MALDI In-Source Decay of Protein

The Mechanism of c-Ion Formation

Abstract

The in-source decay (ISD) phenomenon, the fragmentation at an N–C_α bond of a peptide backbone that occurs within several tens of nanoseconds in the ion-source in matrix-assisted laser desorption/ionization mass spectrometry (MALDI MS), is discussed from the standpoints of the discovery and early publications dealing with MALDI-ISD, the formation of c-ions in energy-sudden desorption/ionization methods, the formation of radical species in a MALDI, model construction for ISD, and matrix materials that are suitable for use in MALDI-ISD. The formation of c-ions derived from peptides and proteins in MALDI-ISD can be rationalized by a mechanism involving intermolecular hydrogen transfer, denoted as the “Takayama’s model” by De Pauw’s group (Anal. Chem. 79: 8678–8685, 2007). It should be emphasized that the model for MALDI-ISD was constructed on the basis of X-ray crystallography and scanning probe microscopy (SPM) analyses of matrix crystals, as well as the use of isotopically-labelled peptides.

INTRODUCTION

Mass spectrometry (MS) is a scientific field that studies gas-phase ions of atoms, molecules and complex clusters. Clarifying the mechanism of ion formation and the fragmentation of gas-phase ions is major basic research interest in the field of MS, as well as having wide applicability to the material, biological, environmental, food, drug and omics sciences. Ionization and fragmentation in the ion-source continues to be one of the most interest events in gas-phase ion chemistry. The mechanism responsible for soft ionization in matrix-assisted laser desorption/ionization (MALDI)^1–3) and electrospray ionization (ESI)^4,5) remains unknown because of its complexity due to several factors such as the form of the matrix crystal being used, laser ablation within a very short timescale and photochemical and radical reactions in MALDI, and solvation between analyte and solvent molecules, ionic separation under a high electric field and droplet explosion in ESI. MALDI events are particularly complex and difficult to understand due to the numerous sequential ion–molecule and radical reactions that can occur in both the solid-phase and dense-gaseous MALDI plumes that are electronically excited with laser photons.

A feature of MALDI MS is radical-induced fragmentation (RIF), referred to as the “in-source decay (ISD)” of peptides and proteins.⁶⁾ MALDI-ISD involves a prompt cleavage that occurs within several tens of nanoseconds (ns) in the MALDI ion-source, and ISD is caused by the radical-sites of neutral species independent of the charge-sites.^7,8) Such radical species are generated by the transfer of hydrogen atoms from the matrix to the protein molecule.^9,10) In general, the radical sites of organic chemicals are react easily with themselves or with other chemicals.

Radical-induced unimolecular decay can be normally observed in analyte molecular ions A^+· generated by the electron ionization (EI) of organic compounds. The analyte ions A^+· degrade within 10⁻⁶ s in the ion-source, owing to the activity of radical sites. RIF plays major role in EI fragmentation such as simple (α)-cleavage. The high reactivity of radical-sites results in intramolecular rearrangements such as the McLafferty hydrogen rearrangement¹¹⁾ and the methylene chain rearrangement^12,13) of fatty acid methyl esters. In those reactions, the radical-site located on the carbonyl oxygen can strongly initiate hydrogen rearrangements from γ-position methylene hydrogens to the oxygen via the formation of a six-membered ring transition state (Scheme 1(a)) and methylene chain rearrangement from the methylene chain to the carbonyl oxygen via backbone ring transition states (Scheme 1(b)). It is very important to recognize that the formation of fragment ions with these rearrangements is complete within microseconds (10⁻⁶ s) or shorter in the EI ion-source and those ring transition states can be attained within a short timescale (10⁻¹⁴–10⁻¹³s) corresponding to molecular vibrational and rotational motions. Furthermore, intramolecular hydrogen and methylene chain rearrangements (Scheme 1) can take place without atomic motion through electron transfer on a very short timescale (10⁻¹⁶ s) corresponding to vertical electronic excitation by means of the Frank–Condon principle. The timescales described above have important meanings in the case of ultraviolet (UV)-MALDI with photochemical processes, as are described in this review article.

Scheme 1. Formation of ring transition states (10⁻¹⁴–10⁻¹³ s) and rearrangements by electron transfer (10⁻¹⁶ s) as a Frank–Condon process (left), radical-induced cleavage (the middle), and the formation of fragment ions at m/z 74 and 255 (right), occurring within 10⁻⁶ or shorter in the EI ion-source.^11–13)

The hydrogen atom (H^·) is the smallest radical species that can participate in intramolecular rearrangements and scrambling in the mass spectrometry of organic molecules. The MALDI-ISD of proteins can start by intermolecular hydrogen transfer from photochemically excited matrix molecules (M*) to protein analytes (A) via hydrogen-bonding between matrix active hydrogens (–OH and –NH₂) and backbone carbonyl oxygen in/on the matrix crystal (1),⁹⁾ and the following radical-induced fragmentation (RIF) (2), ionization (3a) and radical combination/ionization (3b) occur independently of each other.^7–10,14)   

(1)

  

(2)

  

(3a)

  

(3b)

The neutral fragments c and z′ (=z^·+H^·) are ionized by proton transfer from the protonated matrix [M+H]⁺ and to the deprotonated matrix [M−H]⁻.^15–17) The model of intermolecular hydrogen transfer (1) and the following RIF α-cleavage (2) in MALDI-ISD was referred to as “Takayama’s model” by De Pauw’s group,¹⁸⁾ and the model can explain the formation of ISD fragments and provides a powerful strategy for identifying and developing matrix materials that are suitable for use in MALDI-ISD.¹⁸⁾

In this review article, I discuss the discovery of ISD phenomenon in MALDI, radical formation in MALDI, the history of the observation of c-ions, the origin of hydrogen atoms to generate ISD fragments c and z′, the model construction for ISD, and matrix materials suitable for MALDI-ISD. The nomenclature for the neutral fragments generated from MALDI-ISD with 2,5-DHB is shown in Scheme 2.^9,10,14) Fragment ions observed in MALDI-ISD spectra can be understood by the protonation and deprotonation of neutral fragments.

