Higher Resolution LC-MS and MS-MS Analysis of Lipid Extracts Using Benchtop Orbitrap-based Mass Spectrometers and LipidSearch Software David A Peake, 1 Yasuto Yokoi, 2 Reiko Kiyonami 1 and Yingying Huang 1 1 Thermo Fisher Scientific, San Jose, USA; 2 Mitsui Knowledge Industry, Tokyo, JAPAN
Overview Purpose: Benchmark the performance of a new LC-MS/MS system and determine the effect of full scan MS resolution vs. the number of identified lipids in total lipid extracts. Methods: Lipid extracts were profiled using LC-dd-MS/MS analysis using a novel quadrupole-orbitrap mass spectrometer operated at mass resolution of 3, 6, 12 and 24K. Results: The data presented here shows that lipidomic profiling at higher full scan MS resolution provides more identified lipids during an LC-MS run. The higher acquisition rate of MS/MS spectra also contributes to significantly more lipids being identified. Introduction Lipids play a key role in cell, tissue and organ physiology with diseases such as cancer and diabetes which involve disruption of their metabolic enzymes and pathways. Identification of unique lipid biomarkers to distinguish healthy humans compared to those with a disease can have an impact on the early detection of diseases and personalized medicine. Because of the complexity of the lipidome, which includes 8 major categories of lipids, over 8 major classes, 3 sub-classes and thousands of lipid species 1, HPLC MS/MS methods are often used to separate many overlapping isomeric or isobaric molecular ions in biological samples for increased lipid identification coverage. By employing a high resolution accurate mass (HR/AM) platform equipped with a ultra-high field Orbitrap mass spectrometer (Thermo Scientific TM Q Exactive HF TM MS) to detect lipids after HPLC separation, ultra-high resolution MS and MS/MS analysis provides unambiguous identification of lipids in biological samples. However, it is extremely challenging to process all the MS and MS/MS data manually. A sophisticated software with an extensive database is required to identify all detected lipids HPLC MS/MS data. The newly-released Thermo Scientific TM LipidSearch TM software enables automated identification of lipids from biological samples based on its large lipid database which includes >1.5 million lipid ions and predicted fragment ions. In addition, relative quantification of identified lipid precursors can be carried out in the same LC/MS/MS experiment. In this study, we present that hundreds of lipid species can be simultaneously identified and quantified in a single LC/MS-MS experiment by using optimized HPLC separation and HR/AM MS and data-dependent MS 2 conditions. Methods LC-MS Sample Preparation. Bovine brain, heart, liver and yeast total lipid extracts (2.5 mg/ml in Chloroform) were purchased from Avanti Polar Lipids. Dilution series of each lipid extract was prepared by diluting the stock solutions sequentially into 1.25 µg/µl, 5 ng/µl, 25 ng/µl, 125 ng/µl and 5 ng/µl in 5:5 Methanol and Isopropanol (IPA). HPLC Method. A Thermo Scientific Dionex UltiMate 3 Rapid Separation LC (RSLC) system performed separations using the gradient conditions shown in Table 1 2. Mobile phase A was 6:4 Acetonitrile / Water and mobile phase B was 9:1 IPA / Acetonitrile; both A and B contained 1mM ammonium formate and.1% formic acid. The column was an Ascentis Express C18 (Supelco, 2.1 x 1mm, 2.7µm) operated at 55 C, flow rate of 26 µl/min and the injection volume was 