Detection of Neutral Species in the MALDI Plume Using Femtosecond

Article pubs.acs.org/JPCA

Detection of Neutral Species in the MALDI Plume Using Femtosecond Laser Ionization: Quantitative Analysis of MALDI-MS Signals Based on a Semiequilibrium Proton Transfer Model Tatsuro Shirota,†,‡ Masashi Tsuge,§ Yasumasa Hikosaka,∥ Koichi Soejima,¶ and Kennosuke Hoshina*,† †

Faculty of Pharmaceutical Sciences, Niigata University of Pharmacy and Applied Life Sciences, 265-1, Higashijima, Akiha-ku, Niigata 956-8603, Japan ‡ Graduate School of Science and Technology, and ¶Faculty of Science, Niigata University, 8050, Ikarashi 2-no-cho, Nishi-ku, Niigata 950-2181, Japan § Department of Applied Chemistry, National Chiao Tung University, 1001 Ta-Hsueh Road, Hsinchu 30010, Taiwan ∥ Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, 2630 Sugitani, Toyama 930-0194, Japan ABSTRACT: We investigated neutral species in the matrixassisted laser desorption and ionization (MALDI) plume using femtosecond laser ionization spectrometry with simultaneous measurement of the standard MALDI spectrum of the identical MALDI event induced by pulsed UV laser irradiation. The ratio of neutral species in the plume [A]p/[M]p (A = phenylalanine (Phe) or alanine (Ala), M = 2,5-dihydroxybenzoic acid (DHB)) was confirmed to be the same as that of the sample mixture in the range of [A]0/[M]0 = 4 × 10−4−1, indicating the validity of the widely adopted approximation [A]p/[M]p = [A]0/[M]0 in the reaction quotient of the proton transfer reaction MH+ + A ⇄ M + AH+. An effective parameter representing the extent of thermal equilibrium in the thermal proton transfer model is introduced for the first time. Numerical simulation based on this semiequilibrium model successfully reproduced variations of MALDI signal intensities AH+ and MH+ with two parameters: the fraction of ionized matrix a ≤ 10−5 and an effective temperature T = 1200 and 1100 K for Phe/DHB and Ala/DHB systems, respectively. These values show good agreement with those determined previously by different experimental approaches. The extent of thermal equilibrium was determined to be 95% and 98% for Phe/DHB and Ala/DHB systems, respectively, suggesting that the proton transfer reactions almost proceed to their thermal equilibrium.

1. INTRODUCTION Matrix-assisted laser desorption and ionization mass spectrometry (MALDI-MS) is a rapid and highly sensitive detection technique for nonvolatile and fragile molecules, especially for biologically important molecular species such as peptides and proteins.1−4 The sensitivity of MALDI detection has improved to the femtomole range in applications to current proteomics research. In the MALDI measurement, cocrystals of analyte (A) and matrix (M) are prepared on a metallic plate and irradiated with UV laser light to induce desorption and ionization. Typical matrix molecules, such as 2,5-dihydroxybenzoic acid (DHB) used in this study, efficiently absorb UV photons. The rapid temperature rise of the crystal after UV irradiation results in vaporization of heated matrix molecules, including the nondestructive vaporization of analytes. Another important role of the matrix is that it produces protonated matrix MH+ and deprotonated matrix (M−H)− from the MM dimer or two matrix molecules during the desorption process.2,5 In the dense vapor just after desorption, generally called the MALDI plume, MH+ and (M−H)− work as proton donors and acceptors producing standard MALDI signals, AH+ and (A−H)−, © XXXX American Chemical Society

respectively, through proton transfer reactions with vaporized A: MH+ + A ⇄ M + AH+ and (M−H)− + A ⇄ M + (A−H)−. A positive correlation between the AH+ signal intensity and the gas-phase basicity (GB) of A in the positive ion detection mode, and that between (A−H)− and the gas-phase acidity (GA) of A in the negative ion detection mode, confirms that the proton transfer reactions are major processes that produce MALDI signals;6−8 however, whether desorption and proton transfer are completely sequential processes is a subject under discussion.9−11 Determining the concentration of analytes in a MALDI sample from MALDI signals quantitatively is challenging. As seen in the coupled physical and chemical dynamics (CPCD) model, we need a lot of parameters to provide explicit formula connecting prepared sample condition and resultant MALDI signal intensities.12 Even so, the simplest model assuming that proton transfer reactions reach thermal equilibrium has Received: September 22, 2016 Revised: December 8, 2016 Published: December 9, 2016 A

DOI: 10.1021/acs.jpca.6b09591 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

and (ii) secondary ionization in which AH+ and (A−H)− are produced by the following proton transfer reactions

succeeded, to some extent, in the interpretation of the signal ratio [AH+]/[MH+] in MALDI spectra of the positive ion mode.7 In this model, the ratio of nondetectable neutral species is approximated by the sample mixing ratio [A]0/[M]0 because it has been suggested in several reports that a very small fraction of uniformly desorbed A and M, ranging from 10−5 to 10−8, is ionized in the MALDI plume.1,2,13−16 When the reaction quotient is expressed as Q = [M]0/[A]0 × [AH+]/ [MH+], Q must have a constant value corresponding to the equilibrium constant K that is independent of the sample mixing ratio [A]0/[M]0. This also results in a linear relationship between ln([AH+]/[MH+]) and ln([A]0/[M]0) with a slope of 1. The treatment could reproduce variations in [AH+]/[MH+] as a function of [A]0/[M]0 in the low [A]0/[M]0 region for amino acids7 and peptides;17,18 however, Q gradually decreases as [A]0/[M]0 increases, and the deviation becomes significant at [A]0/[M]0 as high as 10−2−1, depending on the analytes and matrices. In this case, a ln([AH+]/[MH+])−ln([A]0/[M]0) plot becomes plateaulike in the high [A]0/[M]0 region and linearly increases in the low [A]0/[M]0 region. In addition, evaluation

