Research Plan

I. Program Summary
  A. Introduction
  B. Background and Broad Context
  C. Significance for Biomedical Science and Disease
  D. Focus Areas
  E. Lipidomics Focus Area
    1. Fatty Acids / Eicosanoids Core
    2. Neutral Lipids Core
    3. Glycerophospholipids Core
    4. Sphingolipids/ Glycosphigolipids Core
    5. Sterols Core
    6. Isoprenoids and Structural Lipidomics Core
  F. Introduction to Cell Biology Focus Area
  G. Introduction to Lipid Detection and Quantitation Focus Area
  H. Introduction to Lipid Synthesis and Characterization Focus Area
  J. Introduction to Informatics Focus Area
  K. Introduction to Administration Focus Area

II. Administrative Management Plan
  A. Structure of LIPID MAPS Organization
  

I. PROGRAM SUMMARY

A. INTRODUCTION

Sequencing of the human genome has opened the way and provided the impetus for building a comprehensive picture of a mammalian cell. Significant efforts are underway in the fields of genomics and proteomics to identify all genes and proteins in a given organism. The goal is a complete map of the genes, gene products and their interaction networks in a functioning cell. The next step in establishing a comprehensive picture of a cell will be to tie the cell's metabolome into the rapidly developing genomic and proteomic maps. A cell's metabolome, however, is such an enormous and complex entity that characterizing it can only be approached in sections. This consortium is now proposing to focus on the lipid section of the metabolome by developing an integrated metabolomic system capable of characterizing the global changes in lipid metabolites ("lipidomics"). Our consortium has developed a Lipid Metabolites and Pathways Strategy, termed LIPID MAPS, that applies a global integrated approach to the study of lipidomics.

Only recently has it been widely recognized by the scientific community that lipids are indeed central to the regulation and control of cellular function and to disease. This insight has greatly enhanced our appreciation of the importance of lipids. Comprising an enormous fraction of cellular metabolites, lipids include such diverse classes as fatty acids, eicosanoids, phospholipids, neutral lipids, sphingolipids, glycosphingolipids, sterols, glycolipids, etc. Because of similarities in the chemical characteristics of this broad spectrum of compounds, the lipid metabolic pathways thus far identified are very complex and often a given lipid can shuttle between a number of different pathways. Changes in the level of one member of a lipid class not only affect the other members of that class, but can also strongly affect members of other classes as well. While investigators have been able to monitor concurrent changes in a few members of a lipid class, no comprehensive strategy has emerged for characterizing the global changes in lipid metabolites. LIPID MAPS fills this gap.

The aim of LIPID MAPS for this grant period is to develop the requisite technology and conduct an integrated research program that will establish lipidomics as a fully developed research field. To accomplish this, we have developed a strategy to identify and quantitate all of the lipids in a cell. The critical first step in achieving this goal will be to set a realistic scope for these studies. Because of the complexity of metabolomics and even lipidomics, we will focus on a single cell type. We have chosen to study macrophages because they carry out many of the functions that are known to be related to lipid metabolism including the secretion of many active lipid second messengers. Macrophages also possess many other characteristics that will aid in tackling the issues presented by this project. These are discussed in detail in Section F below. In addition, all members of the consortium will focus on just one or two stimuli at a time.

A second critical factor is that we will employ a single detection method, mass spectrometry, to qualitatively and quantitatively analyze all lipids. Liquid chromatography/mass spectroscopy (LC/MS) is probably the only analytical technique that, by itself, can separate, identify, and quantitate the vast number of closely related compounds that comprise the lipids as discussed in detail in Section G below.

An important element in LIPID MAPS will be to locate and identify new, minor lipid compounds. While almost all of the major lipids of cells have surely been identified, many minor lipids are still being found. Interestingly, these minor lipids are the ones that appear to have the most potent and profound biological activities and are most often the ones associated with diseases. The LC/MS system is again ideally suited for finding these new minor lipids, even during the course of quantitating the major ones.

We will also employ the tools of genomics and proteomics to identify, characterize and map all of the enzymes and other proteins involved in lipid metabolism and integrate them into the maps and networks of lipidomics.

We have brought together an outstanding group of scientists who will operate ten cores and five bridges in close collaboration. They will employ a common set of biochemical and cell culture protocols that will insure that reproducible experiments will be carried out in each of the consortium laboratories and that the data collected by one unit will be directly comparable to data collected by another. Six of these cores will develop the techniques to quantitate a given class of lipids. To increase the reproducibility and to insure that data from different lipid classes can be compared, each of these cores will be furnished with an identical LC/MS system. Thus, for example, the Glycerophospholipid Core will also be able to quantitate the level of sterols in a given experiment in their laboratory by employing the identical techniques developed in the Sterol Core. This will allow the cross checking of data. The data thus generated will be collected through a common laboratory information management system, will be stored in common databases, and will be analyzed by an Informatics Focus Area, as discussed in Section I below.

The Lipidomics Cores, as discussed in Section E below, will not operate in exactly the same way as traditional "cores", i.e. in providing a narrowly defined function to outside researchers. The duties of each core are to develop the techniques required to quantitate their specific class of lipids, to quantitate the levels and changes in levels of all of these lipids in the macrophages under the conditions set forward by the consortium, and to supply this mass of data to the Bioinformatics and Data Coordination Core for the "new global approach to analysis". However, an equally important duty will be to follow any hypothesis driven leads that they may discover in the course of these experiments. They will interact directly with any other core that is appropriate to test these hypothesis.

LIPID MAPS, by measuring the changes in both major and minor lipid metabolites in response to stimuli and disease and by developing a map of the routes these lipids take in cells, will provide the scientific community with a vision into the coordinated responses of cellular lipids at a level of sensitivity and complexity never before possible. By sharing these findings with the entire community a much broader effort will be possible than through our consortium alone.

The goal of the first five year period is to determine the complete lipidome of the mouse macrophage. Beyond the initial 5 year effort, the long term objective is to record the changes in the lipidome as the macrophage responds to various stimuli and to continue building a database of results and query tools for the LIPID MAPS. At the same time, studies of both human and mouse differentiated tissue macrophages employing these techniques would begin. All of this will be made available to the entire life sciences community and will establish this new global, discovery driven approach in the lipid field that will complement traditional hypothesis driven research approaches.

B. BACKGROUND AND BROAD CONTEXT

Research in the biochemical and medical communities has entered a new era; one that is characterized by the generation of a tremendous amount of data. The volume of this data derives from the use of high throughput technology by various genomic and proteomic studies. It also is due in part to the emerging use of universal systems wide approaches to cell biology in lieu of the traditional hypothesis driven approaches. The new discovery approach focuses on measuring experimental data from a variety of sources, methods, and scales and on measuring as many components of the system as possible, without regard to suspected relevance. Dealing with this volume of data has driven the emergence of the bioinformatics field, which has now developed the tools to collect, store, retrieve, organize and analyze this data. To date, the bulk of this revolution has been centered around genomics and proteomics. The last corner of the triad required to completely describe a cell remains to be tackled, metabolomics. The intellectual challenge for biomedicine in moving this information from raw data to organized, comprehensive knowledge has been variously termed Beyond the Genome, The New Biology, Bringing Genomes to Life, or 21st Century Biology. How challenging is this task?

Over the last decade, various genome projects had to determine a linear sequence with only four nucleotides, and the success of these projects has been a major scientific coup. Proteomics presents a much more challenging task, since the complete set of proteins, built by the combination of twenty amino acids and their posttranslational modifications, must be characterized in different states of development, aging, health, and disease and in response to drugs and other stimuli. Besides differential splicing, the proteome component of cells depends on numerous temporal changes in protein composition and cellular conditions.

