Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 56

Development of Enhanced Analytical Methodology for Lipid Analysis from Sampling to Detection

GIORGIS ISAAC

A Targeted Lipidomics Approach

ISSN 1651-6214ISBN 91-554-6259-6urn:nbn:se:uu:diva-5810

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2005

59

59

This thesis is dedicated in memory of my beloved father Isaac (January 19, 1938-December 19, 1994) a man of endless energy and talent who used to guide, encourage and explain to me the value of education during my child-hood. He was a hard working man always striving to be better and still a family man. He was a proud man and he would do anything possible to take care of his family. With out his dedication I would never be where I am to-day.

Giorgis Isaac May 2005

List of Papers

The following Papers are discussed in this thesis and they are referred to in the text by their Roman numerals:

I Total Lipid Extraction of Homogenized and Intact Lean Fish Muscles Using Pressurized Fluid Extraction and Batch Ex-traction TechniqueGiorgis Isaac, Monica Waldebäck, Ulla Eriksson, Karin E. Markides and Göran Odham, Accepted for publication inJ. Agric. Food Chem., 2005.

II Pressurized Fluid Extraction (PFE) as an Alternative Gen-eral Method for the Determination of Pesticide Residues in Rape SeedTuija Pihlström, Giorgis Isaac, Monica Waldebäck, Bengt-Göran Österdahl and Karin E. Markides, The Analyst, 2002, 70, 554-559.

III Analysis of Phosphatidylcholine and Sphingomyelin Molecu-lar Species From Brain Extracts Using Capillary Liquid Chromatography Electrospray Ionization Mass Spectrome-tryGiorgis Isaac, Dan Bylund, Jan-Eric Månsson, Karin E. Markides and Jonas Bergquist, J. Neurosci. Methods, 2003, 128, 111-119.

IV Brain Lipid Composition in Postnatal Iron-Induced Motor Behavior Alterations Following Chronic Neuroleptic Admini-stration in MiceGiorgis Isaac, Anders Fredriksson, Rolf Danielsson, Per Eriksson and Jonas Bergquist, Submitted to Neurosci. Lett.

V Characterization of Brain Sulfatide Molecular Species in Arylsulfatase A Deficient Mice Using Electrospray Ionization Tandem Mass SpectrometryGiorgis Isaac, Jan-Eric Månsson, Volkmar Gieselmann, Elisa-beth Hansson, Zarah Pernber and Jonas Bergquist, Manuscript.

Reprints of the publications were made with kind permission of the publishers.

The author’s contribution to the Papers:

Paper I: I was responsible for practical work except the NMR part. I wrote the paper in close consultation with the co-authors. Paper II: The experiments were planed by Tuija Pihlström and Monica Waldebäck. I carried out the PFE extraction and contributed to the paper writing.Paper III: Planning the experiments together with Jonas Bergquist and I was responsible for experimental work of the LC/MS experiments. I wrote the paper in close consultation with the co-authors. Paper IV: I was responsible for experimental work and planning of the LC/MS experiments. Paper writing and data analyses were made in col-laboration with co-authors. Paper V: I was responsible for all experimental part in ESI/MS. Paper writing and data analysis was made in collaboration with co-authors.

Contents

1. Introduction.................................................................................................9

2. Lipids and Lipidomics ..............................................................................112.1 Lipids: Definition ...............................................................................112.2. General Classification of Lipids........................................................112.3. Lipidomics: The Next Frontier..........................................................132.4. Membrane Structure and Occurrence of Phospholipids, Sphingomyelin and Sulfatides..................................................................14

2.4.1. Phospholipids.............................................................................162.4.2. Sphingolipids .............................................................................17

3. Pressurized Fluid Extraction .....................................................................193.1. Extraction Using High Diffusion Fluids ...........................................193.2. Theory and Concept of PFE Technique ............................................203.3. PFE Instrumentation..........................................................................213.4. Selected Application and Comparison of the PFE Technique with Conventional Methods .............................................................................23

3.4.1. Total Lipid Extraction Using PFE (Paper I) ..............................233.4.2. PFE as an Alternative Method for Extraction of Pesticide Residues from Fatty Matrix (Paper II).................................................24

4. Analysis of Intact Phosphatidylcholine, Sphingomyelin and Sulfatide Molecular Species Using LC/ESI/MS and ESI/MS/MS...............................26

4.1. Reversed Phase LC of Intact PC and SM..........................................274.2. Mass Spectrometry ............................................................................29

4.2.1. The Mechanism of ESI/MS .......................................................294.2.2. The Principles of ESI/MS of Lipids ..........................................314.2.3. MS Scan Modes.........................................................................31

4.3. Reversed Phase LC/ESI/MS..............................................................35

5. Data Handling ...........................................................................................38

6. Conclusions and Future Aspects ...............................................................39

7. Acknowledgements...................................................................................41

8. Summary in Swedish ................................................................................43

9. References.................................................................................................45

Abbreviations

Ag-LC Silver ion Liquid Chromatography ASA Arylsulfatase A CID Collision Induced Dissociation ESI Electrospray Ionization FA Fatty Acid GC Gas Chromatography GPC Gel Permeation Chromatography LC Liquid Chromatography LLE Liquid-Liquid Extraction MAE Microwave Assisted Extraction MP Mobile Phase MS Mass Spectrometry m/z Charge-to-Mass Ratio PA Phosphatidic Acid PC Phosphatidylcholine PE Phosphatidylethanolamoine PFE Pressurized Fluid Extraction PI Phosphatidylinositol PIS Precursor Ion Scan PL Phospholipid PS Phosphatidylserine RP Reversed Phase SFE Supercritical Fluid Extraction SM Sphingomyelin SP Stationary Phase TLC Thin Layer Chromatography

9

1. Introduction

Many of the traditional sample extraction methods have become the weakest link in modern chemical analytical methods. Some of these methods might be effective but slow, require large amounts of organic solvents and be diffi-cult to automate and hence extremely labour intensive. During the past dec-ade, driving forces such as the need for faster analyses, increased productiv-ity, concern for cost of analysis, safety of laboratory personnel and the re-duction of environmental pollution sources have encouraged chemists to pursue modern sample extraction methods 1, 2.

Pressurized fluid extraction (PFE) is a relatively new automated liquid ex-traction technique which uses elevated temperature and pressure with stan-dard liquid solvents to achieve fast and efficient sample extractions. In this thesis the reliability and efficiency of the PFE technique was utilized for novel application in the extraction of total lipid content from cod and herring muscles (Paper I), pesticides from rape seed (Paper II), and phosphatidylcho-line (PC) and sphingomyelin (SM) molecular species from human brain (Pa-per III). The developed methodology was described and the results were compared to different traditional extraction methods. Furthermore, in Paper III a method was developed for the separation and characterization of the PC and SM molecular species using reversed phase capillary liquid chromatog-raphy (RP-LC) coupled online to electrospray ionization mass spectrometry (ESI/MS). The developed method was applied to investigate lipid composi-tions in mouse brain (Paper IV). In Paper V a tandem MS in precursor ion scan (PIS) mode was used to characterize the sulfatide content in brain tissue.

From an ecological point of view, it is more relevant to express the resi-due level of lipophilic persistent pollutants on fat weight basis than on tradi-tional fresh weight basis 3. Numerous intercalibration laboratories have shown that the original Jensen extraction method for total lipids and persis-tent organic pollutants, gave satisfactory yields when applied to fatty aquatic organisms 4. However, in case of lean fish muscle (fat content below 1%) the lipid yields with that method is about 25% too low and consequently residue levels on lipid bases are correspondingly too high 4. In Paper I, the reliability and efficiency of the PFE technique for the extraction of total lipid content from cod and herring muscle tissue was evaluated and the results were com-pared to both the original and the later modified Jensen extraction methods. The influence of several experimental variables affecting the extraction effi-

10

ciency of the total lipid content with PFE was studied using experimental design.

