Article pubs.acs.org/ac
Tandem Mass Spectrometry Using the Atmospheric Pressure Electron Capture Dissociation Ion Source Damon B. Robb,*,† Jeffery M. Brown,‡ Michael Morris,‡ and Michael W. Blades† †
University of British Columbia, Department of Chemistry, Vancouver, BC V6T 1Z1, Canada Waters Corporation, Stamford Avenue, Altrincham Road, Wilmslow, SK9 4AX, United Kingdom
‡
ABSTRACT: Atmospheric pressure electron capture dissociation (AP-ECD) is an emerging technique capable of being adopted to virtually any electrospray mass spectrometer, without modification of the main instrument. To date, however, because the electron capture reactions occur in the ion source, AP-ECD has been limited by its apparent inability to select precursors prior to fragmentation, i.e., to perform tandem mass spectrometry (MS/MS) experiments. In this paper we demonstrate a novel AP-ECD−MS/MS method using an AP-ECD source on a Xevo G2-S quadrupole time-offlight (Q-TOF) mass spectrometer from Waters Micromass. The method takes advantage of the tendency for electron capture reactions to generate charge-reduced “ECnoD” products, species that have captured an electron and have had a covalent bond cleaved yet do not immediately dissociate into separate products and so retain the mass of the precursor ion. In the method, ECnoD products from the AP-ECD source are isolated in the quadrupole mass filter and induced to dissociate through supplemental activation in the collision cell, and then the liberated ECD fragment ions are mass analyzed using the highresolution TOF. In this manner, true MS/MS spectra may be obtained with AP-ECD even though all of the precursors in the source are subjected to electron capture reactions in parallel. Here, using a late-model Q-TOF instrument otherwise incapable of performing electron-based fragmentation, we present AP-ECD−MS/MS results for a group of model peptides and show that informative, high-sequence-coverage spectra are readily attainable with the method.
E
lectron capture dissociation (ECD)1 and electron transfer dissociation (ETD)2 are important new tools for the structural characterization of peptides and proteins by mass spectrometry. In contrast with conventional collision-induced dissociation (CID), these techniques readily enable the localization of labile post-translation modifications (PTMs), and they provide more extensive and uniform sequence coverage. Further, compared to CID, ECD and ETD are outstanding for their effectiveness in cleaving disulfide bonds, as well as in fragmenting increasingly larger species in top-down analyses. Presently ETD is more widely used than ECD, principally because it can be implemented using common radio frequency (RF) ion traps, unlike ECD, which generally requires a Fourier transform ion cyclotron resonance instrument. ETD is now available as an option on numerous high-performance mass spectrometers; however, in addition to a quadrupole mass analyzer to isolate parent ions and reagent anions, and a highresolution mass analyzer such as a reflectron time-of-flight (TOF) or an Orbitrap to analyze the ETD fragments, these instruments also require a dedicated second ion source for generating anionic ETD reagents and some kind of RF ion trap for carrying out the reactions. Consequently, although they provide unprecedented fragmentation and mass analysis capabilities, these modern ETD-equipped instruments are also highly complex and expensive, and thus may be inaccessible to many researchers wishing to employ electron-based fragmentation. © 2014 American Chemical Society
Atmospheric pressure electron capture dissociation (APECD) is an emerging technique capable of being adopted to virtually any electrospray ionization mass spectrometer.3−5 Because AP-ECD requires only an ion source device, which may easily be retrofit to instruments not originally designed for ECD/ETD experiments, it has the potential to make electronbased fragmentation much more widely accessible; with an APECD source, instrument companies could provide ECD functionality on all of their instruments without impacting existing designs, while researchers without access to a suitable ECD/ETD instrument would have new and more affordable options. AP-ECD has been proven to be effective at generating high-quality ECD spectra of peptides, including phospho- and glycopeptides,5,6 and it has been shown to be fully compatible with liquid chromatography (LC), with on-column detection limits in the low fmol range even when coupled with an earlymodel, relatively low performance instrument.5 In comparison with in vacuo ECD/ETD techniques, APECD has the advantage of being very simple in terms of hardware and method requirements. The essential hardware is composed of a handful of basic machined parts, a photoionization detector (PID) lamp, a dc power supply for the Received: January 22, 2014 Accepted: April 2, 2014 Published: April 2, 2014 4439
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
Figure 1. (A) Illustration of an experimental AP-ECD ion source (3D drawing inset) with a Xevo G2-S Q-TOF from Waters Micromass; (B) crosssectional drawing of the ion source.
source block, e.g., with a low-cost cartridge heater. This is in stark contrast with ETD, for which heating the ions is much more challenging, with the current method of choice being to irradiate the reaction cell with IR photons from a CO2 laser,9,10 an approach having the serious drawback of requiring both a dedicated laser and extensive modification of the instrument.11 Moreover, since AP-ECD is inherently a parallel fragmentation technique,12 it appears to be well-suited to techniques such as “MSE”,13 where all precursor ions are fragmented (normally by
lamp, and an optional heater and temperature controller, while the method requires only adding a photoionizable dopant to the source gas flow and switching on the lamp. Accordingly, with AP-ECD there is no need for ion−ion storage and manipulation schemes that add complexity to the hardware and software/firmware of ETD systems. Further, AP-ECD also has a practical advantage in enabling “activated ion”7−10 experiments, because the temperature of the ions in the ECD reaction region can be elevated simply through heating the surrounding 4440
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
light; Cambridge, U.K.). In operation, multiply charged peptide/protein precursor ions are created within the spray chamber by the enclosed nanospray source. These ions are entrained and transported through the spray chamber by the flow of gas to the downstream source block whose central channel may be irradiated by the photoionization lamp. When the lamp is switched off, precursor ions pass through unaffected and the source operates as a normal nanospray source; when the lamp is switched on, photoelectrons are generated by photoionization of a dopant (typically acetone) added to the AUX gas. Photoelectrons are subsequently captured by the precursor ions in the downstream atmospheric pressure reaction/transport zone, leading to charge reduction and cleavage of covalent bonds for the precursors, either with or without prompt dissociation. Ions exiting the source are then delivered through the vacuum interface of the mass spectrometer. In MS/MS mode, charge-reduced ECnoD reaction products are mass-selected in the quadrupole mass filter and induced to dissociate by supplemental activation in the collision cell, and then the ECD fragments are massanalyzed using the high-resolution TOF. To shield the photoelectron generation region of the source block from the electric field of the nanospray source, a thin wire-mesh screen (88% transmission; tungsten wire, 50 × 50 threads/in., wire o.d. = 0.0012 in.; Unique Wire, Hillside, NJ, USA) is secured to the end of the spray chamber, immediately before the source block entrance. The mesh is held between two thin stainless steel washers, spot-welded together to form a reusable assembly that may be handled and cleaned without damage. The mesh assembly is attached to the spray chamber using a threaded stainless steel cap, with an opening matching the i.d. of the spray chamber, which enables disassembly of the apparatus for cleaning without tools. The mesh assembly is at the same electric potential as the spray chamber and source block, and it confines the electric field of the nanospray source to the spray chamber. Note that the mesh must be cleaned when it becomes contaminated with a film of sample residue, because it can then become positively charged and attract electrons away from the precursor ions, causing a partial or complete loss of ECD performance. Preventative maintenance entails periodic cleaning of the spray chamber and mesh assembly. This can be accomplished simply by brushing the central channel of the spray chamber under running water, followed by rinsing both the spray chamber and the mesh with deionized (DI) water and then methanol, and then air-drying and/or blotting with a laboratory wipe. The required frequency of cleaning depends upon how heavily the source is used: weekly may be acceptable in the case of LC/MS experiments, where sample is only injected at intervals, whereas daily cleaning may be required for infusion experiments where sample is delivered continuously for extended periods; that said, even daily cleaning is of little consequence as the entire routine may be completed in 5 min. To heat the source and thereby both improve peptide ionization efficiency and “activate”7−10 the peptide ions during ECD, a single 1/4 in. o.d., 100 W cartridge heater (model CSH101100/120; Omega, Laval, QC, Canada) was installed in the source block. Temperature control was enabled via an embedded insulated thermocouple (model HSTC-TT-J-24S72-SMPW-CC; Omega) adjacent to the heater and a benchtop temperature controller (model CSC32-J; Omega). To prevent radical electron capture products such as z• ions from forming adducts with oxygen,4,21 the source was
CID) and analyzed simultaneously and postacquisition data processing is used to match product ions with precursors. (Note that for MSE to be viable by ECD or ETD, the relative abundance of analyte ions and reagent electrons or anions must be balanced; with AP-ECD, photoionization is capable of generating a large abundance of electronsμA’s are attainable at atmospheric pressure using a PID lamp14which enables entire precursor populations to be reacted with a duty cycle of up to 100%.) Similarly, because the electrons are generated within a confined spatial volume, through which the precursor ions flow, there are no issues related to the simultaneous trapping and storage of reagent, precursor, and fragment ions of (potentially) disparate m/z and abundance. Nevertheless, despite these advantages, because the electron capture reactions occur in the ion source, AP-ECD has thus far been limited by its apparent major shortcoming: an inability to isolate precursors before fragmentation, i.e., to perform true tandem mass spectrometry (MS/MS) experiments. In this paper we demonstrate a novel AP-ECD−MS/MS method, using an AP-ECD source on a Xevo G2-S quadrupole time-of-flight (Q-TOF) mass spectrometer from Waters Micromass. At the same time, we also describe how a latemodel Q-TOF not originally designed for ECD/ETD experiments may be fitted with an AP-ECD source to enable electronbased fragmentation, without requiring modification of the main instrument. The AP-ECD−MS/MS method takes advantage of the universal tendency for electron capture (and transfer) reactions to generate an abundance of charge-reduced ECnoD products, species which have captured an electron and have had a covalent bond cleaved, yet do not immediately dissociate into separate products and so retain the mass of the precursor. The internal fragments of ECnoD (and ETnoD) species are thought to be held together by noncovalent interactions,7 and it is well-established that their components may be separated by gentle supplemental collisional activation prior to analysis, both to improve ECD/ETD signal intensity and to increase sequence coverage.15−20 In the present method, ECnoD species emanating from the AP-ECD ion source are isolated in the mass-resolving quadrupole and induced to dissociate through supplemental activation in the collision cell, and then the liberated ECD fragment ions are mass analyzed using the high-resolution TOF. In this manner true MS/MS spectra may be obtained with AP-ECD even though all the precursors in the source are subjected to electron-capture reactions in parallel. Here, using a Q-TOF instrument otherwise incapable of performing electron-based fragmentation, we present AP-ECD−MS/MS results for a group of model peptides and show that informative, high-sequencecoverage tandem mass spectra are readily attainable with the method.
