Probing the fragmentation reactions of protonated glycine oligomers via multistage mass spectrometry and gas phase ion molecule hydrogen/deuterium exchange†2

Abstract

An ion trap mass spectrometer equipped with electrospray ionization has been modified to study the structure of protonated polyglycyl peptides G_n (where n = 2–5 glycine residues) and their product ions formed by collision induced dissociation tandem mass spectrometry (CID MS/MS) via the novel application of gas phase ion–molecule hydrogen/deuterium (H/D) exchange reactions. In particular, the structures of the b₂, b₃, b₄, and b₅ ions formed via CID MS/MS from various protonated glycine oligomer precursors have been examined. The b₂ ions, formed from the protonated G₂ and G₃ precursor ions, the b₃ ion from the protonated G₃ precursor, and the b₄ ion from the protonated G₅ ion all undergo CID and gas phase H/D exchange consistent with formation of protonated oxazolone structures previously proposed for b_n-type ions. However, CID MS/MS, MS³, and H/D exchange of the putative b₄ and b₅ arising from the protonated G₄ and G₅ precursor ions, respectively, as well as experiments with various methylated derivatives of G₄, suggest that the major portion of these ions are not b_n ions, but are instead formed via backbone–backbone neighboring group participation reactions remote to the C-terminal amino acid. Efforts to elucidate the mechanisms behind this loss of H₂O are described.

Introduction

Over the past decade mass spectrometry has emerged as an essential tool for the analysis of biologically important molecules. This renaissance in biological mass spectrometry (documented in numerous reviews and books) [1] has largely been fueled by the development of “new” soft ionization methods such as electrospray ionization (ESI) [2] and matrix assisted laser desorption ionization (MALDI) [3]. Primarily, the use of biological mass spectrometry has focused on two areas: (1) to “weigh” the mass of the biomolecule (i.e. to determine its molecular weight); and (2) to induce fragmentation of the molecular ion via tandem mass spectrometric (MS/MS) techniques [4a], [4b] [e.g. collision induced dissociation (CID)], thereby providing primary sequence information. Indeed, the identification of proteins by interpretation of the CID product ion spectra of peptides generated by specific proteolytic enzymes is rapidly becoming routine, particularly when coupled with powerful database search algorithms [5]. With these successes, the mass spectrometry community has now turned its attention to the application of MS for the characterization of (1) noncovalent complexes, and (2) higher order structure such as the secondary, tertiary, or quaternary structures of proteins.

Additionally, several groups have focused their efforts on developing structurally diagnostic tools that are complementary to CID methods. These include (1) the measurement of gas phase thermochemical properties such as gas phase basicities and proton affinities [6], (2) alternate activation methods in tandem mass spectrometry, including surface induced dissociation (SID) [7], photofragmentation [8a], [8b], [8c], electron induced dissociation [9], blackbody infrared dissociation (BIRD) [10], as well as probes of the structures of neutrals such as neutralization–reionization (NR) and neutral fragment reionization (NFR) mass spectrometry [11a], [11b], (3) bimolecular reactions (which can be further classified into (a) ion–molecule reactions [12] and (b) ion–ion reactions [13a], [13b]), (4) physical probes of the gas phase conformational shape of ions (examples include ion mobility [14] and surface impact mass spectrometry [15] that measures the defects of a surface bombarded by an ion) and (5) theoretical molecular modeling methods [16] (including semi-empirical, ab initio, or molecular dynamics techniques).

Ion–molecule reactions [12] are proving to be both popular and elegant probes of the structure of biomolecules since they (1) offer such a huge scope (the combination of different types of ions and neutrals is limitless), (2) may be more useful probes of higher order structure because they involve inherently less energetic reactions than CID (which neccesitates “heating up” the ion to induce fragmentation), and (3) are applicable to a wide range of existing mass spectrometers with little or no modification, thereby obviating the need to invest in new instrumentation. Additionally, many of these novel gas phase ion–molecule reactions are complementary to the more established MS/MS based methods and thus can be used in conjunction with CID in instruments such as ion traps to provide supplementary information. To date, the most extensively used gas phase ion–molecule reactions of biomolecules have been gas phase hydrogen-deuterium (H/D) exchange [6], [17a], [17b] and acid–base reactions involving proton transfer [6].

