Improved Sphingolipidomic Approach Based on Ultra-High

May 20, 2014 - the range m/z 110−1200 with accurate mass measurement of all peaks. A reference ... temperature of 400 °C and a flow rate of 12 L/mi...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Improved Sphingolipidomic Approach Based on Ultra-High Performance Liquid Chromatography and Multiple Mass Spectrometries with Application to Cellular Neurotoxicity Jing-Rong Wang,†,‡ Hongyang Zhang,†,§ Lee Fong Yau,† Jia-Ning Mi,† Stephanie Lee,⊥ Kim Chung Lee,⊥ Ping Hu,§ Liang Liu,*,†,‡ and Zhi-Hong Jiang*,†,‡ †

State Key Laboratory of Quality Research in Chinese Medicine, Macau Institute for Applied Research in Medicine and Health, Macau University of Science and Technology, Taipa, Macau, China ‡ School of Chinese Medicine, Hong Kong Baptist University, Kowloon Tong, Hong Kong, China § School of Chemistry and Molecular Engineering, East-China University of Science and Technology, Shanghai 200237, China ⊥ Agilent Technologies Hong Kong Ltd., North Point, Hong Kong, China S Supporting Information *

ABSTRACT: The emerging field of sphingolipidomics calls for accurate quantitative analyses of sphingolipidome. Existing analytical methods for sphingolipid (SPL) profiling often suffer from isotopic/isomeric interference, leading to the lowabundance, but biologically important SPLs being undetected. In the current study, we have developed an improved sphingolipidomic approach for reliable and sensitive quantification of up to 10 subclasses of cellular SPLs. By integratively utilizing high efficiency chromatographic separation, quadrupole time-of-flight (Q-TOF) and triple quadrupole (QQQ) mass spectrometry (MS), our approach facilitated unambiguous identification of several groups of potentially important but lowabundance SPLs that are usually masked by isotopic/isomeric species and hence largely overlooked in many published methods. The methodology, which featured a modified sample preparation and optimized MS parameters, permitted the measurement of 86 individual SPLs in PC12 cells in a single run, demonstrating great potential for high throughput analysis. The improved characterization, along with increased sensitivity for low-abundance SPL species, resulted in the highest number of SPLs being quantified in a single run in PC12 cells. The improved method was fully validated and applied to a lipidomic study of PC12 cell samples with or without amyloid β peptide (Aβ) treatment, which presents a most precise and genuine sphingolipidomic profile of the PC12 cell line. The adoption of the metabolomics protocol, as described in this study, could avoid misidentification and bias in the measurement of the analytically challenging low-abundance endogenous SPLs, hence achieving informative and reliable sphingolipidomics data relevant to discovery of potential SPL biomarkers for Aβ-induced neurotoxicity and neurodegenerative disease.

S

sphingosine (So), sphingosine-1-phosphate (S1P), ceramide (Cer), and ceramide-1-phosphate (C1P) have been increasingly recognized as bioactive signaling molecules that play important roles in diverse phenotypes and diseases ranging from inflammation, cancer,3 obesity,4 atherosclerosis,5 and neurodegenerative disorders.6

phingolipids (SPLs) are a complex family of compounds that share a common structural feature, the sphingoid base backbone synthesized de novo from serine and a long chain fatty acyl-CoA. The SPLs are found in all eukaryotic cells, where they comprise a small but vital fraction (2−20%) of the membrane lipids. Recent studies have placed this unusually versatile class of membrane lipids at the center of a number of important biological processes, such as cell proliferation/differentiation/apoptosis, migration, membrane trafficking, interactions, and morphology, as well as cellular signaling.1,2 SPL metabolites including © 2014 American Chemical Society

Received: November 21, 2013 Accepted: May 20, 2014 Published: May 20, 2014 5688

dx.doi.org/10.1021/ac5009964 | Anal. Chem. 2014, 86, 5688−5696

Analytical Chemistry

Article

SPLs in PC12 cells has been devised. Specifically, multiple mass spectrometries have been used for comprehensive profiling and high sensitivity quantification of SPLs based on the optimized separation. Quadrupole time-of-flight mass spectrometer (Q-TOF MS) was used for characterizing SPLs based on both accurate mass measurements at the MS and MS/MS level. This together with the isotopic distribution information permitted monitoring of isotopic/ isomeric interferences, which in turn facilitated accurate, reliable and sensitive quantification of SPLs using QQQ in MRM mode.

