New Insights into the Wavelength Dependence of MALDI Mass

Subscriber access provided by EAST TENNESSEE STATE UNIV

Article

New Insights into the Wavelength Dependence of MALDI Mass Spectrometry Marcel Niehaus, Andreas Schnapp, Annika Koch, Jens Soltwisch, and Klaus Dreisewerd Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b01744 • Publication Date (Web): 21 Jun 2017 Downloaded from http://pubs.acs.org on June 25, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

New Insights into the Wavelength Dependence of MALDI Mass Spectrometry

Marcel Niehaus,†,# Andreas Schnapp,†,# Annika Koch,†,# Jens Soltwisch,†,‡ Klaus Dreisewerd†,‡,*

†

Institute for Hygiene, Biomedical Mass Spectrometry, University of Münster, Germany

‡

Interdisciplinary Center for Clinical Research (IZKF), University of Münster, Germany

Key words: MALDI, Matrix, Laser wavelength, Absorption profile, Laser fluence

Running title: Wavelength dependence of MALDI-MS

*

Address correspondence to: Dr. Klaus Dreisewerd, email: [email protected]

#

These authors contributed equally to this study

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 31

Abstract The interplay between the wavelength of the laser and the absorption profile of the matrix constitutes a crucial factor in matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS). Numerous studies have shown that typically best analytical results are obtained if the laser wavelength matches the UV absorption band of the matrix in the solid state well. However, many powerful matrices exhibit peak absorptions which differ notably from the standard MALDI laser wavelengths of 337, 349, and 355 nm, respectively. Here we used two wavelength-tunable lasers to investigate the MALDI wavelength dependence with a selected set of such matrices. We studied 3-hydroxypicolinic acid (3-HPA), 2,4,6trihydroxyacetophenon

(THAP),

dithranol

(1,8-dihydroxy-10H-anthracen-9-on),

2-(4'-

hydroxybenzeneazo)benzoic acid (HABA), and 6-aza-2-thiothymine (ATT). As analyte systems we investigated DNA oligomers (3-HPA), phospholipids (dithranol, THAP, HABA), and non-covalent peptide-peptide and protein-peptide complexes (ATT). We recorded analyte ion and total ion counts as a function of wavelength and laser fluence between 213 and 600 nm. Although the so generated comprehensive heat maps generally corroborated the previously made findings, several fine features became notable. For example, despite of a still high optical absorption exhibited by some of the matrices in the visible wavelength range, ion yields generally dropped strongly, indicating a change in ionization mechanism. Moreover, the non-covalent complexes were optimally detected at wavelengths corresponding to a relatively low optical absorptivity of the ATT matrix, presumably because of ejection of a particular cold MALDI plume. Our comprehensive data shed useful light into the MALDI mechanisms and could assist in further methodological advancement of the technique.


Page 3 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction The desorption/ionization processes in ultraviolet matrix-assisted laser desorption ionization mass spectrometry (UV-MALDI-MS) are critically determined by the physicochemical properties of the used MALDI matrix such as absorption cross section, ionization energies, and proton affinities, on the one hand, and laser irradiation parameters such as wavelength, fluence, and pulse duration, on the other.1-7 The relevance of a sufficient absorption at the laser wavelength was already realized in the early years of the MALDI development and led to the introduction of the two popular matrices 2,5-dihydroxybenzoic acid (DHB) and α-cyano-4-hydroxycinnamic acid (HCCA).8,9 An important factor that needs to be taking into account is that, due to the modified intermolecular forces, absorption profiles in the solution and solid state can differ sizably. A red-shift and broadening of the absorption bands is typically observed for classical MALDI matrices.10,11 A number of studies looked in more detail into the MALDI wavelength dependence. For example, Horneffer et al. investigated six positional DHB isomers and compared discrete laser wavelengths of 266, 308, 337, and 355 nm.12 Good results for the analysis of cytochrome C proteins were consistently obtained only when laser wavelength and individual absorption profiles were overlapping well. For several of the isomers this is ruling out an efficient use of the standard MALDI lasers, namely N2 (337 nm) and frequency-tripled Nd:YLF (349 nm) or Nd:YAG (355 nm) lasers, respectively. Recently, we studied DHB, HCCA and a set of chlorine- and fluorine-substituted derivatives of CCA. Use of a wavelength-tunable dye laser allowed us to apply a narrow wavelength/fluence grid.5,13,14 Our studies revealed that generally the best analytical results were obtained in a few ten nm-wide wavelength window around the individual peak absorption of the matrices in the solid state.5 Taking all studies together, it can be concluded that for the UV-MALDI-MS analysis of peptides and proteins best analytical sensitivities are generally obtained, if the wavelength of 3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 31

the excitation laser corresponds well to the electronic π-π* or n-π* transition bands of the utilized MALDI matrix in the near to middle UV-range. The possible influence of the excitation wavelength in the MALDI process has also been treated in theoretical studies.15,16 Knochenmuss argued that the recorded wavelength dependence could serve to discriminate between pertinent MALDI models. For example, the coupled chemical and physical dynamics (CPCD) model developed by this researcher16,17 predicts a strong dependence of the MALDI ion yields on the efficient excitation of electronic states of the matrix and on pooling processes, whereas other MALDI models such as the polar fluid18,19 and the lucky survivor model20,21 predict more relaxed requirements. Here, the overall energy deposited would be key factors. The number of matrices that have been introduced since the introduction of the MALDI technique in the mid 1980-ies22 must be in the few dozen if not hundreds range. It must be assumed that for a sizable fraction of these compounds, many of which are providing a high analytical potential for certain applications, the absorption profiles will not overlap optimally with the emission wavelengths of the standard MALDI lasers. Therefore, an adjustment of the excitation laser wavelength could in these cases result in an improved analytical performance. To study these possibilities, here we selected five ‘classical matrices’, all of which are exhibiting such non-matching profiles: 6-aza-2-thiothymine (ATT), dithranol (1,8-dihydroxy10H-anthracen-9-on), 2-(4'-hydroxybenzeneazo)benzoic acid (HABA), 3-hydroxypicolinic acid (3-HPA), and 2,4,6-trihydroxyacetophenon (THAP). As analytes we investigated DNA oligomers (3-HPA), phospholipids (dithranol, THAP, HABA), and non-covalent peptidepeptide and protein-peptide complexes (ATT). The examined wavelength range spanned from 213 to 600 nm, thus spanning the UV as well as a part of the visible (VIS) light spectrum. Next to the wavelength, we varied the laser fluence from the ion detection threshold to close 4 ACS Paragon Plus Environment

Page 5 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


to one order of magnitude above. In this way, comprehensive ‘heat maps’ were produced that provide an illustrative overview of the complex signal intensity-wavelength-fluence relationship and enable a straightforward comparison with the solid-state absorption profiles of the compounds.