Even in comparison to proteomics, metabolomics presents a simply staggering challenge. Genes and proteins species each contain a single chemical type. Cellular metabolites represent a diverse range of structural moieties for which a single separation and identification system is unrealistic. In addition, the metabolic profile of a cell changes from second to second. The levels of metabolites, as well as their structure, are in a constant flux that is fundamental to metabolism and life. The levels of the metabolites vary by orders of magnitude. Developing the tools for quantitative analyses are daunting tasks, and will require a variety of separation and analysis systems reflecting the diversity of structural types of metabolites. Collecting, organizing and analyzing this data will be equally daunting. Correlating this data with that of genomics and proteomics will be the ultimate challenge.

Clearly to have any chance at generating a cellular metabolome, it must be divided into more manageable units. The LIPID MAPS Consortium feels that lipids are an advantageous class of molecules with which to start this move toward establishing a complete metabolome of a cell. This class of molecules can be easily separated from other functional classes by simple organic extraction of cells. Because of their organic solubility, good separation techniques exist. A further advantage is that most classes of lipids are metabolically related and result from anabolic and catabolic addition or subtraction of acetyl CoA units of the fatty acid or isoprenoid derived components, which dominate practically all lipids. Thus, we will approach the broader metabolomics issues by focusing on lipid metabolites. We will establish a core for each of the major functional subclasses of lipid metabolites with the goal of identifying and quantitatively monitoring all changes in metabolites by developing new techniques for detection and analysis, and by a systematic analysis on the massive amount of data generated.

We will initially focus on the monocyte/macrophage system, a cell that has been widely studied regarding lipid metabolism. After characterizing the lipid metabolites in the monocyte, there would be the potential to follow changes during the progression to peritoneal, alveolar, microglial, Kupfer cells and other tissue macrophages, and also their progression to foam cells and other disease implicated transformations. We could initially focus on mouse and human cells, both primary cells (which are readily obtainable) as well as established cell lines that have been well studied.

Another strong case for beginning a fully comprehensive characterization of the functional roles of lipids is that existing NIH Centers and Programs include extensive analyses of nucleic acids, proteins and carbohydrates, but lipids remain such an experimental challenge for structural analysis that no large-scale collaboration for their study now exists. A complete picture of the eukaryotic cell and its integration into organismic physiology must include an understanding of the role of lipids in macromolecular machines and supramolecular processes. We believe that the experimental technologies are in hand for a functional characterization of LIPID MAPS and that this functional characterization can be coupled with our understanding of the other macromolecules. Then, by means of computational tools, we and the scientific community as a whole will be able to establish predictive models for homeostasis and drug response in disease states.

C. SIGNIFICANCE FOR BIOMEDICAL SCIENCE AND DISEASE

Lipids play a central role in so many diseases that it is clear that characterizing their metabolic changes with disease and disease models will help in understanding the relationships among various lipid classes. One example, among the lipid agents active as second messengers, is the eicosanoids. The eicosanoids are a family of oxygenated derivatives of 20-carbon polyunsaturated fatty acids that potently mediate a wide variety of physiological and pathophysiological processes. Eicosanoid production by macrophages and macrophage-like cells plays a central role in many inflammatory diseases including rheumatoid arthritis, sepsis, asthma, and inflammatory bowel disease. Other diseases in which eicosanoids play a role include atherosclerosis, Alzheimer's disease, cancer, and stroke. Transient, localized increases in the tissue concentration of eicosanoids influence differentiation, migration and activation of cells in immunity and other integrated physiological responses. A large number of diseases are influenced by the regulation of eicosanoid production. Thus, tracking the metabolism of fatty acids and eicosanoids must play a central role in any comprehensive characterization of lipid metabolism and should identify numerous opportunities for drug intervention in treating these diseases.

To address impacts further upstream, it is essential to monitor changes in phospholipids and sphingolipids as well. Of course, there may be new, yet undiscovered novel lipids that play a role. Of special importance, fatty acids are esterified to cholesterol to form cholesterol esters (a major form of cholesterol in LDL), so this can influence sterol metabolism. How the body gets rid of excess cholesterol and its catabolic regulation are important because controlled elimination is one of the body's chief defenses against cholesterol accumulation and consequent heart attacks. For example, inherited conditions that prevent the synthesis of bile acids can cause accumulation. The integrated approach of monitoring all lipid metabolites under altered metabolic conditions and disease should yield new paradigms for understanding all sorts of disease processes.

D. FOCUS AREAS

Each major activity in the LIPID MAPS Consortium is termed a Focus Area; individual contributing research efforts are comprised of various cores and bridges; see outline below. In subsequent sections, each Focus Area will be discussed. We will provide some details of the collaborative approach we are taking to tackle this scientific problem. We also touch on the scientific expertise each Core/Bridge Director brings to the LIPID MAPS Consortium.

The order of the Focus Areas in the above list is dictated in large part by the RFA design for Glue grants and in the Administrative Management Plan (Section II below), the Focus Areas, Cores and Bridges will be presented in the above order. However, here we will summarize the Focus Areas in a different order dictated more by the scientific thrust of the grant application that has as its centerpiece the Lipidomics Focus Area. We will first discuss Focus Area VI, which consists of six Lipidomics cores that divides the major lipid classes into six groups, see Figure 1. While these groups constitute the classically recognized major groupings of lipids, they also allow us to cover the landscape of lipids. Each core will do the basic analysis of the requisite class of lipids, but also will actively participate in translating and utilizing the results of their own and other cores with application to macrophage biology. This is discussed in Section E. The effectiveness of the lipidomics thrust will be made possible by a uniformity of cellular approaches and integration with functional genomics and contextual proteomics spearheaded by the efforts of Focus Area III., Cell Biology, as described in Section F. The contributions of Focus Area IV, Lipid Detection and Quantitation, will be critical for the Lipidomics cores to achieve their goals and collect useful and comparative data as discussed in Section G. Focus Area V, Lipid Synthesis and Characterization, will enable the Lipidomics cores to carry out their mission by providing standards but also by confirming and extending the utility and understanding of the properties of new lipids discovered in the Lipidomics cores as discussed in Section H. Focus Area II, Informatics, has the challenge of pulling together all of the experimental input and results of the Lipidomics cores and of developing new networks and LIPID MAPS, and most importantly, disseminating them. This will be discussed in Section I. Of course, Focus Area I., Administration, plays a key role in coordinating all of the Focus Areas and efforts, in providing the infrastructure for success, managing the budgetary process and intellectual property issues, in organizing the Advisory, Steering, and Operating Committees, and ensuring full dissemination of the new data as referred to in Section J.

E. INTRODUCTION TO LIPIDOMICS FOCUS AREA

The Lipidomics Focus Area is organized around six cores and two bridges. The thrust of each core will be discussed in turn. The Oxidized Lipids in Macrophages Bridge will characterize various structurally defined oxidized lipids with the goal of determining which would be the appropriate stimulus to employ by the other Lipidomics Cores when that important stimulus is explored. The Lipid Subcellular Localization Bridge will develop methodology for the Lipidomics Cores that will allow them to determine the subcellular distribution of the various lipid classes in the macrophage and then to follow changes in these lipid pools as the cells are stimulated.

1. Fatty Acids / Eicosanoids Core

This core will be directed by Dr. Edward A. Dennis (University of California, San Diego). The complexity of LIPID MAPS is clearly demonstrated by examining the metabolism of fatty acids. Once thought to simply be the building blocks of cellular membranes, phospholipids and their metabolites have been found to be very potent second messengers. As outlined in Section C above concerning their role in disease processes, eicosanoids are particularly notable in this role. Eicosanoids are not stored by cells, but are synthesized in response to cell-specific stimuli. These stimuli activate phospholipase A2s (PLA2s) that in turn increases the concentration of free 20-carbon polyunsaturated fatty acids. The majority of the eicosanoids produced by macrophages are derived specifically from free arachidonic acid (AA). Once freed, AA is converted into over 50 well-known active species, which demonstrates one of the many aspects of the complexity of lipidomics. A second level of complexity stems from the fact that while AA contributes a majority of the major eicosanoids, other polyunsaturated 20-carbon fatty acids also yield related compounds when acted upon by these enzymes.