Critical to the development of techniques to probe the gas phase structure of biomolecules have been advances in the development of mass analyzers that are not only more versatile, but offer ever increasing levels of sensitivity (i.e. decreasing detection limits). Arguably, the most impressive recent advances have been those associated with time of flight [18], Fourier transform ion cyclotron (FTICR) [19a], [19b], [19c] and ion trap [20a], [20b], [20c] mass analyzers. The ion trap, largely because of its unique trapping capabilities that allow multistages of mass spectrometry (MSⁿ) to be performed on mass selected ions, has proved to be a particularly useful instrument. Since the successful coupling of ESI to ion traps [21], these analyzers have been used to study the structure of biomolecules through (1) the measurement of thermochemical quantities such as proton affinities [22a], [22b], (2) MS/MS methods such as CID [23], [24c], (3) gas phase ion–molecule reactions [24a], [24b], [24c], [24d], (4) gas phase ion–ion reactions (i.e. reactions of multiply charged anions or cations with ions of opposite charge [13a], [13b], and (5) the development of hyphenated systems such as ion trap–ion mobility mass spectrometry [25].

Of particular relevance to the work presented here is the ability to combine a number of different events in an ion trap. Because ion trap events (i.e. MSⁿ experiments involving ion isolation, activation, and detection) are tandem-in-time and not tandem-in-space, this instrumentation is particularly suited to performing gas phase ion–molecule chemistry. Thus, it is possible to carry out ion–molecule or ion–ion reactions on mass selected ions either before or after CID reactions [26a], [26b]. For example, the reaction of multiply charged oligonucleotide anions with trimethylsilylchloride results in the formation of trimethylsilylated anions, whose structure can then be interrogated via CID [24d]. Additionally, ion–molecule proton transfer reactions [24c] and ion–ion reactions [27] on CID product ions derived from multiply charged cations have been used to reduce charge states and thereby simplify the product ion spectra.

Aspects of the gas phase ion chemistry of protonated glycine oligomers previously studied by others include (1) determination of proton affinities [28] and gas phase basicities (GB) [29a], [29b], [29c], [29d], [29e], [29f], [29g] (which show an increase in GB as the size of the oligomer increases), (2) examination of the H/D exchange reactions of their [M + H]⁺ ions [30a], [30b], [30c], (3) a study of the reaction between the [M + H]⁺ ion of diglycine and acetonylacetone [31], (4) ion–mobility mass spectrometry [32], (5) molecular modeling [29b], [29c], [3], [32], and (6) CID [33a], [33b], [33c], [33d], [33e], [33f], [33g], [33h], [33i]. Of most relevance to the work described here are the previous studies on the gas phase H/D exchange [30a], [30b], [30c] and CID reactions of protonated glycine oligomers [33a], [33b], [33c], [33d], [33e], [33f], [33g], [33h], [33i].

Previously, gas phase H/D exchange experiments using FTICR MS have shown that in addition to differences in the GB [34a], [34b], (i.e. as the difference in GB between the protonated peptide and D₂O increases, the rate and extent of H/D exchange decreases), the formation of multiple hydrogen bonds between the [M + H]⁺ ion and D₂O, particularly for G₂ and G₃, is an important factor that can effect the extent of H/D exchange. The importance of hydrogen bonding on H/D exchange is exemplified by previous studies that reveal poor H/D exchange yields for protonated glycine, which is unable to form the multiple hydrogen bonds described [30a], [30b], [30c].

The CID reactions of protonated glycine oligomers have also been studied by a number of workers using different types of mass analyzers [33a], [33b], [33c], [33d], [33e], [33f], [33g], [33h], [33i]. Of most relevance are the studies by the groups of Bursey [33a], [33b], [33c], [33e] and Harrison [35c], [35d] using low energy CID in hybrid mass spectrometers and the measurement of activation energies for dissociation by Kebarle [33f] using energy resolved CID in a triple quadrupole.

