Article pubs.acs.org/ac
Top-Down MALDI-In-Source Decay-FTICR Mass Spectrometry of Isotopically Resolved Proteins Simone Nicolardi,* Linda Switzar, André M. Deelder, Magnus Palmblad, and Yuri E.M. van der Burgt Center for Proteomics and Metabolomics, Leiden University Medical Center (LUMC), PO Box 9600, 2300 RC, Leiden, The Netherlands S Supporting Information *
ABSTRACT: An accurate mass measurement of a known protein provides information on potential amino acid deletions and post-translational modifications. Although this field is dominated by strategies based on electrospray ionization, mass spectrometry (MS) methods using matrix-assisted laser desorption/ionization (MALDI) have the advantage of yielding predominantly singly charged precursor ions, thus avoiding peak overlap from different charge states of multiple species. Such MALDI-MS methods require mass measurement at ultrahigh resolution, which is provided by Fourier transform ion cyclotron resonance (FTICR) mass analyzers. Recently, using a MALDI-FTICR-MS platform equipped with a 15 T magnet, we reported on the mass analysis of intact human serum peptides and small proteins with isotopic resolution up to ∼15 kDa and identified new proteoforms from an accurate measurement of mass distances. In the current study, we have used this FTICR system after an upgrade with a novel dynamically harmonized ICR cell, i.e., ParaCell, for mapping isotopically resolved intact proteins up to about 17 kDa and performed topdown MALDI in-source decay (ISD) analysis. Standard proteins myoglobin (m/z-value 16 950) and ribonuclease B (m/z-value 14 900) were measured with resolving powers of 62 000 and 61 000, respectively. Furthermore, it will be shown that (singly charged) MALDI-ISD fragment ions can be measured at isotopic resolution up to m/z-value 12 000 (e.g., resolving power 39 000 at m/z-value 12 000) providing more reliable identifications. Moreover, examples are presented of pseudo-MS3 experiments on ISD fragment ions from RNase B by collisional-induced dissociation (CID).
I
Matrix-assisted laser desorption/ionization (MALDI) insource decay (ISD) MS is an attractive alternative compared to ESI-MS-based methods for sequencing peptides and proteins.15−18 Ideally, the presence of one single proteoform within one MALDI spot is to be preferred to avoid the simultaneous detection of fragment ions from different species. The purification and/or separation (e.g., by LC) of different proteins and/or proteoforms prior to MALDI-ISD MS analysis allows a more confident identification.19 It has been shown that ions produced by MALDI in combination with specific matrices increase their internal energy to the point that fragmentation events are favored prior to extraction from the ionization source.20 These events mainly yield c-, y-, and z-type fragment ions and to a lesser extent w- and a-type ions.16,21−23 The presence of complementary series of fragment ions in a single spectrum results in a more reliable identification of the amino acid sequence and improves the sequence coverage. Moreover, hydrogen-donor MALDI matrices (e.g., 1,5-diaminonaphthalene (1,5-DAN)) catalyze the reduction of disulfide bonds during the
ntact protein analysis by top-down mass spectrometry (MS) allows both the determination of amino acid sequence and the identification of the site of post-translational modifications (PTMs) without the need for proteolytic digestion.1 A comprehensive picture of proteoforms contributes to a more detailed description of the proteome and its interactions, similar to the additional layer of information on the epigenetic code.2 Nowadays, large proteins and even protein complexes can be routinely mass-analyzed by (native) electrospray ionization (ESI)-MS approaches3 and subsequently sequenced using various tandem-MS fragmentation techniques such as collisional-induced dissociation (CID), electron-transfer dissociation (ETD), electron-capture dissociation (ECD), higher energy collisional dissociation (HCD), infrared multiphoton dissociation (IRMPD), or ultraviolet photodissociation (UVPD). Often, these fragmentation strategies provide complementary information on the amino acid sequence and PTMs.4−11 In order to obtain the sensitivity and resolution needed to tackle the complexity of the generated MS- and MS/MS-spectra, top-down analyses of intact proteins require (ultra)high resolution MSinstrumentation with resolving power of approximately 100 000, e.g., Orbitrap, Fourier transform ion cyclotron resonance (FTICR), and time-of-flight (TOF).12−14 © 2015 American Chemical Society
Received: December 18, 2014 Accepted: February 26, 2015 Published: February 26, 2015 3429
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
and mixed with 10 μL of an α-cyano matrix (0.3 g/L in ethanol/ acetone, 2:1). Then, 1 μL of the eluate/matrix mixture was spotted on the target plate. MALDI- and ESI-FTICR Mass Spectrometry. Both MALDI- and ESI-FTICR-MS experiments were performed on a Bruker 15 T solariX XR FTICR mass spectrometer equipped with a CombiSource and a novel ParaCell (Bruker Daltonics, Bremen, Germany). The MALDI-FTICR system was controlled by ftmsControl software and equipped with a Bruker Smartbeam-II Laser System that operated at a frequency of 500 Hz. The “medium” predefined shot pattern was used for the irradiation. All acquisition settings were optimized to reach the sensitivity and resolving power needed for MALDI-FTICR-MS analysis of RNase B and myoglobin and are provided as Supporting Information in Tables S1 and S2, respectively. These settings include parameters with regard to transient length, ion transfer (in the octapole and in the collision hexapole), transfer optics (hexapole), and the ICR cell analyzer (ParaCell). For ISD experiments, two different MALDI-FTICR-MS methods were used (see Tables S1 and S2, Supporting Information) and data were compiled to generate a single MS/MS-spectrum for RNase B and for myoglobin. The online calibration tool was used during the acquisition of the spectra to improve the mass measurement precision. Pseudo-MS3 (i.e., ISD-CID) experiments were performed on selected precursor ions (ISD fragments isolated in the hexapole collision cell) using the Selective Accumulation option set to a minimum intensity of 107 and then fragmented by CID. Direct infusion ESI-FTICR-MS experiments of RNase B were performed using the instrumental settings reported in Table S1, Supporting Information. A quadrupole (Q) was used for precursor ion selection with an isolation window of 8 Da. Direct infusion electrospray ionization (ESI) experiments were carried out at an infusion rate of 2 μL/min using an RNase B solution at 20 μg/mL (in acetonitrile/water 1:1). ETD experiments were performed with accumulation times of 10 s and 100 ms for the analyte and the ETD reagent, respectively. The ETD reaction was allowed for 5 ms, and ETD fragment ions were analyzed in the ICR cell. DataAnalysis Software 4.2 (Bruker Daltonics) was used for the visualization and the calibration of both MALDI- and ESI-FTICR spectra. Theoretical m/z-values of ISD, CID, and ETD fragment ions were obtained using the “MS-product” tool found in http://prospector.ucsf.edu/prospector/mshome.htm.
ionization process and thus allow the identification of ISD fragment ions from regions between cysteines that were connected through S−S bonds, resulting in an increased protein sequence coverage.24,25 The MALDI-ISD spectra are characterized by the presence of extensive series of singly charged fragment ions. Singly charged ions on the one hand facilitate the interpretation of ISD spectra but, on the other hand, require a wide detection range (i.e., up to the mass of the protein) in order to identify a complete fragment ion series. In other words, the detection of MALDI-ISD fragment ions is limited to the m/zrange that is provided by the MS system. Most commonly, MALDI-ISD spectra of peptides and proteins have been acquired on time-of-flight (TOF) mass analyzers.26−29 Although such analyzers exhibit a wide detection range, the sensitivity and resolving power are not sufficient to tackle the complexity of ISD spectra obtained from intact proteins and, consequently, the obtained information is limited to portions of the sequence close to both the N- and C-terminus.30,31 In the current study, we have explored a Fourier transform ion cyclotron resonance (FTICR) platform for the measurement of MALDI-ISD fragments from two different proteins. FTICR systems offer the highest mass accuracy and resolving power and provide MS/MS-information superior to TOF measurements.32 Compared to MALDI-TOF platforms,33−35 FTICR covers a similar detection range, however with full isotopic resolution at much higher m/z-values. Previously, we have reported on MALDI-FTICR-MS methods for the analysis of peptides and small proteins in human serum up to approximately 15 kDa.36,37 Here, identifications of new proteoforms were initially based on accurate mass difference measurements and then confirmed by ESI-FTICR-MS/MS experiments. In the current study, we have aimed for the further development of new ultrahigh resolution MALDI-FTICR-MS methods and the application of these methods for the characterization of intact proteins by ISD experiments with fragment ion detection at isotopic resolution. To this end, both ribonuclease B (RNase B) and myoglobin were subjected to MALDI-ISD-FTICR-MS analysis, and the results were evaluated with regard to previously published data obtained by both ESIMS/MS and MALDI-TOF-MS.
