Lipidomics at the Interface of Structure and Function in Systems Biology

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector Chemistry & Biology Crosstalk

Richard W. Gross1,2,3,4,* and Xianlin Han1,2 1Division of Bioorganic Chemistry and Molecular Pharmacology 2Departments of Medicine 3Developmental Biology Washington University School of Medicine, St. Louis, MO 63110, USA 4Department of Chemistry, Washington University, St. Louis, MO 63105, USA *Correspondence: [email protected] DOI 10.1016/j.chembiol.2011.01.014

Cells, tissues, and biological ﬂuids contain a diverse repertoire of many tens of thousands of structurally distinct lipids that play multiple roles in cellular signaling, bioenergetics, and membrane structure and function. In an era where lipid-related disease states predominate, lipidomics has assumed a prominent role in systems biology through its unique ability to directly identify functional alterations in multiple lipid metabolic and signaling networks. The development of shotgun lipidomics has led to the facile accrual of high density information on alterations in the lipidome mediating physiologic cellular adaptation during health and pathologic alterations during disease. Through both targeted and nontargeted investigations, lipidomics has already revealed the chemical mechanisms underlying many lipid-related disease states.

The Multiple Roles of Lipids (Zhang et al., 2002; Breen et al., 2005). Cantley, 2006; Chiang et al., 2006; Simon in Cellular Function On a third level, lipid membranes serve and Cravatt, 2006). During evolution, Lipids play multiple diverse roles in as molecular scaffolds that promote pro- adaptive alterations in membrane struc- cellular function that must be effectively ductive interactions among membrane- ture and function were selected by integrated into chemical and genetic associated moieties that regulate cellular their integrated effects on cellular metab- networks to allow each cell to fulfill its signaling to facilitate the transmission of olism and signaling. This evolutionary specific biological function. By some biological information across cell mem- process has led to the development of recent estimates, cellular lipids encom- branes, between intracellular compart- multiple membrane-delimited compart- pass tens of thousands of structurally ments or to other cells. Furthermore, the ments (e.g., mitochondria, peroxisomes, distinct compounds (van Meer, 2005; molecular dynamics and physical proper- endoplasmic reticulum, etc.) which Shevchenko and Simons, 2010). On the ties of membrane bilayers are critical perform integrated specialized processes most fundamental level, lipids are the determinants of the activities of trans- that are essential for efficient cellular func- main components of biological mem- membrane proteins such as ion channels tion and adaptation to external perturba- branes where they serve to sequester, and ion pumps (e.g., Schmidt and tions (Figure 1). organize, and distribute the molecular MacKinnon, 2008). Through the use of a entities necessary for life processes. On diverse structural repertoire of lipids to The Growth of Lipidomics a second level, lipids (e.g., fatty acids regulate membrane surface charge, in Systems Biology and triglycerides) are important fuel sour- curvature and molecular dynamics, many Similar to the Systems Biology fields of ces for many cell types. Lipids supply biological functions of membranes can genomics and proteomics, the field of energy for cellular function through oxida- be fulfilled. Thus, the precise covalent lipidomics begins with the identification tion and facilitate metabolic flexibility nature of membrane molecular constitu- and quantitation of lipids that collectively through their ability to store chemical ents within the lipid bilayers, including comprise the lipidome (i.e., the collection energy during times of caloric excess alterations in class, subclass and indi- of lipid molecular species) in each and harvest the energy stored in lipids vidual molecular species are important cell, tissue, or biologic fluid of interest. (e.g., triglycerides) during energy deple- determinants of membrane structure that Through identification and quantitation of tion. Moreover, lipids play prominent roles facilitate many specialized membrane- alterations in lipid molecular species, in the regulation of cellular bioenergetics mediated cellular functions (Figure 1). a high-density matrix containing funda- through integrating oxidative metabolic Finally, many classes of lipids serve as mental information on the metabolic state, programs (Michalik et al., 2006), modu- second messengers of signal transduc- nutritional history and functional status of lating systemic energy balance through tion (e.g., eicosanoids, lysolipids, phos- each cell type can be obtained. Moreover, eicosanoid and lysolipid production phoinositides, endocannabinoids) that the field of lipidomics encompasses the (Vegiopoulos et al., 2010; Skoura and reside in biologic membranes in a latent roles of specific membrane lipid constitu- Hla, 2009), and regulating mitochondrial state which can be activated either by ents in mediating membrane domain electron transport chain flux and coupling hydrolysis and/or covalent transformation formation (e.g., rafts), in facilitating efficiency (e.g., cardiolipin, fatty acids) (e.g., Wolf and Gross, 1985; Shaw and interactions among spatially interwoven

Ion Channels Receptors and Transporters

+ + +

PLA PLC Gx PLD

Signaling Membrane Structure

Cellular Function

Bioenergetics

Mitochondrial H+ H+ H+ Inner Membrane + UCP H F0 c I e- Q III IV β II CPT II α O2 Acyl- F1 carnitine β-Oxidation NADH Electron Transport Chain Acyl-CoA Trifunctional Complex ADP ATP