Although all eicosanoid biosynthesis begins with the release of fatty acid from phospholipids by a PLA2, there are fourteen different classes of PLA2 that could release AA. Most cells express several of these PLA2s. To completely map these systems would require the tracking of the specific fatty acids down to subcellular localizations. Since the eicosanoids are all second messengers, their biosynthesis is governed by the interaction of various agents with cellular receptors. The Dennis laboratory has found numerous activation pathways in the P388D1 macrophage-like cell line that lead to significant PGE2 release. Most likely, other pathways exist that lead to the release of other eicosanoids. Four pathways are relevant to normal functions of macrophages, including responses to inflammation and acute injury. The first is mobilization and PGE2 production involving sequential exposure of the cells to two different stimuli, namely LPS and PAF. The second pathway involves the exposure of macrophages to LPS for long periods, up to 18 hours, which also leads to AA release and PGE2 production. The third class involves responses to zymosan as in yeast infections. In the fourth pathway, diacylglycerol pyrophosphate (DGPP), which is not found in mammals, activates P388D cells; DGPP may elicit a defensive response by mammalian macrophages to the invasion of bacteria or yeast.

The metabolism of fatty acids and the eicosanoid cascade clearly represent a very complex intertwined metabolic system. Furthermore, it is important to know which phospholipid pools contribute the AA and which lysophospholipids result. This will be determined by the Glycerophospholipid Core (Section 3, below). Current experiments tend to follow a single metabolite of this system, usually either AA or PGE2. To understand how this system really works, to understand how the fatty acids flow through these pathways, and to understand how the various stimuli affect this system, this core will quantitate and monitor all fatty acid derived metabolites throughout the course of a given stimulus response.

Severe limitations (in sensitivity and cost) face the use of most commonly employed techniques for detecting fatty acid metabolites via radiolabeling and enzyme immunoassays (EIA), or via the use of monoclonal antibodies specific for each metabolite with subsequent EIA for quantification. Thus, the only currently available technology for this detailed definition of the cascade is the combined use of liquid chromatography and mass spectrometry, by which we will monitor a majority of the fatty acid metabolites in any given experiment without the need for radiolabels or monoclonal antibodies. Subcellular fractionation of the organelles and compositional analysis will allow us to map changes and determine if individual pools of phospholipids are responsible for a given reaction, and which metabolites are produced from each pool.

2. Neutral Lipids Core

This core will be directed by Dr. Robert Murphy (University of Colorado/National Jewish). The neutral lipids include mono, di, and triacylglycerols as well as cholesterol esters. Neutral lipids such as esters of glycerol and cholesterol as well as cholesterol itself are abundant lipid species that are a part of the complex biological systems of lipids present in all cells. These molecules play important roles in normal cellular homeostasis as a storage medium for metabolic energy but are also biochemical precursors and intermediates from which a vast array of other lipids are derived. Some neutral lipids, such as diacylglycerol, play a critical role in signal transduction events. The regulation of triacylglycerol (TAG) content within cells is quite complex and has been the focus of intense study for over 50 years. Attention in the degradation as well as synthesis of glyceryl and cholesterol esters are known to be intimately involved in a large number of pathological processes including obesity, arteriosclerosis, diabetes and alcohol-related diseases, to name just a few. Alteration of cellular status through events of cellular activation and energy demands will undoubtedly induce alteration in neutral lipid composition and likely do so in unpredicted ways.

Some of these complex lipids are quite abundant in the cell; others (eg. DAG) less so, but one common feature is that lipids are constituted by a large number of closely related family members which are termed molecular species. Only recently have specific molecular species within any individual class of neutral lipids been studied in detail, largely because of the lack of changes. Specifically, the analysis of TAG, in the past and even present, has been dominated by approaches which degrade these lipid esters to their constituent fatty acids in order to enable analysis by gas phase techniques such as gas chromatography. The emergence of electro spray and matrix assisted laser desorption ionization techniques now make it possible to analyze these neutral lipids both qualitatively, as to the precise fatty acyl groups esterified, as well as quantitatively even while present in complex mixtures. There is now a new era emerging in neutral lipid analysis where it is possible to study molecular species variation as a consequence of cellular activation.

Since many neutral lipids are present at relatively high concentrations within a cellular membrane or lipid body, the question is not the need to develop or employ highly sensitive techniques, but the challenge is in the development of both qualitative and quantitative tools to measure both relative and absolute changes in a large number of very closely related molecular species. In some instances molecular species are isobaric yet not isomeric which requires a new level of sophistication in glyceryl ester analysis. The new techniques of tandem mass spectrometer, and collisional activation of ions offer considerable promise in dealing with these challenges. The overall goal of this Neutral Lipids Core is to develop robust protocols that can be employed to follow alteration in both abundant neutral lipids, as well as measure the less abundant neutral lipid mediators as a function of cellular status and do so as intact molecules. This information can then be used to understand more fully the integration of the complex biochemical steps involved in lipid ester biosynthesis as well as degradation.

3. Glycerophospholipids Core

This core will be directed by Dr. H. Alex Brown (Vanderbilt University). The plasma membrane of eukaryotic cells is composed of a glycerolipid bilayer of heterogeneous composition and immense chemical complexity. Dynamic changes in membrane lipid composition profoundly affect cell function by providing a rich source of potential cellular messengers and by modulating protein-protein interactions beneath the cell surface. The major glycerophospholipids (GPL) found in membranes of mammalian cells include phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylinositol (PI), phosphatidylserine (PS) and phosphatidylglycerol (PG). These glycerophospholipids also exist as plasmalogens and ether lipids; as multiply phosphorylated species (e.g. poly-PIPs); and with a diverse and specialized array of alkyl chains. Phospholipid metabolism is regulated by cell surface receptors (e.g. G-protein coupled receptors, growth factors, and integrins) in many ways and at many levels: generation of first and second messengers (IP3, PIP2, lysoPA, diacylglycerol, sphingosine, etc.); modifications associated with secretion, migration, plasma membrane shape change; and so far poorly described changes in bilayer structure are needed to regulate the activities of enzymes, channels and transport proteins. As such, multifunctional entities are involved, virtually all areas of biology will benefit from advances in lipidomics.

The Glycerophospholipid Core will be responsible for determining both basal and signaling induced changes in macrophage phospholipids composition. These changes will have implications to both metabolic and signaling pathways. For example, phosphatidic acid is an important second messenger in many cell types, but this lipid also functions as a key intermediate at the branch point of glycerolipid biosynthesis. These intersecting pathways of GPL headgroups, diacylglycerol, and triacylglycerol biosynthesis have been elucidated in pioneering work by Kennedy and other investigators over the latter half of the last century. Yet a complete understanding of how specific changes of a given species influence the global membrane lipid composition of the cell membrane is lacking. As important substrates for (and products of) lipid signaling enzymes (e.g. PLA2, PLC, PLD, PI-3 kinase and lipid phosphatases), GPL lipids directly modulate pathways of cell proliferation, cytoskeletal arrangement, migration, membrane transport (e.g. endocytosis and phagocytosis) as well as many other essential processes. In some cases the in vivo substrates from which important signaling lipids such as phosphatidic acid and lyso-PA are generated have not been unambiguously identified. Using an ESI-MS approach, the Brown Laboratory has identified greater than 150 phospholipids in mammalian cells and described how they change as a result of cell surface receptor stimulation. However, the analysis has been qualitative in nature. A central goal of the LIPID MAPS consortium is to develop a high precision/ high accuracy quantitative analysis procedure for membrane lipids, including changes in GPL composition using stable isotope dilution approaches. Such advances will facilitate the construction of Lipid Arrays, that is, the computational analysis of changes in specific acyl and head group lipid species in response to biological stimuli. The ways in which the relative concentrations of these lipids change in response to cellular signaling potentially provides an extraordinarily rich source of data for classifying cellular metabolic status in addition to evaluating the effects and interactions of incoming signaling information. In addition, changes in lipid composition will implicate the identities (and eventually help to determine subcellular localizations) of the metabolic enzymes involved.