■
EXPERIMENTAL SECTION Chemicals. Bovine pancreatic ribonuclease B (RNase B), 1,5diaminonaphthalene (1,5-DAN; handle with care, check material safety data sheet), myoglobin from equine skeletal muscle, αcyano-4-hydroxycinnamic acid, acetonitrile, and formic acid were purchased from Sigma-Aldrich (St. Louis, MO, USA). Acetone and ethanol were purchased from Merck (Darmstadt, Germany). RNase B and myoglobin were dissolved in Milli-Q water to a final concentration of 5 and 6 mg/mL, respectively, and then stored at −20 °C until further use. MALDI Spotting. For ISD-MS analysis, 1 μL of myoglobin (300 μg/mL in water) was spotted onto a MALDI ground steel target plate (Bruker Daltonics, Bremen, Germany) and mixed on the plate with 1 μL of 1,5-DAN (saturated solution in acetonitrile/water, 1:1) using a 10 μL pipet tip (Rainin tips from Mettler-Toledo, Tiel, Netherlands). Gentle mixing was performed until small crystals were visible by eye; then, the mixture was allowed to dry at room temperature. The same procedure was performed for RNase B (250 μg/mL in water). RNase B was also spotted onto a MALDI AnchorChip (600 μm; Bruker Daltonics, Bremen, Germany). To this end, 2 μL of RNase B (20 μg/mL in acetonitrile/water/formic acid, 50:49.95:0.05%) was transferred into a 1.5 mL Eppendorf tube
■
RESULTS AND DISCUSSION Intact Protein Analysis by MALDI-FTICR-MS. Measurement of a protein’s accurate mass and the determination of mass differences between various proteoforms can lead to the identification of possible PTMs.4,38 In this study, we employed a 15 T MALDI-FTICR system equipped with a dynamically harmonized ICR cell (ParaCell) to measure ultrahigh resolution spectra from small proteins. Isotopically resolved MALDIFTICR-MS spectra of intact ribonuclease B (RNase B) and myoglobin were obtained as is shown in Figure 1. RNase B consists of 124 amino acids, contains four disulfide bonds, and is glycosylated at asparagine(Asn)-34 with two β-1,4-linked Nacetylglucosamines and a varying number of mannose residues (i.e., from 5 to 9).39 Five RNase B glycoforms40 were profiled by MALDI-FTICR-MS as [M + H]+ species (Figure 1A), with an average resolving power of 61 000 (Figure 1C). The baseline resolution of the isotopic distributions allowed the accurate determination of the mass difference between consecutive glycoforms. The average mass difference of 162.10 mass units, observed between the most intense isotopic peak of consecutive 3430
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
intact proteins by MALDI-FTICR-MS needs further investigation and will be presented later. Top-Down Characterization of RNase B by MALDI-ISDFTICR-MS and Comparison with ESI-FTICR-MS/MS. Most commonly, when performing top-down MS/MS-characterization of proteins first, a precursor ion (e.g., using a quadrupole) is selected, either a specific charge state (in case of ESI) or a specific protein species (in case of a mixture). However, in MALDI-ISD experiments, the isolation of a precursor ion is not possible, since fragmentation occurs in the ionization source. Therefore, selection of a single RNase B glycoform for top-down MALDI-ISD-FTICR-MS analysis was not possible and ISD fragmentation was simultaneously performed on all five glycoforms. To this end, RNase B was spotted with 1,5-DAN matrix on a ground steel target plate and measured using two different MALDI-FTICR-MS methods (as described in the Experimental Section). The resulting MALDI-ISD spectra were compiled and are depicted in Figure 2. The most intense peaks detected in the mass range from m/z-value 1000 to m/z-value 4000 (upper panel) were identified as c-type ions while z-, a-, y-, and w-type fragments were detected with lower intensities. Furthermore, it should be noted that all z-type fragments contained an additional hydrogen atom, explicitly indicated with a prime (i.e., z′). In this mass range, the series of c-type fragments ends with c33 (highlighted in the green box). The smallest ISD fragment that allows the identification of the glycosylation site is c34. The calculated mass difference between the c33 and the glycosylated c34 fragment is the sum of the mass of one Asn residue (i.e., 114.0429 Da) and the mass of one of the five possible glycan structures. The smallest and most abundant glycan moiety of RNase B contains two N-acetylglucosamines and five mannoses (GlcNAc2Man5) corresponding to a theoretical mass difference between c33 and c34+GlcNAc2Man5 of 1330.47 Da. A difference of 1330.48 mass units was indeed measured between the c33 fragment (observed m/z-value 3681.6468) and fragment peak detected at m/z-value 5012.1233 (Figure 2 lower panel), thus confirming glycosylation at Asn-34. Interestingly, in the mass range between the c33 and c34+GlcNAc2Man5, z′-, y-, and w-type ions were detected as the most abundant peaks while in the mass range from m/z-value 5000 to m/z-value 6500 glycosylated c-type ions, containing either a GlcNAc2Man5 or GlcNAc2Man6 glycan moiety, were also detected (typical examples of isotopically resolved MALDI-ISD fragment ions of RNase B and their assignments are provided as Supporting Information, Figure S1). In addition, the presence of w-type ions provided information on the amino acid from which these originated,46 and this data was used to validate the identifications of z′-type ions. For example, the mass difference between the peak detected at m/z-value 3849.8493 (z′35) and the peak detected at m/z-value 3936.8857 (z′36) was 87.0364 Da (theoretical value of 87.0320 Da) corresponding to the mass of a serine residue (see Figure 3). This was further corroborated by the measurement of a mass difference of 17.9933 Da between the peak detected at m/z-value 3918.8924 (w36) and the peak at m/ z-value 3936.8857 (z′36). In fact, the theoretical mass difference between a z′- and a w-type ion generated from a serine is 18.0106 Da and corresponds to the sum of the mass of one hydroxyl group lost by α-cleavage after the formation of the z-type radical ion during ISD and one hydrogen.46 All together, MALDI-ISDMS of RNase B resulted in a sequence coverage of 64%. It should be noted that mass measurements were performed from m/zvalue 1012 to avoid the interference from matrix peaks and consequently both the N- and C-terminus were not charac-
Figure 1. Parts of the ultrahigh resolution MALDI-FTICR-MS spectra of RNase B and myoglobin. (A) Five glycoforms of RNase B were observed as singly charged forms in the mass range from m/z-value 14 800 to m/z-value 15 550. (B) Intact myoglobin was measured as the [M + H]+ form and as adduct with 1,5-DAN matrix in the mass range from m/z-value 16 900 to m/z-value 17 270. (C and D) Further enlargements of MALDI-FTICR-MS spectra depicting baseline resolved isotopic distribution of RNase B (i.e., GlcNAc2Man5 form) and myoglobin, respectively, with resolving powers of 61 000 and 62 000, respectively.
glycoforms, is in good agreement with the theoretical mass of a hexose (i.e., mannose) residue (i.e., 162.05 Da). Second, myoglobin, consisting of 153-amino acids,41 was measured by MALDI-FTICR-MS as [M + H]+ at an average resolving power of 62 000 (see Figure 1B,D). To our knowledge, this is the largest protein measured by MALDI-FTICR-MS with baseline isotopic resolution, albeit that such ultrahigh resolution measurements at high m/z-values required long acquisition times (due to the relatively low frequency of 13.56 kHz at m/z-value 16 967.23).42 Thus, RNase B and myoglobin were measured by MALDI-FTICR-MS with time-domain transient lengths of 7.98 and 9.23 s for a single scan, respectively. Furthermore, it should be noted that high resolving powers (e.g., 750 000) can be achieved for small molecules (e.g., m/z-value 600) with shorter time-domain transient length (e.g., 3.35 s) using the same FTICR-MS instrument (data not shown). A higher transient length implies a reduced sensitivity due to collisions of ions and residual gas molecules in the ICR cell. Obviously, at higher magnetic field (e.g., 21 T43), ultrahigh resolution measurements can be performed with shorter transients and may thus result in improved sensitivities. Similarly, measurement with longer timedomain transients and a consequent higher resolution may be performed with an improved ultrahigh vacuum in the ICR cell. Another way to efficiently improve the resolving power is postprocessing the acquired spectra by displaying these in absorption mode.44,45 For the current data, the resolving power of intact myoglobin could be increased up to 77 300 in an absorption-mode spectrum. The beneficial aspect of using an absorption-mode spectrum is further exemplified in identifying fragment ions (see Figure S4, Supporting Information; fragmentation process will be described hereafter). The routine application of absorption-mode spectra for the analysis of small 3431
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
Figure 2. Sequence coverage obtained from the analysis of RNase B by MALDI-ISD-FTICR-MS using 1,5-DAN as MALDI matrix (upper panel). (A and B) Parts of the ultrahigh resolution MALDI-ISD-FTICR-MS spectrum of RNase B glycoforms. The detection of both the c33 and the glycosylated c34 fragment ions (highlighted in the green boxes) allowed the localization of the most abundant glycan moiety (i.e., GlcNAc2Man5) at Asn-34 (indicated in bold in the protein sequence).
