In-Source Decay and Pseudo Tandem Mass Spectrometry

Technical Note pubs.acs.org/ac

In-Source Decay and Pseudo Tandem Mass Spectrometry Fragmentation Processes of Entire High Mass Proteins on a Hybrid Vacuum Matrix-Assisted Laser Desorption Ionization-Quadrupole Ion-Trap Time-of-Flight Mass Spectrometer Lyna Sellami,† Omar Belgacem,‡ Claude Villard,† Matthew E. Openshaw,‡ Pascale Barbier,† and Daniel Lafitte*,† †

Aix-Marseille University, UMR 911 INSERM CRO2, 13385 Marseille, France Shimadzu, Manchester M17 1GP, United Kingdom

‡

S Supporting Information *

ABSTRACT: In-source decay (ISD), although a process known for decades in mass spectrometry, has a renewed interest due to increased theoretical knowledge in fragmentation processes of large biomolecules coupled with technological improvements. We report here an original method consisting of isolating matrix-assisted laser desorption ionization (MALDI)-generated in-source fragments of large proteins and subsequently performing selective fragmentation experiments (up to four cycles) using a hybrid MALDI quadrupole ion-trap time-of-flight mass spectrometer (MALDI-QIT-TOF). This technology takes advantage of keeping high resolution on the selection of precursors and detection of fragments. It allows exhaustive N- and C-terminal sequencing of proteins. In this work, human serum albumin (HSA), β-casein, and recombinant Tau proteins were submitted to in source decay in the MALDI source. The fragments were stored in the ion-trap and submitted to sequential collision-induced dissociation (CID). Finally, ISD and pseudo MSn were performed on oxidized Tau protein and acetylated bovine serum albumin to identify amino acid modifications. This work highlights the potential of the MALDI-QIT-TOF instrument for pseudo MSn strategies and top down proteomics.

M

identify proteins14 and differentiate isoforms of high molecular weight proteins.15 This process is referred to as top-down since proteins are not digested prior to analysis. MALDI-ISD is a faster but still complementary approach to bottom-up assays where proteins are submitted to enzymatic treatment.7,16 The N- and C-termini of proteins, critical portions bearing many post-translational modifications (PTMs), are easily sequenced using ISD while the modifications can be lost due to incomplete coverage after enzymatic digestion in the bottom-up approach. Using this top-down approach, PTMs are preserved while conventional CID is a harsh process that provokes loss of labile PTMs.17−19 Combination of both strategies is often

atrix-assisted laser desorption ionization combined with in-source decay (MALDI-ISD) is a wide spreading approach for protein N- and C-terminal sequencing.1−3 It was first reported by Brown and co-workers in 1995 for sequencing the β chain of bovine insulin and cytochrome C.4 ISD is now used as an example for the characterization of the N- and C-termini of high molecular weight proteins such as antibodies.5 MALDI-ISD is a fragmentation process that occurs in the MALDI plume after desorption/ionization steps.6 Increasing the laser fluence 5−20% above analyte ionization threshold leads to intermolecular transfer of hydrogen atoms between excited matrix molecules and carbonyl oxygen on the peptide backbone.7,8 This fragmentation process occurs via two pathways as described.9 The radical pathway produces preferentially cn- and (zn + 2)-ion series, while the thermal pathway gives yn-, an-, xn-,10 and bn-ion series.9,11−13 This strategy has been used to © 2012 American Chemical Society

Received: January 9, 2012 Accepted: May 17, 2012 Published: May 17, 2012 5180

dx.doi.org/10.1021/ac300072b | Anal. Chem. 2012, 84, 5180−5185

Analytical Chemistry

Technical Note

Figure 1. ISD spectra of HSA and annotated protein sequence with experimental fragments. (A) MALDI-QIT-TOF spectrum of HSA analyzed with DHB matrix. The insert provides the isotopic resolution of the c33 ion. (B) MALDI-TOF spectrum of HSA, the insert is a zoom of the c33 ion. Error plots for both spectra are provided in the Supporting Information (Figure S-2).

