Quantitation of Multiple Sphingolipid Classes Using Normal and Reversed-Phase LC–ESI–MS/MS: Comparative Profiling of Two Cell Lines

Abstract

Sphingolipids are an important class of compounds that regulate signal transduction and other vital cellular processes. Herein, we report sensitive normal and reversed phase LC–MS/MS methods for quantitation of multiple sphingolipid classes. In the normal-phase ESI/MS/MS method, a high content of organic solvents was utilized, which, although it included hexane, ethyl acetate, acetonitrile containing 2% methanol, 1–2% acetic acid, and 5 mM ammonium acetate, resulted in a very efficient electrospray ionization of the ceramides (Cers) and hexosylceramides (MHCers). Three normal-phase LC–MS/MS methods using segmented phases were developed to specifically target Cers, MHCers, or sphingomyelins (SMs). This segmentation scheme increases the number of data points acquired for a given analyte and enhances the sensitivity and specificity of the measurements. Nine separate reversed phase chromatography methods were developed for the three classes of compounds. These assays were used for comparing the levels of Cers, SMs, and MHCers from mouse embryonic fibroblast (pMEF) and human embryonic kidney (HEK293) cells. These findings were then compared with the reported data from RAW264.7 mouse macrophage cells, BHK21 hamster cells, and human plasma and serum samples. The analysis of cell lines, using both normal and reversed phase chromatography, revealed discrimination based on the type of chromatography chosen, while sphingolipid assays of samples containing different amounts of protein showed different results, even after normalizing for protein content. Also, LC/MS/MS profiles were provided for the classes and individual compounds so that they could be used as “molecular profiles” for class or individual sample analysis.

Introduction

Sphingolipids, such as ceramides (Cers), hexosylceramides (MHCers), and sphingomyelins (SMs), are an important class of compounds that play vital roles in cell architecture and function [1, 2]. They are highly conserved and are involved in a number of metabolic functions and signal transduction cascades [3–9]. Abnormalities in these compounds are associated with a variety of human diseases, including inborn errors of metabolism, diabetes, atherosclerosis, arrhythmias, myocardial infarction, Alzheimer’s, and other neurodegenerative diseases [4, 10–24]. Importantly, certain types of human malignancies, including breast, prostate, colon, ovary, and endometrium cancers, exhibit high levels of fatty acid synthase expression and activity, implicating a role for these compounds in the biology of cancer [25–29]. Considering their role in diseases, these Cers, MHCers, and SMs may represent reasonable biomarkers for providing diagnostic or prognostic information [17, 21, 30–33]. For example, sphingolipids can be potentially used as biomarkers for Alzheimer’s disease or other neurodegenerative disorders, disease progression, and therapeutic intervention [30].

Owing to their chemical diversity, no single analytical method is sufficient for studying all the analytes of each of the sub-classes/classes of the lipids. Liquid chromatography–tandem mass spectrometry (LC–MS/MS) methods for sphingolipid analysis, utilizing both electrospray (ESI) and atmospheric pressure chemical ionization (APCI) [34–44], have been reported. It is well established that each class of sphingolipids (i.e., Cers, SMs, and MHCers; the representative compounds with their product ions are shown in Fig. 1), does not separate into individual components by normal-phase chromatography. Instead, all components from a distinct class (e.g., Cers) will move as a single head-group when normal phase chromatography is used. In reversed-phase chromatography, however, each group of compounds (e.g., Cers) tends to separate into individual components [34–39].

Fig. 1.

Structures for some of the representative ceramides, hexosylceramides, and sphingomyelins and their associated specific product ions that were monitored. A core chain for a sphingoid base can be saturated or unsaturated and contains a 2-amino-1,3-dihydroxy moiety, and a notation d18:1^Δ4; for example, a dihydroxy 18 carbon core chain with a double bond on the fourth carbon atom starting from the hydroxyl terminus of the core chain. The notation c18:0 is for the number of carbon atoms on the N-acyl side chain. All d18:1^Δ4 and d18:0 ceramides, hexosylceramides, and sphingomyelins can form specific carbocation product ion fragments of m/z 264 and 266, respectively. Sphingomyelins will form a highly sensitive product ion choline phosphate of m/z 184

Herein we report specific normal-phase and reversed-phase LC–MS/MS methods, where the individual component of analytes from each class of sphingolipids is found to exhibit high sensitivity and good individual peak separations. The sensitivity and specificity of these assays were enhanced because in each method only a few analyte transitions were incorporated that allowed sufficient time to collect more data points on each analyte, and all individual components of analytes exhibited good selectivity. For analysis of analytes by normal-phase ESI/MS utilizing non-polar solvents such as hexane, the organic phase is mixed with 5% isopropanol and ≥50 mM buffer salt [43,44]. In the present context of normal phase chromatography, a high content of organic solvents that includes hexane, ethyl acetate, acetonitrile containing 2% methanol, 1% acetic acid, and 5 mM ammonium acetate was utilized, yet this method resulted in a very efficient electrospray ionization of the ceramides and hexosylceramides. For normal phase assay, three MS/MS methods were developed, and each of these methods was segmented to include only specifically targeted analytes from each sphingolipid class according to their chromatographic retention times. These three methods were sufficient to incorporate the entire range of analytes from the Cers, SMs, and MHCers. Due to the wider chromatographic selectivity of reversed-phase chromatography, a total of nine methods, three for each individual class of compounds, was developed. Subsequently, these optimized analytical methods were applied to quantitate and compare cellular sphingolipid contents from mouse embryonic fibroblasts (pMEF) and human embryonic kidney (HEK293) cells. We also show the evidence of quantitative discrimination related to the type of chromatography used, and the amount of protein present in compared samples.

Experimental Section

Materials

The following were purchased from Avanti Polar Lipids, Inc., Alabaster, AL. N-arachidonoyl-d-erythrosphingosine (Cer d18:1^Δ4 c20:0); N-nervonoyl-d-erythrosphingosine (Cer d18:1^Δ4 c24:1), N-palmitoyl-d-erythro-sphingosyl-phosphorylcholine (SM d18:1^Δ4 c16:0); N-stearoyl-d-erythro-sphingosylphosphorylcholine (SM d18:1^Δ4 c18:0); and LM 6002 internal standard, which is a 25-μM mixture of 10 different sphingolipids; Sphingosine (C17:1 base); Sphinganine (C17:0 base); Sphingosine-1-PO₄ (C17:1 base); Sphinganine-1-PO₄ (C17:0 base); Lactosyl(β) C12 Ceramide; 12:0 Sphingomyelin; Glucosyl(β) C12 Ceramide; 12:0 Ceramide; 12:0 Ceramide-1-PO₄; and 25:0 Ceramide. Reagents and all HPLC-grade solvents were purchased from either Fisher Scientific (Pittsburgh, PA) or Sigma-Aldrich (St. Louis, MO). The HEK293 cell line was obtained from the American Type Culture Collection (ATCC; [#CRL-10892]), designation 293 C18.

HEK293 Cell Growth and Lysis

HEK293 cells were grown in suspension, using serum-free medium. After harvesting, the cells were washed three times in PBS buffer and the cell pellets were stored at −80 °C. To obtain the lysate, cells were resuspended in 50 mM NH₄HCO₃ (pH 7.9), placed in a Dounce homogenizer, and homogenized by 10 gentle strokes. The supernatant was collected after centrifugation at low speed (500×g) for 15 min, and assayed for protein content using a BCA assay.

