Targeted Lipidomics: Discovery of New Fatty Acyl Amides

Tan B, Bradshaw HB, Rimmerman N, Srinivasan H, Yu YW, Krey JF, Monn MF, Chen J-CS-C, Hu S-JS-J, Pickens SR, Walker JM. Targeted Lipidomics: Discovery of New Fatty Acyl Amides. AAPS Journal. 2006; 8(3): E461-E465. DOI: 10.1208/aapsj080354

Bo Tan,¹ Heather B. Bradshaw,¹ Neta Rimmerman,¹ Harini Srinivasan,¹ Y. William Yu,¹ Jocelyn F. Krey,¹ M. Francesca Monn,¹ Jay Shih-Chieh Chen,¹ Sherry Shu-Jung Hu,¹ Sarah R. Pickens,¹ and J. Michael Walker¹

Correspondence to:
J. Michael Walker
Tel: (812) 855-4559
Fax: (812) 855-9032
Email: walkerjm@indiana.edu

The discovery of endogenous fatty acyl amides such as N-arachidonoyl ethanolamide (anandamide), N-oleoyl ethanolamide (OEA), and N-arachidonoyl dopamine (NADA) as important signaling molecules in the central and peripheral nervous system has led us to pursue other unidentified signaling molecules. Until recently, technical challenges, particularly those associated with lipid purification and chemical analysis, have hindered the identification of low abundance signaling lipids. Improvements in chromatography and mass spectrometry (MS) such as miniaturization of high-performance liquid chromatography components, hybridization of multistage mass spectrometers and time-of-flight technology, the development of electrospray ionization (ESI) and of information-dependent acquisition, now permit rapid identification of novel, low abundance, signaling lipids.

Keywords: lipidomics, cannabinoid, fatty acyl amide, information-dependent acquisition, HPLC-MS/MS, nano-HPLC

Since the early 1990s several endogenous lipid derivatives such as anandamide,¹ 2-arachidonoyl glycerol (2-AG),²N-arachidonoyl dopamine (NADA),³ and N-arachidonoyl glycine (NAGly)⁴ have been found in rodent and bovine brains, and their biological actions⁵ and signal transduction pathways have been rigorously studied.^6-8 It is now well established that these fatty acyl amides and esters modulate several physiological processes, including pain sensitivity,⁹ reproduction,¹⁰ immune function,¹¹ and vascular tone¹² among others.

Anandamide was isolated from porcine brains using column chromatography and thin layer chromatography (TLC), and ¹H nuclear magnetic resonance (NMR) and mass spectrometry (MS) spectra were identical to the synthetic material.¹ The physiological functions of anandamide are similar to those of Δ⁹-tetrahydrocannabinol (Δ⁹-THC) in that both activate the cannabinoid receptor 1 (CB1),¹³ and anandamide also acts on the transient receptor potential vanilloid 1 (TRPV1).¹⁴ Homo-γ-linolenoyl ethanolamide and docosatetraenoyl ethanolamide were partially purified from porcine brains on a silica gel column and identified by comparison of the gas chromatography (GC)-MS spectra of the endogenous compounds with those of the synthetic compounds.¹⁵ Palmitoyl ethanolamide (PEA) was originally isolated from soybeans, egg yolk, and peanut meal,¹⁶ and subsequently from mammalian tissue.¹⁷ The structural information was obtained by infrared spectroscopy.¹⁶ Although PEA does not bind the cannabinoid receptors 1 and 2 (CB2) with high affinity, it enhances the effect of anandamide by an unknown mechanism.¹⁸ Oleoyl ethanolamide (OEA) was purified using 3 high-performance liquid chromatography (HPLC) columns and identified by GC-MS.¹⁹ OEA activates the peroxisome proliferator-activated receptor α (PPAR-α)²⁰^,²¹ and TRPV1 receptors,²² while it lacks high affinity for CB1 and CB2 receptors.²⁰^,²¹

NADA and other fatty acyl dopamines were identified by our laboratory³ using solid phase extraction and HPLC-MS/MS. They are found in nervous tissues, especially striatum, hippocampus, cerebellum, and dorsal root ganglion. NADA activates TRPV1 receptors with a potency similar to that of capsaicin; it also binds to CB1 and shows cannabimimetic effects.³N-oleoyldopamine (OLDA) is more potent than NADA at TRPV1 and less potent at CB1.²³ The dopamides of palmitic and stearic acids lack affinity for TRPV1 and CB1 but act in concert with NADA and anandamide to enhance TRPV1-mediated calcium mobilization.²⁴ These findings led to the hypothesis that synthesis and release of NADA and OLDA would cause pain sensations via activation of TRPV1.

