Accepted Manuscript Title: One-step lipid extraction for plasma lipidomics analysis by liquid chromatography mass spectrometry Authors: Yoshinori Satomi, Megumi Hirayama, Hiroyuki Kobayashi PII: DOI: Reference:

S1570-0232(16)31380-0 http://dx.doi.org/10.1016/j.jchromb.2017.08.020 CHROMB 20753

To appear in:

Journal of Chromatography B

Received date: Revised date: Accepted date:

6-12-2016 1-8-2017 15-8-2017

Please cite this article as: Yoshinori Satomi, Megumi Hirayama, Hiroyuki Kobayashi, One-step lipid extraction for plasma lipidomics analysis by liquid chromatography mass spectrometry, Journal of Chromatography Bhttp://dx.doi.org/10.1016/j.jchromb.2017.08.020 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

One-step lipid extraction for plasma lipidomics analysis by liquid chromatography mass spectrometry

1

Yoshinori Satomi, 1Megumi Hirayama and 1Hiroyuki Kobayashi

1

Integrated Technology Research Laboratories, Pharmaceutical Research Division, Takeda

Pharmaceutical Company Limited

Corresponding author: Yoshinori Satomi ([email protected]) TEL: +81-466-32-1526 Address: 26-1, Muraoka-Higashi 2-chome, Fujisawa, Kanagawa, Japan 251-8555

Highlights 

Simple and quick sample preparation for lipidomics analysis



Protein precipitation by alcohol, especially ethanol, was revealed to be suitable sample preparation method for LC/MS based lipidomics analysis



Protein precipitation by acetonitrile was not appropriate for sample preparation method for lipidomics analysis due to the incomplete protein denaturing ability

Abstract In the past decade, various lipidomics methodologies have been developed using mass spectrometry based analytical technologies, enabling wide coverage lipid detection in a quantitative manner. Hence, lipidomics has become a widely-accepted approach for biomarker discovery and mechanism elucidation in both medical and biology research fields; however,

1

there are still technical challenges. In this study, focusing on the sample preparation procedure, a single step deproteinization by a water-soluble organic solvent, such as methanol (MeOH), ethanol (EtOH), isopropanol (IPA) or acetonitrile (ACN), was evaluated and proved to be satisfactory for lipidomics analysis. Moreover, during this investigation ACN deproteinization was revealed to not be an effective method for lipid extraction because lipid decomposition was observed during the protein precipitation process through lipase activation, potentially due to the insufficient protein denaturation. Therefore, excluding ACN, protein precipitation by alcohol was evaluated as the lipid extraction reagent. Moreover, adding the MTBE-MeOH (mMM) method, one of the major liquid-liquid extraction methods for shotgun lipidomics, these four approaches were compared. Lipids were extracted from mouse plasma by these four methods and used for exhaustive lipid profiling by liquid chromatography mass spectrometry (LC/MS) analysis. Comparison of these four methods revealed that alcohol based protein precipitation was a useful sample preparation procedure for LC/MS based lipidomics analysis. Whereas MeOH extraction was appropriate for hydrophilic lipid species, IPA was effective for hydrophobic lipids such as triacylglycerols (TG). In practice, EtOH extraction is thought to be the best approach to cover wide range of lipid species using a simple preparation procedure.

Abbreviations

LC/MS, Liquid chromatography mass spectrometry; IPA, 2-propanol; EtOH, Ethanol; MeOH, Methanol; ACN, Acetonitrile; MTBE, Methyl-tert-butyl ether; TG, Triacylglycerol; DG, Diacylglycerol; MG, Monoacylglycerol; CE, Cholesteryl ester; (L)PC, (Lyso)phosphatidylcholine; (L)PE, (Lyso)phosphatidylethanolamine; (L)PA, (Lyso)phosphatidic acid; (L)PG, (Lyso)phosphatidylglycerol; (L)PI , (Lyso)phosphatidylinositol; Cer, Ceramide; 2

SM, Sphingomyelin; Hex-Cer, Hexosyl (glycolyl or galactosyl) ceramide; FFA, Free fatty acid; CV, Coefficient of variation;

Keywords Lipidomics, liquid chromatography mass spectrometry, sample preparation, alcohol based protein precipitation

