4-phenyl-α-cyanocinnamic acid amide: Screening for a negative ion matrix for MALDI-MS imaging of multiple lipid classes

Article abstract of DOI:10.1021/ac4018154

Matrix-assisted laser desorption/ionization imaging mass spectrometry (MALDI-IMS) has become a method of choice in lipid analysis, as it provides localization information for defined lipids that is not readily accessible with nonmass spectrometric methods. Most current MALDI matrices have been found empirically. Nevertheless, preferential matrix properties for many analyte classes are poorly understood and may differ between lipid classes. We used rational matrix design and semiautomated screening for the discovery of new matrices suitable for MALDI-IMS of lipids. Utilizing Smartbeam- and nitrogen lasers for MALDI, we systematically compared doubly substituted α-cyanocinnamic acid derivatives (R¹-CCA-R²) with respect to their ability to serve as negative ion matrix for various brain lipids. We identified 4-phenyl-α-cyanocinnamic acid amide (Ph-CCA-NH ₂) as a novel negative ion matrix that enables analysis and imaging of various lipid classes by MALDI-MS. We demonstrate that Ph-CCA-NH₂ displays superior sensitivity and reproducibility compared to matrices commonly employed for lipids. A relatively small number of background peaks and good matrix suppression effect could make Ph-CCA-NH₂ a widely applicable tool for lipid analysis.

Full text of DOI:10.1021/ac4018154

Article
pubs.acs.org/ac
4‑Phenyl-α-cyanocinnamic Acid Amide: Screening for a Negative Ion
Matrix for MALDI-MS Imaging of Multiple Lipid Classes
Annabelle Fulop,†,‡,§ Martina B. Porada,†,∥ Christian Marsching,†,‡,§,⊥ Henning Blott,†,∥ Bjorn Meyer,†,‡,§
̈
̈
̈
Suparna Tambe,†,∥ Roger Sandhoff,†,‡,⊥ Hans-Dieter Junker,†,∥ and Carsten Hopf*,†,‡,§
‡
†Applied Research Center in Biomedical Mass Spectrometry (ABIMAS), Instrumental Analysis and Bioanalysis, Mannheim
University of Applied Sciences, Paul-Wittsack-Strasse 10, 68163 Mannheim, Germany
§Institute of Medical Technology, University of Heidelberg and Mannheim University of Applied Sciences, Paul-Wittsack-Strasse 10,
68163 Mannheim, Germany
∥Organic Chemistry, Aalen University of Applied Sciences, Beethovenstrasse 1, 73430 Aalen, Germany
⊥Lipid Pathobiochemistry, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 280, 69120 Heidelberg, Germany
S
* Supporting Information
ABSTRACT: Matrix-assisted laser desorption/ionization imaging mass spec-
trometry (MALDI-IMS) has become a method of choice in lipid analysis, as it
provides localization information for defined lipids that is not readily accessible
with nonmass spectrometric methods. Most current MALDI matrices have been
found empirically. Nevertheless, preferential matrix properties for many analyte
classes are poorly understood and may differ between lipid classes. We used
rational matrix design and semiautomated screening for the discovery of new
matrices suitable for MALDI-IMS of lipids. Utilizing Smartbeam- and nitrogen
lasers for MALDI, we systematically compared doubly substituted α-
cyanocinnamic acid derivatives (R1-CCA-R2) with respect to their ability to
serve as negative ion matrix for various brain lipids. We identified 4-phenyl-α-
cyanocinnamic acid amide (Ph-CCA-NH2) as a novel negative ion matrix that
enables analysis and imaging of various lipid classes by MALDI-MS. We demonstrate that Ph-CCA-NH2 displays superior
sensitivity and reproducibility compared to matrices commonly employed for lipids. A relatively small number of background
peaks and good matrix suppression effect could make Ph-CCA-NH2 a widely applicable tool for lipid analysis.
the very different ionization efficiencies of various lipid species.
It is therefore necessary to choose a matrix that is either
selective for a defined class of lipids16,17 or that facilitates
efficient ionization of a maximum number of lipid classes
independent of their distinct physicochemical properties. In
addition, a suitable matrix should possess a high absorbance at
the emission wavelength of the laser and a low matrix
background.18,19 The presence of alkali metal ions in tissue
slices leads to formation of multiple adducts for each lipid
species in MALDI images in positive ion mode and thus
complicates their interpretation. In contrast, in negative ion
spectra the exclusive presence of [M − H]− ion signals formed
from a large variety of lipids facilitates the analysis.