NAGly was synthesized for structure-activity studies of the CB1 receptor,²⁵ and our laboratory reported the existence of NAGly as a natural constituent of brain, along with N-arachidonoyl-γ-aminobutyric acid and N-arachidonoyl alanine.⁴ These were the first endogenous fatty acyl amino acids identified in mammals. In these studies, Huang et al⁴ used ion exchange and reversed phase solid phase extraction to partially purify the compounds prior to liquid chromatography (LC)-MS/MS on the ion trap and hybrid quadrupole/time-of-flight (QqTOF) instruments. Widely distributed among mammalian tissues, NAGly causes hot plate analgesia in mice and suppresses formalin-induced pain behavior in rats.⁴ Recently, N-arachidonoyl serine was identified in a bovine brain extract by Milman et al.²⁶ This compound triggers vasodilatation of rat mesenteric arteries and abdominal aorta. These findings suggest that other fatty acyl amino acids with similar structures may also exist endogenously and possess similar biological functions.

From the last section, it is clear that there are 2 types of identification strategies used for fatty acyl amides. In the first, extracts are purified to homogeneity and subjected to NMR, infrared (IR), and MS analysis. This approach works well with compounds that are abundant in biological systems but not for those that occur at trace levels. Milligram quantities of compound in pure form are required for ¹³C NMR and microgram quantities for ¹H NMR. This requirement is extremely difficult for compounds that occur in the pmol/g range (eg, anandamide, NAGly), requiring the use of tissue from large species such as cow or pig, and even in these cases, the amounts that can be isolated often do not allow for ¹³C NMR.¹

A second method relies on HPLC-MS/MS methods, where accurate masses are obtained (<10 ppm) and the full fragmentation pattern and HPLC retention time of endogenous material is compared with that of synthetic material. While it is still possible for different chemical species to produce nearly identical mass spectra, high precision mass measurements, coupled with appropriate fragmentation patterns and retention times allow the proposal of the chemical formulae and structures with a reasonable degree of confidence. The amount and the purity of the sample are not as critical as in the first method because recent advances in MS and liquid chromatography permit isolation of molecular species of a particular molecular weight (unit accuracy) followed by product ion scans with trace amounts (eg, 250 fmol on column). Using this approach, small amounts of tissue (eg, 2-6 g) are subjected to a series of liquid-liquid extraction and solid-phase purification, and a QqTOF instrument coupled with nano-HPLC is used to identify the compound.

Whereas the QqTOF performs well for identification of novel compounds, multiple-reaction monitoring (MRM) scan on a triple quadrupole MS coupled with HPLC is the preferred method for quantifying levels of endogenous compounds. This method is useful for pilot studies because the automation of the triple quadrupole with HPLC makes MRM extremely sensitive and easy to use. Although such experiments often yield a single peak, matching of retention times to a standard is required for compound identification. However, MRM experiments cannot replace product ion scans because only one or a few major fragment ions are used in MRM experiments, while a full range mass spectrum is required for identification of new compounds. Although a product ion scan can be performed on a triple quadrupole MS, it requires considerably more (10- to 50-fold) sample and produces reduced mass resolution and accuracy compared with the QqTOF.

Large increases in sensitivity can be obtained using nano-HPLC because the MS signal is proportional to the concentration of the sample, and the smaller column (75 μm inner diameter [id]) requires less sample to produce the same concentration than standard analytical (4.1 mm id) and narrow bore (2.1 mm) columns. Here, the sensitivity increases by the second power of the column diameters with a similar decrease in flow rate. Hence, lowering the column diameter from 4.1 mm to 75 μm requires a drop in the flow rate from 1 mL/min to ~250 nL/min, with a concomitant theoretical increase in sensitivity of 1600-fold. Although the theoretical increase in sensitivity is not achieved in our laboratory, when coupled with nano-HPLC, typically less than 250 fmol of a fatty acyl amide on column is required in the product ion scan mode for QqTOF for identification of new compounds and 100 attomol is required in the MRM scan on the triple quadruple MS for quantification of known compounds. The relatively recent development of nano-HPLC, triple quadrupole, and TOF mass spectrometers provides investigators with markedly improved methods for identification of low abundance signaling lipids.