■
EXPERIMENTAL SECTION AP-ECD Ion Source. Figure 1A is an illustration of the experimental AP-ECD ion source coupled with a Xevo G2-S QTOF mass spectrometer from Waters Micromass (Floats Road, Manchester, U.K.); Figure 1B is a cross-sectional drawing of the source, which is closely modeled after a design described previously.5 In brief, the AP-ECD source is composed of three interconnected subassemblies: a polyimide sprayer plug with an integrated nanospray source; a stainless-steel spray chamber with a central channel and an inlet for an auxiliary (AUX) gas flow; and a stainless steel source block with a dc krypton photoionization lamp (model PKS 100/106; Heraeus Noble4441
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
The scan time per spectrum was 1 s. Each spectrum presented here is the sum of 30 spectra. None of the spectra has been smoothed or centroided, except for melittin’s, which was smoothed using a five point Savitzky-Golay filter and then centroided, followed by charge-state deconvolution with the software’s MaxEnt 3 function (MaxEnt 3 constructs equivalent zero charge spectra from raw mass spectra with ions of multiple charge states).
positioned so that its outlet was on-axis with and very close to (separation ≤ 1 mm) the tip of the sampling cone, the inlet of the instrument’s vacuum interface. The gas exiting the source then swept away the ambient air between it and the cone so that the exposure of radical species to oxygen was minimized. Chemicals. [Glu1]-fibrinopeptide B (“glu-fib”) and substance P were from Sigma-Aldrich (Oakville, Ontario, Canada). Melittin, honey bee (deamidated), was from AnaSpec (San Jose, CA, USA). Phosphorylated cholecystokinin (10−20), glycosylated MUC5AC 3, and glycosylated erythropoietin (EPO) (117−131) were from Protea Biosciences (Morgantown, WV, USA). Methanol (high-performance liquid chromatography (HPLC)-grade) and formic acid (≥95% in water) were from Fisher Scientific (Ottawa, Ontario, Canada). Deionized water was from an in-house generator. The dopant used to promote photoelectron generation was HPLC-grade acetone from Fisher Scientific. Liquid nitrogen boil-off was used for the source gases. For all the samples, the concentration was 1 μM, except for melittin, which was 5 μM. The sample makeup solvent was 50/ 50 methanol/water with 0.1% formic acid. Instrumentation and Methods. The mass spectrometer was an unmodified Xevo G2-S Q-TOF from Waters Micromass, with MassLynx V4.1 software. The sample solutions were infused via an external syringe pump at a flow rate of 0.5 μL min−1. The acetone dopant was delivered continuously at 1.0 μL min−1 via the mass spectrometer’s internal fluidics system. The nanospray voltage was set to 2.2 kV, provided by the capillary high voltage (HV) power supply of the mass spectrometer. The spray chamber and source block offset voltage was set to 200 V by a custom AP-ECD power supply (Electronic Engineering Services, Department of Chemistry, UBC), which also provided the dc voltage (−1.5 kV) for igniting the photoionization lamp (note that the lamp and its power supply were floated at the offset voltage of the spray chamber and source block). The sampling cone voltage of the vacuum interface was set to a low value, 10 V, to avoid CID in the interface region and thus to enhance the transmission of intact ECnoD products. The nebulizer (NEB) gas was from the instrument’s nanoflow gas supply, set to 2.0 bar. The flow rate of the AUX gas was set via an external rotameter to 10 L min−1. The temperature of the AP-ECD source block was 100 °C, while the source temperature of the instrument’s vacuum interface was 80 °C. The experiments involved first acquiring direct parallel fragmentation AP-ECD spectra at both low and high collision energy (CE), 4 and 20 V, respectively. The low-CE spectra were acquired to identify charge-reduced ECnoD products from each sample to be isolated and fragmented in the MS/MS experiments; the high-CE spectra were acquired simply to obtain a record of the parallel fragmentation AP-ECD spectra for each sample. Subsequently, for each sample, for one or more of its charge-reduced precursors, a series of AP-ECD− MS/MS spectra were then acquired as a function of CE, in steps of 5 V, from a low of 10 V until whenever near-complete CID of the precursor was observed. For all the MS/MS spectra presented here, the LM (lowmass) resolution of the quadrupole was set to its default value of 4.7 (instrument units), which enabled the transmission of all the isotopes for each ECnoD precursor. The TOF analyzer was in resolution mode, with its dual-stage reflectron set to provide the highest resolution at the expense of a slight decrease in sensitivity. Continuum (profile) data were acquired exclusively.
■
RESULTS AND DISCUSSION [Glu1]-fibrinopeptide B. To first demonstrate the performance of AP-ECD in its usual direct parallel fragmentation mode, Figure 2A is an AP-ECD−MS spectrum of glu-fib, acquired
Figure 2. AP-ECD spectra of glu-fib: (A) direct parallel fragmentation MS spectrum; (B) MS/MS spectrum of singly charged charge-reduced precursors.