(2)

(M−H)− + A ⇄ M + (A−H)−

(3)

Assuming that proton transfer reactions 2 and 3 occurring in the MALDI plume reach thermal equilibrium, the relationship between amounts of species in the MALDI plume is formulated as follows.21,22 The equilibrium constants for reactions 2 and 3, K+ and K−, respectively, are expressed as K+ =

◦ [M]e [AH+]e = e−ΔG+ / RT + [A]e [MH ]e

(4)

K− =

◦ [M]e [(A−H)− ]e = e−ΔG−/ RT − [A]e [(M−H) ]e

(5)

where subscript e represents the amount of each species in the MALDI plume when proton transfer reactions reach thermal equilibrium and T is the effective temperature of the MALDI plume. ΔG+° and ΔG−° are differences in the standard Gibbs free energy for reactions 2 and 3, respectively, expressed as

o

of the absolute value of Q = e−ΔG /RT in this model has not yet been carried out. We previously proposed that when using amino acids and their esters as analytes, there are two possible failures in the thermal equilibrium model: (i) A and M are not uniformly desorbed resulting in [A]p/[M]p ≠ [A]0/[M]0, where [A]p and [M]p represent densities of A and M in the MALDI plume, and/or (ii) the MALDI plume expands to collision-free conditions before reaching thermal equilibrium, and therefore Q is not equal to K but variable.7,19 In this study, we developed equipment that can probe neutral species A and M in the MALDI plume using near-infrared femtosecond (fs) laser ionization mass spectrometry (MALDIfs-MS) with simultaneous measurement of the standard MALDI spectrum of the identical MALDI event induced by pulsed UV laser irradiation. The ability of fs-laser ionization to suppress fragmentation20 provides reference signals for estimation of [A]p/[M]p, and we first examined the above possibility (i) based on the MALDI-fs-MS measurement. Next, we introduced a semiequilibrium model with a parameter representing the extent of thermal equilibrium in association with possibility (ii). As the CPCD model predicts, the extent of proton transfer must be determined by kinetics, in preference to thermochemically. This parameter effectively convolutes this issue.12 Numerical simulations were performed to reproduce variation of the absolute values of [AH+]/[MH+] as a function of [A]0/[M]0. Optimized parameters for the fraction of ionized matrix and the plume temperature, which characterize the MALDI plume, were compared with those determined previously by different approaches.

ΔG+° = G B°(M) − G B°(A)

(6)

ΔG−° = −GA °(M) + GA °(A)

(7)

where GB°(X) and GA°(X) are the gas-phase basicity and acidity of X at standard conditions, respectively. From eqs 4 and 5, the following relationships are obtained: ln

ΔG+° [AH+]e [A]e = ln − + RT [MH ]e [M]e

(8)

ln

[(A−H)− ]e [A]e ΔG−° = ln − − RT [(M−H) ]e [M]e

(9)

Although the ratio [A]e/[M]e could not be obtained from MALDI-MS measurements, it can be replaced by the analyte/ matrix mixing ratio [A]0/[M]0 at sample preparation, considering that (i) A and M are uniformly desorbed and (ii) a small fraction of desorbed species is ionized. Although only condition (ii) has been demonstrated by several researchers,14−16,2325 the widely adopted approximative relationship between experimentally measurable [AH+]/[MH+] and the sample mixing ratio [A]0/[M]0 and that between [(A−H)−]/ [(M−H)−] and [A]0/[M]0 are obtained as

2. MODEL FOR ION FORMATION MECHANISM IN MALDI 2.1. Thermal Equilibrium Model. It has been widely recognized that the formation of typical MALDI signals, i.e., protonated analyte AH+ and deprotonated analyte (A−H)−, involves a two-step ionization process: (i) primary ionization in which MH+ and (M−H)− are produced from the energetic matrix dimer or two matrix molecules MM (or M + M) → MH+ + (M−H)−

MH+ + A ⇄ M + AH+

ln

ΔG+° [AH+]e [A]0 = ln − + RT [MH ]e [M]0

(10)

ln

[(A−H−)]e [A]0 ΔG−° = ln − − RT [(M−H )]e [M]0

(11)

According to the thermal equilibrium model, the MALDI signal ratio [AH+]/[MH+] and the initial mixing ratio [A]0/[M]0 show a linear relationship with a slope of 1 and an intercept of −ΔG+°/RT in the double logarithmic plot of species in the positive ion detection mode. 2.2. Thermal Semiequilibrium Model. If the MALDI plume expands before reaching thermal equilibrium, the formulation becomes slightly complex because the estimation of transient densities of each species needs additional

(1) B


Article

The Journal of Physical Chemistry A parameters associated with the initial densities and reaction rates. In the present simulation, we introduced a parameter a that represents the initial ratio of MH+ and (M−H)− to M: [MH+]0 = [(M−H)− ]0 = a[M]0

(12)

When the proton transfer reactions reach equilibrium conditions, relative amounts of each species in eqs 4 and 5 are expressed as [M]e = [M]0 + [AH+]e + [(A−H)− ]e − [MH+]0 − [(M−H)− ]0

(13)

[A]e = [A]0 − [AH+]e − [(A−H)− ]e

(14)

[MH+]e = a[M]0 − [AH+]e

(15)

[(M−H)− ]e = a[M]0 − [(A−H)− ]e

(16)

