Introduction

Collision-induced dissociation (CID) combined with electrospray ionization (ESI) [1, 2] is a powerful tool for analyzing biological molecules such as peptides and proteins. Although there are many studies regarding the mechanistic and theoretical aspects of positive-ion low-energy (low-E) CID of peptides [3,4,5,6], negative-ion CID studies have been limited to reports by a small number of groups, namely those of Bowie [7], Harrison [8,9,10], Cassady [11,12,13,14], and Takayama [15, 16]. It is well-known that in positive-ion low-E CID of peptides, the formation of b- and y-ions can be explained by a mechanism involving a mobile proton [4, 5] and residue scrambling [6, 17,18,19]. Furthermore, the low-E CID results in residue-specific enhanced cleavage at the amide bond (C–N) of acidic Asp/Glu/Cys-Xxx residues and Xxx-Pro residues [5]. However, the residue-specific cleavage characteristics of negative-ion CID are quite different from those of positive-ion mode and are as follows (Table 1):

  1. 1.

    Specific cleavage at the N-Cα bond of Xxx-Asp/Asn residues to form a c- and z-ion pair [7, 15]

  2. 2.

    Specific cleavage at the C–N bond of acidic Xxx-Asp/Glu/Cys residues to form b- and y-ions [16]

  3. 3.

    Specific cleavage at the N-Cα bond of Xxx-Ser/Thr residues to form z-ions [20]

  4. 4.

    y1 ion formation due to the acidic carboxyl group at the carboxyl (C)-terminus [12, 15]

Table 1 The Specific Cleavage Residues, Product Ions, and the Product Ions with Intramolecular Hydrogen Transfer in Positive- and Negative-Ion Low-Energy CID of Peptides
Full size table

Here, we use the nomenclature of peptide fragments in as per the Biemann-Roepstorff notation [21, 22], i.e., amino (N)-terminal side a-, b-, and c-ions and C-terminal side x-, y-, and z-ions (Scheme 1), while the nomenclature with hydrogen addition (+H), protonation (+H+), and deprotonation (−H+) to describe c- and y-ions such as [y + 2H]+, [c + 2H]+, [c]−, and [y]− is used according to the proposed nomenclature of Chu et al. [23].

Scheme 1
scheme 1

Simplified Biemann-Roepstorff nomenclature for peptide fragments

Full size image

With respect to the formation of c-ions in negative-ion CID of peptides, the Cassady group has reported that when using alanine heptamers (AAAAAAA) with or without an arginine (R) or lysine (K) residue, a dominant c4 ion is observed in the CID spectra independent of the position of the R and K residues [12]. Furthermore, the same group reported that the c-ions can be produced by cleavage at the N-Cα bond of the backbone region between the third and fourth residues from the C-terminus, i.e., Cm-3 and Cm-4 residues (m represents number of residues in the peptide), respectively, when C-terminus has a carboxyl group, while N-Cα bond cleavage between the Cm-2 and Cm-3 residues (Cm-2,3) preferentially takes place when the C-terminus has an amide group [13]. The preferential cleavage at the N-Cα bond to form c-ions may be explained by proton (H+) abstraction from the backbone amide of the Cm-2,3 and Cm-3,4 residues via 8- and 11-membered ring transition states, respectively. The formation of negative c-ions by Cm-3,4-H+ abstraction via an 11-membered ring transition state is shown in Scheme 2, although Bokatzian-Johnson et al. have proposed the loss of 9-membered ring neutral with intramolecular proton transfer based on the DFT calculation [13]. According to the report of Bokatzian-Johnson et al. [14], it is likely that intramolecular proton abstraction from central amide backbone nitrogen (Nb) via 8- or 11-membered ring transition states excited with collisional activation takes place (Scheme 2). Another characteristic of the negative-ion CID spectra of peptides lacking Asp/Asn/Glu/Cys residues is the formation of y-ions [9, 12, 13]. Both c- and y-ions are produced by intramolecular hydrogen transfer to the backbone amide nitrogen (Nb). It is of interest to examine the source of hydrogen in the generation of c- and y-ions from the standpoints of appropriate and flexible conformations of gas-phase peptide molecules, which enable hydrogens to move to the amide nitrogen.