with the instrument’s quadrupole serving as an RF-only ion guide. To increase sequence coverage and ECD-fragment signal intensity, the collision energy (CE) was set to a relatively high value, 20 V, so that ECnoD products from the source would be likely to dissociate in the collision cell and contribute their fragments to the final AP-ECD spectrum, much as is commonly done with ion traps using the “ETcaD” method.16 Figure 2A shows that when operated in this manner AP-ECD yields extensive sequence ladders of both c′ and z• ions for glu-fib, with every peptide bond being cleaved, providing 100% sequence coverage. The spectrum also includes residual glufib precursors, (M + 2H)2+, and a substantial amount of surviving ECnoD products, (M + 2H)+•, with the latter indicating that the CE could have been raised further to generate additional ECD fragments; however, with increased CE also comes increased CID, particularly for the residual multiply charged precursors, and the prominent series of y ions in Figure 2A indicates this was already a significant process at 20 V. Note that having both CID and ECD/ETD fragments in a spectrum may provide more complete sequence coverage and thus be beneficial;19,20 however, if desired, it is simple to remove the CID fragments from AP-ECD spectra, either by lowering the CE to prevent their formation in the first place or else through background subtraction of a separate spectrum collected with ECD turned off.4 In sum, Figure 2A confirms that AP-ECD−MS with direct parallel fragmentation is capable of yielding informative, high-sequence-coverage ECD spectra. 4442
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
To provide a first demonstration of AP-ECD operating in MS/MS mode, Figure 2B is a spectrum of glu-fib acquired with the quadrupole filter set to m/z 1571.4 to isolate the singly charged charge-reduced products of glu-fib from the mix of ions exiting the source. Compared to Figure 2A, the MS/MS spectrum is much cleaner, being dominated by an extensive sequence ladder of z-type fragments, plus a couple of large ctype fragments, all with high S/N and again providing 100% sequence coverage. From this, it is immediately apparent that informative MS/MS spectra are in fact attainable with AP-ECD, despite the initial cleavage of covalent bonds being performed in parallel. The CE in this case was 30 V, sufficient to release the internal ECD fragments of the majority of the singly charged ECnoD ions, apparently without causing substantial CID of other species. There is, however, a minor series of y ions in the spectrum, situated just to the right of the corresponding z-type ion series, but these are believed to be the products of a secondary ECD pathway,22 not CID fragments, because their intensities relative to those of their z-type neighbors were essentially constant over a wide range of collision energies, up to 50 V (results not shown). Also noteworthy is that the majority of the ECD fragments are present in both the oddand even-electron forms, indicating that H atom migration within the ECnoD species was prevalent prior to their dissociation in the collision cell; this is commonly observed in ECD/ETD experiments, particularly when supplemental activation is employed.16,18,23 The presence of odd-electron c• ions and even-electron z′ ions in the spectrum, in addition to the c′ and z• ions typically formed by direct ECD, is not ideal because it can complicate data analysis;9,24 however, it is also clearly not prohibitive, as a variety of effective data analysis procedures accounting for the phenomenon have been demonstrated (see, for example, refs 16, 17, 20, and 25). A major difference between the AP-ECD−MS and −MS/MS spectra in Figure 2 is that the MS/MS spectrum is much cleaner, in part because it only contains precursor ions from a single charge state. This is a direct result of there being no charge-reduction reactions following precursor selection, unlike in conventional ECD/ETD where the reactions generate a distribution of charge-reduced products, plus neutral loss species, from a single multiply charged precursor. This is significant because the extra peak clusters constitute interferences that can prevent the detection of ECD/ETD fragments of interest (for example, see ref 26). In addition, because only charge-reduced species are subjected to (relatively gentle) collisional activation, conventional CID of residual highly charged precursors is also avoided. In Figure 2A, for example, there are clusters of peaks from both doubly and singly charged charge-reduced species of glu-fib, with neutral loss products, plus the substantial CID background generated mostly from residual triply charged precursors (which are undetected here because they dissociate completely in the collision cell). Contrastingly, in Figure 2B, other than the ECD fragments of interest, there are only peaks from a single charge state of the precursor, with very few CID products. Hence, as noted elsewhere previously,17 the elimination of extraneous chargereduction byproducts may be an inherent advantage of selecting ions to dissociate af ter the EC/ET reactions. Substance P. Figure 3A is an AP-ECD−MS spectrum of substance P, again acquired with the quadrupole serving as an ion guide and with the CE set to 20 V. As seen for glu-fib, an extensive series of ECD fragments is present in the spectrum, in this case also including doubly charged fragments, with all the
Figure 3. AP-ECD spectra of substance P: (A) direct parallel fragmentation MS spectrum; (B) MS/MS spectrum of doubly charged precursors; (C) MS/MS spectrum of singly charged precursors. The series of satellite peaks marked “*” are believed to be due to secondary ECD fragments of doubly charged c4′−c10′, respectively, resulting from two electron capture events.
peptide backbone bonds (other than those N-terminal to proline) having been cleaved. There are also residual doubly charged precursors, (M + 2H)2+, as well as charge-reduced ECnoD species, (M + 2H)+• and (M + 3H)+••, and a background of CID products arising from unreacted doubly and triply charged precursors (the latter being completely dissociated at CE = 20 V). Another notable feature is the chemical background due to impurities in the sample stream, mostly below m/z 400, which can interfere with the detection of smaller ECD fragments unless it is removed by background subtraction (a trivial process when AP-ECD is used in conjunction with LC5). Altogether, Figure 3A is a typical APECD−MS spectrum illustrating the kind of data that is routinely obtained for peptides with the method (which has been included here primarily to facilitate comparisons with the MS/MS spectra). Figure 3B is an AP-ECD−MS/MS spectrum of substance P, acquired with the quadrupole set to m/z 674.8 to isolate (M + 2H)2+/(M + 3H)2+• species from the source. The CE was set to a relatively low value, 10 V, because at higher values CID of (M + 2H)2+ became significant without enhancing the overall ECD signals. Again, the MS/MS spectrum is much cleaner than the corresponding MS spectrum, being dominated by a mixture of singly and doubly charged ECD fragments, plus surviving precursors. As expected, the m/z selection step has eliminated the chemical background, and there are only precursor peaks from one charge state. In this case, though, the sequence coverage is incomplete, as only fragments from the middle of 4443
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
Figure 4. AP-ECD−MS/MS spectra of melittin (doubly charged charge-reduced precursors): (A) raw data; (B) deconvoluted zero-charge spectrum obtained after processing with MaxEnt 3.
containing c2′ and z9′ should survive to reach the collision cell while the corresponding doubly charged ECnoD species do not; we speculate that the difference may be due to differences in the charge states of the original precursors(M + 2H)2+ versus (M + 3H)3+, respectivelyas well as the fact that more highly charged precursor species will have greater internal Coulombic repulsion and thus more open structures and reduced noncovalent interactions for binding ECnoD products. Another significant difference between the MS/MS spectra of substance P lies in the presence of a series of satellite peaks in Figure 3C, marked with asterisks, at m/z values 15 Da below those of c4′ to c10′, which appear to be due to internal fragments formed by secondary fragmentation of primary doubly charged c′ fragments. This conclusion is supported by the results of a separate experiment in which the quadrupole resolution was increased so that only ions of a single m/z would be transmitted: the −15 Da satellite peaks were only present for m/z 1349.6, which allowed for the transmission of (M + 3H)+•• species (products of two electron capture events), whereas they were not present at all for m/z 1348.6, which for substance P’s ECnoD products can only be (M + 2H)+• (products of a single electron capture event). Setting aside further discussion of mechanisms for now, until such time as more in-depth fundamental studies have been performed, the AP-ECD−MS/ MS results for substance P demonstrate again that the method yields informative spectra; however, these results also demonstrate that the sequence coverage may be less than in direct AP-ECD−MS, most likely due to the requirement that the selected ECnoD species be stable enough to reach the collision cell intact, which may be problematic when electron capture occurs near one of the termini.
the molecule are observedthe z9 and c10 fragments present in the MS spectrum, Figure 3A, are missing. This indicates that, for substance P when an electron is captured near the termini, dissociation of the resulting charge-reduced product occurs prior to reaching the collision cell, either promptly in the ion source or else during transit through the vacuum interface and/ or RF ion guides. Plainly, the likelihood that a long-lived ECnoD species will form for a given precursor depends upon where on the molecule the electron is captured, and it seems reasonable to expect that when capture occurs at or near the termini the charge-reduced product will generally be less likely to stay intact, because in this event one of the fragments will be very small and thus have relatively few and/or weak noncovalent interactions with the remainder of the molecule, all else being the same. Also notable in Figure 3B is the series of z• fragments, which are not typically observed with ECD/ETD for substance P (only z9 is common); here, these fragments disappeared rapidly when the CE was raised, indicating that they are unstable species and thus may be difficult to detect under ordinary conditions. Figure 3C is an AP-ECD−MS/MS spectrum of substance P, acquired with the quadrupole set to m/z 1349.6 to isolate singly charged charge-reduced species from the source, primarily (M + 2H)+• and (M + 3H)+••. As with the earlier glu-fib MS/MS spectrum, the CE was set to 30 V, providing enough energy to dissociate the majority of the ECnoD species without causing substantial CID. The spectrum here is dominated by an extensive series of c-type ions, plus surviving charge-reduced precursors, much as in the first MS/MS spectrum of substance P, Figure 3B; in contrast with the latter, however, in Figure 3C fragments from near the N-terminus, c2′ and z9′, are also present. It is not clear to us why singly charged ECnoD species 4444
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
Melittin. To demonstrate the suitability of AP-ECD−MS/ MS for sequencing larger peptides, Figure 4A is a spectrum of melittin acquired with the quadrupole set to m/z 1425.3 Da to transmit doubly charged charge-reduced products. The CE in this case was set to 30 V, again sufficient to break apart the majority of ECnoD species without inducing substantial conventional CID. The raw MS/MS spectrum is complex, with an abundance of both singly and doubly charged c- and ztype ECD fragments. There are also numerous y ions in the spectrum, but their intensities relative to the c and z ions were largely unaffected by CE over a wide range, as noted above for glu-fib, suggesting that these too are minor ECD products. Also, there are only residual precursors from a single-charge state in the spectrum, which in this case is more likely to provide a definite advantage in sequence determination, because for larger peptide/proteins such as melittin it is normal for there to be several charge states of precursors generated from a single highly charged species, all of which can interfere with the detection of the desired ECD fragments. To simplify the raw MS/MS spectrum for melittin, it was processed using the MaxEnt 3 deconvolution function of the MassLynx software. Figure 4B is the resulting deconvoluted zero-charge spectrum, and it reveals that the sequence coverage in this case was excellent, with only one possible cleavage being missed, that being adjacent to the N-terminus (which is consistent with the earlier discussion for substance P regarding fragmentation near the termini). Altogether, because of the high sequence coverage obtained, and the apparent high overall quality of the data, this result indicates that AP-ECD−MS/MS may be well-suited for the analysis of larger peptides/proteins, and thus has the potential to be a valuable tool for middle- and top-down MS analyses. Modified Peptides. To demonstrate the suitability of APECD−MS/MS for analyzing peptides with labile modifications, typically difficult to localize with conventional CID MS/MS, Figure 5 presents spectra obtained for phosphorylated cholecystokinin (10−20), glycosylated MUC5AC 3, and glycosylated erythropoietin (EPO) (117−131). Figure 5A is the spectrum obtained from the singly charged charge-reduced ions of the phosphopeptide, using a CE of 30 V. The sequence coverage is near-complete, with only one possible cleavage (adjacent to the N-terminus) being missed, andimportantlythe phosphorylation is retained on the ECD fragments, enabling its localization directly from the MS/MS spectrum. The latter is significant because the use of supplemental collisional activation of ECnoD products might be expected to dislodge labile modifications, as in conventional CID; the fact that the phosphorylation is completely retained on the fragments here is of course attributable to the relatively low CE required to dissociate the ECnoD species held together by only noncovalent interactions (note that a nominal CE of 30 V is not particularly low for a conventional CID MS/MS experiment, but because here the CID precursors are chargereduced and singly charged the true laboratory-frame collision energy is half what it would be for a typical doubly charged CID precursor). Similarly, Figure 5B is the spectrum obtained for glycosylated MUC5AC 3, obtained from its singly charged charge-reduced precursors using a CE of 25 V. In this case only one possible cleavage is missed (near the C-terminus), and the glycosylation is retained on the ECD fragments, again readily enabling its localization. Finally, Figure 5C is the spectrum obtained for glycosylated EPO, obtained from its singly charged charge-reduced precursors using a CE of 30 V. The sequence
Figure 5. AP-ECD−MS/MS spectra of peptides with labile modifications: (A) phosphorylated cholecystokinin (10−20); (B) glycosylated MUC5AC 3; (C) glycosylated erythropoietin (EPO) (117−131). In each case the modification is retained on the ECD fragments following supplemental collisional activation of the singly charged charge-reduced precursor.