The most frequently used matrices for lipid MALDI-IMS in
negative ion mode are 2,5-dihydroxybenzoic acid (DHB)20 and
9-aminoacridine (9-AA).17,21 However, DHB tends to form
large crystals, which may cause molecular delocalization and
poor spot-to-spot reproducibility,22,23 and it exhibits only poor
analyte sensitivities accompanied by strong background noise in
ipids are major building blocks of biological membranes,
1,2
Land they play key roles in signal transduction pathways
and mechanisms of disease.3,4 Because of these important
functions, the field of lipid analysis and lipidomics has made
considerable progress in recent years.5,6 Matrix-assisted laser
desorption/ionization time-of-flight mass spectrometry
(MALDI-TOF MS)7 in general and MALDI imaging MS
(MALDI-IMS)8,9 in particular, which provides information
about the location of different lipid species, have become
methods of choice for lipid studies.10−14 In a typical MALDI-
IMS experiment, a thin tissue cryosection is mounted onto a
suitable target and covered with a chemical matrix solution that
extracts analytes of interest from the underlying tissue. Some
matrices limit the reproducibility and the spatial resolution of
MALDI imaging experiments because of large or inhomoge-
neous crystal formation on the tissue slices. In addition,
inhomogeneous cocrystallization of analytes with matrices can
lead to the existence of so-called hot spots15 on the sample,
which causes quantitative errors and complicates automated
measurement. Therefore, the role of the matrix is to
cocrystallize with analytes and to absorb laser energy, in
order to promote desorption/ionization. The challenges in
imaging stem from the wide range of lipid concentrations and
Received: June 17, 2013
Accepted: August 28, 2013
Published: August 28, 2013
© 2013 American Chemical Society
9156
dx.doi.org/10.1021/ac4018154 | Anal. Chem. 2013, 85, 9156−9163

Analytical Chemistry
Article
negative ion mode.24 Therefore, it is not an ideal matrix for
lipid detection in negative ion mode. Additional lipid matrices
that have been used for negative ion MALDI-MS are 2,6-
dihydroxyacetophenone (DHA),25,26 2-(2-aminoethylamino)-
5-nitropyridine (AAN),27 and 2-mercaptobenzothiazole
(MBT).28 However, the sublimation of DHA under high-
vacuum conditions29 presents a significant disadvantage for
IMS, whereas the lipid signals of MBT and AAN in negative ion
mode are weaker than those with 9-AA.30
Nearly all matrices in use today have been found empirically.
Among the few exceptions are the rationally designed matrix
1,8-bis(dimethylamino)naphthalene31 for the ionless detection
of small molecules and the matrix 4-chloro-α-cyanocinnamic
acid (Cl-CCA).32 However, the replacement of the hydroxyl
group in the conventional matrix 4-hydroxy-α-cyanocinnamic
acid (HCCA) against the chlorine substituent leads to a
hypsochromic shift. Consequently, increased performance is
only achieved with MALDI-MS instruments with a nitrogen
laser (337 nm) but not in instruments with a frequency-tripled
Nd:YAG laser (355 nm).32 Nevertheless, some substance
properties (e.g., vacuum stability, crystallization) are hard to
plan.
Here, we therefore rationally designed HCCA analogs with
two substitutents (R1-CCA-R2) and performed semiautomated
screening of various substances with respect to their utility as
negative ion matrices for MALDI-MS analysis of brain lipids. In
order to facilitate future use of these novel matrices in diverse
applications, we provide performance measures in positive and
negative ion mode utilizing nitrogen and Nd:YAG lasers as a
resource for the lipid analysis community. Finally, we present 4-
phenyl-α-cyanocinnamic acid amide (Ph-CCA-NH2) as a novel
MALDI-IMS matrix for analysis and imaging of various lipid
classes in negative ion mode.
arbitrary units (au) for the matrix ion peak was observed] were
determined manually for each substance. For automated lipid
measurements the AutoXecute function of the flexControl 3.0
software was used. Spectra were acquired in the mass range
from m/z 400 to 2000 Da with a low mass gate at 370 and 390
Da for negative and positive ion mode, respectively; 4000 laser
shots were accumulated per spot by random walk with 200
shots per position. The laser was set individually for each
substance slightly above the threshold level to get the best
intensity and signal-to-noise result. The MS was calibrated
internally using a list of theoretical masses of detected lipids.
Peak quality parameters were S/N ≥ 3, peak width < 0.2 Da at
80% peak intensity, and resolution > 1000. In case of 30
unsuccessful attempts, measurement was aborted. Spectra were
determined with flexAnalysis 3.4 software (Bruker Daltonics).
Ultraflex I Parameters. Parameters were as used for Autoflex
III, with the following exceptions: source and reflector voltages
were 25 and 26.3 kV, respectively, in positive ion mode and 20
kV and 21 kV, respectively, in negative ion mode.
Statistical examination was done with Microsoft Excel.
Average intensities and standard deviation were calculated of
three replicates [Figures 3 and S2 and S3 (Supporting
Information)]. To compare groups of spectra via gel view,
spectra were analyzed using ClinProTools 2.2 software (Bruker
Daltonics). Groups were generated from four single spectra
[Figures 4 and S4 (Supporting Information)].
Determination of the MSElip. Score.35 To evaluate the
matrix suppression effect (MSE), the MSElip. score, defined as
the summed analyte signals over the ion current from both
matrix and analyte signals, in presence of 20 ng brain total lipid
extract for the matrices 9-AA and Ph-CCA-NH2 was calculated.
Therefore, the number of matrix peaks (∑m) with S/N ≥ 5
and the summed peaks derived from lipid extract (∑lip.) with
S/N ≥ 5 were determined in the presence of 20 ng extract for
each matrix and the score was calculated: MSElip. = ∑lip./
(∑lip. + ∑m).