Since the classes and molecular species of SPLs present in the cellular sphingolipidome are interconnected in terms of their metabolism and homeostasis, changes in expression levels or enzyme activities in any pathway in the network will result in a new homeostasis for the cellular sphingolipidome. As such, profiling the entire sphingolipidome would contribute to a better understanding of the underlying biochemical mechanisms in biological processes and various disease states. This has led to the emergence of the discipline of sphingolipidomics,7−9 a branch of “omics” that predominantly involves comprehensive profiling and comparative quantification of SPLs to discriminate samples and discover biomarkers or patterns among sampled subpopulations. Recently, LC-MS/MS-based approaches with triple quadrupole (QQQ) mass spectrometer in multiple reaction monitoring (MRM) mode have been developed for quantitative analysis of SPLs, demonstrating enhanced sensitivity, and improved accuracy relative to routine LC-MS based approaches.10−16 However, because of the unprecedented complexity in SPL metabolism, there are still several major limitations associated with current LC-MS/MS based methodologies in SPL profiling. A key limitation lies in the fact that only the relatively abundant species of certain subclasses can be detected and accurately quantified, the resulting SPLs analyses thus reflect only the main changes in major SPL content but fail to detect the changes in low-abundance SPLs. However, it has been recognized that bioactive SPLs are often present at low levels of which the subtle changes may result in altered biological function.17,18 In addition, because of the interconversion of SPLs, small changes in upstream high abundance SPLs may result in cascade amplification in the levels of downstream low-abundance metabolites. It is therefore important to develop robust and quantitative analytical methodology for the detection of these low-abundance SPLs to understand and further elucidate the metabolic pathways and cell signaling networks involved in their biological activities. One of the major technical limitations in measuring lowabundance SPLs is the spectral interference arising from numerous isotope and isomeric species. Especially, the isotope contributions from a lower molecular mass moiety to a higher one may become significant when large concentration differences exist between the two species, leading to a failure in assignment or questionable quantification of the low-abundance species. Recently, a two-step procedure based on multidimensional mass spectrometry (MDMS) has been well-developed in which the first step was performed to analyze the abundant and nonoverlapping species, and the second step was carried out to quantify the overlapping and low-abundance species.19−25 Several high resolution mass spectrometry-based approaches were also developed for differentiating isobaric lipids.26−28 However, for conventional LC-MS/MS based method, developing practically achievable approach for resolving overlapped species and detecting low-abundance species is still quite necessary. Besides, because of the large chemical diversity of SPLs, several protocols have described the use of multiple LC systems for separating different classes of SPLs to avoid ion suppression, for example, using normal phase chromatography (LC-NH2 and LC-Si) for analyzing Cers and SMs, and reverse phase chromatography for measuring sphingoid bases. This strategy greatly increase the workload and turnaround time required for analysis, especially for lipidomics studies which typically involves a large number of samples.29,30 With these aforementioned in mind, a versatile sphingolipidomic approach for quantitative analysis of up to 10 subclasses of

■

EXPERIMENTAL SECTION SPLs Extraction. The combined procedure for the extraction of cellular SPLs was established according to the LIPID MAPS protocol reported by Merrill et al. with modifications.16 The cells in each dish (approximate number of 4.0 × 105) were washed twice with PBS and then were scraped from the dish in the residual PBS and transferred into a borosilicate glass tube with polytetrafluoroethylene coated top (tube I). After adding 0.5 mL of MeOH and 0.25 mL of CHCl3, 10 μL of internal standards cocktail (2.5 μM) and 10 μL of C12-sulfatide solution (2.5 μM) were added, and the contents were dispersed in an ultrasonicator at room temperature for 30 s. The mixture was incubated at 48 °C overnight to afford optimal extraction of SPLs (1st extraction).31 Then 75 μL of KOH in MeOH (1M) was added and were incubated in a shaking water bath for 2 h at 37 °C to cleave potentially interfering glycerolipids. After neutralization, four-step extraction of the mixture was performed to yield the SPL extract for analysis. Further details are available in the Supporting Information (Supporting methods and Figure S1-3). Chromatographic Conditions. Chromatographic separation was performed using an Agilent 1290 Infinity UHPLC system (Santa Clara, CA, USA), equipped with a binary solvent delivery system and a standard autosampler. An Agilent Eclipse Plus C18 column (100 × 2.1 mm, 1.8 μm) was used to separate the endogenous SPLs. The mobile phase consisted of (A) MeOH/H2O/HCOOH (60:40:0.2, v/v/v) and (B) MeOH/ IPA/HCOOH (60:40:0.2, v/v/v), both containing 10 mM NH4OAc. A linear gradient was optimized as follows (flow rate, 0.4 mL/min): 0−3 min, 0% to 10% B; 3−5 min, 10% to 40% B; 5−5.3 min, 40% to 55% B; 5.3−8 min, 55% to 60% B; 8−8.5 min, 60% to 80% B; 8.5−10.5 min, 80% to 80% B; 10.5−16 min, 80% to 90% B; 16−19 min, 90% to 90% B; 19−22 min, 90% to 100% B, followed by washing with 100% B and equilibration with 0% B. The injection volume was 2 μL, and the column temperature was maintained at 40 °C for each run. A typical run time was 20 min. Qualitative and Quantitative Analysis on Q-TOF and QQQ. Qualitative analysis was performed using an Agilent ultrahigh definition (UHD) 6550 Q-TOF mass spectrometer. Parameters for the Jet Stream technology included a superheated nitrogen sheath gas temperature of 400 °C and a flow rate of 12 L/min. ESI conditions were as follows: positive ion mode, capillary voltage 4000 V, nozzle voltage 300 V, nebulizer pressure 40 psi, drying gas 6 L/min, gas temperature 300 °C, skimmer voltage 65 V, octapole RF peak 500 V, fragmentor voltage 150 V. The targeted MS/MS collision energy (CE) was set at three different values: 20−60 eV. Mass spectra were recorded across the range m/z 110−1200 with accurate mass measurement of all peaks. A reference solution was nebulized for continuous calibration in positive ion mode using the following reference masses: m/z 121.0509 and 922.0098. The full-scan and MS/MS data were processed with Agilent Mass Hunter Workstation Software. 5689