Scheme 2. Nomenclature for neutral fragments generated by MALDI-ISD with 2,5-DHB matrix.^9,10,14)

DISCOVERY OF MALDI-ISD PHENOMENON AND EARLY RESEARCH PUBLICATIONS

The phenomenon of the ISD of peptides and proteins in MALDI MS was first reported by Brown and Lennon as “fast metastable ion fragmentation,”⁶⁾ and they discovered this phenomenon in the course of the development of a time-of-flight mass spectrometer with a delayed extraction device.¹⁹⁾ They reported the observation of unusual fragment ions c and z+2 assigned by Biemann’s nomenclature.²⁰⁾ Subsequently, several groups reported on the MALDI-ISD data of peptides and proteins from the standpoints of practical,^8,21–30) basic^{7–10,31–33)} and review interests,³⁴⁾ as summarized in Table 1. Recent advances in the fragmentation and theoretical considerations of MALDI-ISD have been reviewed by Asakawa³⁵⁾ and Moon et al.³⁶⁾

Table 1. Early reports of MALDI-ISD (1995–2007).

Author and Journal	Ref.
1995: Brown, Lennon, Anal. Chem. 67: 3990–3999.	6
1996: Brown, Carr, Lennon, J. Am. Soc. Mass Spectrom. 7: 225–232.	31
1997: Brown, Feng, Reiber, Int. J. Mass Spectrom. Ion Processes 169/170: 1–18.	32
1998: Takayama, Tsugita, Int. J. Mass Spectrom. 181: L1–L6.	7
1998: Reiber, Grover, Brown, Anal. Chem. 70: 673–683.	21
1998: Reiber, Brown, Weinberger, Kenny, Bailey, Anal. Chem. 70: 1214–1222.	22
1998: Katta, Chow, Rhode, Anal. Chem. 70: 4410–4416.	23
1999: Lennon, Walsh, Protein Sci. 8: 2487–2493.	24
2000: Takayama, Tsugita, Electrophoresis 21: 1670–1677.	8
2000: Kinumi, Niwa, Matsumoto, Anal. Biochem. 277: 177–186.	25
2001: Takayama, J. Am. Soc. Mass Spectrom. 12: 420–427.	33
2001: Takayama, J. Am. Soc. Mass Spectrom. 12: 1044–1049.	9
2002: Takayama, J. Mass Spectrom. Soc. Jpn. 50: 304–310.	26
2002: Gao, Tsugita, Takayama, Xu, Anal. Chem. 74: 1449–1457.	27
2002: Schnaible et al., Anal. Chem. 74: 4980–4988.	29
2003: Suckau, Resemann, Anal. Chem. 75: 5817–5824.	28
2005: Koecher, Engstrom, Zubarev, Anal. Chem. 77: 172–177.	10
2007: Demeure, Quinton, Gabelica, De Pauw, Anal. Chem. 79: 8678–8685.	18
2007: Quinton et al., J. Proteome Res. 6: 3216–3223.	30
2007: Hardouin, Mass Spectrom. Rev. 26: 672–682.	34

We also found fragment ions c and z+2 reflecting the amino acid sequence in a MALDI mass spectrum of angiotensin I independent of Brown and Lennon’s report,⁶⁾ in the course of experiments directed at the influence of laser fluence on the appearance and peak resolution of an analyte signal [A+H]⁺ in the late 1997s. Regarding fragment ions, we understood that carboxyl (C)-terminal successive truncation occurred by unknown chemical processes in the MALDI ion-source, but immediately after the discovery of sequence-reflected fragment ions we concluded that the observed fragments were generated independent of ionization (protonation and deprotonation) under MALDI conditions,⁷⁾ based on data from both positive- and negative-ion MALDI-ISD experiments of an amino (N)-terminal side acidic peptide porcine pancreastatin33–49 (Pyr-EEEEETAGAPQGLFRG-NH₂). Positive- and negative-ion MALDI-ISD spectra of the peptide merely showed C-terminal side z_n/y_n-ions (n=7–16) and N-terminal side c_n-ions (n=7–16), respectively. This suggested that c- and z′-fragments are generated via specific cleavage at the N–C_α bond of the backbone by the laser-induced excitation of neutral peptides (Scheme 2), while protonation and deprotonation of these fragments took place at the C-terminal side Arg residue and the N-terminal side Glu cluster, respectively.^7,8) That is, the neutral fragments c and z′ as the precursor of c- and z′-ions may be generated directly from neutral peptides, and the neutral fragments are ionized according to the acidic and basic characteristics of the constituent amino acid residues.⁸⁾ Interestingly, c- and z′-ions were characteristically observed when 2,5-dihydroxybenzoic acid (2,5-DHB) was used as the matrix, while the use of an α-cyano-4-hydroxycinnamic acid (CHCA) matrix resulted in the formation of a-, b- and y-ions which are normally observed in collision-induced dissociation (CID) spectra of protonated peptides.²⁰⁾ The c- and z′-ions were rarely observed in energy-sudden desorption/ionization (ESDI) mass spectrometry.

OBSERVATION OF c-IONS IN ENERGY-SUDDEN DESORPTION/IONIZATION METHODS

Unusual fragments such as c- and z+2 ions originate from the cleavage at the N–C_α bond of a peptide backbone that was first reported by Williams et al. in fast-atom bombardment (FAB) with a glycerol matrix and keV order energy of argon atoms.³⁷⁾ They reported the formation of sequence-reflected fragments such as c-, y- and z+2 ions, corresponding to Biemann’s nomenclature, via a hydrogen rearrangement mechanism (Scheme 3). Interestingly, Kenny et al. reported that the liquid secondary ion mass spectrum (LSIMS) of a peptide (SALSSSGEE) with a glycerol matrix gave predominantly c-ions originating from N–C_α bond cleavage, while the CID of the [M+H]⁺ gave merely b- and y-ions originating from peptide bond cleavage.³⁸⁾ Biemann’s group also reported that FAB mass spectra of peptides with a glycerol matrix gave a- and c-ions, while c-ions were never observed in the case of CID spectra,^20,39) and they explained c-ion formation by intramolecular hydrogen rearrangement from side-chain β-carbon hydrogens (Scheme 6(a)). Biemann also reported that c- and a-ions were observed in a FAB mass spectrum of a peptide ACTH18–39, although the spectrum was lacking in c₆ and c₁₈ ions owing to the presence of proline (Pro) residues, i.e., Pro7 and Pro19.²⁰⁾ Mahoney et al. reported that c-ions can be observed in negative-ion FAB mass spectra of peptides with a glycerol matrix.⁴⁰⁾ Furthermore, the observation of c-ions was reported by Vorst et al. by using plasma desorption mass spectrometry (PD MS) with nitrocellulose matrices.⁴⁰⁾ FAB, LSIMS and PD can be referred to as ESDI, a type of high-translational energy particle impinging method, and ESDI results in locally electronic excitation due to keV energy collision on the target molecules. Interestingly, the observation of c-ions in positive- and negative-ion FAB mass spectra of peptides suggests that FAB-ISD occurs independent of ionization, while a detailed mechanism was not presented by any of the above authors.