Substituting eqs 13−16 for eqs 4 and 5 allows the relative values of [MH+]e, [(M−H)−]e, [AH+]e, and [(A−H)−]e to [M]0 to be obtained numerically at given conditions {a, T}. Regarding the reaction rates, we introduced two parameters b+ (0 ≤ b+ ≤ 1) and b− (0 ≤ b− ≤ 1) representing the extent of thermal equilibrium for reactions 2 and 3, respectively. Consequently eqs 4 and 5 are written as Q+ =

deposited on a stainless sample stage. The mixtures were prepared at 13 analyte/matrix mixing ratios in the range from 4 × 10−4 to 1. 3.2. Measurements of MALDI-MS and MALDI-fs-MS. Figure 1 illustrates the MALDI-MS system constructed in our laboratory. The MALDI source equipment (Jordan Co.) and a reflectron-type time-of-flight (TOF) mass spectrometer (Jordan Co.) are connected to the main chamber that is attached to a turbo molecular pump (Osaka Vacuum Co. 450 L/s). The TOF tube is differentially evacuated by another turbo molecular pump (220 L/s). The base pressures of the main chamber and the TOF tube are 5 × 10−5 and 1 × 10−5 Pa, respectively. The third harmonic of output of a Nd:YAG ns-laser (355 nm, 10 ns, 10 Hz, < 1 mJ/pulse, Quanta-Ray INDI; SpectraPhysics) was loosely focused (f = 500 mm) on the sample at an incident angle of 30° to induce MALDI processes. At the sample plate, the focal spot was estimated to be a 2 × 4 mm ellipse resulting in a laser fluence of 130 J/m2 per pulse. A static high voltage typically at +4.5 kV was applied to the sample plate, and accelerated ionic species were detected by the microchannel plate (MCP) through the reflectron-type TOF mass spectrometer. Ion signals converted to digital data (250 MS/s) on an oscilloscope were transferred to a PC and averaged for typically 2000 shots without moving the focal spot. Since ion signals rapidly decreases as number of shots, we averaged data of the first 250 shots to obtain reliable MALDI spectra. In the MALDI-fs-MS measurement, femtosecond laser light (800 nm, 100 fs, 10 Hz, 0.2−0.4 mJ/pulse) from a Ti:sapphire laser was introduced into the main chamber parallel to the sample plate and focused at the center of the MALDI plume by a planoconvex lens ( f = 250 mm) to ionize neutral species after desorption. The laser field intensity at the focal spot, located 2 mm from the sample plate in the present experiment, is estimated to be 1 × 1014 W/cm2. The fs-laser pulses are synchronized with the ns-laser pulses. The delay time between ns-laser and fs-laser pulses was varied in the range of 5−40 μs. For this range, the changes in the ratio of specific ion signals, used for analysis, were within fluctuation error. We set the delay

(1 − 2a)[M]0 + b+[AH+]e + b−[(A−H)− ]e [A]0 − b+[AH+]e − b−[(A−H)− ]e b+[AH+]e a[M]0 − b+[AH+]e

Q− =

Figure 1. Schematic of a MALDI-TOF-MS spectrometer with a postionization source using femtosecond laser irradiation.

(17)

(1 − 2a)[M]0 + b+[AH+]e + b−[(A−H)− ]e [A]0 − b+[AH+]e − b−[(A−H)− ]e b−[(A−H)− ]e a[M]0 − b−[(A−H)− ]e

(18)

where Q+ and Q− are reaction quotients. The relationships between [MH+], [(M−H)−], [AH+], and [(A−H)−] are thus ln

[A]e [AH+] = ln + ln Q + [MH+] [M]e

(19)

ln

[A]e [(A−H)− ] = ln + ln Q − [(M−H)− ] [M]e

(20)

Parameters b+ and b− are effective indices determined by the rate constants of the forward and backward reactions, the varying density, and local homogeneity of the MALDI plume. In the present study, we assumed that b+ and b− are independent of the initial density of each species.

3. EXPERIMENTAL SECTION 3.1. Sample Preparation. Analyte species phenylalanine (Phe) and alanine (Ala) were purchased from Sigma-Aldrich and Wako Pure Chemical Industries, respectively, and used without further purification. Matrix reagent DHB and ultrapure water were purchased from Wako Pure Chemical Industries. MALDI samples were prepared by the dried droplet method26 under atmospheric pressure at room temperature. In brief, 10 μL of analyte solution (0.02−50 μmol/mL in water) and 10 μL of DHB matrix solution (50 μmol/mL in water) were C


Article


Figure 2. (a) MALDI-TOF spectrum of DHB measured by UV ns-laser (355 nm, 10 ns, 10 Hz, ∼ 0.8 mJ/pulse) desorption/ionization. MALDI and MALDI-fs TOF spectra of (b) DHB, (c) [Phe]0/[DHB]0 = 1, (d) [Ala]0/[DHB]0 = 1 samples measured with additional femtosecond laser (800 nm, 100 fs, 10 Hz, 0.2−0.4 mJ/pulse) irradiation. MALDI-fs spectra seen in the range of TOF > 29 μs overlap with MALDI spectra in traces (b)−(d). R+(=C6H5CH2+) represents the side chain of Phe.

Figure 3. (a) MALDI spectra of the Phe/DHB system measured at three initial sample mixing ratios, [Phe]0/[DHB]0 = 0.001(bottom), 0.1(middle), and 0.6(top). The vertical scales are normalized by each DHB+ signal. (b) Corresponding spectral region of MALDI-fs spectra where DHB+ and Phe-derived ions are observed. The vertical scales are normalized by each DHB+ signal.

time to 30 μs so that the overlapping of MALDI and MALDI-fs signals is minimized rather than optimizing the MALDI-fs signal intensities.