Isolation of Primary Mouse Embryonic Fibroblasts (pMEF)

The cells were isolated from pregnant mice as follows. E 10.5 dpc (day post-coitum) embryos were separated from the yolk sac, washed twice with PBS buffer (pH 7.2), and trypsinized for 10 min. The cells were disaggregated by pipetting several times and suspended in 1 mL of pre-warmed (37 °C) DMEM supplemented with 10% FBS. The suspension was allowed to stand for 2–3 min at room temperature, and the visible clumps were removed. The suspension was centrifuged at 500×g for 3 min, and the cell pellet was collected, resuspended in MEF media (DMEM + 10% FBS), and plated on a 6-well plate. The plates were incubated at 37 °C with 3% O₂ and 5% CO₂ [45]. The medium was replaced the next day, and the cells cultured for 4–5 days. Forty-eight hours before extracting the lipids, the cells were shifted to serum-free DMEM medium. Total cell lysate was prepared by homogenizing the cells by applying 20 strokes using a Dounce homogenizer and homogenization buffer (50 mM tris, pH 7.4, 1 mM EDTA, 50 mM NaCl). Cell debris was removed using low-speed centrifugation (500×g for 15 min) and the sample assayed for protein content using BCA assay.

Preparation of Sphingolipid Extracts

Lipid extracts were prepared according to the method described by Merrill et al. [37–39]. In brief, to 50 μL of pMEL cell extract containing ~250 μg of protein, or 1 × 10⁷ of HEK293 cell lysate taken in a 4-mL Teflon-capped glass vial (Lisher scientific, cat # W 224582) were added 0.5 mL of methanol (CH₃OH), 0.250 mL of chloroform (CHC1₃), 50 μL of water (H₂O), and 30 μL of 25 μM solution of LM-6002, providing 750 pmol of each of its individual constituent internal standards. The lipid aggregates were dispersed by sonicating 6 times, using a Branson tip sonicator at an amplitude setting of 21% for 10 s each, with a resting interval of 5 s between pulses. The samples were further sonicated for 1 h, using a bench top sonicator, and incubated overnight at 48 °C with shaking. After cooling the sample to ambient temperature, 75 μL of 1 M methanolic potassium hydroxide (KOH) was added, followed by incubation at 37 °C for 2 h with shaking to hydrolyze glycerophospholipids. The sample solution was divided into two equal aliquots. One aliquot was neutralized by the addition of 7 μL of glacial acetic acid (CH₃COOH) and 2 mL of water. The mixture was extracted twice, using two 1.2-mL volumes of CHC1₃. The lower organic portions were collected, combined, and evaporated to dryness using a Savant Speed Vac Concentrator (GMI, Ramsey, MN). The dried sample was re-suspended in 300–400 μL of 1:3 (v/v) CHCL₃ and normal-phase A (5 mM CH₃COONH₄ dissolved in 20 mL CH₃OH, 15 mL CH₃COOH, 270 mL CH₃CN, 300 mL CH₃COOCH₂CH₃, and 400 mL hexanes). The other half of the sample was evaporated to approximately 25 μL, using a Speed Vac Concentrator, and reconstituted by adding 300–400 μL of 1:1 (v/v) reversed mobile phase A (74/25/1 [v/v/v] H₂O:CH₃OH:HCOOH containing a final concentration of 5 mM HCOONH₄) and reversed mobile phase B (99:1 [v/v] CH₃OH:HCOOH containing a final concentration of 5 mM HCOONH₄). The sample was vortexed for 1 min, followed by centrifugation at 13,200×g (Eppendorf Centrifuge 5415 D) for 2 min, and the supernatant was transferred to a fresh Eppendorf tube.

Liquid Chromatography–Tandem Mass Spectrometry

Liquid chromatography–tandem mass spectrometry (LC–MS/MS) analysis was performed using a TSQ Discovery triple quadrupole mass spectrometer (Thermo Electron Corp., San Jose, CA, USA) equipped with an electrospray ionization (ESI) source. The mass spectrometer was coupled to an Agilent 1100 series HPLC system fitted with a 100-μL/min max flow controller and a 40-μL sample injection loop.

Normal-phase chromatography was performed using a binary system and a 7.5 cm × 3.0 mm × 3 μm Supelcosil LC-NH₂ column operating at a flow rate of 100 μL/min and maintained at 37 °C. Initially, a high flow rate method was tried using this column; however, at the slow flow rate of 100 μL/min, good optimal results were observed. The mobile phase buffer A consisted of 5 mM CH₃COONH₄ dissolved in 20 mL CH₃OH, 10 mL CH₃COOH, 270 mL CH₃CN, 300 mL CH₃COOCH₂CH₃, and 400 mL hexanes. Mobile phase buffer B consisted of 5 mM CH₃COONH₄ in 99:1 (v/v) CH₃OH: CH₃COOH. Approximately 10 μL of sample was injected onto the column and the following gradient conditions were applied: 0% B was held for 5 min, increased to 18% mobile phase B over 1.5 min and held there for 2.5 min; increased to 65% B over 0.5 min and held there for 3.5 min; increased to 98% B over 5 min and held there for 3.5 min; and decreased to 0% B over 0.5 min. The column was re-equilibrated for 8 min prior to the next injection.

Reversed-phase chromatography was performed using a 5 cm × 1 mm i. d. × 5 μm Discovery C18 column (Supelco, Bellefonte, PA), operating at a flow rate of 100 μL/min and maintained at 37 °C. Mobile phase A consisted of 5 mM HCOONH₄ in 74/25/1 (v/v/v) H₂O:CH₃OH:HCOOH, while mobile phase B consisted of 5 mM HCOONH₄ in 99:1 (v/v) CH₃OH:HCOOH. Lor each LC–MS/MS experiment, ca. 10 μL of sample was injected onto the column and was fractionated using the following gradient conditions: 10% B was held for 4 min; increased to 95% B over 3 min and held there for 4 min; increased to 99% B buffer over 6 min and held there for 6 min; decreased to 10% B over 0.5 min. The column was re-equilibrated for 6.5 min prior to the next injection. Both normal-phase and reversed-phase chromatography were performed in main-pass flow in a normal mode on a 40-μL injection volume and without any post-time column flow.

The ESI source conditions were optimized at a LC flow rate of 100 μL/min, using a mixture of LM 6002, brain porcine, and other Cer and SMs, including Cer d18:1^Δ4 c12:0, Cer d18:1^Δ4 c18:1, Cer d18:1^Δ4 c18:0, Cer d18:1^Δ4 c20:0, Cer d18:1^Δ4 c24:1, Cer d18:1^Δ4 c24:0, SM d18:1^Δ4 c12:0, SM d18:1^Δ4 c16:0, and SM d18:1^Δ4 c18:0. The optimized source parameters were as follows: ionization mode, positive; sheath gas pressure, 7 psi; auxiliary gas pressure, 3 (arbitrary units); electrospray needle voltage, 4,500 V; capillary temperature, 300 °C; skimmer offset, −5 V. No in-source fragmentation was observed when the reference compounds were analyzed using direct infusion with skimmer offset values of −7 and less. However, at higher skimmer offset values of 15, ceramides seemed to undergo some water loss to form [M+H−H₂O]⁺ ions.