Scheme 2
scheme 2

An adaptation of Cassady’s mechanism for the formation of c2 ions in negative-ion CID of pentapeptides [13]

Full size image

Here, we examine the source of hydrogen in the formation of c- and y-ions in higher energy collisional dissociation (HCD, corresponding to a conventional low-E CID) experiments with deprotonated peptides [M–H]− generated by ESI MS. To determine the source of hydrogen, peptides labeled with deuterium on the α-carbon, β-carbon, and backbone amide moieties were used. Peaks of c- and y-ions produced with and without intramolecular deuterium (D) transfer were detected and confirmed with a high-resolution mass spectrometer and exact m/z values.

Experimental

Materials

All peptides were purchased from the Peptide Institute (Minoh, Osaka, Japan). The samples used are an alanine pentamer (AAAAA), a lysine pentamer (KKKKK), and a phenylalanine pentamer (FFFFF). The deuterated peptides AAA(d3)AA, AAG(d2)AA, and AAAG(d2)A were supplied from the Peptide Institute. Acetic acid and acetonitrile (HPLC grade) were purchased from Wako Pure Chemicals (Osaka, Japan). Water used in all the experiments was purified with a MilliQ water purification system from Millipore (Billerica, MA, USA). Deuterium oxide (D20), acetonitrile-d3 (CD3CN), and trifluoroacetic acid-d (TFA-d) were purchased from the Sigma-Aldrich (Steinheim, Germany).

Mass Spectrometry and Sample Preparation

The HCD experiments were performed on a Q Exactive Focus Orbitrap mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) equipped with an ESI source. The mass resolving power for precursor ions was 70,000 at m/z 200 in FWHM. The width for selecting precursor ions was ± 0.4 Da. Target gas and collision energy were nitrogen and 10 eV, respectively. The sample was introduced into the ion source with an infusion inlet system at a flow rate of 30 μL/min with nitrogen being used as both nebulizing and drying gas. The samples were prepared as 10 μM solutions with a 1:1 (v/v) mixture of water/methanol with added 0.1% acetic acid. To obtain a deuterated peptide for alanine pentamer AAAAA, the peptide was dissolved in a mixture of D2O and CD3CN (1:1, v/v) and incubated for 24 h at room temperature.

Results and Discussion

Preferential Formation of y1 and c2 Ions of Peptide Pentamers

Negative-ion HCD spectra of deprotonated molecules [M–H]− of three different pentamers AAAAA, KKKKK, and FFFFF are shown in Fig. 1. All the HCD spectra showed considerably intense peaks corresponding to y1 and c2 ions. The c2 ions originate from cleavage at the N–Cα bond between the Cm-3 and Cm-4 residues (Cm-3,4) of the peptides [13], while the y1 ions originate from cleavage at the backbone amide C–N bond of the C-terminal residue [12, 14]. The CID spectra also showed the common product ions of cn (n = 1–4), yn (n = 1–4), cn–H2O, yn–H2O, and yn–CO2. The intense c2 ions can be explained by N–Cα bond cleavage through Cm-3,4–H+ abstraction via an 11-membered ring transition state [13] (Scheme 2). The formation of the intense y1 ions observed in all the HCD spectra can be explained by specific cleavage at the C–N bond of the C-terminal residue due to the influence of the acidic carboxyl group [12, 15].

Figure 1
figure 1

Negative-ion HCD spectra of deprotonated molecules [M–H]− of three different pentamers (a) AAAAA (Mr 373.2), (b) KKKKK (Mr 658.5), and (c) FFFFF (Mr 753.4)

Full size image

Formation of Deuterated c- and y-Ions Confirmed with Deuterated Peptides

In order to examine the possibility of hydrogen transfer from α-carbon (Cα) and β-carbon (Cβ) to form c and y ions, negative-ion HCD spectra of deuterated glycine containing peptides AAG(d2)AA and AAAG(d2)A and a deuterated alanine pentamer AAA(d3)AA were obtained (Fig. 2). Figure 2a, b shows preferential c2 and y1 ion formation accompanied by deuterated products c2(d) and y1(d). Furthermore, c1 and y2 ions were also accompanied by deuterated products c1(d) and y2(d), respectively (Figs. 3 and 4). Although Fig. 2c also showed extraordinarily intense c2 and y1 ion peaks and cn (n = 1–4) and yn (n = 1–4) series ions, the c1, c2, y1, and y2 ions observed were not accompanied by any deuterated peaks of c1(d1), c2(d1), y1(d), and y2(d) due to deuterium transfer from the Cβ. This indicates that the hydrogen to form c1, c2, y1 and y2 ions does not originate from the Cβ of the 3rd Ala(d3) residue. The results obtained above indicate that the hydrogen to form c- and y-ions originates from the sites of the α-carbon (Cα) and/or amide nitrogen (Nb) of the peptide backbone (Scheme 3), and never come from the Cβ sites.