EXPERIMENTAL SECTION
■
Synthesis of α-Cyanocinnamic Acid Derivatives.33
Synthesis was done by Knoevenagel condensation: A mixture
of cyanoacetic acid derivative (1.0 equiv), benzaldehyde
derivative (0.9 equiv), and ammonium acetate (0.15 equiv) in
toluene (20 mL) was refluxed for 2−6 h. After completion of
reaction it was cooled to room temperature and recrystallized
with an acetone/water or methanol/water mixture. In case of
no precipitation the solvent was evaporated completely and
then recrystallized.34 The α-cyanoacetic acid amides were
prepared by refluxing α-cyanoacetic acid ethyl ester with excess
amine in ethanol as a solvent. All synthesized compounds were
purified to homogeneity by repetitive recrystallization or
column chromatography. The α-cyanocinnamic acid derivatives
MALDI-IMS. A frozen Sprague−Dawley rat brain was sliced
into 10 μm coronal sections using a Leica CM1950 cryostat
(Leica Biosystems, Nussloch, Germany) at a chamber temper-
ature of −15 °C and a specimen head temperature of −11 °C.
Cryosections were mounted onto indium tin oxide (ITO)-
coated conductive glass slides (Bruker Daltonics, Bremen,
Germany) and dried for 30 min under vacuum. Identical
amounts of 9-AA [5 mg/mL in acetonitrile/water (80:20, v/v)]
or Ph-CCA-NH2 [5 mg/mL in acetone/water (90:10, v/v)]
were deposited onto one piece of tissue, while the other was
covered with a glass slide. For matrix deposition a SunCollect
MALDI Spotter (SunChrom, Friedrichsdorf, Germany) was
used. Air pressure was 2.5 bar. Matrix was deposited in nine
layers, with medium flow rate set to 10 μL/min (first layer), 15
μL/min (second), 20 μL/min (third), and 25 μL/min (all
others). The distance between tissue and spray head was 25.3
mm. MS images were obtained on the Autoflex III MALDI-
TOF/TOF MS in negative ion-reflector mode in the m/z range
from 600 to 1800 Da at a spatial resolution of 100 μm with 200
laser shots per position. Optimized laser energies for the two
matrices, slightly above the individual threshold level, were
determined before measurement and kept constant during one
region. Regions with different matrix depositions were
measured in random order (Figure 5, 9-AA region first; Figure
S5, Supporting Information, Ph-CCA-NH2 first) using
flexImaging 3.0 software (Bruker Daltonics, Bremen, Ger-
many). Mass filters were chosen with a width of 0.2 Da.
1
were characterized by H and 13C NMR spectroscopy and
MALDI mass spectrometry (see Supporting Information).
MALDI-TOF Mass Spectrometry. Mass spectra were
either acquired on an Autoflex III (355 nm SmartbeamTM
laser; 200 Hz repetition rate, ∼3 × 10−7 hPa) or an Ultraflex I
(337 nm nitrogen laser; 25 Hz, ∼10−7 hPa) (both from Bruker
Daltonics, Bremen, Germany). The Smartbeam laser has a
modulated beam profile. This needs to be taken into account
whenever we refer to it as a 355 nm Nd:YAG laser throughout
the text.
Autoflex III Parameters. Samples were analyzed in reflector
mode with extraction at a source voltage of 19 kV and reflector
voltage of 21 kV in both positive and negative ion mode. For
comparison of matrix substances, laser energy thresholds
[defined as laser intensity, for which an intensity of 1 × 104
9157
dx.doi.org/10.1021/ac4018154 | Anal. Chem. 2013, 85, 9156−9163

Analytical Chemistry
Article
Normalization was passed. The MALDI-MS/MS analysis using
the LIFT cell of the Autoflex III instrument was conducted
directly on the tissue section after MALDI IMS.
ESI QqQ-MS/MS Measurements. Sprague−Dawley rat
brain lipid raw extract was diluted for ESI QqQ-MS/MS
analysis to a concentration corresponding to 0.31 mg wet
weight per milliliter of methanol for negative mode MS and
0.31 mg wet weight per milliliter of methanol with 5 mM
ammonium acetate for positive mode MS. All scans where
performed on a Xevo TQ-S tandem mass spectrometer
(Waters, Milford) by direct infusion with a flow rate of 20
μL min−1. Scan time was 2 min with a scan duration of 3 s. All
spectra were baseline corrected and smoothed with Savitzky−
Golay algorithm twice with a window size of 3. Processing was
done with MassLynks Software 4.1 (Waters, Milford).
RESULTS AND DISCUSSION
Rational Design and Spectroscopic Characterization
of α-Cyanocinnamic Acid Derivatives. Absorption of
■
Figure 2. Brain total lipid extract is an adequate analyte for screening
of a broad range of lipids. (A) MALDI-TOF MS spectra (modulated
Nd:YAG laser) of brain total lipid extract in positive ion mode with
DHB (0.5 M) in methanol or (B) in negative ion mode with 9-AA (5
mg/mL) in 2-propanol/acetonitrile (3:2, v/v) and accordingly (C)
Ph-CCA-NH2 (5 mg/mL) in acetonitrile/water (4:1, v/v).
often high variability in automated data acquisition (own
unpublished observation), and an absorbance minimum at the
laser irradiation wavelength, which leads to a high ionization
energy threshold. Therefore, we used physicochemical proper-
ties of 9-AA like its clogP as guidance for the tailoring of
cinnamic acid derivatives for negative ion mode measurements
of lipid species. We furthermore reasoned that an acid
functionality presumably would be counterproductive for
deprotonation of analytes in negative ion mode. Therefore,
the carboxyl functionality was replaced by neutral amides, to
minimize the influence of matrix acid−base chemistry in the
ionization process and to facilitate the detection of acid/base
sensitive compounds. Use of different amide substituents for
this replacement offered further opportunity to change the
hydrophobicity of the compounds without changing the
principle conjugated system for UV absorption in the
MALDI process.