Collision energies between 23 and 31V provided optimal fragmentation of [M+H]⁺ ions to product ions at m/z 264.2 and 266.2, corresponding to doubly dehydrated product ion fragments for unsaturated and saturated d18:1^Δ4 or d18:0 sphingoid base backbones, respectively, from Cer and MHCer, and m/z 184.1, corresponding to the choline phosphate moiety of sphingomyelin. Collision-induced dissociation was performed using nitrogen gas within Q2, which was offset from Q1 by 10 V. Prior to running SRM experiments, precursor ion scans were performed for m/z 184, 264, and 266 to detect choline containing analytes or sphingolipid constituents with d18:1 or d18:0 sphingolipid base backbone. For experiments conducted using normal-phase chromatography, a total of three single-reaction monitoring (SRM) methods were constructed, each partitioned into 5 segments. Between six and eight analyte transitions from a specific sphingolipid class, along with the corresponding internal standard, were included in each segment to maximize the number of data points and attain maximum sensitivity. The first (3 min) and fifth segments (6 min) of each method contain fictional analyte transitions. In the second segment (3 min), transitions targeting Cer were included, while the third (10.5 min) and fourth segments (7.5 min) contain transitions to measure MHCers and SMs, respectively. Because no analytes were monitored in the fifth segment, a dummy transition was incorporated in this segment. For reversed-phase chromatography experiments, three individual SRM methods comprised of three segments were constructed for each class of compounds; therefore, for three classes of compounds Cers, MHCers, and SMs, nine individual methods were constructed as analytes from each sub-class, spanning a much wider zone. The first and the last segments, spanning 6 and 3 min, respectively, contain the fictional transition, as no analytes were eluted during these time spans, while the second segment (21 min) contains 6–7 analytes and their internal standard transitions.

The acquisition parameters common to all analytes were: scan width (m/z) 0.10; scan time, 0.20 s for each transition; peak width (FWHM) 0.70 for both Q1 and Q3; collision pressure 1.5 mTorr; and skimmer offset at −5 V. Other acquisition parameters and chromatographic retention times are listed in Tables 1, 2, and 3. Data acquisition and analysis were accomplished using Xcalibur software v.2.0.5 (Thermo Electron Corp.). Quantitation data was not corrected for the double bond differences since the degree of unsaturation is limited, due to the existence of up to only two double bonds. However data was corrected for the carbon number difference between a given molecular species and the selected internal standard (i.e., to include the difference in degrees of freedom due to varying carbon chain lengths, which results in different ionization efficiencies) according to the following formula: z₁ = (1 + 0.011n + 0.011² n (n − 1)/2)/(1 + 0.011s + 0.011² s(s − 1)/2) [42].

Table 1.

Retention times and compound-specific SRM parameters for ceramides (Cer), hexosylceramides (MHCer), and sphingomyelins (SM) transitions that are included in the 2nd, 3rd, and 4th segments of the five-segmented normal-phase HPLC–ESI–MS/MS method 1

Analyte	Segment Duration (min)	Precursor [M+H]+ → product ion (m/z)	Retention time		Collision energy (V)	Tube lens (V)
			NP (min)	RP (min)
Segment 1	3.00 (To waste)	999.00 → 950.00 (dummy transition)
Cer in Segment 2	3.00
Cer d18:1Δ4 c12:0 (ISTD)		482.43 → 264.14	4.45	13.75	23	64
Cer d18:1Δ4 c16:0		538.62 → 264.16	4.36	16.64	24	65
Cer d18:0 c16:0		540.67 → 266.16	4.34	17.22	25	67
Cer d18:1Δ4 c18:1		564.52 → 264.16	4.32	13.23	24	69
Cer d18:1Δ4 c18:0		566.52 → 264.17	4.30	18.30	26	74
Cer d18:0 c18:0		568.63 → 266.16	NO	18.81	26	73
MHCer in Segment 3	10.50
MHCer d18:1Δ4 c12:0 (ISTD)		644.51 → 264.08	11.36	13.18	29	74
MHCer d18:1Δ4 c16:0		700.54 → 264.11	9.96	16.46	28	73
MHCer d18:0 c16:0		702.50 → 266.11	NO	17.08	28	73
MHCer d18:1Δ4 c18:0		728.51 → 264.09	9.38	18.15	29	72
MHCer d18:0 c18:0		730.51 → 266.11	NO	NO	29	71
MHCer d18:1Δ4 c20:0		756.52 → 264.10	8.94	19.81	29	70
MHCer d18:0 c20		758.53 → 266.12	NO	NO	29	73
SM in Segment 4	7.50
SM d18:1Δ4 c12:0 (ISTD)		647.49 → 184.01	17.76	13.51	23	79
SM d18:1Δ4 c16:0		703.58 → 184.01	17.59	16.41	27	57
SM d18:0 c16:0		705.56 → 183.97	17.26	17.09	37	71
SM d18:1Δ4 c18:0		731.58 → 183.95	17.52	18.07	31	73
SM d18:0 c18:0		733.59 → 183.95	17.19	18.81	26	53
SM d18:1Δ4 c20:0		759.61 → 183.91	17.44	19.79	27	72
SM d18:0 c20:0		761.62 → 183.98	NO	20.50	28	71
Segment 5	6.00	999.00 → 950.00 (dummy transition)

Table 2.

Retention times and compound-specific SRM parameters for ceramides (Cer), hexosylceramides (MHCer), and sphingomyelins (SM) transitions that are included in the 2nd, 3rd, and 4th segments of the five-segmented normal-phase HPLC–ESI–MS/MS method 2

Analyte	Segment Duration (min)	Precursor [M+H]+ → product ion (m/z)	Retention time		Collision energy (V)	Tube lens (V)
			NP (min)	RP (min)
Segment 1	3.00 (To waste)	999.00 → 950.00 (dummy transition)
Cer in Segment 2	3.00
Cer d18:1Δ4 c12:0 (ISTD)		482.43 → 264.14	4.45	13.76	23	64
Cer d18:1Δ4 c20:0		594.55 → 264.16	4.25	20.02	24	74
Cer d18:0 c20:0		596.64 → 266.19	NI	20.67	25	75
Cer d18:1Δ4 c22:0		622.71 → 264.17	4.21	21.88	26	73
Cer d18:0 c22:0		624.61 → 266.18	NI	22.63	27	74
Cer d18:1Δ4 c24:1		648.59 → 264.19	4.19	21.92	28	71
MHCer in Segment 3	10.50
MHCer d18:1Δ4 c12:0 (ISTD)		644.51 → 264.08	11.36	13.48	29	74
MHCer d18:1Δ4 c22:0		784.53 → 264.12	8.49	21.07	29	68
MHCer d18:0 c22:0		786.59 → 266.11	NO	21.85	29	70
MHCer d18:1Δ4 c24:1		810.59 → 264.12	8.45	21.11	29	73
MHCer d18:0 c24:1		812.64 → 266.11	8.12	NO	29	75
MHCer d18:1Δ4 c24:0		812.65 → 264.11	8.35	23.05	29	72
SM in Segment 4	7.50
SM d18:1Δ4 c12:0 (ISTD)		647.49 → 184.01	17.76	13.38	23	79
SM d18:1Δ4 c22:0		787.64 → 184.01	17.38	21.43	26	72
SM d18:0 c22:0		789.65 → 183.97	16.85	22.34	27	71
SM d18:1Δ4 c23:0*		801.65 → 183.92	17.33	22.48	27	73
SM d18:1Δ4 c24:1		813.66 → 183.95	17.34	21.48	28	75
SM d18:1Δ4 c24:0 or SM d18:0 c24:1		815.68 → 183.98	17.29	23.58	28	74
SM d18:0 c24:0		817.68 → 184.07	16.28	24.76	29	73
Segment 5	6.00	999.00 → 950.00 (dummy transition)

Table 3.