30 μs, several new signals are observed in the subsequent spectral region, and three distinct peaks are assigned as DHBderived signals: C4H4+ (m/z 52), (DHB−H2O)+ (m/z 136), and DHB+ (m/z 154). Hereafter, C4H4+ is written as (DHB−102)+. These three species were also strongly observed in the electron ionization mass spectra of DHB,27,28 confirming the assignment. The signals of intact DHB+, dehydrated (DHB−H2O)+, and (DHB−102)+ are DHB-derived species that can be used as reference signals for estimating relative amounts of neutral DHB in the MALDI plume. Figure 2c and d show the MALDI/MALD-fs TOF spectra for Phe/DHB and Ala/DHB mixtures, respectively. In the MALDI region of Figure 2c, (Phe)H+ (m/z 166), (Phe)Na+ (m/z 188), and (Phe)K+ (m/z 204) are clearly observed, indicating that the MALDI process proceeded normally. Many fragment ion peaks of Phe are detected, in addition to DHB-derived ions, in the MALDI-fs region of Figure 2c. Among these signals, we identified three peaks useful for reference of neutral Phe: (Phe−R)+, (Phe−COOH)+, and R+, where R represents the side chain C6H5CH2. Hereafter, R+ is written as (Phe−74)+. Similarly, (Ala−COOH)+ is observed clearly in the MALDI-fs

4. RESULTS 4.1. Overview of MALDI and MALDI-fs TOF Spectra. Figure 2a shows the MALDI-TOF spectrum of DHB obtained by UV ns-laser irradiation. Typical MALDI signals of protonated DHB, (DHB)H+ (m/z 155), and alkali metal adduct DHBs, (DHB)Na+ (m/z 177), and (DHB)K+ (m/z 193), are assigned, indicating that our MALDI-MS system operates properly, although the (DHB)Na+ and (DHB)K+ signals are unexpectedly strong. A peak from DHB+ (m/z 154) is also observed at an intensity comparable to that of (DHB)H+. Figure 2b shows the MALDI/MALDI-fs TOF spectrum of DHB measured by adding fs-laser irradiation into the MALDI plume generated under the same conditions as those in Figure 2a. The spectral pattern is reasonably identical to that of Figure 2a in the TOF region before 30 μs. After fs-laser irradiation at D


Article


Figure 4. (a) MALDI spectra of the Ala/DHB system measured at three initial sample mixing ratios, [Ala]0/[DHB]0 = 0.001 (bottom), 0.1 (middle), and 0.6 (top). The vertical scales are normalized by each DHB+ signal. (b) Corresponding spectral region of MALDI-fs spectra where (DHB−102)+ and (Ala−COOH)+ are observed. The vertical scales are normalized by each (DHB−102)+ signal.

Figure 5. Logarithmic plots of MALDI-fs signals versus sample mixing ratio. (a) Signal intensity ratio of the side chain R+ (=C6H5CH2+) corresponding to (Phe−74)+ and DHB+ observed in the MALDI-fs spectra. (b) Signal intensity ratio of (Ala−COOH)+ and (DHB−102)+ observed in the MALDI-fs spectra.

region of Figure 2d as a reference signal for Ala, while (Ala)H+ (m/z 90), (Ala)Na+ (m/z 112), and (Ala)K+ (m/z 128) are properly detected in the MALDI region. The reference signals for Phe and Ala are also observed as main fragment peaks in the electron ionization mass spectra29 and in fs-laser ionization mass spectra at a laser intensity of 8 × 1014 W/cm2 with a 100 fs pulse duration,30,31 indicating that these reference peaks come from neutral Phe and Ala. 4.2. Analysis of MALDI-fs Spectra and Ratio of Neutral Species in the MALDI Plume. Typical MALDI and MALDIfs spectra of the Phe/DHB system measured by changing the sample mixing ratio are compared in Figure 3a and b, where the vertical scales are normalized to DHB+, respectively. The corresponding spectra for the Ala/DHB system are shown in Figure 4a and b. As [A]0/[M]0 increases, the amino acidderived peak intensities increase. Signals of (DHB)H+ observed in the MALDI-fs spectra of Figure 4b originate from dissociative ionization of (DHB)2, as demonstrated in fs-laser ionization of carboxylic acid dimers.32 We first analyzed the variation of MALDI-fs signal intensities to examine the validity of the approximation [A]p/[M]p = [A]0/[M]0, which is the fundamental assumption of the thermal equilibrium and semiequilibrium models.

Figure 5a shows logarithmic plots of the signal ratio [(Phe−74)+]/[DHB+] obtained from the MALDI-fs spectra as a function of the sample mixing ratio [Phe]0/[DHB]0. The plots of ln[(Phe−74)+/DHB+] and ln([Phe]0/[DHB]0) have a linear relationship with a slope of 0.93(3), obtained by fitting data in the range of 1 × 10−1 ≤ [Phe]0/[DHB]0 ≤ 1. Considering that the amounts of (Phe−74)+ and DHB+ are proportional to [Phe]p and [DHB]p, respectively, MALDI-fs signals are expressed as [(Phe−74)+ ] = αI n[Phe]p

(21)

[DHB+] = βI m[DHB]p

(22)

where I is the laser intensity, n and m are the numbers of photons necessary for ionization and subsequent fragmentation, and α and β are proportional constants associated with dissociative ionization probability and detection efficiency. From eqs 21 and 22, the relationship ln

⎤ ⎡ α [Phe]p [(Phe−74)+ ] = ln + ⎢ln + (n − m)ln I ⎥ + ⎦ ⎣ β [DHB]p [DHB ] (23)

E


Article


Figure 6. (a) Relative signal intensities of (Phe)H+ (blue circles) and (DHB)H+ (red circles) as a function of sample mixing ratio [Phe]0/[DHB]0. Solid lines show simulations at a = 10−5, b+ = 0.95, and T = 1200 K for (Phe)H+ (blue) and (DHB)H+ (red) by the semiequilibrium model. Broken lines are simulations when proton transfer reaches thermal equilibrium (b+ = 1.00) at a = 10−5 and T = 1200 K. (b) MALDI signal intensity ratio [(Phe)H+]/[(DHB)H+] (circles) as a function of [Phe]0/[DHB]0. Simulated [(Phe)H+]/[(DHB)H+] at a = 10−5, b+ = 0.95, and T = 1200 K (solid line) and at a = 10−5, b+ = 1.00, and T = 1200 K (broken line).