additional peptide bonds were sequenced, resulting in an extended sequence coverage of RNase B of 72%. In the MALDI-ISD spectrum of RNase B (Figure 2), it is furthermore noted that multiple fragment ions are observed in the protein part that contains eight cysteines all involved in disulfide bonds. These four different S−S bridges are reduced due to the hydrogen-donator property of the 1,5-DAN matrix. In a recent top-down study, disulfide bonds in intact RNase B were reduced and carbamidomethylated prior to ESI-FTICR-MS/MS analysis. In this study, a total sequence coverage of 86% was reported, based on a combination of different fragmentation
terized. This lack of coverage can be improved by performing socalled pseudo-MS3 experiments, in which N- and C-terminal regions of a protein can be characterized by MS/MS analysis performed on isolated fragment ions.28,47 With this approach, interfering matrix cluster peaks in the low mass range are removed and the sequence coverage obtained from MALDI-ISDMS experiments can be further extended. For the example of RNase B, the c12- and the w15-MALDI-ISD fragment ions were first isolated in the hexapole collision cell and then fragmented by CID (see the Experimental Section). The resulting pseudo-MS3 spectra are depicted in Figure 4. From these spectra, 15 3432
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
scan as Supporting Information, Figure S3). The identification of ISD fragment ions was based either on the detection of the monoisotopic peak or on the comparison between the theoretical and observed isotopic distribution, or both. Thus, although fragment ions were detected up to m/z-value 16 772, reliable identifications were obtained up to m/z-value 12 000. Similar to the MALDI-ISD data from RNase B, only c- and z′-type ions were annotated in the spectrum. These fragment ions were the most abundant and were measured with resolving powers ranging from 278 000 at m/z-value 1018.45 to 43 000 at m/zvalue 11 998.54. Intact myoglobin itself was measured as an [M + H]+ with a resolving power of 25 600. Note that measurements at higher resolution allowed the detection with baseline isotopic resolution of the intact myoglobin but were inadequate for the detection of ISD fragments due to the lower sensitivity. Previously, myoglobin has been characterized using a MALDIISD-TOF instrument at a moderate resolution resulting in a sequence coverage of 59%.46,50,51 Unfortunately, these experiments suffered from two limitations: (1) the signal-to-noise ratio (S/N) of the observed ISD fragments rapidly decreased with the increased m/z-range; (2) the fragment ions were detected as broad single peaks. Consequently, only high abundant fragment ions were detected and assigned. By using a state-of-the-art 15 T MALDI-FTICR system equipped with a new ParaCell for ICR measurement, the data dramatically improved. The sequence coverage based on c- and z′-type ions was 87%. The ultrahigh resolution MALDI-FTICR measurements allowed the identification of fragment ions that in low resolution MALDI-TOF would have not been detected as a single species. For example, the c32 and the z′33 fragment ions at theoretical m/z-values 3515.8394 and 3517.7420, respectively, were fully resolved with resolving power of 74 000 (see the Supporting Information, Figure S4). Despite such high resolving powers, not all fragment ions could be uniquely assigned, exemplified by a c57 and a z′58 at theoretical m/z-values 6459.3347 and 6462.4224, respectively (see the Supporting Information, Figure S4). These fragment ions were measured at a resolving power of 71 000 whereas theoretically more than 255 000 would be needed to baseline resolve the two isotopic distributions. Nevertheless, although the z′58 was not identified (because of the overlap with the c57), the presence of the K96 was confirmed by the identification of the y58 fragment ion (data not shown). The resolving power of MALDI-FTICR-MS spectra can be improved postacquisition by displaying the spectra in absorption mode instead of in magnitude mode.44,45 Upon performing this for the MALDI-ISD-FTICR-MS spectrum of myoglobin (i.e., obtained using Method 2; see Table S2, Supporting Information), on average, the resolving power increased by almost 3-fold (data not shown). For example, the resolving power at m/z-value 6462.35 (i.e., c57 fragment ion) increased from 70 557 in magnitude mode to 205 165 in absorption mode. This improvement allowed one to resolve the c57 fragment ion from the z′58 (see Figure S4, Supporting Information). It was observed that the signal-to-noise ratio in the absorption-mode spectrum was on average one-third lower (data not shown). It should however be noted that the confidence in identifying fragment ions in the absorption-mode spectra requires further investigation. Recently, Shaw and co-workers have reported on the ESI-MS/ MS characterization of myoglobin using different fragmentation techniques.7 It was shown that the use of ultraviolet photodissociation (UVPD) yields a drastically improved sequence coverage (higher than 93%) of myoglobin compared to CID,
Figure 3. Part of the ultrahigh resolution MALDI-ISD-FTICR-MS spectrum of RNase B. Note that different kinds of fragment ions (e.g., w36, z′36, y36) generated from cleavage of the same peptide bond improve the reliability of the assignments. The mass difference between w- and z-type ions generated from the same peptide bond may be diagnostic for the amino acid involved in the fragmentation. In fact, the mass difference of 17.9933 Da between the peak detected at m/z-value 3918.8924 (w36) and the peak at m/z-value 3936.8857 (z′36) indicates the loss of the hydroxyl group from the Ser-36 (i.e., 17.0027 (OH) + 1.0078 (H) = 18.0011 Da).
techniques (i.e., CID, IRMPD, ECD, ETD).48 Although our MALDI-ISD-FTICR-MS experiments resulted in a lower sequence coverage (i.e., 72% including the pseudo-MS3 data), some complementary information was obtained from the characterization of eight more peptide bonds (namely, Q28− M29, P42−V43, D83−C84, P93−N94, H105−I106, I106−I107, V108− A109, A109−C110). Note that the simultaneous fragmentation of the five different glycoforms during ISD experiments hampered the sequence coverage by increasing the complexity of the spectrum in the mass region higher than m/z-value 6500 (data not shown). Thus, the separation of the different glycoforms prior to ISD analysis may result in a more extensive characterization of RNase B. For comparative purposes, top-down ETD experiments were performed on nonreduced RNase B using the 15 T FTICR-MS system in ESI mode with precursor ion selection of the “15 plus” charge state of RNase B. From the obtained ETD spectrum, it became clear that the presence of disulfide bonds affected the analysis in the way that the fragmentation was limited to “S−S bond free regions” (see Supporting Information, Figure S2). In fact, due to the connection between the different cysteines,49 only fragment ions in the region close to both the N- and the Cterminus were observed. As a result, the sequence coverage was limited to 29% while from the previously reported ETD experiments on the reduced and carbamidomethylated RNase B it was 58%.48 Top-Down Characterization of Myoglobin by MALDIISD-FTICR-MS. As discussed above, both RNase B and myoglobin were measured with isotopic resolution by MALDIFTICR-MS, i.e., up to a mass range of m/z-value 17 000. In the MALDI-ISD analysis of RNase B, however, the interpretation of fragment ions at m/z-values larger than 6500 was hampered by overlapping signals due to the presence of different glycoforms. In the case of myoglobin, MALDI-ISD fragment ions could be assigned up to as high as m/z-value 12 000. To this end, myoglobin was spotted on a ground steel target plate with 1,5DAN and analyzed using two different MALDI-ISD-FTICR-MS methods (see Table S2, Supporting Information). The resulting two spectra were compiled and are depicted in Figure 5 (and full 3433
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
Figure 4. Pseudo-MS3 spectra of RNase B obtained from the analysis of isolated c12 and w15 MALDI-ISD fragment ions by CID and FTICR-MS.
HCD, and ETD, which provided a sequence coverage from a minimum of about 45% (CID) to a maximum of 71% (ETD). Unfortunately, details on the observed fragment ions in UVPD experiments were not reported; thus, the complementarity of ISD to this fragmentation technique could not be determined. One of the difficulties in ESI-MS/MS experiments on multiply protonated intact proteins is the generation of hundreds of product ions that lead to very complex spectra. For example, from the UVPD analysis of five different charge states of myoglobin, more than 1450 product ions were reported. The interpretation of such complex spectra is tedious. On the contrary, MALDI-ISD yields mainly singly charged ions thus
drastically reducing the complexity of the generated spectra. MALDI-ISD MS spectra can be acquired in a much shorter time frame (i.e., seconds) than ESI-MS/MS (i.e., minutes) making this technique more suitable for high-throughput analysis. A final remark is that ISD fragments of myoglobin were measured starting from m/z-value 1012, for similar reasons as discussed for RNase B. The optimization of the ration protein/ matrix and measurements in negative ion mode may lead to a further characterization of regions of the protein sequence near both the N- and the C-terminus.46 3434
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
Figure 5. Sequence coverage obtained from the analysis of myoglobin by MALDI-ISD-FTICR-MS using 1,5-DAN as MALDI matrix (upper panel). (A, B, C, and D) Parts of the ultrahigh resolution MALDI-ISD-FTICR-MS spectrum of myoglobin showing isotopically resolved fragment ions up to m/zvalue 12 000. The full MALDI-ISD-FTICR-MS spectrum of myoglobin is depicted in the Supporting Information, Figure S3.