2 µl. MS Conditions. Thermo Scientific Q Exactive Plus and Q Exactive HF instruments were employed for untargeted lipid profiling experiments using the instrument operating conditions shown in Table 2. Each instrument was operated under optimized conditions providing sufficient scans across the chromatographic peak profile for accurate relative quantification using the HR/AM precursor ion while simultaneously acquiring dd-ms 2 spectra for lipid identification. Data Analysis Software. LipidSearch software was used for lipid identification and relative quantification. TABLE 1. HPLC Gradient Time, min % A % B. 68 32 1.5 68 32 4. 55 45 5. 48 52 8. 42 58 11. 34 66 14. 3 7 18. 25 75 21. 3 97 25. 3 97 25.1 68 32 33. 68 32 Data Processin LC-MS/MS Data Processing 1) Peak Detection. Read raw 2) Identification. Candidate m database >1 6 entries of accu each potential lipid structure a 3) Alignment. The search res window and the results are co 4) Quantification. The accura each identified lipid precursor 5) Statistical Analysis. t-test sample vs. control groups, and FIGURE 1. LipidSearch Work HRAM / Nomin Infusion & LC-M ID based on M DB with 1 6 + li Integration of m KO / WT =.3 p <.53 KO W 2 Highly Sensitive, Robust MS-Based Workflow for Therapeutic Monoclonal Antibody Analysis from Complex Matrices
/MS system and determine the fied lipids in total lipid extracts. S analysis using a novel ass resolution of 3, 6, 12 profiling at higher full scan MS S run. The higher acquisition ore lipids being identified. y with diseases such as cancer enzymes and pathways. althy humans compared to tection of diseases and s 8 major categories of lipids, of lipid species 1, HPLC MS/MS isomeric or isobaric molecular n coverage. By employing a ed with a ultra-high field ctive HF TM MS) to detect lipids /MS analysis provides. S and MS/MS data manually. equired to identify all detected cientific TM LipidSearch TM biological samples based on its and predicted fragment ions. rsors can be carried out in the an be simultaneously nt by using optimized HPLC nditions. TABLE 1. HPLC Gradient Time, min % A % B. 68 32 1.5 68 32 4. 55 45 5. 48 52 8. 42 58 11. 34 66 14. 3 7 18. 25 75 21. 3 97 25. 3 97 25.1 68 32 33. 68 32 Data Processing LC-MS/MS Data Processing Workflow using LipidSearch Software (Figure 1). 1) Peak Detection. Read raw files, MS n and precursor ion accurate masses. 2) Identification. Candidate molecular species are identified by searching a large database >1 6 entries of accurate m/z (lipid precursor and fragment ions) predicted from each potential lipid structure and positive/negative ion adduct. 3) Alignment. The search results for each individual sample are aligned within a time window and the results are combined into a single report. 4) Quantification. The accurate-mass extracted ion chromatograms are integrated for each identified lipid precursor and the peak areas are obtained. 5) Statistical Analysis. t-tests determine significantly differences between lipids in sample vs. control groups, and results are displayed in a whisker plot. FIGURE 1. LipidSearch Workflow. TABLE 2. Orbitrap Operating Conditions HESI Source Q Exactive Plus Q Exactive HF Sheath gas 35 Pos. or Neg. Ion Aux gas 3 Spray volt. 4.2 kv S-Lens 5 Cap. Temp. 32 C Heater temp. 3 C 17.5K, 35K, 7K, 14K Top15 dd-ms 2 R = 35K Pos. 25, 3 Neg. 2, 24, 28 Pos. or Neg. Ion 3K, 6K, 12K, 24K Top2 dd-ms 2 R = 3K Pos. 25, 3 Neg. 2, 24, 28 Results Lipid Identification Results from (1.25 µg/µl) was analyzed by LC/M modes, respectively (Figure 2). Lipi data for all common lipid classes us product ions. For each MS 2 spectru the lipid precursor ion m/z stored in summarized for the identified lipid s experimental data (Figure 3). The speed advantage of Q Exactive illustrated in Figure 4, compared to higher resolution (12K) the Q Exac FIGURE 2. High Resolution (7, Total ion Chromatograms of Bov Positive Ion FIGURE 3. Search Results for m/z.5 mg/ml in Chloroform) were ch lipid extract was prepared L, 5 ng/µl, 25 ng/µl, 125 (IPA). Peak Detect 329.2483 283.264 PC(18:_22:5) Separation LC (RSLC) system wn in Table 1 2. Mobile phase A 9:1 IPA / Acetonitrile; both A rmic acid. The column was an perated at 55 C, flow rate of struments were employed for nt operating conditions shown ed conditions providing for accurate relative ltaneously acquiring dd-ms 2 HRAM / Nominal mass Infusion & LC-MS data ID based on MS n DB with 1 6 + lipid species Integration of multiple runs WT Identify Align 58.347 82.5 FIGURE 4. Increase in the Number Lipid Extract using dd-lcms 2 with 5 QE Plus (top 4 QE HF (top 2 3 d relative quantification. KO / WT =.36 p <.53 KO Quan 2 1 1ng 25ng Thermo Scientific Poster Note PN-64137-ASMS-EN-614S 3
rbitrap Operating Conditions Q Exactive Plus Pos. or Neg. Ion 17.5K, 35K, 7K, 14K Top15 dd-ms 2 R = 35K Pos. 25, 3 Neg. 2, 24, 28 Q Exactive HF Pos. or Neg. Ion 3K, 6K, 12K, 24K Top2 dd-ms 2 R = 3K Pos. 25, 3 Neg. 2, 24, 28 LipidSearch Software (Figure 1). cursor ion accurate masses. are identified by searching a large cursor and fragment ions) predicted from ve ion adduct. idual sample are aligned within a time le report. ion chromatograms are integrated for s are obtained. icantly differences between lipids in yed in a whisker plot. Results Lipid Identification Results from LC/dd-MS 2 Data. Bovine heart total lipid extract (1.25 µg/µl) was analyzed by LC/MS and dd-ms/ms in positive and negative ion modes, respectively (Figure 2). LipidSearch software was used to search the LC-MS data for all common lipid classes using a mass tolerance of ±5 ppm for precursor and product ions. For each MS 2 spectrum matching a predicted fragmentation pattern from the lipid precursor ion m/z stored in the LipidSearch database, search results are summarized for the identified lipid species with m-score indicating the fit to the experimental data (Figure 3). The speed advantage of Q Exactive HF (12K resolution) and TOP2 dd-ms 2 is illustrated in Figure 4, compared to Q Exactive Plus (7K) and TOP15 dd-ms 2. At the higher resolution (12K) the Q Exactive HF obtains 2% more lipid identifications. FIGURE 2. High Resolution (7,) Accurate Mass Positive and Negative Total ion Chromatograms of Bovine Heart Total Lipid Extract. Positive Ion FIGURE 3. Search Results for m/z 88.668, PC(4:5) M+HCO 2 - Relative Abundance Negative Ion 5ng_pos_15K_3kms2_top15_2_12_1 SM: 1 8 6 4 2 5B Improved ID and Relative Q Higher MS resolution lipid pr related lipids overlap even u resolution for identification o two Lyso phospholipid speci during analysis at 2.2 minute needed to separate these tw smaller LPE 18:1 peak (m/z is only partially resolved at 2 required for unequivocal iden lipid species is challenging w The number of lipids identifie results are shown in Table 3 exceed the sum of positive a where lipids are identified in FIGURE 5. Increased Reso PC and PE Lipid Species. R 48.15764 +2.1 ppm 48.15639 R = 43K 48.15664 LPE 18:1 C 23 