Figure 2. Pseudo MSn on HSA protein and annotated sequence with experimental fragments. (A) Pseudo MS3 of the c22-ion (m/z 2541.16). (B) Pseudo MS4 of the b13-ion from the c22-ion of pseudo MS3(m/z 1521.73).

necessary to obtain an exhaustive coverage of the protein sequence.16 For protein identification following ISD, classical database searching using Mascot (http://www.matrixscience.com), for example, or dedicated ISD software in various instrument configurations is performed.1,20,21 However, simple TOF-MS after MALDI-ISD does not generate evidence of the very terminal sequence information. Fragments in the low mass/ charge range are also difficult to obtain due to matrix adducts and/or high background noise. Therefore, pseudo MS3 techniques have been designed to obtain a sequence from ISD fragments. The best example on MALDI tandem MS instruments is called T3 sequencing. Using this approach, Suckau and co-workers easily sequenced RNase B20 and even used de novo sequencing for a 13.6 KDa single heavy chain of an antibody.5 In our study, we combined MALDI-ISD with tandem mass spectrometry on a hybrid MALDI-quadrupole ion trap-time-of-flight (QIT-TOF) instrument to perform pseudo MSn analysis of proteins. We report in this study the top-down analysis of proteins with or without modification. The study of protein mixtures is also reported. Tandem mass spectra were obtained up to pseudo MS5 using the MALDI-QIT-TOF MS. Such an approach is a valuable tool for identification of proteins and efficient characterization of their modifications.

■

EXPERIMENTAL SECTION Protein. HSA, BSA, acetylated BSA, β casein, and cytochrome C were obtained from Sigma Aldrich, and Tau (human hTau40; NCBI Reference Sequence, NP_005901.2) was expressed in pET vector introduced into Escherichia coli BL21(DE3) and purified in house. Stock solutions of 45 μM of proteins were purified on a ZipTip C4 microcolumn from Millipore. Peptides. A peptide calibration mix from LaserBio Laboratories (Sophia-Antipolis, France) was used for the external calibration, containing bradykinin [1-5], angiotensin II, neurotensin, ACTH [18-39], and bovine insulin chain B. Matrixes. α-Cyano-4-hydroxycinnamic acid (CHCA) and 2,5-dihydroxybenzoic acid (DHB) were obtained from LaserBio Laboratories and used without further purification. Matrixes were dissolved in 70% acetonitrile/0.1% TFA (10 mg/mL CHCA and 60 mg/mL DHB). Sample Preparation. Protein stock solutions were purified using a ZipTip C4 microcolumn (ZipTip). The elution step of proteins from the column was performed with matrix solutions. In total, 2 μL of protein solution (approximately 3 μg) was 5181



Technical Note

Figure 3. ISD spectra of Tau protein analyzed with DHB matrix and annotated protein sequence with experimental fragments. (A) MALDI-QITTOF ISD spectrum of Tau. (B) MALDI TOF spectrum of Tau. (C) MALDI QIT-TOF ISD spectrum of oxidized Tau protein. (D) MALDI-TOF spectrum of oxidized Tau.“#”modified residue (+ 16 Da), (##) modified residue (+ 32 Da).

after being reflected on the “pyramid mirror”. The 3D stretched ion-trap acts as a high-voltage electrode and also allows optical access on the target plate by a charge coupled device camera. Two high-voltage Einzel lenses focus ions through the pyramid mirror in the QIT and extend the injected mass range by minimizing the arrival time distribution between ions of different m/z ratios. Ions injected in the QIT are decelerated by a static electric field (nominal 50 V applied to the far end-cap) and by collisions with the pulsed gas. The “rapid rf start-up” technology incorporated allows for switching the rf voltage at full amplitude (5 kV, 500 kHz) within a few tens of nanoseconds, greatly facilitating the trapping processes. Typically, ions are desorbed in the ion source at a pressure in the range of low 10−6 mbar (high vacuum MALDI). They are then transferred into the ion trap where they will undergo collisional cooling by applying a pressure transient environment inside the QIT. This is established by pulsing gas to access higher helium pressures (up to 1.10−2 mbar) required for fast thermalization, without affecting vacuum conditions in the ion source (low 10−6 mbar) and the time-of-flight part of the instrument (around 5 × 10−7 mbar). The back pressure of helium inside the trap is typically 2 × 10−5 mbar.

applied to either the MALDI or the QIT target and allowed to dry at ambient temperature. Protein Oxidation. For protein oxidation, Tau stock solution was mixed with 5 mM NaOCl solution (Across Organics, Belgium). The mixture was incubated for 15 min at 37 °C in a water bath. Following oxidation, the protein was purified using a ZipTip C4 microcolumn (Millipore) and eluted with the matrix. In total, 2 μL of protein solution (approximately 3 μg) was applied either to the MALDI-TOF or MALDI-QITTOF target and allowed to dry at ambient temperature. Mass Spectrometry. MALDI spectra were acquired on a MALDI-TOF and on a high vacuum MALDI-QIT-TOF instrument (Axima Performance and Axima Resonance respectively, Shimadzu, Manchester, U.K.). The Axima control software (Launchpad, version 2.9.2) and GPMAW software (lighthouse data, Denmark) were used for fragment annotation and analysis. Protein identification was performed using Mascot (http://www. matrixscience.com). The MALDI-QIT-TOF instrument is equipped with a nitrogen laser (337 nm wavelength, 3 ns pulse) as well as a 3D iontrap coupled to a gridless reflectron analyzer. The laser beam is directed almost orthogonally to the surface of the MALDI plate 5182