The determination of pesticide residues in fatty matrices such as rape seed is associated with various analytical difficulties. Since the studied pesticides cover a wide range of polarity (Paper II, Table 1) the recovery of very non-polar pesticides from the lipid rich-matrix makes the extraction challenging. Therefore there is a need of developing extraction techniques, which effi-ciently recovers the different classes of pesticides, over the polarity range, with low solvent consumption and rapid extraction time from lipid-rich ma-trices. In Paper II, a PFE multi-method was developed for the determination of pesticide residues in rape seed and the results were compared to the tradi-tional liquid-liquid extraction (LLE) method. The analysis was performed with gas chromatography (GC) using different detectors after clean-up by gel permeation chromatography (GPC).

Disturbances in lipid membrane cause various disorders and therefore the development of methods for the analysis of lipids is essential. Paper III de-scribes a method based on RP capillary LC coupled online to ESI/MS for the analysis of PC and SM molecular species. Furthermore, the extraction effi-ciencies of the traditional Folch method and PFE were compared using hu-man brain samples. In a subsequent study (Paper IV), the developed method was applied for the analysis of PC and SM molecular species in brains from eight groups of mice treated with vehicle, low dose clozapine, high dose clozapine and haloperidol. Both the effects of postnatal iron administration in the lipid composition and behavior; and whether or not the established effects may be altered by subchronic administration of the neuroleptic com-pounds, clozapine and haloperidol were investigated.

In Paper V a tandem MS in PIS mode was used to compare brain sulfatide molecular species accumulation in arylsulfatase A (ASA) knockout (ASA -/-) and wild type control mice (ASA +/+) at different ages. ASA -/- mice were de-veloped as a model of monogenic disease metachromatic leukodystrophy (MLD) with deficiency of the lysosomal enzyme ASA. MLD is a human disease in which sulfatide is accumulated within the lysosomes due to defective en-zyme activity of ASA. As a comparison, the sulfatide molecular species in cultured astrocytes were also characterized.

The aim of this thesis is to describe and discuss method development using the recent extraction technique PFE and analysis of intact PC, SM and sulfatide mo-lecular species using LC/ESI/MS and tandem MS from complex biological sam-ples.

11

2. Lipids and Lipidomics

2.1 Lipids: Definition Before going into details of the different lipid classes (phospholipids (PLs), SM and sulfatide) and their molecular species considered in this thesis, it may be useful to define the term “lipid”. Lipids can not be defined, as it is often done based on their physical properties, as a class of natural com-pounds that share a high degree of solubility in organic solvents. This desig-nation is at the same time both too broad and to narrow because it encom-passes a diversity of structurally and functionally unrelated molecules and excludes partly water soluble lipids such as gangliosides and phosphoinositi-des 5, 6. The alternative definition proposed by W. Christie 6 based on their chemical structure and function seems more adequate i.e. “Lipids are fatty acids (FAs) and their derivatives, and substances related biosynthetically or functionally to these compounds”. By this definition, every biomolecule derived from the condensation of malonyl coenzyme A units and implicated in cell structure and function should be considered as a lipid. Thus FAs are defined as compounds synthesized in nature via condensation of malonyl coenzyme A units by a FA synthase complex. The FAs contain even number of carbon atoms (usually C14-C24), although the synthase can also produce odd and branched chain FAs to some extent when supplied with appropriate precursors. Other substituent groups, including double bonds are normally incorporated into the aliphatic chains later by different enzyme systems 6.

2.2. General Classification of Lipids It is not the aim of this thesis work to give detailed classification of the dif-ferent lipid compounds but rather to show the complexity and structural va-rieties of the different lipids as a base for their diverse characteristics. Unlike proteins and nucleic acids that are linear polymers of amino acids and nu-cleotides respectively, with linkages at only one position, lipids can adopt complex structures. The main lipid classes of animal origin can be classified as shown in Figure 1. This global classification is fairly comprehensive and remains only as a guide. Some lipid compounds classified in one class by some may be classified in another class by others.

12

Figure 1. Lipid classification scheme. This global classification is fairly comprehen-sive but remains only as a guide.

Christie in his recent book 6 classified the lipids into two broad classes which is convenient especially for chromatographic purposes. Simple lipids (glycerols, waxes, free FAs and sterols) are defined as those that, on hy-drolysis, yield at the most two types of primary product per mole; complex lipids (PLs, sphingolipids and glycoglycerolipids) yield three or more pri-mary hydrolysis products per mole. Alternatively, the terms "neutral" and "polar" lipids respectively are used to define these groups, but are at the same time less exact terms.

In a recent review Han and Gross 7 roughly classified the cellular lipids into three large groups including non-polar lipids, polar lipids and metabo-lites based on their relative polarities of the head group regions. Non-polar lipids include predominantly cholesterol and cholesterol esters as well as triacylglycerols. All these classes of lipids consist of a small or weak polar part and a dominant hydrophobic region. Polar lipids are predominately comprised of PLs, sphingolipids, and glycolipids. The last groups of lipids are lipid metabolites, which are derived by enzymatic action on parent lipids or their precursors. Common metabolites include long chain acylcarnitines, nonesterified FA, FA esters, acyl anamide, ceramides, lysolipids, eico-sanoids, diacylglycerols, sphingosine-3-phosphate, etc. Many of the metabo-lites are biologically active second messengers.

13

2.3. Lipidomics: The Next FrontierThe term “lipidomics” has just emerged in the literature in the context of genomics and proteomics 8-12. However lipidomics will certainly pick up momentum by bringing in sophisticated technologies (e.g. recent advance and novel application of LC/ESI/MS and MS/MS) to combat lipid associated diseases such as coronary heart disease, stroke, cancer, diabetes and Alz-heimer’s disease to mention a few disorders. In a sense while genes can be considered to be the blue prints and proteins are the structures built from these blue prints, the lipids provide a functional family that makes the house a home 13. According to Spenner et al. 10 if the entire spectrum of lipids in the biological system is called “lipidome” then mapping this spectrum can be called lipidomics. Thus, lipidomics can be defined as the full characteriza-tion of lipid molecular species and of their biological roles with respect to expression of proteins involved in lipid metabolism and function, including gene regulation. Therefore the first tasks in characterizing the global change in lipids (lipidomics) can be described by the following 10:

New analytical approaches for mapping the lipidome: Global charac-terization of lipid molecular species by LC/MS and development of novel detection techniques. Application of biophysical methods to understand lipid-protein in-teractions: This mainly focus in the interaction of lipid micro- and nanodomains of biological membranes and will be benefited by ad-vanced synthesis of specific lipids, analogs and probes. Identification of lipid network: Identifying the lipid network including lipid mediators, for metabolic and gene regulation and its integration with non-lipid signaling. The challenge here is the integration of lipi-domics with proteomics to study and characterize homeostatic and aber-rant cellular states.

Recently, in the United States the National Institute of Health funded a $ 35 million five year collaborative project between 18 laboratories, entitled “Lipid metabolites and pathway strategy” (www.lipidmaps.org) to determine the lipidome (all the lipids) in the mouse macrophage. These researchers will now work together to analyze the structure and function of lipids 14. As illus-trated in Figure 2, Prestwich 13 suggested that with the solution of the human genome sequence and the growing ability to conduct massive proteomic analyses, lipid constitute the research frontier for this decade.

14

Figure 2. The importance of functional lipidomics 13.Targeted lipidomics focuses on specific lipids rather than the indiscriminate identification of all endogenous lipids.

The first step in global lipidomics is to measure the amount of each dis-tinct chemical entity present in a cells lipidome and identify those alterations in chemical constituents that are associated with disorders 7. In this thesis a targeted lipidomics approach was used to characterize and profile PC and SM molecular species (Papers III and IV) and sulfatide molecular species (Paper V) by LC/MS and MS/MS using different scan modes from complex biological lipid extracts. It is considered a targeted lipidomics approach be-cause it focuses only on PC, SM and sulfatide molecular species rather than the indiscriminate identification of all endogenous lipids.