Experimental Materials Biotinyl-glucagon, human gastrin I (HGI), kemptide, and protein kinase C substrate (PKCS) were from Bachem (Bubendorf, Switzerland), organic solvents from Roth (Karlsruhe, Germany), and dithranol from Seratec (Courville-sur-Eure, France); all other chemicals were from Sigma-Aldrich (Schnelldorf, Germany).

Sample Preparation Solvent systems and molar analyte-to-matrix ratios were optimized for each matrix/analyte system as listed in Table S-1 of the Supporting Information. Analyte and matrix solutions were premixed before spotting 1 µL of solution on the stainless steel sample plate, except for HABA were solutions were mixed ‘on-target’ to improve the co-crystallization. Droplets were allowed to dry at room temperature.

Spectrophotometry Absorption spectra were recorded using a dual beam spectrophotometer (UV-2102PC, Shimadzu, Duisburg, Germany). To record solution phase spectra, matrices were prepared to 10-4 mol/L in deionized water (ATT, 3-HPA, HABA, THAP) or in chloroform:methanol (1:1, v:v; dithranol). Solid state absorption spectra were recorded in diffusive reflection geometry 5 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 31

as described.5 An important prerequisite for recording artefact-free profiles is choosing a suitable analyte:BaSO4 ratio (to obtain sufficient absorption while at the same avoiding saturation effects) and to grind the mixtures thoroughly (to minimize light scattering at crystallite surfaces). We used a dilution series to identify the optimal analyte:BaSO4 ratio and used extensive milling in a ball mill until scattering effects (which lead to a characteristic tailing at the long wavelength side and overall broadening of absorption bands) were minimized (i.e., any further grinding did not result in changes to the profiles). Absolute absorption coefficients α are more difficult to derive by the diffusive reflection geometry and would require measuring a series of differently diluted samples.11 To avoid this complication all data are presented as relative values normalized to the peak absorption. Moreover, the mean laser penetration depth δ = α-1 is plotted in the heat maps. With respect to absolution absorption values previous work has shown that the peak values as determined in solution can serve as a good approximation for the corresponding solid state peak value.10,11

Mass Spectrometer and Laser Systems The prototype orthogonal-extracting time-of-flight mass spectrometer (oTOF-MS), equipped with a modified oMALDI 2TM ion source (AB Sciex, Concord, Canada), has been described elsewhere.23 Except for the non-covalent complexes, for which 1.4 mbar were used to enhance collisional cooling effects, all measurements were performed with a buffer gas pressure (N2) in the region of ion generation of 0.7 mbar. In the experiments with the 3-HPA matrix, a dye laser (FL-2001; Lambda Physik, Göttingen, Germany) providing pulses of ~7 ns duration at a repetition rate frep of 10 Hz was used. The output beam was coupled to the oMALD 2 source via fiber optics and hit the sample at an angle of incidence of 45°.5 A wavelength-tunable optical parametric oscillator (OPO) laser system (versaScan ULD, GWU-Lasertechnik, Erftstadt, Germany; pulse duration, 6 ACS Paragon Plus Environment

Page 7 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


6-7 ns; frep, 10 Hz) was utilized for all other experiments. The OPO output was coupled as a free laser beam using one of two available ports of the oMALDI 2 source. For the experiments with HABA the angle of incidence was 45° for all other matrices it was 60°. To reduce the beam divergence and as a means for compensation of dispersion effects, a telescope was used in all OPO-based experiments. Beam attenuation was achieved with a circular neutral gradient filter (OptoSigma, Santa Ana, CA). Owing to the slightly differing focusing units used and different wavelength ranges covered, the focal spot sizes selected in a particular experiment were between 150 x 300 and 350 x 700 µm2 (see Table S-1 for details). Great care was taken that the focal laser spot size used for a specific matrix was kept as constant as possible. For free beam delivery intensity cross sections were monitored online using a beam splitter and a linear diode array with 15 µm pixel-to-pixel distance (Reticon, Sunnyvale, CA).13 Because the profiles could be fitted reasonably well by Gaussian functions, vertical and horizontal diameters were defined by the 1/e2-definition. The focal diameter of the fiber-delivered laser beam was determined using a knife-edge method and applying the so-called D86 definition.24 Laser pulse energies were monitored on-line using a second beam splitter and a calibrated pyroelectric detector.13 The shot-to-shot pulse energy stabilities of the laser outputs were about 10-15%. All reported Fluences F correspond to the laser pulse energies delivered to the sample divided by the laser spot area A. Mass spectra were acquired in the positive ion mode. The lower cut-off of the transfer quadrupole of the oTOF instrument was set to m/z 130-150, except for detection of the DNA and non-covalent complexes, where it was set to 800 and 3000, respectively. Mass spectra were recorded using tofmulti software (courtesy of V. Spicer and W. Ens, Univ. of Manitoba) with modifications made in-house. Spectra were acquired by irradiating the samples with 300



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 31

or 600 (ATT matrix only) laser pulses for each data point. Care was taken that mostly ‘fresh’ samples areas were irradiated to record each mass spectrum.

MS Data Analysis A peak list was used to define ions of interest. If, based on this list, an ion signal was reported by the tofmulti software, this signal was integrated over a full width at quarter maximum area and centroid intensity results reported to a table.