■
CONCLUSIONS Structural characterization of intact proteins by the top-down ESI-MS/MS approach is commonly performed using different fragmentation techniques that provide complementary information with regard to the amino acid sequence. The combined data results in an improved sequence coverage; however, complete
protein sequence coverage is rarely obtained. In-source decay provides information that in combination with ESI-MS/MS readouts can further extend sequence coverage. So far, this information has been limited by the fact that MALDI-TOF-MS instruments provide suitable resolution and sensitivity limited to m/z-values 4000 up to 5000. In the current study, we have used 3435
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
(12) Marshall, A. G.; Hendrickson, C. L. Annu. Rev. Anal. Chem. 2008, 1, 579−599. (13) Zubarev, R. A.; Makarov, A. Anal. Chem. 2013, 85, 5288−5296. (14) Fornelli, L.; Parra, J.; Hartmer, R.; Stoermer, C.; Lubeck, M.; Tsybin, Y. Anal. Bioanal. Chem. 2013, 405, 8505−8514. (15) Debois, D.; Smargiasso, N.; Demeure, K.; Asakawa, D.; Zimmerman, T.; Quinton, L.; De Pauw, E. In Applications of MALDITOF Spectroscopy; Cai, Z., Liu, S., Eds.; Springer: Berlin Heidelberg, 2013; pp 117−141. (16) Hardouin, J. Mass Spectrom. Rev. 2007, 26, 672−682. (17) Macht, M. TrAC, Trends Anal. Chem. 2013, 48, 62−71. (18) Hanisch, F. G. Anal. Chem. 2011, 83, 4829−4837. (19) Quinton, L.; Demeure, K.; Dobson, R.; Gilles, N.; Gabelica, V.; De Pauw, E. J. Proteome Res. 2007, 6, 3216−3223. (20) Demeure, K.; Gabelica, V.; De Pauw, E. J. Am. Soc. Mass Spectrom. 2010, 21, 1906−1917. (21) Takayama, M. J. Am. Soc. Mass Spectrom. 2001, 12, 1044−1049. (22) Köcher, T.; Engström, Å.; Zubarev, R. A. Anal. Chem. 2005, 77, 172−177. (23) Takayama, M. J. Am. Soc. Mass Spectrom. 2001, 12, 420−427. (24) Fukuyama, Y.; Iwamoto, S.; Tanaka, K. J. Mass Spectrom. 2006, 41, 191−201. (25) Smargiasso, N.; Quinton, L.; De Pauw, E. J. Am. Soc. Mass Spectrom. 2012, 23, 469−474. (26) Rand, K. D.; Bache, N.; Nedertoft, M. M.; Jørgensen, T. J. D. Anal. Chem. 2011, 83, 8859−8862. (27) Ait-Belkacem, R.; Berenguer, C.; Villard, C.; Ouafik, L.; FigarellaBranger, D.; Chinot, O.; Lafitte, D. Proteomics 2014, 14, 1290−1301. (28) Asakawa, D.; Smargiasso, N.; De Pauw, E. Anal. Chem. 2014, 86, 2451−2457. (29) Calligaris, D.; Villard, C.; Terras, L.; Braguer, D.; Verdier-Pinard, P.; Lafitte, D. Anal. Chem. 2010, 82, 6176−6184. (30) Ayoub, D.; Jabs, W.; Resemann, A.; Evers, W.; Evans, C.; Main, L.; Baessmann, C.; Wagner-Rousset, E.; Suckau, D.; Beck, A. mAbs 2013, 5, 699−710. (31) Takayama, M.; Nagoshi, K.; Iimuro, R.; Inatomi, K. Int. J. Mol. Sci. 2014, 15, 8428−8442. (32) Nikolaev, E. N.; Kostyukevich, Y. I.; Vladimirov, G. N. Mass Spectrom. Rev. 2014, DOI: 10.1002/mas.21422. (33) Nakanishi, T.; Okamoto, N.; Tanaka, K.; Shimizu, A. Biol. Mass Spectrom. 1994, 23, 230−233. (34) Solouki, T.; Gillig, K. J.; Russell, D. H. Anal. Chem. 1994, 66, 1583−1587. (35) Solouki, T.; Emmett, M. R.; Guan, S.; Marshall, A. G. Anal. Chem. 1997, 69, 1163−1168. (36) Nicolardi, S.; van der Burgt, Y. E.; Wuhrer, M.; Deelder, A. M. Proteomics 2013, 13, 992−1001. (37) Nicolardi, S.; van der Burgt, Y. E.; Dragan, I.; Hensbergen, P. J.; Deelder, A. M. J. Proteome Res. 2013, 12, 2260−2268. (38) Bondarenko, P.; Second, T.; Zabrouskov, V.; Makarov, A.; Zhang, Z. J. Am. Soc. Mass Spectrom. 2009, 20, 1415−1424. (39) Thaysen-Andersen, M.; Mysling, S.; Højrup, P. Anal. Chem. 2009, 81, 3933−3943. (40) Liang, C.-J.; Yamashita, K.; Kobata, A. J. Biochem. 1980, 88, 51− 58. (41) Dautrevaux, M.; Boulanger, Y.; Han, K.; Biserte, G. Eur. J. Biochem. 1969, 11, 267−277. (42) Amster, I. J. J. Mass Spectrom. 1996, 31, 1325−1337. (43) Kaiser, N.; Weisbrod, C.; Quinn, J.; Blakney, G.; Beu, S.; Chen, T.; Smith, D.; Hendrickson, C.; Marshall, A. Development of an FT-ICR mass spectrometer in preparation for 21 Tesla. 62nd American Society for Mass Spectrometry (ASMS) Conference on Mass Spectrometry and Allied Topics, Baltimore, Maryland, June 15−19, 2014; Oral Presentation. (44) Kilgour, D. P. A.; Wills, R.; Qi, Y.; O’Connor, P. B. Anal. Chem. 2013, 85, 3903−3911. (45) Kilgour, D. P. A.; Neal, M. J.; Soulby, A. J.; O’Connor, P. B. Rapid Commun. Mass Spectrom. 2013, 27, 1977−1982. (46) Asakawa, D.; Smargiasso, N.; De Pauw, E. J. Am. Soc. Mass Spectrom. 2013, 24, 297−300.
an ultrahigh resolution MALDI-FTICR system to perform high mass measurements of ISD fragment ions generated from myoglobin and RNase B proteoforms. Intact, singly charged species were measured with resolving powers of 62 000 (at m/zvalue 16 950) and (at m/z-value 14 900) 61 000 for myoglobin and RNase B, respectively. The high mass resolution and improved sensitivity for higher m/z-values of these MALDI-ISDFTICR measurements yielded a high sequence coverage for both proteins, 87% for myoglobin and 72% for RNase B. For RNase B, complementary sequence information to previously published ESI-MS/MS data was obtained and the use of pseudo-MS3 experiments (CID) on selected ISD fragment ions allowed the characterization of both the N- and C-terminal regions. In the case of myoglobin, MALDI-ISD fragment ions were identified up to m/z-value 12 000 (with resolving power of 39 000 at m/zvalue 12 000) resulting in a sequence coverage of 87%. Although an even higher sequence coverage (93%) was reported using a UVPD in a hybrid linear ion trap Orbitrap mass spectrometer, the advantage of using MALDI-ISD-FTICR-MS to characterize myoglobin is a shorter acquisition time and the generation of less complex spectra.
■
ASSOCIATED CONTENT
S Supporting Information *
Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS The authors thank Dr. David Kilgour for providing absorptionmode spectra of myoglobin using Autophaser. REFERENCES
(1) Siuti, N.; Kelleher, N. L. Nat. Methods 2007, 4, 817−821. (2) Tran, J. C.; Zamdborg, L.; Ahlf, D. R.; Lee, J. E.; Catherman, A. D.; Durbin, K. R.; Tipton, J. D.; Vellaichamy, A.; Kellie, J. F.; Li, M.; Wu, C.; Sweet, S. M. M.; Early, B. P.; Siuti, N.; LeDuc, R. D.; Compton, P. D.; Thomas, P. M.; Kelleher, N. L. Nature 2011, 480, 254−258. (3) Heck, A. J. R. Nat. Methods 2008, 5, 927−933. (4) Tipton, J. D.; Tran, J. C.; Catherman, A. D.; Ahlf, D. R.; Durbin, K. R.; Lee, J. E.; Kellie, J. F.; Kelleher, N. L.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2012, 84, 2111−2117. (5) Han, X.; Jin, M.; Breuker, K.; McLafferty, F. W. Science 2006, 314, 109−112. (6) Tsybin, Y. O.; Fornelli, L.; Stoermer, C.; Luebeck, M.; Parra, J.; Nallet, S.; Wurm, F. M.; Hartmer, R. Anal. Chem. 2011, 83, 8919−8927. (7) Shaw, J. B.; Li, W.; Holden, D. D.; Zhang, Y.; Griep-Raming, J.; Fellers, R. T.; Early, B. P.; Thomas, P. M.; Kelleher, N. L.; Brodbelt, J. S. J. Am. Chem. Soc. 2013, 135, 12646−12651. (8) Madsen, J. A.; Gardner, M. W.; Smith, S. I.; Ledvina, A. R.; Coon, J. J.; Schwartz, J. C.; Stafford, G. C.; Brodbelt, J. S. Anal. Chem. 2009, 81, 8677−8686. (9) Mao, Y.; Valeja, S. G.; Rouse, J. C.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2013, 85, 4239−4246. (10) Ahlf, D. R.; Compton, P. D.; Tran, J. C.; Early, B. P.; Thomas, P. M.; Kelleher, N. L. J. Proteome Res. 2012, 11, 4308−4314. (11) Mihalca, R.; van der Burgt, Y. E. M.; McDonnell, L. A.; Duursma, M.; Cerjak, I.; Heck, A. J. R.; Heeren, R. M. A. Rapid Commun. Mass Spectrom. 2006, 20, 1838−1844. 3436
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry (47) Suckau, D.; Resemann, A. Anal. Chem. 2003, 75, 5817−5824. (48) Bourgoin-Voillard, S.; Leymarie, N.; Costello, C. E. Proteomics 2014, 14, 1174−1184. (49) Smyth, D. G.; Stein, W. H.; Moore, S. J. Biol. Chem. 1963, 238, 227−234. (50) Takayama, M. J. Mass Spectrom. Soc. Jpn. 2002, 50, 304−310. (51) Takayama, M.; Osaka, I.; Sakakura, M. Mass Spectrom. (Tokyo) 2012, 1, A0001.