The Glycerophospholipid Core will have collaborative interactions with each of the Lipidomics cores within the consortium. One of the reasons for choosing to acquire identical instrumentation for each of the cores is to standardize the analytical procedures and develop a strategic approach that will make comprehensive analysis of cellular lipids a reality. This will also facilitate interactions between the cores. Changes in the cellular composition of phospholipids alter the biosynthetic pathways involving neutral lipids (i.e. DAG) and free fatty acids. Phospholipids serve as precursors for some glycolipids and reciprocally phospholipid degradation is regulated by sphingosine derived compounds through modulation of lipid hydrolase activities. Similarly, the substrates from which many eicosanoids and free fatty acid are generated are frequently membrane phospholipids. Establishment of quantitative analysis and unambiguous identification of changes in cellular lipids in response to important biological stimuli (e.g. LPS stimulation) are an essential step in establishing substrate-product relationships in lipid signaling and metabolic pathways. A recent focus of the Brown Laboratory has been to resolve and measure changes in polyphosphate phosphatidylinositol containing lipids (i.e. PI 3,4P2, PI 4,5P2, and PI 3,4,5P3) using a combination of liquid chromatography and mass spectrometric techniques. The ability to make mass measurements of poly-PIPs and identify the acyl composition of these lipid species represents a significant advance in the field of lipid signaling and will facilitate the goal of defining the roles of bioactive lipids in the macrophage. It is highly probable that this improved quantitative analysis and sensitivity in detection will lead to discovery of novel lipid species and illumination of previously unappreciated roles for lipids in human diseases.

4. Sphingolipids/Glycosphingolipids Core

This core will be directed by Dr. Alfred H. Merrill (Georgia Institute of Technology). Sphingolipids are composed of a "ceramide" backbone and headgroup, each of which has dozens to hundreds of structural variants with varying degrees of complexity (for example, macrophages contain ceramides with mainly a sphingosine backbone and several amide-linked fatty acids, which are present as free ceramides, sphingomyelins, glucosylceramides, lactosylceramides, ganglioside GM3's and more complex glycoconjugates). The functions of sphingolipids are also complex, encompassing: interaction with proteins at the cell surface to mediate cell-cell and cell-substratum recognition and to modify receptor behavior; influence on membrane dynamics and the properties of proteins associated with structures often referred to as "rafts" and caveolae (which include receptors and transporters); and, mediation of cell signaling through production of bioactive lipid backbones (ceramide, sphingosine and sphingosine 1-phosphate, inter alia) via both sphingolipid turnover and de novo biosynthesis. To add to the functional complexity, two closely related metabolites (such as sphingosine and sphingosine 1-phosphate) can have opposite regulatory effects (in this case, the former often induces apoptosis whereas the latter can stimulate proliferation); furthermore, changes In sphingolipid metabolism can affect other lipid signaling pathways (for example, ceramides and diacylglycerols are interrelated through the pathway Cer + PC -> SM + DAG, and vice versa). Therefore, to understand the roles of sphingolipids in cell regulation, one must analyze a large number of species with the type of sensitivity and structural resolution that is provided almost exclusively by mass spectrometry.

Many of the sphingolipids of macrophages and other mammalian cell types can be analyzed by recently developed methods using liquid chromatography and electrospray ionization tandem mass spectrometry (LC ESI-MS/MS) with the appropriate internal standards. Furthermore, the Merrill laboratory has applied LC ESI-MS/MS to allow analysis of trace species such as sphingoid bases, lysosphingolipids (including sphingoid base 1-phosphates) and N-methylsphingoid bases using small samples (typically 106 cells) as well as sphingomyelins and neutral glycosphingolipids with two to three carbohydrates. The sensitive analysis of these as well as more complex glycosphingolipids of macrophages can be achieved using two additional instruments: a higher resolution Quadrupole time-of-flight mass spectrometer (the Q-STAR) and an ion trapping instrument (the Q-TRAP) to analyze connectivity during fragmentation of higher order glycoconjugates. Using these instruments, in combination with other biochemical and genetic tools, it will become possible to dissect the complex interrelationships among sphingolipids and their connection to other lipid metabolic (and signaling) pathways.

5. Sterols Core

This core will be directed by Dr. David Russell (UT Southwestern Medical Center School). Sterols, glycerophospholipids, and sphingolipids, represent the three major classes of membrane lipids in mammalian cells. These molecules of very different structure work together to surround cells and their organelles with a semi-permeable barrier that functions as a scaffold for signaling, transport and recognition proteins. Cholesterol and other sterols impart rigidity and they control the microstructure of subdomains within the membrane. The last several decades have seen great progress towards understanding the pathways of sterol synthesis and catabolism that cells use to regulate the levels of this essential but toxic class of membrane lipids. How sterol homeostasis is maintained is of medical as well as scientific importance as disregulation of this process can lead to sterol accumulation and consequent heart attacks and strokes.

All mammalian cells synthesize sterols via a pathway that involves at least 18 enzymes. This pathway also is responsible for the synthesis of biologically active intermediates, in particular the isoprenoids, which play important roles in the biosynthesis of heme, quinones, dolichols, and isoprenylated proteins such as small G proteins. Output from the pathway is regulated by a feedback mechanism that controls the activity of sterol regulatory element binding proteins (SREBPs), which are transcription factors required for the expression of a majority of genes in the cholesterol biosynthetic pathway and many in the fatty acid biosynthetic pathways.

Like synthesis, cholesterol catabolism involves at least 17 enzymes located in each of the major subcellular compartments of hepatocytes as elucidated by the Russell laboratory. Several of these enzymes are expressed in non-hepatic tissues and cells, including macrophages, demonstrating that cholesterol is actively metabolized in many cell types. Unlike synthesis, cholesterol catabolism is regulated by members of the nuclear receptor family of transcription factors via both feedback and feed-forward mechanisms. Several of these receptors are modulated by sterols that play multiple roles in the cell. For example, oxysterols are ligands for the liver X receptor (LXR), substrates for cholesterol catabolism, regulators of SREBPs, and vehicles of sterol transport and secretion.

Many of the pleiotrophic roles of sterols in mammalian cells are recently described, suggesting that additional sterols with unique activities in intracellular signaling, metabolism, and the control of gene expression remain to be discovered. The current application will make innovative use of mass spectrometry and molecular biology to identify and quantitate sterols with new biological functions in a well-defined cell type, the macrophage. Among the questions to be answered include, which sterols are present in the macrophage and in what amounts are they found? How does the macrophage sterol complement change in response to stimulation? What biosynthetic enzymes produce macrophage sterols? How do sterols regulate the output and metabolism of other lipid classes? Success in these endeavors requires a multi-pronged attack involving collaborators with diverse interests and research talents such as those assembled in the LIPID MAPS consortium.