2.4. Membrane Structure and Occurrence of Phospholipids, Sphingomyelin and Sulfatides Contrary to the classical view of the plasma membrane as a homogenous PL assembly that is loaded with proteins, proposed by Singer and Nicholson 15,it is now accepted that the plasma membrane is heterogeneous. Biological membranes are structurally and compositionally complex. The biological complexity is the result of the physicochemical properties of the individual membrane components, which is governed by its composition and chemical environment. Compositional diversity arises through chemical diversity of the membrane constituents. The requirement for a cell membrane to be formed is the existence of lipids. In general, a cell membrane is mainly com-posed of PLs, sphingolipids and sterols. Sterols show little structural varia-tion compared to PLs and sphingolipids, in which structural variation origi-nates through variability of the polar head groups and the non-polar hydro-carbon chains (e.g. carbon chain length, degree of saturation, hydroxylation, branching). This structural variation yields a multitude of similar built mole-cules with diverse physical properties, which together with membrane pro-teins; regulate membrane properties in a dynamic fashion 16. The lipid bi-

15

layer illustrated as a sea of PLs 17, provides the basic structure of the mem-brane and serves as a relatively impermeable barrier to the passage of most water-soluble molecules 18 and it contributes to the structural organization, stability and fluidity of the membrane. The structure of a cell membrane is shown in Figure 3.

Figure 3. Structure of cell membrane. Phospholipids, glycosphingolipids and cho-lesterol are the major constituents of the lipid bilayer in which a variety of mem-brane proteins, such as enzymes, receptors and channels are arranged 19.

Lipids comprise about half of the brain tissue dry weight. Brain contains many complex lipids some of which (gangliosides, cerebrosides, sulfatides, phosphoinositides) were first discovered in the brain, where they are highly enriched in comparison to other tissues. Brain lipids have many physiologi-cal functions. They participate in the function as well as the structure of cel-lular membranes. The biomessenger function of non-membrane lipids (ster-oid hormones and eicosanoids) has long been known. However, in recent years some membrane lipids, such as inositides and PC, which were previ-ously believed to have only a structural role, have also been shown to have important function in the signal transduction process across biological mem-branes 20. These recent discoveries reveal the increasing importance of lipids not only in maintaining membrane structure, but also in the regulation of cellular metabolism and growth, pathological state and neurodegenerative disorders, signal transduction and endomembrane transport 5, 21.

In the next section a detailed description and biological importance of the lipid molecular species considered in this thesis will be provided.

16

2.4.1. Phospholipids Glycerophospholipids or simply phospholipids (PLs) are major constituents of cellular membranes in animal and plant tissues. A typical structure of PL consists of three parts: a glycerol backbone, a polar head group with a phos-phate moiety and two FA chains esterified at the sn-1 and sn-2 positions of the glycerol backbone, Figure 4. Structural and functional diversity are therefore possible within each class of lipids since the head group can be combined with a large pool of FAs that vary in both chain length and degree of unsaturation. This yields similarly built molecules having a large diversity of physical prosperities, which allows the living cell to regulate the intrinsic heterogeneity of membranes in a dynamic fashion. Such heterogeneities are dynamically regulated and can play striking functional roles.

Figure 4. The general structure and classes of phospholipids.

A first level of structural diversity is determined by the presence of spe-cific hydrophilic head groups. Based on the differences in the chemical structures that are linked to the phosphate, PLs can be classified into differ-ent classes. For example, if the specific head group is choline or ethanola-mine the corresponding class of the PL is called PC or phosphatidyletha-nolamine (PE) respectively. Other classes of PL include phosphatidic acid (PA), phosphatidylserine (PS), phosphatidylinositol (PI) and phosphatidyl-glycerol or diphosphatidylglycerol etc., Figure 4. Each of these classes has a

17

specific distribution, not only among different cell types, but also among different organelles of the same cell 5-7.

The complexity of PL molecular species is not only because of variation of polar head groups but also because of the nature of the chain where the FA residues are linked to the sn-1 and sn-2 positions of the glycerol back-bone. These residues vary in chain length as well as the number and location of double bonds. The FA residues determine the shape and flexibility of individual PC molecular species and affect the special properties of mem-brane bilayer. The double bonds introduce kinks into the carbon backbone of the FA; the more double bonds there are, the more curved is the molecule. In general, saturated FAs make membranes more stable and rigid, while the more unsaturated is the FA composition the greater is the degree of flexibil-ity introduced into the membrane 22.

This thesis will focus on the subclass diversity in PC molecular species. PC (commonly termed as lecithin) is usually the most abundant lipid in the membranes of animal tissues and is often a major lipid component of plant membranes and sometimes of microorganisms.

2.4.2. Sphingolipids The second major category of polar lipids are sphingolipids that contain a long chain base (most frequently sphingosine) backbone or its analogs and represent approximately 5-10% of total lipid mass in most brain cells 7, 23-26.Based on the nature of the covalent linkage of the polar moieties to ceramide (i.e N-acylsphingosine), sphingolipids can be classified into SM, sulfatide, cerebroside, ganglioside, glucosylceramide, lactosylceramide and other gly-cosphingolipids that contain multiple sugar rings, Figure 5. Abnormal me-tabolism of sphingolipids could lead to their deposition in neural tissues and results in conditions of sever mental retardation known as sphingolipidoses 26. Therefore, the biochemistry of sphingolipids such as SM (Paper IV) and sulfatide (Paper V) in normal and pathological states has evoked a great deal of interest.

SM is comprised of a choline head group as PC and resembles PC in many of its physical properties and can apparently substitute in part for this in membrane. Both PC and SM are located in the outer leaflet of plasma membranes 27, although they are also found in various other biological struc-tures such as lipoproteins 28. SM consists of a ceramide unit linked at posi-tion 1 to phosphoryl choline and it is found as a major component of the complex lipids of all animal tissues but is not present in plants or microor-ganisms 6. Since different FAs that vary in chain length and number of dou-ble bond may be linked to a long chain base, a great number of molecular species are found in natural extracts, Figure 5. SM is an important constitu-ent in nervous tissue and plasma membrane of higher animals and accounts for about 4.2-12.5% of the PL content in the brain of various species 26. It is

18

actively involved in various cellular processes such as protein trafficking and signal transduction 29.

XO CH3

HN

OH

CO

R

X Class of SphingolipidsH Ceramide

Sphingomyelin

O

HH

HOH

HOSO3

H OH

CH2OH

Sulfatide-

O

HH

HOH

HOH

H OH

CH2OH

Cerebrocide

PO(H3C)3N+H2CH2C

O

O-

Figure 5. The general structure and classes of sphingolipids.

Sulfatide is a major constituent of brain lipids and is found in trace amounts in other tissues. It is an essential glycosphingolipid in myelin in the peripheral as well as the central nervous system 30, constituting 4-6 mol% of the total lipids in adult brain myelin 26. The structure of sulfatide consists of two parts: a polar galactosyl-3-O-sulfate head group facing outward from the plasma membrane and a lipid core, ceramide, composed of a FA and a long chain base, sphin-gosine, Figure 5. Since the FA chain can vary in length, degree of unsatura-tion and hydroxylation, each natural sulfatide contains numerous molecular species. These species determine the interaction of the sulfatide molecule in the cell membrane. Sulfatide mediate diverse biological process including the regulation of cell growth, protein trafficking, signal transduction, cell adhesion, neuronal plasticity and morphogenesis 23-25.

19

3. Pressurized Fluid Extraction

3.1. Extraction Using High Diffusion Fluids The extraction process is one of the most important steps in pretreatment for solid and liquid samples. It is therefore not surprising that a wide range of techniques has been developed for the extraction of these samples. The most widely used method for the extraction of solid samples is Soxhlet extraction. However in recent years various high diffusion fluid techniques such as su-percritical fluid extraction (SFE) 31-33, microwave assisted extraction (MAE) 34, 35 and PFE 36 have been developed as an alternative to other conventional techniques. High diffusion fluids are being created by raising the tempera-ture and controlling the pressure of the liquid. Raising temperature results in increased diffusion coefficients and therefore increased extraction rates re-sulting in shorter extraction times. By extracting the target analyte at ele-vated temperature and pressure the rates of diffusion and desorption can be significantly increased.

