Determination of protein-protein complexes from sparse experimental data

The great majority of cellular processes, including enzyme catalysis, signal transduction, gene expression regulation and the immune response, depend on the formation of transient or permanent macromolecular complexes. In order to understand these processes at the molecular level, the structural biology community is facing the very challenging task to describe the structures and dynamics of hundreds of core macromolecular processes, which include DNA replication by the replisome, DNA transcription by RNA polymerase, protein biosynthesis by the ribosome, protein folding by chaperonins, and protein degradation by the proteasome. Structural information about these complexes at high resolution is normally provided by X-ray crystallography. There are, however, often cases in which it is possible to crystallize the molecular units individually but not as a complex, or in which a crystal can be obtained for a complex but not in a biological relevant conformation. As crystallization is not required for solution nuclear magnetic resonance (NMR) spectroscopy this method can be exploited, at least in principle, in the study protein-protein interactions in such situations. The high-resolution determination of the structures of protein-protein complexes, however, is still a challenging problem for this approach since it can normally provide only a limited amount of structural information at protein-protein interfaces. The objective of this project is to develop a computational platform for determining the structures of protein-protein complexes using sparse experimental information. We have identified chemical shifts from NMR spectroscopy and collisional cross sections from mass spectrometry as the most promising types of experimental data for this purpose, since they are readily measurable, at least with respect to other observables, and comparatively rich in structural information. This application represents our initial step in a long-term plan that we have committed ourselves to in order to provide a standard tool to the structural biology community to routinely determine the structures of large complexes, which are the basic functional units in the cell, rather than the structures of individual proteins or small complexes, which currently represent the vast majority of the structures in the Protein Data Bank.

Since large scale studies, in particular those using affinity purification and mass spectrometry, are proving capable of providing low-resolution information about the nearly complete set of protein-protein complexes in individual cells, we have decided to work on the development of computational methods to determine high-resolution structures from measurements that can be carried out with relative ease, even if they do not provide structural information of optimal quality. In particular, we have identified chemical shifts in nuclear magnetic resonance spectroscopy and collisional cross sections in mass spectrometry as observables that, when used in combination with advanced docking methods, have the potential of resulting in structures at high resolution. The application that we are presenting concerns the development of a software suite, called CamDock, that combines sophisticated ab initio computational docking methods with structural information derived from experimental measurements to determine the structures of protein-protein complexes of biological relevance. The essential ingredients of CamDock consist in a highly efficient representation of protein surfaces in terms of spherical harmonics and the use of cross sections from mass spectrometry and NMR chemical shifts as structural restraints. The latter component builds on our recent demonstration that the structures of the native states of globular proteins can be determined on the basis of the information provided by chemical shifts alone. In addition, the use of collisional cross sections from mass spectrometry will provide invaluable information about the overall shape of the complexes of larger molecular weight that perform crucial functions in human cells, such as DNA duplication and transcription, and protein synthesis and degradation.

This application is the initial step in our long-term goal to provide a standard tool to the structural biology community to determine the structures of large protein complexes, which are the basic functional units in the cell, rather than the structures of individual proteins or those of small complexes, which currently represent the vast majority of the structures in the Protein Data Bank. This goal is particularly relevant considering the fact that more than 80% of human proteins are functional upon forming large complexes of five or more components. In order to achieve this goal, with the present application we seek funding for developing the CamDock method, which is based on the inclusion of sparse data provided by a range of experimental techniques, including NMR spectroscopy, mass spectrometry, cryo-EM and small-angle X-ray scattering (SAXS) into advanced computational docking procedures. As we have shown in our initial studies, the CamDock approach is extremely well suited for integrating in a coherent manner structural information from these sources. Our plan is to address the use of collisional cross section provided by mass spectrometry in the three-year programme that we present, and the use of cryo-EM and SAXS data in a second phase (not covered by the present application) after achieving the goals described here. We believe that the ability to determine the conformations of large protein complexes through the approach that we propose to develop represents a key requirement of the structural biology community and that the support that we requested is essential to help maintaining the UK at the leading edge of structural biology. Our track record shows that we are strongly committed to collaborative research and that we have already established an extensive network of national and international collaborations, which includes scientists in the UK, Germany, Italy, Spain, Japan, Canada and the US. The project that we propose here has an extremely strong interdisciplinary character and we shall collaborate and exchange information, and materials where appropriate, with other research groups with specific expertise. We will accomplish this by either building further on existing collaborations or through the creation of new networks. Due to the interdisciplinary character of the project, the Post-Doctoral Research Associate (PDRA) assigned to this project will be able to increase significantly his/her professional profile. In addition, the PDRA will also gain project management skills, in particular the management of collaborations and supervision of PhD students, which will further him/her to establish in the longer term an independent scientific career. As indicated by our track record, we constantly aim at presenting our results in research journals with high visibility to researchers from a wide range of scientific areas and disciplines, in particular Nature, Science and PNAS, as well as in more specialized journals that aim at readers in our immediate fields and disciplines. We also have an extensive track record in presenting our research results at major international scientific meetings. We will continue to attend and present our findings at such meetings to communicate our results and to establish new scientific collaborations with experimental groups. We have also always kept close contacts between our research and teaching activities at both the graduate and undergraduate levels. We shall continue to communicate our findings in seminars and courses for students at the University of Cambridge, in the UK and worldwide. Furthermore, from the enthusiastic feedback that we have already received from students, we anticipate that the work carried out in this project will help to encourage students to participate in research in the context of Undergraduate, Master and PhD projects.