6. Isoprenoids and Structural Lipidomics Core

This core will be directed by Dr. Christian Raetz (Duke University). There is considerable biochemical evidence for the existence of novel minor lipids. A key observation with regard to all mammalian lipids, including phospholipids, glycolipids and sterols, is extreme molecular heterogeneity. We estimate that there are at least 1000 distinct chemical entities that comprise the "lipidome" of a typical animal cell. Heterogeneity is seen in both the polar and hydrophobic portions of these amazing molecules. When the most sensitive radiochemical labeling techniques are combined with multi-dimensional separation procedures, it is obvious that there are dozens of unidentified minor components, some of which could play key roles in signaling or regulatory networks that remain to be discovered. The recent explosion of information regarding the diversity and function of the multiple types of phosphatidylinositolphosphate derivatives nicely illustrates this point; but it is very likely that many additional lipids falling into known or even novel structural classes are yet to be uncovered.

The Isoprenoids and Structural Lipidomics Core will carry out an in depth analysis of novel lipids in the macrophage using the same powerful technologies already developed by the Raetz laboratory to explore novel minor lipids that function as endotoxin precursors in Gram-negative bacteria. This approach will be applied initially to the full chemical characterization by mass spectrometry and NMR of a lysophosphatidylinositol-like lipid in 32Pi labeled macrophages stimulated with endotoxins. These pilot studies will then lead this core into the systematic analysis of all other unknown minor lipids in macrophages. These efforts will be carried out in collaboration with the various cores of the Lipidomics or Lipid Synthesis Focus Areas, as appropriate. This core will also devise methods to quantify important lipids not covered by the other Lipidomics cores, such as the polysioprene-linked phosphate sugars, certain fat-soluble vitamins and quinones, and the glycophospholipid precursors of the PI glycans.

Genomic evidence likewise strongly supports the existence of novel lipids in procaryotic and eucaryotic cells. Careful inspection of completed genomes indicates the presence of genes encoding novel proteins that might be involved in lipid metabolism. For instance, in E. coli, there are two distant but still functionally uncharacterized orthologs of the glycerol-3-phosphate acyltransferases, two novel members of the phosphatidylglycerophosphate synthase family, and two additional cardiolipin synthase homologues of unknown function. In animal cells, there is similar genomic evidence for additional, as yet uncharacterized, enzymes related to known enzymes of lipid metabolism that likely account for some of the minor unknown lipids discussed above.

There are important biological insights to be gained from the elucidation of novel lipid structures. The clues that will be provided by the elucidation of new minor lipids present in animal cells will immediately facilitate the search for novel enzymatic pathways required for their biosynthesis. The Isoprenoids and Structural Lipidomics Core would be in a strong position to conduct initial enzymatic studies of this kind. The Raetz laboratory would follow the same basic approach used in the case of Gram-negative endotoxins, which resulted in the discovery of more than 25 new enzymes and their structural genes over the past 15 years. The scheme is conceptually simple, and it may be summarized as follows. i) The identification and elucidation of new lipid structures must come first. ii) The new lipid structures provide clues to the existence of plausible new enzymes that might be responsible for their biosynthesis. These "plausible enzymes" can often be deduced using the general rules of mechanistic enzymology derived from studies of unrelated metabolic pathways. The most difficult step is usually the synthesis of the several possible alternative substrate precursors that are needed to assay the proposed new enzyme. The availability of the Lipid Synthesis Focus Area will greatly facilitate this step. iii) Once a new enzyme has been identified by in vitro assay, it can be expression cloned or purified to permit sequencing and matching with the genome. A match to a reading frame of unknown function is the ideal outcome, as it serves to define for the first time the function of that gene. Consequently, the Isoprenoids and Structural Lipidomics Core will contribute in a systematic manner to the functional annotation of the human genome, using the same logic and methods already validated with E. coli endotoxins.

Both NMR spectroscopy and mass spectrometry will be necessary in evaluating the structures of all novel compounds. In some cases, chemical synthesis by the cognizant Focus Area may be necessary for validation. The fractionation and structure determination methods developed previously for bacterial phospholipids and endotoxins are completely general in their applicability to all animal lipid classes, and will now be used for studies with macrophages.

7. Summary and Cross-talk Among Lipid Classes

The six Lipidomics Cores will explore changes and interactions within their assigned class of lipids, but we anticipate finding many examples of cross-talk between and among classes of lipids which we will analyze as well. There are already many examples known of cross-talk in lipid metabolic and signaling pathways. The interactions among lipids can be categorized as three types: physical interactions in membranes and other structures (such as lipoproteins), direct interactions between lipids in coupled metabolic pathways (such as through shared enzymes or precursors), and cross-talk between lipid mediated signaling pathways. Figure 2 gives examples of these using sphingolipids, but many other examples could have been used as well.

In addition, there are a number of interactions that have been linked to macrophage behavior that are important to atherosclerosis, namely: the hydrolysis of SM in LDL (which can be done by treatment with exogenous SMase, or-it is thought-in vivo by a secreted SMase), induces the uptake of LDL by macrophages and increases foam cell formation; and the aforementioned induction of PLA2 by ceramide in macrophages.

These examples underscore the importance of knowing how changes in one lipid metabolism/signaling pathway affects, and is affected by, the behavior of the other lipids, as will become possible to evaluate comprehensively as a result of this Glue grant.

F. INTRODUCTION TO CELL BIOLOGY FOCUS AREA

The Cell Biology Focus Area will be anchored by a Macrophage Biology and Functional Genomics Core directed by Dr. Christopher Glass (University of California, San Diego). In later years he will also direct a bridge to extend and explore applications to Transcriptional Regulation in Macrophages. The macrophage has been chosen as the initial cell type for analysis by the LIPID MAPs consortium for several reasons. First, macrophages and related cell types play critical roles in many aspects of immunity and homeostasis. For example, as effector cells in native immune responses, macrophages can kill invading bacteria through the production of reactive oxygen species, and can take up and kill bacteria, yeast and other pathogens through phagocytic mechanisms. As participants in acquired immune responses, macrophages function to present antigens to T cells, take up antigen/antibody complexes via Fc receptors, and influence the evolution of immune responses through secretion of cytokines, chemokines and other regulatory molecules. Homeostatic roles of macrophages and related cell types include participation in wound healing and regulation of bone turnover. Macrophages also contribute significantly to the pathogenesis of a number of human diseases that involve chronic inflammatory responses, such as atherosclerosis.

Second, lipid metabolism in the macrophage is subject to extreme physiological and pathophysiological programs of regulation. A physiological example is provided by the phagocytosis of senescent red blood cells by macrophages, which ultimately requires the degradation or disposal of all of the components of the red blood cell membranes. Enzymes and transport proteins required for these processes must be induced in order for lipid homeostasis to be maintained. The free cholesterol component must either be esterified, converted to soluble forms, or transported out of the cell to acceptors such as HDL. These homeostatic processes can be overwhelmed or inactivated in disease processes such as atherosclerosis, in which macrophages accumulate massive quantities of cholesterol and take on the appearance of so-called foam cells. An additional example is provide by the response of macrophages to lipopolysaccharide (LPS), a cell wall component of gram negative bacteria. LPS exposure induces the expression of a large number of genes, including those that encode proteins involved in the production of prostaglandins and other classes of eicosanoids.

Third, there are a number of well-established mouse macrophage cell lines as well as standardized protocols for isolating uniform cultures of primary macrophages in large numbers from mice. The availability of these model systems will make it possible for each of the Lipidomics units to identify and characterize major and minor lipids in this cell type under a variety of culture conditions. The Macrophage Biology and Functional Genomics Core will provide standardized protocols and reference cell lines to Participating Investigators. In addition, the Macrophage Biology and Functional Genomics Core will hold workshops for LIPID MAPS participants to enable all participants in this consortium to carry out experiments in macrophages in the most uniform manner possible.

