Instrumental developments of TIMS to make it a powerful tool in solving interesting problems within the areas of health and energy


The function of a protein is intimately linked to its structure. Thus, NMR and x-ray spectroscopy have been invaluable for pharmaceutical research by providing ensemble-averaged structures. However, it is now known that proteins and their complexes are dynamic entities. Indeed, co-existing, transiently populated conformations play crucial roles in some of the fastest growing major diseases in the US—Alzheimer's disease, Type 2 Diabetes, atherosclerosis, cancer—but their detailed structures are not amenable with available biophysical methods that measure ensemble-averaged structures. Hence, in order to develop pharmacological strategies to treat or prevent many diseases, it is imperative to develop novel methods that elucidate co-existing, transient conformations.


Ion mobility spectrometry – mass spectrometry (IMS-MS) methods have recently been shown capable to study such transient protein conformations but still suffer from two major shortcomings: (1) the relationship between the protein structure elucidated by IMS-MS and the biologically active structure in solution is poorly understood and (2) detailed protein structures are largely inaccessible from IMS-MS data. One main thrust of our research is to develop novel IMS-MS methods that are capable to identify solution structures of proteins from IMS-MS measurements in an automatized manner. To this end, we utilize experimental and computational approaches for time-resolved IMS-MS that we recently developed in our lab.

Figure 1: Expected outcome is the ability to accurately elucidate protein structures from IMS-MS cross sections in an automatized manner.

IMS-MS measures an orientation-averaged cross section of a protein. It is thus highly challenging to extract detailed protein structures from only IMS-MS data (Fig. 1). Currently, two broadly distinct approaches are used to structurally interpret IMS-MS data. One approach is to couple structures from traditional methods, e.g. NMR, with IMS-MS data analysis. In our opinion, it is unlikely that such approaches can exploit the full potential of IMS-MS, which is, to elucidate structures for exactly those systems where traditional methods fail. Other approaches couple theoretical calculations with IMS-MS experiments. Currently, these approaches suffer from charge-state dependent protein dynamics in the gas phase and, thus, yield ambiguous structures. Moreover, these approaches can be computationally prohibitive for proteins and are not automatized.

Advantages of TIMS and future work

Our group develops methods to accurately elucidate protein structures from IMS-MS data alone, i.e. without considering data from other techniques, and in an automated manner. The central aspect of our approach is that we integrate experimental and computational methods for time-resolved IMS-MS, namely trapped IMS-MS and PSA/LCPA.

Figure 2: Illustration of the LCPA method

The projection superposition approximation (PSA) and local collision probability approximations (LCPA) are computational methods to estimate a momentum transfer cross section of an analyte ion. These methods are fast and accurate and enable high-throughput calculation of cross sections for large systems, including proteins and their complexes. Our group maintains a webservice for accessing the PSA algorithm at (free, registration required). The website also provides a detailed description of what the PSA does and how it works. More recently, our group developed the LCPA method, which accounts for the momentum transfer process more accurately than the PSA. As depicted in Figure 2, the LCPA method takes a molecular structure (2A) and calculates the ion-neutral potential energy surface U(r) that governs the collisions between the analyte ion and the neutral buffer gas particles (2B). On the basis of this potential energy surface U(r), a collision probability function τ(ε,r) is calculated (2C, where ε is the kinetic energy of the collision). Ray tracing and Delaunay triangulations are then carried out in order to reconstruct a collision surface (2D) from the collision probability function τ(ε,r). The cross section of this collision surface can then be easily calculated by known projection procedures. Because the LCPA explicitly accounts for the details of the ion-neutral interaction potential, it is a more flexible method than the PSA and can be applied to electronically diverse buffer gases. We are actively working on the LCPA.