3437
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Top-Down MALDI-In-Source Decay-FTICR Mass Spectrometry of Isotopically Resolved Proteins Simone Nicolardi,* Linda Switzar, André M. Deelder, Magnus Palmblad, and Yuri E.M. van der Burgt Center for Proteomics and Metabolomics, Leiden University Medical Center (LUMC), PO Box 9600, 2300 RC, Leiden, The Netherlands S Supporting Information *
ABSTRACT: An accurate mass measurement of a known protein provides information on potential amino acid deletions and post-translational modifications. Although this field is dominated by strategies based on electrospray ionization, mass spectrometry (MS) methods using matrix-assisted laser desorption/ionization (MALDI) have the advantage of yielding predominantly singly charged precursor ions, thus avoiding peak overlap from different charge states of multiple species. Such MALDI-MS methods require mass measurement at ultrahigh resolution, which is provided by Fourier transform ion cyclotron resonance (FTICR) mass analyzers. Recently, using a MALDI-FTICR-MS platform equipped with a 15 T magnet, we reported on the mass analysis of intact human serum peptides and small proteins with isotopic resolution up to ∼15 kDa and identified new proteoforms from an accurate measurement of mass distances. In the current study, we have used this FTICR system after an upgrade with a novel dynamically harmonized ICR cell, i.e., ParaCell, for mapping isotopically resolved intact proteins up to about 17 kDa and performed topdown MALDI in-source decay (ISD) analysis. Standard proteins myoglobin (m/z-value 16 950) and ribonuclease B (m/z-value 14 900) were measured with resolving powers of 62 000 and 61 000, respectively. Furthermore, it will be shown that (singly charged) MALDI-ISD fragment ions can be measured at isotopic resolution up to m/z-value 12 000 (e.g., resolving power 39 000 at m/z-value 12 000) providing more reliable identifications. Moreover, examples are presented of pseudo-MS3 experiments on ISD fragment ions from RNase B by collisional-induced dissociation (CID).
I
Matrix-assisted laser desorption/ionization (MALDI) insource decay (ISD) MS is an attractive alternative compared to ESI-MS-based methods for sequencing peptides and proteins.15−18 Ideally, the presence of one single proteoform within one MALDI spot is to be preferred to avoid the simultaneous detection of fragment ions from different species. The purification and/or separation (e.g., by LC) of different proteins and/or proteoforms prior to MALDI-ISD MS analysis allows a more confident identification.19 It has been shown that ions produced by MALDI in combination with specific matrices increase their internal energy to the point that fragmentation events are favored prior to extraction from the ionization source.20 These events mainly yield c-, y-, and z-type fragment ions and to a lesser extent w- and a-type ions.16,21−23 The presence of complementary series of fragment ions in a single spectrum results in a more reliable identification of the amino acid sequence and improves the sequence coverage. Moreover, hydrogen-donor MALDI matrices (e.g., 1,5-diaminonaphthalene (1,5-DAN)) catalyze the reduction of disulfide bonds during the
ntact protein analysis by top-down mass spectrometry (MS) allows both the determination of amino acid sequence and the identification of the site of post-translational modifications (PTMs) without the need for proteolytic digestion.1 A comprehensive picture of proteoforms contributes to a more detailed description of the proteome and its interactions, similar to the additional layer of information on the epigenetic code.2 Nowadays, large proteins and even protein complexes can be routinely mass-analyzed by (native) electrospray ionization (ESI)-MS approaches3 and subsequently sequenced using various tandem-MS fragmentation techniques such as collisional-induced dissociation (CID), electron-transfer dissociation (ETD), electron-capture dissociation (ECD), higher energy collisional dissociation (HCD), infrared multiphoton dissociation (IRMPD), or ultraviolet photodissociation (UVPD). Often, these fragmentation strategies provide complementary information on the amino acid sequence and PTMs.4−11 In order to obtain the sensitivity and resolution needed to tackle the complexity of the generated MS- and MS/MS-spectra, top-down analyses of intact proteins require (ultra)high resolution MSinstrumentation with resolving power of approximately 100 000, e.g., Orbitrap, Fourier transform ion cyclotron resonance (FTICR), and time-of-flight (TOF).12−14 © 2015 American Chemical Society
Received: December 18, 2014 Accepted: February 26, 2015 Published: February 26, 2015 3429
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
and mixed with 10 μL of an α-cyano matrix (0.3 g/L in ethanol/ acetone, 2:1). Then, 1 μL of the eluate/matrix mixture was spotted on the target plate. MALDI- and ESI-FTICR Mass Spectrometry. Both MALDI- and ESI-FTICR-MS experiments were performed on a Bruker 15 T solariX XR FTICR mass spectrometer equipped with a CombiSource and a novel ParaCell (Bruker Daltonics, Bremen, Germany). The MALDI-FTICR system was controlled by ftmsControl software and equipped with a Bruker Smartbeam-II Laser System that operated at a frequency of 500 Hz. The “medium” predefined shot pattern was used for the irradiation. All acquisition settings were optimized to reach the sensitivity and resolving power needed for MALDI-FTICR-MS analysis of RNase B and myoglobin and are provided as Supporting Information in Tables S1 and S2, respectively. These settings include parameters with regard to transient length, ion transfer (in the octapole and in the collision hexapole), transfer optics (hexapole), and the ICR cell analyzer (ParaCell). For ISD experiments, two different MALDI-FTICR-MS methods were used (see Tables S1 and S2, Supporting Information) and data were compiled to generate a single MS/MS-spectrum for RNase B and for myoglobin. The online calibration tool was used during the acquisition of the spectra to improve the mass measurement precision. Pseudo-MS3 (i.e., ISD-CID) experiments were performed on selected precursor ions (ISD fragments isolated in the hexapole collision cell) using the Selective Accumulation option set to a minimum intensity of 107 and then fragmented by CID. Direct infusion ESI-FTICR-MS experiments of RNase B were performed using the instrumental settings reported in Table S1, Supporting Information. A quadrupole (Q) was used for precursor ion selection with an isolation window of 8 Da. Direct infusion electrospray ionization (ESI) experiments were carried out at an infusion rate of 2 μL/min using an RNase B solution at 20 μg/mL (in acetonitrile/water 1:1). ETD experiments were performed with accumulation times of 10 s and 100 ms for the analyte and the ETD reagent, respectively. The ETD reaction was allowed for 5 ms, and ETD fragment ions were analyzed in the ICR cell. DataAnalysis Software 4.2 (Bruker Daltonics) was used for the visualization and the calibration of both MALDI- and ESI-FTICR spectra. Theoretical m/z-values of ISD, CID, and ETD fragment ions were obtained using the “MS-product” tool found in http://prospector.ucsf.edu/prospector/mshome.htm.
ionization process and thus allow the identification of ISD fragment ions from regions between cysteines that were connected through S−S bonds, resulting in an increased protein sequence coverage.24,25 The MALDI-ISD spectra are characterized by the presence of extensive series of singly charged fragment ions. Singly charged ions on the one hand facilitate the interpretation of ISD spectra but, on the other hand, require a wide detection range (i.e., up to the mass of the protein) in order to identify a complete fragment ion series. In other words, the detection of MALDI-ISD fragment ions is limited to the m/zrange that is provided by the MS system. Most commonly, MALDI-ISD spectra of peptides and proteins have been acquired on time-of-flight (TOF) mass analyzers.26−29 Although such analyzers exhibit a wide detection range, the sensitivity and resolving power are not sufficient to tackle the complexity of ISD spectra obtained from intact proteins and, consequently, the obtained information is limited to portions of the sequence close to both the N- and C-terminus.30,31 In the current study, we have explored a Fourier transform ion cyclotron resonance (FTICR) platform for the measurement of MALDI-ISD fragments from two different proteins. FTICR systems offer the highest mass accuracy and resolving power and provide MS/MS-information superior to TOF measurements.32 Compared to MALDI-TOF platforms,33−35 FTICR covers a similar detection range, however with full isotopic resolution at much higher m/z-values. Previously, we have reported on MALDI-FTICR-MS methods for the analysis of peptides and small proteins in human serum up to approximately 15 kDa.36,37 Here, identifications of new proteoforms were initially based on accurate mass difference measurements and then confirmed by ESI-FTICR-MS/MS experiments. In the current study, we have aimed for the further development of new ultrahigh resolution MALDI-FTICR-MS methods and the application of these methods for the characterization of intact proteins by ISD experiments with fragment ion detection at isotopic resolution. To this end, both ribonuclease B (RNase B) and myoglobin were subjected to MALDI-ISD-FTICR-MS analysis, and the results were evaluated with regard to previously published data obtained by both ESIMS/MS and MALDI-TOF-MS.