H 47 NO 7 P +.1 ppm R = 83K 48.15713 R t = 2.2 min 15K 3K 6K 12K 48.5 48.1 48.15 48.2 48.25 m/z TABLE 3. Comparison of L Bovine Brain, Heart, Liver Identify Align 329.2483 283.264 PC(18:_22:5) 58.347 82.5862 Match Details FIGURE 4. Increase in the Number of Lipid Species Identified in Bovine Heart Lipid Extract using dd-lcms 2 with Q Exactive HF Compared to Q Exactive Plus 5 QE Plus (top 15) 4 QE HF (top 2) 3 2 1 1ng 25ng 5ng 1ng 25ng Lipid Bovine Brain Class Neg Pos Merged CL LPC 13 34 4 PC 49 152 242 LPE 14 15 24 PE 88 73 171 LPS 6 4 7 PS 35 24 63 LPG 2 2 PG 8 2 9 LPI 4 2 4 PI 19 19 33 PA 8 13 SM 34 41 62 So 4 5 Cer 2 31 34 CerG1 2 33 37 CerG2 2 2 ChE 5 5 ZyE DG 26 35 TG 86 147 CoQ 1 1 Total 284 552 936 4 Highly Sensitive, Robust MS-Based Workflow for Therapeutic Monoclonal Antibody Analysis from Complex Matrices
Bovine heart total lipid extract in positive and negative ion was used to search the LC-MS nce of ±5 ppm for precursor and dicted fragmentation pattern from atabase, search results are re indicating the fit to the tion) and TOP2 dd-ms 2 is 7K) and TOP15 dd-ms 2. At the % more lipid identifications. s Positive and Negative ipid Extract. Relative Abundance :5) M+HCO 2 - Negative Ion Improved ID and Relative Quantitation Results Obtained with Q Exactive HF Higher MS resolution lipid profiling translates into improved identification when closely related lipids overlap even under chromatographic conditions. The effect of MS1 resolution for identification of lipid species at the same retention time is illustrated for two Lyso phospholipid species, 18:1 LPE and 16:p LPC. These two lipids overlap during analysis at 2.2 minutes at m/z in positive ion of 48.3. The mass resolution needed to separate these two lipids is illustrated in Figure 5. At a resolution of 1K, the smaller LPE 18:1 peak (m/z 48.385) is overlapped by LPC 16:p (m/z 48.3449) and is only partially resolved at 22K. The 6K setting giving an actual resolution of 42K is required for unequivocal identification of both M+H ions. Thus, identification of minor lipid species is challenging without sufficient mass resolution, leading to fewer ID s. The number of lipids identified in positive and negative ion, and the aligned (merged) results are shown in Table 3. Note that the number of lipid species after alignment may exceed the sum of positive and negative ion species because of the correlation step where lipids are identified in multiple samples and isomers are present at the MS 2 level. FIGURE 5. Increased Resolution Improves Identification of Overlapping Lyso PC and PE Lipid Species. Resolution (m/z 2) = 15K, 3K, 6K and 12K 5ng_pos_15K_3kms2_top15_2_12_1 3/28/14 3:31:6 SM: 1 8 6 4 5B 48.15764 R t = 2.2 min +2.1 ppm 48.15639 R = 43K 48.15664 LPE 18:1 C 23 H 47 NO 7 P +.1 ppm R = 83K 15K 3K 6K 48.371 48.3946 48.34658 48.34361 48.34458 48.34454 2 48.15713 48.3854 12K 48.5 48.1 48.15 48.2 48.25 48.3 48.35 48.4 48.45 m/z TABLE 3. Comparison of Lipid Species Identified in Total Lipid Extracts of Bovine Brain, Heart, Liver and from Yeast NL: 2.58E5 5ng_pos_15K_3kms2_top15_2_12 _1#1162-1333 +3.6 ppm RT: 2.13-2.43 AV: 35 R = 1K NL: 2.43E5 5ng_pos3k_3kms2_top15_2_12 _1#187-1253 -2.6 ppm RT: 2.1-2.29 AV: 28 T: FTMS R = + 22K p ESI Full ms NL: 2.97E5 5ng_pos6k_3kms2_top15_2_12 LPC 16:p _1#11-1263 C RT: 2.14-2.42 AV: 25 24 H 51 NO 6 P -.6 ppm NL: 2.82E5 R = 86K 5ng_pos12k_3kms2_top15_2_12 _1#999-1144 RT: 2.14-2.43 AV: 18 TABLE 4. Typical Composition o Total Lipid