dx.doi.org/10.1021/ac5009964 | Anal. Chem. 2014, 86, 5688−5696

Analytical Chemistry

Article

been suggested by previous studies. To achieve more subtle gradient of the strong elution solvent, we tested different combinations of MeOH/IPA. Finally, MeOH/IPA (60:40) was found to be the lowest composition for eluting out all species and an optimal composition for separating SMs and Cers, the most diverse species. More importantly, enhanced signal intensity for the low-abundance species, that is, Sa, So, and Cers was observed when using MeOH/IPA (60:40). Subsequently, we optimized the mobile phase with focus on the gradient for elution of SMs and Cers. The finalized mobile phase facilitated the baseline separation of SMs with varied unsaturation degree, while providing adequate resolution of other species as validated by PC12 cell samples. On the basis of the optimized separation, a linear correlation between carbon numbers vs retention time was observed for Cer, SM, and HexCer. Moreover, linear regression model was established by plotting unsaturation degree of SMs and retention time with good fitting (r2 > 0.98 for each series), suggesting a capability for predicting the retention time of unsaturated SM, which could be further employed as a supporting evidence for the identification of unsaturated SMs (Supporting Information, Figure S5). However, as elution gradient was employed to resolve individual species on reverse-phase column, changes in the components of the mobile phase with the gradient progress may cause differential ionization efficiency of SPLs eluted at different retention time.20,21 Hence we examined the ionization efficiency of SM and Cer standards under different gradient of mobile phase, revealing that ionization of SM was more resistant to the changes of mobile phase composition, whereas Cers with long N-acyl chain was sensitive to the changes (Supporting Information, Figure S6). Improved Characterization and Quantification of Dihydrosphingomyelins (DHSMs) that are Overlapped by Isotopic Ions of Sphingomyelins (SMs). SMs are prone to fragment into a polar headgroup, resulting in a high abundant product ion of the phosphocholine headgroup at m/z 184 and with very little fragments derived from backbone cleavage. This headgroup-specific fragment was therefore employed for both identification and quantification of SMs. However, when using this headgroupdependent product ion, overlapping signals from the exact mass of the [M + 2] isotopic ions of acyl-chain-matched SMs might compromise reliable characterization, and could result in isotopic interference in quantification in LC-MS/MS analysis when using MRM technique. In mammalian cells, SMs with a sphingosine (d18:1) backbone are the dominant species, whereas DHSMs with sphinganine (d18:0) as a backbone are present in relatively low abundance. As a result, the [M + 2] isotopic ions of SMs could seriously compromise the characterization and quantification of DHSMs that are a minor, but biologically important species. Because of the very similar elemental composition of DHSMs and SMs, the overlapped ions from these two species would not be fully resolved even at the 40 000 full width at half-maximum (fwhm) resolving power available with the Q-TOF mass spectrometer employed in this study. The optimized chromatographic separation was found to be essential for accurate characterization and quantification of these species. For example, under nonoptimized chromatographic conditions, the chromatographic peak of SM (d18:0/16:0) completely coeluted with that of SM (d18:1/16:0) (Figure 1A), leading to an overlap of the [M + H]+ of SM (d18:0/16:0) (calculated m/z = 705.5905) and [M+2+H]+ of SM (d18:1/16:0) (calculated m/z = 705.5812) signals (Figure 1). This phenomenon of signal overlap in turn led to an unexpected decrease in mass accuracy for these ions

Quantitative analysis was carried out using an Agilent 6460 QQQ mass spectrometer (Santa Clara, CA, USA). The Jet Stream parameters were a superheated nitrogen sheath gas temperature of 400 °C and a flow rate of 12 L/min. ESI conditions were optimized as follows: positive ion mode, capillary voltage 3500 V, nozzle voltage 300 V, nebulizer pressure 40 psi, drying gas 6 L/min, and gas temperature 300 °C. To ensure the maximal sensitivity, a two segment scan was adopted to monitor separately the transitions of analytes. The optimized parameters such as characteristic transitions (precursor ion → product ion), fragmentor voltages, and CE values selected for each individual compound are shown in Supporting Information Table S2. Establishment of Personal SPL Database. A personal SPL database based on the LIPID MAPS information was established in the Agilent Mass Hunter Personal Compound Database and Library (PCDL) software, whereby 4318 SPLs, which were recorded online (3 May 2013), were included. The database combined with UHPLC-Q-TOF MS enabled reliable screening and identification of SPLs in the biological samples. Sphingolipidomic Application and Data Analysis. The sphingolipidomic approach was applied to PC12 cell samples with (models, n = 10) or without Aβ treatment (controls, n = 9). The reports for quantitative analysis were established using Agilent Mass Profiler Professional (MPP) software, where concentration unit unification and internal standards exclusion were performed. The resulting 3-D matrix, including SPL names (variables), sample names (observations), and concentrations were imported into SIMCA-P+ 13.0 software (Umetrics, Umea, Sweden) for multivariate statistical analysis. Principal component analysis (PCA) was used to visualize general clustering between controls and models. Partial least-squares discriminant analysis (PLS-DA) was carried out to identify the differentially expressed SPLs responsible for the separation. Statistical analyses of the differentially expressed lipids were performed by using T-test. P values less than 0.05 were considered significant.