■
EXPERIMENTAL SECTION Chemicals. Bovine pancreatic ribonuclease B (RNase B), 1,5diaminonaphthalene (1,5-DAN; handle with care, check material safety data sheet), myoglobin from equine skeletal muscle, αcyano-4-hydroxycinnamic acid, acetonitrile, and formic acid were purchased from Sigma-Aldrich (St. Louis, MO, USA). Acetone and ethanol were purchased from Merck (Darmstadt, Germany). RNase B and myoglobin were dissolved in Milli-Q water to a final concentration of 5 and 6 mg/mL, respectively, and then stored at −20 °C until further use. MALDI Spotting. For ISD-MS analysis, 1 μL of myoglobin (300 μg/mL in water) was spotted onto a MALDI ground steel target plate (Bruker Daltonics, Bremen, Germany) and mixed on the plate with 1 μL of 1,5-DAN (saturated solution in acetonitrile/water, 1:1) using a 10 μL pipet tip (Rainin tips from Mettler-Toledo, Tiel, Netherlands). Gentle mixing was performed until small crystals were visible by eye; then, the mixture was allowed to dry at room temperature. The same procedure was performed for RNase B (250 μg/mL in water). RNase B was also spotted onto a MALDI AnchorChip (600 μm; Bruker Daltonics, Bremen, Germany). To this end, 2 μL of RNase B (20 μg/mL in acetonitrile/water/formic acid, 50:49.95:0.05%) was transferred into a 1.5 mL Eppendorf tube
■
RESULTS AND DISCUSSION Intact Protein Analysis by MALDI-FTICR-MS. Measurement of a protein’s accurate mass and the determination of mass differences between various proteoforms can lead to the identification of possible PTMs.4,38 In this study, we employed a 15 T MALDI-FTICR system equipped with a dynamically harmonized ICR cell (ParaCell) to measure ultrahigh resolution spectra from small proteins. Isotopically resolved MALDIFTICR-MS spectra of intact ribonuclease B (RNase B) and myoglobin were obtained as is shown in Figure 1. RNase B consists of 124 amino acids, contains four disulfide bonds, and is glycosylated at asparagine(Asn)-34 with two β-1,4-linked Nacetylglucosamines and a varying number of mannose residues (i.e., from 5 to 9).39 Five RNase B glycoforms40 were profiled by MALDI-FTICR-MS as [M + H]+ species (Figure 1A), with an average resolving power of 61 000 (Figure 1C). The baseline resolution of the isotopic distributions allowed the accurate determination of the mass difference between consecutive glycoforms. The average mass difference of 162.10 mass units, observed between the most intense isotopic peak of consecutive 3430
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
intact proteins by MALDI-FTICR-MS needs further investigation and will be presented later. Top-Down Characterization of RNase B by MALDI-ISDFTICR-MS and Comparison with ESI-FTICR-MS/MS. Most commonly, when performing top-down MS/MS-characterization of proteins first, a precursor ion (e.g., using a quadrupole) is selected, either a specific charge state (in case of ESI) or a specific protein species (in case of a mixture). However, in MALDI-ISD experiments, the isolation of a precursor ion is not possible, since fragmentation occurs in the ionization source. Therefore, selection of a single RNase B glycoform for top-down MALDI-ISD-FTICR-MS analysis was not possible and ISD fragmentation was simultaneously performed on all five glycoforms. To this end, RNase B was spotted with 1,5-DAN matrix on a ground steel target plate and measured using two different MALDI-FTICR-MS methods (as described in the Experimental Section). The resulting MALDI-ISD spectra were compiled and are depicted in Figure 2. The most intense peaks detected in the mass range from m/z-value 1000 to m/z-value 4000 (upper panel) were identified as c-type ions while z-, a-, y-, and w-type fragments were detected with lower intensities. Furthermore, it should be noted that all z-type fragments contained an additional hydrogen atom, explicitly indicated with a prime (i.e., z′). In this mass range, the series of c-type fragments ends with c33 (highlighted in the green box). The smallest ISD fragment that allows the identification of the glycosylation site is c34. The calculated mass difference between the c33 and the glycosylated c34 fragment is the sum of the mass of one Asn residue (i.e., 114.0429 Da) and the mass of one of the five possible glycan structures. The smallest and most abundant glycan moiety of RNase B contains two N-acetylglucosamines and five mannoses (GlcNAc2Man5) corresponding to a theoretical mass difference between c33 and c34+GlcNAc2Man5 of 1330.47 Da. A difference of 1330.48 mass units was indeed measured between the c33 fragment (observed m/z-value 3681.6468) and fragment peak detected at m/z-value 5012.1233 (Figure 2 lower panel), thus confirming glycosylation at Asn-34. Interestingly, in the mass range between the c33 and c34+GlcNAc2Man5, z′-, y-, and w-type ions were detected as the most abundant peaks while in the mass range from m/z-value 5000 to m/z-value 6500 glycosylated c-type ions, containing either a GlcNAc2Man5 or GlcNAc2Man6 glycan moiety, were also detected (typical examples of isotopically resolved MALDI-ISD fragment ions of RNase B and their assignments are provided as Supporting Information, Figure S1). In addition, the presence of w-type ions provided information on the amino acid from which these originated,46 and this data was used to validate the identifications of z′-type ions. For example, the mass difference between the peak detected at m/z-value 3849.8493 (z′35) and the peak detected at m/z-value 3936.8857 (z′36) was 87.0364 Da (theoretical value of 87.0320 Da) corresponding to the mass of a serine residue (see Figure 3). This was further corroborated by the measurement of a mass difference of 17.9933 Da between the peak detected at m/z-value 3918.8924 (w36) and the peak at m/ z-value 3936.8857 (z′36). In fact, the theoretical mass difference between a z′- and a w-type ion generated from a serine is 18.0106 Da and corresponds to the sum of the mass of one hydroxyl group lost by α-cleavage after the formation of the z-type radical ion during ISD and one hydrogen.46 All together, MALDI-ISDMS of RNase B resulted in a sequence coverage of 64%. It should be noted that mass measurements were performed from m/zvalue 1012 to avoid the interference from matrix peaks and consequently both the N- and C-terminus were not charac-
Figure 1. Parts of the ultrahigh resolution MALDI-FTICR-MS spectra of RNase B and myoglobin. (A) Five glycoforms of RNase B were observed as singly charged forms in the mass range from m/z-value 14 800 to m/z-value 15 550. (B) Intact myoglobin was measured as the [M + H]+ form and as adduct with 1,5-DAN matrix in the mass range from m/z-value 16 900 to m/z-value 17 270. (C and D) Further enlargements of MALDI-FTICR-MS spectra depicting baseline resolved isotopic distribution of RNase B (i.e., GlcNAc2Man5 form) and myoglobin, respectively, with resolving powers of 61 000 and 62 000, respectively.
glycoforms, is in good agreement with the theoretical mass of a hexose (i.e., mannose) residue (i.e., 162.05 Da). Second, myoglobin, consisting of 153-amino acids,41 was measured by MALDI-FTICR-MS as [M + H]+ at an average resolving power of 62 000 (see Figure 1B,D). To our knowledge, this is the largest protein measured by MALDI-FTICR-MS with baseline isotopic resolution, albeit that such ultrahigh resolution measurements at high m/z-values required long acquisition times (due to the relatively low frequency of 13.56 kHz at m/z-value 16 967.23).42 Thus, RNase B and myoglobin were measured by MALDI-FTICR-MS with time-domain transient lengths of 7.98 and 9.23 s for a single scan, respectively. Furthermore, it should be noted that high resolving powers (e.g., 750 000) can be achieved for small molecules (e.g., m/z-value 600) with shorter time-domain transient length (e.g., 3.35 s) using the same FTICR-MS instrument (data not shown). A higher transient length implies a reduced sensitivity due to collisions of ions and residual gas molecules in the ICR cell. Obviously, at higher magnetic field (e.g., 21 T43), ultrahigh resolution measurements can be performed with shorter transients and may thus result in improved sensitivities. Similarly, measurement with longer timedomain transients and a consequent higher resolution may be performed with an improved ultrahigh vacuum in the ICR cell. Another way to efficiently improve the resolving power is postprocessing the acquired spectra by displaying these in absorption mode.44,45 For the current data, the resolving power of intact myoglobin could be increased up to 77 300 in an absorption-mode spectrum. The beneficial aspect of using an absorption-mode spectrum is further exemplified in identifying fragment ions (see Figure S4, Supporting Information; fragmentation process will be described hereafter). The routine application of absorption-mode spectra for the analysis of small 3431
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
Figure 2. Sequence coverage obtained from the analysis of RNase B by MALDI-ISD-FTICR-MS using 1,5-DAN as MALDI matrix (upper panel). (A and B) Parts of the ultrahigh resolution MALDI-ISD-FTICR-MS spectrum of RNase B glycoforms. The detection of both the c33 and the glycosylated c34 fragment ions (highlighted in the green boxes) allowed the localization of the most abundant glycan moiety (i.e., GlcNAc2Man5) at Asn-34 (indicated in bold in the protein sequence).