We synthesized 12 bisubstituted α-cyanocinnamic acid
derivatives (R1-CCA-R2; Figure 1A) and measured their UV
absorption in solution and solid state (Table S1B, Supporting
Information). UV absorbance profiles of MALDI matrices can
be determined as barium sulfate pellets or as sublimed
films.37,38 Solid-phase measurements, in general, show a
broadening of absorbance bands. Furthermore, it was shown
that oscillator strengths in solution and solid phase are
comparable, when optical absorbance coefficients are deter-
mined of sublimed MALDI matrix films. Therefore, solution
data can only be relied on within a factor of about 2 to provide
data for modeling the solid film matrices.38 In this work, we
(re)crystallized matrix substances on a fiber optics to
approximate conditions on a MALDI target plate. In summary,
Figure 1. Structures of synthesized α-Cyanocinnamic acid derivatives
and UV−vis absorption of new matrix Ph-CCA-NH2. (A) Structures of
synthesized α-cyanocinnamic acid derivatives (R1-CCA-R2) and (B)
UV−vis absorption of Ph-CCA-NH2 in solution [c = 624 μM in
acetonitrile/water (4:1, v/v)] and in solid state (Asolution = 0.91Asolid).
photons at the operational wavelength of the laser is an
absolute requirement for any MALDI matrix. Since replace-
ment of the hydroxyl group in HCCA against proton-affine
substituents such as chlorine leads to a hypsochromic shift,32
matrices like Cl-CCA are not very well suited for applications
using MALDI-MS with frequency-tripled Nd:YAG lasers.
Therefore, we sought to design novel cinnamic acid matrix
compounds that avoided this hypsochromic shift and
consequently could be useful in lipid analysis applications
such as MALDI-IMS that today are often done with Nd:YAG
lasers, in particular, Smartbeam lasers with a modulated beam
profile. To this end, we used 4-chloro-α-cyanocinnamic acid
(Cl-CCA), 4-hydroxy-α-cyanocinnamic acid (HCCA), and 4-
phenyl-α-cyanocinnamic acid (Ph-CCA), which differ in their
pKa values, their molecular surface area (MSA), and hydro-
phobicity (clogP and clogD), as starting points for our design
approach (Table S1, Supporting Information).
For analysis of lipids by MALDI-MS in negative ion mode,
basic matrices such as 9-AA are preferred.10,17 This matrix in
particular features a relatively high clogP, which may enhance
detection of hydrophobic lipid species. 9-AA has some
drawbacks though, such as its unfavorable toxicology profile,36
9158
dx.doi.org/10.1021/ac4018154 | Anal. Chem. 2013, 85, 9156−9163

Analytical Chemistry
Article
Figure 3. Ph-CCA-NH2 is a versatile matrix with superior characteristics in negative ion mode. Mean peak intensities of selected lipids measured
with 0.5 μg brain total lipid extract with α-cyanocinnamic acid matrices [5 mg/mL in acetonitrile/water (4:1, v/v)] or 9-AA [5 mg/mL in 2-
propanol/acetonitrile (3:2, v/v)] in negative ion mode, measured with a Nd:YAG laser. Average intensities and standard deviation were calculated of
three replicates. m/z values of negatively ionized lipids were (A) 888.6 [SM4s(42:2) − H]−, (B) 904.6 [SM4s(h42:2 − H)]−, (C) 885.6 [PI(38:4) −
H]−, and (D) 766.5 [PE(38:4) − H]− and 750.5 [PE(p38:4) − H]−.
all α-cyanocinnamic acid derivatives showed UV-absorption in
the solid state in the desired range of wavelengths. According to
known UV/vis measurements of α-cyanocinnamic acid
derivatives,38,37 a more intense bathochromic shift as well as
peak broadening was observed for the solid state measure-
ments, as illustrated for Ph-CCA-NH2 as an example (Figure
1B). Therefore, maximal absorption in the solid state is not
reported as a fixed wavelength in Table S1B (Supporting
Information), but as a range of maximal light absorption.
We next attempted automated evaluation of multiple
matrices by MALDI-TOF MS. Although the variability of ion
signals depends on laser irradiance (i.e., intensity, time and
wavelength of laser irradiation), the main obstruction for
automation was the inhomogeneity of the matrix surface.39 We
achieved more homogeneous crystallization by dissolving all
HCCA derivatives in 80% acetonitrile, and then overlaying the
air-dried lipid solution with them (i.e., PrimaDrop sample
preparation method)40 (Figure S1, Supporting Information).
Crystal formation was still not uniform for all matrix
substances. In some cases small crystals or needles were
observed, but ring formation was avoided and automated data
acquisition was possible.