Figure 1. The Pleiotropic Roles of Lipids in Cellular Function Lipids fulfill multiple roles in cellular function including cellular signaling (top left) through: (1) harboring latent second messengers of signal transduction that are released by phospholipases (PLA, PLC, and PLD enzymes); (2) covalent transformation of membrane lipids into biologically active ligands by kinases (e.g., PI 3,4,5 triphosphate); (3) providing molecular scaffolds for the assembly of protein complexes mediating receptor/effector coupling (e.g., G protein-coupled receptors); and (4) coupling the vibrational, rotational, and translational energies and dynamics of membrane lipids to transmembrane proteins such as ion channels and transporters (top right) thereby facilitating dynamic cooperative lipid-protein interactions that collectively regulate transmembrane protein function. More- over, lipids play essential roles in mitochondrial cellular bioenergetics (bottom) through the use of fatty acids as substrates for mitochondrial b-oxidation (bottom left) that result in the production of reducing equivalents (e.g., NADH). The chemical energy in NADH is harvested through oxidative phosphorylation whose flux is tightly regulated by mitochondrial membrane constituents including cardiolipins which modulate electron transport chain (ETC) supercomplex formation. A second mechanism modulating mitochondrial energy production is the dissipation of the proton gradient by the transmembrane flip-flop of fatty acids in the mitochondrial inner membrane bilayer as well as the fatty acid-mediated regulation of uncoupling proteins (UCP).

networks of signaling proteins (e.g., G disease states. Lipidomics has now The Challenges of Early Lipid protein receptor-effector coupling), and matured into a field that has already iden- Analytical Techniques in providing dynamic highly specialized tified biomarkers predictive of disease Historically, progress in lipid research has molecular scaffolds for the construction states, alterations in the profiles of lipids been hindered by the difficulties inherent of microscopic and macroscopic chemi- that reflect disease severity and can be in the identification of lipid structure cal assemblies necessary for life pro- used to determine treatment efficacy (e.g., the chemical nature and regiospeci- cesses (e.g., Klose et al., 2010). (Nomura et al., 2010; Porter et al., 2010; ficity of aliphatic chains in each lipid The large majority of diseases in indus- Bartz et al., 2007). Thus, through the class and subclass) and quantitation of trialized societies in the 21st century, comprehensive analysis of alterations in individual molecular species. Over many such as diabetes, obesity, atheroscle- lipid molecular species and their abun- decades multiple approaches have rosis, myocardial infarction, and stroke, dance, the field of lipidomics produces been applied to the separation, quantita- to name a few, are lipid-related disorders a high-density array of diagnostic, thera- tion and characterization of lipids in- (e.g., Unger, 2002; Unger et al., 2010; peutic and mechanistic information which cluding thin layer chromatography, gas Dixon, 2010). Thus, lipidomics offers a has greatly impacted our understanding chromatography, NMR, and HPLC in con- direct and unique perspective into the of the complex roles of lipids in health junction with a variety of complementary pathologic alterations in cellular regula- and disease (e.g., Nomura et al., 2010; procedures including chemical hydro- tory networks that promote lipid-mediated Mancuso et al., 2009; Liu et al., 2010). lysis, regiospecific enzymatic cleavage