Extracts (Avanti Lipid Wt % Brain Heart Liver Yeas PA 3 1 1 LPC 1 PC 1 5 42 19 LPE 2 PE 1 PG 17 7 22 5 LPI 1 PI 2 3 8 13 PS 11 4 CL 2 Chol 7 Neutrals 5 2 Unknown 59 32 Total 1 1 1 45 FIGURE 6. Composition of Lipid Identified in Bovine Heart ompared to Q Exactive Plus 1ng 25ng Lipid Bovine Brain Bovine Heart Bovine Liver Yeast Class Neg Pos Merged Neg Pos Merged Neg Pos Merged Neg Pos Merged CL 9 9 1 2 2 2 LPC 13 34 4 14 28 3 18 36 39 13 25 4 PC 49 152 242 78 141 253 61 137 294 26 58 92 LPE 14 15 24 14 8 17 15 1 17 9 9 12 PE 88 73 171 92 45 165 74 51 248 19 19 34 LPS 6 4 7 1 1 2 7 1 7 3 3 PS 35 24 63 37 6 53 23 1 65 24 6 38 LPG 2 2 4 4 5 5 1 1 PG 8 2 9 26 4 35 17 5 48 1 3 1 LPI 4 2 4 5 1 5 9 3 9 4 2 5 PI 19 19 33 31 25 55 34 33 98 27 41 55 PA 8 13 3 3 5 11 9 3 18 SM 34 41 62 37 39 56 37 39 91 6 7 12 So 4 5 2 2 5 5 3 3 Cer 2 31 34 1 17 18 5 41 46 2 2 CerG1 2 33 37 1 1 9 9 2 3 CerG2 2 2 3 3 ChE 5 5 5 5 ZyE 4 2 DG 26 35 32 45 29 44 27 31 TG 86 147 174 329 155 248 161 241 CoQ 1 1 5 5 3 3 4 3 Total 284 552 936 351 528 187 31 568 1297 153 374 67 Conclusion Four different total lipid extra lipidomics experiments, sepa Higher resolution and speed identification and relative qu At least 2% more lipid ID s Orbitrap. LipidSearch provides autom single LC-ddMS 2 experimen lipid species. References 1. LIPID MAPS comprehensiv Lipid Res. 29, 5, S9-S1 2. Lipidomics profiling by high dissociation fragmentation: cardiolipins and monolysoc 94 949. dx.doi.org/1.12 Lipid Search is a registered trademark of MKI All other trademarks are the property of Therm This information is not intended to encourage intellectual property rights of others. Thermo Scientific Poster Note PN-64137-ASMS-EN-614S 5
btained with Q Exactive HF proved identification when closely onditions. The effect of MS1 e retention time is illustrated for LPC. These two lipids overlap of 48.3. The mass resolution igure 5. At a resolution of 1K, the d by LPC 16:p (m/z 48.3449) and ing an actual resolution of 42K is ons. Thus, identification of minor esolution, leading to fewer ID s. ive ion, and the aligned (merged) f lipid species after alignment may because of the correlation step omers are present at the MS 2 level. ification of Overlapping Lyso 15K, 3K, 6K and 12K TABLE 4. Typical Composition of Total Lipid Extracts (Avanti Lipids) Wt % Brain Heart Liver Yeast PA 3 1 1 LPC 1 PC 1 5 42 19 LPE 2 PE 1 PG 17 7 22 5 LPI 1 PI 2 3 8 13 PS 11 4 CL 2 Chol 7 Neutrals 5 2 Unknown 59 32 Total 1 1 1 45 Composition of Total Lipid Extracts Composition information for the total lipid extracts listed on Avanti Polar Lipids website are shown in Table 4. The LC-MS 2 results for each lipid extract injected in duplicate were searched using LipidSearch, and positive and negative ion data were aligned within a.4 min window (Table 3). The peak areas for each lipid class were summed and the relative amounts of each lipid class were plotted as % Area from the full scan MS (Figure 6). As expected from the 5 wt % of neutral lipids, the bovine heart extract contained the highest amount of TG (65% by peak area), and the yeast extract contained 61% TG. Brain, liver and yeast extracts contained the most PC. 48.4 48.45 NL: 2.58E5 5ng_pos_15K_3kms2_top15_2_12 _1#1162-1333 +3.6 ppm RT: 2.13-2.43 AV: 35 R = 1K NL: 2.43E5 5ng_pos3k_3kms2_top15_2_12 _1#187-1253 -2.6 ppm RT: 2.1-2.29 AV: 28 T: FTMS R = + 22K p ESI Full ms NL: 2.97E5 5ng_pos6k_3kms2_top15_2_12 LPC 16:p _1#11-1263 C RT: 2.14-2.42 AV: 25 24 H 51 