Retention times and compound-specific SRM parameters for ceramides (Cer), hexosylceramides (MHCer), and sphingomyelins (SM) transitions that are included in the 2nd, 3rd, and 4th segments of the five-segmented normal-phase HPLC–ESI–MS/MS method 3

Analyte	Segment duration (min)	Precursor [M+H]+ → product ion (m/z)	Retention time		Collision energy (V)	Tube lens (V)
			NP (min)	RP (min)
Segment 1	3.00 (To waste)	999.00 → 950.00 (dummy transition)
Cer in Segment 2	3.00
Cer d18:1Δ4 c12:0 (ISTD)		482.43 → 264.14	4.45	13.80	23	64
Cer d18:1Δ4 c24:0		650.61 → 264.15	4.16	24.15	28	79
Cer d18:0 c24:1		650.61 → 266.17	NI	22.01	28	73
Cer d18:0 c24:0		652.81 → 266.18	4.17	25.19	28	74
Cer d18:1Δ4 c26:1		676.83 → 264.18	4.15	24.19	27	75
Cer d18:0 c26:1		678.82 → 266.19	NI	NI	29	75
Cer d18:0 c26:0		680.78 → 266.18	NI	NI	29	78
MHCer in Segment 3	10.50
MHCer d18:1Δ4 c12:0 (ISTD)		644.51 → 264.08	11.36	13.25	29	74
MHCer d18:0 c24:0		814.63 → 266.11	8.12	22.96	29	70
MHCer d18:1Δ4 c26:1		838.63 → 264.12	8.15	22.94	30	73
MHCer d18:0 c26:1		840.64 → 266.11	7.79	NI	30	70
MHCer d18:1Δ4 c26:0		840.63 → 264.12	NI	25.48	30	69
MHCer d18:0 c26:0		842.63 → 266.11	NI	NI	31	72
SM in Segment 4	7.50
SM d18:1Δ4 c12:0 (ISTD)		647.49 → 184.00	17.76	13.46	23	79
SM d18:1Δ4 c25:1*		827.68 → 184.07	17.37	22.50	29	74
SM d18:0 c25:0		829.68 → 183.94	17.34	24.77	29	75
SM d18:1Δ4 c26:1		841.69 → 183.93	17.28	23.55	28	77
SM d18:1Δ4 c26:0 or SM d18:0 c26:1		843.69 → 183.97	17.25	26.18	28	76
SM d18:0 c26:0		845.70 → 184.12	17.35		29	75
Segment 5	6.00	999.00 → 950.00 (dummy transition)

NP Normal phase; RP reversed phase; NO not observed; NI not identified; V Volts; min minutes; m/z mass to charge ratio

The internal standard LM 6002 was added to each sample at the beginning of sample preparation, and the ratio of peak areas of an analyte versus internal standard was calculated. An advantage of this quantitative technique is that ionization suppression effects of the matrix, and the variation in ionization efficiency between standards and analytes are compensated.

Results and Discussion

Normal-Phase Liquid Chromatography–Tandem Mass Spectrometry

Ceramides, SMs, and MHCers can be resolved and quantitated using normal-phase liquid chromatography, since the separation is based on head group characteristics rather than on the presence of different lengths of fatty-N-acyl side chains [46, 47]. Therefore, the components of an individual sphingolipid class elute together when separated by normal-phase liquid chromatography, and a single peak, primarily due to lipophilic (hydrophobic) interaction, is displayed. Thus, all Cers elute with similar retention times, as do MHCers and SMs, etc., which are displayed as single peaks but at different retention times. Accordingly, the quantitative analysis of the head group classes by normal-phase chromatography is not capable of distinguishing the isotopic ¹³C contributions from [M+H+1]⁺ and [M+H+2]⁺ and other isomeric species, since it does not provide a distinction between the compounds containing an unsaturated (sphingosine) and/or saturated (sphinganine) base structures, as all the components are displayed at a single retention time. The MS/MS mass filters that are used to identify these analytes cannot differentiate between co-eluting compounds with overlapping m/z values, unless the transition ions used for SRM are widely different for each analyte that was monitored. Therefore, normal-phase chromatography is typically employed to quantitate total Cers or SMs inclusive of all low- and high-abundant species, rather than individual components of Cers and SMs. In normal-phase chromatography, although all lipid classes are separated, components of the same lipid class elute together; thus, there exists the possibility of inter-molecular interactions from individual components within the lipid class that may lead to some ion suppression. In comparative analysis, when performed, for example, between wild and mutant species, such effects during analysis do cancel each other out.

In the present non-polar LC–MS/MS method presented here, many of the components from each class of compounds (e.g., MHCer d18:1^Δ4c16:0 [retention time, 9.96 min, Fig. 2, panel j], MHCer d18:1^Δ4c18:0 [rt: 9.41 min, Fig. 2, panel l], MHCer d18:1^Δ4c20:0 [rt: 8.91 min, Fig. 2, panel n] or between cer d18:1^Δ4c16:0 [rt: 4.33 min, Fig. 2, panel c] and cer d18:1^Δ4c18:0 [rt: 4.30, Fig. 2, panel f]) were well resolved. This method is also capable of differentiating between saturated (sphinganine-containing) and unsaturated (sphingosine-containing) isomorphous components that differ only by two hydrogen atoms (e.g., SM d18:1^Δ4c18:0 [rt: 17.52 min, Fig. 2, panel t], and SM d18:0 c18:0 [rt: 17.16 min, Fig. 2, panel u]). The present method exhibits picogram-level sensitivity for each lipid class.

Fig. 2.

Extracted ion chromatogram from SRM scans of precursor → double-dehydrated carbocation product fragment for the ceramides (transitions from molecular ions of 482–568), hexosylceramides (transitions from molecular ions of 644–758), and sphingomyelins (transitions from molecular ions of 644–758), including the respective internal standards by method 1 of the normal-phase liquid chromatography–tandem mass spectrometry

Three SRM-MS assays that rely on normal-phase separation to analyze Cers, SMs, and MHCers from mammalian cells, as described in the Experimental Section, were developed. Using these methods, as shown in Figs. 2, 3, and 4 and Tables 1, 2, and 3, Cers appear at retention times between 4.1 and 4.5 min, while MHCers and SMs elute over wider retention time ranges, (7.7–11.5 and 16.2–17.8 min, respectively). Compounds containing the sphingoid base backbone that differ by two hydrogen atoms, but contain equivalent numbers of all other atoms, are fully resolved for MHCers and reasonably well to partially resolved for SMs by using this method. A few trends are clearly discernable within the data. As the number of methylene groups within the Cers, MHCers, and SMs increase, the retention times decrease. For example, Cer d18:1^Δ4 c16:0 appears at a retention time of 4.33 min (Fig. 2, panel C), while Cer d18:1^Δ4 c18:0 appears at a retention time of 4.30 min (Fig. 2, panel f). Compounds with unsaturated sphingoid base backbones and their saturated counterparts containing the same number of non-hydrogen atoms but differing by only two hydrogen atoms, always appear as pairs, with the saturated compound eluting first. For example, SM d18:1^Δ4 c22:0 (Fig. 3, panel Q) and SM d18:0 c22:0 (Fig. 3, panel r) appears at a retention time of 17.38 and 16.85 min, respectively. For compounds with an equivalent number of non-hydrogen atoms, the saturated compound elutes fastest, followed by the one in which the double bond is present on the sphingoid base backbone. When double bonds are present both on the sphingoid base and the fatty-N-acyl moiety, then such a compound elutes last. This is illustrated in the example for the following series of compounds with 18 carbon atoms on the sphingoid chain and 24 carbon atoms in the N-acyl fatty acid side chain, where the order of elution is: d18:0 c24:0 < d18:1^Δ4 c24:0 < d18:0 c24:1 < d18:1^Δ4 c24:1. Finally, the retention times shown for compounds possessing an unsaturation in both their sphingoid base backbone and side chain are similar to the compound that contains the same sphingoid base backbone but with two fewer methylene groups and does not possess any unsaturation in its N-acyl side chain. For example, the retention time exhibited for MHCer d18:1^Δ4 c24:1 (Rt: 8.45 min, Fig. 3, panel 1) is similar to that of MHCer d18:1^Δ4 c22:0 (Rt: 8.49 min. Fig. 3, panel j), or SM d18:1^Δ4 c24:1 (Rt: 17.34 min. Fig. 3, panel t) is similar to that of SM d18:1^Δ4 c22:0 (Rt: 17.38 min. Fig. 3, panel q).

