Ion Mobility Mass Spectrometry as a Potential Tool To Assign Disulfide

Mar 19, 2013 - Laboratory of Mass Spectrometry, GIGA-R, Department of Chemistry, University of Liege, Liege, Belgium. ‡. iBiTec-S, SIMOPRO, Commissa...
1 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Ion Mobility Mass Spectrometry as a Potential Tool To Assign Disulfide Bonds Arrangements in Peptides with Multiple Disulfide Bridges Julien Echterbille,† Loïc Quinton,† Nicolas Gilles,‡ and Edwin De Pauw*,† †

Laboratory of Mass Spectrometry, GIGA-R, Department of Chemistry, University of Liege, Liege, Belgium iBiTec-S, SIMOPRO, Commissariat à l’Energie Atomique, Gif-sur-Yvette, France

‡

S Supporting Information *

ABSTRACT: Disulfide bridges play a major role in defining the structural properties of peptides and proteins. However, the determination of the cysteine pairing is still challenging. Peptide sequences are usually achieved using tandem mass spectrometry (MS/MS) spectra of the totally reduced unfolded species, but the cysteine pairing information is lost. On the other hand, MS/MS experiments performed on native folded species show complex spectra composed of nonclassical ions. MS/MS alone does not allow either the cysteine pairing or the full sequence of an unknown peptide to be determined. The major goal of this work is to set up a strategy for the full structural characterization of peptides including disulfide bridges annotation in the sequence. This strategy was developed by combining ion mobility spectrometry (IMS) and collision-induced dissociation (CID). It is assumed that the opening of one S−S bridge in a peptide leads to a structural evolution which results in a modification of IMS drift time. In the presence of multiple S−S bridges, the shift in arrival time will depend on which disulfide(s) has (have) been reduced and on the shape adopted by the generated species. Due to specific fragmentations observed for each species, CID experiments performed after the mobility separation could provide not only information on peptide sequence but also on the localization of the disulfide bridges. To achieve this goal, synthetic peptides containing two disulfides were studied. The openings of the bridges were carried out following different experimental conditions such as reduction, reduction/alkylation, or oxidation. Due to disulfide scrambling highlighted with the reduction approaches, oxidation of S−S bonds into cysteic acids appeared to be the best strategy. Cysteine connectivity was then unambiguously determined for the two peptides, without any disulfide scrambling interference.

D

detected thanks to a loss of H2S2, resulting in the formation of two dehydroalanine residues in place of the previous disulfide bridge.13 It is worth noting that asymmetric fragmentation can also be observed in negative ions fragmentation spectra.14,15 Disulfide bonds have been studied by electron-mediated dissociation techniques [electron-capture dissociation (ECD) and electron-transfer dissociation (ETD)] due to their high electron affinity.16−21 Wu et al. demonstrated the possibility to decipher complicated disulfide bond patterns of recombinant proteins22,23 from a combination of CID and ETD experiments. Recently, Cole et al. studied 13 intrachain disulfide-linked peptides by ETD. They demonstrated that both backbone and disulfide bridges were cleaved during a single ETD event.24 Moreover, a spectral signature of new c/z-type ions (c − 33/z + 33, c + 32/z − 32) was observed providing a large sequence information. Matrix-assisted laser desorption ionization (MALDI) mass spectrometry has also been considered to characterize peptides that contain disulfide bonds. Disulfide bonds show fragmentation

isulfide bonds have a prominent role in the stabilization of the structure and the redox reactivity of peptides and proteins.1,2 They are encountered in many families of proteins such as albumins, insulins, antibodies, or also toxins from animal venoms. If peptide sequences are obtained through different mass spectrometry based approaches,3 the determination of the cysteine pairing remains a real analytical challenge. Although several papers describe the detection and the fragmentation of disulfide bridges in peptides using mass spectrometry,4−8 none of them proposes a solution to determine the cysteine pairing in peptide containing several S−S bonds. Low-energy collisioninduced dissociation (CID) has been successfully used to fragment peptide toxins containing a single disulfide bond.9 The authors observed an atypical fragmentation explained by the presence of a proline within the S−S loop. Xxx-Pro peptide bond becomes more fragile, fragments, and undergoes asymmetric fragmentation of the disulfide bond,10 leading to a disulfohydryl y-ion type (mass shift of +32 amu) and a dehydroalanine b-ion type (mass shift of −34 amu).8,11 High-energy CID induces not only the cleavage of the peptide backbone12 but also fragmentation of the amide bonds included in the disulfide loop. Negative ion mode CID has been also employed to assign disulfide bridges. In this case, intramolecular disulfide bond is © 2013 American Chemical Society