SFE is a selective extraction method that incorporates thermal and solu-bility driven mechanism using supercritical fluids such as carbon dioxide (CO2), dinitrogen oxide, sulfurhexaflouride, methanol, freons and water. In practice, more than 90% of all analytical SFE is performed with CO2 for several practical reasons 32, 37. Apart from having low critical pressure (74 bar) and temperature (32oC), CO2 is chemically inert, available in high purity at relatively low cost, nonflammable, is easily removed from the extract and environment and analyst friendly when used for analytical purpose. How-ever, SFE has not become the “universal technique” due to some inherent disadvantages. Method development is not straight forward, the polarity of supercritical CO2 is poor for the extraction of polar analytes and the tech-nique is highly matrix dependent. Despite its shortcomings, SFE has suc-cessfully been applied for the determination of fat content in foods 38-42,pesticide residue 43-45 and total poly aromatic hydrocarbons (PAH) and polychlorinated biphenyls (PCB) 44, 46-51.

MAE is another high diffusion fluid extraction technique which uses mi-crowave energy to heat solvents in contact with a sample in order to partition analytes from the sample matrix into the solvent. To be able to absorb mi-crowave, either the solvent or the matrix has to be dipole with a high dielec-tric constant 52-54. It is also believed that high efficiency is obtained due to the destruction of the matrix macrostructure 55. MAE could be performed in

20

a closed or open vessel heating system. The major advantages achieved us-ing the MAE technique are reduced extraction times, improved extraction efficiencies and reduced solvent consumption.

PFE, also been known as pressurized liquid extraction (PLE), pressurized solvent extraction (PSE), accelerated solvent extraction (ASE) or enhanced solvent extraction (ESE), is considered to be an extension of the SFE tech-nique in which CO2 is replaced by organic solvents to solve potential polar-ity troubles 56. In the following section detailed discussions about the theory of the PFE technique, instrumentation, capabilities and limitations as well as selected applications from Papers I and II will be provided.

3.2. Theory and Concept of PFE Technique PFE is a relatively new automated liquid extraction technique which uses elevated temperature and pressure with standard liquid solvents to achieve fast and efficient sample extractions 36, 57. The technique has been devel-oped in an effort to reduce the large solvent consumption, long extraction times and solvent exposure of operator compared to traditional liquid extrac-tion techniques.

Liquid solvents at elevated temperature and pressure provide enhanced extraction capabilities compared to their use at room temperature and atmos-pheric pressure. This is because increased temperature can disrupt the strong solute-matrix interactions caused by van der Waal´s forces, hydrogen bond-ing and dipole attractions of the solute molecules and active sites on the ma-trix 36, 56. Higher temperature also decreases the viscosity and surface ten-sion of solvents, allowing better penetration of solvents into the matrix parti-cles. Both lower viscosity and lower surface tension facilitate better contact of the analytes with the solvent and hence enhance the extraction 36, 56, 58.Thus, the elevated temperature used in PFE increases the capacity of sol-vents to solubilize analytes, improves mass transfer and hence increases extraction rates.

The purpose of higher pressure in PFE is mainly to keep the solvents as liquid even at a temperature above their atmospheric boiling points. Pressur-ized flow would also aid in the solubilization of air bubbles so that the sol-vent penetrates the sample matrix more easily 36. Pressure has minor impor-tance for the resulting recovery in PFE as has been shown in several studies 58-62. However improved extraction might be achieved for analytes trapped in water blocked pores, as has been demonstrated by increased recoveries of polycyclic aromatic hydrocarbons from a model matrix consisted of wet silica 63. This could be explained by the fact that the organic solvents are being forced into the pores at high pressure.

As a result of these combined effects, time and solvent consumption are significantly reduced (Papers I and II) and improved (Paper I) or comparable

21

(Paper II) efficiencies are achieved when using PFE technique compared to classical solvent extraction methods. Method development is easy with PFE compared for example to SFE. Temperature and solvent are the two most important parameters that affect extraction recovery of target analytes. In Paper I, the extraction temperature, solvent and time were optimized using experimental design. Co-extracted materials like lipids are the main problem to be considered when using high temperature to extract lipophilic target analytes. In Paper II, the temperature was optimized to reduce the co-extracted lipids. However, even at the optimized temperature an additional clean-up step was still required. If a particular extraction solvent works well for the traditional methods, in most cases that solvent is also a good choice for PFE 58-62. In Paper II, the same solvent as in the LLE method was used which simplified method development.

Other parameters that can affect recovery include static time, particle size, the flush volume and the number of static cycles used in the extraction. The influence of particle size on the recovery of analytes using PFE has been reported by Björklund et al. 64. They found that smaller particle size results in increased extraction efficiency. Also in Paper I homogenized cod muscle resulted in higher total lipid recovery compared to intact muscle. This is because finely divided solids are more easily dissolved and extracted due to their large surface area to volume ratio.

3.3. PFE Instrumentation The commercial PFE instrument used in this study is a fully automated se-quential extraction system. The system (ASE® 200, Dionex Co., Sunnyvalle, CA, USA) can operate with up to 24 sample containing extraction cells and up to 26 collection vials plus an additional 4 vial positions for rinse 65. Fig-ure 6 shows a schematic diagram of PFE. To perform an extraction, a filter paper is inserted into the extraction cell, followed by the sample. The filled extraction cells are loaded on a carrousel, as well as collection vials are loaded in the lower collection vial carrousel. A robotic arm transfers selected cell separately into the oven for extraction. The oven is maintained at chosen operating temperature, ranging from room temperature to 200oC. After the cell is in place in the oven, the pump immediately begins to deliver solvent to the extraction cell. The static valve closes to allow pressurization of the cell. Because the solvent expands as it heats, the pressure in the cell will increase when the static valve closes. When the pressure reaches 1.5 MPa above the set point, the static valve opens rapidly to relieve the pressure and then closes again. The pump also delivers fresh solvent to the cell to return the pressure to the set point value. This addition of fresh solvent during PFE is analogous to fresh solvent dripping down from the condenser onto the extraction thimble during Soxhlet extraction 58. The first period of extraction

22

is called the heat up time, as the cell contents are heated by the oven to the chosen operating temperature to reach thermal equilibrium. Compared to MAE in which the solvent is heated directly by microwave radiation, heating in PFE is a convection process in which the extraction cell is heated and heat is transferred to the enclosed solvent. Once thermal equilibrium has been reached the extraction enters a static period with a user-chosen duration. Typical static times are 5-10 min but they can vary from 1 to 99 min depend-ing on the optimum extraction time for the particular sample. After the static time, fresh solvent is flushed through the cell to remove extracted analytes. The flush amount can vary from 5 to 150% of the extraction cell volume (with 60% as most common).

Figure 6. Schematic diagram of Pressurized Fluid Extraction system, Dionex ASE200.

If required, multiple extractions can be performed by programming from one to five static cycles. The flush volume is divided by the number of static cycles so that fresh solvent is present at the beginning of each static cycle.

Following the final solvent flush, compressed nitrogen is used to force all solvent and remaining target analytes from the lines and cell, into the collec-tion vial. After the completion of the purge step, the cell is returned to the carrousel and the next sample is taken to the oven to begin the next extrac-tion. Current instrumentation enables unattended operation of extraction systems without user intervention if the sample matrix is simple and does not block the solvent line (Paper II).

23

3.4. Selected Application and Comparison of the PFE Technique with Conventional Methods 3.4.1. Total Lipid Extraction Using PFE (Paper I) There is some uncertainty regarding the effectiveness of total lipid content extraction using the traditional methods in lean fish muscle (fat content be-low 1%) where polar PLs dominate. Since the concentration of organic pol-lutants should be reported on both fresh and lipid weight basis, it is crucial that the used extraction procedures lead to correct results for the pollutants as well as for the lipids. In this study, the efficiency of the PFE technique was investigated for the extraction of total lipid content from cod and herring muscle tissue. The results were compared to original and modified Jensen extraction methods.