NO 6 P -.6 ppm NL: 2.82E5 R = 86K 5ng_pos12k_3kms2_top15_2_12 _1#999-1144 RT: 2.14-2.43 AV: 18 in Total Lipid Extracts of Bovine Liver Yeast g Pos Merged Neg Pos Merged 1 2 2 2 8 36 39 13 25 4 1 137 294 26 58 92 5 1 17 9 9 12 4 51 248 19 19 34 7 1 7 3 3 3 1 65 24 6 38 5 5 1 1 7 5 48 1 3 1 9 3 9 4 2 5 4 33 98 27 41 55 5 11 9 3 18 7 39 91 6 7 12 5 5 3 3 5 41 46 2 2 9 9 2 3 3 3 5 5 4 2 29 44 27 31 155 248 161 241 3 3 4 3 568 1297 153 374 67 FIGURE 6. Composition of Lipid Extracts by LC-HRMS Peak Area (%) Conclusion Brain Liver Yeast Heart Four different total lipid extracts were profiled and are suitable for benchmarking lipidomics experiments, separation conditions and instrumental optimization. Higher resolution and speed of Q Exactive HF provides more confident identification and relative quantification in a single positive and negative ion run. At least 2% more lipid ID s were obtained using a ultra-high field benchtop Orbitrap. LipidSearch provides automated identification of more than 5 lipid species in a single LC-ddMS 2 experiment, and alignment of multiple samples gives over 1 lipid species. References 1. LIPID MAPS comprehensive classification system for lipids. E Fahy et al., J. Lipid Res. 29, 5, S9-S14. doi: 1.1194/jlr.R895-JLR2. 2. Lipidomics profiling by high-resolution LC-MS and high-energy collisional dissociation fragmentation: focus on characterization of mitochondrial cardiolipins and monolysocardiolipins. S Bird, et al., Anal. Chem. 211, 83, 94 949. dx.doi.org/1.121/ac12598u. Lipid Search is a registered trademark of MKI and Ascentis Express is a registered trademark of Sigma-Aldrich. All other trademarks are the property of Thermo Fisher Scientific and its subsidiaries. This information is not intended to encourage use of these products in any manners that might infringe the intellectual property rights of others. PO64137-EN 614S Heart Yeast Liver Brain 7 6 5 4 3 2 1 6 Highly Sensitive, Robust MS-Based Workflow for Therapeutic Monoclonal Antibody Analysis from Complex Matrices
www.thermoscientific.com 214 Thermo Fisher Scientific Inc. All rights reserved. ISO is a trademark of the International Standards Organization. Lipid Search is a registered trademark of MKI and Ascentis Express is a registered trademark of Sigma-Aldrich. All other trademarks are the property of Thermo Fisher Scientific and its subsidiaries. This information is presented as an example of the capabilities of Thermo Fisher Scientific products. It is not intended to encourage use of these products in any manners that might infringe the intellectual property rights of others. Specifications, terms and pricing are subject to change. Not all products are available in all countries. Please consult your local sales representative for details. Africa +43 1 333 5 34 Australia +61 3 9757 43 Austria +43 81 282 26 Belgium +32 53 73 42 41 Canada +1 8 53 8447 China 8 81 5118 (free call domestic) 4 65 5118 Denmark +45 7 23 62 6 Europe-Other +43 1 333 5 34 Finland +358 9 3291 2 France +33 1 6 92 48 Germany +49 613 48 114 India +91 22 6742 9494 Italy +39 2 95 591 Japan +81 45 453 91 Latin America +1 561 688 87 Middle East +43 1 333 5 34 Netherlands +31 76 579 55 55 New Zealand +64 9 98 67 Norway +46 8 556 468 Russia/CIS +43 1 333 5 34 Thermo Fisher Scientific, San Jose, CA USA is ISO 91:28 Certified. Singapore +65 6289 119 Spain +34 914 845 965 Sweden +46 8 556 468 Switzerland +41 61 716 77 UK +44 1442 233555 USA +1 8 532 4752 PN-64137-EN-614S