilitate the use of ATTRACT-EM by the community. A practical based on these interfaces has already been given in the “Algorithms in structural bioinformatics” Winter school. A preliminary web interface has already been developed and is available at http://www.attract.ph.tum.de. All developed methodology is available under open-source licenses.