additional peptide bonds were sequenced, resulting in an extended sequence coverage of RNase B of 72%. In the MALDI-ISD spectrum of RNase B (Figure 2), it is furthermore noted that multiple fragment ions are observed in the protein part that contains eight cysteines all involved in disulfide bonds. These four different S−S bridges are reduced due to the hydrogen-donator property of the 1,5-DAN matrix. In a recent top-down study, disulfide bonds in intact RNase B were reduced and carbamidomethylated prior to ESI-FTICR-MS/MS analysis. In this study, a total sequence coverage of 86% was reported, based on a combination of different fragmentation
terized. This lack of coverage can be improved by performing socalled pseudo-MS3 experiments, in which N- and C-terminal regions of a protein can be characterized by MS/MS analysis performed on isolated fragment ions.28,47 With this approach, interfering matrix cluster peaks in the low mass range are removed and the sequence coverage obtained from MALDI-ISDMS experiments can be further extended. For the example of RNase B, the c12- and the w15-MALDI-ISD fragment ions were first isolated in the hexapole collision cell and then fragmented by CID (see the Experimental Section). The resulting pseudo-MS3 spectra are depicted in Figure 4. From these spectra, 15 3432
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
scan as Supporting Information, Figure S3). The identification of ISD fragment ions was based either on the detection of the monoisotopic peak or on the comparison between the theoretical and observed isotopic distribution, or both. Thus, although fragment ions were detected up to m/z-value 16 772, reliable identifications were obtained up to m/z-value 12 000. Similar to the MALDI-ISD data from RNase B, only c- and z′-type ions were annotated in the spectrum. These fragment ions were the most abundant and were measured with resolving powers ranging from 278 000 at m/z-value 1018.45 to 43 000 at m/zvalue 11 998.54. Intact myoglobin itself was measured as an [M + H]+ with a resolving power of 25 600. Note that measurements at higher resolution allowed the detection with baseline isotopic resolution of the intact myoglobin but were inadequate for the detection of ISD fragments due to the lower sensitivity. Previously, myoglobin has been characterized using a MALDIISD-TOF instrument at a moderate resolution resulting in a sequence coverage of 59%.46,50,51 Unfortunately, these experiments suffered from two limitations: (1) the signal-to-noise ratio (S/N) of the observed ISD fragments rapidly decreased with the increased m/z-range; (2) the fragment ions were detected as broad single peaks. Consequently, only high abundant fragment ions were detected and assigned. By using a state-of-the-art 15 T MALDI-FTICR system equipped with a new ParaCell for ICR measurement, the data dramatically improved. The sequence coverage based on c- and z′-type ions was 87%. The ultrahigh resolution MALDI-FTICR measurements allowed the identification of fragment ions that in low resolution MALDI-TOF would have not been detected as a single species. For example, the c32 and the z′33 fragment ions at theoretical m/z-values 3515.8394 and 3517.7420, respectively, were fully resolved with resolving power of 74 000 (see the Supporting Information, Figure S4). Despite such high resolving powers, not all fragment ions could be uniquely assigned, exemplified by a c57 and a z′58 at theoretical m/z-values 6459.3347 and 6462.4224, respectively (see the Supporting Information, Figure S4). These fragment ions were measured at a resolving power of 71 000 whereas theoretically more than 255 000 would be needed to baseline resolve the two isotopic distributions. Nevertheless, although the z′58 was not identified (because of the overlap with the c57), the presence of the K96 was confirmed by the identification of the y58 fragment ion (data not shown). The resolving power of MALDI-FTICR-MS spectra can be improved postacquisition by displaying the spectra in absorption mode instead of in magnitude mode.44,45 Upon performing this for the MALDI-ISD-FTICR-MS spectrum of myoglobin (i.e., obtained using Method 2; see Table S2, Supporting Information), on average, the resolving power increased by almost 3-fold (data not shown). For example, the resolving power at m/z-value 6462.35 (i.e., c57 fragment ion) increased from 70 557 in magnitude mode to 205 165 in absorption mode. This improvement allowed one to resolve the c57 fragment ion from the z′58 (see Figure S4, Supporting Information). It was observed that the signal-to-noise ratio in the absorption-mode spectrum was on average one-third lower (data not shown). It should however be noted that the confidence in identifying fragment ions in the absorption-mode spectra requires further investigation. Recently, Shaw and co-workers have reported on the ESI-MS/ MS characterization of myoglobin using different fragmentation techniques.7 It was shown that the use of ultraviolet photodissociation (UVPD) yields a drastically improved sequence coverage (higher than 93%) of myoglobin compared to CID,
Figure 3. Part of the ultrahigh resolution MALDI-ISD-FTICR-MS spectrum of RNase B. Note that different kinds of fragment ions (e.g., w36, z′36, y36) generated from cleavage of the same peptide bond improve the reliability of the assignments. The mass difference between w- and z-type ions generated from the same peptide bond may be diagnostic for the amino acid involved in the fragmentation. In fact, the mass difference of 17.9933 Da between the peak detected at m/z-value 3918.8924 (w36) and the peak at m/z-value 3936.8857 (z′36) indicates the loss of the hydroxyl group from the Ser-36 (i.e., 17.0027 (OH) + 1.0078 (H) = 18.0011 Da).
techniques (i.e., CID, IRMPD, ECD, ETD).48 Although our MALDI-ISD-FTICR-MS experiments resulted in a lower sequence coverage (i.e., 72% including the pseudo-MS3 data), some complementary information was obtained from the characterization of eight more peptide bonds (namely, Q28− M29, P42−V43, D83−C84, P93−N94, H105−I106, I106−I107, V108− A109, A109−C110). Note that the simultaneous fragmentation of the five different glycoforms during ISD experiments hampered the sequence coverage by increasing the complexity of the spectrum in the mass region higher than m/z-value 6500 (data not shown). Thus, the separation of the different glycoforms prior to ISD analysis may result in a more extensive characterization of RNase B. For comparative purposes, top-down ETD experiments were performed on nonreduced RNase B using the 15 T FTICR-MS system in ESI mode with precursor ion selection of the “15 plus” charge state of RNase B. From the obtained ETD spectrum, it became clear that the presence of disulfide bonds affected the analysis in the way that the fragmentation was limited to “S−S bond free regions” (see Supporting Information, Figure S2). In fact, due to the connection between the different cysteines,49 only fragment ions in the region close to both the N- and the Cterminus were observed. As a result, the sequence coverage was limited to 29% while from the previously reported ETD experiments on the reduced and carbamidomethylated RNase B it was 58%.48 Top-Down Characterization of Myoglobin by MALDIISD-FTICR-MS. As discussed above, both RNase B and myoglobin were measured with isotopic resolution by MALDIFTICR-MS, i.e., up to a mass range of m/z-value 17 000. In the MALDI-ISD analysis of RNase B, however, the interpretation of fragment ions at m/z-values larger than 6500 was hampered by overlapping signals due to the presence of different glycoforms. In the case of myoglobin, MALDI-ISD fragment ions could be assigned up to as high as m/z-value 12 000. To this end, myoglobin was spotted on a ground steel target plate with 1,5DAN and analyzed using two different MALDI-ISD-FTICR-MS methods (see Table S2, Supporting Information). The resulting two spectra were compiled and are depicted in Figure 5 (and full 3433
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
Figure 4. Pseudo-MS3 spectra of RNase B obtained from the analysis of isolated c12 and w15 MALDI-ISD fragment ions by CID and FTICR-MS.
HCD, and ETD, which provided a sequence coverage from a minimum of about 45% (CID) to a maximum of 71% (ETD). Unfortunately, details on the observed fragment ions in UVPD experiments were not reported; thus, the complementarity of ISD to this fragmentation technique could not be determined. One of the difficulties in ESI-MS/MS experiments on multiply protonated intact proteins is the generation of hundreds of product ions that lead to very complex spectra. For example, from the UVPD analysis of five different charge states of myoglobin, more than 1450 product ions were reported. The interpretation of such complex spectra is tedious. On the contrary, MALDI-ISD yields mainly singly charged ions thus
drastically reducing the complexity of the generated spectra. MALDI-ISD MS spectra can be acquired in a much shorter time frame (i.e., seconds) than ESI-MS/MS (i.e., minutes) making this technique more suitable for high-throughput analysis. A final remark is that ISD fragments of myoglobin were measured starting from m/z-value 1012, for similar reasons as discussed for RNase B. The optimization of the ration protein/ matrix and measurements in negative ion mode may lead to a further characterization of regions of the protein sequence near both the N- and the C-terminus.46 3434
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
Figure 5. Sequence coverage obtained from the analysis of myoglobin by MALDI-ISD-FTICR-MS using 1,5-DAN as MALDI matrix (upper panel). (A, B, C, and D) Parts of the ultrahigh resolution MALDI-ISD-FTICR-MS spectrum of myoglobin showing isotopically resolved fragment ions up to m/zvalue 12 000. The full MALDI-ISD-FTICR-MS spectrum of myoglobin is depicted in the Supporting Information, Figure S3.