A major goal will be to understand the mechanisms by which extracellular signals regulate lipid metabolism, and how specific lipid products influence cell biology and patterns of gene expression. This Focus Area will characterize the relationship between lipid metabolism and gene expression using microarray technology and associated bioinformatics approaches. In conjunction with the Bioinformatics Core, the Macrophage Biology and Functional Genomics Core will develop custom microarrays for analysis of all genes involved in lipid synthesis, binding, transport and catabolism in response to physiological and pathophysiological stimuli, a technique already developed in the Glass laboratory. A commercial vendor will fabricate these arrays so that they can also be made available to the general scientific community. In addition, the Macrophage Biology and Functional Genomics Core will profile transcriptional responses of macrophage cell lines and primary macrophages to biologically active lipids that are discovered by the Lipidomics cores and bridges. Data derived from these experiments will be provided to the Bioinformatics Core facility for dissemination and use in the development of lipid networks.

The Macrophage Biology and Functional Genomics Core will also be responsible for developing reagents for RNA interference that will allow Participating Investigators to knockdown expression of genes involved in lipid metabolism. As a complement to this approach, the Macrophage Biology and Functional Genomics Core will provide an entry point for the analysis of macrophages derived from knockout strains of mice in which genes of interest to members of the LIPID MAPS consortium have been disrupted by homologous recombination.

An example of experimental design that will be established by the Macrophage Biology and Functional Genomics Core is provided in Figure 3. Data will be collected over a broad range of time points to characterize processes that result from actions of preformed proteins as well as processes that result from changes in gene expression. Each Lipidomics laboratory will obtain solvent extracts for lipid analysis and whole cell extracts for protein analysis using a uniform collection strategy. Isolation of RNA for microarray experiments will focus on longer time scales over which changes in gene expression occur.

G. INTRODUCTION TO LIPID DETECTION AND QUANTITATION FOCUS AREA

The Lipid Detection and Quantitation Focus Area will be anchored by an LC/Mass Spectrometric Analysis Core directed by Dr. Robert Murphy (National Jewish Medical Center/University of Colorado School of Medicine). This core will also serve as the Lipidomics Core for Neutral Lipids (described in Section I.E.1). In addition, Dr. Murphy will direct the Novel Detection Techniques Development Bridge that will aim to expand the functioning of this core as well as the other Lipidomics cores.

1. LC/Mass Spectrometric Approach. The overall goal to both qualitatively and quantitatively assess the changes occurring for lipid substances present within a cell is, with little exaggeration, a daunting task. Nevertheless, this goal is clearly achievable due to the major advances, which have taken place in the discipline of mass spectrometry, which can now be applied to lipid analysis. While a great deal of publicity has been generated, and correctly so, in the capabilities of new techniques in mass spectrometry to identify and measure differentially expressed proteins in cells, proteomics is not the only avenue of biomedical research exploding with new ideas, ancillary approaches, and novel applications afforded by the new generation of mass spectrometers. We now have passed the "imagine" barrier of biomolecule analysis where volatility limited analysis by mass spectrometry. An enormous potential for specificity, sensitivity, and precise quantitation have become a reality for the biochemist.

The important advances that have emerged within the past decade are in three fundamental areas: the first is with the techniques of electrospray ionization and matrix assisted laser desorption (MALDI) (which received the Nobel Prize in Chemistry for 2002). These techniques are now the sine qua non methods to induce nonvolatile molecules into a mass spectrometer, including all biologically derived lipid substances identified to date. There are no known exceptions. Electrospray ionization has the unique advantage of easily interfacing directly with the powerful separation technique of high pressure liquid chromatography (HPLC). MALDI is unsurpassed in efficiency of ion formation and robust operation with mixtures of biological molecules.

The second major advance has been in the development of tandem mass spectrometers and trapping mass spectrometers that have enabled direct study of ions on the millisecond or less time scale after collisional activation and generation of product ions. Structural details can be readily gleaned by MS/MS experiments with picomole or less quantities of lipids. Time-of-flight analyzers (TOF) are uniquely suited for MALDI experiments and high-resolution measurements are now possible with lipid-derived ions at a mass error of a few parts per million, enabling elemental compositions to be accurately determined.

A third major advance has been the database management strategies and informatics in general that can deal with the extensive amount of information generated in the mass spectrometric experiments in a very short period of time. One cannot minimize the importance of the explosive growth of informatics and data handling protocols to enable the investigator to realistically examine the wealth of information gathered.

The optimism of this consortium in stating that it is possible to assess lipid composition and changes in lipid composition in a realistic manner largely is the result of the structural organization of this assembled group. We propose a parallel development of specific techniques in six Lipidomics core laboratories which have extensive experience within their own unique area of lipid biochemistry. For example, experts in steroid biochemistry and sphingolipid biochemistry will be charged with developing the qualitative and quantitative tools for their target class of lipids. This essential component of parallel development will be the result of each laboratory having an identical mass spectrometer (6 mass spectrometers) used for different classes of lipids to cover the broad area of lipid biomolecules present within the cell, specifically the macrophage.

Furthermore, the instrument requested is the tandem quadrupole time-of-flight instrument (QqTOF), which has superb sensitivity and enables all product ions formed after collisional activation of the precursor ion to be collected during the TOF analysis. This feature adds an important dimension, not possible in the tandem quadrupole or even trapping instruments. Also, the requested QqTOF has been engineered for both electrospray and MALDI ionization capability. It is important to note that the Murphy laboratory, Core E, has been involved for approximately 30 years in the analysis of lipid substances by mass spectrometry and the study of the ion chemistry of these molecules. This laboratory will serve a critical role in assisting each of the other centers to best utilize their QqTOF. The growth of mass spectrometry as an industry within the last ten years has resulted in the development of computer tools to facilitate operation of the mass spectrometer for both qualitative and quantitative analysis. While this new generation of machines was developed for protein and drug metabolite analysis, the specific advances are directly applicable to the analysis of lipid substances. More importantly, it will not be necessary that each core director need to become an expert in ion chemistry.

The overall LIPID MAPS project will center around the use of mass spectrometry as a fundamental tool for analysis of lipid substances present within the macrophage. Chromatographic separations and specifically HPLC, both normal phase (separation by polarity) and reversed phase (separation by lipophilicity), will be developed and interfaced to mass spectrometry to deal with lipids as they are presented in complex mixtures. When lipids are extracted from the macrophage, they can range from major components (triacylglycerols, phospholipids, cholesterol, and cholesterol esters) to constituents present at trace levels such as the lipid mediators derived from the arachidonic acid or from sphingolipids. The ability to carry out multi-component analyses will require advances in separation science to deal with these complexities. Thus, multiple stage chromatography will most likely be the rule rather than the exception in the analysis of lipids as they dynamically change within the cell.

2. Quantitation of Lipids. Furthermore, there is a fundamental requirement for high precision, accuracy, and sensitivity in the measurement of certain lipids, which are present in very low levels and play an important role as mediating intra- and intercellular events. On the other hand, the analysis of the more abundant complex lipids do not require the ultimate in sensitivity or precision since they are typically present in concentrations several orders of magnitude higher than lipid mediators. Thus, the general strategies for the quantitative analysis of lipid substances will depend upon the specific lipid being analyzed and specific demands for precision and accuracy. For example, a modest change in the concentration of a particular triacyglycerol species may require an analytical accuracy of 10-20 % to detect. However, the generation of a few picomoles of prostaglandin E2 within a cell would require stable isotope dilution quantitative analysis for measurement.

Therefore, the second theme of this consortium is that quantitative analysis will be carried out with two general strategies, one involving stable isotope tracers as internal standards for high precision and accuracy, and the second, using structural homologs as internal standards for moderate precision and accuracy.