Consistent with the literature for Sprague−Dawley rat
brain,42,43 m/z values of the most prominent peaks in positive
ion mode corresponded to phosphatidyl choline (PC) isoforms
{m/z 734.6, [PC(32:0) + H]+; m/z 760.6, [PC(34:1) + H]+;
and m/z 788.6, [PC(36:1) + H]+; Figure 2A}; peaks
corresponding to sphingomyelin (SM) isoforms {m/z 731.6,
[SM(36:1) + H]+} and phosphatidyl ethanolamine (PE)
species {m/z 746.6, [PE(36:1) + H]+ and m/z 768.6,
[PE(38:4) + H]+} were also observed. Despite a 3-fold excess
of PE over PC in the brain extract, PE detection was suppressed
by PC in the extract,44 resulting in very small signals. In
negative ion mode MS spectra of brain lipid extract (Figure
2B,C), the most prominent lipids between m/z 400 and 2000
were hydroxylated (h) and nonhydroxylated sulfatides (SM4s
isoforms) that are known to be selectively ionized by 9-AA with
355 nm laser,16,17 whereas the same matrix substance appears
to show no preference for a special lipid species in case of a
nitrogen laser.44,45 Nevertheless, signals of phospholipids such
as phosphatidyl inositols (PI) and phosphatidyl serines (PS),
negatively charged 1,2-diacyl-PE species (PE) and plasmalo-
genic (p) PE species as well as gangliosides (GM1) that occur
in rat brain tissue25,43,46 were also detected in brain total lipid
extract, in particular with Ph-CCA-NH2 matrix (Figure 2C).
We concluded that this commercially available extract was a
suitably complex analyte mix that was useful for matrix
screening.
Since we were interested in screening candidate matrix
substances against a complex analyte mix containing a broad
range of lipid classes, we tested brain total lipid extract and used
DHB as a reference matrix for positive ion mode and 9-AA for
negative ion mode MALDI-MS. The LIPID MAPS Database41
was used for preliminary assignment of m/z values to lipids.
Semiautomated Screening of α-Cyanocinnamic Acid
Amide Derivatives Reveals Ph-CCA-NH2 as a Promising
9159
dx.doi.org/10.1021/ac4018154 | Anal. Chem. 2013, 85, 9156−9163

Analytical Chemistry
Article
DHB, well-known for detection of DPPC in positive ion
mode,47 were confirmed with PC species in brain total lipid
extract (Figure S2A/C, Supporting Information). The sulfatides
SM4s(42:2) and SM4s(h42:2) as well as phospholipids
PI(36:4), PE(38:4), and PE(p38:4) were chosen as exemplary
lipids in negative ion mode. As expected, 9-AA with a Nd:YAG
laser resulted in high, but variable, signal intensities for
sulfatides, but it did not work well for phospholipids (Figure
2B and 3). Furthermore, all cinnamic acid derivatives displayed
matrix properties for lipids in negative ion mode. Taking all
measurements with a 355 nm laser together (Figure 3A-D), Ph-
CCA-NH2 is always a good matrix and sometimes the best,
suggesting broad applicability of Ph-CCA-NH2 in lipid MALDI
MS, because of the outstanding intensity for PI as well as good
results for sulfatides and the less intense PE species.
Whereas the comparison of both lasers showed a comparable
trend in positive ion mode, the intensity pattern with nitrogen
laser in negative ion was very different (Figure S3, Supporting
Information). Surprisingly, measurements at 337 nm with the
newly synthesized α-cyanocinnamic acid derivatives resulted in
relatively low ion intensities compared with 9-AA. Con-
sequently, for detection of negatively charged lipids with
nitrogen laser MALDI instruments, 9-AA matrix may still be
advised.
Overall, our screening results suggested that, in line with UV
measurements in the solid state, most of the α-cyanocinnamic
acid amide derivatives we synthesized had MALDI matrix
properties for lipids and that their lipid preference and
selectivity warrant further investigation. Replacement of the
carboxyl with an amide group had no negative effect on matrix
performance, with the α-cyanocinnamic acid amides HCCA-
NH2 and particularly Ph-CCA-NH2 yielding promising results
with a Smartbeam laser.
Figure 4. Reduced matrix noise and higher signal intensities result in a
higher matrix suppression effect with Ph-CCA-NH2. (A) Gel view of
four groups of four negative ion MALDI spectra each. (Examples of
single spectra are shown in Figure S4, Supporting Information.)
Spectra were recorded with Ph-CCA-NH2 (5 mg/mL in 90% acetone)
matrix or 9-AA [5 mg/mL in 2-propanol/acetonitrile (3:2, v/v)]
matrix. Matrix background spectra as well as spectra recorded with 20
ng of brain total lipid extract show less matrix noise in combination
with Ph-CCA-NH2 matrix. (B) Total number of peaks resulting from
lipids signals and matrix background (S/N ≥ 5) and number of lipid
peaks (S/N ≥ 5) with Ph-CCA-NH2 or 9-AA. The MSElip. score was
calculated for both matrices in the presence of 20 ng of brain total lipid
extract.
Ph-CCA-NH2 Displays Higher Matrix Suppression and
Higher Versatility in Negative Ion MALDI-IMS. Desorption
and ionization of matrix molecules during analyte detection
causes background peaks in the low-mass region (m/z < 1000)
because of matrix fragmentation and cluster formation. These
background signals can be suppressed by sufficient analyte
ions.48,49 It has been noted that this matrix suppression effect
(MSE) is not observed with all matrices at the same ratio and is
generally reduced if the crystallization is inhomogeneous.49
Besides the matrix:analyte ratio, the most adjustable factor is
laser intensity. The MSE is maximal near the MALDI laser
intensity threshold, because a stronger laser pulse generates
more matrix ions that require more analytes to quench them.
Nevertheless, it is practically impossible to prepare samples
with an appropriate matrix:analyte ratio, since the analyte
concentrations in biological samples are often unknown. To
circumvent this limitation, it is desirable that the matrix of
choice already promotes the MSE at low amounts of sample.
