Using Collision Cross Section Distributions to Assess the Distribution of Collision Cross Section Values

Careful transfer of ions into the gas-phase permits the measurement of protein structures, with ion mobility-mass spectrometry, which provides shape and stoichiometry information. Collision cross sections (CCS) can be obtained from measurements made of the ions mobility through a given gas, and such structural information once obtained should also permit interlaboratory comparisons. However, until recently, there was not a recommended standard form for the reporting of such measurements. In this study, we explore the use of collision cross section distributions to allow comparisons of IM-MS data for commonly analyzed proteins. We present measurements from seven proteins across three IM-MS configurations, namely, an Agilent 6560 IMQToF, a Waters Synapt G2 possessing a TWIMS cell and a modified Synapt G2 possessing an RF confining linear field drift cell. Mobility measurements were taken using He and N-2 as the drift gases. To aid comparability across instruments and best assess the corresponding gas-phase conformational landscapes of the protein standards, we present the data in the form of averaged CCS distributions. For experiments carried out in N-2, CCS values for the most compact ion conformations have an interinstrument variability of <= 3%, and the total CCS distributions are generally similar across platforms. For experiments carried out in He, we observe the total CCS distributions to follow the same trend as observed in N-2, while CCS for the most compact ion conformations sampled on the 6560 are systematically smaller by up to 10% than those observed on the G2. The calibration procedure (for TWIMS) yields TWCCS for native-like proteins which are largely similar to those obtained on DTIMS instruments. We collate previously reported values of CCS for these proteins in the form of histograms which bear a remarkable similarity to the CCS distributions, reflecting the conformational heterogeneity of proteins and also how conformer populations can be altered on transfer from solution to the detector. This gives concern for some caution when calibrating sample protein drift times simply with single numeric CCS values.