Technical Note

Pressure transients are experimentally determined by choosing a threshold where internally excited ions do not undergo dissociation. Finally, ions are extracted into the TOF analyzer by applying a potential of 10 kV. Figure S1 in the Supporting Information is a schematic of the instrument. A detailed description of the mass spectrometer can be found elsewhere.22 The MALDI-TOF instrument is a TOF/ReTOF instrument equipped with a nitrogen laser (337 nm wavelength, 3 ns pulse) and a collision cell allowing ions to collide with helium gas without any deceleration (20 keV kinetic energy). A double wire ion gate provides a precursor ion selection resolution of up to 400 (FWHM). A detailed description of the instrument can be found elsewhere.23 For MALDI-QIT-TOF experiments, ISD spectra were externally calibrated using peptide standards of LaserBio Laboratories. Peptides were prepared in the same DHB matrix as respective proteins. ISD spectra from intact proteins were acquired in positive ion mode. The low-mass ion deflector cutoff was set to 800 Da. The laser fluence was raised 10−30% beyond the threshold. A total of 1000 shots were accumulated for each spectrum. For MSn sequencing of the proteins, ISD fragment ions were further fragmented using the CID mode. The precursor ion selector was set to the standard resolution (1%), spectra were obtained after CID activation and by keeping the same laser fluence as used for ISD. A total of 1300 shots were accumulated for pseudo MS3 spectra and 1600 laser shots for pseudo MS4 spectra. The mass of HSA and Tau proteins were confirmed using MALDI-TOF MS in the linear positive ion mode. ISD spectra of proteins were acquired in the linear positive mode with an accelerating potential of 20 keV. The ion extraction delay was set to 80 ns.

■

RESULTS AND DISCUSSION To obtain exhaustive protein sequence coverage MALDI-ISD should be combined with efficient MSn strategies. The MALDIQIT-TOF technology was used to fragment various proteins. Ions were produced in the MALDI source by ISD then transferred to the ion-trap and submitted further to multiple cycles of MSn using CID. In Figure 1A, top-down spectrum of HSA was obtained using ISD with DHB as a matrix. Increasing laser fluence by approximately 30% produced extensive protein fragmentation. ISD fragments were measured between m/z 800 and 4000 in the positive mode. The cn- series ion (c9−c33) and bn-series ion (b8−b17) were obtained. As previously described, HSA protein fragments via the radical pathway, and therefore we should not obtain b- or y-ion fragments.6,20 The same analysis was performed on a MALDI-TOF instrument and no b-ions were observed (Figure 1B). The combination of MALDI-ISD with QIT enhanced b-ion intensity, and this is explained by CID like processes in the ion-trap. These b-ion fragments were also observed with the ion trap-orbitrap hybrid mass analyzer on bovine serum albumin.24 As explained by Papanastasiou et al., one of the main problems of the first versions of MALDI ion-trap instruments was uncontrollable fragmentation due to the transfer from the vacuum MALDI source to the ion-trap.22 This was largely solved by thermalization and stabilization of metastable MALDI ions in the trap under a pressure transient environment established by pulsing a gas prior to ion introduction. In our case, the presence of b-ions in the ion-trap could be explained by a longer time prior to detection. Indeed it could be considered as a c-ion minus 17 Da. The loss of neutrals in the ion-trap device is something very common (more time is given to the c-ions to fragment). This is not the case in standard TOF instrument where the experiment

Figure 4. Pseudo MSn performed on oxidized Tau. (A) Pseudo MS3 of the c16#-ion (m/z 1860.82); the insert on the left shows the methionine oxidation and the characteristic neutral loss of the sulfenic acid (− 64 Da) by CID, (B) Pseudo MS4 of the b12#-ion (m/z 1477.63). (C) Pseudo MS5 of the [y438#-b12] (m/z 1277.53). “#” modified residues.