Fig. 3.

Extracted ion chromatogram from SRM scans of precursor → double-dehydrated carbocation product fragment for the ceramides (transitions from molecular ions of 594–648), hexosylceramides (transitions from molecular weight of 784–812), and sphingomyelins (transitions from molecular ions of 787–817), including the respective internal standards by method 2 of the normal-phase liquid chromatography–tandem mass spectrometry

Fig. 4.

Extracted ion chromatogram from SRM scans of precursor → double-dehydrated carbocation product fragment for the ceramides (transitions from molecular ions of 650–679), hexosylceramides (transitions from molecular weight of 814–840), and sphingomyelins (transitions from molecular ions of 827–845), including the respective internal standards by method 3 of the normal-phase liquid chromatography–tandem mass spectrometry

Sphingolipid analysis of RAW264.7 cells using normal-phase LC–MS/MS has been reported previously by Shaner et al. [35]; however, individual analytes from each sphingolipid class were not chromatographically resolved. In their study, ceramides and hexosylceramides eluted as single peaks, as did SMs and lactosylceramides. If the components from each class of analytes are not resolved, there will be considerable overlap of not only the compounds with the same nominal masses such as isomeric and isobaric species, but also, due to the [M+H+1]⁺ and [M+H+2]⁺ carbon ¹³C isotopic contributions from 2 Da lower sphingoid core unsaturated compounds to a 2 Da higher sphingoid base saturated molecular species. This problem is compounded when the unsaturated analogue is more abundant than its corresponding saturated sphingoid base analogue. Theoretically, the isotopic contribution from [M+H+1]⁺ and [M+H+2]⁺ can be calculated based on the number and kind of atoms present or the intensity of the peak observed, and for the [M+H+2]⁺, should be in the theoretical range ca. 6–10%, depending upon the number of atoms present; in addition, the double ¹³C isotope must be contained in both the precursor and product ion. In the present experiments, the isotopic contributions to the saturated sphingoid base dihydro species due to [M+H+2]⁺ from the corresponding sphingoid base dehydro species (e.g., un-saturated d-18 sphingoid base with product ion of m/z 264.2), are found to be 6–10% of the amount of the unsaturated species, even if for example m/z 266.2 of the saturated d-18 sphingoid base product ion transition is monitored by SRM. Thus, the dehydro species almost always seems to exhibit a lingering (residual) small d-saturated product ion transition coming from the sphingoid base skeleton, similar to that arising from the dihydro species, or could as well be due to some isobaric contribution. Hence, unless the d-unsaturated and d-saturated analogues are chromatographically resolved, or if the retention times for both species are the same, the existence of the dihydro-saturated species can only be considered if the peak area of the saturated dihydro compound is more than 5–10% of the corresponding dehydro-unsaturated species.

The LC–MS/MS data for sphingolipids obtained from the analysis of HEK293 and MEF cells are presented in Tables 4 and 5, respectively. Table 4 presents the data obtained using normal-phase chromatography to analyze the HEK293E cells, while Table 5 provides comparative data obtained by both normal-phase and reversed-phase chromatography to analyze MEF cells. The quantitative data indicates that the ceramide levels in HEK293E cell present are ca. 10 nmol, while the levels for MHCers and SMs are <50 nmol each. Therefore, the fatty acid contribution from these sphingolipid classes is <200 nmol in this cell analysis.

Table 4.

Normal-phase liquid chromatography selected reaction monitoring data obtained for the evaluation of sphingomyelins (SM), ceramides (Cer), and hexosylceramides (MHCer) from HEK293 cells

Analyte	SM			Cer			MHCer
	Mean ± SD (pmol/1 × 107 cells)	CV (%)	Analyte (%)	Mean ± SD (pmol/1 × 107 cells)	CV (%)	Analyte (%)	Mean ± SD (pmol/1 × 107 cells)	CV (%)	Analyte (%)
d18:1Δ4 c16:0	14,467.8 ± 372.1	2.7	44.7	2,220.8 ± 109.1	4.9	24.6	4,464.1 ± 102.8	2.3	9.2
d18:0 c16:0	594.6 ± 17.5	3.1	1.8	NO	–	–	15.2 ± 0.8	5.2	0.0
d18:1Δ4 c18:0	2,885.1 ± 102.4	3.8	8.9	324.3 ± 16.9	5.2	3.6	772.1 ± 20.9	2.7	1.6
d18:1Δ4 c18:1	NO	–	–	20.5 ± 1.1	5.6	0.1	NO	–	–
d18:0 c18:0	185.5 ± 5.0	2.9	0.6	NO	–	–	6.6 ± 0.1	1.6	0.0
d18:1Δ4 c20:0	505.9 ± 17.1	3.7	1.6	131.7 ± 10.1	7.7	1.5	815.2 ± 32	3.9	1.7
d18:0 c20:0	115.4 ± 3.4	3.2	0.4	NO	–	–	NO	–	–
d18:1Δ4 c22:0	2,805.9 ± 152.7	6.0	8.7	1,053.5 ± 30.9	2.9	11.7	10,011.5 ± 283.9	2.8	20.6
d18:0 c22:0	187.7 ± 10.2	6.1	0.6	NO	–	–	28.6 ± 0.9	3.1	0.1
d18:1Δ4 c23:0	50.2 ± 3.0	6.6	0.2	NO	–	–	NO	–	–
d18:1Δ4 c24:1	6,445.2 ± 340.7	5.9	19.9	2,279.2 ± 42.0	1.8	25.2	6,814.7 ± 275.5	4.0	14.0
d18:1 c24:0	5,656.5 ± 257.1	5.1	17.5	2,868.4 ± 113.9	3.2	31.8	24,291.2 ± 1,122.6	4.6	49.9
d18:0 c24:0	597.4 ± 39.0	7.4	1.8	73.0 ± 5.0	6.8	0.8	NO	–	–
d18:1Δ4 c26:1	241.1 ± 23.9	11.5	0.7	51.6 ± 4.2	8.1	0.6	322.8 ± 30.4	9.4	0.7
d18:1Δ4 c26:0	216.8 ± 19.3	10.3	0.7	4.8 ± 0.6	12.1	0.1	1,116.5 ± 97.3	8.7	2.3
d18:0 c26:1	26.9 ± 1.9	8.2	0.1	NO	–	–	37.8 ± 3.2	8.4	0.1
Total in the Cells	32,389.6 ± 1,193.3	3.7		9,027.7 ± 286.9	3.1		48,696.2 ± 1,446.3	3.0

CV is based on triplicate run of samples. NO not observed; NI not identified

Table 5.