■

RESULTS AND DISCUSSION Integrated Approach for Comprehensive Profiling and Accurate Quantification of SPLs. To overcome the limitations associated with the use of mass spectrometers with only unit mass resolving power (e.g., ion traps or triple quadrupoles), as in previous LC-MS/MS approaches, we developed an integrated approach for comprehensive profiling and improved quantification of SPLs by utilizing UHPLC-Q-TOF in conjunction with UHPLC-QQQ (Supporting Information, Figure S4). The first and foremost objective of this approach was to achieve improved identification of low-abundance SPLs in which their MS signals might be masked by isotopic, isomeric or isobaric overlaps. The identified SPLs were subsequently subjected to LC-MS/MS analysis with QQQ in MRM mode to achieve sensitive quantification. Parallel experiments with Q-TOF and QQQ were performed during the optimization of the chromatographic separation, and these were found to be crucial for reliable identification and quantification. Optimized Chromatographic Separation of SPLs. Liquid chromatography condition was optimized to achieve adequate separation of isotopic species while providing resolving of other species. Separation was carried out on a common C18 column, which was proved to be superior to C8 and HILLIC column. MeOH/isopropanol (IPA) was demonstrated to be a more suitable combination in comparison with ACN/IPA as it provides enhanced separation of SMs with varied unsaturation degree. Addition of 10 mM NH4OAc and 0.2% HCOOH were proved to be essential for the separation of most species, as have 5690

dx.doi.org/10.1021/ac5009964 | Anal. Chem. 2014, 86, 5688−5696

Analytical Chemistry

Article

Figure 1. Improved characterization and quantification of SPLs supported by optimized separation. Extracted compound chromatograms [ECC, an extracted ion chromatogram (EIC) computed using all of the masses in the compound spectrum] of SM (d18:1/16:0) and SM (d18:0/16:0) (left panel of A and B) and their high resolution mass spectra obtained on Q-TOF (middle panel of A and B, red boxes around the mass and its isotope showed the predicted isotope distribution) under nonseparated (A) and baseline separated conditions (B). MRM chromatograms for each species obtained under the same chromatographic conditions shown in the right panel of A and B. Compound identification results under different separation conditions were given in C. aDenotes the ratio of peak area determined in MRM mode using QQQ. bCalculated using the following equation: [Peak area measured for m/z 705 → 184 − Predicted peak area of [M + 2] isotopic ion of SM (d18:1/16:0)]/Peak area of SM (d18:1/16:0). The area derived from isotopic ion was predicted by calculating the [M + 2] isotope abundance of the molecular formula of SM (d18:1/16:0) (subtract the elemental composition of headgroup) by using isotope distribution tool built into the Agilent Mass Hunter software.

corresponding to both species (Figure 1C). Although this did not impact on the characterization of SM (d18:1/16:0), identification of SM (d18:0/16:0) tended to be problematic, although the high isotopic abundance of the ion at m/z 705 indicated the

presence of this species (Figure 1A). By contrast, baseline separation of the two species resulted in significantly improved mass accuracy of the ion at m/z 705 (Figure 1B, C), facilitating unambiguous identification of these two species. 5691

dx.doi.org/10.1021/ac5009964 | Anal. Chem. 2014, 86, 5688−5696

Analytical Chemistry

Article

The improved identification of SMs and DHSMs achieved via optimized separation was found to have especially merits during quantitative analysis with QQQ mass spectrometer in MRM mode. In the example of SM (d18:1/16:0) and the corresponding DHSM under nonoptimized chromatographic condition, MRM monitoring of the transition m/z 705 → 184 resulted in a single peak with the same retention time to that of m/z 703 → 184, suggesting an existence of isotopic interference (Figure 1A, right panel). With the optimized separation, the “real” peak for DHSM can be confidently differentiated from the interfering peak derived from the SM counterpart on QQQ platform (Figure 1B, right panel), which cannot be accomplished accurately in LC-MS/MS analysis with a QQQ analyzer solely. Meanwhile, since overlapped isotopic ions lead to a decreased mass accuracy and altered isotope distribution, a rigorous identification of SM (d18:0/16:0) based on high mass accuracy and accurate isotope distribution can ensure elimination of isotopic interference in return, and therefore permit accurate quantification of the low-abundance DHSM. More importantly, optimized separation was found to be crucial for the identification of minor DHSMs, which otherwise could be totally masked by isotopes of the corresponding SMs. This can be exemplified by the characterization and quantification of SM (d18:0/15:0) (Supporting Information, Figure S7). Using the integrated approach, 9 DHSMs with N-acyl chains ranging from C14 to C24 were successfully identified and accurately quantified, based on baseline separation with their SM counterparts (Supporting Information, Figure S8). For unambiguous identification of the dihydrosphingolipid in PC12 cells, accurate mass, retention time and complete MS/MS patterns of these species were compared with those of model compounds, dihydroceramide standards and synthesized dihydrosphingomyelins by using sphingomyelin standards with varied chain length of (C16−C24) as starting materials. Dihydrosphingolipid usually exists at low levels in cells, except for the unusual DHSM enrichment in the human lens,32 and in the human immunodeficiency virus (HIV)-1 envelope.33,34 DHSM interacts with cholesterol35 with higher affinity than acyl-chain-matched SM and form lipid domains different from and more rigid than SM-rich lo phases,36 suggesting that DHSMs are a group of low-abundant but biologically important SPLs. Accurate quantification of these highly important species will provide a new analysis tool for exploring the role of DHSMs and the biological function of the corresponding enzymes. Differentiation of Isomeric SPLs. Another issue in the identification of SPLs is the differentiation of isomeric species having the identical molecular elemental composition and differing only in their acyl chain composition. With optimized separation, this kind of isomeric species can be differentiated using MS/MS with Q-TOF. As can be seen from Figure 2A, a pair of isomers appeared at retention times of 15.6 and 15.8 min, each peak having exactly the same protonated molecular mass of m/z 634, but their targeted MS/MS spectra revealed characteristic fragment ions corresponding to the sphingosine (d18:1) (m/z 264.2680) and the sphingosine (d18:2) (m/z 262.2522) backbone, respectively. Thus, the two isomers were assigned to be Cer (d18:1/23:1) and Cer (d18:2/23:0). The assignment of the two isomers was further confirmed by an MRM experiment in which the transition pair, m/z 634 → 264 and 634 → 262, corresponded to a single peak with specific retention time. Compared with the Cers, characterization of SMs is relatively difficult because they are more prone to be fragmented into polar head groups. By employing relatively high CE (40−60 eV),