Although lipids present in the cell differ greatly in their structural, chemical, and physical properties, they typically occur in families composed of conjugates of lipid moieties and functional groups in a combinatorial fashion. This was evident, for example, in the various fatty acyl ethanolamides, of which anandamide is but one product, and a minor one.²⁷ Precursor ion and neutral loss scans available with the triple quadrupole MS are the common tool to identify a group of lipids with similar structures. These scanning modes require relatively large amount of samples compared with MRM, which is the bottleneck for identification of trace lipids. To improve throughput for proteomics, a method was needed to identify as many putative compounds as possible in one run, while maintaining high sensitivity. Information-dependent (or data-dependent) acquisition (IDA)²⁸ was developed to provide this capability. Currently, IDA methods are typically available for QqTOF, triple quadrupole, and ion trap mass spectrometers. No baseline separation of analytes is required, and the mass accuracy and sensitivity is not compromised.²⁹ With appropriate improvements in offline data analysis, IDA has proven useful for lipidomic approaches.³⁰

A cycle of IDA contains a survey scan and several corresponding MS/MS events. The general mechanism is shown in Figure 1. Ions found in the survey scan are examined using a series of switch criteria. The computer selects ions for further experimentation based upon such factors as intensity, presence in an inclusion list, absence from an exclusion list, recency of occurrence, isotope ratio, and a fixed mass difference between pairs of ions. A full scan without fragmentation of the molecular ion is the most common survey scan but neutral loss, precursor ion, and MRM scans can also trigger the product ion scan.

As noted above, IDA was developed for and is widely used in proteomics and metabolomics research. The authors are aware of only one report on the use of IDA for lipidomics.³⁰ In that report, IDA was used to identify 35 known triacylglycerols and 90 glycerophospholipids in a lipid extract from Caenorhabditic elegans.³⁰ While it was possible to identify a large number of high abundance lipids by this approach, the use of infusion rather than an HPLC separation would greatly limit the number of signaling lipids that could be identified. Detection of compounds within the nanomolar concentration range was based on using 20 to 30 seconds as the accumulation time. This approach would not be practical with HPLC separation because only one ion per LC peak would be possible, thereby eliminating analysis of coeluting lipids. The authors’ suggestion for analysis of low abundance compounds was to use an inclusion list. While a logical step, this presupposes the structure of the compound to be studied, and it should be noted that the peak intensities obtained with HPLC would be much greater, relative to background noise, than that obtained with flow injection. Hence, the detection limits of low abundance compounds would still be much lower with flow injection than with LC separation.

We developed a sensitive IDA method that was successfully applied to identification of new lipids in a biological sample. Methanol extracts of rodent brains were purified by solid phase extraction on ion exchange, reversed phase, and normal phase columns, and the fractions were injected into a nano-LC/hybrid quadrupole/TOF MS using nano-electrospray ionization (Applied Biosystems/MDS Sciex QStar, Foster City, CA) at a flow rate of 250 nL/min. Data were acquired with IDA triggered by TOF masses obtained in Analyst Software (Applied Biosystems/MDS Sciex). The previously selected target ions were excluded for 30 seconds. Isotope ions were excluded throughout the run, and none of the inclusion and exclusion list was used. One TOF scan and the following MS/MS scan from one IDA run are shown in Figure 2. The data were analyzed with a computer program constructed in house, which compared spectral peaks from the MS with theoretical masses based on the formulae combinatorially of 15 fatty acids and 23 amino acids and other small endogenous amines capable of forming amides. An initial match required the spectra from the IDA experiment to contain the mass of the molecular ion and a fragment mass corresponding to the amino acid. Subsequent inspection of matches was performed to confirm the presence of other predicted fragments. The MS/MS spectra of the brain extract were then compared with a library of synthetic N-acyl amino acids (>100). The results obtained from the computer program were consistent with those obtained from inspection of the mass spectra. One run of an IDA experiment generated ~1000 mass spectra from which more than 40 acyl amino acids were identified. Of these, more than 30 were novel compounds and had structures similar to previously identified fatty acyl amides. It is possible that these novel endogenous compounds have important biological signaling functions.