Received: December 19, 2012 Accepted: March 19, 2013 Published: March 19, 2013 4405

dx.doi.org/10.1021/ac303686w | Anal. Chem. 2013, 85, 4405−4413

Analytical Chemistry

Article

Concerning α-CnIa peptide, the capillary and the sampling cone voltages were set at 1.2 kV and 25 V, respectively. The source temperature was 80 °C. For the mobility separation of partially reduced peptide, a pressure of 2.639 mbar in N2 was used for the IMS cell, 46.4 μbar in the trap device, and 0.583 mbar in the He cell. For the partially reduced and alkylated peptide and the partially oxidized species, the pressure was increased to 2.715 mbar in the mobility cell, 57.5 μbar in the trap device, but reduced to 0.249 mbar in the He cell. The wave parameters were adjusted in order to obtain the best separation (wave height 40, 36, and 38 V and wave velocity 1100, 950, and 950 m/s for the partially reduced peptide, partially reduced and alkylated, and partially oxidized peptide, respectively). The accelerating voltage of the transfer cell (filled with Ar) needed for efficient CID experiment was 23.5 V in the positive mode and 45 V during negative mode analysis. Concerning α-GI peptide, the same kinds of conditions were used for the ion mobility and mass spectrometry. Briefly, the capillary and the sampling cone voltages were fixed at 1 kV and 25 V, respectively. The source temperature was 80 °C. For the mobility separation of partially reduced toxin, a pressure of 2.921 mbar in N2 was used for the IMS cell, 61.2 μbar in the trap device, and 0.263 mbar in the He cell. For the partially reduced and alkylated peptide, the pressure was decreased to 2.752 mbar in the mobility cell, 50.4 μbar in the trap device, and 0.250 mbar in the for the He cell. In the case of partially oxidized peptide, pressure was 2.627 mbar in the mobility cell, 48.9 μbar in the trap device, and 0.242 mbar in the for the He cell. The wave parameters for the best separation of species were the following: wave height 40, 38, and 33.5 V and wave velocity 1300, 1050, and 1000 m/s for the partially reduced toxin, partially reduced and alkylated, and partially oxidized toxin, respectively. The accelerating voltage of the transfer cell (filled with Ar) needed for efficient CID experiment was 23 V in the positive mode and 33 V during negative mode analysis. Each spectrum was recorded after an acquisition time of 25 min. Arrival time distributions were analyzed with MassLynx 4.1 software (Waters Co., Manchester, U.K.). MS/MS spectra were interpreted manually with the help of fragmentation simulators. Reduction Procedure. Partial reductions of toxins were performed using a slight excess of tris(carboxyethyl)phosphine (TCEP).42 A volume of 10 μL of 10 μM solution of peptide was incubated with 10 μL of TCEP 400 μM at 56 °C for 30 min at pH 4.5 or 2. The mixture of native, reduced, and partially reduced toxins was purified using ZipTip C18 (Millipore, Billerica, MA, U.S.A.) according to the manufacturer’s protocol (http://www. millipore.com/userguides/tech1/pr02358) and eluted by 20 μL of 50/50 methanol/formic acid 0.1% (v/v). This solution was directly loaded into the nESI needles for IMS analysis. Reduction and Alkylation Procedure. With Iodoacetamide. A volume of 3 μL of 50 μM solution of toxin was incubated with 10 μL of DTT 5 mM, 2 μL of IAA 100 mM, and 5 μL of NH4HCO3 500 mM buffer at pH 8.5 and 56 °C for 30 min in the dark. Samples were purified using ZipTip C18 and eluted by 20 μL of 50/50 methanol/formic acid 0.1% (v/v). This solution was directly loaded into the nESI needles for IMS analysis. With N-Ethylmaleimide. A volume of 3 μL of 50 μM solution of toxin was incubated with 10 μL of TCEP 400 μM, 2 μL of NEM 100 mM, and 5 μL of formic acid 0.1% at pH 4.5 and 56 °C for 30 min in the dark. Samples were purified using ZipTip C18 and eluted by 20 μL of 50/50 methanol/formic acid 0.1% (v/v).