and spectrophotometric assays in at- including the human heart (Hazen et al., the complexity of lipids in mammalian tempts to determine the diversity of lipid 1993). The use of HPLC followed by direct cell membranes (Han and Gross, 1994), structures in biological systems and quan- ionization of intact nonvolatile lipids and the roles of phospholipase-mediated titate their abundance. However, these analysis by mass spectrometry heralded cleavage of arachidonoyl containing lipids strategies required multiple sequential the beginning of a new era in under- in modulating the gating currents of steps each of which possessed limited standing the pleiotropic roles of lipids in voltage-dependent ion channels (Gubi- sensitivity and accuracy that collectively cellular function. tosi-Klug et al., 1995), the presence of resulted in the propagation of errors. chain elongation of acylcarnitine molec- Although the utility of GC-MS for the Electrospray Ionization as an ular species during myocardial ischemia analysis of volatile nonpolar lipids has Enabling Technology for Deep (Ford et al., 1996), the identification of provided a robust platform for analysis Penetrance into the Lipidome a-oxidation in adipocytes (Su et al., of volatile lipids for many years, the over- Over the last decade, several different 2004), and dynamic alterations in cardio- whelming majority of cellular membrane strategies for analysis of cellular lipi- lipin content and molecular species constituents are nonvolatile charged domes have been developed. Here, we distribution that occur during diabetes moieties that are not accessible by GC- will first focus on electrospray ionization- and obesity that lead to mitochondrial MS methods. Thus, progress in lipidomics mass spectrometry (ESI-MS) and its dysfunction (Han et al., 2007). was largely dependent on the develop- robust contribution to the development The dramatic increase in sensitivity ment of a mass spectrometric approach of the field and subsequently discuss resulting from ESI for mass spectrometric for the identification and quantitation of other emerging methodologies later in analysis of lipids clearly represented an charged nonvolatile lipids. the text. Currently, the most commonly enabling event in the growth and employed ionization method for lipido- development in the field of lipidomics. The Emergence of Sequential HPLC mics is ESI-MS which was developed by Moreover, the robust increase in signal and Mass Spectrometry as the late John Fenn (Fenn et al., 1989)in to noise during ESI readily led to applica- a Principal Strategy for Lipidomic his Nobel Prize winning work. Electro- tion of tandem mass spectrometric Analyses spray ionization is effected by passage approaches for lipid analysis (Han and The development of fast atom bombard- of a solution containing analytes Gross, 1995) allowing discrimination of ment-mass spectrometry (FAB-MS) in through a narrow orifice at a high electric the multiple isobaric and isomeric molec- the early 1980s allowed the relatively potential (typically 4 kV) resulting in the ular species present at each m/z value. soft ionization and desorption of intact formation of charged droplets in the ioni- The assignment of the aliphatic chain nonvolatile charged lipids from a variety zation chamber. The use of ESI improved composition and the regiospecificity of of chemical matrices resulting in the ionization efficiencies for nonvolatile aliphatic chains could be directly deter- identification of the m/z of the molecular charged lipids by two to three orders of mined by ESI MS/MS through the differ- ion. Using HPLC separations (either magnitude in comparison to FAB-MS ential fragmentation kinetics of aliphatic straight phase, reversed phase, or multi- methods (Han and Gross, 1994). The ioni- moieties present at the sn-1 and sn-2 dimensional orthogonal separations [e.g., zation efficiency of most lipids during the positions (Han and Gross, 1996). Two sequential SCX and RPHPLC]) followed electrospray process is largely dependent basic approaches for ESI MS/MS have by FAB-MS, the molecular ions corre- on the charge density present in each been developed, each with context sponding to individual lipid molecular lipid. Thus, the charge state and magni- dependent strengths and limitations. species in mammalian tissues could be tude of the dipole in the lipid analyte In the traditional approach, lipids are identified (Gross, 1984). Thus, in conjunc- are the predominant contributors to its separated by HPLC and directly sprayed tion with regiospecific enzymatic hydro- interactions with the electric field in into the ESI ion source for mass spectro- lysis, the identification and quantitation the ion source and thus represent the metric analysis by molecular ion moni- of nonvolatile lipid molecular species primary determinants of the ionization toring in conjunction with product ion was possible. This approach demon- efficiency of each analyte. Accordingly, analysis, selected reaction monitoring strated that canine myocardial sarco- among molecular classes comprised of (SRM), or other fragmentation strategies. lemmal membranes are composed similarly charged or zwitterionic lipids, In a second approach, termed Shotgun predominantly of plasmalogen molecular the ionization efficiencies of individual Lipidomics, direct infusion of organic constituents, that the major storage depot molecular species of lipids are largely extracts of tissues, cells, and biologic for arachidonic acid in myocardium was independent of the aliphatic chain length fluids into the mass spectrometer is per- plasmalogen molecular species and that within each lipid class analyzed in the formed. Direct infusion of organic specialized intracellular membrane com- low-concentration regime. In cases where extracts allows marked increases in the partments were composed of discrete analytes of interest do not carry an signal to noise ratio of molecular ions sets of phospholipid classes, subclasses, inherent charge, ESI can be affected corresponding to individual molecular and individual molecular species (Gross, through adduction with other anions or species. Concurrently, direct infusion 1984, 1985). Utilization of HPLC followed cations in solution to induce charge facilitates the utilization of a wide variety by FAB-MS provided abundant informa- separation to facilitate ionization. The of informative fragmentation strategies tion on the specialized lipid compositions use of ESI and mass spectrometric anal- that are not limited by the transient elution present in diverse cell types, subcellular ysis of molecular ions directly from lipid of individual lipid molecular species compartments, and many organ systems extracts led to many new insights into during column chromatography.