Finally, investments in the development of new strategies for the analysis of both abundant as well as trace lipids using newer methods of mass spectrometry will drive future applications of total lipid analysis. All quantitative and qualitative analyses will be carried out in the QqTOF. This particular instrument has the additional capability of storing ions in the collision cell to substantially increase the sensitivity of the mass spectrometric measurements. The use of this instrument with electrospray ionization will permit lipid analysis in both high precision, stable isotope dilution quantitative analysis of lipid mediators, and the molecular species analysis of abundant complex lipids. When fitted with the matrix assisted laser desorption source, this instrument can, in subsequent years, lead to the implementation of techniques developed to improve sample throughput. At the present time, a LC/MS/MS instrument using electrospray ionization has the capacity for only a limited number of samples it can analyze in a given day. The major limitation of this technique is the time it takes to carry out the separation of components by HPLC and the fact that many samples must be analyzed multiple times as both positive and negative ions. This requirement for multiple analyses adds considerable overhead in time that reduces the efficiency of the mass spectrometric approach, not to mention the need to obtain larger samples. Reduction of this limitation will be the major focus of new proposals in technology development under auspices of a bridge project.

The general outline of five centers implementing mass spectrometry for their specific lipid target will be quite similar. The sixth center will be charged with the discovery and structure elucidation of novel lipids as well as the routine analysis of glycolipids. It will take some time following initiation of the grant to have the staff undergo appropriate training in the operation of each instrument. The remainder of the first year will be used to qualitatively assess those lipids present or produced by the macrophage and therefore, targets for quantitative analysis. Specific assays for each of the target lipids will involve either the use of stable isotope dilution strategies or homolog internal standards, establishing standard curves using both primary and secondary standards, and an assessment of the analytical figures of merit for each of the individual quantitative assays for their target molecules. It will be critical that the variability in the assay be established and must be lower than the expected biological variability of each individual lipid target.

The second year will require considerable development of extraction methods relevant to the class of lipids each core will investigate. Overall, an attempt will be made to cooperatively work towards a common extraction scheme for the majority of lipids. This general approach could also be more directly implemented by other investigators outside of the consortium. It is clear, however, that specific lipids will likely require a unique extraction/purification strategy to remove the target lipid substance from the biological matrix. A second major goal will be the development of purification strategies from the macrophage to optimize the relative purity of their target lipid. This is an absolutely critical point since electrospray ionization sensitivity depends a great deal upon sample purity. It is clear that complete suppression of ions can take place in a crude extract, leading to erroneous results, if appropriate care is not paid to the purification (largely by normal and reversed phase HPLC) of the target compounds. During this year, a critical milestone will be an assessment of the variability of lipid composition within the resting macrophage and an assessment of those variables in cell culture, which alter lipid composition of both major and minor components. In subsequent years when the biology of the macrophage will be the major focus of activities, implementation of advances of MALDI ionization to increase sample throughput will take place in each of the laboratories.

Much remains to be done in the application of newer techniques of mass spectrometry to the wide range of lipid substances that can be isolated from the macrophage. Response factors for most lipid components need to be determined in the QqTOF by either electrospray and/or MALDI ionization. The response factors will be the key in being able to carry out quantitative analysis of multiple components as in complex mixtures since it will be technically feasible only to add a few internal standards to normalize the instrument response and glean quantitative information. When a small number of lipid mediators are the target of quantitative analysis, stable isotope dilution with its attendant superb analytical figures of merit can be used to great advantage. Finally, the flow of information from the mass spectrometer to the database must be optimized. A meaningful display of the data must be developed since the biochemical investigator must interact with this data. Thus, a close collaboration between experts in the Informatics Focus Area as well as those in generating mass spectrometric data must be established.

3. Data Presentation and Display. The recent advances in mass spectrometric capabilities, ionization, and chromatographic separation have brought the field of lipid biochemistry to a point where it is feasible to consider qualitative and quantitative assessment of a vast array of lipid substances synthesized and present within a cell. Furthermore, it will be possible through a careful design of separation techniques to look at both major components as well as trace lipids which carry quite different levels of information, but nonetheless, quite important for a global understanding of events taking place within a cell and within a cell compartment. Finally, the use of common strategies with identical mass spectrometers will enable this team of investigators to rapidly capitalize on advances in techniques while still having a user-friendly instrument. High throughput analysis of targeted lipid entities is anticipated as an important outcome.

One of the particular challenges of the analysis of neutral lipids, glycerophospholipids, and sphingolipids is the total number of molecular species to be quantitated and the requirement to examine in detail the extensive set of quantitative information generated from each biological experiment. In the past, reports of lipid molecular species, either phospholipids or TAGs, have been in the form of tables reporting average concentrations or mole % with attendant statistical estimates of analytical errors. We propose to use an alternative strategy to present data largely based upon the tools employed in gene chip experiments. In this manner, the quantitative levels of neutral lipids will be ex-pressed as a color square (with red being the most abundant and blue being the least abundant) (see Figure 4). Changes in molecular species would then be seen as changes in color as a function of experimental condition. Another display will normalize each of the observed molecular species to a single color (mid-range) (Figure 5) for the control cell situation and express the change in a positive or negative direction by a change in expressed color for each of the individual molecular species in the color chart array. In this way, changes in both minor as well as major molecular species could be more readily observed. Of course the ability to extract the exact quantitative numbers (based upon mass spectrometry) for each of the experimental test conditions can be readily recalled.

4. Data Analysis: Statistical Considerations. The efficacy of a statistical test to distinguish between different values of a parameter of interest is referred to as the power of the test. In generating new computational software for analyzing glycerophospholipid changes in mast cells the Brown laboratory wanted to determine the magnitude of the difference that could be measured within a given probability function. That is, given a differential of particular magnitude, the test will detect this difference with a calculated certainty. This function involves the other parameters that drive the underlying characteristics of the distribution of the test statistic. Some of these parameters are under the control of the investigators, for example, the number of repetitions within the various groups of interest.

The Brown laboratory has found that to ascertain changes in phospholipid composition in cell surface receptor stimulation in mast cells five (5) repetitions are sufficient (with a time course of 6 to 8 points). Other parameters are not under the control of the investigators and can usually be considered to be unknown a priori. Common practice involves designing tests with a desired rate of false alarm production, usually referred to as the alpha value of the test, which is familiar to most users of statistical testing. Given a desired alpha value, an attempt is made to optimize the ability of the test to detect a desired difference with an acceptable probability against the practical problems in sample collection and data analysis such as cost, time, computing power, etc. There is no reason to believe that measurements of lipid changes in macrophages will be appreciably different. However, we anticipate that initial fractionation of lipid samples by LC will increase the number of samples six-fold, but will simultaneously improve the precision of quantitation for a given sample.

The main difficulty presented in the detection of differences between test conditions within mass spectrometry data, is that the methodology currently employed involves the detection of relative changes in a normalized collection of data sets. This creates the problem of determining beforehand what level of normalized differential between various treatments is desirable to detect. Moreover, it is necessary to examine the output from several hundred (after parsing) individual signals within each data set. Therefore, there are actually hundreds of individual power functions to attempt to optimize against. Nonetheless, it has been possible to locate many statistically significant changes in the normalized output signals looking at data sets involving a total of 48 runs. This data was organized into two treatment groups where three repetitions were observed at eight time course points. Though only a preliminary analysis, this suggests that a large number of collection points are not necessary for reasonable results.

With the incorporation of internal standards into the samples, improved data transformation functions will yield a better understanding of the magnitude of changes that are desirable to detect. This still does little to ameliorate the problem of optimizing the large number of power functions within the analysis. Each of the various signals examined will likely exhibit a different variance in its underlying distribution, and therefore require a different number of observations to detect a given differential, even when this differential is considered as a function of the magnitude in the basal condition. This difficulty can be handled in several ways. If the investigator has a set of particular m/z values that are of interest, preliminary data runs can be used to obtain an estimate of the variability within these signals, and the required number of observations can be calculated from a worst-case scenario type analysis. If, however, the analysis is to be used in a screening format, the best method for handling the power problem may be to determine an acceptable percentage of signals one would like to have the ability to detect a given differential from a function of the basal magnitude. In either case preliminary results suggest the required number of samples would not be unreasonable in most cases.