Figure 2
figure 2

Negative-ion HCD spectra of deuterated glycine G(d2) and alanine A(d3) containing alanine clusters (a) AAG(d2)AA (Mr 361.3), (b) AAAG(d2)A (Mr 361.3), and (c) AAA(d3)AA (Mr 376.2)

Full size image
Figure 3
figure 3

The c2 and y2 ion formation in negative-ion HCD spectra of deuterated glycine G(d2) containing alanine clusters (a) AAG(d2)AA (Mr 361.3) and (b) AAAG(d2)A (Mr 361.3)

Full size image
Figure 4
figure 4

The c1 and y1 ion formation in negative-ion HCD spectra of deuterated glycine G(d2) containing alanine clusters (a) AAG(d2)AA (Mr 361.3) and (b) AAAG(d2)A (Mr 361.3)

Full size image
Scheme 3
scheme 3

The formation of c- and y-ions due to intramolecular hydrogen transfer from the backbone Cα and Nb sites

Full size image

In order to confirm the intramolecular hydrogen transfer from the sites of the Cα and/or Nb of the peptide backbone, the negative-ion HCD spectrum of the de-deuteronated analyte [M(d7)–D]− at m/z 378.2264 formed from a deuterated alanine pentamer AAAAA-d7 obtained by the incubation in D2O/CD3CN solution was obtained as shown in Fig. 5. The AAAAA having seven active hydrogens was deuterated with the rate of 72.7% (data not shown). The HCD spectrum preferentially showed y1(d), y1(d2), c2(d3), and y2(d3) ions, as shown in the insets of Fig. 5. The peaks at m/z 159.0996 and 160.1059 corresponding to c2(d1) and c2(d2) observed may be due to hydrogen and/or residue scrambling, although the mechanism is unclear.

Figure 5
figure 5

Negative-ion HCD spectrum of the precursor ion [M(d7)–D]− formed from deuterated alanine pentamer AAAAA-d7 (Mr 380.4). The insets represent enlarged spectra of c1, y1, c2, and y2 ions

Full size image

Source of Hydrogen for the Formation of c- and Y-Ions

The intramolecular hydrogen transfer to form c- and y-ions in negative-ion CID spectra of peptides has been studied by Harrison [8]. From negative-ion CID spectra of tripeptides AAG and AAG-d5 (Fig. 3 in the report of Harrison [8]), the rate of intramolecular hydrogen (deuterium) transfer to form c1 and y1 ions can be estimated as shown in Scheme 4. The rate of H (D) transfer estimated from the report of Harrison [8] indicates that 68.2% of deuterium in the c1 ion originates from the backbone amide nitrogen (Nb) between Ala2 and Gly3 or C-terminal carboxyl group (Scheme 4b), while 31.8% of the hydrogen originates from the Cα sites of Ala2 and Gly3. On the other hand, 74.3% of the deuterium in the y1 ion originates from Nb between Ala1 and Ala2 and/or the N-terminal amino group (Scheme 4a), while 25.7% originates from the Cα sites of Ala1 and Ala2. The intramolecular deuterium transfer from Nb to another Nb site to form c1(D) in Scheme 4a may be explained by a 5- or 8-membered ring transition states (Scheme 4b).

Scheme 4
scheme 4

(a) The rate of hydrogen (deuterium) transfer for negative c1 and y1 ions in the paper of Harrison [8] and proposed mechanisms of (b) c1 ion formation via 5- or 8-membered ring transition states