Figure 2. Differentiation of SPL isomers by targeted MS/MS. (A) Two peaks observed in extracted ion chromatograms (EIC) of m/z 634.6133 (with accurate mass window of 5 ppm) showed characteristic product ion corresponding to Cer (d18:1/23:1) (m/z 264.2680, and its [M + 1] isotopic ion at m/z 265.2714, red boxes around the Mass and its isotope showed the predicted isotope distribution) and Cer (d18:2/23:0) (m/z 262.2522) in targeted MS/MS at respective time point. (B) Targeted MS/MS of the isomeric SMs with [M + H]+ at m/z 717.5905 yield characteristic product ion of SM (d19:1/16:0) at m/z 278.2876 [corresponding to sphingosine (d19:1) backbone] and SM (d18:1/17:0) at m/z 264.2675 [corresponding to sphingosine (d18:1) backbone], respectively.

targeted MS/MS provided an enhanced characterization of most SMs, including several pairs of isomers. For example, for the candidate SM (d18:1/17:0), the extracted ion chromatogram of m/z 717.5905 (accurate mass window 5 ppm) yielded two peaks at 11.6 and 11.7 min, respectively. Targeted MS/MS of m/z 717 at respective time points gave product ions corresponding to the sphingosine (d19:1) backbone (m/z 278.2876) and the sphingosine (d18:1) backbone (m/z 264.2675) (Figure 2B), providing evidence for the identification of these two peaks as SM (d19:1/16:0) and SM (d18:1/17:0). This enabled subsequent quantification of these two isomers by the MRM technique. It should be noted that identification of these multiple-peaks with identical elemental composition is almost impossible in infusion mode using a precursor ion scan with m/z 184 as the product ion and subsequent MRM experiment. Based on the targeted MS/MS, 47 SMs were identified with characteristic product ions, among which, four pairs of isomers were successfully differentiated (Supporting Information, Table S1). This facilitated accurate quantification of respective species in MRM mode. Discovery of Novel Highly Unsaturated Sphingolipids. The high-resolution spectra and MS/MS afforded by TOF, together with optimized chromatographic separation, also enabled us to discover a series of SMs with highly unsaturated N-acyl chain, among which six SMs, i.e., SM (d18:2/20:2), SM (d18:2/22:2), SM (d18:2/22:3), SM (d18:1/23:3), SM (d18:1/25:3), and SM (d18:1/26:3) have hitherto not been reported (Supporting Information, Table S1), while the others have been reported to be presented in plasma or breast milk.37−39 These highly unsaturated SMs were characterized based on their accurate 5692

dx.doi.org/10.1021/ac5009964 | Anal. Chem. 2014, 86, 5688−5696

Analytical Chemistry

Article

Figure 3. Characterization of a series of SMs with the same total carbon number but with increasing degree of unsaturation by using Q-TOF MS (A, B) and quantification of the SMs in MRM mode (C). In MRM mode, each unsaturated SM could generate an isotopic interference on SMs with less unsaturation sites. When the level of the SM was much lower than that of SM with one more unsaturation site, the isotopic interference could be very significant, as exemplified by SM (d18:0/22:0).