Heat Maps Signal-intensity/wavelength/fluence contour plots were generated using OriginPro 2016 (OriginLab, Northampton, MA, USA). On the basis of the experimental wavelength/fluence grid (black dots in the heat maps) the program interpolates the signal intensities. The so produced false color-encoded ‘heat maps’ provide an intuitive presentation of the complex data set, however some care should be taken in the data interpretation. For instance, in some of the heat maps ‘intensity islands’ are notable for individual wavelength/fluence data points. Generally, such ‘outliers’ are due to the experimental error (e.g., scatter in ion abundances) and not representing a true physical effect.

Safety Hazard Note Protective goggles have to be worn when working with free beams of class 4 lasers.

Results 8 ACS Paragon Plus Environment

Page 9 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Absorption Profiles in the Liquid and Solid State Absorption profiles of the five matrices in the liquid and solid state, respectively, are plotted in Figure 1. Wavelengths of peak absorption and molar decadic extinction coefficients at these wavelengths (determined in solution) are listed in Table S-2. Comparing the two aggregation states, a red-shift of the peak absorption values in the solid state by 10-20 nm and, typically, some peak tailing towards longer wavelengths is notable for all compounds. The latter is particular strong for HABA and dithranol, providing a sizable absorption in the visible part of the light spectrum.

3-HPA Matrix 3-HPA is the classical matrix for MALDI mass spectrometry of nucleotide polymers (DNA and RNA).25 Although frequency-tripled (355 nm) and quadrupled (266 nm) Nd:YAG lasers were initially used as light source,25,26 it was realized early on that the absorption profile of 3HPA lends itself more to the use of the 337 nm line of N2 laser.27 Our spectrophotometric measurements place the peak absorption of 3-HPA in the solid state at 320 nm, in the liquid phase it is located at 295 nm (black lines in Figure 1). The absorption band within which at least 50% of the peak absorption is obtained ranges in the solid state from about 290 to 350 nm. Using four DNA oligomers as analyte system, we scanned this range with 5 nm steps and recorded mass spectra at a narrowly placed set of fluence points. A heat map displaying the summed molecular ion signals ([M + H]+, [M + Na]+, [M + K]+, multiple adducts) is shown in Figure 2a. The lowest ion detection threshold fluence is found at 305 nm. To both shorter and longer wavelengths, the ion detection threshold fluence rises monotonously. The small island of high signal intensity discernible at the 305 nm data point and low laser fluences can presumably be attributed to the experimental 9 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 31

uncertainty. Several sample spots were irradiated and 3-HPA is moreover known to produce particularly pronounced “sweet spot effects”. At all wavelength points, DNA signals rise exponentially with fluence before a plateau region is reached, and molecular ion signals eventually start to decrease due to increasing fragmentation reactions at even higher fluence values. Except for the island point at 350 nm, maximum ion counts are obtained at about three times the ion detection threshold This finding reflects the general MALDI characteristics recorded in numerous studies.2,28,29 ‘Individual’ heat maps for the summed [M + H]+ and alkali metal adduct ions, respectively are plotted in Figure S-2). These heat maps point to the occurrence of different ionization routes. Notably, ion yields for the protonated molecules are in tendency stronger toward smaller wavelengths and low to medium laser fluence areas, whereas alkali metal adducts are in tendency produced predominantly toward higher wavelengths and if the fluence is raised to (MALDI-untypically) high values. More detailed follow-up studies with different matrix systems are however needed to further study these relevant aspects in detail. Next to the molecular ion counts, also the total ion count (TIC) can serve as a measure to probe the ionization efficiency. The heat map of the 3-HPA-derived TIC (Figure S-1a) all in all resembles those of the combined analyte ion counts quite closely. Naturally, a fragmentation-related signal drop at high fluence values is not found in the TIC. Another measure providing insight into the underlying mechanisms is the ratio between the analyte ion count and the TIC. For simplicity, we define this ratio here as ‘analyte ion yield (AY)’. Figure 2b reveals that the overall produced charges are transferred to intact DNA molecules particularly efficiently at low laser fluences. An area of maximized AY is achieved between 290 and 320 nm. However, in these regions relatively low total ion counts are generated and some care should be exercised in interpreting these findings.


Page 11 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Three mass spectra acquired at 305 nm (wavelength of lowest ion detection threshold), 320 nm (wavelength of maximum absorption), and 340 nm (close to the standard MALDI wavelength of 337 nm), are plotted in Figure S-3. All spectra were recorded at individually optimized fluences, corresponding to a maximized AY. Qualitatively similar mass spectra were obtained at the three wavelength points. However, a (slightly) improved signal (within a factor of 2) is obtained if the matrix absorbs the laser energy most efficiently. In conclusion, our data show that upon fluence adjustment mass spectra of similar quality can be obtained with the 3-HPA matrix over a wide wavelength range from about 290 to 340 nm. Use of an adjusted wavelength (e.g., 305 nm) would provide an approximate improvement by a factor of 2 to 3 compared to 337 nm. These findings corroborate previous observations,25,27 but for the first time put them in the context of a systematically recorded, comprehensive ion signal/wavelength/fluence data set.

THAP Matrix THAP was first introduced for the analysis of DNA and RNA,30 but is today utilized mostly for MALDI mass spectrometry of lipids.31 The spectrophotometric measurements place the peak absorption of THAP at 280 nm in solution and at 300 nm in the solid state (blue lines in Figure 1a,b). The solid state profile shows a considerable tailing to both the long and short wavelength side. Compared to the peak value at 300 nm, at 337 nm the absorption is reduced by about 25%. We used phosphatidic acid (PA(18:1/18:1)) and phosphatidylglycerol (PG(18:1/18:1)) as analytes for a wavelength study spanning from 220 to 340 nm. A step size of 10 nm (20 nm along the flanks of the absorption profiles) was used. A heat map representing the summed intensity of all alkali metal adduct ions of both phospholipids is shown in Figure 3a. 11 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 31