Bearing the MSE in mind, we examined formation of
background matrix signals and their suppression by the
matrices Ph-CCA-NH2 and 9-AA near their laser intensity
thresholds in the absence or presence of 20 ng of brain total
lipid extract (Figure 4A). In the absence of lipid extract,
characteristic matrix cluster formation was observed in the case
of 9-AA, whereas the matrix spectra of Ph-CCA-NH2 showed
fewer background signals (Figures 4A and Figure S4B,
Supporting Information). The majority of the latter were
suppressed in the presence of lipid extract. This MSE was not
observed in the case of 9-AA matrix with 20 ng of lipid extract
(Figure S4A, Supporting Information). The MSElip. score with
Negative Ion Matrix for Preferential Use with 355 nm
Lasers. In order to take different laser thresholds of different α-
cyanocinnamic acid amide derivatives into account, but to also
avoid experimenter bias, the laser energy was set manually to a
value slightly above the threshold level for each matrix and each
MALDI instrument (for best intensity and signal-to-noise ratio)
and data was then acquired automatically. The α-cyanocin-
namic acid derivatives and reference matrices were compared
by measuring the ion intensities of selected brain lipid extract
species.
In the positive ion mode the intensities of lipid species
PC(32:0), PC(34:1), and SM(36:1), generated with the α-
cyanocinnamic acid derivatives and DHB as reference, were
compared (Figure S2, Supporting Information). Since the
comparison of absolute intensities between two different
MALDI instruments at two sites is not useful, the patterns of
relative ion intensities were considered for the respective
instruments. From this, HCCA-NH2 showed promise for lipid
detection in positive ion mode primarily when combined with a
355 nm Smartbeam laser. This matrix displayed an improved
signal intensity for SM(36:1) (Figure S2B, Supporting
Information) and comparable intensities for PC species (vs
DHB). In addition, comparable performances of Cl-CCA and
9160
dx.doi.org/10.1021/ac4018154 | Anal. Chem. 2013, 85, 9156−9163

Analytical Chemistry
Article
Figure 5. Ph-CCA-NH2 matrix enables MALDI-IMS of various lipid species. Imaging was performed on two adjacent cryosections of Sprague−
Dawley rat brain on the same ITO slide. Data was acquired in negative ion-reflector mode at a spatial resolution of 100 μm with 200 laser shots per
position. Prior to measurement, the same amount of 9-AA [5 mg/mL in acetonitrile/water (80:20, v/v)] (left side) or Ph-CCA-NH2 [5 mg/mL in
acetone/water (90:10, v/v)] (right side) was deposited onto one of the tissue sections with a SunCollect MALDI spotter. Imaging was carried out
while the laser energy was kept constant and slightly above the individual threshold level for one region. The 9-AA region was measured first. Mass
filters were chosen with a width of 0.2 Da.
20 ng extract was 4 times improved with Ph-CCA-NH2 when
compared to 9-AA (Figure 4B).
selector window range has to be expanded. Hence,
fragmentation of other lipids is possible. Consequently, the
unambiguous assignment of m/z 821.6 in the case of 9-AA
matrix was not possible, although a distinct image was obtained.
To confirm that the named lipids, to which imaged m/z values
had been matched, were indeed present in rat brain, we
analyzed crude extract of the corresponding rat brain region by
ESI-QqQ-MS/MS in a series of precursor and neutral loss
scanning experiments (Figure 6), demonstrating the presence
of all lipids imaged with Ph-CCA-NH2.
We therefore reasoned that the small number of background
peaks in the m/z range 400−2000 may imply that Ph-CCA-
NH2 be useful in the combined detection of sulfatides and
other lipids in IMS experiments. We therefore acquired two
independent IMS data sets using rat brain cryosections. In both
of them one piece of tissue was coated with 9-AA and the other
with Ph-CCA-NH2. In one experiment the 9-AA-coated region
was measured first (Figure 5), in the other it was measured last
(Figure S5, Supporting Information). The images were in good
agreement with the intensities displayed in Figure 3. The
reported selectivity of 9-AA for sulfatides [SM4s(42:2) and
SM4s(h42:2)] was confirmed.17,21 In contrast, Ph-CCA-NH2
also enabled imaging and localizing other lipids such as
phosphatidyl inositols [e.g., PI(36:4) and PI(38:4)] and even
the phosphatidyl ethanolamines [PE(38:4) and PE(p38:4)] or
the phosphatidyl glycerol species [PG(34:1) and PG(40:6)].
For further validation of the identity of lipids represented in
Figures 5 and S5 (Supporting Information), on tissue MS/MS
spectra were acquired for the above-named lipid isoforms with
Ph-CCA-NH2 matrix (Table S2, Supporting Information).