Figure 7. (a) Relative signal intensities of (Ala)H+ (blue circles) and (DHB)H+ (red circles) as a function of sample mixing ratio [Ala]0/[DHB]0. Solid lines show simulations at a = 10−5, b+ = 0.98, and T = 1100 K for (Ala)H+ (blue) and (DHB)H+ (red) by the semiequilibrium model. Broken lines are simulations when proton transfer reaches thermal equilibrium (b+ = 1.00) at a = 10−5 and T = 1100 K. (b) MALDI signal intensity ratio [(Ala)H+]/[(DHB)H+] (circles) as a function of [Ala]0/[DHB]0. Simulated [(Ala)H+]/[(DHB)H+] at a = 10−5, b+ = 0.98, and T = 1100 K (solid line) and at a = 10−5, b+ = 1.00, and T = 1100 K (broken line).

m/z as (Ala−COOH)+. These results indicate that the approximated eqs 10 and 11, in which [A]e/[M]e is replaced by [A]0/[M]0, are applicable in the high and low [A]0/[M]0 regions, confirming that a negligible fraction of uniformly desorbed A and M is ionized as mentioned in the experimental section. As a consequence, we excluded the possibility that the plateau or decrease in Q+ seen in the ln([AH+]/[MH+])− ln([A]0/[M]0) plots in previous reports7,17−19 is attributed to a failure of this approximation. 4.3. Quantitative Analysis of MALDI Signals Based on a Thermal Semiequilibrium Model. Figure 6a shows logarithmic plots of the MALDI signal intensity of (Phe)H+ and (DHB)H+ relative to their sum, taken from the same spectra as those used for plotting the MALDI-fs signals in Figure 5a, as a function of [Phe]0/[DHB]0. The signal intensity of (Phe)H+ increases as [Phe]0/[DHB]0 increases; consequently (DHB)H+ decreases in the high [Phe]0/[DHB]0 region, indicating that the proton transfer reaction from (DHB)H+ to Phe proceeds to some extent. This trend gradually decelerates, and the signal intensities of (Phe)H+

is obtained. Therefore, the linear relationship with a slope of 1 seen in Figure 5a means that Phe and DHB are uniformly desorbed, even in the high [Phe]0/[DHB]0 region. Therefore, since the [A]p/[M]p = [A]0/[M]0 is satisfied in the low [A]0/ [M]0 region, we can also apply this approximation in the high [A]0/[M]0 region. Similarly, Figure 5b shows a logarithmic plot of [(Ala− COOH)+]/[(DHB−102)+] as a function of the sample mixing ratio [Ala]0/[DHB]0. We adopted (DHB−102)+ as a reference peak for the Ala/DHB system because the offset of ln([(Ala− COOH)+]/[(DHB−102)+]) plots is not affected by unintended small changes in the laser intensity. In contrast, the offset of ln([(Ala−COOH)+]/[DHB+]) is affected by the laser intensity, meaning m ≠ n in eq 23 in the case of (Ala− COOH)+ and DHB+. The slope 1.08(8) obtained from fitting data in the range of 1 × 10−1 ≤ [Ala]0/[DHB]0 ≤ 1 also means [Ala]p/[DHB]p = [Ala]0/[DHB]0. The constant value of the signal intensity ratio (Ala−COOH)+/(DHB−102)+ in the [Ala]0/[DHB]0 region below 1 × 10−1 comes from a small DHB-derived peak, possibly COO+, appearing with the same F


Article


detection,25 and 10−8−10−9 for DHB using a rotatable quadrupole mass spectrometer.15,16 Although the ratio depends on species and laser fluence, it is in the range of 10−3−10−5 at conditions applicable to MALDI detection of analytes. Our result a ≤ 10−5 might be consistent with the reported value 10−8−10−9 for DHB after following considerations. We also calculated ion-to-neutral ratio of analytes [AH+]e/[A]e by performing simulations with the optimized temperature T and parameter b+ while changing the parameter a; the values are [(Phe)H+]e/[Phe]e = 6 × 10−3, 6 × 10−4, 6 × 10−5, and 6 × 10−6 at a = 10−5, 10−6, 10−7, and 10−8, respectively, at [A]0/ [M]0 = 10−4. The corresponding values for Ala/DHB are [(Ala)H+]e/[Ala]e = 9 × 10−3, 9 × 10−4, 9 × 10−5, and 9 × 10−6. Comparison of these [AH+]e/[A]e values with the reported values 10−3−10−5 leads to the conclusion that a ≤ 10−5 determined from our simulations corresponds to the ionto-neutral ratio 10−6−10−8 for DHB. MALDI Plume Temperature. MALDI plume expansion is recognized to have two stages:3 the early plume where insource decay occurs in the dense gaseous phase with a temperature Tearly, and the late plume where post-source decay proceeds under collision-free conditions with a cooled temperature Tlate. The temperatures are estimated to be in the range of Tearly = 785−1000 K37−39 and Tlate = 400−430 K38−40 for DHB, based on the internal energy determined from fragmentation patterns. The proton transfer from MH+ to A is thought to occur in the early plume.3 This was confirmed in the report by Kim et al., where the precision of ln([AH+]/ [MH+])−ln([A0]/[M0]) plots was improved by selecting MALDI spectra with Tearly.18 Our estimation of 1100−1200 K as well as our previous report7 is higher than Tearly suggesting that the proton transfer proceeds prior to in-source decay of protonated species. Extent of Proton Transfer. The value of the extent of proton transfer b+ = 0.95−0.98 indicates that the proton transfer approach thermal equilibrium when plume expands to collisionfree conditions. The values of b+ are determined by the rate constants of the forward and backward reactions 2 under varying physical and chemical conditions of MALDI plume12 and therefore can be dependent on the analytes. The similar values determined for the Phe/DHB and Ala/DHB systems suggest that the proton transfer rates from (DHB)H+ to these amino acids are almost the same. This is probably due to their exothermal reaction pathway involving the same proton transfer processes, i.e., from the COOH group of DHB to the NH2 group of the amino acid, resulting in similar activation energies.