Maximum analyte ion signals are found in a broad wavelength range from 260 to 320 nm. Despite of a similar absorption value, analyte ion counts are sizably reduced at the shortest tested wavelength point of 220 nm. In contrast, the TIC shows a more symmetric behavior (Figure S-1b), suggesting that the loss in molecular ion signals could be due to increased fragmentation reactions. Another explanation would be photodynamic reactions and loss of matrix ions into products that do not contribute to subsequent analyte ionization. The AY maximizes along the long wavelength side up to 340 nm (Figure 3b). Comparing the MS data with the absorption profiles of THAP reveals an overall close correlation. Notably, the local minimum in optical absorption that is found at 240 nm for both aggregation states is reflected well by the MS data. Three mass spectra that were recorded at 240, 310 and 340 nm, corresponding to wavelengths of minimum and maximum solid state absorption, and to about the standard N2 laser wavelength, respectively, are plotted in Figure S-4. The spectra demonstrate that, apart for the different total signal intensities, qualitatively similar ion profiles are produced within the probed wavelength range. In conclusion, our data show that an improved MALDI-MS performance for the analysis of lipids with the THAP matrix can be achieved upon wavelength adjustment. Optimal results are found between 280 and 320 nm. Utilization of the N2 laser wavelength would require elevated laser fluences and this would, for the given test system, not fully rescue the analyte ion counts.

Dithranol Matrix Dithranol was first introduced as a matrix for the analysis of technical polymers.32 Due to its ability to ionize lipids in both the positive and negative ion modes and straightforward to 12 ACS Paragon Plus Environment

Page 13 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


obtain micro-crystalline sample morphologies, the compound has recently found a high interest for MALDI-MS imaging applications.33,34 In the liquid state, dithranol exhibits a strongly modulated absorption profile with three distinct maxima at 255, 285, and 355 nm (red lines in Figure 1a,b). In the solid state, two distinct absorption bands of comparable amplitude are found, centering around 260 and 375 nm, respectively. In addition, a high absorptivity is also found around 220 nm. To investigate wavelength-dependent fine effects that are possibly associated with these features, we applied a step size of 20 nm over a total range from 220 to 440 nm. The same PG and PA mix system as above was used as analyte system. The heat map for the analyte ion counts (sum of all alkali metal adduct-derived PG and PA ions) is presented in Figure 4a. The two clearly discernible regions of maximal ion counts correspond well to the two absorption maxima in the solid state (solid grey line in the figure). At 220 nm, high ion counts are only obtained for the TIC (Figure S-1c). Like for the THAP matrix, the AY shows a broadened plateau region (Figure 4b). As a consequence, the ‘dip’ around the absorption minimum at 310 nm, which is clearly resolved in total analyte and overall ion counts, is obliterated. Another notable finding is the precipitous ‘loss’ of ion signals for wavelengths in excess of 400-420 nm, despite of a still sizable absorptivity. Or in other words, the gradual reduction in absorption cross section can here not be compensated by raising the laser fluence. Similar observations were already made by Chen et al. in their classical study with a set of cinnamic acid derivatives.18 Mass spectra recorded at 220, 260, 340, and 380 nm, corresponding to the local absorption maxima and to about the N2 laser emission line, respectively, are displayed in Figure S-5. At 220 nm, an elevated level of background signals is found in the low m/z range, at the other wavelength points qualitatively rather similar mass spectra are obtained.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 31

In conclusion, our data show that dithranol is useful for MALDI mass spectrometry over a particular wide wavelength range. Optimal results are found around the local absorption maxima of 260 and 380 nm, however N2 and frequency-tripled neodymium-doped glass lasers would be almost equally well usable.

HABA Matrix HABA was introduced as a MALDI matrix in 1993 and has been identified as particular useful for the analysis for larger proteins, in particular glycoproteins.35 This could indicate a particular softness of the MALDI process with this compound. Several studies also described the applicability of HABA for further compound classes including technical polymers.353535,36 From all compounds investigated in our study, HABA exhibits the largest change in its absorption profile upon phase transition. In the liquid state, a broad absorption band is centering around 350 nm (grey line in Figure 1a) with some tailing extending towards 500 nm and above. Due to the conjugation between the aromatic ring systems linked by an azo group, aromatic azo compounds generally exhibit a high intensity π–π* absorption band in the UV range, in case of HABA maximizing at 350 nm.37 It is also known that the UV-VIS absorption spectra of trans-azobenzenes are characterized by a band in the region of 440-480 nm originating from symmetry-forbidden n–π* transitions occurring at the central nitrogen atoms.38 Moreover, there is evidence of participation of the azo-hydrazone tautomerism.39 The additional band in the solid state centering at 480 nm (grey line in Figure 1b) is attributed to the protonated form in the prevailed equilibrium of azo- and hydrazone-form of the molecule. All effects together produce a very broad absorption band in the solid state that spans from 360 nm to 500 nm, with sizable additional tailing towards 300 nm and 600 nm, respectively.


Formatted: Supe

Page 15 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


We covered this range with a step size of 20 nm and by use of phosphatidylcholine PC(16:0/16:0) as analyte. Three mass spectra, acquired at wavelengths of 340, 400, and 480 nm, are presented in Figure S-6. Despite of using only one analyte species, the spectra are highly complex. Next to protonated monomeric PC molecules and their sodium and potassium adducts, PC dimers are detected in high abundance, as well as matrix adducts formed with both monomers and dimers. Interestingly, matrix adducts are only found if potassium serves as the charge carrier. This finding can presumably be explained by a particular high affinity of HABA towards K+. Furthermore, also a series of fragment ions is detected, including species resulting from the neutral loss of a 16:0 fatty acid residue and that of the phosphorylcholine head group (HG). Comparing the three mass spectra, strongly increased signal levels for all described ion species are notable in the 400 nm data. Also, the matrix ion background is clearly elevated. At the other two wavelength points, the spectra show elevated levels of HG loss (340 nm) and of dimer formation (480 nm), respectively. In contrast to the 400 nmspectrum, matrix ions are essentially absent at the latter wavelength points. To further characterize these interesting wavelength-dependent fine features, all identified ion signals were assembled in a data evaluation table. Heat maps representing the analyte ion counts of molecular PC ions (including molecular monomeric and dimeric ions, adduct species and fragments) are plotted in Figure 5a. These data confirm that the highest ion counts were obtained in a comparatively narrow wavelength region between about 380 to 440 nm, only. The same result is obtained for the TIC (Figure S-1d). This wavelength window would correspond to absorption by both the azo- and hydrazone-form of HABA. For wavelengths above ~460 nm, a precipitous increase in the laser fluence (by a factor of ≥3.5) was required to generate useful mass spectra. The increased energy deposition was accompanied by visible material ablation per laser pulse. A heat map displaying the relative abundance ratio of dimeric ions, normalized to all analyte-derived ion counts, is shown in 15 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 31