Hereby, m/z 885.6 was identified as [PI(18:0;20:4) − H]− and
m/z 821.5 as [PG(18:0;22:6) − H]−. The other assignments
are illustrated in Table S2 (Supporting Information).50 On
tissue MS/MS fragmentation with 9-AA was only possible for
sulfatides and for PI(18:0;20:4), whereas no difference in the
fragmentation pattern generated with either Ph-CCA-NH2 or 9-
AA was observed, as exemplified for [PI(18:0;20:4) − H]− (m/
z 885.6) (Figure S6, Supporting Information). In general, the
wide precursor ion selector window of the Autoflex III MS
complicates fragmentation analysis. For less intense signals the
CONCLUSION
■
In this study, with the goal of identifying a novel matrix for lipid
MALDI-IMS applications, we have conducted a semiautomated
comparison of different α-cyanocinnamic acid amide derivatives
(R1-CCA-R2) using brain total lipid extract as analyte
containing a broad spectrum of lipids. Automated measurement
was installed to avoid experimenter bias. The homogeneous
crystallization necessary for such measurements could be
achieved by PrimaDrop sample preparation. Comparison of
different matrix substances demonstrated that replacement of
the carboxyl group with an amide group is a promising method
to improve matrix performance for lipids. Superior to any other
matrix tested in this study in negative ion mode with 355 nm
Smartbeam laser, 4-phenyl-α-cyanocinnamic acid amide dis-
played features of MALDI matrix with improved sensitivity and
reproducibility and a small number of background peaks in the
mass range of lipids. In conclusion, Ph-CCA-NH2 may be an
attractive alternative matrix of choice for negative ion mode
detection of various less abundant lipids by MALDI-MS.
9161
dx.doi.org/10.1021/ac4018154 | Anal. Chem. 2013, 85, 9156−9163

Analytical Chemistry
Article
REFERENCES
■
(1) Phillips, R.; Ursell, T.; Wiggins, P.; Sens, P. Nature 2009, 459,
379−85.
(2) O’Donnell, V. B.; Murphy, R. C. Blood 2012, 120, 1985−92.
(3) Brewer, J. W. Biochim. Biophys. Acta 2013, 1831, 642−51.
(4) Vance, J. E.; Tasseva, G. Biochim. Biophys. Acta 2013, 1831, 543−
54.
(5) Han, X.; Yang, K.; Gross, R. W. Mass Spectrom. Rev. 2012, 31,
134−78.
(6) Harkewicz, R.; Dennis, E. A. Annu. Rev. Biochem. 2011, 80, 301−
25.
(7) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299−301.
(8) Stoeckli, M.; Farmer, T. B.; Caprioli, R. M. J. Am. Soc. Mass
Spectrom. 1999, 10, 67−71.
(9) Spengler, B.; Hubert, M. J. Am. Soc. Mass Spectrom. 2002, 13,
735−48.
(10) Fuchs, B.; Suss, R.; Schiller, J. Prog. Lipid Res. 2010, 49, 450−75.
(11) Gode, D.; Volmer, D. A. Analyst 2013, 138, 1289−315.
(12) Fernandez, J. A.; Ochoa, B.; Fresnedo, O.; Giralt, M. T.;
Rodriguez-Puertas, R. Anal. Bioanal. Chem. 2011, 401, 29−51.
(13) Colsch, B.; Jackson, S. N.; Dutta, S.; Woods, A. S. ACS Chem.
Neurosci. 2011, 2, 213−22.
(14) Woods, A. S.; Colsch, B.; Jackson, S. N.; Post, J.; Baldwin, K.;
Roux, A.; Hoffer, B.; Cox, B. M.; Hoffer, M.; Rubovitch, V.; Pick, C.
G.; Schultz, J. A.; Balaban, C. ACS Chem. Neurosci. 2013, 4, 594−600.
(15) Tholey, A.; Heinzle, E. Anal. Bioanal. Chem. 2006, 386, 24−37.
(16) Cheng, H.; Sun, G.; Yang, K.; Gross, R. W.; Han, X. J. Lipid Res.
2010, 51, 1599−609.
(17) Marsching, C.; Eckhardt, M.; Grone, H. J.; Sandhoff, R.; Hopf,
C. Anal. Bioanal. Chem. 2011, 401, 53−64.
(18) Karas, M.; Bachmann, D.; Hillenkamp, F. Anal. Chem. 1985, 57,
2935−9.
(19) Zenobi, R.; Knochenmuss, R. Mass Spectrom. Rev. 1998, 17,
337−66.
(20) Wang, H. Y.; Jackson, S. N.; Post, J.; Woods, A. S. Int. J. Mass
Spectrom. 2008, 278, 143−9.
Figure 6. Validation of molecular peak assignment in MALDI-IMS by
ESI QqQ-MS/MS. Identification of all lipid species assigned in
MALDI IMS (Figure 5) in rat brain raw extract either by specific
neutral-loss (NL) scanning in positive ion mode for (A) PE (+NL
141) and (B) PG (+NL 172) or by specific parent ion scanning in
negative ion mode for (C) PI (− parents 241) and (D) SM4s (−
parents 97).
(21) Cerruti, C. D.; Benabdellah, F.; Laprevote, O.; Touboul, D.;
Brunelle, A. Anal. Chem. 2012, 84, 2164−71.
(22) Rujoi, M.; Estrada, R.; Yappert, M. C. Anal. Chem. 2004, 76,
1657−63.
(23) Bouschen, W.; Spengler, B. Int. J. Mass Spectrom. 2007, 266,
129−37.
(24) Schiller, J.; Suss, R.; Fuchs, B.; Muller, M.; Petkovic, M.;
Zschornig, O.; Waschipky, H. Eur. Biophys. J. 2007, 36, 517−27.
(25) Jackson, S. N.; Wang, H. Y.; Woods, A. S. J. Am. Soc. Mass
Spectrom. 2007, 18, 17−26.