coverage is again extensive with only one possible cleavage being missed (near the C-terminus), and the glycosylation is retained on the ECD fragments. Hence, because it provides extensive sequence coverage without requiring highly energetic collisions, AP-ECD−MS/MS appears to be well-suited for analyzing peptides with labile modifications.
■
CONCLUSION The results presented here demonstrate that true tandem MS spectra may be readily obtained from a mixture of precursor ions using an AP-ECD source fitted to a Q-TOF mass spectrometer. In the AP-ECD−MS/MS method, ECnoD product ions generated in the ion source become the precursor ions to be selected by the quadrupole for supplemental CID, ultimately generating ECD fragment ions for mass analysis. The sequence coverage obtained with the method appears to be very high, >90% for the model peptides tested here (based upon the percentage of peptide bonds cleaved, excepting those N-terminal to proline). Note that the cleavages missed tend to be at or near the termini, indicating that the ECnoD species selected for supplemental collisional activation may be less likely to survive to reach the collision cell when initial electron capture occurs near one of the termini of the precursor, which does constitute a limitation of the technique. On the other hand, an attractive feature of the method is that ions from only one charge state of precursor are present in the final spectrumi.e., there is no distribution of charge-reduced precursors, with attendant neutral loss peaks, to interfere with the detection of the desired ECD fragments; we expect this to provide a significant advantage for the middle- and top-down MS analyses of larger peptides/proteins. Finally, the results for peptides with labile modifications demonstrate that these are retained on the ECD fragments at the relatively low CE values required for collisional dissociation of ECnoD species, enabling their direct localization. 4445
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
■
Article
(23) Nishikaze, T.; Takayama, M. J. Am. Soc. Mass Spectrom. 2010, 21, 1979−1988. (24) Savitski, M. M.; Kjeldsen, F.; Nielsen, M. L.; Zubarev, R. A. J. Am. Soc. Mass Spectrom. 2007, 18, 113−120. (25) Sun, R.-X.; Dong, M.-Q.; Song, C.-Q.; Chi, H.; Yang, B.; Xiu, L.Y.; Tao, L.; Jing, Z.-Y.; Liu, C.; Wang, L.-H.; Fu, Y.; He, S.-M. J. Proteome Res. 2010, 9, 6354−6367. (26) Fornelli, L.; Parra, J.; Hartmer, R.; Stoermer, C.; Lubeck, M.; Tsybin, Y. O. Anal. Bioanal. Chem. 2013, 405, 8505−8514.
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Major funding for this work was from an “InventionTools, Techniques and Devices” Catalyst Grant from the Institute of Genetics (IG) of the Canadian Institutes for Health Research (CIHR). Additional funding was from a Discovery Grant from the Natural Sciences and Engineering Research Council (NSERC) of Canada. Jason Rogalski of UBC is acknowledged for helpful discussions.
■
REFERENCES
(1) Zubarev, R. A.; Kelleher, N. L.; McLafferty, F. W. J. Am. Chem. Soc. 1998, 120, 3265−3266. Cooper, H. J.; Hakansson, K.; Marshall, A. G. Mass Spectrom. Rev. 2005, 24, 201−222. (2) Syka, J. E. P.; Coon, J. J.; Schroeder, M. J.; Shabanowitz, J.; Hunt, D. F. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 9528−9533. (3) Robb, D. B.; Rogalski, J. C.; Kast, J.; Blades, M. W. Rapid Commun. Mass Spectrom. 2010, 24, 3303−3308. (4) Robb, D. B.; Rogalski, J. C.; Kast, J.; Blades, M. W. J. Am. Soc. Mass Spectrom. 2011, 22, 1699−1706. (5) Robb, D. B.; Rogalski, J. C.; Kast, J.; Blades, M. W. Anal. Chem. 2012, 84, 4221−4226. (6) Robb, D.; Rogalski, J.; Carter, D.; Blades, M.; Kast, J. In Proceedings of the 59th ASMS Conference: Denver, CO, 2011. (7) Horn, D. M.; Ge, Y.; McLafferty, F. W. Anal. Chem. 2000, 72, 4778−4784. (8) Pitteri, S. J.; Chrisman, P. A.; McLuckey, S. A. Anal. Chem. 2005, 77, 5662−5669. (9) Ledvina, A. R.; McAlister, G. C.; Gardner, M. W.; Smith, S. I.; Madsen, J. A.; Schwartz, J. C.; Stafford, G. C., Jr.; Syka, J. E. P.; Brodbelt, J. S.; Coon, J. J. Angew. Chem., Int. Ed. 2009, 48, 8526−8528. (10) Ledvina, A. R.; Rose, C. M.; McAlister, G. C.; Syka, J. E. P.; Westphall, M. S.; Griep-Raming, J.; Schwartz, J. C.; Coon, J. J. J. Am. Soc. Mass Spectrom. 2013, 24, 1623−1633. (11) Rose, C. M.; Russell, J. D.; Ledvina, A. R.; McAlister, G. C.; Westphall, M. S.; Griep-Raming, J.; Schwartz, J. C.; Coon, J. J.; Syka, J. E. P. J. Am. Soc. Mass Spectrom. 2013, 24, 816−827. (12) Purvine, S.; Eppel, J.-T.; Yi, E. C.; Goodlett, D. R. Proteomics 2003, 3, 847−850. (13) Silva, J. C.; Gorenstein, M. V.; Li, G.-Z.; Vissers, J. P. C.; Geromanos, S. J. Mol. Cell. Proteomics 2006, 5, 144−156. (14) Robb, D. B.; Blades, M. W. J. Am. Soc. Mass Spectrom. 2006, 17, 130−138. (15) Pitteri, S. J.; Chrisman, P. A.; Hogan, J. M.; McLuckey, S. A. Anal. Chem. 2005, 77, 1831−1839. Han, H.; Xia, Y.; McLuckey, S. A. Rapid Commun. Mass Spectrom. 2007, 21, 1567−1573. Han, H.; Xia, Y.; Yang, M.; McLuckey, S. A. Anal. Chem. 2008, 80, 3492−3497. (16) Swaney, D. L.; McAlister, G. C.; Wirtala, M.; Schwartz, J. C.; Syka, J. E. P.; Coon, J. J. Anal. Chem. 2007, 79, 477−485. (17) Wu, S.-L.; Huhmer, A. F. R.; Hao, Z.; Karger, B. L. J. Proteome Res. 2007, 6, 4230−4244. (18) Xia, Y.; Han, H.; McLuckey, S. A. Anal. Chem. 2008, 80, 1111− 1117. Hamidane, H. B.; Chiappe, D.; Hartmer, R.; Vorobyev, A.; Moniatte, M.; Tsybin, Y. O. J. Am. Soc. Mass Spectrom. 2009, 20, 567− 575. (19) Campbell, J. L.; Hager, J. M.; Le Blanc, J. C. Y. J. Am. Soc. Mass Spectrom. 2009, 20, 1672−1683. (20) Liu, C.-W.; Lai, C.-C. J. Am. Soc. Mass Spectrom. 2011, 22, 57− 66. (21) Xia, Y.; Chrisman, P. A.; Pitteri, S. J.; Erickson; McLuckey, S. A. J. Am. Chem. Soc. 2006, 128, 11792−11798. (22) Simons, J. Chem. Phys. Lett. 2010, 484, 81−95. 4446
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Tandem Mass Spectrometry Using the Atmospheric Pressure Electron Capture Dissociation Ion Source Damon B. Robb,*,† Jeffery M. Brown,‡ Michael Morris,‡ and Michael W. Blades† †
University of British Columbia, Department of Chemistry, Vancouver, BC V6T 1Z1, Canada Waters Corporation, Stamford Avenue, Altrincham Road, Wilmslow, SK9 4AX, United Kingdom
‡
ABSTRACT: Atmospheric pressure electron capture dissociation (AP-ECD) is an emerging technique capable of being adopted to virtually any electrospray mass spectrometer, without modification of the main instrument. To date, however, because the electron capture reactions occur in the ion source, AP-ECD has been limited by its apparent inability to select precursors prior to fragmentation, i.e., to perform tandem mass spectrometry (MS/MS) experiments. In this paper we demonstrate a novel AP-ECD−MS/MS method using an AP-ECD source on a Xevo G2-S quadrupole time-offlight (Q-TOF) mass spectrometer from Waters Micromass. The method takes advantage of the tendency for electron capture reactions to generate charge-reduced “ECnoD” products, species that have captured an electron and have had a covalent bond cleaved yet do not immediately dissociate into separate products and so retain the mass of the precursor ion. In the method, ECnoD products from the AP-ECD source are isolated in the quadrupole mass filter and induced to dissociate through supplemental activation in the collision cell, and then the liberated ECD fragment ions are mass analyzed using the highresolution TOF. In this manner, true MS/MS spectra may be obtained with AP-ECD even though all of the precursors in the source are subjected to electron capture reactions in parallel. Here, using a late-model Q-TOF instrument otherwise incapable of performing electron-based fragmentation, we present AP-ECD−MS/MS results for a group of model peptides and show that informative, high-sequence-coverage spectra are readily attainable with the method.