The variables extraction time, temperature and extraction solvent were optimized and optimum conditions were found to be 115oC with 2-propanol/n-hexane (65/35, v/v) as a first and n-hexane/diethyl ether (90/10, v/v) as a second solvent. Extraction time had no significant effect on total lipid recovery and was kept at 10 min. PFE gave significantly higher lipid yields (0.62%, RSD=3.1, n=5) compared to original (0.38%, RSD=6.3, n=5) and modified (0.55%, RSD=2.8, n=5) Jensen methods from homogenized wet cod muscle. The efficiency of modified Jensen was also further con-firmed in comparison to original Jensen by replacing acetone with isopropa-nol and n-hexane with diethyl ether earlier reported by Jensen et al. 66. A sub-portion of the freeze dried samples did not show channeling and gave clear extracts however the low lipid yield (0.41%, RSD=2.5, n=5) could be possibly due to difference in the extraction solvent composition as the wet sample contains water as indigenous solvent and loss of lipids during freeze drying process. PFE and original Jensen were also compared for the extrac-tion of total lipid content from herring. PFE gave significantly higher yield (7.8%, RSD=8.5, n=5) compared to original Jensen method (6.9%, RSD=2.2, n=5).

The amount of total lipid extracted depends on the way the sample is treated prior to extraction. To investigate this effect on the total lipid yield homogenized and intact tissues were used and the results were compared to the modified Jensen method. Significantly higher total lipid yields were achieved with PFE [0.68%, RSD=3.7, n=5 using IPR/n-Hx (65/35, v/v) and n-Hx/DEE (90/10, v/v)] and [0.68%, RSD=8.5, n=5 using IPR/DEE (25/10, v/v) and n-Hx/DEE (90/10, v/v)] than the modified Jensen method (0.62%, RSD=1.4, n=5) from a homogenized wet cod muscle.

Using PFE, the solvent mixture IPR/DEE (25/10, v/v) used by Jensen etal. 66 together with n-Hx/DEE (90/10, v/v) gave significantly higher yield than IPR/n-Hx (65/35, v/v) along with n-Hx/DEE (90/10, v/v) for intact cod muscle (0.51%, RSD=18 n=5) compared to (0.36%, RSD=2.9 n=5) respec-

24

tively. However, when using homogenized tissue the statistical t-test showed no significant difference in yields between the two extraction solvents (Paper I, Table 5).

In conclusion, PFE gave about 10% higher yield in comparison to the conventional batch extraction methods. Infrared and NMR spectroscopy study suggested that the additional 10% lipid-like material is rather homoge-nous and is PL origin with head groups of inositidic and/or glycosidic nature. It is difficult to find comparable data on similar biological tissue samples with low lipid content as most PFE extractions were performed on high fat tissues 67-69.

Despite the clean-up step, using PFE technique both the extraction time and solvent consumption were reduced by about 50% compared to original and modified Jensen methods.

3.4.2. PFE as an Alternative Method for Extraction of Pesticide Residues from Fatty Matrix (Paper II) In Paper II, a PFE multi-residue method has been validated for the determi-nation of 25 pesticides and metabolites that cover a wide range of polarity from rape seed. The reliability and efficiency of the PFE technique was in-vestigated comparing the recoveries with conventional LLE technique. Fi-nally incurred samples containing vinclozolin and iprodione were extracted and compared using LLE, PFE and SFE. The extracts obtained were analyzed on a GC using different detectors after clean-up by GPC.

When pesticide residues have to be determined from fatty matrices the ex-traction is associated with several analytical difficulties. Since the matrix is complex, like in the case of rape seed with a high lipid content of 43%, the recovery of very nonpolar fat soluble pesticides from the lipid-rich matrix is troublesome. Clean-up of the extracts using GPC has shown to be necessary in order to remove the co-extracted lipids. Purification of the extracts by GPC resulted in a considerable decrease in the lipid content, Figure 7. Thus, the interpretation of the chromatograms was simplified and the lifetime of the GC column could be extended. Alumina was tested as an adsorbent to retain the co-extracted lipids by directly adding alumina inside the cell. De-spite the decrease in lipid content in the extracts, alumina as a fat retainer was abandoned since most of the target analytes were also strongly ad-sorbed.

25

Figure 7. Residual lipids in rape seed (mg g-1) after PFE extraction and GPC puri-fication (n=3) at 600C, 10 min, hexane saturated with acetonitrile,10 MPa, 60% flush volume and 1 min N2 purge. The error bars represent RSD value.

The quantification was performed using standards dissolved in residue free rape seed extract to eliminate the possible matrix effect on the recovery. Most recoveries were within the acceptable range, 70-110% (Paper II, Table 2).

A comparison of extraction efficiency between the PFE and LLE was conducted. The comparison of the extraction scheme is shown in Paper II, Figure 3. Both methods recover the studied pesticides within acceptable levels, with overall recoveries of 96% for LLE and 91% for PFE. The LLE technique gave more precise results with lower RSD values. When using fortified samples, there is a risk that the method does not imply the true re-covery since the spiking on samples is performed with small amounts and the penetration and binding of analytes into the matrix is not fully controlled. Therefore, the PFE method was tested using residues from agriculturally incurred samples and the results were compared to the LLE and SFE meth-ods. The concentrations obtained are in good agreement with low RSD val-ues of 1-4%, Table 1. Using PFE, both the extraction time and solvent con-sumption were substantially reduced by 90% compared to LLE. PFE also allows multiple samples to be extracted sequentially in an automated system. However the main disadvantage of the current PFE method is that a sample clean-up is still required as an additional step.

Table 1. Extraction of pesticide residues from incurred samples of rape seed with LLE, SFE and PFE determined by using GC/ECD and GC/MS/ITD.

Found concentration/mg kg-1 (RSD, %)

Sample Pesticide LLE SFE PFE 1 (n=4) Vinclozolin 0.02 (1.1) 0.02 (2.7) 0.02 (4.3)

Iprodione 0.21 (2.5) 0.15 (1.7) 0.14 (3.2) 2 (n=2) Vinclozolin 0.10 0.10 0.10

050

100150200250300350400450500

Before PFE After PFE After GPC

Resid

ual l

ipid

s (m

g g

-1)

26

4. Analysis of Intact Phosphatidylcholine, Sphingomyelin and Sulfatide Molecular Species Using LC/ESI/MS and ESI/MS/MS

The interest in the analysis of lipids in general and phosphatidylcholine, sphingomyelin and sulfatides in particular is continuously increasing due to the importance of these molecules related to various disorders. Phosphati-dylcholine, sphingomyelin and sulfatide molecular species exist in nature as a complex mixture of closely related components which differ in the fatty acid chain length, degree of unsaturation and hydroxylation. Based on the fatty acid composition of the molecules each class can be separated into mo-lecular species 6. These species differ greatly in their chemical and biologi-cal properties. They determine the interaction of the molecule in the cell membrane and their identification and characterization has been of great interest. Various analytical methods have been employed to separate and analyze individual molecular species from its “intact” (underivatized or un-modified) form. The general advantages analyzing a lipid in its intact form have been described as 6, 70, 71:

Elimination of the time consuming derivatization step Possibility to study and follow biosynthesis, metabolism, turnover and transport of the molecular species Avoiding the danger of rearrangement of the FA chains during derivati-zationUsefulness in the study of the immunogenicity of sphingolipid (sulfati-des)

This section of the thesis is intended to give an overview of the analytical methods such as RP-LC, ESI/MS and ESI/MS/MS (Paper V) and RP-LC/ESI/MS and RP-LC/ESI/MS/MS (Papers III and IV) used for the analy-sis of molecular species of intact PLs and sphingolipids (mainly PC, SM and sulfatide molecular species). The advantages and drawbacks of each analyti-cal method will be discussed.

27

4.1. Reversed Phase LC of Intact PC and SM Thin layer chromatography (TLC), LC, and GC are common techniques used for separation of intact or derivatized PL molecular species. A review of derivatization methods for PL and SM molecular species analysis using these methods was recently presented by Wang et al. 72 and Toyo'oka 73.Although most chemical derivatizations improve selectivity, resolution and sensitivity, they are often very time and reagent consuming.