H. INTRODUCTION TO LIPID SYNTHESIS AND CHARACTERIZATION FOCUS AREA

The Lipid Synthesis and Characterization Focus Area will consist of one Core of the same title. However, this core has three distinct, but interrelated sections namely: (A) Lipid Standards and Production Chemistry directed by Dr. Walter Shaw (Avanti Polar Lipids), (B) Novel Lipid Synthetic Design directed by Dr. Michael VanNieuwenhze (University of California, San Diego), and (C) Novel Lipid Biophysical Characterization directed by Dr. Stephen H. White (University of California, Irvine).

This Focus Area will respond to new challenges presented by the Lipidomics Focus Area, in terms of producing lipid standards for mass spectrometry and new materials of the necessary quality and quantity, and in advancing the methodology itself. This core is well equipped for studies advancing organic synthesis in its own right through the total synthesis of novel lipid structures. In addition, there will be ample need for synthetic methods development to allow the efficient synthesis of general lipids of biological importance necessary for the Lipidomics Cores to function optimally. Lipids offer an interesting breadth of synthetic challenges to complement more typical syntheses. The scope of the molecular diversity of lipids cannot be precisely specified at the outset, but it is anticipated that enough surprises still exist in lipid and lipo-conjugate structure to require a specific effort in the organic synthesis of lipids with Participating Investigators who are experts in the art of organic chemistry. Whenever possible, the synthetic approach will provide computer aided design of the new molecular architectures followed by controlled directed chemical synthesis and then process development for large scale production will be carried out by the Shaw laboratory with the expertise of Avanti Polar Lipids Inc. The unequivocal elucidation and proof of structure remains crucial for the LIPID MAPS effort. The wealth of structural diversity represented by lipids also offers research challenges for the spectroscopic analysis of advanced synthetic lipid intermediates as well as products.

We have access to all of the instrumentation needed to carry out such analyses. In cases where the synthesis of novel lipids reveal less well-understood phenomena, we will follow up the mechanistic and biophysical implications of these findings. Overall, this Focus Area will provide routine lipid products by efficient synthetic routes, design syntheses for the novel lipids uncovered by the LIPID MAPS Consortium, and follow up the biomedically relevant organic and medicinal chemistry signaled by these compounds. For every new lipid synthesized, the biophysical characterization section of this core will completely assess its characteristics and behavior.

I. INTRODUCTION TO INFORMATICS FOCUS AREA

The Informatics Focus Area will be anchored by the Bioinformatics and Data Coordination Core directed by Dr. Shankar Subramaniam (University of California, San Diego). In later years, he will also direct the LIPID MAPS Networks Bridge that will develop new approaches to utilize the massive data accumulated by the Bioinformatics Core.

While the Glue grant programs such as the one proposed are centered on systems and multidisciplinary approaches to biology, their success (or failure) will critically depend upon informatics and computational cores. Akin to other Glue grants, the proposed project will generate large volumes of data pertaining to lipids and lipid-guided interactions and networks. These will serve to provide a key missing feature in modern cell systems research. There are two major challenges posed by this project to the Bioinformatics Core. First, unlike macromolecular sequence or structure data, the data collected in this project will be heterogeneous, complex, unstructured and occasionally contradictory. Making biological sense of this data and converting diverse data into testable biological hypotheses will be the first challenge for the Bioinformatics Core. The cell functions as an integrated entity. The combination of data and biological knowledge of all the cellular components is essential for this integrated systems view and to map quantitatively cellular input to response. Integration will remain a serious challenge and will require novel strategies not currently available in the bioinformatics area.

The challenges of modern systems biology can be deconvoluted into three parts: obtaining the components, also known as the "parts-list" in a context-specific manner, deciphering interactions between components and reconstructing the complex biochemical network and modeling of the network in terms of input-response characteristics. The LIPID MAPS project will have the additional burden of integrating lipid components and interactions with other macromolecular networks. While several biochemical pathways have been reconstructed through painstaking experimentation, some involve lipid molecules, there is no comprehensive map of a lipid-guided networks and that will be an important goal and anticipated outcome of this project.

This project is complex. The bioinformatics developments for this project by themselves would serve the community immensely and advance biological science. Our ability to undertake a project of this magnitude comes from extensive past experience in building large-scale bioinformatics infrastructures. The Biology Workbench, developed in the Subramaniam laboratory almost a decade ago, is a highly valued infrastructure in the large biomedical community. The bioinformatics infrastructure developed for the Alliance for Cellular Signaling, a project which is at least as complex as the LIPID MAPS project, has already provided the community with next generation biology-computer science interface paradigms. In both cases the infrastructure, software and applications have been made available freely to the scientific community and the bioinformatics effort has a track record or excellent dissemination.

In this project, we plan to utilize our past experience to carry out three tasks. In addition to coordination and communication of the data by building a Laboratory Information Management System (LIMS), development of seamless dissemination, query and analysis interfaces, we plan to advance the bioinformatics research area concerning LIPID MAPS. We plan to develop a detailed ontology for lipids and lipid-associated phenomena (a task not carried out to date by the community), build an object-relational database of all pertinent data, innovate new tools and methods for data analysis, integration and mining and finally develop entirely novel strategies for reconstructing lipid networks.

Dissemination of the results to the broader community is an essential feature of all Glue grant research programs, in light of the breadth of resources made available and the importance of such major research challenges in biomedicine. UCSD and the San Diego Supercomputer Center have considerable experience with such dissemination efforts and is aware of the importance of timely release of information, through our participation as the Bioinformatics Core (and dissemination activities) in the Joint Center for Structural Genomics and the Glue grant, the Alliance for Cell Signaling. Our colleagues have created Molecule Pages, peer reviewed by Nature, to ensure outstanding expertise in curation of findings on the specific new signaling proteins and new results on known molecules. This builds on a decade of computational biology and bioinformatics at UCSD, which through its Supercomputer Center, pioneered many applications at the interface of computing and biology. As such, while the early years of LIPID MAPS will frequently require a service component aspect for dissemination and bioinformatics, as the results from the cores and bridges grows and more information is established on the interplay of various lipid classes, more research and analysis oriented approaches will need to be established to allow and sophisticated querying across the various data sets ultimately enabling modeling approaches to complement experimental ones. By setting up a standard cell system and experimental conditions, and implementing a common infrastructure for data, the actual information infrastructure needed will be far more straightforward than for most of the integration issues addressed in bioinformatics.

The Informatics Focus Area will contain the Bioinformatics and Data Coordination Core and the LIPID MAPS Network Bridge that proposes to work on LIPID MAPS networks. The Data Coordination section of the Core will deal with development of a Laboratory Information Management System (LIMS), the acquisition and processing of data in the informatics server, and the creation of a public dissemination and communication system on the Web. The Bioinformatics section of the Core will deal with the design of structured databases for all the LIPID MAPS data and annotations, processing and analysis of LIPID MAPS data, and the creation of databases for lipids and proteins involved in lipid metabolism.

The LIPID MAPS Network Bridge will deal with the development of novel methods of analysis of lipidomics data, the creation of lipid metabolism and signaling networks, and the reconstruction of LIPID MAPS for quantitative analysis of input-response in macrophage cells.

J. INTRODUCTION TO ADMINISTRATION FOCUS AREA

The Administrative Focus Area will consist of one Administrative Core and will be directed by Dr. Edward A. Dennis (University of California, San Diego) as Principle Investigator of this grant. The contributions of the Administrative Focus Area to the overall efforts is summarized in Sections II and III below.

II. ADMINISTRATIVE MANAGEMENT PLAN

A. STRUCTURE OF LIPID MAPS ORGANIZATION