Figure 5b. These data further evidence the change in desorption/ionization mechanisms occurring between about 440 to 460 nm. Essentially a mirror image is obtained if all fragment-derived ion signals are treated in the same way (Figure 5c). [M + H]+, [M + Na]+, and [M + K]+ signals of PCs produced with the HABA matrix showed only a minor tendency toward an enhanced generation of alkali metal adducts for longer wavelengths (data not shown), ruling out that the reduced fragmentation rate is a result of a lower relative abundance of less stable protonated PC molecules. In conclusion, our data reflect the complex photodynamic properties of HABA. Optimal MALDI ion yields are obtained between about 380 nm and 440 nm, which would coincide with excitation of both the hydrazone- and azo-configurations. N2 and frequencytripled Nd:YAG lasers are suitable for MALDI-MS with this matrix, although they are not the optimal choice for achieving a maximized sensitivity. Use of visible laser light resulted in a change in the MALDI mechanisms and a reduced fragmentation yield for the tested PC ions.

Analysis of Non-Covalent Peptide-Peptide and Peptide-Protein Complexes with the ATT Matrix Owing to its near-neutral pH of 5.4, ATT is regarded as a matrix of choice for the MALDIMS analysis of non-covalent complexes.40,41 Our spectrophotometric measurements place the peak absorption of ATT in both the solution and solid state at 265 nm (green lines in Figure 1a,b). At 337 nm, the absorptivity in the solid state is reduced by about 50%. ATT is thus a particular striking example of a MALDI matrix where one of the standard laser wavelengths deviates by as much as 72 nm from the peak absorption. An obvious question to ask is whether this reduction constitutes a


Page 17 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


disadvantage (because of potentially reduced ion yields) or whether it could not rather provide the basis for its analytical success. To investigate these considerations, we started out by using two peptide-peptide systems as previously analyzed by Zehl and Allmaier.40 Specifically, we used a 1:1:1-molar mixture of human gastrin I (HGI; Pyr-GPWLEEEEEAYGW) and its potential binding partners kemptide (LRRASLG) and protein kinase C substrate (PKCS, VRKRTLRRL); the binding motifs are underlined. Our wavelength study covered the range from 220 to 390 nm in steps of 10 nm; in addition, the minimum wavelength setting of the OPO laser of 213 nm was included. A mass spectrum recorded at 240 nm, which is providing the optimum for the detection of the intact complexes (see below), is shown in Figure 6a. At this wavelength, both HGI-kemptide and HGI-PKCS complexes are detected, along with unbound monomeric peptide units and a few peptide-matrix adducts. Control experiments (e.g., by adding nonbinding peptides and acidifying the analyte solution) corroborated the specificity of the binding and ruled out that the observed ions are the result of unspecific gas phase processes (data not shown). The most striking evidence for the specificity is given by the wavelength/fluence dependence of molecular ion signals (Figure 6b). Whereas the highest [M + H]+ signals of the unbound peptides are produced within the main absorption band of ATT in the solid state (Figure 6c), for the peptide-peptide complexes a vastly different scenario is found. These ions are detected best when the laser excitation wavelength corresponds to a sizably reduced optical absorptivity of the ATT matrix (by about 50%), namely between 220-250 nm and, in a small(er) range, around 330 nm. We hypothesize that the correspondingly higher laser penetration depths (displayed by solid grey line in Figure 6b,c) generate softer material ejection conditions, which could for 17 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 31

instance resemble those observed in infrared (IR-)MALDI.42 At the same time, a still sufficiently high ‘electronic’ absorption of the matrix seems to form a pre-requisite for the necessary protonation step in MALDI. An optimal compromise appears to be achieved if the ATT absorption corresponds to about 50% of its peak value. Notably, under these conditions even the use of very short 220-230 nm wavelengths did not quantitatively dissociate the peptide complexes. The absorption profile of the complex (determined in solution) is depicted in Figure 6b,c as a dashed grey line. Although the heat map for the TIC (Figure S-1e) reflects that for the unbound peptides, a slight red-shift of the latter (reflecting a relatively increased analyte ion signal above 290 nm) is visible. Next, we studied a streptavidin-biotinyl-glucagon-complex; biotinyl-glucagon (BG) exhibiting a molecular weight (MW) of 3,709 g/mol instead of biotin (MW, 244.3 g/mol) was chosen to obtain a larger mass increment. Streptavidin (MW, 52 kg/mol)43 consists of 4 identical subunits (SU; MW 13 kg/mol), each of which can bind one biotinyl-glucagon ligand.44 Binding of the ligand stabilizes the binding between the subunits.45 A mass spectrum acquired at 320 nm shows signals of the SU-BG complex with sizable abundance (Figure S-7a). The also observed SU-BG2 complex can be explained by the binding position between the subunits, so that biotin can bind to the SU from two different directions. In addition, also weak signals of SU2-BG and SU2-BG2 are found. Trimers or the full tetramer were not detected under our conditions but have been detected in electrospray ionization (ESI-)MS studies.43 Heat maps for the SU-BG complex and for unbound SU, respectively, are shown in Figure S-7b. Different to the peptide-peptide complexes above, the protein-peptide complexes are optimally detected only along the long wavelength shoulder of the ATT absorption band. At 240 nm, only a low signal level is obtained for low laser fluences. Signals for unbound SU are also found at 250 and 260 nm but again with lower intensity than in the higher wavelength 18 ACS Paragon Plus Environment

Page 19 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


range. This finding can be rationalized by direct photodissociation of the protein. The absorption profile of streptavidin (recorded in solution) is presented as a dashed grey line in Figure S-7b,c. In conclusion, MALDI-MS of non-covalent complexes with an ATT matrix is best performed with laser wavelengths that are corresponding to a reduced optical absorption of the matrix. At the same time a sufficient electronic excitation is necessary to ensure a sufficient ionization mechanism. An optimal is found at about 50% of the peak absorption value. This renders the use of N2 lasers (337 nm) particular well suitable whereas the 355 nmline of the Nd:YAG laser falls outside of the optimal range.

