Article

Cite This: Anal. Chem. XXXX, XXX, XXX-XXX pubs.acs.org/ac

A Comprehensive High-Resolution Targeted Workflow for the Deep

Profiling of Sphingolipids
Bing Peng,† Susan T. Weintraub,‡ Cristina Coman,† Srigayatri Ponnaiyan,§ Rakesh Sharma,∥,⊥
Björn Tews,∥,⊥ Dominic Winter,§ and Robert Ahrends*,†
†
Leibniz-Institut für Analytische Wissenschaften-ISAS-e.V., 44227 Dortmund, Germany
‡
Department of Biochemistry and Structural Biology, The University of Texas Health Science Center at San Antonio, San Antonio,
Texas 78229, United States
§
Institute for Biochemistry and Molecular Biology, University of Bonn, 53113 Bonn, Germany
∥
Schaller Research Group, University of Heidelberg and DKFZ, 69120 Heidelberg, Germany
⊥
Molecular Mechanisms of Tumor Invasion, DKFZ, 69120 Heidelberg, Germany
*
S Supporting Information

ABSTRACT: Sphingolipids make up a highly diverse group

of biomolecules that not only are membrane components but
also are involved in various cellular functions such as signaling
and protein sorting. To obtain a quantitative view of the
sphingolipidome, sensitive, accurate, and comprehensive
methods are needed. Here, we present a targeted reversed-
phase liquid chromatography−high-resolution mass spectrom-
etry-based workflow that significantly increases the accuracy of
measured sphingolipids by resolving nearly isobaric and
isobaric species; this is accomplished by a use of (i) an
optimized extraction procedure, (ii) a segmented gradient, and
(iii) parallel reaction monitoring of a sphingolipid specific
fragmentation pattern. The workflow was benchmarked
against an accepted sphingolipid model system, the RAW 264.7 cell line, and 61 sphingolipids were quantified over a dynamic
range of 7 orders of magnitude, with detection limits in the low femtomole per milligram of protein level, making this workflow
an extremely versatile tool for high-throughput sphingolipidomics.

S phingolipids represent 10−30% of the total cellular lipid

content1,2 and are associated with four main functions as a
consequence of their physicochemical properties. First,
sphingolipids can exert a major influence on the physical
properties of membranes, including membrane thickness,
stiffness, and curvature.3 Second, they are involved in protein
sorting and the organization of membrane protein distribution
and activity such as in lipid rafts.3−5 Third, sphingolipids play a
major role in cellular signaling.6 Of particular importance are
water-soluble sphingolipids, such as sphingosine, their
phosphorylated analogues, and short-chain ceramides, because
they can initiate signaling of various effector molecules,
including phosphatases, kinases, and G protein-coupled
receptors.7−9 Fourth, the polar headgroups of many complex
sphingolipids that are part of the outer membrane leaflet
participate in a variety of different lipid−protein binding events
and mediate host−pathogen interactions and adhesion
processes.10 Besides these primary functions, sphingolipid
metabolism is tightly coupled to the metabolism of
glycerophospholipids and cholesterol, allowing for mutual
sensing and control.11
Sphingolipids make up a highly diverse group of lipid species
containing a sphingoid base backbone assembled out of a linear
alkane and a 1,3-dihydroxy-2-amino headgroup. Sphingolipids
as a class are structurally very heterogeneous as a consequence
of the combination of various headgroup species, different
numbers of hydroxyl groups, and the variation of the fatty acyl
chains. The high degree of diversity, the wide dynamic range of
sphingolipid concentrations in tissues and cells, interference
with detection due to the presence of abundant phospholipids,
and the existence of many isobaric and nearly isobaric
sphingolipid species (mass difference of <0.5 Da at the triple-
quadrupole mass spectrometer) often complicate their
identification.
During the past decade, major advancements in sphingolipid
analysis have been made by hyphenating LC techniques such as
reversed-phase (RP) and hydrophilic interaction liquid
chromatography (HILIC) with selected reaction monitoring
(SRM) mass spectrometry (MS).12−14 In addition, develop-
Received: September 1, 2017
Accepted: October 17, 2017
Published: October 17, 2017