Scheme 3. Fragments originated from the N–C_α bond cleavage presented by Williams et al.³⁷⁾

I also conducted FAB experiments of angiotensin I in an attempt to observe c-ions using a glycerol matrix. The resulting FAB mass spectrum gave peaks corresponding to a- and c-ions, as shown in Fig. 1 (unpublished data). Although the detailed mechanism responsible for this is not clear, the possibility that c-ions are formed by the N–C_α bond cleavage of peptide backbone via hydrogen transfer from matrix active hydrogens (hydroxyl groups of glycerol) to the carbonyl oxygens on the peptide backbone cannot be excluded. It has been reported that FAB with keV order particle beams brings about electronic excitations corresponding to electron impact energies of 30 eV or higher.⁴²⁾ Such a high-energy characteristic of FAB may result in the production of a- and c-ions in the FAB ion source.

Fig. 1. Positive-ion FAB-ISD spectrum of angiotensin I (DRVYIHPFHL, M_r 1296.5) with glycerol matrix, obtained with an EB type mass spectrometer JEOL JMS-AX505W (JEOL, Tokyo, Japan).

According to Biemann’s nomenclature (Scheme 4), c-, y- and z+2 ions can be formed by intramolecular hydrogen rearrangements that occur during the CID experiments of peptide. This type of origin of hydrogen atoms is similar to the electron-capture dissociation (ECD) of multiply protonated proteins [A+nH]ⁿ⁺ leading to the formation of c- and z-ions by intramolecular hydrogen rearrangements via multiply-charged radical ions [A+nH]^(n−1)+· generated by electron-capture,⁴³⁾ although the production of hydrogen atoms essentially differs from CID experiments. It would be of interest to clarify the formation of hydrogen atoms and to determine the origin of hydrogens under ESDI conditions.

Scheme 4. Biemann’s nomenclature for peptide fragments.²⁰⁾

HYDROGEN ATOM AND RADICAL SPECIES FORMATION IN MALDI WITH 2,5-DHB

Scott et al. reported evidence that high levels of hydrogen atoms are generated from 2,5-DHB matrix when irradiated with nitrogen laser photons (337 nm) by performing the post-ionization (with a Nd : YAG laser) of neutral components in the MALDI plume ablated from a 2,5-DHB crystal surface.⁴⁴⁾ The experiments provided direct evidence for the presence of large amounts of hydrogen atoms in the MALDI plume with larger velocities (5,000 m/s) than the matrix species, as illustrated in Scheme 5. The formation of radical species such as hydrogen atoms (H^·) and 2,5-DHB radicals (M^·) has been also reported by the Hillenkamp group^45,46) and Bourcier et al.⁴⁷⁾ in MALDI experiments with 2,5-DHB. When 2,5-DHB is used as a matrix, the MALDI plume generated with ablation is made up of large amounts of neutral species such as matrix molecules M, hydrogen atoms H^· and DHB radicals M^·, as well as small amounts of matrix ions [M+H]⁺ and [M−H]⁻ with the ratio of ions/neutrals in the range of 10⁻⁹ to 10⁻⁸.⁴⁸⁾

Scheme 5. MALDI plume constituting of abundant matrix neutrals M, matrix radicals M^·, hydrogen atoms H^· and a few ionic species [M+H]⁺ and [M−H]⁻.

Zubarev’s group pointed out that a combination of reactions of z^· radicals with radical species (H^· and M^·) occur in the MALDI plume.¹⁰⁾ They reported that a C-terminal side peptide radical z^· binds to both a hydrogen atom H^· and a 2,5-DHB radical M^· to form z-related species z′ (=z^·+H^·) and a z-matrix (=z^·+M^·) in the MALDI plume. A similar mechanism regarding the production of z′ and z-matrix (Scheme 2) was reported by De Pauw’s group with 1,5-diaminonaphthalene (1,5-DAN) matrix.⁴⁹⁾ It is important to recognize that a series of combination reactions of radical species can be completed within several tens or hundreds of nanoseconds (corresponding to delayed extraction time) in the ion source, due to the great amounts of radical and neutral species in the MALDI plume.

MODEL CONSTRUCTION FOR MALDI-ISD OF PROTEIN

a) The origin of hydrogen atoms in forming c-ions

I attempted to construct a model for MALDI-ISD by using a model peptide RLG (RLGNQWAVGDLAE, M_r 1428.6) and its corresponding deuterium-labeled peptide RLG(Ad3, Gd2) (RLGNQWA(d3)VG(d2)DLAE, M_r 1433.6).⁹⁾ It is most important to clarify the origin of hydrogen atoms to generate c-ions (Scheme 4) for a better understanding of the mechanism of the MALDI-ISD phenomenon. Although the formation of c- and z-ions in ECD can be explained by intramolecular hydrogen rearrangement,⁴³⁾ in MALDI-ISD there are at least three possible sources of hydrogen atoms from either protein ions or 2,5-DHB matrix as follows:

1. 1. Intramolecular rearrangement of active or non-active hydrogen atoms,
2. 2. Intramolecular rearrangement of hydrogens formed by electron capture in multiply protonated peptides under MALDI conditions,
3. 3. Intermolecular transfer of active hydrogens from 2,5-DHB matrix molecules.

However, it is difficult to determine the origin of hydrogen atoms in hydrogen rearrangements, because active hydrogens can migrate from remote sites and some scrambling occurs. In this study, the ISD experiments were limited to three possible hydrogen sources to explain the c-ion formation:

• 1-1. Hydrogen(s) of the alpha-carbon (C_α) on the backbone (Scheme 6a),
• 1-2. Hydrogen(s) of the side-chain beta-carbon (C_β) (Scheme 6a),
• 3′. Active hydrogens (Ph–COOH and Ph–OH) of 2,5-DHB molecules (Scheme 6b).