Scheme 1. Biosynthesis and Metabolic Pathways for Sphingolipidsa

a

SPT = serine palmitoyl transferase; 3-KR = 3-keto-dihydrosphingosine reductase; CerS = ceramide synthase; Cdase = ceramidase; DES = dihydroceramide desaturase; SphK = sphinganine kinase; SMase = sphingomyelinase; SMS = sphingomyelin synthase; SPP = Sphingosine phosphate phosphatase; SK = Sphingosine kinase.

for [M + Na]+ of SM (d18:1/22:0) (calculated m/z 809.6507). Under this case, [M + H]+ was assigned based on the observation of corresponding [M + Na]+. Moreover, as [M + Na]+ is not prone to fragment into structure-specific product ions, hence MRM signal can be employed as evidence for discrimination of [M + H]+ and [M + Na]+ ions of SM (Supporting Information, Figure S9).

mass, isotope distribution and diagnostic fragment ion at m/z 184, 264 or 262, as well as subsequent LC-MS/MS analysis in MRM mode (Figure 3). It is worth noting that ions of these species could experience overlap by the adduct ions of SPLs with shorter N-acyl chains. For example, the [M + H]+ of SM (d18:1/ 24:3) (calculated m/z 809.6531) coincides exactly with the m/z 5693

dx.doi.org/10.1021/ac5009964 | Anal. Chem. 2014, 86, 5688−5696

Analytical Chemistry

Article

Figure 4. Quantitative analysis of SPLs in PC12 cells. The amounts of these SPLs were measured by UHPLC-QQQ MS using 10 internal standards (n = 9).

as described in the Experimental Section. The LODs (peak-to-peak calculation with signal-to-noise ratio of 3) for all standards were at the sub pmol (0.003−0.87 pmol) level (Supporting Information Table S3−4), indicating that the proposed method is highly sensitive and linear over a wide dynamic range. A crucial issue for an “-omic” approach is the suitability for quantification of alterations in the level of individual species. To address this issue, we carried out an experiment in which exogenous and endogenous species were spiked into the PC12 samples with varied amount. Correlation between the added and determined amount of individual species was examined (Supporting Information, Figure S12), and demonstrated a good applicability of our approach for quantification of changes over a wide range. Quantification of Sphingolipids in PC12 Cells and Cellular Neurotoxicity. As part of an ongoing study,41 the validated method was applied to PC12 cell samples. Levels for all 86 species of SPLs that were identified in the PC12 cell line were measured using the developed method. The results (Figure 4) illustrated several interesting points. (1) The relative abundance of different subclasses of SPLs varied greatly, with SM being the most abundant species, followed by Cer, HexCer, ST, C1P, sphingoid bases and their 1-phosphates. (2) SPLs with an N-acyl chain length of C16 were the most abundant species within each subclass except for ST of which the C18 acyl chain was the dominant species. (3) The amounts of SPLs were generally several orders of magnitude higher than that of the corresponding DHSMs. (4) DHSMs and most highly unsaturated SMs were minor species, existing in the PC12 cells at levels much lower than other species.

The major characteristic of these highly unsaturated SMs unveiled in this study is their sphingosine backbone and up to three unsaturated sites in the N-acyl chain, resulting in up to four unsaturated sites in total within the structure. Four d18:2 SMs with a highly unsaturated N-acyl chain were also assigned, in which the total degree of unsaturation reached five. Existence of this series of unsaturated SMs could introduce even greater complexity in the isotope overlap issue, thereby compromising SPL profiling and quantification based on isotope correction (Figure 3). However, it should be noted that for unsaturated fatty acid there might be isomers with different double bond locations, which need to be characterized by other approaches.40 Optimized MRM Conditions for LC-MS for All Subclasses of SPLs in PC12 Cells. The ESI conditions as well as the MS/MS fragmentations (including characteristic precursor/product ion pairs, fragmentor voltages and CE values) for each internal standard were optimized for quantification of biological SPLs (Supporting Information Table S2). Notably, sulfatide (ST) has been conventionally analyzed in negative ion mode with the loss of sulfate (m/z 96.9). Our experiment suggested that STs can be detected in positive MRM mode with significantly enhanced sensitivity (Supporting Information Figure S10). Based on the comprehensive profiling of SPLs with Q-TOF and optimized MRM conditions for each subclass of SPLs, the UHPLC-MS/MS method was developed for the 86 identified SPLs. The MRM chromatograms of all identified SPLs in the PC12 cell line are shown in the Supporting Information (Figure S11), together with the relevant data for each MRM transition (Supporting Information Table S1). Validation of MRM profiling was performed 5694

dx.doi.org/10.1021/ac5009964 | Anal. Chem. 2014, 86, 5688−5696

Analytical Chemistry

Article

quantification from which more accurate reliable and meaningful metabolic conclusions can be drawn.

As an application of the sphingolipidomic approach in cellular neurotoxicity, PC12 cells with (models) or without (controls) Aβ treatment were analyzed in parallel. Multivariate statistical analysis, including unsupervised PCA and supervised PLS-DA methods, were used to differentiate the two groups. Twentyeight potential biomarkers were identified among which several low abundance species such as DHSM, odd N-acyl chain species, and highly unsaturated SMs were included (Supporting Information, Figure S13 and Table S5). This result showed that application of the improved sphingolipidomic approach could afford comprehensive information on the changes of individual SPLs in response to Aβ treatment, which in turn could provide crucial clues for the identification of key enzyme isoforms involved in the SPL metabolism associated with the Aβ-induced neurotoxicity.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as indicated in text. This material is available free of charge via the Internet at http://pubs.acs.org/.

■

AUTHOR INFORMATION

Corresponding Authors

*Tel: +853-8897 2238. Fax: +853-2882 5886. E-mail: lliu@must. edu.mo. *Tel.: +853-8897 2777. Fax: +853-2882 5886. E-mail: zhjiang@ must.edu.mo.

■

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. J.-R.W. and H.Z. contributed equally to the work.