in postsource and in-source decay.25−27 The identification of fragments can lead to the determination of the disulfide reticulation pattern.28−30 All these studies are helpful to understand the behavior of disulfide bonds during fragmentation experiments. However, they remain powerless when the peptides bear several disulfides bonds and when the cysteine pairings are unknown. Off-line techniques have, however, been developed to elucidate disulfides patterns. Most of them are based on (multi)enzymatic digestion of isolated peptides and high-performance liquid chromatography (HPLC) separation of the resulting fragments to detect cross-linked peptides.4−8,29,31,32 The analysis of cysteine pairings is made even more difficult by an interfering phenomenon called disulfide scrambling.33−38 Disulfide bond rearrangement could be promoted by several factors as, for instance, the presence of free sulfhydryl groups at neutral or alkaline pH. This phenomenon has been observed during various experiments such as enzymatic cleavage at slightly alkaline pH (tryptic digestion), reduction of peptides (e.g., insulin) with dithiothreitol in phosphate or ammonium bicarbonate buffer,4,36 or heat-stressed stability analysis of monoclonal antibodies.38 Disulfide scrambling has also been observed during the MALDI process thanks to a photoinduced radical recombination.15 In this study, ion mobility mass spectrometry39 (IMS) is combined to CID to provide in one step not only the sequence of structured peptides but also their disulfide pairings. Peptide toxins containing two disulfide bonds were reduced in solution or reduced and alkylated to generate a mixture of oxidized, reduced, and partially reduced forms (corresponding to the reduction of one of the two bridges).40,41 The isobaric species are separated by IMS according to their shapes since a progressive increase of the collisional cross section was expected after partial or full reduction. Separated ions are online subjected to CID to obtain characteristic tandem mass spectrometry (MS/MS) spectra of each species. As the nature of the fragments obtained is directly linked to the disulfide pattern of the peptide, a critical study of the fragmentation spectra can lead to the determination of the cysteines pairings. However, disulfide scrambling was observed for each peptide, leading to the loss of the investigated cysteine pairing. Partial oxidation of cysteines into cysteic acids was then considered. Since the opening of disulfide coincides with the addition of a protecting group (oxygen atoms), scrambling of disulfides was totally avoided, allowing the full characterization of the peptides, from their sequences to the cysteine pairings.

■

MATERIALS AND METHODS Materials. Dithiothreitol (DTT) was purchased from Affymetrics USB Products (Cleveland, OH, U.S.A.). Toxins α-CnIa and α-GI, tris(carboxyethyl)phosphine (TCEP), ammonium hydrogen carbonate (NH4HCO3), iodoacetamide (IAA), N-ethylmaleimide (NEM), 3-chloroperoxybenzoic acid (3-CPBA), and formic acid (FA) were purchased from Sigma-Aldrich (Steinheim, Germany), and methanol (MeOH) was obtained from Biosolve (Valkenswaard, The Netherlands). All these chemicals were used without further purification. Ion Mobility and Mass Spectrometry. All experiments were carried out on a Waters Synapt-G2 HDMS (Waters Company, Manchester, U.K.) instrument in positive and negative ionization mode in the so-called “resolution mode”. A volume of 7.5 μL of sample was loaded into the tip capillary and sprayed via the nano-ESI source of the spectrometer (∼110 nL/min). Borosilicate nano-ESI (nESI) emitters were purchased from Proxeon (Odense, Denmark) and from New Objective (Woburn, MA, U.S.A.). 4406

dx.doi.org/10.1021/ac303686w | Anal. Chem. 2013, 85, 4405−4413

Analytical Chemistry

Article

Table 1. Sequences and Modifications Observed in the Two-Peptide Toxin Used in This Study conotoxin

sequence

mass (Da)

cysteine connectivity

other PTMs

α-CnIa

GRC1C2HPAC3GKYYSC4*

1541.6

* = C-ter amidation

α-GI

EC1C2NPAC3GRHYSC4*

1436.5

C1−C3 C2−C4 C1−C3 C2−C4

* = C-ter amidation

(data not shown). It corresponds to a double fragmentation of the peptide bond at each side of the same amino acid (Scheme 1, pathway 5). For the totally reduced peptide (arrival time of 7.38 ms), large b-, y-, and a-ions are detected, driving to the full characterization of the peptide sequence (data not shown). Fragmentation mass spectra of partially reduced species were analyzed to obtain information on cysteine pairings thanks to classical fragments (y-, b-, or a-ions) but also to their homologous showing a loss of 1 or 34 Da or a gain of 32 Da (Scheme 1).14 The study of fragment ions extracted from arrival time of 5.93 ms allows determining that the only relevant possibility explaining the fragmentation profile is the presence of a linkage between Cys2 and Cys4 (Figure 2). First, ions due to residue losses are visible at m/z of 691.41 (loss of Y), 609.85 (loss of YY), 724.43 (loss of P), 545.69 (loss of KYY), and 566.21 (loss of YYS). These fragments lead us to identify Cys4 as involved in a disulfide bridge, which is confirmed by other fragment ions (b5 − 1, b6 − 1, and b7 − 1 as well as y9 − 1), related to the symmetric cleavage of the disulfide bond. The ions a3 (m/z 289.14) and y11 demonstrates the involvement of Cys2 instead of Cys1 in the bond. The presence of these ions is critical to assign correctly the disulfides as Cys1 and Cys2 are vicinal. Combined information attests that the remaining disulfide for this species is Cys2− Cys4, and consequently, the other disulfide bond is Cys1−Cys3. CID spectrum of the second conformation corresponding to a semireduced peptide was extracted from the ATD (6.31 ms). In this fragmentation spectrum, a large series of b-type ions and y-type ions are found (b8−b13 and y1−y3 and y6). It is consequently possible to conclude that Cys4 is not included in the disulfide bridge. Second, several [yi − 1]-type ions are visible (y7 − 1 to y9 − 1 and y11 − 1) with [bj − 1]-type ions (b5 − 1 to b7 − 1). The position of these ions indicates that Cys1−Cys3 is maintained in this conformation. The critical ion to conclude that Cys2 is not included in the disulfide bridge is the y11 − 1 ion. The importance of the detection of y11 and b3 ions is explained in Figure 3. The distinction between each possible conformation based on detection of particular fragments is highlighted on the same figure. The fragmentation spectrum related to the third semireduced peptide was extracted from the peak at 6.66 ms. This spectrum contains enough information to determine that the remaining disulfide in this conformation is between Cys3 and Cys4 due to b2, b4−b7 ions and y7−y9, y11, and y12 ions. This means that neither Cys1 nor Cys2 can be included in a disulfide bond. All this information confirms that a nonexpected disulfide bridge between Cys3 and Cys4 was formed during the experiment and linked to a probable disulfide reformation process. Case of α-GI (Two Disulfide Bonds). α-GI is a 13 amino acid peptide containing four cysteine residues reticulated into two disulfide bridges (Cys1−Cys3 and Cys2−Cys4). Once the partial reduction was performed, ion mobility analysis revealed four contributions in the ATD of the triply charged ions (see Supporting Information Figure S-1). For the m/z ratio corresponding to the partially reduced peptides, two contributions are detected, which seems to indicate that no scrambling phenomenon has occurred in this case (see Supporting Information Figure S-2).

