Prerequisites for supplying complementary high-resolution mass spectrometry data in RCM publications


In a 2005 Editorial of this journal,1 the then Editor-in-Chief Dr Robert Boyd outlined several important minimum requirements and standards for proposing ion fragmentation mechanisms and assigning product ion structures from experimental measurements. He emphasized that “the Editors of RCM have concluded that papers that are mainly devoted to postulation of 'fragmentation mechanisms', involving hypothesized ion structures, must provide a level of experimental evidence that justifies the amount of speculative discussion. Papers that do not meet this requirement in the opinion of the Editors will be returned to the authors.” Obviously, such requirements could include high-resolution, accurate mass measurements for deducing elemental compositions of precursor and product ions, multi-generation genealogical mapping experiments using MSn to establish the connectivities among ions, labelling and H/D exchange experiments, comparisons of dissociation behaviours of unknown peaks with those of commercial or synthesized standard compounds, correlations of MS/MS spectra with those of class analogues or derivatives, computational approaches, database searches, etc. Of course, not all of these experimental methods will be available or are even appropriate for a specific application. It is not even 10 years ago that many scientists had only low-resolution mass spectrometers at their disposal, usually of QqQ or QIT design, and sometimes medium resolution TOF or QqTOF instruments. High-resolution mass spectrometers, in particular FTICR-MS instruments, were available only in specialized laboratories. This situation has changed dramatically in recent years. Very sophisticated high-resolution hybrid mass spectrometers (Orbitrap, QqTOF, QIT-TOF) are now in widespread use in many clinical, biological and environmental laboratories, partly because instrument prices have come down but more importantly because instrument designs and instrument control software have been improved to a point where non-specialized personnel can obtain meaningful results without detailed mass spectrometry knowledge. Mass measurements are routinely possible with measurement uncertainties in the low ppm or even upper ppb range in both full-scan and tandem MS modes; very often instruments are now part of fully automated high-throughput workflows with minimal user intervention and data interpretation.

The easy access that scientists today have to these instruments has prompted the Editors of RCM to reiterate the importance of reporting 'appropriate' evidence for proposed molecular structures. Naturally, this raises the question of how much evidence is required to accurately propose an ion fragmentation pathway and to determine elemental formulae of precursor and product ions? Equally important is the question, when do we consider an unknown compound in a complex matrix to be 'identified' or the presence of a particular target substance unambiguously confirmed?

Obviously, this depends significantly on the type of the analytical application. For example, is the unknown compound a product of metabolism or degradation of a known parent compound? When available, confirmation with authentic analytical standards provides unambiguous assignment, and is required for publication. Or is it an entirely unknown compound? There is no easy answer to these questions but several helpful comments and suggestions can be made to authors.

It is often said that a 1 ppm mass measurement accuracy is sufficient for determining the elemental formula of an unknown small molecule using a properly calibrated instrument,2 even though it was also convincingly shown that such an accuracy alone is not sufficient to exclude enough candidates with complex elemental compositions (C, H, N, S, O, P, and potentially F, Cl, Br and Si).3 The use of isotopic abundance patterns as a single further constraint can remove >95% of false-positive candidates in these cases.3 Papers dealing with the mass spectrometric analysis of drugs and their metabolites usually fall in this category and authors describing metabolite identification and/or MS/MS fragmentation patterns should include accurate mass data in their papers, in addition to low-resolution MS/MS or MSn analyses. Unbiased metabolomics applications require accurate identification of multiple unknown compounds in very complex samples. These analyses are usually applied in conjunction with metabolite databases. The Metabolomics Society has defined minimum standards for compound identification,4, 5 including the necessity not only to assign a chemical name to a compound, but also to provide at least one further database annotation (e.g., PubChem or preferably the Independent Chemical Identifier, InChI). The minimum accepted 'proof' of metabolite identification is two characteristic, independently matching substance parameters, e.g., retention time and accurate mass or accurate mass and MS/MS spectrum, compared with a database. The Editors endorse these guidelines for RCM publications describing metabolomics applications. In addition, we strongly support the use of the terminology and statistical procedures for accurate mass measurements as recently outlined by Brenton and Godfrey.6

Unique elemental formulae cannot be established with 1 ppm measurement accuracy for larger molecules and other techniques have been developed for this purpose. For example, a top down/bottom up method7 computes elemental formulae from consecutive neutral losses (<500 u), starting with the intact precursor ion and down to the lowest nth generation precursor ion of m/z <500 using MSn. The precursor and product ion elemental formulae are computed by summing the elemental formulae for the lowest precursor ion together with the elemental formulae of the consecutive neutral losses. Papers that describe identifications of proteins based on mass spectrometric data must ensure that these meet the criteria described by Taylor and Goodlett.8

The number of papers describing the mass spectrometric analysis of natural products has steadily increased over the last several years. We would like to point authors to a review of FTICR-MS applications for natural products by Feng and Siegel.7 Several important techniques and strategies for natural product discovery by high-resolution MS have been summarized by the authors. As Orbitrap and the latest-generation QqTOF instruments venture into FTICR territory with respect to resolution and mass accuracy, many of the described experiments apply more universally. For example, the use of isotope patterns was mentioned above, but the procedure can be extended to accurate mass measurements of m/z differences between the heavier isotope peaks and the mono-isotopic peak (e.g., Δ = 1.99705 u for 35Cl/37Cl, Δ = 1.99795 u for 79Br/81Br).7 Also, the intact precursor species can be readily assigned from measuring the exact mass difference between the protonated and deprotonated molecules and adduct ion species (Δ = 2.01455 u between [M + H]+ and [M–H]; Δ = 21.98194 u between [M + H]+ and [M + Na]+).7

A closely related discussion is also required for defining minimum journal requirements for proof of presence of compounds in samples, e.g., for the confirmatory analysis of drugs of abuse. Indeed, typical questions are: how many ions are required for unambiguous confirmation in MRM analyses? And which ions should be chosen? These questions are beyond the scope of this Editorial and are the subject of intense regulatory activities.9, 10 RCM authors are particularly encouraged to study chapter 9 of Boyd et al.'s excellent treatment of the subject,11 where mass spectral identification criteria are evaluated and fitness for purpose of mass spectrometry methods is discussed.

In conclusion, establishing standardized guidelines for structural elucidation and compound identification is difficult and depends on the specific application. But as described above, today's easier access to high-resolution mass spectrometric analysis raises the bar for structural elucidation papers. Editors will now routinely ask authors of papers dealing with structure analysis or identification routines to additionally provide accurate mass data for their papers if they are not submitted with the initial submission, unless other conclusive means of structure proof are provided such as outlined in the opening paragraph of this Editorial.