Comparative normal-phase and reversed-phase liquid chromatography-tandem mass spectrometric (SRM) data obtained for the evaluation of sphingomyelins (SM), ceramides (Cer), and hexosylceramides (MHCer) from mouse embryonic cells at the 1-mg protein level

Analyte	Normal Phase LC/MS/MS			Reversed Phase LC/MS/MS
	Mean ± SD (1 mg Protein Level)	CV (%)	Analyte (%)	Mean ± SD (1 mg Protein Level)	CV (%)	Analyte (%)
SM
SM d18:1Δ4 c16:0	5,732.0 ± 402.9	7.0	55.1	6,804.4 ± 245.4	3.6	50.6
SM d18:0 c16:0	208.4 ± 28.9	13.7	2.0	357.7 ± 6.7	1.9	2.7
SM d18:1Δ4 c18:0	534.9 ± 59.8	11.2	5.1	682.4 ± 25.1	3.7	5.1
SM d18:0 c18:0	7.0 ± 0.9	12.4	0.1	14.6 ± 1.4	9.4	0.1
SM d18:1Δ4 c20:0	198.2 ± 13.8	7.0	1.9	230.9 ± 15.3	6.6	1.7
SM d18:0 c20:0	NI	–		5.8 ± 0.5	8.4	0.1
SM d18:1Δ4 c22:0	442.0 ± 37.2	8.4	4.2	618.2 ± 25.9	4.2	4.6
SM d18:0 c22:0	NI	–		12.6 ± 0.8	6.4	0.1
SM d18:1Δ4 c23:0	81.8 ± 4.1	5.0	0.8	111.5 ± 12.1	10.8	0.8
SM d18:1Δ4 c24:1	2,166.9 ± 212.3	9.8	20.8	3,404.2 ± 156.1	4.6	25.3
SM d18:1Δ4 c24:0	1,012.9 ± 102.5	10.1	9.7	1,157.2 ± 68.8	5.9	8.6
SM d18:1Δ4 c26:0	19.2 ± 3.6	14.9	0.2	56.2 ± 1.8	3.2	0.4
Total in pmol	10,403.3 ± 705	7.0		13,456.6 ± 402.1	3.0
Cer
d18:1Δ4 c16:0	62.8 ± 0.9	1.5	25.3	182.7 ± 1.6	0.9	48.3
d18:0 c16:0	1.6 ± 0.1	8	0.6	NI	–	–
d18:1Δ4 c18:1	1.4 ± 0.1	6.5	0.6	13.5 ± 0.8	5.7	3.6
d18:1Δ4 c18:0	13.5 ± 0.1	0.5	5.4	20.4 ± 1.2	5.8	5.4
d18:1Δ4 c20:0	5.8 ± 0.11	0.3	2.3	8.5 ± 0.8	9.6	2.2
d18:1Δ4 c22:0	26.7 ± 0.7	2.5	10.7	20.5 ± 1.3	6.5	5.4
d18:1Δ4 c24:1	66.2 ± 0.3	0.5	26.6	68.7 ± 4.9	6.4	18.2
d18:1Δ4 c24:0	61.8 ± 0.9	1.4	25.3	62.1 ± 4.8	5.1	16.4
d18:1Δ4 c26:1	7.8 ± 0.2	2.4	3.1	1.7 ± 0.1	6.2	0.4
Total in pmol	248.6 ± 1.4	1.6		378.1 ± 21.8	4.8
MHCer
d18:1Δ4 c16:0	257.4 ± 39.9	15.5	30.3	481.2 ± 5.0	1.0	52.7
d18:0 c16:0	NI	–		NI
d18:1Δ4 c18:0	39.7 ± 5.1	12.9	4.7	20.5 ± 0.4	1.7	2.2
d18:0 c18:0	NI	–		NI
d18:1Δ4 c20:0	28.0 ± 3.5	12.5	3.3	16.4 ± 2.1	12.8	1.8
d18:1Δ4 c22:0	92.3 ± 8.6	9.3	10.9	66.6 ± 7.9	11.9	7.3
d18:0 c22:0	NI	–	–	NI		–
d18:1Δ4 c24:1	165.4 ± 19.1	11.5	19.5	127.2 ± 6.4	5.0	13.9
d18:0 c24:1	3.9 ± 0.5	12.1	0.5	NO		–
d18:1Δ4 c24:0	256.7 ± 31.3	12.2	30.2	197.2 ± 11.9	6.0	21.6
d18:1Δ4 c26:1	5.8 ± 0.7	11.5	0.7	3.8 ± 0.2	6.6	0.4
Total in pmol	849.3 ± 109.2	12.8		912.7 ± 24.4	2.7

Comparative Sphingolipid Measurements

Normal-phase LC–MS/MS methods were developed to quantify and compare the levels of SMs, Cers, and M in HEK293 and MEF cells (Tables 4, 5). In both cell types, d18:1^Δ4 c16:0 represents ~45–55% of the total amount of SM, followed by d18:1^Δ4 c24:1 (~20%) and d18:1^Δ4 c24:0 (~17.5 and 10% in HEK293 and MEF cells, respectively). The major Cer present in HEK293 cells is d18:1^Δ4 c24:0 (~32% of total), followed by d18:1^Δ4 c16:0 (25%), d18:1^Δ4 c24:1 (25%), and d18:1^Δ4 c22:0 (17%). In MEF cells, a similar pattern was observed with d18:1^Δ4 c16:0, d18:1^Δ4 c24:1, d18:1^Δ4 c24:0, and d18:1^Δ4 c22:0 constituting about 80–90% of the total Cer content.

The measurement of monohexosylceramide, i.e., hexosylceramide (MHCer), levels in HEK293 cells using normal-phase chromatography, shows d18:1 c24:0 as the dominant MHCer (~50% of total), followed by d18:1^Δ4 c22:0 (~21%), d18:1^Δ4 c24:1 (~14%), and d18:1^Δ4 c16:0 (~10%), respectively. In MEFs, d18:1^Δ4 c16:0 (30%) and 18:1^Δ4 c24:0 (30%) are the dominant MHCers, followed by d18:1^Δ4 c24:1 (20%) and d18:1^Δ4 c22:0 (11%). Overall, the Cer, MHCer, and SM profiles of the two cell types are similar, with the predominant species being the d18:1^Δ4 c16:0, d18:1^Δ4 c24:0, d18:1^Δ4 c24:1, and d18:1^Δ4 c22:0 molecules. Importantly, the sphingolipids mixture of internal standard containing C17 or C12 bases (LM6002) are unnatural synthetic lipids, and consequently, no endogenous signals of these categories were observed from the cell lines.

In the study by Shaner et al. [35], the most abundant SM in RAW264.7 mouse macrophage cells was d18:1^Δ4 c16:0, followed by d18:0 c16:0 and d18:1^Δ4 c24:1. These results are similar to what we observed for HEK293 and MEF cells, except that we found much lower levels of d18:0 c16:0. The levels of Cer d18:1^Δ4 c24:0, d18:1^Δ4 c16:0, and d18:1^Δ4 c24:1 found using our method, are also similar to those reported by Merrill et al. While we and Merrill et al. found that MHCer d18:1^Δ4 c16:0 was the most abundant MHCer in MEF cells, this lipid was the third most abundant species in HEK293 cells. Previously Koivusalo et al. [47] found only three SMs in the BHK21 hamster cells, SM, d18:1^Δ4 c16:0 (65% of total sphingolipid content), SM d18:1^Δ4 c24:1 (20%), and SM d18:1^Δ4 c24:0 (15%), similar to the top three SMs observed in HEK293 and MEF cells. Other SMs in their studies might not have been observed, perhaps due to the lack of the sensitivity of their assay. Hammad et al. [48] performed the sphingolipidomic analysis of the serum and plasma samples of both healthy human males and females, and these authors reported d18:1^Δ4 c16:0, and d18:1^Δ4 c24:1 as the two most abundant sphingomyelins and monohexosylceramides, and d18:1^Δ4 c24:0, d18:1^Δ4 c24:1 as the two most abundant ceramide species. Thus, from all these quantitative analyses of different cells and tissues, we can say with some confidence that d18:1^Δ4 c16:0 and d18:1^Δ4 c24:1 might be the top sphingomyelin candidates, and Cer d18:1^Δ4 c24:0, d18:1^Δ4 c16:0, and d18:1^Δ4 c24:1 might be the top three ceramide species present in many types of mammalian cells; and therefore, comparison of these dominant top sphingolipids in healthy and disease cells might provide some insights into the prognosis of health and disease. It should be noted that nerve tissues that are not studied here might have some other Cers and SMs, such as those that contain hydroxyl fatty acids in their chains as the dominating species.