Scheme 6. Three possible hydrogen sources, (a) α- and β-carbons of peptide and (b) matrix active hydrogens, for generating c-ions in MALDI-ISD. (c) Another possible hydrogen source originating from electron capture in multiply-protonated peptides under MALDI conditions.

In order to confirm the possibility of intramolecular hydrogen rearrangements 1-1 and 1-2, the deuterium labeled peptide RLG(Ad3, Gd2) described above was used.

Figure 2 shows MALDI-ISD spectra of the model peptides RLG and RLG (Ad3, Gd2). If the intramolecular hydrogen rearrangements occur as shown in Scheme 6(a), the m/z value of the c₆ ion originating from the deuterated model RLG (Ad3, Gd2) shifts to 1 Da larger than that of the c₆ ion at m/z 773 of the model peptide RLG, by intramolecular deuterium rearrangement from the C_β methyl group of A(d3). Similarly, the c₈ ion of RLG (Ad3, Gd2) will shift to 4 Da larger than that of c₈ ion at m/z 943 of RLG, as the result of the deuterium rearrangement from the C_α methylene of G(d2). However, no shifts in m/z values were observed in the ISD spectra, as shown in Fig. 2. Here it can be concluded that the hydrogen atom involved in the formation of c-ions did not originate at least from any hydrogens located on alpha- and beta-carbon positions. So that, it is most likely that the origin of hydrogen atoms in generating c-ions are from the 2,5-DHB matrix active hydrogens generated via a photochemically induced mechanism,⁹⁾ as shown in Scheme 6(b).

Fig. 2. MALDI-ISD spectra of (a) a model peptide RLGNQWAVGDLAE (M_r 1428.6) and its deuterium-labeled peptide RLGNQWA(d3)VG(d2)DLAE (M_r 1433.6) obtained with 5,1-ANL matrix. Asterisk indicates metastable peak.

According to the scenario regarding the origin of hydrogen atoms described above, MALDI-ISD experiments were performed for the deuterium labeled peptide RLG(A-d3, G-d2) deuterated with D₂O, using a deuterium labeled matrix 2,5-DHB-d3 and deuterated solvents D₂O and CH₃COOD.⁹⁾ The results obtained suggest that both a deuterium atom D^· and a deuteron D⁺ to generate c₃ (RLG-d11, m/z 355) and c₄ (RLGN-d14, m/z 472) ions were originated from the matrix 2,5-DHB-d3. Zubarev’s group also reported that an ISD hydrogen transfer is not mediated by the electron capture of multiply-charged peptides (Scheme 6(c)), but the origin of the hydrogen atom involved in generating c-ions in MALDI-ISD is from the matrix by the mechanism shown in Scheme 6(b).¹⁰⁾ Consequently, it can be concluded that the origin of hydrogen atoms involved in forming c-ions in the MALDI-ISD of peptides and proteins are from the 2,5-DHB matrix.

b) X-ray and scanning probe microscopy (SPM) analysis of 2,5-DHB crystal with and without protein incorporation

It is important to note that only a couple of matrix materials produce analytes and ISD ions of peptides and proteins, and that 2,5-DHB is unique in its ability to form abundant ISD ions.³²⁾ In order to construct a model for MALDI-ISD and to clarify the ISD phenomenon, it is necessary to understand the nature of the interactions between 2,5-DHB and protein molecules, based on appropriate information of the 2,5-DHB crystal (Fig. 3). Regarding this, Strupat et al. reported an X-ray crystallographic investigation of 2,5-DHB with and without the incorporation of protein into the matrix crystal.⁵⁰⁾ They presented a concrete image for the crystal growth of 2,5-DHB. That is, the 2,5-DHB matrix without protein incorporation forms a hydrophobic planar network as the hydrophobic crystal plane (011) by intermolecular hydrogen-bonding, as shown in the illustration of Scheme 7, and furthermore, the stacking of the (011) crystal planes forms a (100) surface along the direction of crystal growth. The crystal surface plane (100), which is hydrophilic in nature due to phenolic and carboxylic hydroxyl groups, is normally irradiated by UV laser photons in MALDI experiments. This is of importance from the standpoint of the interaction between matrix and protein molecules that an abundance of phenolic hydroxyl groups (Ph–OH) are exposed from the 2,5-DHB matrix (100) crystal surface. Strupat et al. observed an important fact that the 2,5-DHB crystal structure did not change, even after protein incorporation, while the (100) crystal surface became smoother than that without protein incorporation.⁵⁰⁾

Fig. 3. Microscope images of 2,5-DHB matrix crystal without any additives obtained with a microscope BX51 (Olympus, Tokyo, Japan).

Scheme 7. Schematic illustration of the 2,5-DHB matrix hydrophobic crystal plane (011) formed by an intermolecular hydrogen bonding network and hydrophobic interactions.⁵⁰⁾

On the other hand, the structures of the crystal (100) surface plane of the 2,5-DHB matrix with and without peptide incorporation for ACTH18–39 were observed by using scanning probe microscopy (SPM).⁵¹⁾ In the case of the crystal surface without a peptide, right-angled layered steps along the crystal growth direction were observed (Fig. 4(a)), while the layered steps disappeared and the (100) crystal surface formed a smooth structure after peptide incorporation (Fig. 4(b)). From X-ray crystallographic data⁵⁰⁾ and SPM data⁵¹⁾ of the 2,5-DHB crystals with and without analyte incorporation, it appears that the (100) crystal surface is covered with the analyte peptide or protein molecules. Therefore, intermolecular hydrogen-bonding between matrix active-hydrogens (Ph–OH and/or Ph–COOH) exposed from the (100) surface and carbonyl oxygens on the peptide backbone,^9,51) would be more likely, as illustrated in Scheme 8.

Fig. 4. Scanning probe microscopic images of the crystal surface of 2,5-DHB matrix (a) without and (b) with analyte ACTH18–39 molecules, obtained with a SPM-9500J2 (Shimadzu, Kyoto, Japan). The layered steps formed on the (100) plane surface of 2,5-DHB disappeared from the crystal as the result of the incorporation of peptide molecules. The arrow indicates the direction of crystal growth by the stacking of the (011) crystal planes.