This solution was directly loaded into the nESI needles for IMS analysis. Oxidation Procedure. A volume of 3 μL of 50 μM solution of peptide was incubated with 10 μL of 3-CPBA 75 μM and 7 μL of water during 3 h at room temperature and in the dark. Samples were purified using ZipTip C18 and eluted by 20 μL of 50/50 methanol/water. This solution was directly loaded into the nESI needles for IMS analysis.

■

RESULTS AND DISCUSSION The two peptides of interest contain two disulfide bridges (Table 1). It is important to note that peptide toxins often show adjacent cysteines in their sequences, which add a supplementary difficulty to decipher disulfide patterns, since the CID experiment must be efficient to obtain characteristic ions resulting from the cleavage of the peptide bond between the two cysteine residues. Experiments 1: Partial Reduction of the Two Peptides. Case of α-CnIa (Two Disulfide Bonds). Two different charge states are detected in the spectra (doubly and triply charged peptides). The triply charged species was chosen for structure determination since the separation obtained in the arrival time distribution (ATD) was better than for the doubly charged species. The sample after reduction is more complex than expected (Figure 1). The extraction of the ATD for the native folded peptide (m/z 515.567) indicates two contributions suggesting two conformations. This result is quite unexpected since only one contribution was detected before the reduction step (Figure 1). The mobility signal at 5.58 ms, which corresponds to the most compact structure, was identified to be the native peptide by comparison with the drift time of the peptide before reduction (Figure 1).The isobaric compound detected at 6.15 ms was not clearly identified. A possible explanation could be a random (noncontrolled) reformation of the disulfides during the reduction reaction, the oxygen dissolved within the solvent playing the role of the oxidant (redox potential of a single disulfide bridge, −0.22 V/SHE;43 redox potential of O2 (pH = 7), 0.815 V/SHE). In other words, this species could correspond to the peptide with another cysteine pairing. The peak at 7.38 ms is less ambiguous since it was identified as the fully reduced peptide (M = 1545.6 Da), corresponding to the larger cross section. In the case of the partially reduced species (m/z 516.243), the ATD reveals three contributions suggesting the presence of three different conformations, each corresponding to a partially reduced form (only one disulfide reduced). While two conformations were expected, the detection of three contributions can be explained by different ways. As discussed above, a disulfide reformation process can be suspected. Another assumption could also be the highlight of two conformations for the same species, as has already been seen in proteins.44 To rationalize these observations, each of the compounds was fragmented to generate characteristic fragments, aiming the localization of disulfide linkage. The CID spectrum of the native peptide (arrival time 5.58 ms) shows unusual amino acids fragmentation, within the loops 4407

dx.doi.org/10.1021/ac303686w | Anal. Chem. 2013, 85, 4405−4413

Analytical Chemistry

Article

Figure 1. Up: Before partial reduction, the native toxin shows only one contribution in the mobilogram suggesting that α-CnIa has has only one conformation in the gas phase. Down: Total arrival time distribution (ATD) of the partially reduced α-CnIa. The mass spectrum extracted clearly shows a mixture of oxidized, reduced, and semireduced species. From each contribution in the mass spectrum, ATD of each species can be extracted. Some unexpected results are obtained from the ATD extraction of each m/z ratio. The ATD of the oxidized form (in blue) reveals two conformations for a single value of m/z ratio. Semireduced forms (in purple) are, for their part, present in three conformations according to the extracted ATD. However, regarding the ATD of the reduced species (in red), and as expected, only one conformation is detected.

