Subunit architecture of non-covalent complexes isolated directly from the cells.

Previous mass spectrometry approaches although successful in proposing numerous complexes have relied on separating all the proteins and generating peptides from them to identify the component proteins. While that approach is excellent at providing a catalogue of proteins it is unable to reveal whether or not the proteins from stable complexes that can be extracted intact. The approaches that we are developing rely upon the proteins being assembled, so that they can be extracted as an intact complex. The protein assemblies extracted using our methods also bring with them associations with nucleic acids and cofactors and reveal whether or not all of the proteins are present all of the time. We have chosen an important set of complexes involved in processing RNA to develop these ideas and from this research hope to discover new interactions. We also plan to use mass spectrometry to uncover how the proteins are organised within the structure of the assembly by using various methods to generate smaller subcomplexes. Since there will be some proteins that occur in more than one subcomplex this will allow us to piece them together to discover the overall architecture of the whole complex. Once we have this information a major question is in what orientation or shape are these subunits organised? To address this question we are developing methods that can measure the cross section of ions directly in the mass spectrometer. The method relies upon the movement of ions through a device that measures their speed as a function of their cross sectional area. This will enable us to distinguish basic shapes. We would like to extend this further to distinguish more subtle differences in shape. Taken together information about the number of subunits, their possible arrangements and overall shape we hope that we will be able to start to predict molecular structures of some of the macromolecular machines and their interactions that have so far remained elusive in cells.

Recent developments have shown that we can extract, and maintain intact within a mass spectrometer, the cap binding complex with copy numbers of approximately 12 000 molecules per cell. We have also developed MS instrumentation that enables us to maintain large cellular complexes intact or to raise their internal energy to partially dissociate them. These approaches, together the opportunity to form sub-complexes in solution, have allowed us to establish the complete subunit architecture of the exosome, the molecular machine responsible for degradation of RNA in yeast. Through generation of over 30 different complexes employing three different 'bait' proteins and with little previous experimental data we used computational methods to derive the entire subunit architecture. The outcome of these successes have prompted us to attempt ever more complicated assemblies for which very little structural information exists. While preliminary data for these assemblies are very promising they also serve to highlight the shortfalls of our current approaches. Specifically it is often impossible to rely upon the intact mass to provide the identity of the subunit. The extensive truncations and modifications of many similarly sized subunits render this an insoluble problem with our current methodology. We therefore need to correlate intact mass with sequence information to identify unambiguously subunits within an intact complex. We also need to understand more fully the generation of sub-complexes, either using chaotropic agents or thermal dissociation, as well as the gas phase dissociation of these assemblies. Very recently we have shown that it is possible to ascertain the shape of the complexes from ion mobility mass spectrometry and computational methods. We believe that application of this approach to endogenous complexes to obtain experimental validation of our proposed 3D interaction models would provide a very powerful adjunct to structural biology.