PL and sphingolipid analysis typically falls into two parts. The more gen-eral analysis determines each PL classes (e.g. PC, PE, PS or PI) including SM based on differences on the hydrophilic portion of the lipid. The order in which the lipid classes elute and separate depends primarily on the polarity of the head group adsorption to the normal phase material such as silica gel, 74-77, Figure 8A. More specific analysis characterize the membrane of one class, which share the common hydrophilic head group but differ in their aliphatic residue (e.g. PC, SM and sulfatides, Papers III-V). Most PL mo-lecular species analyses are carried by RP-LC of intact or after derivatiza-tion. In this section only analysis of intact PC and SM molecular species will be discussed.

One or two dimensional silver ion TLC has been used in the past to sepa-rate PC and SM molecular species. However, further analysis by LC and/or MS is required for specific molecular species analysis because of its low resolution and poor reproducibility. Silver ion liquid chromatography (Ag-LC) either silica gel impregnated with silver nitrate or by binding silver ion to an ion exchange phase has been used for lipid molecular species analysis 6, 78. The principle of the method is that silver ions interact reversibly with the pi electrons of the double bonds (cis more strongly than trans) to form polar complexes. The greater the number of double bonds in the FA chain, the stronger the complex formation and the longer it is retained 6, 78-80. It enables the separation of the molecular species according to the number, geometry and position of the double bonds in FA chain 6, 78. The mecha-nisms of the interaction between Ag-LC and unsaturated FA have been stud-ied in detail by Nikolova-Damyanova et al. 79. They concluded that silver ion simultaneously interacts with the double bonds of the FA and other elec-tron rich moiety such as the ester group. In general Christie and his co-workers made the greatest progress in developing Ag-chromatographic sys-tems 6, 81-83.

28

Hydrophilic head group

Hydrophobic FA chainR1 R2

(B) Reversed phase (A) Normal phase

Figure 8. Separation of PLs using liquid chromatography A) PL class separation using normal phase. B) PL molecular species separation using RP. The separation is mainly due to specific hydrophobic interaction of the FA chain with the RP.

RP-LC is widely used for the separation of lipid molecular species. In Pa-pers III and IV using a RP column 200x0.2 mm (Paper III) and 200x0.5 mm (Paper IV) packed with YMC-GEL ODS A C18 120 Å, 5 µm, the PC and SM were separated into their molecular species. The separation is mainly based on the hydrophobic interaction of the FA tail group with the RP pack-ing material, Figure 8B. PC and SM have higher polarity of the head group in comparison for e.g. to PE. As the polarity of the head group increases the efficiency of the chromatography decreases due to interactions between po-lar groups of molecules immobilized in the stationary phase (SP) and those dissolved in the mobile phase (MP) as well as between the polar groups of the molecular species and any exposed silanols on the RP column 84. This problem can be reduced by the appropriate choice of SP and MP. Among the different MPs studied a mixture of methanol/tetrahydrofuran/water (80:15:5) in 0.1% formic acid at a flow rate of 2 µL/min provided the best separation of the molecular species. It has the advantage that isocratic elution and there-fore a simple and inexpensive LC pump can be used. Furthermore, the ab-sence of non-volatile components (like salts) in the MP allows subsequent identification by ESI/MS and isocratic elution reduces the ESI fluctuation considerably that could be caused by changes in gradient elution. It was found that the elution order of the molecular species within each class de-pends mainly on the length and the number of double bond on the FA chain. Thus the longer and the more saturated the FA chain, the longer the retention time of molecular species. Each double bond reduces the apparent FA chain length by approximately less than two carbon atoms 85, 86.

29

The main problem of the developed method was a shift in retention time of the molecular species after about 40-50 injections of the real sample ex-tract. This could possibly be due to the lipid materials and artifacts built up in the column. It was thus necessary to wash and equilibrate the column for 5 min with the MP before the next injection could be performed. The shift in retention time may not be considered as a main problem if MS is used as a detector. In Papers III and IV, MS was used as a detector and the identity of the chromatographic peaks could be easily confirmed even if there would be a shift in retention time. For the real sample analysis in Paper IV the column was changed to a larger internal diameter (0.5 mm, at a flow rate of 15 µL/min) in order to avoid frequent plugging of the column, variation in re-tention time from run to run and to accommodate the large number of runs from the crude lipid extracts. Moreover, switching the MS between positive and negative mode for a while after a certain number of runs helped to stabi-lize the electrospray and improve the chromatographic peak shape by dump-ing any material build-up on the tip of the capillary transfer line (i.d. 50 µm). All the columns used in this thesis were slurry packed in our laboratory at a fraction of the cost of commercial columns and hence very cost effective.

4.2. Mass Spectrometry Mass spectrometry is a powerful detection technique that enables separation and characterization of substances according to their mass-to-charge ratio (m/z). Its essential components include a sample inlet, ion source, mass ana-lyzer, detector and data handling system. The combination of sensitivity, selectivity, speed and its ability for structural information of unknown lipid compound makes MS an ideal method for intact lipid molecular species analysis. It can be performed by direct infusion (Paper V) or coupled to a separation technique such as RP-LC (Papers III-IV).

4.2.1. The Mechanism of ESI/MS ESI/MS, first conceived in the 1960s by chemistry professor Malcolm Dole but was put into practice by Fenn et al. 87, 88, is a technique for transforma-tion of ions from a liquid to a gas phase. It is a simple method that operates at atmospheric pressure and ambient temperature. The principal mechanism of ESI is under investigation and still not fully understood. Many theories of the physical processes involved in ion generation during ESI have been pro-posed and validated in some detail 88-91. The principles of the ESI process can be schematically represented as shown in Figure 9. Initially, the LC ef-fluent (Papers III and IV) or direct infusion (Paper V) containing the ana-lytes of interest is pumped into the ESI ion source through a capillary tubing. The narrowed orifice at the end of the capillary tubing and the mechanical

30

forces imparted as the solution passes through the narrow orifice facilitate the formation of sprayed small droplets in the ionization chamber. If an elec-tric potential (approximately 2-5 kV) is applied between the end of the capil-lary tube and the entry into the mass analyzer, because of oxida-tion/reduction processes, the droplets can carry net charges and are directed into the mass analyzer by the applied electric field. This applied potential could be either positive or negative depending on the substance to be ana-lyzed. As shown in Figure 9, if a positive electric potential is applied to the end of a capillary tube and a negative electric potential is present at the en-trance of the mass analyzer in the positive-ion mode, droplets carry net posi-tive charges. The resulting electric field causes charge separation of ions in the solution, with positive ions moving towards the liquid surface and nega-tive ions migrating back into the liquid, Figure 9. During flight, the droplets are desolvated, by passage through a curtain gas. Thus, during desolvation, the coulombic force between ions is dramatically increased. Once this force exceeds the surface tension of the solvent, the droplets explode to form smaller droplets. This cycle is iteratively repeated until molecular ions are generated prior to their entrance into the mass analyzer 7, 89. The exact mechanism of ion formation, whether it is by ion evaporation (ion-evaporation model) or by complete solvent removal (charge-residue model) from the charged droplet is under debate, and probably different mechanisms may apply in different situations.

Figure 9. Schematic diagram of the principles of electrospray ionization in positive ion mode.

31

4.2.2. The Principles of ESI/MS of Lipids The soft ionization MS techniques such as fast atom bombardment (FAB) 92-94, field desorption (FD) 95, thermospray (TS) 96-99, matrix assisted laser desorption ionization (MALDI) 100, 101 and ESI (Papers III-V) have the abil-ity to ionize lipid molecules with out causing extensive fragmentation. Re-cently ESI has become the preferred and widely used method to ionize bio-logically important lipid molecules. In a recent review Han and Gross 7,classified the lipid classes into four groups based on their electrical propensi-ties which have been exploited for ESI/MS analyses of lipids. These include:

Anionic lipids: containing one or more net negative charge(s) at neutral pH such as sulfatide, PS, PI, phosphatidylglycerol, etc. Weak anionic lipids: carrying a net negative charge at alkaline pH such as PE, lysoPE, ceramide etc. Neutral polar lipids: are polar but electrically neutral lipids at both neu-tral and alkaline pH. This group of lipids includes choline containing lip-ids (PC, lysoPC and SM), diacylglycerol, triacylglycerol etc. Special lipids: varying in their electrical propensity. Includes acylcar-nitines, cholesterol and its derivatives including cholesterol ester and sterols, etc.