865.8 891.8 Multidimensional Mass 100 TAG IS structures within biologic Spectrometry-Based Molecular 849.8 samples can be defined di- 50 Ion Shotgun Lipidomics 837.8 rectly from their organic ex- 811.8 The large majority of cellular Relative Intensity (%) 0 tracts. For example, if neutral lipids are comprised of linear 863.9 loss scanning of aliphatic 837.9 combinations of aliphatic chains present in triglycerides chains and polar head groups for all major biologically 809.8 NL254.1 that are covalently linked to a (16:1 FA) 865.9 occurring aliphatic chains is glycerol or sphingosine back- performed, then extending 839.9 bone. Accordingly, determi- a perpendicular line from any 811.8 nation of the head group NL256.2 m/z value of the full mass structure as well as the nature (16:0 FA) 887.8 scan identifies the nature 861.9 and regiospecificity of the and relative abundance of aliphatic chains represents a fragment ions at each cross 835.8 formal structural identification NL278.2 peak (Han and Gross, 2001) (18:3 FA) 889.9 of the overwhelming majority 863.9 (Figure 2). Bioinformatic array of lipids. To exploit the limited analysis of these cross peaks degrees of freedom in lipid 835.8 can then be used to identify NL280.2 structure, in conjunction with (18:2 FA) 891.9 the structure of each indi- chemically predictable frag- 865.9 vidual isobaric lipid molecular mentation patterns during species present at each m/z 837.9 tandem mass spectrometry, NL282.2 value in the full mass scan. (18:1 FA) we developed multidimen- 893.9 In this two dimensional array, sional mass spectrometry- 867.9 the first dimension comprises based Shotgun Lipidomics 839.8 the molecular ions (shown NL284.2 (MDMS-SL) (Han and Gross, (18:0 FA) 907.9 on the x axis), while the 2005) that has served as a 881.8 second dimension corre- critical technology to under- Triacylglycerols of Blocks Building sponds to the mass of indi- 855.9 935.9 stand the multiple roles of NL304.2 vidual aliphatic chains that (20:4 FA) lipids in biologic function. 935.9 are subject to neutral loss MDMS-SL is comprised of scanning. These two-dimen- four component parts 911.9 sional mass spectrometric NL330.2 including: (1) multiplexed (22:5 FA) arrays have units of mass in extractions and chemistries; 750 800850 900 950 each dimension and thus are (2) intrasource separation of m/z analogous in both concept lipids based upon their and principle to 2D-NMR intrinsic electrical propensi- Figure 2. Electrospray Ionization Mass Spectrometric Analysis of plots which derive structural ties; (3) multidimensional Extracts of Murine White Adipose Tissue by MDMS-SL information from through- Bligh and Dyer extracts of white adipose tissue from mice fed a high-fat diet mass spectrometry; and (4) were prepared as previously described (Han and Gross, 2005) and directly space or through-bond bioinformatic array analysis. infused into the ESI ion source. Positive-ion ESI mass spectra identified connectivities present in the MDMS-SL begins with a mul- multiple molecular ions in the full mass scan (top row; x axis) that were quan- targeted molecule, but have tiplexed series of extractions titated through ratiometric comparisons with internal standard (tri 17:1 triacyl- units of frequency on each glycerol, m/z 849.8) after corrections for isotope abundance and acyl chain and chemistries that are opti- length and unsaturation effects. Tandem mass spectrometric analysis was axis. Thus, these two-dimen- mized to collectively result in performed using neutral loss scanning for the indicated naturally occurring sional maps of lipid structure the integrated analysis of aliphatic chains. All scans were normalized to the base peak of the individual have been referred to as spectrum. Through bioinformatic analysis of the ion counts in each row (x axis over 30 lipid classes leading scan) and column (y axis scans), the compositional identities of each individual two-dimensional mass spec- to the identification and quan- molecular species can be determined and their relative abundance can be trometry (2D-MS) since they titation of many hundreds to quantified. are entirely analogous to 2D- thousands of lipid molecular NMR spectroscopy. More- species directly from lipid over, additional dimensions extracts. Through alterations in the pH of effectively replacing the need for chroma- can be accessed through precursor ion the infusate solution, the highly selective tography prior to mass spectrometric or neutral loss scanning of multiple sets ionization of groups of specific lipid analysis. Through construction of multidi- of informative fragment ions under differ- classes based on their intrinsic electrical mensional arrays of full mass scans of ential conditions (e.g., polar head groups, properties could be effectively achieved molecular ions with precursor ion scan- varied collision energies) thereby within the ion source (Han et al., 2004). ning or neutral loss scanning of aliphatic providing a high-density matrix of struc- This process, now known as intrasource chains, polar head groups, and other tural information in n-dimensional space. separation, results in the resolution of lipid informative product ions, a comprehen- These multidimensional spectra are constituents in the ion source, thereby sive description of the diversity of lipid ideally suited to bioinformatic array