1

Introduction

In the past decade, lipidomics has been established as a subset of the metabolomics research field with a focus on non-water soluble metabolites in the body. The extensive roles of lipids as structural components, energy storage, or with their own biological activities have attracted researchers to discover new mechanisms or biomarkers using lipidomics technologies [1–10]. The diverse function of lipids is attributed to the variety structure of the molecular species. The complexity of the lipid world made lipidomics analysis challenging; however, advancement in recent years has encouraged effective use of lipidomics technology for a wide range of applications. As a broad category, there are two types of methodologies for lipidomics analysis, both of which are mass spectrometry based. One is a method that directly introduces a lipid fraction into the mass spectrometer without chromatographic separations, called shotgun lipidomics [11–19]. Taking advantage of the molecular discrimination ability and high-sensitivity of mass spectrometers, shotgun lipidomics enabled high-throughput, robust, and quantitative analysis. Although a snapshot of the lipid profile can be quickly obtained by this approach, constituents in the lipid fraction interfere with the detection of some components. The second method employs liquid chromatography mass spectrometry (LC/MS) aimed at in-depth 3

analysis for observing a wider range or higher sensitivity for molecular species [18,20–27]. The diverse conditions for a liquid chromatography system, selection of a separation column, mobile phase composition, or gradient conditions, enables development of a variety of methods depending on the target molecules of interest; however, it takes a longer run time to cover a broad range of molecular species, and sometimes establishing suitable chromatographic separation for every target molecule is challenging. Therefore, in order to develop the best analytical conditions, a huge effort is often required. Considering the advantages, disadvantages, needs, and interests, for an analytical technique multiple approaches can be applied for lipidomics research. Despite the variety of analytical methods, the pre-analytical sample preparation process has mostly relied on traditional means. For example, a liquid-liquid extraction (LLE) with an optimized blend of chloroform-methanol-water, developed by Bligh  Dyer [28] or Folch [29], have been acknowledged as the best lipid extraction methods. On the other hand, recent reports demonstrated other ideas, such as the MTBE-MeOH [30,31] and BUME (butanol-methanol) methods [32], as easy-to-use approaches for lipidomics analyses that provide almost the equivalent quality of traditional methods. In practice, the lipid fraction extracted by these LLE methods is considered to be best for shotgun lipidomics analysis because of the selectivity and extraction efficiency of the lipid fraction; however, for LC/MS based lipidomics analyses, the LLE samples cannot be directly injected to the liquid chromatography instrument, especially in the case of a reverse-phase system, to obtain proper chromatographic separation. Therefore, solvent reconstitution, dry-up and re-dissolving with a water-soluble organic solvent is necessary for LC/MS based lipidomics analyses. The extra preparation steps cause additional sample preparation time and affect the analytical quality and molecular coverage. As an alternative method, protein precipitation by a water-soluble organic solvent can be used for lipid 4

extraction; however, the non-lipid fraction is generally extracted in the same sample. The potential of protein precipitation methods was discussed in a previous report to evaluate the benefit for shotgun lipidomics analysis [33], but there still showed disadvantages for acquiring a wide coverage of lipid profiling, probably because of interference from the abundant non-lipid fraction. This disadvantage can be avoided by using LC/MS based lipidomics analysis. In this study, protein precipitation by different water-soluble organic solvents was evaluated through comparison with MTBE based lipid extraction to determine which would be the best method for LC/MS based lipidomics analysis.

2. 2.1.

Materials and methods Nomenclature for lipids

For acylglycerol and phospholipids, one “acyl” group is denoted by “a” and one “ether” group is denoted by “e.” For example, PC(aa-34:3) is a phosphatidylcholine consists of two fatty acyl groups whose total number of carbons and double bonds in the fatty acids is 34 and 3, respectively. All sphingolipids were identified as having the C18 sphingoid base structure, and the number of carbons and double bonds in the N-acyl group is denoted. For example, Cer(24:1) is a ceramide composed of C18 sphingosine and an N-acyl derivative of the C24:1 fatty acid.

2.2.

Chemicals and reagents

Acetonitorile (LC/MS grade), methanol (LC/MS grade), ethanol (99.5%, HPLC grade), 2-propanol (LC/MS grade), and the 2 mol/L ammonia isopropanol solution, and 0.2 M disodium dihydrogen ethylenediaminetetraacetic acid (EDTA-2Na) solution were purchased from Wako Pure Chemical (Osaka, Japan). Mouse plasma and liver were purchesed from Cosmo Bio Co., Ltd. (Tokyo, Japan). LPC(a-18:1), LPE(a-18:1), LPA(a-18:1), LPG(a-18:1), LPS(a-18:1), 5

LPI(a-18:1), PC(a-18:1/a-18:1), PE(a-18:1/a-18:1), PA(a-18:1/a-18:1), PG(a-18:1, a-18:1), PS(a-18:1, a-18:1), PI(a-18:1/a-18:1), Cer(24:0), SM(18:1), and sphinganine were purchased from Avanti Polar Lipids (Alabaster, AL, USA) Glycocholic acid and taurocholic acid were purchased from Sigma-Aldrich (St. Louis, MO, USA). CE(a-18:0) and DG(a-18:/a-18:0) were purchased from Wako Pure Chemical (Osaka, Japan). MG(a-16:0), TG(a-18:1/a-18:1/a-18:1) and Cholic acid were purchased from Tokyo Chemical Industry (Tokyo, Japan). Acylcarnitine (C18:0) was purchased from Larodan Fine Chemicals (Malmö, Sweden). FFA(C18:0) was purchased from MP biochemicals (Santa Ana, CA, USA).

2.3.