© XXXX American Chemical Society A DOI: 10.1021/acs.analchem.7b03576

Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
ment of software tools for parameter optimization, identi-
fication, quantification, and visualization, such as Multiquant15
and Skyline-modified “lipidomics”,16 has made it possible to
analyze many different sphingolipid species. However, many
isobaric and nearly isobaric sphingolipids cannot be differ-
entiated from one another using a low-resolution/low-mass
accuracy triple-quadrupole mass spectrometer and/or relying
on single transitions as the identifier and quantifier, even with
sophisticated LC separations. To obtain a more precise picture
of which molecules are present in a sample, the analysis of
sphingolipids should be conducted by monitoring the exact
mass of the precursor and multiple fragments, including the
long-chain base (LCB), fatty acyls, and/or class specific
fragments, thereby allowing annotation at the fatty acyl species
level.17 Therefore, we developed an improved workflow that
includes an optimized extraction procedure, a segmental linear
chromatographic gradient, and a high-resolution (HR)/high-
mass accuracy full scan/parallel reaction monitoring (PRM)
approach for a comprehensive sphingolipid analysis. This
enabled us to profile sphingolipids in response to autophagy in
RAW 264.7 cells. The lipid profile of this cell line is published,
and the cells are used by the LIPID MAPS consortium as a
standard cell line for method development, enabling us to apply
a standardized cell culture protocol and directly benchmark our
method.18
■ EXPERIMENTAL SECTION
Materials. Chemicals and reagents were obtained from the
following sources: formic acid, ammonium formate, acetic acid
(HAc), potassium hydroxide (KOH), MS-grade phosphoric
acid (85−90%), and tert-butyl methyl ether (MTBE) from
Sigma-Aldrich (Steinheim, Germany); MS-grade acetonitrile
(ACN) and methanol (MeOH) from Biosolve (Valkenswaard,
The Netherlands); sodium chloride (NaCl) and isopropanol
(IPA) from Merck (Darmstadt, Germany); sodium dodecyl
sulfate (SDS) from Roth (Karlsruhe, Germany); tris-
(hydroxymethyl)aminomethane (Tris) from Applichem
(Darmstadt, Germany); and Kdo2-Lipid A (KLA), monosulfo
galactosyl ceramide (GalaCerS) d18:1/12:0, and ceramide/
sphingoid internal standard mixture II (CerMix) consisting of
sphingosine d17:1, sphinganine d17:0, sphingosine-1-P d17:1,
sphinganine-1-P d17:0, sphingomyelin d18:1/12:0, ceramide
(Cer) d18:1/12:0, glucosylceramide d18:1/12:0 (GlcCer,
internal standard for HexCer), lactosylceramide (LacCer, as
the internal standard for DiHexCer) d18:1/12:0, and ceramide-
1-P (CerP) d18:1/12:0 in ethanol from Avanti Polar Lipids
(Alabaster, AL). Ultrapure water (18 MΩ cm at 25 °C) was
obtained from an Elga Labwater system (Lane End, U.K.). The
bicinchoninic acid assay (BCA) was obtained from Thermo
Fisher Scientific (Rockford, IL).
Cell Culture. Mouse macrophages (RAW 264.7; ATCC,
Manassas, VA) were treated with KLA according to the
protocol provided by LIPID MAPS (www.lipidmaps.org/
protocols); the detailed protocol is described in the Supporting
Information. Three biological replicates each for untreated and
treated cells were prepared from a single batch of cells. The cell
pellets were transferred into a 2 mL polypropylene tube
(Eppendorf, Hamburg, Germany) prior to extraction.
Western Blot. Western blot was performed as described by
Barth et al.36 Further information can be found in the
Supporting Information.
Sample Preparation. Four different extraction methods
without and with alkaline treatment were tested to identify the
B DOI: 10.1021/acs.analchem.7b03576
Article
extraction protocol enabling the best determination of low-
abundance sphingolipid species. Alkaline hydrolysis of the
esters was applied to reduce the chromatographic interference
of glycerophospholipids and glycerolipids.19 Chloroform/
methanol (CHCl3/MeOH) extraction and alkaline CHCl3/
MeOH extraction were conducted as previously described by
Shaner et al.12 Detailed information is given in the Supporting
Information. MTBE extraction was performed following the
protocol of Matyash et al.20 (Supporting Information). Alkaline
MTBE extraction is based on the MTBE protocol of Matyash
et al.20 with small modifications. After the first incubation, 97.5
μL of 1 M KOH in MeOH was added and the mixture
incubated in the thermomixer for 2 h at 37 °C. After the
mixture was cooled to 25 °C, 2 μL of HAc was added to
neutralize the solution. Afterward, 188 μL of water was added,
and the samples were centrifuged at 10000g for 10 min at 4 °C.
The lipid extract was stored at −80 °C prior to further analysis.
High-Performance Liquid Chromatography (HPLC)
Gradients. Gradients previously utilized for sphingolipid
analysis were chosen for comparison21−23 (Table S1). In
addition, a segmented linear gradient was designed with
GradientOptimizer by inputting the retention times of
sphingolipid species that were obtained from a linear gradient
(initial 30% B, held at 30% B from 0.0 to 3.0 min, 30 to 75% B
from 3.0 to 15.0 min, 75 to 100% B from 15.0 to 17.0 min, and
100% B from 17.0 to 25.0 min).24
LC−MS/MS. Evaluation of Extraction Protocols. Initial
extraction comparison was accomplished with a QTRAP 6500
instrument (Applied Biosystems, Darmstadt, Germany) that
was equipped with an electrospray ion source (Turbo V ion
source). The detailed information about ESI source settings is
provided in the Supporting Information. The collision energy
was optimized for each sphingolipid class by direct infusion of a
corresponding standard. The scheduled SRM detection window
was set to 2 min, and the cycle time was set to 2 s. Data were
acquired with Analyst version 1.6.2 (AB Sciex, Concord, ON).
Skyline (64-bit, 3.5.0.9319) was used to calculate the lipid
transition list, visualize results, integrate signals over the time,
and quantify all lipids that were detected by MS.16
Sphingolipid Analysis. High-performance liquid chromatog-
raphy−electrospray ionization tandem mass spectrometry
(HPLC−ESI-MS/MS) was conducted using an UltiMate
3000 instrument (Thermo Fischer Scientific, Darmstadt,
Germany) in conjunction with a Q Exactive Plus mass
spectrometer (Thermo Fisher Scientific, Bremen, Germany).
Chromatographic separation was accomplished with an
Ascentis Express C18 column (150 mm × 2.1 mm, 2.7 μm;
Supelco, Bellefonte, PA) fitted with a guard cartridge (50 mm ×
2.1 mm, 2.7 μm; Supelco). The temperatures of the
autosampler and the column oven were set at 10 and 60 °C,
respectively. All gradients and LC parameters can be found in
Table S1. The injector needle was automatically washed with
30% B and 0.1% phosphoric acid prior to each injection.
Samples were injected in a volume of 5 μL. The use of
phosphoric acid did not affect LC or MS instrumentation over a
running period of three years. The Q Exactive Plus instrument
was configured to perform a high-resolution MS full scan (HR-
FS) and PRM in one measuring cycle. Full scan settings were as
follows: m/z 250−900 [positive ion detection, resolution of
70000 (m/z 200)]; AGC (automatic gain control), 3 × 106;
maximum injection time, 50 ms. For PRM, spectra were
acquired as follows: isolation windows, m/z 0.5 [mass
resolution of 17500 (m/z 200); AGC, 1 × 105; maximum
Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry Article