for eicosanoid analyses (e.g., Bollinger et al., 2010)(Figure 3). One limitation of MDMS-based SL is the limits of detection and quantitation that are feasible for low abundance species. Currently, the lower limits of detection for lipids in favorable cases using MDMS-SL are in the high attomole to low femtomole level. Although the MDMS-SL approach takes advantage Figure 3. Enhanced Shotgun Lipidomics Approaches of the 4 orders of magnitude dynamic The analytic power of shotgun lipidomics can be extended through exploiting the unique chemical char- range inherent in mass spectrometry, we acteristics of specific lipid classes. These include the derivatization of lipid moieties to increase signal intensity and/or engender a mass shift to facilitate mass spectrometric analyses, the use of a M+1/2 iso- explicitly point out that enrichment ap- topologue approach for doubly negatively charged cardiolipins, multiplexed extractions including liquid/ proaches are necessary for examination liquid partitioning, liquid/solid partitioning (solid phase extraction), and/or alkaline hydrolysis to enrich for of extremely low abundance lipids which sphingolipids or ether lipids. In some tissues (e.g., adipose tissue), the removal of nonpolar lipids by hexane extraction is judicious prior to the subsequent analysis of polar lipid constituents. shotgun lipidomics is unable to access with the sensitivities of currently available mass spectrometers. Chromatographic analysis allowing the comprehensive iden- device thereby avoiding the need for chro- separation, liquid-liquid partitioning, solid tification and quantitation of many matography. A second advantage of phase extraction, and other enrichment hundreds to thousands of lipid molecular MDMS-SL is through direct infusion methods can greatly extend the limits of species. This strategy is also well suited large increases in S/N can be achieved detection for extremely low abundance for quantitation of individual molecular in comparison to chromatographic ap- molecular species. In our hands, chro- species through utilization of ratiometric proaches where limited amounts of time matographic enrichment of many different comparisons using a two-step approach. for signal acquisition and tandem mass classes of low abundance lipids is most This two-step procedure optimally uses spectrometric analyses are available. A effectively accomplished by orthogonal the advantages of both exogenously third benefit of MDMS-SL is that it does chromatographies employing an initial added internal standards as well as not require previous knowledge of the straight phase (or SCX) separation that re- endogenous lipid constituents as stan- compound(s) present in the mixture to solves lipid classes followed by RPHPLC dards for tandem mass spectrometry by identify transitions, but rather interrogates to affect the resolution of lipid subclasses using endogenous molecular species all mass space that contains naturally and individual molecular species within whose mass content is known with high occurring aliphatic chains, head groups, each class. Even with orthogonal chroma- precision through their intensities in the and signature fragmentation ions through tographies, multiple molecular species full mass scan. Through this tandem precursor ion and neutral loss scanning, are typically present at each m/z value mass spectrometric approach, the thereby allowing for the identification of that require analysis by tandem mass dynamic range of quantitation is extended previously unknown compounds. A fourth spectrometric approaches. A second by over two orders of magnitude and is use of MDMS-SL is the benefit from anal- limitation of MDMS-SL is the difficulty in accomplished with greater precision ysis of peak contours in multidimensional the localization and the stereochemistry through exploiting multiple additional space that lead to multiple bioinformatic of the olefinic linkages in aliphatic chains comparisons using endogenous molec- advantages that facilitate refinements in which can be accessed through multi- ular species as standards. Finally, we point quantitation through corrections for peak stage fragmentation. A third limitation of out the utility of alterations in the ionization shape, mass offset, and isotopologue MDMS-SL is the inability to discriminate conditions, ramping of collision energies, analysis. A fifth advantage of MDMS-SL enantiomeric lipids. Through chiral chro- examination of fragmentation kinetics is that it is ideally suited to exploit the matography, Blair et al. resolved eicosa- and n-dimensional scanning of informative distinctive chemical characteristics of noid enantiomers for direct mass spectro- fragment ions that each provide additional many lipid classes. Prominent examples metric analysis (Lee and Blair, 2009). dimensions to substantiate structural include the use of the M+1/2 isotopologue Whether direct infusion, chromatographic identification and quantitation of the abun- approach for cardiolipin analyses (Han enrichment, partitioning, or derivatization dance of identified lipid molecular species. et al., 2006), the use of Fmoc derivatization is used, many of the strategic advantages for the analysis of ethanolamine contain- of multidimensional mass spectrometry Advantages and Limitations ing constituents (Glaser and Gross, can be readily accessed. of Multidimensional Mass 1995), the use of specifically deuterated Spectrometry-Based Shotgun amine selective reagents for ratiometric The Future Challenges Lipidomics quantitation through precursor ion scan- of Lipidomics The obvious advantage of MDMS-SL is ning (Zemski Berry et al., 2009), alkaline The future challenges of lipidomics in- the direct acquisition of high-density infor- hydrolysis to greatly enhance penetrance clude the development of chemistries and mation on lipid structure and abundance into the sphingolipidome (Jiang et al., technologies to increase the penetrance directly from biological extracts using the 2007) and charge-switch derivatization into the lipidome, identification of the mass spectrometer as a separations methods to greatly increase the sensitivity subcellular compartments or domains