Sample preparation

Mouse plasma (20 μL) was mixed with 180 μL of a water-soluble organic solvent (IPA, EtOH and MeOH) in an ice-cold 1.5 mL tube then vortexed for 1 min. After centrifugation at 15,000 rpm and 4 °C for 5 minutes, the supernatant was used as the sample solution and used for liquid chromatography/mass spectrometry (LC/MS) analysis. For the LLE method, a modified MTBE-MeOH (mMM) method was applied. Mouse plasma (20 μL) was mixed with 150 μL of MeOH and 500 μL of MTBE in an ice-cold glass vial then vortexed for 1 min. After a 1 h incubation, 125 μL of distilled water was added and centrifuged at 15,000 rpm and 4 °C for 5 minutes. The organic layer (upper phase) was collected and transferred to an ice-cold 1.5 mL tube. The lower phase was mixed with 100 μL of MTBE-MeOH-water (10:3:3). After centrifugation, the organic layer (upper phase) was transferred to the ice-cold 1.5 mL tube and dried under a dried nitrogen stream to complete dryness. The dried sample was reconstituted 200 μL of EtOH, then introduced for liquid chromatography/mass spectrometry analysis. Mouse liver was homogenized with 9 times the volume (v/w) of an ice-cold water-soluble

6

organic solvent (ACN and EtOH) using a ball-mill, and centrifuged at 15,000 rpm and 4 °C for 5 minutes. The supernatant was applied to LC/MS analysis.

2.4.

Liquid chromatography/mass spectrometry

The liquid chromatography/mass spectrometry system consisted of an Ultimate3000 RSLC system (Thermo Fisher Scientific inc., Sunnyvale, CA, USA) and a LTQ Orbitrap XL mass spectrometer (Thermo Fisher Scientific inc., Sun Jose, CA, USA). Liquid chromatography separation was performed at 60 °C using a reverse phase column, XBridge BEH C18 (2.1 × 30 mm, 2.5 μm, 130 Å, Waters co., Milford, MA, USA), at a flow rate of 0.5 mL/min. The mobile phases consisted of water containing 0.01% acetic acid, 1 mM ammonia, and 10 μM EDTA-2Na (mobile phase A), and a mixture of ethanol and isopropanol (1:1) containing 0.001% acetic acid and 0.2 mM ammonia (mobile phase B). The EDTA in the mobile phase improved chromatographic separations of the phosphor group containing lipids, such as LPA [34]. A gradient elution was used for the lipid separation by the following program. The ratio of the mobile phase B was started at 2% B for 1 min, increased to 100% B at 8 min, maintained at 100% B for 2 min, then decreased to 2% B and kept constant for 2 min. Mass spectrometry analysis was performed separately in both positive and negative ionization modes. The eluent from liquid chromatography was introduced directly for electrospray ionization using a heated electrospray ionization probe (H-ESI2, Thermo Fisher Scientific inc., San Jose, CA, USA) with a spray voltage at 4 kV for negative or 5 kV for positive ionization mode, and a vaporizer temperature at 250 °C. A capillary temperature was set at 300C with a capillary voltage at －30 V for negative or 30 V for positive ionization mode. Full mass spectra (MS) were acquired by an Orbitrap ranging from m/z 200 to m/z 2000 with a mass resolving power of 60000 FWHM (Full width at half maximum) at m/z 400. The 7

top 5 highest signals in the full mass spectra were sequentially introduced to a LTQ ion trap cell to obtain product ion spectra (MS/MS) by collision activated dissociation (CAD).

2.5.

Peak extraction from LC/MS raw data

The raw data from the LC/MS were processed by Expressionist Refiner MS software (ver.8.2, Genedata AG, Basel, Switzerland). First, the raw data were aligned to the Grid activity using the adaptive grid method with a scan count of 3. Then, background chemical noise was removed using the Chemical Noise Subtraction activity by setting the parameters to 999 of the RT window with a 90% quantile. The chemical noise that randomly appeared in the data were removed by the RT structure removal method and the m/z structure removal method by setting 4 scans for both the minimum RT and m/z length. The retention time of all LC/MS data were aligned by the RT alignment method using the pairwise alignment based tree method with parameters of 50 scans for the RT search interval, 3 mDa for the m/z window, and 0.1 min for the RT window. Afterwards, peaks were detected by the Peak Detection activity using the curved-based peak detection method with parameters of 80% for the refinement threshold, 2 for the consistency threshold, 3 mDa for the smoothing window, 0.1 min for the summation window, 0.05 min for the minimum peak size, 5 mDa for the maximum merge distance, and 1% for the gap/peak ratio of the maximum intensity profiling. The detected peaks were merged into isotopic ion clusters of individual peaks by the Isotope Clustering activity using parameters of 0.02 min for the RT tolerance and 5 mDa for the m/z tolerance. The peak intensities of the isotope clusters were exported as a tab-separated file along with the monoisotopic m/z value and retention time information, and further data processing was performed using Excel.