Mass spectrometric approaches to the identification of lipid signaling molecules are reviewed here. Compared with proteomics, there are fewer applications of lipidomics, indicating difficulties associated with the identification of endogenous lipids. Specific purification and HPLC/MS-MS methods have been designed and proved highly efficient to identify the members of one family of lipids, the fatty acyl amides. Some members of this family have demonstrated important biological functions and play a role in pain, immune function, reproduction, and appetite.

The authors are grateful for the support of the National Institute on Drug Abuse (grant numbers DA-018224, DA-020402, and F32-DA-016825), the Gill Center for Biomolecular Science, Indiana University, Bloomington, IN, and the Lilly Foundation Inc, Indianapolis, IN. The authors also acknowledge Dr David K. O’Dell for synthesis of all the standards in the library used in this work.

1. Devane WA, Hanus L, Breuer A, et al. Isolation and structure of a brain constituent that binds to the cannabinoid receptor. Science. 1992;258:1946-1949.
PubMed

2. Mechoulam R, Ben-Shabat S, Hanus L, et al. Identification of an endogenous 2-monoglyceride, present in canine gut, that binds to cannabinoid receptors. Biochem Pharmacol. 1995;50:83-90.
PubMed DOI: 10.1016/0006-2952(95)00109-D

3. Huang SM, Bisogno T, Trevisani M, et al. An endogenous capsaicin-like substance with high potency at recombinant and native vanilloid VR1 receptors. Proc Natl Acad Sci USA. 2002;99:8400-8405.
PubMed DOI: 10.1073/pnas.122196999

4. Huang SM, Bisogno T, Petros TJ, et al. Identification of a new class of molecules, the arachidonyl amino acids, and characterization of one member that inhibits pain. J Biol Chem. 2001;276:42639-42644.
PubMed DOI: 10.1074/jbc.M107351200

5. Bradshaw HB, Walker JM. The expanding field of cannabimimetic and related lipid mediators. Br J Pharmacol. 2005;144:459-465.
PubMed DOI: 10.1038/sj.bjp.0706093

6. Elphick MR, Egertova M. The neurobiology and evolution of cannabinoid signalling. Philos Trans R Soc Lond B Biol Sci. 2001;356:381-408.
PubMed DOI: 10.1098/rstb.2000.0787

7. Piomelli D. The molecular logic of endocannabinoid signalling. Nat Rev Neurosci. 2003;4:873-884.
PubMed DOI: 10.1038/nrn1247

10. Bradshaw HB, Rimmerman N, Krey JF, Walker JM. Sex and hormonal cycle differences in brain levels of pain-related cannabimimetic lipid mediators. Am J Physiol Regul Integr Comp Physiol [serial online]. 2006;Mar:.
DOI: 10.1152/ajpregu.00933.2005

11. Croxford JL, Yamamura T. Cannabinoids and the immune system: potential for the treatment of inflammatory diseases? J Neuroimmunol. 2005;166:3-18.
PubMed DOI: 10.1016/j.jneuroim.2005.04.023

12. Ralevic V, Kendall DA, Randall MD, Smart D. Cannabinoid modulation of sensory neurotransmission via cannabinoid and vanilloid receptors: roles in regulation of cardiovascular function. Life Sci. 2002;71:2577-2594.
PubMed DOI: 10.1016/S0024-3205(02)02086-6

13. Crawley JN, Corwin RL, Robinson JK, Felder CC, Devane WA, Axelrod J. Anandamide, an endogenous ligand of the cannabinoid receptor, induces hypomotility and hypothermia in vivo in rodents. Pharmacol Biochem Behav. 1993;46:967-972.
PubMed DOI: 10.1016/0091-3057(93)90230-Q

14. Ross RA. Anandamide and vanilloid TRPV1 receptors. Br J Pharmacol. 2003;140:790-801.
PubMed DOI: 10.1038/sj.bjp.0705467

15. Hanus L, Gopher A, Almog S, Mechoulam R. Two new unsaturated fatty acid ethanolamides in brain that bind to the cannabinoid receptor. J Med Chem. 1993;36:3032-3034.
PubMed DOI: 10.1021/jm00072a026

16. Kuehl FA, Jr, Jacob TA, Ganley OH, Ormond RE, Meisinger MAP. The identification of N-2 (hydroxyethyl)-palmitamide as a naturally occurring anti-inflammatory agent. J Am Chem Soc. 1957;79:5577-5578.
DOI: 10.1021/ja01577a066