The structures of peptides can be probed by studying the CID reactions of their [M + nH]ⁿ⁺ [36a], [36b] or [M − H]⁻ ions [37]. Generally, the major “sequence” ions formed upon CID of a protonated peptide ion are the N-terminally truncated y_n type and complementary C-terminally truncated b_n type ions [36a], [36b]. Mechanisms for the formation of these structurally relevant “sequence” ions (i.e. b- and y-type ions) [36a], [36b] from singly or multiply protonated peptides have been proposed [35a], [35b], [35c], [35d], [35e], [35f], [35g], [35h], [38a], [38b], including the structures of the neutral products [39a], [39b], [39c], using the concept of the “mobile” proton [38a], [38b]. Harrison et al. [35c] have shown that the protonated N-benzoyl-glycyl-glycine peptide ion fragments via loss of the C-terminal glycine residue to form a 2-phenyl-5-oxazolone product ion. They proposed that the mechanism for the formation of this product ion, by nucleophilic attack at the protonated carbonyl of the second amide bond by the preceding carbonyl followed by proton transfer and bond cleavage, could be a general one that operates for other protonated peptide ions. This general mechanism is illustrated for the formation of the b₃ product ion from a protonated tetraglycine peptide (Scheme 1, Pathway 1A). Others have also proposed similar mechanisms [35e]. According to the mechanism shown in Scheme 1, a truncated peptide is formed as the complementary neutral to this protonated oxazolone. However, if proton transfer occurs within the ion–molecule pair, a neutral oxazolone and an N-terminally truncated molecular ion corresponding to a y_n-3 ion would be formed (Scheme 1, Pathway 2). B_n-type ions can fragment further by loss of CO to produce a_n-type immonium ions (Scheme 1, Pathway 1B) or by direct fragmentation to a_n-1 type, or smaller b-type ions [35c], [35d], [35f], [35h]. A-type ions may also be formed directly by fragmentation of the intact peptide ion [35f].

Four different structures have been considered for the b₂ ions of simple peptides [structures (A)–(D) are shown for the b₂ ion of G₂ in Diagram 1] [33a], [33b], [33c], [33d], [33e], [33f], [33g], [33h], [33i]. One of the challenges to unequivocally assigning any one of these structures to the b₂ ion is the independent “synthesis” of each of these structures in order to compare the MS/MS spectra to that of the b₂ ion formed in the fragmentation of a peptide [M + H]⁺ ion. Many have ruled out the open chain acylium ion (A) as the b₂ product ion structure, because b₁ ions are not stable for simple aliphatic amino acids [33i]. The diketopiperazine structure (B) has been ruled out on the basis that it exhibits a different CID MS/MS product ion spectrum to that of an equivalent b₂ ion [33g], [33h], [39a]. Until recently, this had left the oxazolone (C) as the default structure. However, based solely upon ab initio calculations, Eckart et al. have suggested an immonium ion (D) as yet another possible structure [33h].

Alternate mechanisms have also been proposed for the formation of y_n-1 and y_n-2 product ions (Schemes 2 and 3, respectively). Both product ions may potentially be formed via nucleophilic attack at either the protonated first or second carbonyl by the N-terminal amino group, followed by elimination of a neutral aziridinone [35a], [35b] (Scheme 2) or diketopiperazine (Scheme 3), respectively. Evidence for the diketopiperazine neutral loss upon formation of the y_n-2 ion has been provided using NFR mass spectrometry [39a], [39b], [29c].

The losses of small neutrals such as H₂O, NH₃, and CO from the protonated precursor ion have also been the subject of some recent interest. The product ions corresponding to these “nonsequence” losses do not lead to structurally relevant information regarding the primary sequence and have therefore generally been ignored. However, further interrogation of these product ions might provide useful additional structural information. Recent results suggest that competing neighboring group participation reactions [40] involving side chain–backbone interactions [41a], [41b], [41c], [41d], side chain–side chain interactions [42a], [42b], as well as the effect of secondary and tertiary structures [43] may have a significant effect on the formation of “sequence” versus “nonsequence” ions. Thus, in light of our previous results with cysteine containing peptides, (which suggested that the loss of H₂O may be in competition with the formation of sequence ions via neighboring group side chain–backbone participation reactions [41b], [41c]), we were interested in determining whether the “nonsequence” loss of H₂O could compete with “sequence” ion formation in the absence of any side chain interactions. Simple polyglycyl peptides are particularly suited to such studies in that complications arising from side chain reactivities are removed and thus only the key backbone chemistry is probed. Potentially, bimolecular ion–molecule reactions could be employed to study the structures of the product ions formed via CID and could provide valuable additional mechanistic insights into CID fragmentation processes. Therefore, in this article we have employed gas phase H/D exchange ion–molecule reactions in a modified ion trap in combination with collision induced dissociation to study the structure of protonated glycine containing peptides and their product ions.