2.6.

Lipid identification 8

Standard reagents can be used for structure determination of observed lipids based on the consistency of the retention time and m/z values. For other lipids, m/z values observed in the Orbitrap mass spectrometer were accurate enough to speculate the identity of the structure from the lipid database. Therefore, additional product ion information (MS/MS spectrum) can be a determiner for structure identification. For example, the constituent fatty acid can be identified by fatty acid anion fragments in a MS/MS spectrum of a negative ion. Also, neutral losses of fatty acids can be used as evidence that indicates the fatty acid composition, an important signature for identification of acylglycerols. Lipid class specific fragment ions were also used as the criteria [35][22]. PC, LPC, and SM generate the characteristic product ion, m/z 184, indicating existence of a phosphocholine group. PS and LPS generates neutral losses of serine (-87) or phosphoserine (-185) from the [M-H]- and [M+H]+ ions, respectively. Cholesteryl ester generates the characteristic product ion, m/z 369, from the [M+NH4]+ ion, which indicates the existence of a cholesterol moiety. MS/MS spectra of glycosphingolipids show characteristic ions with sequential degradation of the sugar chain and oxonium ions of the constituent sugars. For lipids without standard reagents and MS/MS spectra, the following criteria were used for structure determination. Ion species profile can be a characteristic indicator for estimating the lipid class, especially when both a positive and negative ion pattern were match those of a predicted structure obtained from the lipid database [22]. Another criterion was alignment of retention times in a same lipid class. The retention time interval in a same lipid class exhibits a consistent pattern based on the constituent fatty acid structure [22,24,35]. An increment number of carbons in the fatty acid almost constantly delays the retention time. On the other hand, the retention time becomes faster by increasing the number of double bonds in the fatty acid. Therefore, retention time can be predicted in a same group of lipids using information about the fatty acid composition. 9

3.

Results and Discussion

3.1.

Enhancement of lipase activity by acetonitrile extraction

Among water-soluble organic solvents, ACN, MeOH, EtOH, and IPA can be used as protein precipitation reagents; however, ACN was found to not be a suitable reagent for lipid extraction because it potentially activated some enzymes during the protein precipitation process. In Figure 1, lipid extraction by ACN from mouse liver exhibited significantly higher intensities of FFA, MG, and DG than those of EtOH extraction. Although a high extraction efficiency of ACN was expected, co-extraction with a lipase inhibitor, Orlistat, decreased the peak intensities depending on the dose, approaching the lipid profile of EtOH extraction. In contrast, the TG level from ACN extraction was increased by co-extraction with Orlistat in a dose dependent manner, suggesting that ACN extraction decomposed TG in mouse liver during homogenization through activation of lipases, potentially hepatic lipase. Additionally, Orlistat inhibited the enzymatic activity. ACN has been believed to activate enzymatic activity in the proteomics field. ACN solutions are known to enhance trypsin activity, improving the coverage of tryptic peptides [36]. The mechanism could be the same as enhancement of lipase activity during lipid extraction. This is presumably because of the high dipole moment of ACN, which might weaken the protein denaturing capabilities. With remaining or weakly loosened conformation of proteins, and with exposed substrates in the tissue homogenate, some enzymatic reactions are temporarily activated. Therefore, in this study ACN was excluded as a lipid extraction reagent.

3.2.

Comparison of neutral lipid extraction

10

Protein precipitation by different types of alcohols, such as IPA, EtOH, and MeOH, was evaluated as a sample preparation method for LC/MS-based lipidomics analysis. These methods were compared to a modified MTBE-MeOH (mMM) extraction; a lipid fraction enriched in the MTBE-MeOH layer was dried and reconstituted in EtOH for the sample to be injectable into reverse-phase LC/MS system. Lipids were extracted from mouse plasma by these four methods, then LC/MS analysis was performed using a LTQ-Orbitrap mass spectrometer in both positive and negative ionization modes. Overall, chromatographic patterns were all similar except for when MeOH was used (Supplementary Figure 1). Drastic differences were found in peaks around 11–11.5 min in positive ionization mode (Supplementary Figure 1A, C, E, G, I). The major components of these peaks were TG and CE. MeOH extraction exhibited weaker intensities for these molecules, consistent with the weaker hydrophobicity of MeOH compared to the other solvents. Figure 2A shows differences in peak area of the TG components extracted by the five methods, where the peak area of each molecule was normalized by the average peak area from IPA extraction. All TGs extracted by MeOH were lower than in other methods. Notably, TGs with saturated fatty acids were insoluble in the MeOH extraction; however, extraction efficiencies were improved as an increment of the number of double bonds in TG in the MeOH extraction. Since double bonds in fatty acids decrease the hydrophobicity, polyunsaturated TGs were thought to be rather amenable to MeOH extraction. In total, IPA, EtOH, and the mMM methods were considered to be the best approach for TG extraction. A similar trend was observed for CE (Supplementary Figure 2B), consistent with the hydrophobicity of TG and CE, whose retention times are almost same. On the other hand, DG and MG can be extracted by MeOH with almost equal or better extraction efficiencies relative to IPA, EtOH, and mMM method (Figure 2B, Supplementary Figure 2A). The hydrophobicity of DG and MG is weaker than that of TG due to the number of fatty acyl chains and the exposed 11

hydroxy group. In addition, the mMM method was shown to be not the best approach for neutral lipid extraction, especially for MG. This result is because the reconstitution process by EtOH determines the solubility and MTBE is too hydrophobic to completely extract MG. In total, lipid extraction by alcohols, especially EtOH, was thought to be an appropriate method for LC/MS based lipidomics analysis.