Figure 1. Integrated approach for sphingolipid profiling, including (A) alkaline extraction for sphingolipids, (B) an optimized gradient for LC
separation, (C) parallel reaction monitoring and full scan for MS scanning, and (D) critical criteria for identification.

Figure 2. Comparison of total ion chromatograms (TICs) of (A) non-alkaline and (B) alkaline MTBE extraction for sphingolipid analysis.
Abbreviations: PA, phosphatidic acid; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PS, phosphatidylserine; PG, phosphatidylglycerol;
DG, diacylglycerol; TG, triacylglycerol; SM, sphingomyelin; Cer, ceramide; HexCer, hexosyl ceramide.

injection time, 100 ms]. The normalized collision energy
(NCE) was optimized for each sphingolipid class by direct
infusion of a corresponding standard. An inclusion list of
precursors was used for scheduled PRM data acquisition
(Tables S2 and S3).
classes.26−28 In recent years, MTBE extraction has been shown
to deliver similar or even better recoveries of most lipid species,
including sphingolipids.20 However, original protocols were
designed to isolate all lipid classes together, with the
consequence that there are frequently false positive identi-

■ RESULTS AND DISCUSSION

SRM of lipid extracts using a triple-quadrupole mass
spectrometer has been the method of choice for targeted
sphingolipid analysis. However, technical limitations of the
instrumentation in combination with the high level of
complexity of lipid extracts compromise accurate identification
and quantification of sphingolipids. The major issues associated
with this strategy are interference (ion suppression and isotope
overlapping) from abundant phospholipids, nearly isobaric
sphingolipid species that cannot be distinguished on the basis
of low resolution masses, common fragment ions derived from
key structural features (such as the LCB), the overlap of natural
isotope clusters due to varied degrees of unsaturation among
isobars, and chromatographic co-elution of isobaric sphingoli-
pids. To overcome these challenges, we present here an
improved workflow that includes an optimized extraction
procedure, a segmental linear HPLC gradient, and a HR-FS/
PRM approach, resulting in comprehensive sphingolipid
analysis (Figure 1).
Establishment of a Comprehensive Profiling Strategy
for Sphingolipids. Comparison of Extraction Methods for
Sphingolipids. An effective extraction strategy is essential for
obtaining a comprehensive sphingolipid profile. Over five
decades, CHCl3/MeOH extraction has been the most widely
used protocol for lipid analysis,25 although several modified
methods have been reported for selected sample types and lipid
C DOI: 10.1021/acs.analchem.7b03576
fications for sphingomyelin (SM) due to the presence of
glycerophosphatidylcholines that have the same phosphocho-
line headgroup and overlap with the natural isotope pattern of
SM when analyzed on a triple-quadrupole mass spectrometer.
An alkaline hydrolysis step has been used in conjunction with
CHCl3/MeOH or MTBE extraction as a way to hydrolyze all
ester bonds and deplete the samples of phospholipids and/or
glycerolipids as well as reduce matrix effects.12,20 Because of an
increased solubility in water, the hydrolyzed fragments are not
retained on the stationary phase. To evaluate the consequences
of alkaline hydrolysis, we benchmarked four different extraction
methods against each other for extraction of sphingolipids from
RAW 264.7 cells: CHCl3/MeOH extraction with and without
alkaline hydrolysis and MTBE with and without alkaline
hydrolysis. We chose SM d18:1/16:0, a common sphingomye-
lin species, to demonstrate the resolution of overlap with PC
species. Two characteristic fragments were selected for
monitoring: the phosphocholine headgroup (m/z 184.1) and
the LCB (m/z 264.3). While there was no significant difference
in the intensity of the LCB fragment because of the inclusion of
alkaline hydrolysis, the chromatographic traces of the
phosphocholine fragment derived from SM were considerably
improved by eliminating co-eluting PCs (Figure S1A,B). As can
be seen from the total ion chromatogram (TIC) derived from
the high-resolution full scan (HR-FS) in Figure 2, hundreds of
co-eluting phospholipids and glycerolipids between 6 and 10
min were eliminated by alkaline hydrolysis.18 Although the
Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry Article

