All-at-Once MS

In recent years, fantastic technological advances in mass spectrometry (MS) have driven the emergence of fields such as proteomics and metabolomics. MS also plays a pivotal role in pharmacology, environment and forensic sciences by providing two kinds of information: molecular mass and structure (through fragmentation). Whereas masses of molecules can be obtained in a single spectrum, structures are painfully obtained one by one, after selection (either manually or automatically) of each component of interest in the mixture.

In contrast, two-dimensional (2D) techniques allow the simultaneous acquisition of both mass and structural information, whatever the number of molecules, opening up a completely new way to work with complex samples – and without reliance on prior separation by chromatography. 2D Fourier transform techniques have revolutionized nuclear magnetic resonance (NMR) since their introduction by Ernst in 1974, paving the way for the analysis of complex samples, such as isolated proteins but also blood or urine. No surprise that they have also been awarded two Nobel prizes. Among mass analyzers, the Fourier-transform ion cyclotron resonance (FT-ICR) spectrometer has the best capability to separate individual molecules, reaching a resolution of ten thousand (distinguishing one electron at 20,000 Dalton) and a mass precision of 0.1 ppb. The principle of 2D FT-ICR was established in the late eighties by the joint efforts of two groups at the École Polytechnique Fédérale de Lausanne (Switzerland) – the Geoffrey Bodenhausen NMR team and later Tino Gäumann’s FT-ICR team. Unfortunately, 2D FT-ICR did not follow in the footsteps of 2D NMR as the initial version suffered from three main drawbacks: loss of resolution caused by in-cell fragmentation by collision induced dissociation, difficulty in data treatment at full resolution, and intense scintillation noise.

From 2010 onwards, we revisited 2D FT-ICR with two NMR groups – Geoffrey Bodenhausen (who had moved to the Ecole Normale Supérieure in Paris) and Marc-André Delsuc (University of Strasbourg) and introduced solutions to those problems by using gas-free fragmentation modes like IRMPD and ECD, developing data treatment to handle files of several gigabytes, developing improved pulse sequences and by introducing an innovative algorithm based on mathematical sparsity theory for noise reduction (1). Currently, a 2D acquisition takes approximately the same time as an LC-MS run to obtain a 2D spectrum with unit mass resolution for precursor ions and FT-ICR ion resolution for fragments. With these new tools, our group and that of Peter O’Connor’s at the University of Warwick (UK) demonstrated that 2D FT-ICR is now able to analyze complex mixtures of peptides starting from the classic Cytochrome C digestion mixture containing a dozen peptides and up to a very complex digest of a cell line, also allowing detection of polar peptides that were lost in nanoLC separations. Moreover, whatever the sample complexity, 2D FT-ICR is a completely ‘data independent analysis’ – acquisition from the full mixture takes place without chromatography separation and without triggered MS/MS acquisition.

However, the resolution of precursor ions in 2D FT-ICR is limited by the number of scans acquired in one experiment: currently 2048 (2k) to 8192 (8k) or 0.5 to 4 hour acquisition times, respectively. For many applications, especially in the field of environmental or petroleum analysis, high resolution of precursors is highly desirable but requires at least 128k points. The standard solution used for accelerating acquisition in 2D techniques based on Fourier Transform is non-uniform sampling (NUS). NUS consists of skipping points in the dimension with the higher cost in term of acquisition time, which, in 2D FT-ICR, is the first dimension (precursors). In order to obtain the desired structural information, missing points have to be reconstructed through data treatment – in NMR, data is processed using a maximum entropy algorithm. Unfortunately, the time dependence of this algorithm precludes its use in mass spectrometry. We recently reported the development of NUS acquisition in 2D FT-ICR MS and data reconstruction based on a denoising algorithm we published in 2014 (2). We were able to acquire “square” 2D FT-ICR spectrum with the same resolution on precursors and fragments using 32k points at NUS 4 for precursors and 128k for fragments, which exhibits resolution over 20 000 at m/z 400 in both dimensions (3). We are currently applying our NUS protocol to various samples: complex digests of proteins, lipid mixtures that are not easily separated by chromatography, and beverage flavors.

Much remains to be done in 2D FT-ICR MS to join the performance level of 2D NMR: design of new pulse sequences for MS3 and MSn experiments, including the development of new algorithms for speeding up data treatment and expansion into new applications. But 2D FT-ICR MS, which relies only on pulse sequence and so works on all commercial FT-ICR mass spectrometers, is already a powerful tool. Indeed, through data-independent acquisition it can provide a full MS/MS map at high resolution for fragments and precursors of complex mixtures without chromatographic separation, solving a challenge in mass spectrometry.