3.3.

Comparison of phospholipid extraction

PC was extracted by alcohol with almost equivalent efficiency or slightly better than the mMM method, irrespective of the fatty acid structure (Figure 3A). The same trend was observed in PE and PI (Supplementary Figure 3). LPC showed almost the same trend as PC, but the extraction efficiency of the mMM method was worsened compared to PC (Figure 3B). Moreover, with regards to lysophospholipids, the head group had a significant impact on the solubility. In fact, large differences within the extraction methods were found in LPI; MeOH was the best reagent, EtOH was the second best, and mMM was the worst method (Figure 3C). Lysophospholipids are a group of phospholipids with a single fatty acyl chain, which weakens the hydrophobicity and increases the weight of the head group with respect to solubility of the reagent, as shown in the result of LPI. The chemical character of the head group determined the degree of solubility differences (Supplementary Figure 4); however, the common finding was that MeOH was the best reagent for lysophospholipid extraction. Taken together, alcohol extraction is sufficient, and better than the mMM method, for phospholipid extraction. Specifically, MeOH exhibited the best extraction efficiency for both phospho and lysophospholipids.

3.4.

Comparison of ganglioside extraction 12

As we discussed previously, alcohol extraction demonstrated equal or better efficiencies for neutral lipids and phospholipids. The same was true for Cer, SM, and Hex-Cer, whose extraction efficiencies were almost the same as those of the phospholipids (Supplementary Figure 5). In addition, GM2 and GM3 gangliosides were also extracted with almost equal extraction efficiencies by all alcohols, while the mMM method exhibited around half the peak intensities of alcohol extraction (Figure 4A, and Supplementary Figure 6). On the other hand, GD1 ganglioside went together with MeOH and EtOH, showing a drastically higher extraction capability than those of IPA and mMM extraction (Figure 4B). GD1 ganglioside has two sialic acid groups, which make the molecular character hydrophilic and potentiate the affinity to MeOH and EtOH, whereas GM2 and GM3, which are composed of a single sialic acid, do not contribute to the extraction differences.

3.5.

Extraction of hydrophilic lipids

The impact of the molecular character on extraction efficiency was clearly observed in acylcarnitine extraction. Extraction efficiencies of acylcarnitine varied in the order of hydrophilicities of the molecules and extraction reagents. Acylcarnitine (C5:0) is one of the most hydrophilic molecules, which consist of a quaternary amine and carboxylic acid beside a fatty acyl group that, in this analysis, eluted at 2.5 min in the LC/MS/MS chromatogram. While the mMM method hardly detected acylcarnitine (C5:0), MeOH, and EtOH extraction observed the molecule, showing 10 and 5 times higher intensity than that of IPA, respectively. The extraction efficiencies were gradually approaching the same level of all methods as the number of carbons in the fatty acyl chain increased, and acylcarnitine (C18:0) was extracted at almost an equal level in all alcohols; however, the mMM method did not reached the same level as the alcohol extraction method. Considering the above results, the mMM method was inclined to be 13

nearly insoluble to lipids that are composed of hydrophilic moieties. This trend was clearly observed in bile acids, which were extracted by all alcohols but not by the mMM method (Figure 6A). The same was true for sphingosine-1-phosphate and sphinganine-1-phosphate for the same reason, and MeOH extraction showed the highest intensities. Basically, alcohol extraction was good enough for almost all lipid species, including cholesterol, cholesterol sulfate, sphingosine, sphinganine, and corticosterone, and the mMM method was not the best approach likely due to the indigenous extraction capability and uneven reconstitution efficiency, increasing analytical variability. In fact, CV values from the mMM method were relatively larger than those of other methods (Supplementary Figure 8).

4.