Figure 3. Comparison of LC methods for sphingolipid separation. Panels from left to right present ion chromatograms at the precursor and fragment
level. Mass errors are indicated with color-coded text next to corresponding signals. For analysis, reversed-phase chromatography [Ascentis Express
C18 column (150 mm × 2.1 mm, 2.7 μm)] and full scan PRM methods on an Q Exactive Plus instrument are applied. (A) Gradient I used MeOH/
IPA/H2O solvent compositions from ref 23 to separate sphingolipids. (B and C) Gradients II and III used the same ACN/IPA/H2O solvent
composition but different gradients. The parameters and the gradient composition are listed in Table S1. ① indicates lipid species Cer d18:1/24:1,
while ② indicates Cer d18:2/24:0.

alkaline solution hydrolyzes amides at a slower rate, precautions
should be taken and experiments carefully planned and
designed. On the basis of a comparison of internal standards
with and without base hydrolysis, no significant hydrolysis was
observed for the main sphingolipid classes Cer, HexCer,
DiHexCer, and SM (Figure S2). A notable advantage of the
MTBE protocol is that it is less time-consuming than CHCl3/
MeOH extraction (only ∼4 h for the preparation of lipid
extracts compared to at least 12 h), and it permits easier
collection of the organic phase and precipitated proteins.29
Quantitative results obtained from inclusion of internal
standards indicated that MTBE alone or combined with the
alkaline treatment consistently resulted in an extraction
efficiency that is better than that of CHCl3/MeOH or
CHCl3/MeOH extraction with alkaline treatment for the
more hydrophobic species with longer fatty acids such as SM
d18:1/24:0 (7859 ± 614.7 pmol/mg of protein vs 886.9 ±
88.18 pmol/mg of protein) and Cer d18:1/24:0 (1890 ± 169.9
pmol/mg of protein vs 189.7 ± 29.91 pmol/mg of protein).
For relatively hydrophilic species such as So 18:1, MTBE
extraction showed the same extraction ability as CHCl3/MeOH
extraction (Figure S1C). The results of CHCl3/MeOH
extraction with alkaline hydrolysis yielded the same concen-
tration as MTBE with base hydrolysis (Figure S1C), and
MTBE extraction with or without base hydrolysis exhibited an
overall lower level of variability. Addition of a base hydrolysis
step to the MTBE method eliminates lipid species that co-elute
D DOI: 10.1021/acs.analchem.7b03576
with sphingolipids by RP-HPLC, thereby reducing the level of
background interference and yielding better signal-to-noise
ratios (Figure S1A,B), which is particularly important for
identification and quantification of low-abundance sphingoli-
pids. As such, this is the extraction method of choice for
sphingolipid analysis. The preliminary sphingolipid analysis was
accomplished by scanning known sphingolipid species in RAW
264.7 cells in SRM mode using a QTRAP 6500 instrument.
Although extraction efficiency is independent of the mass
spectrometer used for the analysis, SRM-based quantification
on a triple quadrupole is limited for nearly isobaric species. It is
also challenging to scan for all potential lipid-derived fragment
ions of sphingolipids due to a limitation in the number of
transitions that can be monitored in SRM mode. Furthermore,
the lack of high-resolution/high-mass accuracy information for
the precursors analyzed on a triple-quadrupole instrument can
hinder correct interpretation of results. Therefore, subsequent
investigations were accomplished with HRMS by applying FS
and PRM.
Comparison of HPLC Methods for Sphingolipid Analysis.
An efficient HPLC separation is essential for increasing the
specificity and sensitivity of detection of sphingolipids. On the
basis of the results of previous studies,21−23 we evaluated
different reversed-phase separation strategies that are currently
used in the field for sphingolipid analysis (Figure 3). The
results from two (Table S1) of these are reported here. All
three combinations performed similarly for HexCer and
Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
DiHexCer species, while I [resolution (R), 100%] and III (R,
115−409%) yielded separation that was better than that of II
(R, 91−230%) for ceramides. III offered baseline separation
that was better than that of I for Cer d18:1/24:1 and Cer
d18:2/24:0 (R, 116%). III also offered the best separation for
SM with the least co-elution of isobaric species, as verified by
the elution pattern of characteristic SM fragments and R
(Figure 4 and Figures S3 and S4). The results of our HPLC
Figure 4. Comparison of LC methods for sphingomyelin. (A) Ion
chromatogram of the d18:1 LCB for SM d18:1/16:0 (①) and
interfering precursor ion (②). Mass errors are colored red. (B) PRM
MS/MS spectrum of ① and ②.
gradient evaluations indicated that use of a segmented linear
gradient (III) is optimal and faster for sphingolipid separation
and is likely to be an effective option for lipidomics in general.
MS Approach for Sphingolipid Profiling. The use of a high-
resolution/high-mass accuracy mass spectrometer is particularly
important for assuring the validity of targeted sphingolipid
analyses because of the presence of multiple nearly isobaric
species (Figure 5A). For example, differentiation of Cer d18:1/