Symmetric and asymmetric cleavage ions of the disulfide link are also observed (b9 − 34, b10 − 1, and b12 − 34 ions and y2 − 1 and y6 − 1 ions). All this information confirms that a nonexpected disulfide bridge between Cys3 and Cys4 was formed during the experiment and linked to a reoxidation process. Several experimental conditions were explored in order to avoid the cysteines pairing scrambling. The simplest method was to increase the acidity of the reaction buffer to pH 2. Under these conditions, the presence of a thiol is favored over a negative reactive thiolate on the cysteine side chain and it can be expected that the scrambling phenomenon will be lessened. However, the consequent ATDs of both toxins presents again several conformations. Fragmentation spectra confirm that the rearrangement

However, the CID spectrum extracted from the contribution at 7.92 ms shows clues leading to the unexpected connectivity Cys2− Cys3. Indeed, a series of a- and b-type ions and y-type ions are found (a10, a11, and b2, b9, and y1, y2, y3, y6, y11, and y12), signifying that Cys4 is not a part of a disulfide as well as Cys1 since b2 and y11 ions are found meaning that Cys2 and Cys3 are connected. Additional evidence is given by fragments produced by a double fragmentation of the peptide bond (loss of Pro and Ala) or from a symmetric cleavage of the disulfide link (y9 − 1, y10 − 1, and b3 − 1). The second contribution in the ATD (8.41 ms) corresponds to the Cys3 and Cys4 pairing. This is clearly highlighted by the presence of a2, b4, and y7−y12 ions, indicating that Cys1 and Cys2 are not included into a disulfide bond. 4408

dx.doi.org/10.1021/ac303686w | Anal. Chem. 2013, 85, 4405−4413

Analytical Chemistry

Article

Figure 2. CID spectra extracted from each contribution of the semireduced species ATD. Thanks to the typical fragments found in each spectrum, it is possible to determine the connectivity between cysteines and highlight a scrambling of disulfide bonds. (A) The y11 ion allows us to rule out the involvement of cysteine 1 in the remaining disulfide bridge. Other ions are present corresponding to a symmetric dissociation of the disulfide (following pathway 2 in Scheme 1). Double arrows (in purple) symbolize a double dissociation of the peptide bond resulting in the loss of an amino acid (Scheme 1, pathway 5). All these ions are clues to identify cysteine 3 as not being part of a disulfide bridge. (B) As for the panel A, classical b- and y-type ions are observed; double fragmentations and symmetric dissociation of the remaining disulfide reveal the combination of cysteine 1 and 3 in the bridge. (C) In this case, classical b- and y-type ions are as many clues about the absence of cysteines in the bridge found in this conformation. For this last conformation and based on the linkage pattern described in the literature for this toxin, it is clear that a scrambling phenomenon is occurring.

to block free thiols as quickly as possible. After infusion with nano-ESI, the ion corresponding to the partially reduced/ alkylated species (mass of the peptide + 114 Da) was selected in the quadrupole before the ion mobility separation and then fragmented. ATD obtained for both toxins studied here are given in the Supporting Information (Figures S-3 and S-4 for the partial

phenomenon still occurs. Disulfide configurations deduced from these data are in agreement with those obtained at pH 4 (data not shown). The idea was then to circumvent the rearrangement phenomenon by alkylating the sulfhydryl groups with IAA. Alkylating agent was added immediately in the reaction medium 4409

dx.doi.org/10.1021/ac303686w | Anal. Chem. 2013, 85, 4405−4413

Analytical Chemistry

Article

Scheme 1. Fragmentation Pathways for Peptides Containing Disulfide Bondsa

a

Fragments obtained are classical b- or y-type ions, but some gain or loss of masses is observed. A loss of 1 amu compared to the mass of the reduced species corresponds to a symmetrical fragmentation between the two sulfurs (pathway 2). A loss of 34 Da is observed in the case of asymmetrical fragmentation between a carbon and a sulfur driving to a loss of a sulfur atom and a dehydroalanine formation (pathways 1 and 3). On the other hand, a gain of 32 Da stands for addition of a sulfur atom at the other side of the disulfide bond after the asymmetric fragmentation (pathways 1 and 3). For pathway 5, a double fragmentation of the peptide bond induces the loss of the amino acid (lysine residue, for instance).