Conclusion In summary, alcohol extraction was revealed to be a feasible approach for lipidomics

analysis. In contrast, the mMM method gave a relatively negative impression in this study for recovery and reproducibility for all lipid spices, likely due to low recovery during the reconstitution process. Moreover, for hydrophilic lipids, such as short-chain acylcarnitines or bile acids, the recovery could be improved by MeOH reconstitution. One possible drawback of the alcohol extraction method is the complexity of the extract, which includes abundant non-lipid components. Therefore, alcohol extraction is likely not suitable for shotgun lipidomics analysis, causing a severe matrix effect that prevents sensitive and selective analysis; however, LC/MS based lipidomics analysis enabled wide coverage lipid analysis with high reproducibility, partially because liquid chromatography separation circumvents the influence of the additional components from alcohol extraction. Alcohol extraction is a simple and quick sample preparation method for lipidomics analysis, but differences of the extraction efficiencies between the alcohols exists. MeOH extraction has a 14

preference for hydrophilic lipids, whereas IPA extraction is inclined to hydrophobic lipids, such as TG. Actually, the water content could influence extraction efficiency. In this study, a 90% alcohol solution was used to ensure the protein denaturation, but an 80% alcohol solution could alter the extraction efficiencies (data not shown). Past report also demonstrated that protein precipitation by IPA was an excellent sample preparation method for lipidomics analysis [37], in which the reuslt of methanol precipitation was somewhat worse than the result of this study, presumably because the alcohol ratio during protein precipitation was 75%. Thus, a sample preparation protocol should be carefully optimized based on needs of the target molecular species. Overall, alcohol extraction can be used for LC/MS based lipidomics analysis with high coverage and high reproducibility. In this study, ACN was proved to not be suitable as a protein denaturing reagent since the denaturing capability is too weak to stably extract lipid species. One question was whether MeOH extraction caused same effect as ACN. In our study, there were no critical differences between EtOH and MeOH extraction except for TG and CE. Moreover, MeOH homogenization was traditionally applied to the first step of lipid extraction for the Bligh-Dyer and Folch methods, and even for MTBE-MeOH extraction. We currently concluded that MeOH extraction is a suitable lipid extraction method, especially for hydrophilic lipid species; however, MeOH cannot cover very hydrophobic lipids. Therefore, we propose EtOH extraction as a superior lipid extraction approach for LC/MS based lipidomics analysis, covering wide range of lipid species from short-chain acylcarnitine to triacylglycerols.

Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. 15

Acknowledgement We greatly acknowledge Ayako Okamoto for technical assistance of this work.

16

References [1]

F.B. Cordeiro, T.R. Cataldi, K.J. Perkel, L. do Vale Teixeira da Costa, R.C. Rochetti, J. Stevanato, et al., Lipidomics analysis of follicular fluid by ESI-MS reveals potential biomarkers for ovarian endometriosis, Journal of Assisted Reproduction and Genetics. (2015). doi:10.1007/s10815-015-0592-1.

[2]

Y.-Y. Zhao, H. Miao, X.-L. Cheng, F. Wei, Lipidomics: Novel insight into the biochemical mechanism of lipid metabolism and dysregulation-associated disease, Chemico-Biological Interactions. 240 (2015) 220–238. doi:10.1016/j.cbi.2015.09.005.

[3]

X. Han, S. Rozen, S.H. Boyle, C. Hellegers, H. Cheng, J.R. Burke, et al., Metabolomics in Early Alzheimer’s Disease: Identification of Altered Plasma Sphingolipidome Using Shotgun Lipidomics, PLoS ONE. 6 (2011) e21643. doi:10.1371/journal.pone.0021643.

[4]

P. Proitsi, M. Kim, L. Whiley, M. Pritchard, R. Leung, H. Soininen, et al., Plasma lipidomics analysis finds long chain cholesteryl esters to be associated with Alzheimer’s disease., Translational Psychiatry. 5 (2015) e494. doi:10.1038/tp.2014.127.

[5]

M. Mapstone, A.K. Cheema, M.S. Fiandaca, X. Zhong, T.R. Mhyre, L.H. MacArthur, et al., Plasma phospholipids identify antecedent memory impairment in older adults., Nature Medicine. 20 (2014) 415–8. doi:10.1038/nm.3466.

[6]

Y. Tajima, M. Ishikawa, K. Maekawa, M. Murayama, Y. Senoo, T. Nishimaki-Mogami, et al., Lipidomic analysis of brain tissues and plasma in a mouse model expressing mutated human amyloid precursor protein/tau for Alzheimer’s disease., Lipids in Health and Disease. 12 (2013) 68. doi:10.1186/1476-511X-12-68.

17

[7]

J. Ji, A.E. Kline, A. Amoscato, A.K. Samhan-Arias, L.J. Sparvero, V. a Tyurin, et al., Lipidomics identifies cardiolipin oxidation as a mitochondrial target for redox therapy of brain injury, Nature Neuroscience. 15 (2012) 1407–1413. doi:10.1038/nn.3195.

[8]

R. Puri, M. Duong, K. Uno, Y. Kataoka, S.J. Nicholls, The emerging role of plasma lipidomics in cardiovascular drug discovery., Expert Opinion on Drug Discovery. 7 (2012) 63–72. doi:10.1517/17460441.2012.644041.

[9]

M. Orešič, T. Seppänen-Laakso, D. Sun, J. Tang, S. Therman, R. Viehman, et al., Phospholipids and insulin resistance in psychosis: a lipidomics study of twin pairs discordant for schizophrenia, Genome Medicine. 4 (2012) 1. doi:10.1186/gm300.