E
lectron capture dissociation (ECD)1 and electron transfer dissociation (ETD)2 are important new tools for the structural characterization of peptides and proteins by mass spectrometry. In contrast with conventional collision-induced dissociation (CID), these techniques readily enable the localization of labile post-translation modifications (PTMs), and they provide more extensive and uniform sequence coverage. Further, compared to CID, ECD and ETD are outstanding for their effectiveness in cleaving disulfide bonds, as well as in fragmenting increasingly larger species in top-down analyses. Presently ETD is more widely used than ECD, principally because it can be implemented using common radio frequency (RF) ion traps, unlike ECD, which generally requires a Fourier transform ion cyclotron resonance instrument. ETD is now available as an option on numerous high-performance mass spectrometers; however, in addition to a quadrupole mass analyzer to isolate parent ions and reagent anions, and a highresolution mass analyzer such as a reflectron time-of-flight (TOF) or an Orbitrap to analyze the ETD fragments, these instruments also require a dedicated second ion source for generating anionic ETD reagents and some kind of RF ion trap for carrying out the reactions. Consequently, although they provide unprecedented fragmentation and mass analysis capabilities, these modern ETD-equipped instruments are also highly complex and expensive, and thus may be inaccessible to many researchers wishing to employ electron-based fragmentation. © 2014 American Chemical Society
Atmospheric pressure electron capture dissociation (APECD) is an emerging technique capable of being adopted to virtually any electrospray ionization mass spectrometer.3−5 Because AP-ECD requires only an ion source device, which may easily be retrofit to instruments not originally designed for ECD/ETD experiments, it has the potential to make electronbased fragmentation much more widely accessible; with an APECD source, instrument companies could provide ECD functionality on all of their instruments without impacting existing designs, while researchers without access to a suitable ECD/ETD instrument would have new and more affordable options. AP-ECD has been proven to be effective at generating high-quality ECD spectra of peptides, including phospho- and glycopeptides,5,6 and it has been shown to be fully compatible with liquid chromatography (LC), with on-column detection limits in the low fmol range even when coupled with an earlymodel, relatively low performance instrument.5 In comparison with in vacuo ECD/ETD techniques, APECD has the advantage of being very simple in terms of hardware and method requirements. The essential hardware is composed of a handful of basic machined parts, a photoionization detector (PID) lamp, a dc power supply for the Received: January 22, 2014 Accepted: April 2, 2014 Published: April 2, 2014 4439
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
Figure 1. (A) Illustration of an experimental AP-ECD ion source (3D drawing inset) with a Xevo G2-S Q-TOF from Waters Micromass; (B) crosssectional drawing of the ion source.
source block, e.g., with a low-cost cartridge heater. This is in stark contrast with ETD, for which heating the ions is much more challenging, with the current method of choice being to irradiate the reaction cell with IR photons from a CO2 laser,9,10 an approach having the serious drawback of requiring both a dedicated laser and extensive modification of the instrument.11 Moreover, since AP-ECD is inherently a parallel fragmentation technique,12 it appears to be well-suited to techniques such as “MSE”,13 where all precursor ions are fragmented (normally by
lamp, and an optional heater and temperature controller, while the method requires only adding a photoionizable dopant to the source gas flow and switching on the lamp. Accordingly, with AP-ECD there is no need for ion−ion storage and manipulation schemes that add complexity to the hardware and software/firmware of ETD systems. Further, AP-ECD also has a practical advantage in enabling “activated ion”7−10 experiments, because the temperature of the ions in the ECD reaction region can be elevated simply through heating the surrounding 4440
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
light; Cambridge, U.K.). In operation, multiply charged peptide/protein precursor ions are created within the spray chamber by the enclosed nanospray source. These ions are entrained and transported through the spray chamber by the flow of gas to the downstream source block whose central channel may be irradiated by the photoionization lamp. When the lamp is switched off, precursor ions pass through unaffected and the source operates as a normal nanospray source; when the lamp is switched on, photoelectrons are generated by photoionization of a dopant (typically acetone) added to the AUX gas. Photoelectrons are subsequently captured by the precursor ions in the downstream atmospheric pressure reaction/transport zone, leading to charge reduction and cleavage of covalent bonds for the precursors, either with or without prompt dissociation. Ions exiting the source are then delivered through the vacuum interface of the mass spectrometer. In MS/MS mode, charge-reduced ECnoD reaction products are mass-selected in the quadrupole mass filter and induced to dissociate by supplemental activation in the collision cell, and then the ECD fragments are massanalyzed using the high-resolution TOF. To shield the photoelectron generation region of the source block from the electric field of the nanospray source, a thin wire-mesh screen (88% transmission; tungsten wire, 50 × 50 threads/in., wire o.d. = 0.0012 in.; Unique Wire, Hillside, NJ, USA) is secured to the end of the spray chamber, immediately before the source block entrance. The mesh is held between two thin stainless steel washers, spot-welded together to form a reusable assembly that may be handled and cleaned without damage. The mesh assembly is attached to the spray chamber using a threaded stainless steel cap, with an opening matching the i.d. of the spray chamber, which enables disassembly of the apparatus for cleaning without tools. The mesh assembly is at the same electric potential as the spray chamber and source block, and it confines the electric field of the nanospray source to the spray chamber. Note that the mesh must be cleaned when it becomes contaminated with a film of sample residue, because it can then become positively charged and attract electrons away from the precursor ions, causing a partial or complete loss of ECD performance. Preventative maintenance entails periodic cleaning of the spray chamber and mesh assembly. This can be accomplished simply by brushing the central channel of the spray chamber under running water, followed by rinsing both the spray chamber and the mesh with deionized (DI) water and then methanol, and then air-drying and/or blotting with a laboratory wipe. The required frequency of cleaning depends upon how heavily the source is used: weekly may be acceptable in the case of LC/MS experiments, where sample is only injected at intervals, whereas daily cleaning may be required for infusion experiments where sample is delivered continuously for extended periods; that said, even daily cleaning is of little consequence as the entire routine may be completed in 5 min. To heat the source and thereby both improve peptide ionization efficiency and “activate”7−10 the peptide ions during ECD, a single 1/4 in. o.d., 100 W cartridge heater (model CSH101100/120; Omega, Laval, QC, Canada) was installed in the source block. Temperature control was enabled via an embedded insulated thermocouple (model HSTC-TT-J-24S72-SMPW-CC; Omega) adjacent to the heater and a benchtop temperature controller (model CSC32-J; Omega). To prevent radical electron capture products such as z• ions from forming adducts with oxygen,4,21 the source was
CID) and analyzed simultaneously and postacquisition data processing is used to match product ions with precursors. (Note that for MSE to be viable by ECD or ETD, the relative abundance of analyte ions and reagent electrons or anions must be balanced; with AP-ECD, photoionization is capable of generating a large abundance of electronsμA’s are attainable at atmospheric pressure using a PID lamp14which enables entire precursor populations to be reacted with a duty cycle of up to 100%.) Similarly, because the electrons are generated within a confined spatial volume, through which the precursor ions flow, there are no issues related to the simultaneous trapping and storage of reagent, precursor, and fragment ions of (potentially) disparate m/z and abundance. Nevertheless, despite these advantages, because the electron capture reactions occur in the ion source, AP-ECD has thus far been limited by its apparent major shortcoming: an inability to isolate precursors before fragmentation, i.e., to perform true tandem mass spectrometry (MS/MS) experiments. In this paper we demonstrate a novel AP-ECD−MS/MS method, using an AP-ECD source on a Xevo G2-S quadrupole time-of-flight (Q-TOF) mass spectrometer from Waters Micromass. At the same time, we also describe how a latemodel Q-TOF not originally designed for ECD/ETD experiments may be fitted with an AP-ECD source to enable electronbased fragmentation, without requiring modification of the main instrument. The AP-ECD−MS/MS method takes advantage of the universal tendency for electron capture (and transfer) reactions to generate an abundance of charge-reduced ECnoD products, species which have captured an electron and have had a covalent bond cleaved, yet do not immediately dissociate into separate products and so retain the mass of the precursor. The internal fragments of ECnoD (and ETnoD) species are thought to be held together by noncovalent interactions,7 and it is well-established that their components may be separated by gentle supplemental collisional activation prior to analysis, both to improve ECD/ETD signal intensity and to increase sequence coverage.15−20 In the present method, ECnoD species emanating from the AP-ECD ion source are isolated in the mass-resolving quadrupole and induced to dissociate through supplemental activation in the collision cell, and then the liberated ECD fragment ions are mass analyzed using the high-resolution TOF. In this manner true MS/MS spectra may be obtained with AP-ECD even though all the precursors in the source are subjected to electron-capture reactions in parallel. Here, using a Q-TOF instrument otherwise incapable of performing electron-based fragmentation, we present AP-ECD−MS/MS results for a group of model peptides and show that informative, high-sequencecoverage tandem mass spectra are readily attainable with the method.