17. Bachur NR, Masek K, Melmon KL, Udenfriend S. Fatty acid amides of ethanolamine in mammalian tissues. J Biol Chem. 1965;240:1019-1024.
PubMed

18. Calignano A, La Rana G, Giuffrida A, Piomelli D. Control of pain initiation by endogenous cannabinoids. Nature. 1998;394:277-281.
PubMed DOI: 10.1038/28393

19. Fontana A, Di Marzo V, Cadas H, Piomelli D. Analysis of anandamide, an endogenous cannabinoid substance, and of other natural N-acylethanolamines. Prostaglandins Leukot Essent Fatty Acids. 1995;53:301-308.
PubMed DOI: 10.1016/0952-3278(95)90130-2

20. Fu J, Gaetani S, Oveisi F, et al. Oleylethanolamide regulates feeding and body weight through activation of the nuclear receptor PPAR-alpha. Nature. 2003;425:90-93.
PubMed DOI: 10.1038/nature01921

21. Guzman M, Lo Verme J, Fu J, Oveisi F, Blazquez C, Piomelli D. Oleoylethanolamide stimulates lipolysis by activating the nuclear receptor peroxisome proliferator-activated receptor alpha (PPAR-alpha). J Biol Chem. 2004;279:27849-27854.
PubMed DOI: 10.1074/jbc.M404087200

22. Ahern GP. Activation of TRPV1 by the satiety factor oleoylethanolamide. J Biol Chem. 2003;278:30429-30434.
PubMed DOI: 10.1074/jbc.M305051200

23. Chu CJ, Huang SM, De Petrocellis L, et al. N-oleoyldopamine, a novel endogenous capsaicin-like lipid that produces hyperalgesia. J Biol Chem. 2003;278:13633-13639.
PubMed DOI: 10.1074/jbc.M211231200

24. De Petrocellis L, Chu CJ, Moriello AS, Kellner JC, Walker JM, Marzo VD. Actions of two naturally occurring saturated N-acyldopamines on transient receptor potential vanilloid 1 (TRPV1) channels. Br J Pharmacol. 2004;143:251-256.
PubMed DOI: 10.1038/sj.bjp.0705924

25. Sheskin T, Hanus L, Slager J, Vogel Z, Mechoulam R. Structural requirements for binding of anandamide-type compounds to the brain cannabinoid receptor. J Med Chem. 1997;40:659-667.
PubMed DOI: 10.1021/jm960752x

26. Milman G, Maor Y, Abu-Lafi S, et al. N-arachidonoyl L-serine, an endocannabinoid-like brain constituent with vasodilatory properties. Proc Natl Acad Sci USA. 2006;103:2428-2433.
PubMed DOI: 10.1073/pnas.0510676103

27. Hansen HS, Lauritzen L, Moesgaard B, Strand AM, Hansen HH. Formation of N-acyl-phosphatidylethanolamines and N-acetylethanolamines: proposed role in neurotoxicity. Biochem Pharmacol. 1998;55:719-725.
PubMed DOI: 10.1016/S0006-2952(97)00396-1

28. Josephs JL. Detection and characterization of fumonisin mycotoxins by liquid chromatography/electrospray ionization using ion trap and triple quadrupole mass spectrometry. Rapid Commun Mass Spectrom. 1996;10:1333-1344.
PubMed DOI: 10.1002/(SICI)1097-0231(199608)10:11<1333::AID-RCM620>3.0.CO_2-F

29. Scholz K, Dekant W, Volkel W, Pahler A. Rapid detection and identification of N-acetyl-L-cysteine thioethers using constant neutral loss and theoretical multiple reaction monitoring combined with enhanced product-ion scans on a linear ion trap mass spectrometer. J Am Soc Mass Spectrom. 2005;16:1976-1984.
PubMed DOI: 10.1016/j.jasms.2005.08.003

30. Schwudke D, Oegema J, Burton L, et al. Lipid profiling by multiple precursor and neutral loss scanning driven by the data-dependent acquisition. Anal Chem. 2006;78:585-595.
PubMed DOI: 10.1021/ac051605m

31. Saghatelian A, Trauger SA, Want EJ, Hawkins EG, Siuzdak G, Cravatt BF. Assignment of endogenous substrates to enzymes by global metabolite profiling. Biochemistry. 2004;43:14332-14339.
PubMed DOI: 10.1021/bi0480335