time frame is in the hundreds of microseconds range. The time spent by an ion in the QIT is in the milliseconds range. Various HSA c-ions were submitted to MSn fragmentation (up to pseudo MS4) allowing the sequencing of the c22-ion. Figure 2A shows the pseudo MS3 spectrum of the c22-ion obtained by colission induced dissociation (CID). The precursor ion selector was set to the standard resolution (70). The ion-trap acts as both a precursor selector and collision cell during MS/MS analysis which is facilitated via CID and allowed the detection of MSn isotopically resolved peaks (resolution = 10 000 FWHM). Upon CID, two major peaks were measured at m/z 1406.70 and 1521.72 corresponding, respectively, to the b12 (after a lysine residue) and the b13 (after aspartic residue) ions and their related species with the loss of NH3 and H2O. Peaks were detected with a resolution around 10 000 (FWHM) for the b13-ion enabling easy reselection for another stage of MSn. Other low intensity peaks were resolved; b16, b17, and y21. 5183



Technical Note

Figure 5. Application of the approach on protein mixture (cytochrome C, BSA, β casein). (A) ISD spectrum of the protein mixture, (B) pseudo MS3 of m/z 1460.80, (C) pseudo MS4 of m/z 1361.87 D) Mascot score after combination of the pseudo MS3 and pseudo MS4 research.

Figure 2B shows the pseudo MS4 spectrum of the b13-ion obtained by CID. The bn-ion series ranged from b8 to b12 and yn-ion series y4-y7, y9-y11. Internal ion fragments and ion species −H2O and/or −NH3 were also detected. The resolution was around 8000 (FWHM) for the y10 fragment. Pseudo MS3 can also be performed using MALDI TOF/TOF instruments.20 Tandem TOF instruments require an additional TOF mass analyzer for each step of analysis, which limits the number of cycles for spatial considerations. Pseudo MS3 has also been observed on an intermediate-vacuum MALDI orbitrap with relatively good efficiency, accuracy, and resolving power but no pseudo MS4 has been shown so far,24 probably due to the signal-to-noise ratio of pseudo MS3 fragments. Although thermalized and stabilized under a pressure transient environment, we believe that the metastable character of ions on an ion-trap instrument coupled to a high-vacuum MALDI ion source explains the easy internal activation of ions and the acceleration of the kinetics of unimolecular dissociation. We then applied our strategy on recombinant Tau, a 42 kDa protein involved in Alzheimer’s disease.25 This protein, in vitro oxidized for other purposes, is a good model to show the interest of our approach for characterization of modified proteins. We first performed ISD on the unmodified Tau protein using DHB as a matrix (Figure 3A). Various ion series were detected over the mass range m/z 800 to 4000. N-terminus fragments were cn-series ions (c7−c36) and bn-series ions (b8−b30). The yn-series ions (y12−y34) were obtained from the C-terminus. To make sure that no proteolytic fragments interfered with our ISD data, the same samples were analyzed using CHCA at lower laser fluence. No peaks that could correspond to proteolytic fragments were detected in the low

mass range (see Figure S-3 in the Supporting Information). The same analysis was performed using a MALDI-TOF instrument and no b-ions were observed (Figure 3B). ISD preserves PTMs, therefore in the case of oxidized species, typically no loss of sulfenic acid is observed during conventional MALDI TOF ISD (Figure 3D). The result is the same using MALDI-QIT-TOF and labile modifications are preserved during analysis in the ion-trap (Figure 3C). From the N-terminus, c-ion series (c7, c8, and c9), c-ions with a mass shift of (+16 Da) corresponding to the chemical modification of the Met10 to its sulfoxide form (c10#, c11#, c12#, c15#, and c16#) and c-ions with a mass shift of (+32 Da) corresponding to the sulfone form of the same methione (c10##, c11##, c12##, c15##, and c16##) were detected. The bn-ion (b7, b8) completed the panel. C-terminus fragments were yn-ions (y17, y20−y22) and y-ions with a mass shift of (+16 Da) corresponding to the oxidation of Met418 (y23#, y26#, y27#, and y29# ion). The c16# ion was selected for pseudo MS3, and the chemical modification on Met10 was identified, as expected (Figure 4A). Fragment ions show the characteristic neutral loss of the sulfenic acid (−64 Da) from the methionine sulfoxide b11# − 64 Da and b12# − 64 Da. Pseudo MS4 was performed by selecting the b12# ion as the precursor (Figure 4B). We eventually reached the limit of our system with pseudo MS5 on the [y438#-b12] (Figure 4C). The same top down approach was performed on acetylated BSA (Figure S-4 in the Supporting Information). The pseudo MS4 allowed the selection and the characterization of the acetylation on a specific lysine (K12). Such a conclusion was not possible from MS3 alone. Finally, we performed ISD on a mixture of three proteins BSA, β casein, and cytochrome C (Figure 5A); one ISD fragment 5184



Technical Note

(m/z 1460.80) was subjected to pseudo MS3 (Figure 5B) and pseudo MS4 (Figure 5C), and only a Mascot search combining pseudo MS3 and pseudo MS4 data allowed the identification of the casein (Figure 5D).