Full size image

In a similar manner, the rate of intramolecular hydrogen (deuterium) transfer in the HCD spectra of Figs. 2, 3, and 4 can be estimated from the peak intensity. The rate of intramolecular deuterium transfer from the Cα of Gly3(d2) and Gly4(d2) residue to form c1 and c2 ions is represented in Scheme 5. It is hypothesized that the c1 and c2 ions for AAG(d2)AA are formed by intramolecular deuterium transfer from the Cα via 6- and 3-membered ring transition states, respectively (Scheme 5 upper). The c1(d) and c2(d) ions for AAAG(d2)A are formed by deuterium transfer via 9- and 6-membered ring transition states, respectively (Scheme 5 lower). This indicates that about 39–49% of hydrogens which form c1 and c2 ions originate from the Cα of the Gly residue, while the residual hydrogens (51–61%) come from the Cα of Ala and the backbone amide nitrogen (Nb). Considering the result of Harrison [8] and our results of Schemes 4 and 5, it may be concluded that the major source of hydrogen in the formation of c-ions is both the Nb and the Ala(Cα) sites of peptides. In contrast, the formation of y1(d) and y2(d) ions can be explained by intramolecular deuterium transfer from the Cα of the Gly(d2) residue via 7- and 4-membered ring transition states with rates of 29–60% (Scheme 6). The major source of hydrogens in the formation of y-ions may also originate from the backbone amide (Nb), as shown in Schemes 4a and 6.

Scheme 5
scheme 5

The rate of deuteron transfer from the Cα of Gly3(d2) and Gly4(d2) residues of deuterated peptides AAG(d2)AA and AAAG(d2)A in the formation of c1 and c2 ions

Full size image
Scheme 6
scheme 6

The rate of deuterium transfer from the Cα of Gly3(d2) and Gly4(d2) residues of deuterated peptides AAG(d2)AA and AAAG(d2)A in the formation of y1 and y2 ions

Full size image

In the negative-ion HCD spectrum of [M(d7)–D]− showed in the insets of Fig. 5, the peak intensities of the y1(d) and y1(d2) indicate that 69.1% of deuterium in the y1(d2) ion originates from the amide nitrogens (Nb) and/or the N-terminal amino group (Scheme 7), while 30.9% of the hydrogen originates from the Cα sites. The peak intensities of the c1(d2) and c1(d3) indicate that 41.5% of deuterium in the c1(d3) ion originates from the Nb sites and/or the C-terminal carboxyl group, while 58.5% of the hydrogen originates from the Cα sites. In a similar manner, the rates of intramolecular D (H) transfer to form y2(d3) and c2(d4) ions can be estimated from the insets of Fig. 5, as shown in Scheme 7.

Scheme 7
scheme 7

The rate of deuterium transfer from the amide nitrogen (Nb) and N-terminal amino group of deuterated peptides AAG(d2)AA, AAAG(d2)A in the formation of y1 and y2 ions

Full size image

The rates of intramolecular deuterium transfer from the Cα sites of Gly3(d2) for AAG(d2)AA and Gly4(d2) for AAAG(d2)A and from the Nb sites of AAAAA-d7 residues to form c1, c2, c3, y1, and y2 ions in negative-ion HCD spectra (Figs. 2, 3, 4, and 5) are summarized in Table 2.

Table 2 Rate (%) of the Intramolecular Deuterium Transfer from the Cα Sites of Gly3(d2) for AAG(d2)AA and Gly4(d2) for AAAG(d2)A and from the Nb Sites of AAAAA-d7 Residues to Form c1, c2, c3, y1, and/or y2 Ions in Negative-Ion HCD Spectra
Full size table

Conclusions

The characteristics of residue-specific cleavage and product ions in negative-ion CID of peptides are quite different from those of positive-ion CID, as summarized in Table 1. However, intramolecular hydrogen transfer to the amide nitrogen (Nb) for the formation of c- and y-ions is common to both positive- and negative-ion CID, and it is of interest from the standpoints of the conformation and flexibility of gas-phase peptide ions suitable for hydrogen transfer. The use of the deuterated peptides AAA(d3)AA, AAG(d2)AA, AAAG(d2)A and AAAAA-d7 in negative-ion HCD experiments gave information about the source of hydrogen in the formation of c- and y-ions. The results obtained indicated that the major source of hydrogen, over 50% in the rate of intramolecular hydrogen transfer to form c- and y-ions in the peptides AAG(d2)AA and AAAG(d2)A, is the backbone amide nitrogen (Nb), while another source is the backbone Cα sites. The hydrogen did not originate from the Cβ sites. In the case of the peptide AAAAA-d7, the hydrogen to form y-ions comes from the Nb sites with the rate of 69%, while the hydrogen to form c1 and c2 ions comes from the Nb sites with the rate of 41.5 and 14.4%, respectively. For the intramolecular hydrogen transfer to form negative c- and y-ions, it is suggested that deprotonated peptides [M–H]− transiently form at least 3-, 4-, 6-, 7-, 8-, and 9-membered ring transition states, indicating the flexibility of gas-phase peptides.