Scheme 8. Schematic illustration of intermolecular hydrogen-bonding between phenolic hydroxyl groups exposed from the hydrophilic (100) surface plane of the 2,5-DHB matrix crystal (lower) and the carbonyl oxygens of the peptide backbone (upper).

c) The formation of c fragments by intermolecular hydrogen transfer via hydrogen-bonding between matrix active hydrogens and backbone carbonyl oxygens

Hydrogen-bonding interactions between matrix active-hydrogens and backbone carbonyl oxygens (Scheme 8) can explain the MALDI-ISD phenomenon by assuming a photochemical intermolecular hydrogen transfer from the matrix to the analyte.^9,10) The photochemical intermolecular hydrogen transfer in UV-MALDI can be understood by an electron transfer within a very short time without atomic motion (Scheme 1 left). Hydrogen transfer (rearrangement or abstraction) between a carbonyl oxygen and adjacent hydrogens results in the formation of radical species and subsequent degradation such as the Norrish Type II reaction.⁵²⁾ As a result, therefore, carbonyl oxygens of organic compounds have a high affinity for hydrogen atoms due to the radical characteristics of the photochemically excited state transition n–π*, so that hydrogen atom abstraction from the matrix to the peptide backbone carbonyl oxygens can easily occur via hydrogen-bonding between matrix and peptide molecules that are excited with UV laser photons, as shown in Scheme 9. This type of hydrogen transfer may be due to a very short electron transfer time without any hydrogen atom movement, similar to a McLafferty hydrogen rearrangement (Scheme 1(a) left). The subsequent MALDI-ISD to form c-ions is completed within several tens of nanoseconds or less without intramolecular hydrogen scrambling in the MALDI ion-source.^9,53)

Scheme 9. MALDI-ISD processes involved in generating a c fragment. (a) Very fast electron transfer via hydrogen-bonding between matrix active hydrogen and the peptide backbone induced with laser photons, (b) the formation of 2,5-DHB phenoxy radicals, hydrogen atoms and hypervalent peptide radicals, and (c) the formation of ISD products c and z^· by α-cleavage.⁹⁾

According to the mechanism of intermolecular hydrogen transfer for MALDI-ISD presented in Schemes 8 and 9, Demeure et al. have attempted to present a strategy for the rational selection of the optimum MALDI-ISD matrix materials,¹⁸⁾ and they determined the hydrogen-donating characteristics of several matrices and gave the order of the ability, i.e., picolinic acid (PA)>1,5-diaminonaphthalene (1,5-DAN)>2,5-DHB>sinapinic acid (SA)>α-cyano-4-hydroxy-cinnamic acid (CHCA). We also reported on the hydrogen-donating properties of matrix materials as follows, 1,5-DAN>5-amino-salicylic acid (5-ASA)>2,5-DHB>SA≒CHCA.⁵⁴⁾ Demeure et al. applied a mechanistic model (Scheme 9) to top-down proteomics using MALDI-ISD MS.^18,55)

It should be emphasized from the experimental and rational results described above that the role of the X-ray crystallography and SPM analysis of matrix crystals is of great importance for developing a model construction of MALDI-ISD. Regarding this model, Asakawa et al. reported that the use of crystal phase matrices such as 2,5-DHB and 1,5-DAN resulted in abundant ISD ions via efficient hydrogen transfer, while hydrogen transfer was suppressed when a liquid or amorphous phase of a mixture of 2,5-DHB and 1,5-DAN was used.⁵⁶⁾ This observation strongly supports the model illustrated in Schemes 8 and 9.

MATRIX MATERIALS SUITABLE FOR MALDI-ISD: 5-POSITION AS KEY FUNCTIONAL GROUPS OF ISD

Brown et al. found that the most suitable matrix for the formation of ISD fragment ions was 2,5-DHB.³¹⁾ I also attempted to estimate the ability of producing abundant ISD fragment ions of angiotensin I, substance P, and ACTH18–39 by using dihydroxybenzoic acids (2,3-DHB, 2,4-DHB, 2,5-DHB, 2,6-DHB, 3,4-DHB, 3,5-DHB), trihydroxybenzoic acids (2,3,4-THB, 2,4,6-THB, and 3,4,5-THB) and 2-hydroxy-5-methoxybenzoic acid (2,5-HMBA) (Scheme 10),³³⁾ and it was finally concluded that 2,5-DHB was best matrix for MALDI-ISD experiments under following conditions:

Scheme 10. Matrix materials (molar absorptivity for ε337 nm) used for the MALDI-ISD experiments of peptides.³²⁾ Dihydroxybenzoic acids (2,3-DHB, 2,4-DHB, 2,5-DHB (ε337: 4250), 2,6-DHB(ε337: 700), 3,4-DHB, 3,5-DHB (ε337: 100)), trihydroxybenzoic acids (2,3,4-THB, 2,4,6-THB, 3,4,5-THB), and 2,5-HMBA(ε337: 5000).

1. 1. A hydroxyl group at the 2-position is an effective function for the formation of MALDI signals and ISD fragment ions, because the 2-position hydroxyl group can form a six-membered ring by hydrogen-bonding with the 1-position carboxyl group.
2. 2. A large molar absorptivity may be a requirement for ISD fragment ion formation.
3. 3. The 5-hydroxy group in 2,5-DHB has an important structural role in the formation of abundant ISD fragment ions. ISD experiments with a 5-methoxy group containing 2,5-HMBA matrix with a higher molar absorptivity of the nitrogen laser (337 nm) than that of 2,5-DHB did not show any ISD ion formation, as shown in Fig. 5.

Fig. 5. MALDI-ISD spectra of RLG (RLGNQWAVGDLAE, M_r 1428.6) obtained with (a) 2,5-DHB, and (b) 2,5-HMBA.

It should be noted from the above descriptions that the hydroxyl group at the 5-position in 2,5-DHB has an especially important structural and functional role in the formation of abundant ISD fragment ions. The important points in identifying suitable matrix materials for MALDI-ISD are to consider both the high molar absorptivity (ε337) and the appropriate positions of functional groups such as phenolic hydroxyl groups (Ph–OH) and anilinic amino groups (Ph–NH₂) which have a high capability for releasing hydrogen atoms. As a matter of fact, we identified matrix materials suitable for MALDI-ISD by considering such 5-position functional groups,^54,57) as shown in Scheme 11 left. On the other hand, Smargiasso et al. reported that 2-aminobenzamide (2-AB) and 2-aminobenzoic acid (2-AA) were suitable for both positive- and negative-ion MALDI-ISD of peptides and proteins (Scheme 11 middle).⁵⁸⁾ This indicates that the presence of anilinic amino group(s) efficiently induces the release of hydrogen atoms from amino groups.