where alterations are manifest during throughput analyses with rapid repetition subcellular compartments, identify un- cellular adaptation or disease, and the rate lasers that are unencumbered by known lipids and greatly facilitate our development of bioinformatic approaches the time constraints normally encoun- understanding of membrane structure to define changes in lipid metabolic tered using fluid-based methods for and function. networks that can predict the onset of mass spectrometry. The use of MALDI/ Nuclear magnetic resonance (NMR) disease, identify disease progression, TOF-TOF mass spectrometry for lipido- spectroscopy has been used to identify and evaluate treatment efficacy. The mic analysis of biologic tissues has a wide variety of lipids in biological development of increasingly sophisti- recently been further developed for systems (e.g., Lindon and Nicholson, cated databases that contain the exact high throughput lipidomics where many 2008 and references therein). Mountford mass, fragmentation patterns, and HPLC dozens of samples can be directly and colleagues used 1H NMR to distin- elution characteristics of each lipid analyzed within minutes facilitating the guish invasive breast cancer from benign molecular species is essential for the large scale application of lipidomics lesions on the basis of increased total continued development of lipidomics and to human disease through this high choline phospholipid species (MacKinnon in its application to biomedical research. throughput approach (Sun et al., 2008). et al., 1997). While NMR is a nondestruc- Currently, there are three main databases Moreover, laser or beam ionization tive and nonselective technique that that are routinely utilized for lipidomics methods provide a platform for innovative provides unique information pertaining to research including METLIN (Smith et al., chemical strategies for matrix construc- molecular structure and dynamics, its 2005), The Human Metabolome Database tion that can be exploited for selective modest sensitivity and resolving power (Wishart et al., 2008), and LIPID MAPS ionization (e.g., the selective ionization/ to distinguish individual chemical species (Sud et al., 2007). Through the use of ESI desorption of discrete lipid classes) or in comparison to mass spectrometry has in conjunction with time of flight mass can be used for nonbiased identification limited utilization of NMR in lipidomics spectrometry, Shevchenko et al. used and quantitation of analytes. In elegant investigations. NMR interpretation is also a shotgun lipidomics strategy in conjunc- work, Siuzdak et al. engineered nano- complicated by the considerable number tion with custom software to identify structured surfaces to trap fluorinated of spin-coupled multiplets that result in over 250 lipids from yeast (Ejsing et al., siloxanes that desorb analytes upon spectral crowding. Recently this problem 2009). Dennis et al. identified over 400 laser-induced vaporization (e.g., Northen has been addressed by using two-dimen- lipids in RAW 267.2 cells using the et al., 2007). Through multiplexing differ- sional J-resolved NMR spectroscopy LIPID MAPS database and demonstrated ent surface-based technologies, doping to visualize metabolite chemical shifts alterations in multiple lipid classes and with selected cavitands, and the con- and J-couplings along different spectral molecular species after cellular stimula- struction of engineered charge transfer dimensions and thereby increase peak tion (Dennis et al., 2010). Siuzdak et al. complexes to facilitate class-specific ioni- dispersion (Ludwig and Viant, 2010). An identified over 150 lipids in stem cells zation, the future growth of MALDI TOF/ advantage of magnetic resonance during differentiation and found increases TOF mass spectrometry is assured. approaches is the ability to analyze lipids in the unsaturation index of aliphatic Moreover, recent success with MALDI noninvasively in intact cells and tissues chains that alter membrane structure imaging (Reyzer and Caprioli, 2005), without losing chemical information about and dynamics during differentiation desorption ESI (DESI) imaging (Takats the analyte environment (Beckonert et al., (Yanes et al., 2010). Collectively, these et al., 2004) techniques, and nanostruc- 2010). A review of the methods used studies demonstrate the dynamic role of tured surfaces (Patti et al., 2010) has led in high-resolution magic-angle spinning alterations in lipid molecular species to new insights into the spatial distribution NMR, their application to lipidomics and content and membrane composition of lipids in tissues, but at present these their potential biomedical uses have that serve to facilitate cellular differentia- methods are limited by constraints on been discussed recently (Fonville et al., tion and cellular activation. Through spatial resolution. A prominent issue in 2010). Moreover, selective recoupling of orchestrating changes in membrane lipid lipidomics is the identification of methods dipolar and chemical-shift interactions composition, molecular structure, and to delineate alterations in membrane removed by magic-angle spinning in the membrane dynamics the execution of composition in discrete subcellular loci solid state allows for the characterization intrinsic biologic programs (e.g., differenti- or within specific membrane domains. of regulatory interactions, dynamics, and ation) or responses to external perturba- Thus, the development of methods which ion channels within biological membranes tions necessary for life processes are precisely localize changes in specific (Guillion and Schaefer, 1989; deAzevedo facilitated. compartments or domains in intact cells et al., 1999; Kim et al., 2009; Cady et al., Many other methods for ionization of is a fundamental goal of future lipidomics 2010). Since NMR is a nondestructive lipids for mass spectrometry have been research. The use of secondary ion mass technique, it can be used to interrogate used with great success. In pioneering spectrometry using ion beams has greatly alterations in lipid structure and dynamics work, Marshall et al. demonstrated the enhanced spatial resolution to the order in biochemically functioning cells. Thus, it utility of matrix assisted laser desorption/ of 100 nm that can identify lipids in is anticipated that future developments ionization (MALDI) for lipid analysis in specific membrane domains (Winograd will provide critical information on alter- conjunction with high mass accuracy and Garrison, 2010). Additional techno- ations in membrane structure and func- using FT-ICR (Marto et al., 1995). The logic advances are anticipated in these tion through the synergistic application use of laser-based methods have the fields that will define spatially privileged of a variety of solution state and solid- obvious advantage of allowing high domains within cell membranes and state NMR approaches.