[10]

D.L. Gorden, P.T. Ivanova, D.S. Myers, J.O. McIntyre, M.N. VanSaun, J.K. Wright, et al., Increased diacylglycerols characterize hepatic lipid changes in progression of human nonalcoholic fatty liver disease; comparison to a murine model., PloS One. 6 (2011) e22775. doi:10.1371/journal.pone.0022775.

[11]

K. Yang, H. Cheng, R.W. Gross, X. Han, Automated lipid identification and quantification by multidimensional mass spectrometry-based shotgun lipidomics, Analytical Chemistry. 81 (2009) 4356–4368. doi:10.1021/ac900241u.

[12]

M.A. Surma, R. Herzog, A. Vasilj, C. Klose, N. Christinat, D. Morin-Rivron, et al., An automated shotgun lipidomics platform for high throughput, comprehensive, and quantitative analysis of blood plasma intact lipids, European Journal of Lipid Science and Technology. (2015) n/a-n/a. doi:10.1002/ejlt.201500145.

[13]

X. Han, R.W. Gross, Shotgun lipidomics: Electrospray ionization mass spectrometric analysis and quantitation of cellular lipidomes directly from crude extracts of biological samples, Mass Spectrometry Reviews. 24 (2005) 367–412. doi:10.1002/mas.20023. 18

[14]

K. Schuhmann, R. Herzog, D. Schwudke, W. Metelmann-Strupat, S.R. Bornstein, A. Shevchenko, Bottom-up shotgun lipidomics by higher energy collisional dissociation on LTQ orbitrap mass spectrometers, Analytical Chemistry. 83 (2011) 5480–5487. doi:10.1021/ac102505f.

[15]

B. Simons, D. Kauhanen, T. Sylvänne, K. Tarasov, E. Duchoslav, K. Ekroos, Shotgun Lipidomics by Sequential Precursor Ion Fragmentation on a Hybrid Quadrupole Time-of-Flight Mass Spectrometer, Metabolites. 2 (2012) 195–213. doi:10.3390/metabo2010195.

[16]

J. a. Bowden, J.T. Bangma, J.R. Kucklick, Development of an Automated Multi-Injection Shotgun Lipidomics Approach Using a Triple Quadrupole Mass Spectrometer, Lipids. 49 (2014) 609–619. doi:10.1007/s11745-014-3903-x.

[17]

M. Wang, Y. Huang, X. Han, Accurate mass searching of individual lipid species candidates from high-resolution mass spectra for shotgun lipidomics, Rapid Communications in Mass Spectrometry. 28 (2014) 2201–2210. doi:10.1002/rcm.7015.

[18]

M. Ståhlman, C.S. Ejsing, K. Tarasov, J. Perman, J. Borén, K. Ekroos, High-throughput shotgun lipidomics by quadrupole time-of-flight mass spectrometry, Journal of Chromatography B. 877 (2009) 2664–2672. doi:10.1016/j.jchromb.2009.02.037.

[19]

K. Schuhmann, R. Almeida, M. Baumert, R. Herzog, S.R. Bornstein, A. Shevchenko, Shotgun lipidomics on a LTQ Orbitrap mass spectrometer by successive switching between acquisition polarity modes., Journal of Mass Spectrometry : JMS. 47 (2012) 96– 104. doi:10.1002/jms.2031.

[20]

X. Gao, Q. Zhang, D. Meng, G. Isaac, R. Zhao, T.L. Fillmore, et al., A reversed-phase capillary ultra-performance liquid chromatography-mass spectrometry (UPLC-MS) 19

method for comprehensive top-down/bottom-up lipid profiling., Analytical and Bioanalytical Chemistry. 402 (2012) 2923–33. doi:10.1007/s00216-012-5773-5. [21]

K. Sandra, A.D.S. Pereira, G. Vanhoenacker, F. David, P. Sandra, Comprehensive blood plasma lipidomics by liquid chromatography/quadrupole time-of-flight mass spectrometry, Journal of Chromatography A. 1217 (2010) 4087–4099. doi:10.1016/j.chroma.2010.02.039.

[22]

J. Hummel, S. Segu, Y. Li, S. Irgang, J. Jueppner, P. Giavalisco, Ultra Performance Liquid Chromatography and High Resolution Mass Spectrometry for the Analysis of Plant Lipids, Frontiers in Plant Science. 2 (2011) 1–17. doi:10.3389/fpls.2011.00054.

[23]

J. Kim, C.L. Hoppel, Comprehensive approach to the quantitative analysis of mitochondrial phospholipids by HPLC-MS., Journal of Chromatography. B, Analytical Technologies in the Biomedical and Life Sciences. 912 (2013) 105–14. doi:10.1016/j.jchromb.2012.10.036.