This part will only focus on the ESI/MS behavior of the choline containing lipids PC and SM (Papers III and IV) and sulfatide (Paper V).

Both PC and SM are choline containing species and are closely related in their ESI properties. They are characterized by the presence of quaternary nitrogen atom whose positive charge is neutralized by the negative charge of the phosphate group. The quaternary nitrogen atom readily forms an abun-dant [M+H]+ molecular ion by ESI in positive ion mode because the phos-phate anion can be protonated during the electrospray process. Furthermore, when various alkali metal ions such as Na+ are present in the electrospray solvent sodiated species [M+Na]+ are observed in a positive mode. In Paper III both the [M+H]+ accompanied by sodium adduct [M+Na]+ molecular ions were observed.

Sulfatide contains a polar galactosyl-3-sulfate that carries a negative charge and can be detected in a negative mode ESI. Abundant negative ion [M+H]- molecular species were produced by ESI/MS (Paper V).

4.2.3. MS Scan Modes A number of mass analyzers are available, e.g. quadrupole, ion trap, time of flight, ion cyclotron resonance and sector instruments, which separate charged species in vacuum depending on their m/z ratio. Apart from separa-tion, the mass analyzers can also be used for fragmentation of the charged

32

species. The most widely used are the triple quadrupole instrument (Papers III-V). In a triple quadrupole instrument the middle field free quadrupole focuses and transmits almost all ions and may be used as a collision cell for fragmentation reaction. The type of fragmentation occurring in quadrupole instruments is referred to as collision induced dissociation (CID). As a result of the collision, the internal energy of the ion may be increased by conver-sion of kinetic energy into internal energy 102, 103. The choice of collision gas, its pressure and the collision energy affect the degree of fragmentation. The fragment ions are then analyzed in the second mass analyzer. In addition to structural information tandem MS provides higher sensitivity and reduced chemical background by selecting different scan modes. In Paper III, in or-der to obtain structural information of the PC and SM molecular species, tandem MS was performed at optimized collision energies on the triple quadrupole instrument. Product ion scans of standard PC and SM of the [M+H]+ and [M+Na]+ ions were performed, where the first quadrupole is locked to a selected m/z ratio and the third is scanning the daughter ions pro-duced in the collision cell. Figures 10 and 11 show the product ion mass spectra for the standard PC16:0/18:1 species of the [M+H]+ ion at m/z 760.6 and the [M+Na]+ ion at m/z 782.6, respectively. The fragment ions are con-sistent with the fragment structures proposed in the literature 104, 105. The CID product ion spectrum differs depending on the choice of the parent ion. The CID product ion spectrum of [M+H]+ produced few and low abundant fragment ions in comparison to the [M+Na]+. However, both tandem spectra of [M+H]+ and [M+Na]+ ions contain a characteristic intense ion at m/z 184, corresponding to the head group, phosphocholine [H2PO4CH2CH2N(CH3)3]+.CID of the [M+H]+ ion produced low intensity fragment ions at m/z 577, 504 and 478 due to loss of [HPO4CH2CH2N(CH3)3], sn-1 FA (16:0) and sn-2 FA (18:1), respectively. The fragment ions at m/z 522 and 496 resulted from the loss of alkyl ketenes [CH3(CH2)13CH=C=O] and [CH3(CH2)7CH=CH(CH2)6CH=C=O], respectively.

The CID spectra of sodiated standard PC16:0/18:1 species is shown in Figure 11. Neutral loss of trimethylamine, N(CH3)3 [M+Na-59]+ yielded an ion at m/z 723. Ions at m/z 599 [M+Na-183]+ and 577 [M+Na-204]+ are through the loss of [HPO4CH2CH2N(CH3)3] and [NaPO4CH2CH2N(CH3)3]respectively. Neutral loss of the sn-1 FA substituents as free FA [M+Na-256]+ or as sodium salt [M+Na-278]+ are observed at m/z 526 and 504, re-spectively. In addition, neutral loss of the sn-2 FA substituents as free FA [M+Na-282]+ or as sodium salt [M+Na-304]+ are observed at m/z 500 (weak signal) and 478, respectively. The fragment ions at m/z 239 and 265 repre-sent the acylium ions of [CH3(CH2)14C=O+] and [CH3(CH2)7CH=CH(CH2)7C=O+], respectively. Losses of trimethylamine plus palmitic acid or oleic acid are observed at m/z 441 and 467 respectively. As confirmed by Hsu et al. 104 using lithiated adducts of standard PC16:0/18:1 and PC18:1/16:0 comparison of the relative abundance of the

33

ions at m/z 441 and 467 in Figure 11 indicates that the abundance of the ions reflecting loss of trimethylamine plus the sn-1 substituent exceeds that of the ion reflecting loss of trimethylamine plus the sn-2 substituent. This permits the assignment of FA moieties as sn-1 or sn-2. However Ho and Huang 105

suggest that an average of over 40 spectra is required to produce a consistent abundance ratio that lead to correct assignment of the FA moieties. It can be concluded that the sodium adduct greatly improved the information content of the CID spectra by yielding diagnostic fragment ions that provide more detailed information about the structure of the PC in comparison to [M+H]+.

PIS can also allow the selective detection of PC and SM molecular spe-cies among other PL classes by head group scanning (by scanning for pre-cursor of m/z 184). However this scan mode is non specific between PC and SM and do not yield information about FA substituent, hence it is difficult to assign the spectra unless a separate CID product ion scan is performed in parallel.

Figure 10. Product ion mass spectra of the [M+H]+ from PC16:0/18:1 at m/z 760.6 using nitrogen as a collision gas.

Figure 11. Product ion mass spectra of [M+Na]+ from PC16:0/18:1 at m/z 782.6 using nitrogen as a collision gas.

0

50

100

0 100 200 300 400 500 600 700 800

184

760.6 (M+H)+

400 450 500 550 600

478504

496

522577

m/z

0

50

100

0 100 200 300 400 500 600 700 800

184

760.6 (M+H)+

400 450 500 550 600

478504

496

522577

0

50

100

0 100 200 300 400 500 600 700 800

184

760.6 (M+H)+

0

50

100

0 100 200 300 400 500 600 700 800

184

760.6 (M+H)+

184

760.6 (M+H)+

400 450 500 550 600

478504

496

522577

400 450 500 550 600

478504

496

522577

400 450 500 550 600

478504

496

522577478

504

496

522577

m/z

184782.6 (M+Na)+

577

599 723

146

0

50

100

0 100 200 300 400 500 600 700 800

200 250 300 350 400 450 500 550

467

441

526

265239 504

478

m/z

184782.6 (M+Na)+

577

599 723

146

184782.6 (M+Na)+

577

599 723

146

0

50

100

0 100 200 300 400 500 600 700 800

200 250 300 350 400 450 500 550

467

441

526

265239 504

478

0

50

100

0 100 200 300 400 500 600 700 800

200 250 300 350 400 450 500 550

467

441

526

265239 504

478

200 250 300 350 400 450 500 550

467

441

526

265239 504

478

467

441

526

265239 504

478

m/z

34

When single quadrupole instrumentation is used, fragmentation reaction can be induced in the source region by elevating the cone-to-skimmer poten-tial difference. Protonated molecules desorbed from the electrosprayed drop-lets are accelerated between the cone and skimmer and can undergo CID in collision reaction with residual gas molecules 106. This process is called up-front fragmentation or nozzle-skimmer fragmentation. In Papers III and IV up-front fragmentation was performed in order to confirm the assignment of the chromatographic peaks from the biological sample. The orifice potential was varied between the optimal settings for quantitation (30 V) and up-front fragmentation (120 V) for structural confirmation. Figures 12A and B repre-sent the up-front fragmentation mass spectra of the deuterium labeled 1,2-Dipalmitoyl-sn-Glycero-3-Phosphocholine (d4-16:0/16:0-PC) standard at orifice potentials of 30 V and 120 V respectively. The 1,2-Dipalmitoyl-sn-Glycero-3-Phosphocholine (d4-16:0/16:0-PC) was deuterium labeled at its head group. When a low orifice potential was applied the [M+H]+ ion m/z738.6 was dominant and used for quantification. The spectrum obtained at the higher orifice potential showed substantial fragmentation as indicated by the presence of the dominant phosphocholine ion at m/z 188.6 (a fragment of the deuterium labeled head group IS). The results obtained from the real sample extracts also showed a dominant [M+H]+ ion (in this case m/z 760.6, Figure 12C) at low orifice potential, and at high orifice potential the phos-phocholine fragment ion m/z 184.6, becomes the dominant peak in the mass spectra, Figure 12D.