Discussion In line with previously reported wavelength studies our data corroborate the relevance of the laser excitation conditions in the MALDI process. On the basis of the comprehensive data set essentially three wavelength ranges can be differentiated. MALDI with wavelengths below about 240-250 nm appears to be generally accompanied by reduced analyte ion counts, even if the optical absorption is high. This could be either caused by an increased analyte fragmentation or by photochemical side reactions preventing efficient charge transfer. A notable exception is given by the analysis of the peptide-peptide complexes with ATT for which particular high ion yields are found between 230-240 nm. We attribute this finding to a change in material ejection mechanism, beneficial for detection of the weakly bound complexes and outperforming a possible gradual inset of direct photodissociation. Within the middle to near UV-range (and for some matrices extending slightly into the visible light range), the best performance is generally found if the laser wavelength matches the π-π* and n- π* absorption bands of the specific matrices. In some cases (like for dithranol), 19 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 31

the TIC almost exactly follows the solid state absorption, in others (like for 3-HPA) a mixture between the liquid and solid state absorption profiles is observed. Generally, the areas of relatively high analyte ion counts extend to somewhat higher wavelengths than those for the TIC. Knochenmuss previously accounted this difference to the increased number of photons being available at the same fluence and, all in all, more enhanced secondary ionization processes.15,16 To some extent (roughly up to a 50% reduction) a decreased absorptivity can in the main UV wavelength range typically be compensated for by raising the laser fluence. This feature relaxes experimental requirements and, importantly, allows for the utilization of N2 (and in some cases frequency-tripled Nd:YAG or Nd:YLF) lasers, even if their emission lines may be off the peak absorption. In the visible wavelength range, ion yields typically show a strong decline, even if the optical absorption is still high and also material ejection is visibly occurring. Here, the loss in absorptivity typically cannot be compensated for by raising the fluence. This suggests that the population of certain (higher energetic) electronic states of the matrix is required for efficient ionization. A certain exception is given by the HABA matrix with its peculiar electronic configurations. With this compound sizable analyte ion counts could even be generated in excess of 500 nm. Our measurements with HABA constitutes one of the few examples of successful VIS-MALDI reported in the literature.46,47 At the same time, despite of its broad absorption profile highest ion counts are obtained in a comparatively narrow wavelength window between 380-420 nm. This is indicative of the relevance of the actual electronic configuration of the matrix and its excitation in the overall MALDI processes. The studied wavelength ranges included experimentally determined 2-photon ionization thresholds for THAP and dithranol at 295 nm (hν = 4.22 eV) and 305 nm (4.09 eV), respectively, and the 3-photon threshold for HABA at 450 nm (2.77 eV); all data are for free molecules.48 The 2-photon threshold for ATT can be estimated to be close to 4.2 eV, such that 20 ACS Paragon Plus Environment

Page 21 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


this data point would also be included in the respective wavelength scan. All thresholds are denoted in Figure S-1 by dashed vertical lines. The visible drop in the TIC found for the four matrices and wavelength in excess of the threshold values points to the relevance of efficient electronic excitation in the ionization processes. Our combined data demonstrate the high complexity of the MALDI process and that different scenarios, presumably including molecular desorption at high absorption/low fluence conditions and more sizable material ejection at high fluence/low absorptivity conditions,13,49 can generate comparable mass spectra. A multitude of secondary reactions in the MALDI plume will generally drive the system into a final state (as visible in the mass spectra) that is typically dominated by the thermodynamic properties of the material such as proton affinities. For several of the here studied analyte/matrix systems increased analyte ion counts were obtained upon wavelength adjustment. However, only samples with fixed analyte matrix ratios were so far investigated. A follow up study, currently under way in the authors’ laboratory, will explore the limits of detection (as the analytically most crucial parameter) in dependence on the laser wavelength. Next to using a wavelength-tunable laser—which however comes with limitations on the pulse repetition rate and also with higher maintenance requirements compared to standard MALDI lasers—a compensation for the wavelength mismatch might be achieved by use of binary matrix systems where the second component would be highly absorbing at one of the standard MALDI wavelengths. These results will be reported elsewhere. Another question to ask is how much the wavelength-dependent MALDI mechanisms under the here applied collisional cooling conditions might differ from MALDI performed with the more classical axial-TOF instruments where the ion source is run at high vacuum. For example, it is generally known that MALDI o-TOF instruments enable the use of much 21 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 31

higher laser fluences before a loss in the molecular ion signals sets in. Follow-up experiments with both o-TOF and axial-TOF geometries to look into this matter are currently also under way and are to be reported elsewhere. These studies will in particular look into the internal energy of the generated MALDI plumes using thermometer molecules.

Conclusion In amendment to previous wavelength studies we studied analyte systems other than standard peptides and proteins and selected matrices with proven MALDI performance but notably ‘non-matching’ absorption profiles. Moreover, a particular wide wavelength and laser fluence range was covered, allowing for a relatively direct comparison of the wavelength- and fluence-dependent MS performance with the absorption profiles of the MALDI matrices. The data could indicate ways for further advancement of the MALDI method and serve to question or corroborate possible mechanistic routes, an exciting field that could not be covered in the present article due to space restrictions.

Acknowledgements We are thankful to GWU-Lasertechnik, U. Röhling, T. Deilmann and M. Wiese for technical assistance. Financial support by the German Research Foundation (grants DR 416/8-1 and DR 416/8-2) is gratefully acknowledged.