[24]

J.M. Choi, T.-E. Kim, J.-Y. Cho, H.J. Lee, B.H. Jung, Development of lipidomic platform and phosphatidylcholine retention time index for lipid profiling of rosuvastatin treated human plasma., Journal of Chromatography. B, Analytical Technologies in the Biomedical and Life Sciences. 944 (2014) 157–65. doi:10.1016/j.jchromb.2013.10.029.

[25]

C. Buré, S. Ayciriex, E. Testet, J.-M. Schmitter, A single run LC-MS/MS method for phospholipidomics, Analytical and Bioanalytical Chemistry. 405 (2013) 203–213. doi:10.1007/s00216-012-6466-9.

[26]

S. Becker, L. Kortz, C. Helmschrodt, J. Thiery, U. Ceglarek, LC-MS-based metabolomics in the clinical laboratory., Journal of Chromatography. B, Analytical

20

Technologies in the Biomedical and Life Sciences. 883–884 (2012) 68–75. doi:10.1016/j.jchromb.2011.10.018. [27]

Y.-Y. Zhao, S.-P. Wu, S. Liu, Y. Zhang, R.-C. Lin, Ultra-performance liquid chromatography–mass spectrometry as a sensitive and powerful technology in lipidomic applications, Chemico-Biological Interactions. 220 (2014) 181–192. doi:10.1016/j.cbi.2014.06.029.

[28]

D.W. Bligh EG, A rapid method of total lipid extraction and purification, Canadian Journal of Biochemistry and Physiology. 37 (1959) 911–917.

[29]

S.S.G. Folch J, Lees M, A simple method for the isolation and purificatin of total lipids from animal tissues, Journal of Biological Chemistry. 226 (1957) 497–509.

[30]

S.K. Abbott, A.M. Jenner, T.W. Mitchell, S.H.J. Brown, G.M. Halliday, B. Garner, An Improved High-Throughput Lipid Extraction Method for the Analysis of Human Brain Lipids, Lipids. 48 (2013) 307–318. doi:10.1007/s11745-013-3760-z.

[31]

V. Matyash, G. Liebisch, T. V. Kurzchalia, a. Shevchenko, D. Schwudke, Lipid extraction by methyl-tert-butyl ether for high-throughput lipidomics, The Journal of Lipid Research. 49 (2008) 1137–1146. doi:10.1194/jlr.D700041-JLR200.

[32]

L. Löfgren, M. Ståhlman, G.-B. Forsberg, S. Saarinen, R. Nilsson, G.I. Hansson, The BUME method: a novel automated chloroform-free 96-well total lipid extraction method for blood plasma., Journal of Lipid Research. 53 (2012) 1690–700. doi:10.1194/jlr.D023036.

[33]

D.Y. Lee, T. Kind, Y.-R. Yoon, O. Fiehn, K.-H. Liu, Comparative evaluation of extraction methods for simultaneous mass-spectrometric analysis of complex lipids and

21

primary metabolites from human blood plasma, Analytical and Bioanalytical Chemistry. 406 (2014) 7275–7286. doi:10.1007/s00216-014-8124-x. [34]

Y. Sato, T. Nakamura, K. Aoshima, Y. Oda, Quantitative and wide-ranging profiling of phospholipids in human plasma by two-dimensional liquid chromatography/mass spectrometry., Analytical Chemistry. 82 (2010) 9858–64. doi:10.1021/ac102211r.

[35]

T. Yamada, T. Uchikata, S. Sakamoto, Y. Yokoi, E. Fukusaki, T. Bamba, Development of a lipid profiling system using reverse-phase liquid chromatography coupled to high-resolution mass spectrometry with rapid polarity switching and an automated lipid identification software, Journal of Chromatography A. 1292 (2013) 211–218. doi:10.1016/j.chroma.2013.01.078.

[36]

M.B. Strader, D.L. Tabb, W.J. Hervey, C. Pan, G.B. Hurst, Efficient and specific trypsin digestion of microgram to nanogram quantities of proteins in organic-aqueous solvent systems, Analytical Chemistry. 78 (2006) 125–134. doi:10.1021/ac051348l.

[37]

M.H. Sarafian, M. Gaudin, M.R. Lewis, F.-P. Martin, E. Holmes, J.K. Nicholson, et al., Objective Set of Criteria for Optimization of Sample Preparation Procedures for Ultra-High Throughput Untargeted Blood Plasma Lipid Profiling by Ultra Performance Liquid Chromatography−Mass Spectrometry, (2014). doi:10.1021/ac500317c.

22

Figure 1. Differences in EtOH (white bar), ACN (light gray bar), ACN + 5 μM Orlistat (dark gray bar), and ACN + 50 μM Orlistat (black bar) extractions on (A) FFA, (B) MG, (C) DG, and (D) TG from mouse liver. Peak area was normalized to the average peak area of EtOH extraction. Bar graphs are displayed with mean  standard deviations. Statistical significances between ACN and EtOH were evaluated by Welch’s t-test (* p