Scheme 11. 5-position constitution salicylic acids, 5-ASA, 5-FSA and 5-NSA matrices for MALDI-ISD of peptide (left), 2-aminobenzoic acid (2-AA) and amide (2-AB) matrices (the middle), and 1,5-position constitution naphthalene derivatives, 1,5-DAN, 1,5-DHN, and 5,1-ANL matrices suitable for the positive- and negative-ion MALDI-ISD of intact proteins (right).

Furthermore, the recent report of a new matrix, 5-amino-1-naphthol (5,1-ANL),^59,60) suitable for both positive- and negative-ion MALDI-ISD of intact proteins including phosphorylated-peptides and -proteins has enabled a top-down analysis of intact proteins to be performed by using MALDI-ISD/CID experiments.^61,62) The 5,1-ANL and related-derivatives 1,5-diaminonaphthalene (1,5-DAN)^30,63) and 1,5-dihydroxynaphthalene (1,5-DHN)⁶⁰⁾ are useful for MALDI-ISD of proteins (Scheme 11 right). The 5,1-ANL matrix is particularly useful and has a superior capability for the comprehensive top-down analysis of intact proteins.¹⁴⁾ Crystals of 5,1-ANL, 1,5-DAN, and 1,5-DHN with and without cytochrome c incorporation are shown in Fig. 6.

Fig. 6. Crystals of 5-postion constituted naphthalene derivatives (a) 5,1-ANL, (b) 1,5-DHN, and (c) 1,5-DAN, without analyte (upper) and with cytochrome c incorporation (lower).

CONCLUSION

The MALDI-ISD of peptides and proteins can be understood by dividing the overall process into three steps. The first step can be considered to be the photochemical intermolecular abstraction of hydrogen by carbonyl oxygens of the peptide backbone from environmental matrix chemicals via hydrogen-bonding (Scheme 9(a)). The second step is the prompt simple (α)-cleavage of neutral radical analytes (Scheme 9(b)). The last step, the formation of ISD products (Scheme 9(c)) and ionization of the neutral species takes place via proton transfer reactions between neutral ISD fragments c, z′, z-M (Scheme 2) and matrix ions ([M+H]⁺ and [M–H]⁻) in the MALDI-plume. The ionization occurs at both basic and acidic functional groups according to the gas-phase basicity and acidity of the neutral peptide fragments. In the case of intact proteins, both positive- and negative-ion ISD spectra merely produce N-terminal side c-ions.^61,62) Therefore, MALDI-ISD can be used for the top-down sequencing of unknown and post-translational modified proteins.^{8,21–30,61,62)} It should be stressed that the model for MALDI-ISD was constructed on the basis of X-ray crystallography findings and SPM analyses of matrix crystals, as well as the use of isotope-labelled peptides.

Notes

Mass Spectrometry Society of Japan bestowed the 2016 MSSJ AWARD for Distinguished Contribution in MS on the author.

Abbreviations: ISD, in-source decay; SPM, scanning probe microscopy; RIF, radical-induced fragmentation; ESDI, energy-sudden desorption/ionization; 5,1-ANL, 5-amino-1-naphthol

Acknowledgments

The author gratefully acknowledges the financial supports of the MATSUDA fund of JEOL (JEOL Ltd., Tokyo, Japan), the Creation of Innovation Centers for Advanced Interdisciplinary Research Area Program in the Special Coordination Fund for Promoting Science and Technology, and the Grant-in-Aid for Scientific Research (C) (18550078, 20550081, 23550101) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