References


Page 23 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. Kochenmuss, R.; Zenobi, R. Chem. Rev. 2003, 103, 441–452. 2. Dreisewerd, K. Chem. Rev. 2003, 103, 395-426. 3. Menzel, C; Dreisewerd, K.; Berkenkamp, S.; Hillenkamp, F. J. Am. Soc. Mass Spectrom. 2002, 13, 975-984. 4. Jaskolla, T. W.; Lehmann, W.-D.; Karas, M. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 12200-12205. 5. Soltwisch, J.; Jaskolla, T. W.; Hillenkamp, F.; Karas, M.; Dreisewerd, K. Anal. Chem. 2012, 84, 6567-6576. 6. Soltwisch, J.; Kettling, H.; Vens-Cappell, S.; Wiegelmann, M.; Müthing, J.; Dreisewerd, K. Science 2015, 348, 211-215. 7. Koch, A.; Schnapp, A.; Soltwisch, J.; Dreisewerd, K. Int. J. Mass Spectrom., in press, DOI: 10.1016/j.ijms.2016.11.007 8. Beavis, R. C.; Chaudhary, T.; Chait, B. T. Org. Mass Spectrom. 1992, 27, 156-158. 9. Strupat, K.; Karas, M.; Hillenkamp, F. Int. J. Mass Spectrom. Ion Processes 1991, 111, 89102. 10. Heise, T. W.; Yeung, E. S. Anal. Chim. Acta 1995, 299, 377.385. 11. Lai, Y.-H.; Wang, C.-C.; Wen Chen, C.; Liu, B.-H.; Hsien Lin, S.; Tseh Lee, Y.; Wang, Y.-S. J. Phys. Chem. B 2012, 116, 9635-9643 12. Horneffer, V.; Dreisewerd, K.; Lüdemann, H.-C.; Hillenkamp, F.; Läge, M.; Strupat, K. Int. J. Mass Spectrom. 1999, 185, 859-870. 13. Soltwisch, J.; Jaskolla, T. W.; Dreisewerd, K. J. Am. Soc. Mass Spectrom. 2013, 24, 14771488. 14. Wiegelmann, M.; Soltwisch, J.; Jaskolla, T. W.; Dreisewerd, K. Anal. Bioanal. Chem. 2013, 405, 6925-6932. 15. Knochenmuss, R. Analyst 2014, 139, 147-156. 23 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 31

16. Knochenmuss, R. Annu. Rev. Anal. Chem. 2016, 9, 365-385. 17. Knochenmuss, R. J. Mass Spectrom. 2013, 48, 998-1004. 18. Chen, X.; Carroll, J. A.; Beavis, R. C. J. Am. Soc. Mass Spectrom. 1998, 9, 885–891. 19. Chu, K. Y.; Lee, S.; Tsai, M.-T.; Lu, I-C.; Dyakov, Y. A.; Lai, Y. H.; Lee, Y.-T.; Ni, C.K. J. Am. Soc. Mass Spectrom. 2014, 25, 310-318. 20. Jaskolla, T. W.; Karas, M. J. Am. Soc. Mass Spectrom. 2011, 22, 976-988. 21. Krüger, R.; Karas, M. Chem. Rev. 2003, 103, 427–439 22. Karas, M.; Bachmann, D.; Bahr, U.; Hillenkamp, F. Int. J. Mass Spectrom. Ion Processes 1987, 78, 53-68. 23. Soltwisch, J.; Souady, J.; Berkenkamp, S.; Dreisewerd, K. Anal. Chem. 2009, 81, 29212934. 24. Soltwisch, J.; Dreisewerd, K. Rapid Commun. Mass Spectrom. 2011, 25, 1266-1270. 25. Wu, K. J.; Steding, A.; Becker, C. H. Rapid Commun. Mass Spectrom. 1993, 7, 142-146. 26. Liu, Y.-H.; Bai, J.; Zhu, Y.; Liang, X.; Lubman, D. M.; Siemieniak, D.; Venta, P. J. Rapid Commun. Mass Spectrom. 1995, 9, 735-743. 27. Little, D. P.; Cornish, T. J.; O’Donnell, J. O.; Braun, A.; Cotter, R. J.; Köster, H. Anal. Chem. 1997, 69, 4540-4546. 28. Dreisewerd, K.; Schürenberg, M.; Karas, M.; Hillenkamp, F. Int. J. Mass Spectrom. Ion Processes 1995, 141, 127-148. 29. Rohlfing, A.; Leisner, A.; Hillenkamp, F.; Dreisewerd, K. J. Phys. Chem. C 2010, 114, 5367-5381. 30. Pieles, U.; Zürcher, W.; Schär, M.; Moser, H. Nucleic Acids Res. 1993, 21, 3191-3196. 31. Stübiger, G.; Belgacem, O. Anal. Chem. 2007, 79, 3206-3213. 32. Montaudo, G.; Samperi, F.; Montaudo, M. S. Prog. Polym. Sci. 2006, 31, 277-357. 33. Le, C. H.; Han, J.; Borchers, C. H. Anal. Chem. 2012, 84, 8391-8398. 24 ACS Paragon Plus Environment

Page 25 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


34. Kettling, H.; Vens-Cappell, S.; Soltwisch, J.; Pirkl, A.; Haier, J.; Müthing, J.; Dreisewerd, K. Anal. Chem. 2014, 86, 7798-7805. 35. Juhasz, P.; Costello, C. E.; Biemann, K. J. Am. Soc. Mass Spectrom. 1993, 4, 399-409. 36. Soeriyadi, A. H.; Whittaker, M. R.; Boyer, C.; Davis, T. P. J. Polym. Sci., Part A: Polym. Chem. 2013, 51, 1475-1505. 37. Green, N. M. Biochem. J. 1965, 94, 23c-24c. 38. Hamon, F. Djedaini-Pilard, F.; Barbot, F.; Len, C. Tetrahedron 2009, 65, 10105-10123. 39. Cinar, M.; Yildiz, N.; Karabacak, M.; Kurt, M. Spectrochim. Acta, Part A 2013, 105, 8087. 40. Zehl, M.; Allmaier, G. Rapid Commun. Mass Spectrom. 2003, 17, 1931-1940. 41. Woods, A. S.; Huestis, M. A. J. Am. Soc. Mass Spectrom. 2001, 12, 88-96. 42. Dybvik, A. I.; Norberg, A. L.; Schute, V.; Soltwisch, J.; Peter-Katalinić, J.; Vårum, K. M.; Eijsink, V. G. H.; Dreisewerd, K.; Mormann, M.; Sørlie, M. Anal. Chem. 2011, 83, 40304036. 43. Eckart, K.; Spiess, J. J. Am. Soc. Mass Spectrom. 1995, 6, 912-919. 44. Chaiet, L.; Wolf, F. J. Arch. Biochem. Biophys. 1964, 106, 1-5. 45. Sano, T.; Pandori, M. W.; Smith, C. L.; Cantor, C. R. In Advances in biomagnetic separation; Uhlén, M.; Hornes, E.; Olsvik, Ø., Eds: Natick, 1994; pp 21-29. 46. Hu, X. K.; Lacey, D.; Li, J.; Yang, C.; Loboda, A. V.; Lipson, R. H. Int. J. Mass Spectrom. 2008, 278, 69-74. 47. Tang, K; Allman, S. L.; Jones, R. B.; Chen, C. H. Org. Mass Spectrom. 1992, 27, 13891392. 48. Hoteling, A.; Nichols, W.; Giesen, D.; Lenhard, J.; Knochenmuss, R. Eur. J. Mass Spectrom. 2006, 12, 345-358.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 31