Summary K., and Shevchenko, A. (2009). Proc. Natl. Acad. raclough, B.H., Bilous, M., and Mountford, C.E. 106 204 The past decades have witnessed the Sci. USA , 2136–2141. (1997). Radiology , 661–666. rapid growth and development of lipido- Fenn, J.B., Mann, M., Meng, C.K., Wong, S.F., Mancuso, D.J., Kotzbauer, P., Wozniak, D.F., mics, which has now become an essential and Whitehouse, C.M. (1989). Science 246, Sims, H.F., Jenkins, C.M., Guan, S., Han, X., 64–71. Yang, K., Sun, G., Malik, I., et al. (2009). J. Biol. part of Systems Biology. Multidimen- Chem. 284, 35632–35644. sional mass spectrometry-based shotgun Fonville, J.M., Maher, A.D., Coen, M., Holmes, E., lipidomics has now been developed into Lindon, J.C., and Nicholson, J.K. (2010). Anal. Marto, J.A., White, F.M., Seldomridge, S., and Chem. 82, 1811–1821. Marshall, A.G. (1995). Anal. Chem. 67, 3978– a robust technology that can be used 3984. to provide rapid access to the mecha- Ford, D.A., Han, X., Horner, C.C., and Gross, R.W. (1996). Biochemistry 35, 7903–7909. Michalik, L., Auwerx, J., Berger, J.P., Chatterjee, nisms underlying diverse disease states. V.K., Glass, C.K., Gonzalez, F.J., Grimaldi, P.A., Continuing advances in ionization tech- Glaser, P.E., and Gross, R.W. (1995). Biochemistry Kadowaki, T., Lazar, M.A., O’Rahilly, S., et al. 34, 12193–12203. 58 nologies, use of high mass accuracy (2006). Pharmacol. Rev. , 726–741. lipidomics, novel fragmentation strate- Gross, R.W. (1984). Biochemistry 23, 158–165. Nomura, D.K., Long, J.Z., Niessen, S., Hoover, H.S., Ng, S.-W., and Cravatt, B.F. (2010). Cell gies, improvements in enrichment ap- 24 Gross, R.W. (1985). Biochemistry , 1662–1668. 140, 49–61. proaches, and enhanced spatial resolution in conjunction with NMR-based Gubitosi-Klug, R.A., Yu, S.P., Choi, D.W., and Northen, T.R., Yanes, O., Northen, M.T., Marri- Gross, R.W. (1995). J. Biol. Chem. 270, 2885– approaches will collectively lead to the nucci, D., Uritboonthai, W., Apon, J., Golledge, 2888. S.L., Nordstrom, A., and Suizdak, G. (2007). Nature development of a comprehensive under- 449, 1033–1036. Guillion, T., and Schaefer, J. (1989). J. Magn. standing of the pleiotropic roles of lipids 81 Reson. , 196. Patti, G.J., Shriver, L.P., Wassif, C.A., Woo, H.K., in cellular function. Uritboonthal, W., Apon, J., Manchester, M., Porter, Chemistry & Biology invites your Han, X., and Gross, R.W. (1994). Proc. Natl. Acad. F.D., and Siuzdak, G. (2010). Neuroscience 170, Sci. USA 91, 10635–10639. comments on this topic. Please write 858–864. Han, X., and Gross, R.W. (1995). J. Am. Soc. Mass to the editors at [email protected]. Porter, F.D., Scherrer, D.E., Lanier, M.H., Lang- 6, 1202–1210. made, S.J., Molugu, V., Gale, S.E., Olzeski, D., Sidhu, R., Dietzen, D.J., Fu, R., et al. (2010). ACKNOWLEDGMENTS Han, X., and Gross, R.W. (1996). J. Am. Chem. Cholesterol oxidation products are sensitive Soc. 118, 451–457. and speciﬁc blood-based biomarkers for Nie- 2 This work was supported by the National Institutes Han, X., and Gross, R.W. (2001). Anal. Biochem. mann-Pick C1 disease. Sci. Transl. Med. , of Health Grants PO1HL57278 (R.W.G.), 295, 88–100. 56ra81. RO1HL41250 (R.W.G.), and RO1-AG31675 (X.H.). Han, X., and Gross, R.W. (2005). Mass Spectrom. Reyzer, M.L., and Caprioli, R.M. (2005). J. Pro- 4 Rev. 24, 367–412. teome Res. , 1138–1142. REFERENCES Han, X., Yang, J., Cheng, H., Ye, H., and Gross, Schmidt, D., and MacKinnon, R. (2008). Proc. Natl. 105 Bartz, R., Li, W.-H., Venables, B., Zehmer, J.K., R.W. (2004). Anal. Biochem. 330, 317–331. Acad. Sci. USA , 19276–19281. Roth, M.R., Welti, R., Anderson, R.G., Liu, P., and 441 Chapman, K.D. (2007). J. Lipid Res. 48, 837–847. Han, X., Yang, K., Yang, J., Cheng, H., and Gross, Shaw, J., and Cantley, L.C. (2006). Nature , R.W. (2006). J. Lipid Res. 47, 864–879. 424–430. Beckonert, O., Coe, M., Keun, H.C., Wang, Y., Ebbels, T.M.D., Holmes, E., Lindon, J.C., and Han, X., Yang, J., Yang, K., Zhao, Z., Abendschein, Shevchenko, A., and Simons, K. (2010). Nat. Rev. 11 Nicholson, J.K. (2010). Nat. Protoc. 5, 1019–1032. D.R., and Gross, R.W. (2007). Biochemistry 46, Mol. Cell Biol. , 593–598. 6417–6428. Bollinger, J.G., Thompson, W., Lai, Y., Oslund, Simon, G.M., and Cravatt, B.F. (2006). J. Biol. 281 R.C., Hallstrand, T.S., Sadilek, M., Turecek, F., Hazen, S.L., Hall, C.R., Ford, D.A., and Gross, R.W. Chem. , 26465–26472. and Gelb, M.H. (2010). Anal. Chem. 82, 6790– (1993). J. Clin. Invest. 91, 2513–2522. 50 6796. Skoura, A., and Hla, T. (2009). J. Lipid Res. , Jiang, X., Cheng, H., Yang, K., Gross, R.W., and S293–S298. Breen, E.P., Gouin, S.G., Murphy, A.F., Haines, Han, X. (2007). Anal. Biochem. 371, 135–145. L.R., Jackson, A.M., Pearson, T.W., Murphy, Smith, C.A., O’Maille, G., Want, E.J., Qin, C., P.V., and Porter, R.K. (2005). J. Biol. Chem. 281, Kim, S.J., Singh, M., and Schaefer, J. (2009). J. Trauger, S.A., Brandon, T.R., Custodio, D.E., 2114–2119. Mol. Biol. 391, 414–425. Abagyan, R., and Siuzdak, G. (2005).