CONCLUSIONS The majority of LC-MS/MS methodologies published for SPLs analysis to date involved chromatographic separation at the subclass level coupled to tandem MS with unit mass resolving power. These methodologies are effective for the detection of several subclasses of high-abundance SPLs, but not comprehensive enough to profile the complexity of SPLs that arises from the metabolic interconnections of bioactive SPLs. Numerous highly important but low-abundant SPLs were clearly undetected using such analytical strategies, resulting in a lack of information on low-abundance SPLs, which are involved in basic biochemical mechanisms. Isotopic interferences and isomeric interferences represent one of the major barriers for detecting such lowabundance SPLs. By using a combined analytical strategy, several groups of low-abundance SPLs, which would otherwise be masked by isotopic/isomeric species, were unambiguously characterized by high-resolution Q-TOF MS with the support of optimized chromatographic separation of the structurally similar SPLs. This new approach distinguished DHSMs from the [M + 2] isotope of the corresponding SMs; differentiated isomeric SMs which differs in their N-acyl chain composition, and also resulted in the discovery of a series of highly unsaturated SMs, some of which to date have not been reported. This improved specificity in identification, in turn, facilitated accurate quantification of all low-abundance species with elimination of isotopic/isomeric interference. In total 86 SPLs in PC12 cells could be successfully analyzed in a single run. As a consequence, the use of ultrahigh performance chromatographic separation, high-resolution MS and high sensitivity MS/MS for highthroughput, high specificity and high sensitivity for sphingolipidomic analyses is strongly recommended. For application the method was applied to a comparative study of PC12 cell samples with or without Aβ treatment, which plays a key role in the pathogenesis of Alzheimer disease by inducing neurotoxicity and cell death. Collectively, our improved sphingolipidomic approach based on UHPLC-Q-TOF MS and UHPLC-QQQ MS provided enhanced characterization and quantification of SPLs from which more reliable and meaningful metabolic conclusions can be drawn. Research also highlighted the issue of isomeric and isotopic interference in SPLs analysis, in particular low-abundance species, should be carefully evaluated in sphingolipidomic studies to avoid errors in characterization and subsequent quantification. It has been proposed that improved quantitative lipidomic measurements will become a standard clinical tool to ensure reliable diagnostics.42 Our improved sphingolipidomic approach based on UHPLC-Q-TOF MS and UHPLC-QQQ MS has provided enhanced characterization and

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by the Macao Science and Technology Development Fund, Macau Special Administrative Region (039/2011/A2 to Z.-H.J.).

■

REFERENCES

(1) Lahiri, S.; Futerman, A. H. Cell. Mol. Life Sci. 2007, 64, 2270−2284. (2) Merrill, A. H., Jr.; Wang, M. D.; Park, M.; Sullards, M. C. Trends Biochem. Sci. 2007, 32, 457−468. (3) Ryland, L. K.; Fox, T. E.; Liu, X.; Loughran, T. P.; Kester, M. Cancer. Biol. Ther. 2011, 11, 138−149. (4) Shah, C.; Yang, G.; Lee, I.; Bielawski, J.; Hannun, Y. A.; Samad, F. J. Biol. Chem. 2008, 283, 13538−13548. (5) Hojjati, M. R.; Li, Z.; Zhou, H.; Tang, S.; Huan, C.; Ooi, E.; Lu, S.; Jiang, X. C. J. Biol. Chem. 2005, 280, 10284−10289. (6) He, X.; Huang, Y.; Li, B.; Gong, C. X.; Schuchman, E. H. Neurobiol. Aging 2010, 31, 398−408. (7) Sullards, M. C.; Allegood, J. C.; Kelly, S.; Wang, E.; Haynes, C. A.; Park, H.; Chen, Y.; Merrill, A. H., Jr. Methods Enzymol. 2007, 432, 83− 115. (8) Merrill, A. H., Jr.; Stokes, T. H.; Momin, A.; Park, H.; Portz, B. J.; Kelly, S.; Wang, E.; Sullards, M. C.; Wang, M. D. J. Lipid Res. 2009, 50 (Suppl), S97−102. (9) Li, J.; Hoene, M.; Zhao, X.; Chen, S.; Wei, H.; Haring, H. U.; Lin, X.; Zeng, Z.; Weigert, C.; Lehmann, R.; Xu, G. Anal. Chem. 2013, 85, 4651−4657. (10) Lieser, B.; Liebisch, G.; Drobnik, W.; Schmitz, G. J. Lipid Res. 2003, 44, 2209−2216. (11) Jiang, X.; Han, X. J. Lipid Res. 2006, 47, 1865−1873. (12) Liebisch, G.; Drobnik, W.; Reil, M.; Trumbach, B.; Arnecke, R.; Olgemoller, B.; Roscher, A.; Schmitz, G. J. Lipid Res. 1999, 40, 1539− 1546. (13) Hsu, F. F.; Turk, J. J. Am. Soc. Mass Spectrom. 2000, 11, 437−449. (14) Zarei, M.; Kirsch, S.; Muthing, J.; Bindila, L.; Peter-Katalinic, J. Anal. Bioanal. Chem. 2008, 391, 289−297. (15) Ikeda, K.; Shimizu, T.; Taguchi, R. J. Lipid Res. 2008, 49, 2678− 2689. (16) Merrill, A. H., Jr.; Sullards, M. C.; Allegood, J. C.; Kelly, S.; Wang, E. Methods 2005, 36, 207−224. (17) Marsolais, D.; Rosen, H. Nat. Rev. Drug Discovery 2009, 8, 297− 307. (18) Hannun, Y. A.; Obeid, L. M. Nat. Rev. Mol. Cell Biol. 2008, 9, 139− 150. (19) Han, X. Biochim. Biophys. Acta 2010, 1801, 774−783. 5695