Impact of the Sample Protein Amount on Sphingolipids Quantitation

When different batches of MEF cell samples containing different amounts of protein were collected and processed for NP-LC/MS/MS, the quantitative data was irreproducible, even after normalization based on protein concentration. Initially, it was thought that this discrepancy in data might be due to a subtle variation in the cell preparation and collection technique, and consequently in the sample handling; hence, the samples might have varying matrix type of effects. However, this discrepancy might also be due to the different protein levels that were present in the samples when different amount of samples were taken; and if this is true, then at what range of protein content in the sample will the sample give consistent quantitative estimation of analytes. To test these hypotheses, aliquots of MEF cells containing different amount of proteins (i.e., 400, 250, 100, 50, and 25 μg of total protein levels) were processed for LC–MS analysis. The sphingolipid data for SMs and Cers along with their percentages are presented in Table 6. These samples were prepared identically and then processed at the same time. It is very clear from Table 6 that as the absolute amount of protein in the samples decreases, the total amount of Cer measured increases. At moderately high protein content (400 μg), the Cer levels measured were only half as great compared to those found in the sample containing the lowest amount of protein (25 μg). However, it is apparent that the overall percentages of the individual Cer components remain similar. Ceramide levels in the samples containing 50–250 μg of protein exhibited minimal differences. To measure Cer in cells, the absolute range of protein levels should be between 50 and 300 μg; otherwise, the data will lose quantitative integrity in terms of absolute amounts.

Table 6.

Comparative liquid chromatography selected reaction monitoring data obtained for the evaluation of sphingomyelin (SM), and ceramides (CER) from mouse embryonic fibroblast cells that contain different amount of absolute protein. Data is normalized to the 1-mg protein level

	Amount of Protein
	400 μg		250 μg		100 μg		50 μg		25 μg
	pmol	%	pmol	%	pmol	%	pmol	%	pmol	%
Ceramides
Cer d18:1Δ4 c16:0	108.0	51.6	134.8	48.6	137.3	45.3	145.0	43.8	174.0	41.5
Cer d18:1Δ4 c18:1	1.5	0.7	1.4	0.5	3.9	1.3	2.5	0.7	2.9	0.7
Cer d18:1Δ4 c18:0	10.2	4.9	12.2	4.4	14.1	4.7	17.1	5.2	32.6	7.8
Cer d18:1Δ4 c20:0	3.3	1.6	4.3	1.6	6.0	2.0	9.4	2.8	13.8	3.3
Cer d18:1Δ4 c22:0		4.2	12.0	4.3	14.3	4.7	15.7	4.8	15.7	3.7
Cer d18:1Δ4 c24:1	44.0	21.0	62.5	22.6	74.3	24.5	83.6	25.3	117.2	28.0
Cer d18:1Δ4 c24:0	32.2	15.4	48.8	17.6	51.4	17.0	54.8	16.6	63.1	15.0
Cer d18:1Δ4 c26:1	1.0	0.5	1.1	0.4	1.3	0.4	2.5	0.7	NO	–
Total	209.1		277.1		302.7		330.6		419.2
Sphingomyelins
SM d18:1Δ4 c16:0	6,618.0	59.3	6,084.2	55.2	7,123.0	59.9	7,421.4	64.3	8,677.7	64.2
SM d18:0 c16:0	174.8	1.6	199.9	1.8	194.1	1.6	155.9	1.4	206.7	1.5
SM d18:1Δ4 c18:0	448.3	4.0	569.0	5.2	540.3	4.5	536.0	4.6	622.6	4.6
SM d18:0 c18:0	9.4	0.1	9.0	0.1		0.1	6.1	0.1	6.4	0.0
SM d18:1Δ4 c20:0	141.1	1.3	184.7	1.7	172.8	1.5	160.2	1.4	182.0	1.3
SM d18:0 c20:0	45.9	0.4	37.6	0.3	NO	–	NO	–	NO	–
SM d18:1Δ4 c22:0	411.6	3.7	470.0	4.3	440.1	3.7	403.2	3.5	472.1	3.5
SM d18:1Δ4 c23:0	80.8	0.7	89.9	0.8	82.6	0.7	69.7	0.6	85.6	0.6
SM d18:1Δ4 c24:1	2,284.4	20.5	2,410.1	21.8	2,253.3	18.9	2,053.6	17.8	2,376.8	17.6
SM d18:1Δ4 c24:0	818.9	7.3	861.4	7.8	967.6	8.1	648.1	5.6	828.0	6.1
SM d18:0 c24:0	22.9	0.2	5.1	0.0	5.1	0.0	NO	–	NO	–
SM d18:1Δ4 c25:1	45.2	0.4	43.6	0.4	45.5	0.4	34.6	0.3	32.8	0.2
SM d18:1Δ4 c25:0	22.0	0.2	25.2	0.2	22.6	0.2	17.0	0.1	NO	–
SM d18:1Δ4 c26:1	41.9	0.4	40.7	0.4	35.8	0.3	29.6	0.3	32.9	0.2
Total amount	11,165.2		11,030.5		11,891.7		11,535.3		13,523.6

For SM, the data shows consistency between 50 and 400 μg of protein, whereas at 25 μg of protein, the total SM levels appear to be relatively high; however, some of the signals were of low intensity and showed poor peak shape. Therefore, if sphingolipids such as Cers, MHCers, and SMs are being measured in comparative samples, it is important to begin with samples containing similar amounts of total protein. The recommended protein level is between 50 and 300 μg.

It is known that analyzing lipids by direct infusion into the mass spectrometer at high concentrations (i.e., 10 pmol/μL or ca. 150 pmol/μL of total lipids), the relative response decreases strongly to increasing acyl chain length [47]. At lower concentrations (i.e., 5 or 60 pmol/μL of total lipids), however, the response is essentially linear with the chain length as lipid-lipid interactions and lipids-to-solution droplets crowding at the ESI inter-phase are minimized [42, 47]. It should, however, be noted that use of chromatography will essentially dilute this process because of the separation of analytes, and at a given duration of time not many analytes co-elute. This is particularly true for RP chromatography, where individual analytes from each class are well separated. Even with NP chromatography, all the sphingolipid classes viz. ceramides, hexosylceramides, and sphingomyelins are well separated into bands appearing at vastly different retention times, with many individual analytes from each class partially to fully resolved. Further, the volume of sample solution injected on to a column in this study is ca. 10–20 μL from a total sample volume of ca. 700–800 μL divided between the NP and RP chromatography; therefore, given the amount of protein assayed in these samples, the lipid contents in this injected volume might not be too high at all.