■
EXPERIMENTAL SECTION AP-ECD Ion Source. Figure 1A is an illustration of the experimental AP-ECD ion source coupled with a Xevo G2-S QTOF mass spectrometer from Waters Micromass (Floats Road, Manchester, U.K.); Figure 1B is a cross-sectional drawing of the source, which is closely modeled after a design described previously.5 In brief, the AP-ECD source is composed of three interconnected subassemblies: a polyimide sprayer plug with an integrated nanospray source; a stainless-steel spray chamber with a central channel and an inlet for an auxiliary (AUX) gas flow; and a stainless steel source block with a dc krypton photoionization lamp (model PKS 100/106; Heraeus Noble4441
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
The scan time per spectrum was 1 s. Each spectrum presented here is the sum of 30 spectra. None of the spectra has been smoothed or centroided, except for melittin’s, which was smoothed using a five point Savitzky-Golay filter and then centroided, followed by charge-state deconvolution with the software’s MaxEnt 3 function (MaxEnt 3 constructs equivalent zero charge spectra from raw mass spectra with ions of multiple charge states).
positioned so that its outlet was on-axis with and very close to (separation ≤ 1 mm) the tip of the sampling cone, the inlet of the instrument’s vacuum interface. The gas exiting the source then swept away the ambient air between it and the cone so that the exposure of radical species to oxygen was minimized. Chemicals. [Glu1]-fibrinopeptide B (“glu-fib”) and substance P were from Sigma-Aldrich (Oakville, Ontario, Canada). Melittin, honey bee (deamidated), was from AnaSpec (San Jose, CA, USA). Phosphorylated cholecystokinin (10−20), glycosylated MUC5AC 3, and glycosylated erythropoietin (EPO) (117−131) were from Protea Biosciences (Morgantown, WV, USA). Methanol (high-performance liquid chromatography (HPLC)-grade) and formic acid (≥95% in water) were from Fisher Scientific (Ottawa, Ontario, Canada). Deionized water was from an in-house generator. The dopant used to promote photoelectron generation was HPLC-grade acetone from Fisher Scientific. Liquid nitrogen boil-off was used for the source gases. For all the samples, the concentration was 1 μM, except for melittin, which was 5 μM. The sample makeup solvent was 50/ 50 methanol/water with 0.1% formic acid. Instrumentation and Methods. The mass spectrometer was an unmodified Xevo G2-S Q-TOF from Waters Micromass, with MassLynx V4.1 software. The sample solutions were infused via an external syringe pump at a flow rate of 0.5 μL min−1. The acetone dopant was delivered continuously at 1.0 μL min−1 via the mass spectrometer’s internal fluidics system. The nanospray voltage was set to 2.2 kV, provided by the capillary high voltage (HV) power supply of the mass spectrometer. The spray chamber and source block offset voltage was set to 200 V by a custom AP-ECD power supply (Electronic Engineering Services, Department of Chemistry, UBC), which also provided the dc voltage (−1.5 kV) for igniting the photoionization lamp (note that the lamp and its power supply were floated at the offset voltage of the spray chamber and source block). The sampling cone voltage of the vacuum interface was set to a low value, 10 V, to avoid CID in the interface region and thus to enhance the transmission of intact ECnoD products. The nebulizer (NEB) gas was from the instrument’s nanoflow gas supply, set to 2.0 bar. The flow rate of the AUX gas was set via an external rotameter to 10 L min−1. The temperature of the AP-ECD source block was 100 °C, while the source temperature of the instrument’s vacuum interface was 80 °C. The experiments involved first acquiring direct parallel fragmentation AP-ECD spectra at both low and high collision energy (CE), 4 and 20 V, respectively. The low-CE spectra were acquired to identify charge-reduced ECnoD products from each sample to be isolated and fragmented in the MS/MS experiments; the high-CE spectra were acquired simply to obtain a record of the parallel fragmentation AP-ECD spectra for each sample. Subsequently, for each sample, for one or more of its charge-reduced precursors, a series of AP-ECD− MS/MS spectra were then acquired as a function of CE, in steps of 5 V, from a low of 10 V until whenever near-complete CID of the precursor was observed. For all the MS/MS spectra presented here, the LM (lowmass) resolution of the quadrupole was set to its default value of 4.7 (instrument units), which enabled the transmission of all the isotopes for each ECnoD precursor. The TOF analyzer was in resolution mode, with its dual-stage reflectron set to provide the highest resolution at the expense of a slight decrease in sensitivity. Continuum (profile) data were acquired exclusively.
■
RESULTS AND DISCUSSION [Glu1]-fibrinopeptide B. To first demonstrate the performance of AP-ECD in its usual direct parallel fragmentation mode, Figure 2A is an AP-ECD−MS spectrum of glu-fib, acquired
Figure 2. AP-ECD spectra of glu-fib: (A) direct parallel fragmentation MS spectrum; (B) MS/MS spectrum of singly charged charge-reduced precursors.
with the instrument’s quadrupole serving as an RF-only ion guide. To increase sequence coverage and ECD-fragment signal intensity, the collision energy (CE) was set to a relatively high value, 20 V, so that ECnoD products from the source would be likely to dissociate in the collision cell and contribute their fragments to the final AP-ECD spectrum, much as is commonly done with ion traps using the “ETcaD” method.16 Figure 2A shows that when operated in this manner AP-ECD yields extensive sequence ladders of both c′ and z• ions for glu-fib, with every peptide bond being cleaved, providing 100% sequence coverage. The spectrum also includes residual glufib precursors, (M + 2H)2+, and a substantial amount of surviving ECnoD products, (M + 2H)+•, with the latter indicating that the CE could have been raised further to generate additional ECD fragments; however, with increased CE also comes increased CID, particularly for the residual multiply charged precursors, and the prominent series of y ions in Figure 2A indicates this was already a significant process at 20 V. Note that having both CID and ECD/ETD fragments in a spectrum may provide more complete sequence coverage and thus be beneficial;19,20 however, if desired, it is simple to remove the CID fragments from AP-ECD spectra, either by lowering the CE to prevent their formation in the first place or else through background subtraction of a separate spectrum collected with ECD turned off.4 In sum, Figure 2A confirms that AP-ECD−MS with direct parallel fragmentation is capable of yielding informative, high-sequence-coverage ECD spectra. 4442
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
To provide a first demonstration of AP-ECD operating in MS/MS mode, Figure 2B is a spectrum of glu-fib acquired with the quadrupole filter set to m/z 1571.4 to isolate the singly charged charge-reduced products of glu-fib from the mix of ions exiting the source. Compared to Figure 2A, the MS/MS spectrum is much cleaner, being dominated by an extensive sequence ladder of z-type fragments, plus a couple of large ctype fragments, all with high S/N and again providing 100% sequence coverage. From this, it is immediately apparent that informative MS/MS spectra are in fact attainable with AP-ECD, despite the initial cleavage of covalent bonds being performed in parallel. The CE in this case was 30 V, sufficient to release the internal ECD fragments of the majority of the singly charged ECnoD ions, apparently without causing substantial CID of other species. There is, however, a minor series of y ions in the spectrum, situated just to the right of the corresponding z-type ion series, but these are believed to be the products of a secondary ECD pathway,22 not CID fragments, because their intensities relative to those of their z-type neighbors were essentially constant over a wide range of collision energies, up to 50 V (results not shown). Also noteworthy is that the majority of the ECD fragments are present in both the oddand even-electron forms, indicating that H atom migration within the ECnoD species was prevalent prior to their dissociation in the collision cell; this is commonly observed in ECD/ETD experiments, particularly when supplemental activation is employed.16,18,23 The presence of odd-electron c• ions and even-electron z′ ions in the spectrum, in addition to the c′ and z• ions typically formed by direct ECD, is not ideal because it can complicate data analysis;9,24 however, it is also clearly not prohibitive, as a variety of effective data analysis procedures accounting for the phenomenon have been demonstrated (see, for example, refs 16, 17, 20, and 25). A major difference between the AP-ECD−MS and −MS/MS spectra in Figure 2 is that the MS/MS spectrum is much cleaner, in part because it only contains precursor ions from a single charge state. This is a direct result of there being no charge-reduction reactions following precursor selection, unlike in conventional ECD/ETD where the reactions generate a distribution of charge-reduced products, plus neutral loss species, from a single multiply charged precursor. This is significant because the extra peak clusters constitute interferences that can prevent the detection of ECD/ETD fragments of interest (for example, see ref 26). In addition, because only charge-reduced species are subjected to (relatively gentle) collisional activation, conventional CID of residual highly charged precursors is also avoided. In Figure 2A, for example, there are clusters of peaks from both doubly and singly charged charge-reduced species of glu-fib, with neutral loss products, plus the substantial CID background generated mostly from residual triply charged precursors (which are undetected here because they dissociate completely in the collision cell). Contrastingly, in Figure 2B, other than the ECD fragments of interest, there are only peaks from a single charge state of the precursor, with very few CID products. Hence, as noted elsewhere previously,17 the elimination of extraneous chargereduction byproducts may be an inherent advantage of selecting ions to dissociate af ter the EC/ET reactions. Substance P. Figure 3A is an AP-ECD−MS spectrum of substance P, again acquired with the quadrupole serving as an ion guide and with the CE set to 20 V. As seen for glu-fib, an extensive series of ECD fragments is present in the spectrum, in this case also including doubly charged fragments, with all the
Figure 3. AP-ECD spectra of substance P: (A) direct parallel fragmentation MS spectrum; (B) MS/MS spectrum of doubly charged precursors; (C) MS/MS spectrum of singly charged precursors. The series of satellite peaks marked “*” are believed to be due to secondary ECD fragments of doubly charged c4′−c10′, respectively, resulting from two electron capture events.