(16) Pflieger, D.; Przybylski, C.; Gonnet, F; et al. Mol. Cell. Proteomics 2010, 9, 593−610. (17) Lennon, J. J.; Walsh, K. A. Protein Sci. 1999, 8, 2487−2493. (18) Kellie, J. F.; Tran, J. C.; Lee, J. E.; Ahlf, D. R.; Thomas, H. M.; Ntai, I.; et al. Mol. BioSyst. 2010, 6, 1532−1539. (19) McLafferty, F. W.; Breuker, K.; Jin, M.; Han, X.; Infusini, G.; Jiang, H. FEBS J. 2007, 274, 6256−68. (20) Suckau, D.; Resemann, A. Anal. Chem. 2003, 75, 5817−5824. (21) Zimmerman, T. A.; Debois, D.; Mazzucchelli, G.; Bertrand, V.; De Pauw, E. Anal. Chem. 2011, 83, 6090−6097. (22) Papanastasiou, D.: Belgacem, O.; Montgomery, H.; Sudakov, M. Raptakis, E. A. Practical Aspects of Trapped Ion Mass Spectrometry, Vol. IV; CRC Press: Boca Raton, FL, 2010; Chapter 19, pp 791−822. (23) Belgacem, O.; Bowdler, A.; Brookhouse, I.; Brancia, F. L.; Raptakis, E. Rapid Commun. Mass Spectrom. 2006, 20, 1653−1660. (24) Scheffler, K.; Strupat, K. Thermo Scientific Application Note, 2010; 30218 (25) Vermersch, P.; Frigard, B.; Delacourte, A. Acta Neuropathol. 1992, 85, 48−54.

■

CONCLUSIONS MALDI-ISD combined with QIT-TOF MS analysis is a real improvement for extensive characterization of biomolecule sequences. This hybrid MALDI-QIT-TOF instrument provides good resolution for daughter ions and a high degree of fragmentation. Combination of these two aspects allows new MSn steps with effective selection of precursor ions and fragmentation up to pseudo MS5. This work open new ways for precise characterization of protein modification when multiple sites are concerned or when the modification type is complex. This can be crucial, for example, for glycosylated protein characterization. This approach is also extremely useful for protein identification of species in protein mixtures combining pseudo MS3 and MS4 data, for example. In source decay on tissue often generates very few fragments, and subsequent MSn is necessary for protein identification.

■

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: daniel.lafi[email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by Aix Marseille Université, Kratos Analytical, l’Association de Recherche Contre le Cancer (ARC), Fondation Sport Santé et Développement Durable, Cancéropole PACA

■

REFERENCES

(1) Hardouin, J. Mass Spectrom Rev. 2007, 26, 672−82. (2) Macht, M. Bioanalysis 2009, 1, 1131−48. (3) Calligaris, D.; Villard, C.; Lafitte, D. J. Proteomics 2011, 10, 920− 34. (4) Brown, R. S.; Lennon, J. J. Anal. Chem. 1995, 67, 3990−3999. (5) Resemann, A.; Wunderlich, D.; Rothbauer, U.; Warscheid, B.; Leonhardt, H; et al. Anal. Chem. 2010, 15 (82), 3283−92. (6) Demeure, K.; Quinton, L.; Gabelica, V.; De Pauw, E. Anal. Chem. 2007, 79, 8678−85. (7) Takayama, M. J. Am. Soc. Mass Spectrom. 2001, 12, 1044−1049. (8) Brown, R. S.; Feng, J.; Reiber, D. C. Int. J. Mass Spectrom. Ion Processes 1997, 169/170, 1−18. (9) Demeure, K.; Gabelica, V.; De Pauw, E. A. J. Am. Soc. Mass Spectrom. 2010, 21, 1906−1917. (10) Asakawa, D.; Takayama, M. J. Am. Soc. Mass Spectrom. 2011, 1224−1233. (11) Takayama, M. J. Am. Soc. Mass Spectrom. 2001, 12, 420−427. (12) Takayama, M. J. Mass Spectrom. Soc. Jpn. 2002, 50, 304−310. (13) Köcher, T.; Engström, A.; Zubarev, R. A. Anal. Chem. 2005, 77, 172−177. (14) Suckau, D.; Resemann, A. J. Biomol. Tech. 2009, 20, 258−262. (15) Calligaris, D.; Villard, C.; Terras, L.; Braguer, D.; VerdierPinard, P.; Lafitte, D. Anal Chem. 2010, 82, 6176−6184. 5185


In-Source Decay and Pseudo Tandem Mass Spectrometry

Recommend Documents