23:0 [molecular weight (MW) of 635.6216] from Cer d18:1/
h22:1 (MW of 635.5852) requires a resolving power of >17000
and a mass accuracy better than 57 ppm. However, even though
the two nearly isobars can be clearly identified in MS1, they
fragment together because of the 0.5 Da isolation window
during precursor ion selection in PRM. Because there are
multiple product ions in the majority of MS2 spectra of
sphingolipids (Table S4 and Figure S5), the use of
fragmentation patterns in conjunction with an accurate mass
is the most effective way to accurately identify a sphingolipid
species (Figure 5B). SRM has been the method of choice in
sphingolipid analyses, but use of SRM can be problematic
because the full MS2 profile is not acquired. In particular, if
only the transition generating the LCB is monitored on a triple-
quadrupole mass spectrometer, it is not possible to detect the
presence of nearly isobaric sphingolipids (e.g., HexCer d18:1/
16:0 and CerP d18:1/22:1 that contain the same LCB
fragment). This can be remedied through the use of PRM.
The complete fragmentation pattern (Figures S5 and S6) can
be utilized for confident identification within an individual
sphingolipid class, supporting the distinction of isobaric
precursors such as Cer d18:1/22:0 and Cer d18:0/22:1 (Figure
E DOI: 10.1021/acs.analchem.7b03576
Article
5C). The existence of fragment-level isobars with overlapping
isotopes is an additional confounding problem that cannot be
solved chromatographically. As such, a targeted method was
established in concert with high-resolution MS to detect and
quantify the various molecular species (Figure 5D).
Confident identification of sphingolipids at fatty acid scan
species level17 requires the use of at least two or three
fragments: the class specific fragment (if available), the LCB,
and the fatty acid to unambiguously identify the lipid molecule.
High-resolution survey scans in conjunction with PRM offer
the possibility of determining the exact mass of the intact
sphingolipid and a complete MS2 spectrum to validate the
sphingolipid identity. For identification, it is essential that the
mass errors of precursor ions and/or fragment ions be below 5
ppm and that there be a match of the fragmentation pattern to
that of an authentic standard. For Cer, HexCer, DiHexCer,
sulfatides, and gangliosides, a group of fragments is associated
with the LCB (Figure S5, red font). We call this group the
“LCB pattern”, with the fragments designated as W′, W″, and
W′-CHO. These fragments are generated in the course of HCD
(higher-energy collisional dissociation) and can be used to
identify sphingolipids using MS2 spectra (Figures S4 and S8).
Knowledge that the ratios of fragment intensities (W′:W″:W′-
CHO) should be approximately 10:100:15 for an unsaturated
LCB and 2:50:100 for a saturated LCB is important for
determining if there is co-isolation or co-elution of any
sphingolipid species (Figure S5 and Table S5). The ratio
between different transition intensities, the so-called qualifier
ion ratio (QIR), is used to support unequivocal molecule
verification.30,31 The known QIR value (derived from analysis
of an authentic standard) should always be achieved by well-
separated lipids. It is worth noting, on the basis of analysis of
authentic standards, we have found that this set of fragments is
not generated for SM and CerP. Therefore, for SM and CerP,
along with the exact mass and the LCB fragments, additional
ions are required to identify and differentiate these two lipid
classes from other sphingolipids: the phosphocholine head-
group for SM (m/z 184.0733) and neutral loss of the
phosphate group of CerP (loss of 97.9769 Da) (Figure S7).
For SM identification, a further fragment can be observed when
the correct collision energy is used (34 V for HCD on our Q
Exactive Plus instrument). Under these conditions, a low-
intensity fatty acid carboxylate anion [m/z 199.1705 for SM
d18:1/12:0 (m/z 199.1698 calculated, +3.5 ppm) and m/z
255.2330 for SM d18:1/16:0 (m/z 255.2324 calculated, +2.4
ppm)] is detected in negative ion mode (Figure S8). This
fragment is presumably generated via an intramolecular
rearrangement (retro-heteroene reaction) involving the oxygen
on the LCB, because the close agreement between calculated
and observed masses clearly indicates the presence of two
oxygens.32 Although the intensity for the carboxylate fragment
is relatively low, it is detectable for medium- to high-abundance
species (such as SM d18:1/16:0) and provides valuable
evidence for structure verification.
Sphingolipid Profile in RAW 264.7 Cells. To illustrate
the effectiveness of our sphingolipid workflow, we analyzed the
sphingolipids in RAW 264.7 cells and assessed the effect of
KLA on sphingolipid metabolism. KLA treatment of RAW
264.7 cells has been previously reported to induce autophagy
and to increase the levels of Cer by specifically activating the
Toll-like 4 receptor (TLR-4).33,34 During autophagy, micro-
tubule-associated protein 1 light chain 3 (LC3) is involved in
the formation of autophagosomal membranes. LC3 has two
Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry Article