peptide backbone bonds (other than those N-terminal to proline) having been cleaved. There are also residual doubly charged precursors, (M + 2H)2+, as well as charge-reduced ECnoD species, (M + 2H)+• and (M + 3H)+••, and a background of CID products arising from unreacted doubly and triply charged precursors (the latter being completely dissociated at CE = 20 V). Another notable feature is the chemical background due to impurities in the sample stream, mostly below m/z 400, which can interfere with the detection of smaller ECD fragments unless it is removed by background subtraction (a trivial process when AP-ECD is used in conjunction with LC5). Altogether, Figure 3A is a typical APECD−MS spectrum illustrating the kind of data that is routinely obtained for peptides with the method (which has been included here primarily to facilitate comparisons with the MS/MS spectra). Figure 3B is an AP-ECD−MS/MS spectrum of substance P, acquired with the quadrupole set to m/z 674.8 to isolate (M + 2H)2+/(M + 3H)2+• species from the source. The CE was set to a relatively low value, 10 V, because at higher values CID of (M + 2H)2+ became significant without enhancing the overall ECD signals. Again, the MS/MS spectrum is much cleaner than the corresponding MS spectrum, being dominated by a mixture of singly and doubly charged ECD fragments, plus surviving precursors. As expected, the m/z selection step has eliminated the chemical background, and there are only precursor peaks from one charge state. In this case, though, the sequence coverage is incomplete, as only fragments from the middle of 4443
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
Figure 4. AP-ECD−MS/MS spectra of melittin (doubly charged charge-reduced precursors): (A) raw data; (B) deconvoluted zero-charge spectrum obtained after processing with MaxEnt 3.
containing c2′ and z9′ should survive to reach the collision cell while the corresponding doubly charged ECnoD species do not; we speculate that the difference may be due to differences in the charge states of the original precursors(M + 2H)2+ versus (M + 3H)3+, respectivelyas well as the fact that more highly charged precursor species will have greater internal Coulombic repulsion and thus more open structures and reduced noncovalent interactions for binding ECnoD products. Another significant difference between the MS/MS spectra of substance P lies in the presence of a series of satellite peaks in Figure 3C, marked with asterisks, at m/z values 15 Da below those of c4′ to c10′, which appear to be due to internal fragments formed by secondary fragmentation of primary doubly charged c′ fragments. This conclusion is supported by the results of a separate experiment in which the quadrupole resolution was increased so that only ions of a single m/z would be transmitted: the −15 Da satellite peaks were only present for m/z 1349.6, which allowed for the transmission of (M + 3H)+•• species (products of two electron capture events), whereas they were not present at all for m/z 1348.6, which for substance P’s ECnoD products can only be (M + 2H)+• (products of a single electron capture event). Setting aside further discussion of mechanisms for now, until such time as more in-depth fundamental studies have been performed, the AP-ECD−MS/ MS results for substance P demonstrate again that the method yields informative spectra; however, these results also demonstrate that the sequence coverage may be less than in direct AP-ECD−MS, most likely due to the requirement that the selected ECnoD species be stable enough to reach the collision cell intact, which may be problematic when electron capture occurs near one of the termini.
the molecule are observedthe z9 and c10 fragments present in the MS spectrum, Figure 3A, are missing. This indicates that, for substance P when an electron is captured near the termini, dissociation of the resulting charge-reduced product occurs prior to reaching the collision cell, either promptly in the ion source or else during transit through the vacuum interface and/ or RF ion guides. Plainly, the likelihood that a long-lived ECnoD species will form for a given precursor depends upon where on the molecule the electron is captured, and it seems reasonable to expect that when capture occurs at or near the termini the charge-reduced product will generally be less likely to stay intact, because in this event one of the fragments will be very small and thus have relatively few and/or weak noncovalent interactions with the remainder of the molecule, all else being the same. Also notable in Figure 3B is the series of z• fragments, which are not typically observed with ECD/ETD for substance P (only z9 is common); here, these fragments disappeared rapidly when the CE was raised, indicating that they are unstable species and thus may be difficult to detect under ordinary conditions. Figure 3C is an AP-ECD−MS/MS spectrum of substance P, acquired with the quadrupole set to m/z 1349.6 to isolate singly charged charge-reduced species from the source, primarily (M + 2H)+• and (M + 3H)+••. As with the earlier glu-fib MS/MS spectrum, the CE was set to 30 V, providing enough energy to dissociate the majority of the ECnoD species without causing substantial CID. The spectrum here is dominated by an extensive series of c-type ions, plus surviving charge-reduced precursors, much as in the first MS/MS spectrum of substance P, Figure 3B; in contrast with the latter, however, in Figure 3C fragments from near the N-terminus, c2′ and z9′, are also present. It is not clear to us why singly charged ECnoD species 4444
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
Article
Melittin. To demonstrate the suitability of AP-ECD−MS/ MS for sequencing larger peptides, Figure 4A is a spectrum of melittin acquired with the quadrupole set to m/z 1425.3 Da to transmit doubly charged charge-reduced products. The CE in this case was set to 30 V, again sufficient to break apart the majority of ECnoD species without inducing substantial conventional CID. The raw MS/MS spectrum is complex, with an abundance of both singly and doubly charged c- and ztype ECD fragments. There are also numerous y ions in the spectrum, but their intensities relative to the c and z ions were largely unaffected by CE over a wide range, as noted above for glu-fib, suggesting that these too are minor ECD products. Also, there are only residual precursors from a single-charge state in the spectrum, which in this case is more likely to provide a definite advantage in sequence determination, because for larger peptide/proteins such as melittin it is normal for there to be several charge states of precursors generated from a single highly charged species, all of which can interfere with the detection of the desired ECD fragments. To simplify the raw MS/MS spectrum for melittin, it was processed using the MaxEnt 3 deconvolution function of the MassLynx software. Figure 4B is the resulting deconvoluted zero-charge spectrum, and it reveals that the sequence coverage in this case was excellent, with only one possible cleavage being missed, that being adjacent to the N-terminus (which is consistent with the earlier discussion for substance P regarding fragmentation near the termini). Altogether, because of the high sequence coverage obtained, and the apparent high overall quality of the data, this result indicates that AP-ECD−MS/MS may be well-suited for the analysis of larger peptides/proteins, and thus has the potential to be a valuable tool for middle- and top-down MS analyses. Modified Peptides. To demonstrate the suitability of APECD−MS/MS for analyzing peptides with labile modifications, typically difficult to localize with conventional CID MS/MS, Figure 5 presents spectra obtained for phosphorylated cholecystokinin (10−20), glycosylated MUC5AC 3, and glycosylated erythropoietin (EPO) (117−131). Figure 5A is the spectrum obtained from the singly charged charge-reduced ions of the phosphopeptide, using a CE of 30 V. The sequence coverage is near-complete, with only one possible cleavage (adjacent to the N-terminus) being missed, andimportantlythe phosphorylation is retained on the ECD fragments, enabling its localization directly from the MS/MS spectrum. The latter is significant because the use of supplemental collisional activation of ECnoD products might be expected to dislodge labile modifications, as in conventional CID; the fact that the phosphorylation is completely retained on the fragments here is of course attributable to the relatively low CE required to dissociate the ECnoD species held together by only noncovalent interactions (note that a nominal CE of 30 V is not particularly low for a conventional CID MS/MS experiment, but because here the CID precursors are chargereduced and singly charged the true laboratory-frame collision energy is half what it would be for a typical doubly charged CID precursor). Similarly, Figure 5B is the spectrum obtained for glycosylated MUC5AC 3, obtained from its singly charged charge-reduced precursors using a CE of 25 V. In this case only one possible cleavage is missed (near the C-terminus), and the glycosylation is retained on the ECD fragments, again readily enabling its localization. Finally, Figure 5C is the spectrum obtained for glycosylated EPO, obtained from its singly charged charge-reduced precursors using a CE of 30 V. The sequence
Figure 5. AP-ECD−MS/MS spectra of peptides with labile modifications: (A) phosphorylated cholecystokinin (10−20); (B) glycosylated MUC5AC 3; (C) glycosylated erythropoietin (EPO) (117−131). In each case the modification is retained on the ECD fragments following supplemental collisional activation of the singly charged charge-reduced precursor.