49. Wiegelmann, M.; Dreisewerd, K.; Soltwisch, J. J. Am. Soc. Mass Spectrom. 2016, 27, 1952-1964.


Page 27 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure Captions Figure 1. Absorption profiles of the investigated matrices in (a) solution and (b) the solid state. Figure 2. Heat maps showing the wavelength-dependence of MALDI-MS with a 3-HPA matrix. (a) Summed ion counts for 5, 10, 17, and 28 DNA oligomers, and (b) analyte ion yield. The solid grey lines display the laser penetration depth (δ = α-1) in the solid state, black dots the experimental data points. Figure 3. Heat maps showing the wavelength-dependence of MALDI-MS with a THAP matrix. (a) Summed ion counts for PA and PG, and (b) analyte ion yields. Solid grey lines: laser penetration depth. Figure 4. Heat maps showing the wavelength dependence of MALDI-MS with a dithranol matrix. (a) Summed ion counts for molecular PA and PG signals, and (b) analyte ion yields. Solid grey lines: laser penetration depth. Figure 5. Heat maps showing the wavelength dependence of MALDI-MS with a HABA matrix. (a) Summed ion counts for all PC-derived ion species. (b) Ratio of all adduct ion signals like dimers ([2M + H]+, [2M + Na]+, [2M + K]+) and analyte-matrix adduct species (such as [M + HABA + K]+ and [2M + HABA + K]+) to the total analyte-derived ion count. (c) Ratio of the ion counts of fragment ion signals ([M - HG + H]+, [M - FA + H]+, [M - FA + Na]+ to the total analyte-derived ion count displayed in a. Solid grey lines: laser penetration depth; FA: fatty acid; HG; head group. Figure 6.

UV-MALDI-MS analysis of non-covalent HGI-kemptide and HGI-PKCS

complexes with an ATT matrix. (a) Mass spectrum recorded at 240 nm and a fluence of 65 J m-2. Colored symbols denote the different ion species. (b) Analyte ion counts for gastrinPKCS complex and (c) for the unbound peptide monomer sub-units. Solid grey lines: laser penetration depth; dashed grey line: absorption profile of the peptides in solution – decadic molar extinction coefficients are denoted on the right axis. 27 ACS Paragon Plus Environment


Figures

Extinction coefficient / L mol-1 cm-1

(a)

solution phase 337 nm

20000

355 nm

3-HPA THAP dithranol HABA ATT

10000

0 300

400

500

600

Wavelength / nm

(b)

solid phase 337 nm

355 nm

Absorption / a. u.

1.0

0.5 3-HPA THAP dithranol HABA ATT

0.0 300

400

500

600

Wavelength / nm

-2

600

analyte ion count 2E05 (a) 00 2x10 5

analyte ion yield 0.2 (b) 00.0 0.2

8

400

6

200

4 2 300

320

340

300

320

340

Wavelength / nm

Penetration depth / a. u.

Figure 1

Fluence / J m

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 31

Figure 2


0

2x10

analyte ion yield 0.18 (b) 0.04 0.04

10

0.18

200

5

100 0 225 250 275 300 325

Wavelength / nm


Fluence / J m

300

analyte ion count 2E05 (a) 0 5

225 250 275 300 325

Fluence / J m

-2

600

analyte ion count analyte ion yield 4E04 (b)0.01 0.08 (a) 0 4 0

0.01

4x10

10

0.08

400 5 200

250 300 350 400

250 300 350 400

0

Wavelength / nm


Figure 3

Figure 4

(a) 0

2x104

0

adduct ratio

(b)

2E04

fragment ratio

0

(c)

1

0

1

0

1

1x100

0

1000

500

300

4

2

400

500

600

300

400

500

600

300

400

500


analyte-derived ion count

Fluence / J m-2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


-2

Page 29 of 31

600

Wavelength / nm

Figure 5



matrix matrix fragment gastrin I kemptide PKCS

-2 6000 λ = 240 nm, F = 64 J m

4000 x10

2000

0 500

1000

1500

2000

2500

3000

m/z

(b)

gastrin I - PKCS complex

300

0

1E03 0

1125

(c)

unbound peptides

0

9E05 0

2E5

86400

200

Extinction coefficient / L mol-1 cm-1

Intensity / a. u.

(a)

Fluence / J m-2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 31

1E5

100 0 220 260 300 340 380 220 260 300

340 380

Wavelength / nm

Figure 6


1 2 3 4 5 6 7

oToF - MS

/ J m-2 Fluence Fluence / J m-2

600 600

(a)

00

4E04 4E04

0 dithranol 4x104

400 400 200 200

ACS Paragon Plus Environment

wavelength-tunable laser

250 300 300 350 400 400 250 350

Wavelength / nm

Laser penetration depth / a. u.

analyte count analyte ion ion count MALDI Page 31 of 31Analytical Chemistry

(b

New Insights into the Wavelength Dependence of MALDI Mass

Recommend Documents