Information Figures S-5 and S-6 for the partial reduction and alkylation of α-CnIa and α-GI, respectively). The ideal way to completely prevent the scrambling phenomenon could be the opening of the disulfide simultaneously to the modification of the thiol group. Oxidation of disulfides into sulfonic acids seems to be an interesting way to study as the opening of the disulfide coincides with the addition of an oxygen atom. Oxidation. Case of α-CnIa. Disulfide bonds are opened thanks to the addition of oxygen atoms on the sulfur leading to the formation of a sulfonate function (R−CH2−SO3H). This particular group allowed performing nano-ESI infusion in the negative mode. After a selection of the ion corresponding to the peptide with two sulfonate groups in the quadrupole (M + 96 Da), the ATD showed only one conformation, whereas two conformations could be expected: one corresponding to the first disulfide opened and the other corresponding to the second disulfide opened. A possible explanation could be a difference in terms of electrochemical potential for oxidation so that only one of the bridges may be oxidized. In order to corroborate this hypothesis, additional experiments have to be done. For instance, a longer time of reaction (or a higher concentration of 3-CPBA) should give access to more oxidized species. A selection of the m/z ratios corresponding to the peptide with seven or eight additional oxygen atoms (peptide toxin with two sulfonate groups and with a partial oxidation of a disulfide bond) could be instructive about the way in which oxygen atoms are added on disulfides during the oxidation process. The CID spectrum was extracted from this arrival time (9.07 ms) and is shown in Figure 4. In this one, the only possibility of a disulfide linkage consists on a bond between Cys1 and Cys3. First, a loss of H2S2, usual for disulfide-linked peptides in negative mode,13 is observed. Due to this loss, classical fragmentation is allowed to occur on basis of the subsequent peptide containing two dehydroalanine residues. The position of these unusual amino acids in the sequence allows the assignment of the S−S bridge. Several b- and y-type ion masses match with the peptide containing a disulfide between Cys1 and Cys3,

Figure 3. Importance of the presence of y11 and/or b3 ions in order to assign connectivity between cysteines. Depending on which disulfide bond is still formed, y11 and b3 ions will be detected or not. Likewise, ions belonging to the “release part” of the peptide could be detected as indicated in the different panels.

reduction and alkylation of α-CnIa and α-GI, respectively). Clearly, multiple conformations are detected meaning that scrambling phenomenon occurred even if the alkylation is performed in the same time than partial reduction. As the alkylation reaction with IAA requires an alkaline pH for an efficient progress, it can also promote the scrambling of disulfide bonds. An alkylation in acidic conditions was then tested with NEM.45,46 Unfortunately, this approach gives also rise to contributions in the ATD appearing as non-native disulfide bridges configuration for both studied peptides (Supporting 4410

dx.doi.org/10.1021/ac303686w | Anal. Chem. 2013, 85, 4405−4413

Analytical Chemistry

Article

Figure 4. Oxidation of toxins with 3-chloroperoxybenzoic acid leads to the formation of several partially oxidized species. The semioxidized species of each toxin is selected in the quadrupole of the instrument. (A) α-CnIa toxin: it reveals only one contribution in the ATD. The CID spectrum extracted from this single contribution allows concluding the connectivity of the cysteines. The cysteines connected by the remaining disulfide bond are colored in purple; the indication in brackets means that during the CID experiment a loss of H2S2 occurred leading to the formation of two dehydroalanines. (B) α-GI toxin: the same kinds of results are obtained in this case with same conclusions.

transformed into dehydroalanine residues, and oxidized in sulfonate Cys2 and Cys4. Additional internal fragments are detected by taking into account these modifications, reinforcing our hypothesis. All this information leads us to attest that the disulfide bond formed in this conformation was Cys1−Cys3. Case of α-GI. Partial oxidation of cysteines was also investigated for toxin α-GI. The m/z ratio corresponding to the

peptide toxin with two sulfonate groups provided the ATD as shown in Figure 4. As with the previous peptide, a single contribution was obtained for this ATD, signifying the presence of a single conformation (6.46 ms). The CID spectrum was extracted from the arrival time of 6.46 ms and is shown in Figure 4. Once again, Cys1 and Cys3 linkage is the most relevant possibility of binding in this conformation. Indeed, H2S2 loss takes place again, 4411

dx.doi.org/10.1021/ac303686w | Anal. Chem. 2013, 85, 4405−4413

Analytical Chemistry

Article

Notes

inducing a conventional fragmentation of the peptide containing two dehydroalanine residues. Several b-type and y-type ions perfectly match with this peptide, if the two dehydroalanines are localized at position 2 and 7 (in place of Cys1 and Cys3) and the two cysteic acids in position 3 and 13 (in place of Cys2 and Cys4). Some intense signals were consequently assigned to internal fragments, confirming our hypothesis. Partial oxidation is an interesting alternative to the reduction of peptide to avoid disulfide scrambling in solution. The addition of an oxygen atom coincides with the rupture of disulfide bond, blocking immediately the reactive sulfur. Moreover, the oxidation opens only one of the disulfide bridges. The strategy allowed the assignment of the cysteine pairings of the two studied peptides.

The authors declare no competing financial interest.