dx.doi.org/10.1021/ac5009964 | Anal. Chem. 2014, 86, 5688−5696

Analytical Chemistry

Article

(20) Han, X.; Yang, K.; Gross, R. W. Mass Spectrom. Rev. 2012, 31, 134−178. (21) Yang, K.; Zhao, Z.; Gross, R. W.; Han, X. J. Chromatogr. B 2009, 877, 2924−2936. (22) Wang, M.; Hayakawa, J.; Yang, K.; Han, X. Anal. Chem. 2014, 86, 2146−2155. (23) Han, X.; Gross, R. W. Expert Rev. Proteom. 2005, 2, 253−264. (24) Han, X.; Yang, J.; Cheng, H.; Ye, H.; Gross, R. W. Anal. Biochem. 2004, 330, 317−331. (25) Han, X.; Yang, K.; Cheng, H.; Fikes, K. N.; Gross, R. W. J. Lipid Res. 2005, 46, 1548−1560. (26) Schuhmann, K.; Herzog, R.; Schwudke, D.; Metelmann-Strupat, W.; Bornstein, S. R.; Shevchenko, A. Anal. Chem. 2011, 83, 5480−5487. (27) Schwudke, D.; Schuhmann, K.; Herzog, R.; Bornstein, S. R.; Shevchenko, A. CSH Perspect. Biol. 2011, 3, a004614. (28) Fhaner, C. J.; Liu, S.; Ji, H.; Simpson, R. J.; Reid, G. E. Anal. Chem. 2012, 84, 8917−8926. (29) Bielawski, J.; Szulc, Z. M.; Hannun, Y. A.; Bielawska, A. Methods 2006, 39, 82−91. (30) Scherer, M.; Leuthauser-Jaschinski, K.; Ecker, J.; Schmitz, G.; Liebisch, G. J. Lipid Res. 2010, 51, 2001−2011. (31) van Echten-Deckert, G. Methods Enzymol. 2000, 312, 64−79. (32) Byrdwell, W. C.; Borchman, D.; Porter, R. A.; Taylor, K. G.; Yappert, M. C. Invest. Ophthalmol. Vis. Sci. 1994, 35, 4333−4343. (33) Brugger, B.; Glass, B.; Haberkant, P.; Leibrecht, I.; Wieland, F. T.; Krausslich, H. G. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 2641−2646. (34) Lorizate, M.; Brugger, B.; Akiyama, H.; Glass, B.; Muller, B.; Anderluh, G.; Wieland, F. T.; Krausslich, H. G. J. Biol. Chem. 2009, 284, 22238−22247. (35) Kusumi, A.; Koyama-Honda, I.; Suzuki, K. Traffic 2004, 5, 213− 230. (36) Vieira, C. R.; Munoz-Olaya, J. M.; Sot, J.; Jimenez-Baranda, S.; Izquierdo-Useros, N.; Abad, J. L.; Apellaniz, B.; Delgado, R.; MartinezPicado, J.; Alonso, A.; Casas, J.; Nieva, J. L.; Fabrias, G.; Manes, S.; Goni, F. M. Chem. Biol. 2010, 17, 766−775. (37) Blaas, N.; Schuurmann, C.; Bartke, N.; Stahl, B.; Humpf, H. U. J. Agric. Food Chem. 2011, 59, 6018−6024. (38) Quehenberger, O.; Armando, A. M.; Brown, A. H.; Milne, S. B.; Myers, D. S.; Merrill, A. H.; Bandyopadhyay, S.; Jones, K. N.; Kelly, S.; Shaner, R. L.; Sullards, C. M.; Wang, E.; Murphy, R. C.; Barkley, R. M.; Leiker, T. J.; Raetz, C. R.; Guan, Z.; Laird, G. M.; Six, D. A.; Russell, D. W.; McDonald, J. G.; Subramaniam, S.; Fahy, E.; Dennis, E. A. J. Lipid Res. 2010, 51, 3299−3305. (39) Li, J.; Hu, C.; Zhao, X.; Dai, W.; Chen, S.; Lu, X.; Xu, G. J. Chromatogr. A 2013, 1320, 103−110. (40) Yang, K.; Zhao, Z.; Gross, R. W.; Han, X. Anal. Chem. 2011, 83, 4243−4250. (41) Zhang, H.; Wang, J. R.; Yau, L. F.; Ho, H. M.; Chan, C. L.; Hu, P.; Liu, L.; Jiang, Z. H. Mol. BioSyst. 2012, 8, 3208−3215. (42) Shevchenko, A.; Simons, K. Nat. Rev. Mol. Cell Biol. 2010, 11, 593−598.

5696

dx.doi.org/10.1021/ac5009964 | Anal. Chem. 2014, 86, 5688−5696