Figure 5. Current challenges in sphingolipid analysis. (A) In challenge I, nearly isobaric species from low-resolution- and mass accuracy-derived data,
e.g., from triple-quadrupole MS, are not always sufficient to resolve closely related species. (B) In challenge II, near isobars are sharing fragment ions.
The chemical formulas (C40H83ON2P3, C43H76O5N2, C44H76O6, and C43H77O3N2P) are examples that also fit the detected precursor masses within a
5 ppm mass error. Collecting multiple fragments rather to rely on a single diagnostic fragment reduces false positive identification. Liquid
chromatographic separation is required. (C) In challenge III, isobars overlap with isotopes. The PRM experiment via high-resolution MS showed
disagreement with SRM via triple-quadrupole MS. The PRM overcomes SRM limitation by providing full MS2 spectra with multiple fragments with
high resolution and mass accuracy to calculate transition intensity ratios (QIRs) by peak areas (a and b). (D) In challenge IV, isobaric species are at
the precursor and fragment level. This challenge can be overcome only with the best possible resolution at LC, a full scan, and a PRM approach.
Abbreviations: M, exact mass; EIC, extracted ion chromatogram.

Figure 6. KLA-induced autophagy in RAW 264.7 cells. (A) The unmodified and lipidated version of LC3 (LC3-II), as autophagy protein markers, is
detected by Western blot. Tubulin staining serves as a loading control for the LC3-II protein. (B) Fold change in the relative amount of LC3-II in
control and KLA-treated cells (based on densitometric quantification of three biological replicates). (C) Selected regulated sphingolipid species. All
sphingolipids are listed in Figure S9. All experiments were performed in biological triplicate. *p < 0.05.