■

REFERENCES

(1) Creighton, T. E. BioEssays 1988, 8 (2−3), 57−63. (2) Matsumura, M.; Signor, G.; Matthews, B. W. Nature 1989, 342 (6247), 291−293. (3) Zhang, Y., Fonslow, B. R., Shan, B., Baek, M.-C., Yates, J. R., Chem. Rev. 2013, in press. (4) Gorman, J. J.; Wallis, T. P.; Pitt, J. J. Mass Spectrom. Rev. 2002, 21 (3), 183−216. (5) Janecki, D. J.; Nemeth, J. F. J. Mass Spectrom. 2011, 46 (7), 677− 688. (6) Jones, M. D.; Patterson, S. D.; Lu, H. S. Anal. Chem. 1998, 70 (1), 136−143. (7) Kallestad, J. C.; Shoyab, M.; Linsley, P. S. J. Biol. Chem. 1991, 266 (14), 8940−8945. (8) Mormann, M.; Eble, J.; Schwöppe, C.; Mesters, R.; Berdel, W.; Peter-Katalinić, J.; Pohlentz, G. Anal. Bioanal. Chem. 2008, 392 (5), 831−838. (9) Thakur, S. S.; Balaram, P. Rapid Commun. Mass Spectrom. 2007, 21 (21), 3420−3426. (10) Papayannopoulos, I. A. Mass Spectrom. Rev. 1995, 14 (1), 49−73. (11) Lioe, H.; OrsHair, R. J. Am. Soc. Mass Spectrom. 2007, 18 (6), 1109−1123. (12) Bean, M. F.; Carr, S. A. Anal. Biochem. 1992, 201 (2), 216−226. (13) Bilusich, D.; Bowie, J. H. Mass Spectrom. Rev. 2009, 28 (1), 20−34. (14) Chrisman, P. A.; McLuckey, S. A. J. Proteome Res. 2002, 1 (6), 549−557. (15) Zhao, L.; Almaraz, R.; Xiang, F.; Hedrick, J.; Franz, A. J. Am. Soc. Mass Spectrom. 2009, 20 (9), 1603−1616. (16) Chrisman, P.; Pitteri, S.; Hogan, J.; McLuckey, S. J. Am. Soc. Mass Spectrom. 2005, 16 (7), 1020−1030. (17) Duan, X.; Engler, F. A.; Qu, J. J. Mass Spectrom. 2010, 45 (12), 1477−1482. (18) Chowdhury, S.; Munske, G.; Ronald, R.; Bruce, J. J. Am. Soc. Mass Spectrom. 2007, 18 (3), 493−501. (19) Cooper, H. J.; Håkansson, K.; Marshall, A. G. Mass Spectrom. Rev. 2005, 24 (2), 201−222. (20) Wiesner, J.; Premsler, T.; Sickmann, A. Proteomics 2008, 8 (21), 4466−4483. (21) Zubarev, R. A.; Kruger, N. A.; Fridriksson, E. K.; Lewis, M. A.; Horn, D. M.; Carpenter, B. K.; McLafferty, F. W. J. Am. Chem. Soc. 1999, 121 (12), 2857−2862. (22) Wu, S.-L.; Jiang, H.; Hancock, W. S.; Karger, B. L. Anal. Chem. 2010, 82 (12), 5296−5303. (23) Wu, S.-L.; Jiang, H.; Lu, Q.; Dai, S.; Hancock, W. S.; Karger, B. L. Anal. Chem. 2008, 81 (1), 112−122. (24) Cole, S.; Ma, X.; Zhang, X.; Xia, Y. J. Am. Soc. Mass Spectrom. 2012, 23 (2), 310−320. (25) Crimmins, D. L.; Saylor, M.; Rush, J.; Thoma, R. S. Anal. Biochem. 1995, 226 (2), 355−361. (26) Qiu, X.; Cui, M.; Li, H.; Liu, Z.; Liu, S. Rapid Commun. Mass Spectrom. 2007, 21 (21), 3520−3525. (27) Zhou, J.; Ens, W.; Poppe-Schriemer, N.; Standing, K. G.; Westmore, J. B. Int. J. Mass Spectrom. Ion Processes 1993, 126 (0), 115− 122. (28) Schnaible, V.; Wefing, S.; Resemann, A.; Suckau, D.; Bücker, A.; Wolf-Kümmeth, S.; Hoffmann, D. Anal. Chem. 2002, 74 (19), 4980− 4988.