coverage is again extensive with only one possible cleavage being missed (near the C-terminus), and the glycosylation is retained on the ECD fragments. Hence, because it provides extensive sequence coverage without requiring highly energetic collisions, AP-ECD−MS/MS appears to be well-suited for analyzing peptides with labile modifications.
■
CONCLUSION The results presented here demonstrate that true tandem MS spectra may be readily obtained from a mixture of precursor ions using an AP-ECD source fitted to a Q-TOF mass spectrometer. In the AP-ECD−MS/MS method, ECnoD product ions generated in the ion source become the precursor ions to be selected by the quadrupole for supplemental CID, ultimately generating ECD fragment ions for mass analysis. The sequence coverage obtained with the method appears to be very high, >90% for the model peptides tested here (based upon the percentage of peptide bonds cleaved, excepting those N-terminal to proline). Note that the cleavages missed tend to be at or near the termini, indicating that the ECnoD species selected for supplemental collisional activation may be less likely to survive to reach the collision cell when initial electron capture occurs near one of the termini of the precursor, which does constitute a limitation of the technique. On the other hand, an attractive feature of the method is that ions from only one charge state of precursor are present in the final spectrumi.e., there is no distribution of charge-reduced precursors, with attendant neutral loss peaks, to interfere with the detection of the desired ECD fragments; we expect this to provide a significant advantage for the middle- and top-down MS analyses of larger peptides/proteins. Finally, the results for peptides with labile modifications demonstrate that these are retained on the ECD fragments at the relatively low CE values required for collisional dissociation of ECnoD species, enabling their direct localization. 4445
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
Analytical Chemistry
■
Article
(23) Nishikaze, T.; Takayama, M. J. Am. Soc. Mass Spectrom. 2010, 21, 1979−1988. (24) Savitski, M. M.; Kjeldsen, F.; Nielsen, M. L.; Zubarev, R. A. J. Am. Soc. Mass Spectrom. 2007, 18, 113−120. (25) Sun, R.-X.; Dong, M.-Q.; Song, C.-Q.; Chi, H.; Yang, B.; Xiu, L.Y.; Tao, L.; Jing, Z.-Y.; Liu, C.; Wang, L.-H.; Fu, Y.; He, S.-M. J. Proteome Res. 2010, 9, 6354−6367. (26) Fornelli, L.; Parra, J.; Hartmer, R.; Stoermer, C.; Lubeck, M.; Tsybin, Y. O. Anal. Bioanal. Chem. 2013, 405, 8505−8514.
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Major funding for this work was from an “InventionTools, Techniques and Devices” Catalyst Grant from the Institute of Genetics (IG) of the Canadian Institutes for Health Research (CIHR). Additional funding was from a Discovery Grant from the Natural Sciences and Engineering Research Council (NSERC) of Canada. Jason Rogalski of UBC is acknowledged for helpful discussions.
■
REFERENCES
(1) Zubarev, R. A.; Kelleher, N. L.; McLafferty, F. W. J. Am. Chem. Soc. 1998, 120, 3265−3266. Cooper, H. J.; Hakansson, K.; Marshall, A. G. Mass Spectrom. Rev. 2005, 24, 201−222. (2) Syka, J. E. P.; Coon, J. J.; Schroeder, M. J.; Shabanowitz, J.; Hunt, D. F. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 9528−9533. (3) Robb, D. B.; Rogalski, J. C.; Kast, J.; Blades, M. W. Rapid Commun. Mass Spectrom. 2010, 24, 3303−3308. (4) Robb, D. B.; Rogalski, J. C.; Kast, J.; Blades, M. W. J. Am. Soc. Mass Spectrom. 2011, 22, 1699−1706. (5) Robb, D. B.; Rogalski, J. C.; Kast, J.; Blades, M. W. Anal. Chem. 2012, 84, 4221−4226. (6) Robb, D.; Rogalski, J.; Carter, D.; Blades, M.; Kast, J. In Proceedings of the 59th ASMS Conference: Denver, CO, 2011. (7) Horn, D. M.; Ge, Y.; McLafferty, F. W. Anal. Chem. 2000, 72, 4778−4784. (8) Pitteri, S. J.; Chrisman, P. A.; McLuckey, S. A. Anal. Chem. 2005, 77, 5662−5669. (9) Ledvina, A. R.; McAlister, G. C.; Gardner, M. W.; Smith, S. I.; Madsen, J. A.; Schwartz, J. C.; Stafford, G. C., Jr.; Syka, J. E. P.; Brodbelt, J. S.; Coon, J. J. Angew. Chem., Int. Ed. 2009, 48, 8526−8528. (10) Ledvina, A. R.; Rose, C. M.; McAlister, G. C.; Syka, J. E. P.; Westphall, M. S.; Griep-Raming, J.; Schwartz, J. C.; Coon, J. J. J. Am. Soc. Mass Spectrom. 2013, 24, 1623−1633. (11) Rose, C. M.; Russell, J. D.; Ledvina, A. R.; McAlister, G. C.; Westphall, M. S.; Griep-Raming, J.; Schwartz, J. C.; Coon, J. J.; Syka, J. E. P. J. Am. Soc. Mass Spectrom. 2013, 24, 816−827. (12) Purvine, S.; Eppel, J.-T.; Yi, E. C.; Goodlett, D. R. Proteomics 2003, 3, 847−850. (13) Silva, J. C.; Gorenstein, M. V.; Li, G.-Z.; Vissers, J. P. C.; Geromanos, S. J. Mol. Cell. Proteomics 2006, 5, 144−156. (14) Robb, D. B.; Blades, M. W. J. Am. Soc. Mass Spectrom. 2006, 17, 130−138. (15) Pitteri, S. J.; Chrisman, P. A.; Hogan, J. M.; McLuckey, S. A. Anal. Chem. 2005, 77, 1831−1839. Han, H.; Xia, Y.; McLuckey, S. A. Rapid Commun. Mass Spectrom. 2007, 21, 1567−1573. Han, H.; Xia, Y.; Yang, M.; McLuckey, S. A. Anal. Chem. 2008, 80, 3492−3497. (16) Swaney, D. L.; McAlister, G. C.; Wirtala, M.; Schwartz, J. C.; Syka, J. E. P.; Coon, J. J. Anal. Chem. 2007, 79, 477−485. (17) Wu, S.-L.; Huhmer, A. F. R.; Hao, Z.; Karger, B. L. J. Proteome Res. 2007, 6, 4230−4244. (18) Xia, Y.; Han, H.; McLuckey, S. A. Anal. Chem. 2008, 80, 1111− 1117. Hamidane, H. B.; Chiappe, D.; Hartmer, R.; Vorobyev, A.; Moniatte, M.; Tsybin, Y. O. J. Am. Soc. Mass Spectrom. 2009, 20, 567− 575. (19) Campbell, J. L.; Hager, J. M.; Le Blanc, J. C. Y. J. Am. Soc. Mass Spectrom. 2009, 20, 1672−1683. (20) Liu, C.-W.; Lai, C.-C. J. Am. Soc. Mass Spectrom. 2011, 22, 57− 66. (21) Xia, Y.; Chrisman, P. A.; Pitteri, S. J.; Erickson; McLuckey, S. A. J. Am. Chem. Soc. 2006, 128, 11792−11798. (22) Simons, J. Chem. Phys. Lett. 2010, 484, 81−95. 4446
dx.doi.org/10.1021/ac5002959 | Anal. Chem. 2014, 86, 4439−4446