and (DHB)H+ seem to have constant values in the high [Phe]0/[DHB]0 region. The ratio of (Phe)H+/(DHB)H+ plotted in Figure 6b as a function of [Phe]0/[DHB]0 shows a linear increase with a slope of 1 in the low [Phe]0/[DHB]0 region and a plateau in the high [Phe]0/[DHB]0 region. If the proton transfer reactions reach thermal equilibrium, the relationship between ln([(Phe)H+]/ [(DHB)H+]) and ln([Phe]0/[DHB]0) follows eq 11, showing a linear increase with a slope of 1 in any [Phe]0/[DHB]0 range as long as the approximation [Phe]p/[DHB]p = [Phe]0/[DHB]0 is applicable. We demonstrated the validity of the approximation in the present [Phe]0/[DHB]0 range by simultaneous measurement of MALDI and MALDI-fs spectra as described in the previous section. Therefore, the plateaulike behavior suggests that the MALDI plume expanded before the proton transfer reaches thermal equilibrium. The situation is qualitatively the same in the case of the Ala/DHB system, as seen in Figure 7a and b. We examined whether the semiequilibrium model reproduces the variation in four data sets, that is, (Phe)H+ and (DHB)H+ in the Phe/DHB system, and (Ala)H+ and (DHB)H+ in the Ala/DHB system. The GA and GB values used in the simulation are GA(Phe) = 1379.0 kJ/mol,33 GA(Ala) = 1398.5 kJ/mol,33 GA(DHB) = 1329.4 kJ/mol,34 GB(Phe) = 887.4 kJ/mol,35 GB(Ala) = 863.6 kJ/mol,35 and GB(DHB) = 822 kJ/mol.36 Among simulated results calculated by 341 combinations of {a, T} and b+ with a = 10−8−10−2, b+ = 0.7− 1.0 at intervals of 0.01, T = 800−1300 K at intervals of 50 K, the conditions of {a ≤ 10−5, b+ = 0.95, T = 1200 K} and {a ≤ 10−5, b+ = 0.98, T = 1100 K} show the lowest error sums of squares for Phe/DHB and Ala/DHB, respectively. The results of the simulation reproduce the observed data well, as shown by the solid lines in Figures 6 and 7. The optimized parameters independently determined for Phe/DHB and Ala/DHB show good agreement with each other, supporting the validity of the present analysis. Regarding the parameter a, we are able to determine only its upper limit 10−5 since simulations with a ≤ 10−5 converged within experimental error. For comparison, simulated results at thermal equilibrium (b+ = 1.00) are drawn by broken lines. The differences in behaviors from observation become significant as [A]0/[M]0 increases; i.e., there is no plateau region in ln([AH+]/[MH+])−ln([A]0/[M]0) plots.

5. DISCUSSION 5.1. Parameters of the MALDI Plume. We obtained optimized values, the fraction of protonated and deprotonated matrix, a ≤ 10−5, the extent of thermal equilibrium, b+ = 0.95− 0.98, and the average temperature of proton transfer reactions, T = 1100−1200 K, which reproduce the variation in AH+ and MH+ signal intensities with a sample mixing ratio A0/M0. We discuss their validities by comparing with the values previously reported. Fraction of Ionized Species. The fraction of ionized matrix in the MALDI plume, which affects AH+ yield through proton transfer, has been evaluated using several methods. The fractions of ionized species or ion-to-neutral ratios in the MALDI plume are estimated to be 10−4−10−5 for analyte in diphenylamine/ferulic acid(FA) and diphenylamine/sinapic acid(SA) systems using VUV ionization,14 10−3−10−4 for analyte in the insulin/SA system by evaluation of removed molecules per laser shot,23 10−5 for FA matrix by high-precision weight measurement using a quartz crystal microbalance,24 10−5 for 3-hydroxypicolinic acid matrix by laser-induced fluorescence

6. CONCLUSION We developed MALDI-TOF-MS equipment that can detect neutral species in the MALDI plume using fs-laser ionization while simultaneously collecting a standard MALDI spectrum of the same MALDI event. The MALDI-fs spectra demonstrated that the analyte and matrix are uniformly desorbed in the MALDI plume (i.e., [A]p/[M]p = [A]0/[M]0) for the Phe/ DHB and Ala/DHB systems in the mixture range of [A]0/[M]0 ≤ 1. This is the first experimental evidence to support the validity of the approximation [A]p/[M]p = [A]0/[M]0 in reaction quotients of proton transfer reactions in the MALDI plume. We can thus exclude the possibility that the decrease in reaction quotients at high [A]0/[M]0 when assuming thermal equilibrium is not attributed to failure of the approximation. We introduced a semiequilibrium model to reproduce the observed behaviors of MH+ and AH+ signal intensities. By G