■
CONCLUSIONS Structural characterization of intact proteins by the top-down ESI-MS/MS approach is commonly performed using different fragmentation techniques that provide complementary information with regard to the amino acid sequence. The combined data results in an improved sequence coverage; however, complete
protein sequence coverage is rarely obtained. In-source decay provides information that in combination with ESI-MS/MS readouts can further extend sequence coverage. So far, this information has been limited by the fact that MALDI-TOF-MS instruments provide suitable resolution and sensitivity limited to m/z-values 4000 up to 5000. In the current study, we have used 3435
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry
(12) Marshall, A. G.; Hendrickson, C. L. Annu. Rev. Anal. Chem. 2008, 1, 579−599. (13) Zubarev, R. A.; Makarov, A. Anal. Chem. 2013, 85, 5288−5296. (14) Fornelli, L.; Parra, J.; Hartmer, R.; Stoermer, C.; Lubeck, M.; Tsybin, Y. Anal. Bioanal. Chem. 2013, 405, 8505−8514. (15) Debois, D.; Smargiasso, N.; Demeure, K.; Asakawa, D.; Zimmerman, T.; Quinton, L.; De Pauw, E. In Applications of MALDITOF Spectroscopy; Cai, Z., Liu, S., Eds.; Springer: Berlin Heidelberg, 2013; pp 117−141. (16) Hardouin, J. Mass Spectrom. Rev. 2007, 26, 672−682. (17) Macht, M. TrAC, Trends Anal. Chem. 2013, 48, 62−71. (18) Hanisch, F. G. Anal. Chem. 2011, 83, 4829−4837. (19) Quinton, L.; Demeure, K.; Dobson, R.; Gilles, N.; Gabelica, V.; De Pauw, E. J. Proteome Res. 2007, 6, 3216−3223. (20) Demeure, K.; Gabelica, V.; De Pauw, E. J. Am. Soc. Mass Spectrom. 2010, 21, 1906−1917. (21) Takayama, M. J. Am. Soc. Mass Spectrom. 2001, 12, 1044−1049. (22) Köcher, T.; Engström, Å.; Zubarev, R. A. Anal. Chem. 2005, 77, 172−177. (23) Takayama, M. J. Am. Soc. Mass Spectrom. 2001, 12, 420−427. (24) Fukuyama, Y.; Iwamoto, S.; Tanaka, K. J. Mass Spectrom. 2006, 41, 191−201. (25) Smargiasso, N.; Quinton, L.; De Pauw, E. J. Am. Soc. Mass Spectrom. 2012, 23, 469−474. (26) Rand, K. D.; Bache, N.; Nedertoft, M. M.; Jørgensen, T. J. D. Anal. Chem. 2011, 83, 8859−8862. (27) Ait-Belkacem, R.; Berenguer, C.; Villard, C.; Ouafik, L.; FigarellaBranger, D.; Chinot, O.; Lafitte, D. Proteomics 2014, 14, 1290−1301. (28) Asakawa, D.; Smargiasso, N.; De Pauw, E. Anal. Chem. 2014, 86, 2451−2457. (29) Calligaris, D.; Villard, C.; Terras, L.; Braguer, D.; Verdier-Pinard, P.; Lafitte, D. Anal. Chem. 2010, 82, 6176−6184. (30) Ayoub, D.; Jabs, W.; Resemann, A.; Evers, W.; Evans, C.; Main, L.; Baessmann, C.; Wagner-Rousset, E.; Suckau, D.; Beck, A. mAbs 2013, 5, 699−710. (31) Takayama, M.; Nagoshi, K.; Iimuro, R.; Inatomi, K. Int. J. Mol. Sci. 2014, 15, 8428−8442. (32) Nikolaev, E. N.; Kostyukevich, Y. I.; Vladimirov, G. N. Mass Spectrom. Rev. 2014, DOI: 10.1002/mas.21422. (33) Nakanishi, T.; Okamoto, N.; Tanaka, K.; Shimizu, A. Biol. Mass Spectrom. 1994, 23, 230−233. (34) Solouki, T.; Gillig, K. J.; Russell, D. H. Anal. Chem. 1994, 66, 1583−1587. (35) Solouki, T.; Emmett, M. R.; Guan, S.; Marshall, A. G. Anal. Chem. 1997, 69, 1163−1168. (36) Nicolardi, S.; van der Burgt, Y. E.; Wuhrer, M.; Deelder, A. M. Proteomics 2013, 13, 992−1001. (37) Nicolardi, S.; van der Burgt, Y. E.; Dragan, I.; Hensbergen, P. J.; Deelder, A. M. J. Proteome Res. 2013, 12, 2260−2268. (38) Bondarenko, P.; Second, T.; Zabrouskov, V.; Makarov, A.; Zhang, Z. J. Am. Soc. Mass Spectrom. 2009, 20, 1415−1424. (39) Thaysen-Andersen, M.; Mysling, S.; Højrup, P. Anal. Chem. 2009, 81, 3933−3943. (40) Liang, C.-J.; Yamashita, K.; Kobata, A. J. Biochem. 1980, 88, 51− 58. (41) Dautrevaux, M.; Boulanger, Y.; Han, K.; Biserte, G. Eur. J. Biochem. 1969, 11, 267−277. (42) Amster, I. J. J. Mass Spectrom. 1996, 31, 1325−1337. (43) Kaiser, N.; Weisbrod, C.; Quinn, J.; Blakney, G.; Beu, S.; Chen, T.; Smith, D.; Hendrickson, C.; Marshall, A. Development of an FT-ICR mass spectrometer in preparation for 21 Tesla. 62nd American Society for Mass Spectrometry (ASMS) Conference on Mass Spectrometry and Allied Topics, Baltimore, Maryland, June 15−19, 2014; Oral Presentation. (44) Kilgour, D. P. A.; Wills, R.; Qi, Y.; O’Connor, P. B. Anal. Chem. 2013, 85, 3903−3911. (45) Kilgour, D. P. A.; Neal, M. J.; Soulby, A. J.; O’Connor, P. B. Rapid Commun. Mass Spectrom. 2013, 27, 1977−1982. (46) Asakawa, D.; Smargiasso, N.; De Pauw, E. J. Am. Soc. Mass Spectrom. 2013, 24, 297−300.
an ultrahigh resolution MALDI-FTICR system to perform high mass measurements of ISD fragment ions generated from myoglobin and RNase B proteoforms. Intact, singly charged species were measured with resolving powers of 62 000 (at m/zvalue 16 950) and (at m/z-value 14 900) 61 000 for myoglobin and RNase B, respectively. The high mass resolution and improved sensitivity for higher m/z-values of these MALDI-ISDFTICR measurements yielded a high sequence coverage for both proteins, 87% for myoglobin and 72% for RNase B. For RNase B, complementary sequence information to previously published ESI-MS/MS data was obtained and the use of pseudo-MS3 experiments (CID) on selected ISD fragment ions allowed the characterization of both the N- and C-terminal regions. In the case of myoglobin, MALDI-ISD fragment ions were identified up to m/z-value 12 000 (with resolving power of 39 000 at m/zvalue 12 000) resulting in a sequence coverage of 87%. Although an even higher sequence coverage (93%) was reported using a UVPD in a hybrid linear ion trap Orbitrap mass spectrometer, the advantage of using MALDI-ISD-FTICR-MS to characterize myoglobin is a shorter acquisition time and the generation of less complex spectra.
■
ASSOCIATED CONTENT
S Supporting Information *
Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS The authors thank Dr. David Kilgour for providing absorptionmode spectra of myoglobin using Autophaser. REFERENCES
(1) Siuti, N.; Kelleher, N. L. Nat. Methods 2007, 4, 817−821. (2) Tran, J. C.; Zamdborg, L.; Ahlf, D. R.; Lee, J. E.; Catherman, A. D.; Durbin, K. R.; Tipton, J. D.; Vellaichamy, A.; Kellie, J. F.; Li, M.; Wu, C.; Sweet, S. M. M.; Early, B. P.; Siuti, N.; LeDuc, R. D.; Compton, P. D.; Thomas, P. M.; Kelleher, N. L. Nature 2011, 480, 254−258. (3) Heck, A. J. R. Nat. Methods 2008, 5, 927−933. (4) Tipton, J. D.; Tran, J. C.; Catherman, A. D.; Ahlf, D. R.; Durbin, K. R.; Lee, J. E.; Kellie, J. F.; Kelleher, N. L.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2012, 84, 2111−2117. (5) Han, X.; Jin, M.; Breuker, K.; McLafferty, F. W. Science 2006, 314, 109−112. (6) Tsybin, Y. O.; Fornelli, L.; Stoermer, C.; Luebeck, M.; Parra, J.; Nallet, S.; Wurm, F. M.; Hartmer, R. Anal. Chem. 2011, 83, 8919−8927. (7) Shaw, J. B.; Li, W.; Holden, D. D.; Zhang, Y.; Griep-Raming, J.; Fellers, R. T.; Early, B. P.; Thomas, P. M.; Kelleher, N. L.; Brodbelt, J. S. J. Am. Chem. Soc. 2013, 135, 12646−12651. (8) Madsen, J. A.; Gardner, M. W.; Smith, S. I.; Ledvina, A. R.; Coon, J. J.; Schwartz, J. C.; Stafford, G. C.; Brodbelt, J. S. Anal. Chem. 2009, 81, 8677−8686. (9) Mao, Y.; Valeja, S. G.; Rouse, J. C.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2013, 85, 4239−4246. (10) Ahlf, D. R.; Compton, P. D.; Tran, J. C.; Early, B. P.; Thomas, P. M.; Kelleher, N. L. J. Proteome Res. 2012, 11, 4308−4314. (11) Mihalca, R.; van der Burgt, Y. E. M.; McDonnell, L. A.; Duursma, M.; Cerjak, I.; Heck, A. J. R.; Heeren, R. M. A. Rapid Commun. Mass Spectrom. 2006, 20, 1838−1844. 3436
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
Article
Analytical Chemistry (47) Suckau, D.; Resemann, A. Anal. Chem. 2003, 75, 5817−5824. (48) Bourgoin-Voillard, S.; Leymarie, N.; Costello, C. E. Proteomics 2014, 14, 1174−1184. (49) Smyth, D. G.; Stein, W. H.; Moore, S. J. Biol. Chem. 1963, 238, 227−234. (50) Takayama, M. J. Mass Spectrom. Soc. Jpn. 2002, 50, 304−310. (51) Takayama, M.; Osaka, I.; Sakakura, M. Mass Spectrom. (Tokyo) 2012, 1, A0001.
3437
DOI: 10.1021/ac504708y Anal. Chem. 2015, 87, 3429−3437