Cady, S.D., Schmidt-Rohr, K., Wang, J., Soto, Klose, C., Ejsing, C.S., Garcia-Saez, A.J., Kalser, Su, X., Han, X., Yang, J., Mancuso, D.J., Chen, J., C.S., DeGrado, W.F., and Hong, M. (2010). Nature H.J., Surma, M.A., Shevchenko, A., Schwille, P., Bickel, P.E., and Gross, R.W. (2004). Biochemistry 43 463, 689–693. and Simons, K. (2010). J. Biol. Chem. 285, , 5033–5044. 30224–30232. Chiang, K.P., Niessen, S., Saghatelian, A., and Sud, M., Fahy, E., Cotter, D., Brown, A., Dennis, Cravatt, B.F. (2006). Chem. Biol. 13, 1041–1050. Lee, S.H., and Blair, I.A. (2009). BMB Reports 42, E.A., Glass, C.K., Merrill, A.H., Jr., Murphy, R.C., 401–410. Raetz, C.R.H., Russell, D.W., and Subramaniam, deAzevedo, E.R., Hu, W.-G., Bonagamba, T.J., S. (2007). Nucleic Acids Res. 35, D527–D532. and Schmidt-Rohr, K. (1999). J. Am. Chem. Soc. Lindon, J.C., and Nicholson, J.K. (2008). Annu. http://www.lipidmaps.org/. 11, 8411–8412. Rev. Anal. Chem. 1, 45–69. Sun, G., Yang, K., Zhao, Z., Guan, S., Han, X., and Dennis, E.A., Deems, R.A., Harkewicz, R., Quehen- Liu, Y., Chen, Y., Momin, A., Shaner, R., Wang, E., Gross, R.W. (2008). Anal. Chem. 80, 7576–7585. berger, O., Brown, H.A., Milne, S.B., Myers, D.S., Bowen, N.J., Matyunina, L.V., DeEtte Walker, L., Glass, C.K., Hardiman, G., Reichart, D., et al. McDonald, J.F., Sullards, M.C., and Merrill, A.H., Takats, Z., Wiseman, J.M., Gologan, B., and (2010). J. Biol. Chem. 285, 39976–39985. Jr. (2010). Mol. Cancer 9, 186–199. Cooks, R.G. (2004). Science 306, 471–473.

Dixon, J.B. (2010). Mol. Cell. Endocrinol. 316, Ludwig, C., and Viant, M.R. (2010). Phytochem. Unger, R.H. (2002). Annu. Rev. Med. 53, 319–336. 104–108. Anal. 21, 22–32. Unger, R.H., Clark, G.O., Scherer, P.E., and Ejsing, C.S., Sampaio, J.L., Surendranath, V., MacKinnon, W.B., Barry, P.A., Malycha, P.L., Gil- Orci, L. (2010). Biochim. Biophys. Acta 1801, Duchoslav, E., Ekroos, K., Klemm, R.W., Simons, lett, D.J., Russell, P., Lean, C.L., Doran, S.T., Bar- 209–214.

van Meer, G. (2005). EMBO J. 24, 3159–3165. Wishart, D.S., Knox, C., Guo, A.C., Eisner, R., Desponts, C., Ding, S., and Siuzdak, G. (2010). Young, N., Gautam, B., Hau, D.D., Psychogios, Nat. Chem. Biol. 6, 411–417. Vegiopoulos, A., Muller-Decker, K., Strzoda, D., N., Dong, E., Bouatra, S., et al. (2008). Nucleic Schmitt, I., Chichelnitskiy, E., Ostertag, A., Berriel 37 Acids Res. , D603–D610. http://www.hmdb.ca/. Zemski Berry, K.A., Turner, W.W., VanNieu- Diaz, M., Rozman, J., Hrabe de Angelis, M., wenhze, M.S., and Murphy, R.C. (2009). Anal. Nusing, R.M., et al. (2010). Science 328, 1158– Wolf, R.A., and Gross, R.W. (1985). J. Biol. Chem. Chem. 81, 6633–6640. 1161. 260, 7295–7303.

Winograd, N., and Garrison, B.J. (2010). Annu. Yanes, O., Clark, J., Wong, D.M., Patti, G.J., Zhang, M., Mileykovskaya, E., and Dowhan, W. Rev. Phys. Chem. 61, 305–322. Sanchez-Ruiz, A., Benton, H.P., Trauger, S.A., (2002). J. Biol. Chem. 277, 43553–43556.