F DOI: 10.1021/acs.analchem.7b03576
Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
forms, cytosolic LC3-I and membrane-bound LC3-II.35 LC3-II
is a lipidated form of LC3 that has been shown to be an
autophagosomal marker in mammals, providing evidence that
autophagy is induced after KLA treatment.33 We monitored
lipidated LC3-I and LC3-II by Western blot after KLA
treatment and used tubulin as a loading control.36 Our results
are in good agreement with earlier studies that observed an
increase in the level of LC3-II (Figure 6B) and an increase in
the LC3-II:LC3-I ratio between treatment and control (Figure
6A). This indicates induction of autophagosome formation in
KLA-treated RAW264.7 cells. We found that the levels of most
of the Cer species were increased after KLA treatment relative
to SM (Figure 6C and Figure S9). Through use of our
workflow, we identified more sphingolipid species with high-
quality spectra and high confidence than in previous
reports.12,18 We were able to detect and quantify 50% more
Cer species than earlier studies did.12,18 Cer levels were also
altered after KLA treatment (Figures 6C and 7). The number
Figure 7. Comparison of identified lipid species in the literature and
this work. Black bars represent data from the study presented here and
light gray and dark gray bars data from earlier studies. We were able to
detect and quantify 50% more Cer species than earlier studies did. The
number of HexCer, SM, and CerP species identified by our workflow
is lower than in other reports. However, there is a high level of
confidence associated with our identifications, because they are based
on high-resolution/high-mass accuracy precursor and fragment ion
data in addition to complete fragmentation information at the MS2
level.
of HexCer, SM, and CerP species identified by our workflow is
technically lower than in other reports. However, there is a high
level of confidence associated with our identifications, because
they are based on high-resolution/high-mass accuracy precursor
data and the complete tandem mass spectral pattern. In
contrast, only the LCB (i.e., one fragment) that is shared
among many sphingolipid species was used for sphingolipid
identification in the referenced papers. Use of the sphingolipid
fragmentation pattern in combination with high-resolution/
high-mass accuracy precursor measurements increases the
reliability substantially, because many false positive lipids are
removed from consideration. In total, we identified and
quantified 61 sphingolipid species within a dynamic range of
7 orders of magnitude with detection ranging to the low
femtomole per milligram protein level. In agreement with
earlier studies of the same cell type,18,37 sulfatides were not
identified after KLA treatment.
G DOI: 10.1021/acs.analchem.7b03576
Article
■ CONCLUSIONS
A RPLC/HRMS-based workflow has been developed that
significantly increases the accuracy in sphingolipidomics by
resolving isobaric and nearly isobaric species through use of an
efficient MTBE-based extraction approach that includes alkaline
hydrolysis that removes interfering lipids, a segmented linear
gradient that has been tailored for sphingolipids and effectively
separates closely eluting sphingolipid species, and an MS
strategy that includes HRMS survey scans and PRM MS
detection. Our workflow provides enhancements in sample
preparation and analysis that yield a higher level of confidence
for sphingolipid identification and quantification. Application of
the workflow to analysis of sphingolipids in RAW 264.7 cells
demonstrates its utility for samples of biological origin.
■
ASSOCIATED CONTENT
*
S Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.anal-
chem.7b03576.
Tables S1−S5, Figures S1−S9, and additional informa-
tion as mentioned in the text (PDF)
■ AUTHOR INFORMATION
Corresponding Author
*Leibniz-Institut für Analytische Wissenschaften-ISAS-e.V.,
Otto-Hahn-Str .6b, 44227 Dortmund, Germany. Phone: +49-
231-1392-4173. Fax: +49-231-1392-4173. E-mail: robert.
ahrends@isas.de.
ORCID
Cristina Coman: 0000-0002-3771-2410
Dominic Winter: 0000-0001-6788-6641
Robert Ahrends: 0000-0003-0232-3375
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
This study was supported by the Ministerium für Innovation,
Wissenschaft und Forschung des Landes Nordrhein-Westfalen,
the Senatsverwaltung für Wirtschaft, Technologie und For-
schung des Landes Berlin, and the Bundesministerium für
Bildung und Forschung and by the BMBF de.NBI program
(code 031L0108A). The authors thank Dr. Christian Hellmuth
for valuable comments on the manuscript.
■ REFERENCES
(1) van Meer, G.; de Kroon, A. I. P. M. J. Cell Sci. 2011, 124, 5−8.
(2) van Meer, G.; Lisman, Q. J. Biol. Chem. 2002, 277, 25855−25858.
(3) Trajkovic, K.; Hsu, C.; Chiantia, S.; Rajendran, L.; Wenzel, D.;
Wieland, F.; Schwille, P.; Brugger, B.; Simons, M. Science 2008, 319,
1244−1247.
(4) Lingwood, D.; Simons, K. Science 2010, 327, 46−50.
(5) Aït Slimane, T.; Hoekstra, D. FEBS Lett. 2002, 529, 54−59.
(6) Meikle, P. J.; Summers, S. A. Nat. Rev. Endocrinol. 2016, 13, 79−
91.
(7) Fyrst, H.; Saba, J. D. Nat. Chem. Biol. 2010, 6, 489−497.
(8) Hannun, Y. A.; Obeid, L. M. Nat. Rev. Mol. Cell Biol. 2008, 9,
139−150.
(9) Rosen, H.; Gonzalez-Cabrera, P. J.; Sanna, M. G.; Brown, S.
Annu. Rev. Biochem. 2009, 78, 743−768.
(10) Lopez, P. H.; Schnaar, R. L. Curr. Opin. Struct. Biol. 2009, 19,
549−557.
Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry Article

(11) Breslow, D. K.; Weissman, J. S. Mol. Cell 2010, 40, 267−279.