REFERENCES

• 1) M. Karas, D. Bachmann, U. Bahr, F. Hillenkamp. Matrix-assisted ultraviolet laser desorption of non-volatile compounds. Int. J. Mass Spectrom. Ion Process. 78: 53–69, 1987.
• 2) K. Tanaka, H. Waki, Y. Ido, S. Akita, Y. Yoshida, T. Yoshida. Protein and polymer analysis up to m/z 10,000 by laser ionization TOF-MS. Rapid Commun. Mass Spectrom. 2: 151–153, 1988.
• 3) M. Karas, F. Hillenkamp. Laser desorption ionization of proteins with molecular masses exceeding 10,000 daltons. Anal. Chem. 60: 2299–2301, 1988.
• 4) C. M. Whitehouse, R. N. Dreyer, M. Yamashita, J. B. Fenn. Electrospray interface for liquid chromatographs and mass spectrometers. Anal. Chem. 57: 675–679, 1985.
• 5) J. B. Fenn, M. Mann, C. K. Meng, S. F. Wong, C. M. Whitehouse. Electrospray ionization for mass spectrometry of large biomolecules. Science 246: 64–71, 1989.
• 6) R. S. Brown, J. J. Lennon. Sequence-specific fragmentation of matrix-assisted laser-desorbed protein/peptide ions. Anal. Chem. 67: 3990–3999, 1995.
• 7) M. Takayama, A. Tsugita. Does in-source decay occur independent of the ionization process in matrix-assisted laser desorption? Int. J. Mass Spectrom. 181: L1–L6, 1998.
• 8) M. Takayama, A. Tsugita. Sequence information of peptides and proteins with in-source decay in matrix assisted laser desorption/ionization-time of flight-mass spectrometry. Electrophoresis 21: 1670–1677, 2000.
• 9) M. Takayama. N–C_α bond cleavage of the peptide backbone via hydrogen abstraction. J. Am. Soc. Mass Spectrom. 12: 1044–1049, 2001.
• 10) T. Köcher, A. Engström, R. A. Zubarev. Fragmentation of peptides in MALDI in-source decay mediated by hydrogen radicals. Anal. Chem. 77: 172–177, 2005.
• 11) F. W. McLafferty. Mass spectrometric analysis molecular rearrangements. Anal. Chem. 31: 82–87, 1959.
• 12) I. Vidavsky, R. A. Chorush, P. Longevialle, F. W. McLafferty. Functional group migration in ionized long-chain compounds. J. Am. Chem. Soc. 116: 5865–5872, 1994.
• 13) M. Takayama. A terminal product ion in the fragmentation of methyl stearate under electron ionization conditions. J. Mass Spectrom. Soc. Jpn. 46: 139–142, 1998.
• 14) M. Takayama, K. Nagoshi, R. Iimuro, K. Inatomi. Access of hydrogen-radicals to the peptide-backbone as a measure for estimating the flexibility of proteins using matrix-assisted laser desorption/ionization mass spectrometry. Int. J. Mol. Sci. 15: 8428–8442, 2014.
• 15) R. Zenobi, R. Knochenmuss. Ion formation in MALDI mass spectrometry. Mass Spectrom. Rev. 17: 337–366, 1998.
• 16) K. Dreisweld. The desorption process in MALDI. Chem. Rev. 103: 395–426, 2003.
• 17) R. Knochenmuss, R. Zenobi. MALDI ionization: The role of in-plume processes. Chem. Rev. 103: 441–452, 2003.
• 18) K. Demeure, L. Quinton, V. Gabelica, E. De Pauw. Rational selection of the optimum MALDI matrix for top-down proteomics by in-source decay. Anal. Chem. 79: 8678–8685, 2007.
• 19) R. S. Brown, J. J. Lennon. Mass resolution improvement by incorporation of puled ion extraction in a matrix-assisted laser desorption/ionization linear time-of-flight mass spectrometer. Anal. Chem. 67: 1998–2003, 1995.
• 20) K. Biemann. Contributions of mass spectrometry to peptide and protein structure. Biomed. Environ. Mass Spectrom. 16: 99–111, 1988.
• 21) D. C. Reiber, T. A. Grover, R. S. Brown. Identifying proteins using matrix-assisted laser desorption/ionization in-source fragmentation data combined with database searching. Anal. Chem. 70: 673–683, 1998.
• 22) D. C. Reiber, R. S. Brown, S. Weinberger, J. Kenny, J. Bailey. Unknown peptide sequencing using matrix-assisted laser desorption/ionization and in-source decay. Anal. Chem. 70: 1214–1222, 1998.
• 23) V. Katta, D. T. Chow, M. F. Rohde. Applications of in-source fragmentation of protein ions for direct sequence analysis by delayed extraction MALDI-TOF mass spectrometry. Anal. Chem. 70: 4410–4416, 1998.
• 24) J. J. Lennon, K. Walsh. Locating and identifying posttranslational modifications by in-source decay during MALDI-TOF mass spectrometry. Protein Sci. 8: 2487–2493, 1999.
• 25) T. Kinumi, H. Niwa, H. Matsumoto. Phosphopeptide sequencing by in-source decay spectrum in delayed extraction matrix-assisted laser desorption ionization time-of-flight mass spectrometry. Anal. Biochem. 277: 177–186, 2000.
• 26) M. Takayama. Susceptible region of myoglobins to in-source decay using matrix-assisted laser desorption/ionization coupled with delayed extraction reflectron time-of-flight mass spectrometer. J. Mass Spectrom. Soc. Jpn. 50: 304–310, 2002.
• 27) J. Gao, A. Tsugita, M. Takayama, L. Xu. A programmable fragmentation analysis of proteins by in-source decay in MALDI-TOF mass spectrometry. Anal. Chem. 74: 1449–1457, 2002.
• 28) D. Suckau, A. Resemann. T³-sequencing: Targeted characterization of the N- and C-termini of undigested proteins by mass spectrometry. Anal. Chem. 75: 5817–5824, 2003.
• 29) V. Schnaible, S. Wefing, A. Resemann, D. Suckau, A. Buecker, S. Wolf-Kuemmel, D. Hoffmann. Screening for disulfide bonds in proteins by MALDI in-source decay and LIFT-TOF/TOF-MS. Anal. Chem. 74: 4980–4988, 2002.
• 30) L. Quinton, K. Demeure, R. Dobson, N. Gilles, V. Gabelica, E. De Pauw. New method for characterizing highly disulfide-bridged peptides in complex mixtures: Application to toxin identification from crude venoms. J. Proteome Res. 6: 3216–3223, 2007.
• 31) R. S. Brown, B. L. Carr, J. J. Lennon. Factors that influence the observed fast fragmentation of peptides in matrix-assisted laser desorption. J. Am. Soc. Mass Spectrom. 7: 225–232, 1996.
• 32) R. S. Brown, J. Feng, D. C. Reiber. Further studies of in-source fragmentation of peptides in matrix-assisted laser desorption–ionization. Int. J. Mass Spectrom. Ion Process. 169/170: 1–18, 1997.
• 33) M. Takayama. In-source decay characteristics of peptides in matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. J. Am. Soc. Mass Spectrom. 12: 420–427, 2001.
• 34) J. Hardouin. Protein sequence information by matrix-assisted laser desorption/ionization in-source decay mass spectrometry. Mass Spectrom. Rev. 26: 672–682, 2007.