(26) Jackson, S. N.; Wang, H. Y.; Woods, A. S.; Ugarov, M.; Egan, T.;
Schultz, J. A. J. Am. Soc. Mass Spectrom. 2005, 16, 133−8.
(27) Lorkiewicz, P.; Cecilia Yappert, M. J Mass Spectrom. 2009, 44,
137−43.
ASSOCIATED CONTENT
■
S
* Supporting Information
Additional information as noted in the text. This material is
available free of charge via the Internet at http://pubs.acs.org.
(28) Astigarraga, E.; Barreda-Gomez, G.; Lombardero, L.; Fresnedo,
O.; Castano, F.; Giralt, M. T.; Ochoa, B.; Rodriguez-Puertas, R.;
Fernandez, J. A. Anal. Chem. 2008, 80, 9105−14.
(29) Wang, H. Y.; Jackson, S. N.; Woods, A. S. J. Am. Soc. Mass
Spectrom. 2007, 18, 567−77.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: c.hopf@hs-mannheim.de. Phone: +49 (0)621/292-
6802. Fax: +49 (0)621/292-6420.
(30) Teuber, K.; Schiller, J.; Fuchs, B.; Karas, M.; Jaskolla, T. W.
Chem. Phys. Lipids 2010, 163, 552−60.
Notes
The authors declare no competing financial interest.
(31) Shroff, R.; Rulisek, L.; Doubsky, J.; Svatos, A. Proc. Natl. Acad.
Sci U. S. A. 2009, 106, 10092−6.
(32) Jaskolla, T. W.; Lehmann, W. D.; Karas, M. Proc. Natl. Acad. Sci.
U. S. A. 2008, 105, 12200−5.
ACKNOWLEDGMENTS
■
This work was supported by the Baden-Wurttemberg Ministry
of Science and Culture (INST 874/2-1 LAGG to C.H.), by a
joined grant (“ZAFH ABIMAS”) from ZO IV by the
̈
(33) Cope, A. C. J. Am. Chem. Soc. 1937, 59, 2327−30.
(34) Jaskolla, T.; Karas, M. Patent in http://www.patent-de.com/
20090108/DE102007040251B3.html, 2009.
Landesstiftung Baden-Wurttemberg and the Europaischer
̈
̈
(35) McCombie, G.; Knochenmuss, R. Anal. Chem. 2004, 76, 4990−
7.
(36) Sakore, T. D.; Jain, S. C.; Tsai, C. C.; Sobell, H. M. Proc. Natl.
Acad. Sci. U. S. A. 1977, 74, 188−92.
Fonds fur regionale Entwicklung (EFRE; to H.D.J., R.S., and
̈
C.H.), and by a grant by the Phospholipid-Forschungszentrum
e.V., Heidelberg, Germany (to C.H.).
9162
dx.doi.org/10.1021/ac4018154 | Anal. Chem. 2013, 85, 9156−9163

Analytical Chemistry
Article
(37) Soltwisch, J.; Jaskolla, T. W.; Hillenkamp, F.; Karas, M.;
Dreisewerd, K. Anal. Chem. 2012, 84, 6567−76.
(38) Allwood, D. A.; Dreyfus, R. W.; Perera, I. K.; Dyer, P. E. Rapid
Commun. Mass Spectrom. 1996, 10, 1575−8.
(39) Jensen, O. N.; Mortensen, P.; Vorm, O.; Mann, M. Anal. Chem.
1997, 69, 1706−14.
(40) Ruh, H.; Sandhoff, R.; Meyer, B.; Gretz, N.; Hopf, C. Anal.
Chem. 2013, 85, 6233−40.
(41) Fahy, E.; Cotter, D.; Sud, M.; Subramaniam, S. Biochim. Biophys.
Acta, Mol. Cell Biol. Lipids 2011, 1811, 637−47.
(42) Jackson, S. N.; Wang, H. Y.; Woods, A. S. J. Am. Soc. Mass
Spectrom. 2005, 16, 2052−6.
(43) Jackson, S. N.; Wang, H. Y.; Woods, A. S. Anal. Chem. 2005, 77,
4523−7.
(44) Petkovic, M.; Schiller, J.; Muller, M.; Benard, S.; Reichl, S.;
́
̈
Arnold, K.; Arnhold, J. Anal. Biochem. 2001, 289, 202−16.
(45) Fuchs, B.; Bischoff, A.; Suss, R.; Teuber, K.; Schurenberg, M.;
Suckau, D.; Schiller, J. Anal. Bioanal. Chem. 2009, 395, 2479−87.
(46) Woods, A. S.; Jackson, S. N. AAPS J. 2006, 8, E391−5.
(47) Jaskolla, T.; Fuchs, B.; Karas, M.; Schiller, J. J. Am. Soc. Mass
Spectrom. 2009, 20, 867−74.
(48) Knochenmuss, R.; Dubois, F.; Dale, M. J.; Zenobi, R. Rapid
Commun. Mass Spectrom. 1996, 10, 871−7.
(49) Knochenmuss, R.; Karbach, V.; Wiesli, U.; Breuker, K.; Zenobi,
R. Rapid Commun. Mass Spectrom. 1998, 12, 529−34.
(50) Hsu, F. F.; Turk, J. J. Am. Soc. Mass Spectrom. 2000, 11, 986−99.
9163
dx.doi.org/10.1021/ac4018154 | Anal. Chem. 2013, 85, 9156−9163