Sphingolipid Analysis Using Reversed-Phase Liquid Chromatography–Tandem Mass Spectrometry

We also developed reversed-phase LC–MS/MS methods to quantitate Cers, MHCers, and SMs. The SRM profiles are shown in Figs. 5, 6, and 7, respectively. Reversed-phase chromatography has higher selectivity and specificity for these compounds than normal phase chromatography, owing to its superior resolution of these molecules. The separation in reversed-phase chromatography is based on the hydrophobicity of the fatty-N-acyl side chains, rather than on head group separation of these molecules. For each class of analytes, three SRM methods were constructed, each employing between 5 and 7 analyte transitions, enabling the analysis of 18–20 molecules within each class. Excellent signals were observed for all SMs (Fig. 7); however, signals for MHCer were very poor for samples containing low and moderate levels of these compounds (Fig. 6). Some of the elution trends observed when separating the compounds using normal-phase chromatography were reversed in reversed-phase chromatography, e.g., as the number of methylene groups increased in each class of sphingolipids, their retention times increased, while the d-unsaturated and d-saturated still appeared as pair, the dihydro species appearing at a higher retention time, compared to the corresponding dehydro compound.

Fig. 5.

Extracted ion chromatogram from SRM scans of precursor → double-dehydrated carbocation product fragment for all the ceramides, as observed by three reversed-phase methods. All three reversed-phase methods contain only transitions from ceramides. Each method also comprises a transition from the ceramide internal standard

Fig. 6.

Extracted ion chromatogram from SRM scans of precursor → double-dehydrated carbocation product fragment for all the hexosylceramides (MHCer), as observed using three reversed-phase methods. All three reversed-phase methods contain only transitions from MHCer. Each method also comprises a transition from the MHCer internal standard

Fig. 7.

Extracted ion chromatogram from SRM scans of precursor → double-dehydrated carbocation product fragment for all the sphingomyelins (SM), as observed by three reversed-phase methods. All three reversed-phase methods contain only transitions from SM. Each method also comprises a transition from the SM internal standard

Obtaining the necessary sensitivity for some of the low-abundant species, particularly for hexosylceramides in reversed-phase chromatography, may require injecting larger amounts of material onto the column. This need can be a limiting factor when little biological sample is available. Furthermore, the interaction of Cers, MHCers, and SMs with the solvents used in reversed- and normal-phase chromatography differ. Owing to their choline head group, SMs have stronger dipole-dipole and H-bonding type of interactions with reversed-phase solvents, compared to weaker van der Waals, London-London and other dispersion forces with normal-phase solvents. Consequently, SMs resolve better under reversed-phase conditions. Although ESI normally does not ionize efficiently under the high content of non-polar organic solvents, such as hexane, CHCl₃, ethyl acetate or ether, but in the presence of 5 mM CH₃COONH₄ dissolved in 20 mL CH₃OH and containing 1–2% CH₃COOH, the ceramides and MHCers seem to exhibit good ionization efficiency and good sensitivity, in comparison to their ionization efficiency with reversed-phase solvents. It should be noted that the run times in both normal- and reversed-phase chromatography are almost identical (ca. 22 min); however, in normal-phase chromatography, each of the three classes of sphingolipids is displayed within a narrow elution time window of its own: Cers elute between 3.7 and 4.5 min, MHCers between 8 and 12 min, and SMs between 16 and 19 min, thus making it possible to combine analytes from each sub-class together into different segments to run them concurrently in each method. The individual MRM pair that uniquely identifies each species can then be summed to yield the total class species. In reversed-phase chromatography compounds from each class of sphingolipids are well resolved, with retention time ranging widely between 12 and 27 min. Further, because of a large elution window between the internal standard that was used and the respective analytes of that class, each sub-class of sphingolipids was run separately, instead of juxtaposing the peaks over each other chromatographically. Because of this larger elution window, as, for example, there is more than 10 min difference between the elution time of Cer d18:1^Δ4 c12:0 internal standard and the analyte Cer d18:1^Δ4 c24:1, there remains a strong possibility for ionization differences between earlier and later eluting analytes, in comparison to the use of only one either earlier or later eluting internal standard, which might result in a disparity of data between the normal- and reversed-phase chromatography. However, data from Table 5 shows the calculated amounts of individual Cers, MHCers, and SMs obtained by both reversed- and normal-phase chromatography, indicating the overall bias is less for SMs. For Cers and MHCers, the earlier eluting analytes exhibit higher amounts in comparison to the use of only one earlier eluting internal standard for each. Thus, it seems prudent to use more than two internal standards that span the entire elution window for each class of analytes when reversed-phase chromatography is performed.

As the NP and RP-LC/MS/MS quantitation results were inconsistent between the two, a solution containing four reference standards, two each of ceramides and sphingomyelins, viz. Cer d18:1^Δ4 c20:0, Cer d18:1^Δ4 c24:1, SM d18:1^Δ4 c16:0, and SM d18:1^Δ4 c18:0, were prepared at the 1-μg/mL level and then spiked in equal amounts and studied by both NP and RP-LC/MS/MS. The peak areas obtained by NP chromatography for these reference standards were 2.85 × 10⁷, 3.19 × 10⁸, 4.64 × 10⁸, and 2.41 × 10⁸, while under reversed-phase conditions, the respective peak areas were 9.63 × 10⁷, 1.17 × 10⁷, 2.01 × 10⁹, 1.04 × 10⁹; these results also suggest different ionization efficiencies in NP and RP chromatograph. Thus, data obtained with these spiked reference standards also suggested the similar discriminatory results between the NP and RP chromatography. Recovery between 70–80% and 90–95% were observed for samples processed by normal-phase and reversed-phase, respectively, using the comparative calculations of the peak areas of individual components from neat and LM6002 prespiked sample.

A series of reliable, efficient, and highly sensitive LC–MS (SRM) methods using normal-phase and reversed-phase liquid chromatography was developed to quantitate mammalian Cers, MHCers, and SMs containing saturated and unsaturated sphingoid base cores. In the context of normal phase chromatography a high content of organic solvents still provided efficient electrospray ionization of the ceramides and hexosylceramides. Using non-polar solvents in normal phase chromatography, even low abundant Cers having similar retention times were implicitly resolved, while reversed-phase chromatography was used to resolve the Cer compounds explicitly. SMs could be resolved using either type of chromatography, however, reversed-phase chromatography provided better specificity, selectivity, and sensitivity. Hexosylceramides were best resolved and quantitated using normal phase chromatography. Because the interaction of Cers, MHCers, and SMs differs between normal-phase and reversed-phase solvents and there is a possibility of some ion suppression effects, it is essential to link the quantitative results to the type of chromatography utilized. An absolute total amount of protein amounts between 100 and 300 μg is essential to get consistent quantitative results for all the sphingolipids. Overall, the quantitative results show that for MEF and HEK293 cells, more than 75% of the Cers, MHCers, and SMs exist as d18:1^Δ4 c16:0, d18:1^Δ4 c24:1, and d18:1^Δ4 c24:0.

Abbreviations

Cers

Ceramides

MHCers

Hexosylceramides or monohexosylceramides

SMs

Sphingomyelins

LC–MS/MS

Liquid chromatography–tandem mass spectrometry

Rt

Retention time

MA and AA

Manually and automatic integrated areas

MRM

Multiple reaction monitoring

d-saturated

The sphingoid base core without any double bonds

d-unsaturated

Sphingoid base core containing double bonds

Notation d18:1^Δ4 c16:0

A dihydroxy 18 carbon sphingoid base core structure with a double bond on the fourth carbon atom starting from the hydroxyl terminus of the core chain, and the c16:0 represents the number of carbon atoms linked to the amino side chain