• 35) D. Asakawa. Principles of hydrogen radical mediated peptide/protein fragmentation during matrix-assisted laser desorption/ionization mass spectrometry. Mass Spectrom. Rev., 2014. DOI: 10.1002/mas.21444
• 36) J. H. Moon, S. Yoon, Y. J. Bae, M. S. Kim. Formation of gas-phase peptide ions and their dissociation in MALDI: Insights from kinetic and ion yield studies. Mass Spectrom. Rev. 34: 94–115, 2015.
• 37) D. H. Williams, C. Bradley, G. Bojesen, S. Santiklarn, L. C. E. Tayler. Fast atom bombardment mass spectrometry: A powerful technique for the study of polar molecules. J. Am. Chem. Soc. 103: 5700–5704, 1981.
• 38) P. T. M. Kenny, W.-H. Xu, K. Tachibana, H. Naoki. in Proceedings of the 2nd Japan–China Joint Symposium on Mass Spectrometry (1987), pp. 177–180.
• 39) K. M. Downard, K. Biemann. Amino acid sequence prerequisites for the formation of c_n ions. J. Am. Soc. Mass Spectrom. 4: 874–881, 1993.
• 40) J. F. Mahoney, J. Perel, S. A. Ruatta, P. A. Martino, S. Husain, T. D. Lee. Massive cluster impact mass spectrometry: A new desorption method for the analysis of large biomolecules. Rapid Commun. Mass Spectrom. 5: 441–445, 1991.
• 41) H. J. Vorst, M. W. E. M. van Tilborg, P. A. van Veelen, U. R. Tjaden, J. van der Greef. Sequence-informative fragmentation of peptides up to a molecular weight of 4.6 kDa in plasma-desorption mass spectrometry. Rapid Commun. Mass Spectrom. 4: 202–205, 1990.
• 42) M. Takayama, T. Fukai, T. Nomura. Matrix effect on the extent of fragmentation in molecular ion M^+· produced under fast atom bombardment (FAB) conditions. Int. J. Mass Spectrom. Ion Process. 89: Rl, 1989.
• 43) R. A. Zubarev, N. L. Kelleher, F. W. McLafferty. Electron capture dissociation of multiply charged protein cations. A nonergodic process. J. Am. Chem. Soc. 120: 3265–3266, 1998.
• 44) C. T. J. Scott, C. Kosmidis, W. J. Jia, K. W. D. Ledingham, R. P. Singhal. Formation of atomic hydrogen in matrix-assisted laser desorption ionization. Rapid Commun. Mass Spectrom. 8: 829–832, 1994.
• 45) K. Strupat, M. Karas, F. Hillenkamp. 2,5-Dihydroxybenzoic acid: A new matrix for laser desorption–ionization mass spectrometry. Int. J. Mass Spectrom. Ion Process. 111: 89–102, 1991.
• 46) H. Ehring, M. Karas, F. Hillenkamp. Role of photoionization and photochemistry in ionization processes of organic molecules and relevance for matrix-assisted laser desorption ionization mass spectrometry. Org. Mass Spectrom. 27: 472–480, 1992.
• 47) S. Bourcier, S. Bouchonnet, Y. Hoppilliard. Ionization of 2,5-dihydroxybenzoic acid (DHB) matrix-assisted laser desorption ionization experiments and theoretical study. Int. J. Mass Spectrom. 210/211: 59–69, 2001.
• 48) I.-C. Lu, K. Y. Chu, C.-Y. Lin, S.-Y. Wu, Y. A. Dyakov, J.-L. Chen, A. Gray-Weale, Y.-T. Lee, C.-K. Ni. Ion-to-neutral ratios and thermal proton transfer in matrix-assisted laser desorption/ionization. J. Am. Soc. Mass Spectrom. 26: 1242–1251, 2015.
• 49) K. Demeure, V. Gabelica, E. A. De Pauw. New advances in the understanding of the in-source decay fragmentation of peptides in MALDI-TOF-MS. J. Am. Soc. Mass Spectrom. 21: 1906–1917, 2010.
• 50) K. Strupat, J. Kampmeier, V. Horneffer. Investigations of 2,5-DHB and succinic acid as matrices for UV and IR MALDI. Part II: Crystallographic and mass spectrometric analysis. Int. J. Mass Spectrom. Ion Process. 169/170: 43–50, 1997.
• 51) M. Takayama. The characteristics of in-source decay in mass spectrometric degradation methods—Hydrogen-Attachment Dissociation (HAD)—. J. Mass Spectrom. Soc. Jpn 50: 337–349, 2002.
• 52) P. J. Wagner. Type II Photoelimination and photocyclization of ketones. Acc. Chem. Res. 4: 168–177, 1971.
• 53) N. Bache, K. D. Rand, P. Roepstorff, T. J. D. Jørgensen. Gas-phase fragmentation of peptides by MALDI in-source decay with limited amide hydrogen (¹H/²H) scrambling. Anal. Chem. 80: 6431–6435, 2008.
• 54) M. Sakakura, M. Takayama. In-source decay and fragmentation characteristics of peptides using 5-aminosalicylic acid as a matrix in matrix-assisted laser desorption/ionization mass spectrometry. J. Am. Soc. Mass Spectrom. 21: 979–988, 2010.
• 55) K. Demeure, L. Quinton, V. Gabelica, E. De Pauw. Top-down proteomics using matrix enhanced ISD. Proceedings 55th ASMS Conference on Mass Spectrometry, WOA am-09:55, Indianapolis, 3–7 June 2007.
• 56) D. Asakawa, D. Calligaris, N. Smargiasso, E. De Pauw. Ultraviolet laser induced hydrogen transfer reaction: Study of the first step of MALDI in-source decay mass spectrometry. J. Phys. Chem. B 117: 2321–2327, 2013.
• 57) D. Asakawa, M. Takayama. C_α–C bond cleavage of the peptide backbone in MALDI in-source decay using salicylic acid derivative matrices. J. Am. Soc. Mass Spectrom. 22: 1224–1233, 2011.
• 58) N. Smargiasso, L. Quinton, E. De Pauw. 2-Aminobenzamide and 2-aminobenzoic acid as new matrices inducing radical mediated in-source decay of peptides and proteins. J. Am. Soc. Mass Spectrom. 23: 469–474, 2012.
• 59) M. Takayama. Flexible Xxx–Asp/Asn and Gly–Xxx residues of equine cytochrome c in matrix-assisted laser desorption/ionization mass spectrometry. Mass Spectrom. (Tokyo) 1: A0007, 2012.
• 60) I. Osaka, M. Sakai, M. Takayama. 5-Amino-1-naphthol, a novel 1,5-naphthalene derivative matrix suitable for matrix-assisted laser desorption/ionization in-source decay of phosphorylated peptides. Rapid Commun. Mass Spectrom. 27: 103–108, 2013.
• 61) M. Takayama, S. Sekiya, R. Iimuro, S. Iwamoto, K. Tanaka. Selective and nonselective cleavages in positive and negative CID of the fragments generated from in-source decay of intact proteins in MALDI-MS. J. Am. Soc. Mass Spectrom. 25: 120–131, 2014.
• 62) S. Sekiya, K. Nagoshi, S. Iwamoto, K. Tanaka, M. Takayama. Timeframe dependent fragment ions observed in in-source decay experiments with β-casein using MALDI MS. J. Am. Soc. Mass Spectrom. 26: 1588–1598, 2015.
• 63) Y. Fukuyama, S. Iwamoto, K. Tanaka. Rapid sequencing and disulfide mapping of peptides containing disulfide bonds by using 1,5-diaminonaphthalene as a reductive matrix. J. Mass Spectrom. 41: 191–201, 2006.