CONCLUSION The aim of this work was to bring the proof of concept of a new method for disulfide bond assignment of peptides based on IMS combined to CID experiments. For native species, only one contribution is detected in ATD leading to the conclusion that only one fold was present within the limits of IMS resolution. In the CID spectra, fragments are observed including double cleavages of the peptide backbone. We first used reducing agents to obtain the mixture of oxidized, semireduced, and fully reduced disulfide bridges. When analyzing the mixture, several contributions are detected in the ATD. The fully reduced peptides give only one ATD. In the case of oxidized and partially reduced forms, the number of detected ATDs is higher than expected, linked to an interfering phenomenon called disulfide scrambling. CID spectra of IMS-separated partially reduced peptides show specific ions due to the symmetric or asymmetric dissociation of the disulfide bond. These ions allow assigning the connectivity between cysteines. To avoid this interference, two one-step reduction/alkylation of cysteines experiments were performed, one in basic and the other in acidic conditions. Classical y- and b-type ions with a shift in mass due to alkylating group were also observed as well as expected ions due to symmetric and asymmetric cleavage of the disulfide bond. Scrambling of disulfide bonds was still observed. In order to block disulfide scrambling, the oxidation of cysteines into cysteic acids was performed. The ion mode was turned to negative. The oxidation opens only one of the disulfide bridges. A single contribution is detected in the ATDs showing a specificity in the oxidation reaction. Scrambling is not observed. In addition to classical ions, CID spectra show the loss of H2S2, transforming cysteines of the remaining disulfide bridge into dehydroalanine. This allows assigning the cysteines’ connectivities. Our future work will focus on the more complex structures (three or more disulfide bridges) for which the number of possible conformations is larger. ASSOCIATED CONTENT

S Supporting Information *

Time distributions (ATDs) and CID spectra obtained for semireduced α-GI conotoxin and also semireduced and alkylated α-GI and α-CnIa conotoxins. This material is available free of charge via the Internet at http://pubs.acs.org.

■

ACKNOWLEDGMENTS

The authors acknowledge the FRS-FNRS for its financial support for the instrumentation, the Fonds Européen de développement regional (FEDER), and the Walloon region for financial support. We also acknowledge the Research Council of the University of Liege for support through the “Fonds spéciaux de recherche”

■

■

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. 4412

dx.doi.org/10.1021/ac303686w | Anal. Chem. 2013, 85, 4405−4413

Analytical Chemistry

Article

(29) Tie, J.-K.; Mutucumarana, V. P.; Straight, D. L.; Carrick, K. L.; Pope, R. M.; Stafford, D. W. J. Biol. Chem. 2003, 278 (46), 45468− 45475. (30) Qin, J.; Chait, B. T. Anal. Chem. 1997, 69 (19), 4002−4009. (31) Horn, N. A.; Hurst, G. B.; Mayasundari, A.; Whittemore, N. A.; Serpersu, E. H.; Peterson, C. B. J. Biol. Chem. 2004, 279 (34), 35867− 35878. (32) Leal, W. S.; Nikonova, L.; Peng, G. FEBS Lett. 1999, 464 (1−2), 85−90. (33) Brgles, M.; Kurtović, T.; Halassy, B.; Allmaier, G.; MarchettiDeschmann, M. J. Mass Spectrom. 2011, 153−162. (34) Fagerquist, C. K. Rapid Commun. Mass Spectrom. 2004, 18 (6), 685−700. (35) Gilbert, H. F. Methods Enzymol. 1995, 251, 8−28. (36) Morris, H. R.; Pucci, P.; Panico, M.; Marino, G. Biochem. J. 1990, 268 (3), 803−806. (37) Wang, X.; Kumar, S.; Singh, S. Pharm. Res. 2011, 28 (12), 3128− 3144. (38) Wang, Y.; Lu, Q.; Wu, S.-L.; Karger, B. L.; Hancock, W. S. Anal. Chem. 2011, 83 (8), 3133−3140. (39) Creaser, C. S.; Griffiths, J. R.; Bramwell, C. J.; Noreen, S.; Hill, C. A.; Thomas, C. L. P. Analyst 2004, 129 (11), 984−994. (40) Favreau, P.; Krimm, I.; Le Gall, F.; Bobenrieth, M.-J.; Lamthanh, H.; Bouet, F.; Servent, D.; Molgo, J.; Menez, A.; Letourneux, Y.; Lancelin, J.-M. Biochemistry 1999, 38 (19), 6317−6326. (41) Guddat, L. W.; Martin, J. A.; Shan, L.; Edmundson, A. B.; Gray, W. R. Biochemistry 1996, 35 (35), 11329−11335. (42) Burns, J. A.; Butler, J. C.; Moran, J.; Whitesides, G. M. J. Org. Chem. 1991, 56 (8), 2648−2650. (43) Jocelyn, P. C. Eur. J. Biochem. 1967, 2 (3), 327−331. (44) Shelimov, K. B.; Clemmer, D. E.; Hudgins, R. R.; Jarrold, M. F. J. Am. Chem. Soc. 1997, 119 (9), 2240−2248. (45) Paulech, J.; Solis, N.; Cordwell, S. J. Biochim. Biophys. Acta, Proteins Proteomics 2013, 1834 (1), 372−379. (46) Smyth, D. G.; Nagamatsu, A.; Fruton, J. S. J. Am. Chem. Soc. 1960, 82 (17), 4600−4604.

4413

dx.doi.org/10.1021/ac303686w | Anal. Chem. 2013, 85, 4405−4413