Article


(12) Knochenmusss, R. The Coupled Physical and Chemical Dynamics Model of MALDI. Annu. Rev. Anal. Chem. 2016, 9, 365− 385. (13) Puretzky, A. A.; Geohegan, D. B.; Hurst, G. B.; Buchanan, M. V.; Luk’yanchuk, B. S. Imaging of Vapor Plumes Produced by Matrix Assisted Laser Desorption: A Plume Sharpening Effect. Phys. Rev. Lett. 1999, 83, 444−447. (14) Mowry, C. D.; Johnston, M. V. Simultaneous Detection of Ions and Neutrals Produced by Matrix-Assisted Laser Desorption. Rapid Commun. Mass Spectrom. 1993, 7, 569−575. (15) Tsai, M.-T.; Lee, S.; Lu, I.-C.; Chu, K. Y.; Liang, C.-W.; Lee, C. H.; Lee, Y. T.; Ni, C.-K. Ion-to-Neutral Ratio of 2,5-Dihydroxybenzoic Acid in Matrix-Assisted Laser Desorption/Ionization. Rapid Commun. Mass Spectrom. 2013, 27, 955−963. (16) Lu, I.-C.; Chu, K. Y.; Lin, C.-Y.; Wu, S.-Y.; Dyakov, Y. A.; Chen, J.-L.; Gray-Weale, A.; Lee, Y.-T.; Ni, C.-K. Ion-to-Neutral Ratios and Thermal Proton Transfer in Matrix-Assisted Laser Desorption/ Ionization. J. Am. Soc. Mass Spectrom. 2015, 26, 1242−1251. (17) Ahn, S. H.; Bae, Y. J.; Moon, J. H.; Kim, M. S. Matrix Suppression as a Guideline for Reliable Quantification of Peptides by Matrix-Assisted Laser Desorption Ionization. Anal. Chem. 2013, 85, 8796−8801. (18) Park, K. M.; Bae, Y. J.; Ahn, S. H.; Kim, M. S. A Simple Method for Quantification of Peptides and Proteins by Matrix-Assisted Laser Desorption Ionization Mass Spectrometry. Anal. Chem. 2012, 84, 10332−10337. (19) Tsuge, M.; Hoshina, K. Effect of Esterification on MALDI−MS Detection Sensitivity for Amino Acids. Int. J. Mass Spectrom. 2011, 300, 39−43. (20) DeWitt, M. J.; Levis, R. J. Near-Infrared Femtosecond Photoionization/Dissociation of Cyclic Aromatic Hydrocarbons. J. Chem. Phys. 1995, 102, 8670−8673. (21) Knochenmuss, R. Positive/Negative Ion Ratios and In-Plume Reaction Equilibria in MALDI. Int. J. Mass Spectrom. 2008, 273, 84− 86. (22) Hillenkamp, F.; Wafler, E.; Jecklin, M. C.; Zenobi, R. Positive and Negative Analyte Ion Yield in Matrix-Assisted Laser Desorption/ Ionization Revisited. Int. J. Mass Spectrom. 2009, 285, 114−119. (23) Ens, W.; Mao, Y.; Mayer, F.; Standing, K. G. Properties of Matrix-Assisted Laser Desorption Measurements with a Time-toDigital Converter. Rapid Commun. Mass Spectrom. 1991, 5, 117−123. (24) Quist, A. W.; Huth-Fehre, T.; Sundqvist, B. U. R.; Vertes, A. Total Yield Measurements in Matrix-Assisted Laser Desorption Using a Quartz Crystal Microbalance. Rapid Commun. Mass Spectrom. 1994, 8, 149−154. (25) Puretzky, A. A.; Geohegan, D. B. Gas-Phase Diagnostics and LIF-Imaging of 3-Hydroxypicolinic Acid Maldi-Matrix Plumes. Chem. Phys. Lett. 1998, 286, 425−432. (26) Karas, M.; Bahr, U.; Gieβmann, U. Matrix-Assisted Laser Desorption Ionization Mass Spectrometry. Mass Spectrom. Rev. 1991, 10, 335−357. (27) Bourcier, S.; Bouchonnet, S.; Hoppilliard, Y. Ionization of 2,5dihydroxybenzoic Acid (DHB) Matrix-Assisted Laser Desorption Ionization Experiments and Theoretical Study. Int. J. Mass Spectrom. 2001, 210/211, 59−69. (28) Hsu, H. C.; Lu, I.-C.; Lin, P.-H.; Dyakov, Y. A.; Bagchi, A.; Lin, C.-Y.; Hung, S.-W.; Lee, Y.-T.; Ni, C.-K. Does Decarboxylation Make 2,5-Dihydroxybenzoic Acid Special in Matrix-Assisted Laser Desorption/Ionization? Rapid Commun. Mass Spectrom. 2014, 28, 1082− 1088. (29) Junk, G.; Svec, H. The Mass Spectra of the α-Amino Acids. J. Am. Chem. Soc. 1963, 85, 839−845. (30) Calvert, C. R.; Belshaw, L.; Duffy, M. J.; Kelly, O.; King, R. B.; Smyth, A. G.; Kelly, T. J.; Costello, J. T.; Timson, D. J.; Bryan, W. A.; Kierspel, T.; Rice, P.; Turcu, I. C. E.; Cacho, C. M.; Springate, E.; Williams, I. D.; Greenwood, J. B. LIAD-fs Scheme for Studies of Ultrafast Laser Interactions with Gas Phase Biomolecules. Phys. Chem. Chem. Phys. 2012, 14, 6289−6297.

numerical simulation with three parameters (fraction of ionized species a, extent of thermal equilibrium b+, and temperature T), we successfully reproduced the variation of MH+ and AH+ signal intensities with similar optimized parameter sets a ≤ 10−5, b+ = 0.95, and T = 1200 K for the Phe/DHB system, and a ≤ 10−5, b+ = 0.98, and T = 1100 K for the Ala/DHB system. The parameter b+ = 0.95−0.98 means that the proton transfer approaches thermal equilibrium when the MALDI plume expands to collision-free conditions. The parameters a ≤ 10−5 and T = 1100−1200 K are consistent with those previously determined by different approaches, indicating the validity of the present analyses based on a semiequilibrium model. Applying the present method to MALDI signal analysis and gathering parameters obtained from other species in both positive and negative modes will be a promising approach for quantitatively understanding MALDI processes.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Telephone and fax: +81-25025-5276. ORCID

Kennosuke Hoshina: 0000-0002-0613-1920 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The present research was supported by a Grant-in-Aid for Scientific Research (C) (JSPS KAKENHI Grant Number JP 23550026) from the Japan Society for the Promotion of Science.