(12) Shaner, R. L.; Allegood, J. C.; Park, H.; Wang, E.; Kelly, S.;
Haynes, C. A.; Sullards, M. C.; Merrill, A. H., Jr J. Lipid Res. 2009, 50,
1692−1707.
(13) Basit, A.; Piomelli, D.; Armirotti, A. Anal. Bioanal. Chem. 2015,
407, 5189−5198.
(14) Scherer, M.; Bottcher, A.; Schmitz, G.; Liebisch, G. Biochim.
Biophys. Acta, Mol. Cell Biol. Lipids 2011, 1811, 68−75.
(15) Weir, J. M.; Wong, G.; Barlow, C. K.; Greeve, M. A.; Kowalczyk,
A.; Almasy, L.; Comuzzie, A. G.; Mahaney, M. C.; Jowett, J. B.; Shaw,
J.; Curran, J. E.; Blangero, J.; Meikle, P. J. J. Lipid Res. 2013, 54, 2898−
2908.
(16) Peng, B.; Ahrends, R. J. Proteome Res. 2016, 15, 291−301.
(17) Liebisch, G.; Vizcaino, J. A.; Kofeler, H.; Trotzmuller, M.;
Griffiths, W. J.; Schmitz, G.; Spener, F.; Wakelam, M. J. J. Lipid Res.
2013, 54, 1523−1530.
(18) Dennis, E. A.; Deems, R. A.; Harkewicz, R.; Quehenberger, O.;
Brown, H. A.; Milne, S. B.; Myers, D. S.; Glass, C. K.; Hardiman, G.;
Reichart, D.; Merrill, A. H., Jr.; Sullards, M. C.; Wang, E.; Murphy, R.
C.; Raetz, C. R.; Garrett, T. A.; Guan, Z.; Ryan, A. C.; Russell, D. W.;
McDonald, J. G.; et al. J. Biol. Chem. 2010, 285, 39976−39985.
(19) Merrill, A. H., Jr.; Sullards, M. C.; Allegood, J. C.; Kelly, S.;
Wang, E. Methods 2005, 36, 207−224.
(20) Matyash, V.; Liebisch, G.; Kurzchalia, T. V.; Shevchenko, A.;
Schwudke, D. J. Lipid Res. 2008, 49, 1137−1146.
(21) Bird, S. S.; Marur, V. R.; Sniatynski, M. J.; Greenberg, H. K.;
Kristal, B. S. Anal. Chem. 2011, 83, 940−949.
(22) Hu, C.; van Dommelen, J.; van der Heijden, R.; Spijksma, G.;
Reijmers, T. H.; Wang, M.; Slee, E.; Lu, X.; Xu, G.; van der Greef, J.;
Hankemeier, T. J. Proteome Res. 2008, 7, 4982−4991.
(23) Wang, J. R.; Zhang, H.; Yau, L. F.; Mi, J. N.; Lee, S.; Lee, K. C.;
Hu, P.; Liu, L.; Jiang, Z. H. Anal. Chem. 2014, 86, 5688−5696.
(24) Moruz, L.; Pichler, P.; Stranzl, T.; Mechtler, K.; Kall, L. Anal.
Chem. 2013, 85, 7777−7785.
(25) Bligh, E. G.; Dyer, W. J. Can. J. Biochem. Physiol. 1959, 37, 911−
917.
(26) Steinhauer, J.; Gijon, M. A.; Riekhof, W. R.; Voelker, D. R.;
Murphy, R. C.; Treisman, J. E. Mol. Biol. Cell 2009, 20, 5224−5235.
(27) Ejsing, C. S.; Bilgin, M.; Fabregat, A. PLoS One 2015, 10,
No. e0144817.
(28) Zhang, T.; Walensky, L. D.; Saghatelian, A. ACS Chem. Biol.
2015, 10, 1398−1403.
(29) Coman, C.; Solari, F. A.; Hentschel, A.; Sickmann, A.; Zahedi, R.
P.; Ahrends, R. Mol. Cell. Proteomics 2016, 15, 1453−1466.
(30) Hellmuth, C.; Weber, M.; Koletzko, B.; Peissner, W. Anal.
Chem. 2012, 84, 1483−1490.
(31) Delatour, T.; Mottier, P.; Gremaud, E. J. Chromatogr. A 2007,
1169, 103−110.
(32) Demarque, D. P.; Crotti, A. E. M.; Vessecchi, R.; Lopes, J. L. C.;
Lopes, N. P. Nat. Prod. Rep. 2016, 33, 432−455.
(33) Sims, K.; Haynes, C. A.; Kelly, S.; Allegood, J. C.; Wang, E.;
Momin, A.; Leipelt, M.; Reichart, D.; Glass, C. K.; Sullards, M. C.;
Merrill, A. H., Jr J. Biol. Chem. 2010, 285, 38568−38579.
(34) MacKichan, M. L.; DeFranco, A. L. J. Biol. Chem. 1999, 274,
1767−1775.
(35) Kabeya, Y.; Mizushima, N.; Ueno, T.; Yamamoto, A.; Kirisako,
T.; Noda, T.; Kominami, E.; Ohsumi, Y.; Yoshimori, T. EMBO J. 2000,
19, 5720−5728.
(36) Barth, S.; Glick, D.; Macleod, K. F. J. Pathol. 2010, 221, 117−
124.
(37) Koberlin, M. S.; Snijder, B.; Heinz, L. X.; Baumann, C. L.;
Fauster, A.; Vladimer, G. I.; Gavin, A. C.; Superti-Furga, G. Cell 2015,
162, 170−183.

H DOI: 10.1021/acs.analchem.7b03576
Anal. Chem. XXXX, XXX, XXX−XXX