■

REFERENCES

(1) Dreisewerd, K. The Desorption Process in MALDI. Chem. Rev. 2003, 103, 395−425. (2) Knochenmuss, R. Ion Formation Mechanisms in UV-MALDI. Analyst 2006, 131, 966−986. (3) Bae, Y. J.; Kim, M. S. A Thermal Mechanism of Ion Formation in MALDI. Annu. Rev. Anal. Chem. 2015, 8, 41−60. (4) Lu, I.-C.; Lee, C.; Lee, Y.-T.; Ni, C.-K. Ionization Mechanism of Matrix-Assisted Laser Desorption/Ionization. Annu. Rev. Anal. Chem. 2015, 8, 21−39. (5) Zenobi, R.; Knochenmuss, R. Ion Formation in MALDI Mass Spectrometry. Mass Spectrom. Rev. 1998, 17, 337−366. (6) Nishikaze, T.; Takayama, M. Cooperative Effect of Factors Governing Molecular Ion Yields in Desorption/Ionization Mass Spectrometry. Rapid Commun. Mass Spectrom. 2006, 20, 376−382. (7) Tsuge, M.; Hoshina, K. Investigation of Protonation Efficiency for Amino Acids in Matrix-Assisted Laser Desorption/Ionization. Bull. Chem. Soc. Jpn. 2010, 83, 1188−1192. (8) Kinsel, G. R.; Yao, D.; Yassin, F. H.; Marynick, D. S. Equilibrium Conditions in Laser-Desorbed Plumes: Thermodynamic Properties of α-Cyano-4-hydroxycinnamic Acid and Protonation of Amino Acids. Eur. J. Mass Spectrom. 2006, 12, 359−367. (9) Chen, X.; Carroll, J. A.; Beavis, R. C. Near-Ultraviolet-Induced Matrix-Assisted Laser Desorption/Ionization as a Function of Wavelength. J. Am. Soc. Mass Spectrom. 1998, 9, 885−891. (10) Niu, S.; Zhang, W.; Chait, B. T. Direct Comparison of Infrared and Ultraviolet Wavelength Matrix-Assisted Laser Desorption/ Ionization Mass Spectrometry of Proteins. J. Am. Soc. Mass Spectrom. 1998, 9, 1−7. (11) Chu, K. Y.; Lee, S.; Tsai, M.-T.; Lu, I.-C.; Dyakov, Y. A.; Lai, Y. H.; Lee, Y.-T.; Ni, C.-K. Thermal Proton Transfer Reactions in Ultraviolet Matrix-Assisted Laser Desorption/Ionization. J. Am. Soc. Mass Spectrom. 2014, 25, 310−318. H


Article

The Journal of Physical Chemistry A (31) Duffy, M. J.; Kelly, O.; Calvert, C. R.; King, R. B.; Belshaw, L.; Kelly, T. J.; Costello, J. T.; Timson, D. J.; Bryan, W. A.; Kierspel, T.; et al. Fragmentation of Neutral Amino acids and Small Peptides by Intense, Femtosecond Laser Pulses. J. Am. Soc. Mass Spectrom. 2013, 24, 1366−1375. (32) Amada, M.; Sato, Y.; Tsuge, M.; Hoshina, K. Near-Infrared Femtosecond Laser Ionization of the Acetic Acid Dimer. Chem. Phys. Lett. 2015, 624, 24−28. (33) O’Hair, R. A. J.; Bowie, J. H.; Gronert, S. Gas Phase Acidities of the α Amino Acids. Int. J. Mass Spectrom. Ion Processes 1992, 117, 23− 36. (34) Breuker, K.; Knochenmuss, R.; Zenobi, R. Gas-Phase Basicities of Deprotonated Matrix-Assisted Laser Desorption/Ionization Matrix Molecules. Int. J. Mass Spectrom. 1999, 184, 25−38. (35) Harrison, A. G. The Gas-Phase Basities and Proton Affinities of Amino Acids and Peptides. Mass Spectrom. Rev. 1997, 16, 201−217. (36) Mormann, M.; Bashir, S.; Derrick, P. J.; Kuck, D. Gas-Phase Basicities of the Isomeric Dihydroxybenzoic Acids and Gas-Phase Acidities of Their Radical Cations. J. Am. Soc. Mass Spectrom. 2000, 11, 544−552. (37) Gabelica, V.; Schulz, E.; Karas, M. Internal Energy Build-Up in Matrix-Assisted Laser Desorption/Ionization. J. Mass Spectrom. 2004, 39, 579−593. (38) Yoon, S. H.; Moon, J. H.; Kim, M. S. A Comparative Study of In- and Post-Source Decays of Peptide and Preformed Ions in MatrixAssisted Laser Desorption Ionization Time-of-Flight Mass Spectrometry: Effective Temperature and Matrix Effect. J. Am. Soc. Mass Spectrom. 2010, 21, 1876−1883. (39) Bae, Y. J.; Moon, J. H.; Kim, M. S. Expansion Cooling in the Matrix Plume Is Under-Recognized in MALDI Mass Spectrometry. J. Am. Soc. Mass Spectrom. 2011, 22, 1070−1078. (40) Moon, J. H.; Yoon, S. H.; Kim, M. S. Temperature of Peptide Ions Generated by Matrix-Assisted Laser Desorption Ionization and Their Dissociation Kinetic Parameters. J. Phys. Chem. B 2009, 113, 2071−2076.

I


Detection of Neutral Species in the MALDI Plume Using Femtosecond

Recommend Documents