Figure 12. Up-front fragmentation of the deuterium labeled 1,2-Dipalmitoyl-sn-Glycero-3-Phosphocholine (d4-16:0/16:0) standard and 16:0/18:1 from human brain extract. (A) d4-16:0/16:0 at low orifice potential (30 V) (B) d4-16:0/16:0 at high orifice potential (120 V), (C) 16:0/18:1 from human brain extract at low orifice potential (30 V), and (D) 16:0/18:1 from human brain extract at high orifice poten-tial (120 V).

Intensity (cps)

0

5000000

150 250 350 450 550 650 750 850

m/z

(M+H)+

738.6 APhosphocholine ion

188.6

Intensity (cps)

0

3000000

150 250 350 450 550 650 750 850

m/z

D

(M+H)+

760.6

Phosphocholine ion 184.6

Intensity (cps)

0

4500000

160 260 360 460 560 660 760 860 960

m/z

(M+H)+

760.6

Phosphocholine ion 184.6

C

Intensity (cps)

0

5000000

150 250 350 450 550 650 750 850

m/z

B

(M+H)+

738.6

Phosphocholine ion188.6

35

In Paper V, sulfatide molecular species were analyzed by PIS. PIS is based on independent operation of two mass analyzers positioned inside a tandem MS at both sides of the collision cell. The second analyzer (quadru-pole Q3) is usually set to transmit ions with a certain constant m/z, while the first mass analyzer (quadrupole Q1) is scanning. Ions passing through the first mass analyzer Q1 fragment in the collision cell are recorded by the de-tector only if they produce a fragment with specific m/z at which the second analyzer has been set. Therefore, PIS selectively detects precursor ions that produce the same characteristic fragment ion upon their CID even if a very complex mixture is injected into the mass spectrometer. In Paper V, the ac-quisition of precursor ion spectra of the sulfate fragment ion allowed detec-tion of major sulfatide molecular species. CID was performed with nitrogen as a collision gas. The intense signal at m/z 97 corresponds to the HSO4

- ion. Thus, sulfatide can be detected by scanning for the precursors of m/z 97 in the negative ion mode 107-109. First, synthetic IS 14:0-sulfatide was used to optimize the instrument operational parameters and intense [M+H]- ion was obtained at m/z 750.6 that corresponds to the 14:0-sulfatide IS.

The developed method was applied to compare endogenous sulfatide mo-lecular species accumulation in arylsulfatase A knockout (ASA -/-) developed as a model of metachromatic leukodystrophy and wild type control mice (ASA +/+) at different ages. An alteration in the composition of sulfatide molecular spe-cies was observed between the ASA -/- and ASA +/+ mice with accretion of sulfatide mainly with stearic acid in the ASA -/- mice (15% higher at the older ages). In cultivated astrocytes the proportion of short chain FAs (C16-C18) was 60%, with C18 FAs constituting 40%.

4.3. Reversed Phase LC/ESI/MS RP-LC is widely used for the separation of intact lipid molecular species using different detectors. Phospholipids and sphingolipids lack chromopho-res that permit direct specific spectrophotometric detection. Some of the common detectors used for intact molecular species analysis are ultra violet (UV) 85, 86, 110-112, evaporative light scattering detector (ELSD) 81, 113, 114

and MS (Papers III-V). LC with UV detection offers a fast and convenient method however sever limitation are imposed on selecting the MP since UV absorption of underivatized PL molecules only occur near 200 nm depending on the degree of unsaturation of the FA moieties. UV detection in the 200-210 nm range creates problem for quantification because the detector re-sponse is not the same for various molecular species and most of the com-monly used MPs absorb strongly in this region 112, 115. In ELSD detectors the column effluent is nebulized to an aerosol which is subsequently evapo-rated, leaving the solute components as fine droplets. These droplets are illuminated by a laser and the amount of scattered light is measured. Such

36

detectors are universal in their applicability, because they respond to any lipid molecular species that does not evaporate 115. Moreover both UV and ELSD do not give structural information of the molecular species and con-firmation of each peak identity must be established using other techniques such as GC/MS or tandem MS. The advantages of UV and ELSD are that they are simple, inexpensive and versatile. The ELSD has an advantage over UV in that the response is largely independent of the number and configura-tion of double bonds in the FA chains, thus making quantification more straight forward 70, 115.

In order to analyze molecular species of lipids from complex biological samples, development of a more efficient technique which can simultane-ously provide unambiguous structural information and sensitivity is required. An alternative method to overcome some of the above difficulties in the analysis of intact lipid molecular species is to use LC coupled with ESI/MS. It is one of the powerful techniques used today in the analysis of intact polar lipid molecular species. The separation power of LC into either different lipid classes or molecular species within a class together with selective MS detection makes the simultaneous determination of protonated molecules and the identification of several structural elements possible 116. The LC/MS system used in Papers III and IV is shown in Figure 13.

Figure 13. Experimental set-up of the LC/MS and LC/MS/MS system used in Papers III and IV. Pump: Jasco PU-980 operated at constant pressure, Injector: a 60 nL and 200 nL internal loop, Column: 200x0.2 mm (Paper III) and 200x0.5 mm (Paper IV) packed with 5 µm particles C18 (ODS A, YMC-GEL), capillary transfer line 50 µm i.d.

LC/MS is still limited to conditions that are suitable for ESI/MS condi-tions. There are restrictions on pH, choice of solvent, solvent additives and flow rates for LC in order to achieve optimal MS sensitivity 90. It is impor-tant to select a solvent system that renders chromatographic resolution and is compatible with ESI/MS performance. In Papers III and IV a mixture of methanol/tetrahydrofuran/water (80:15:5) in 0.1% formic acid was found to be a suitable MP at a flow rate of 2 µL/min during isocratic elution. By re-ducing the capillary LC and using lower flow rates the sensitivity can be

37

enhanced due to lower flow rates and the smaller i.d. capillaries resulting in smaller droplets and increased charge to volume ratio (increased analyte concentration). In Papers III and IV, using a capillary column, the PC and SM molecular species were separated successfully, Figure 14. Reconstructed ion chromatogram using standards of PC and SM at 60 nL injection volume in single ion monitoring mode indicated a base line separation of the species. The single ion monitoring mode showed minimum detectable quantity as low as 8 fmol using deuterium labeled 1,2-Dipalmitoyl-sn-Glycero-3-Phosphocholine (d4-16:0/16:0) IS. It also showed good linearity (r2=0.9991) over two orders of magnitude in concentration range. More over, as the col-umn i.d. is decreased, the lower volumetric flow rate through a capillary column would minimize consumption of chemicals in the MP and amount of sample consumed to perform an analysis. In Papers III and IV, 60 and 200 nL volume was injected respectively, using an internal loop injector, which gave an added advantage when working with a limited amount of sample. In Paper IV, the developed method was applied to investigate the effect of postnatal iron administration in lipid composition (PC and SM molecular species) in brains of eight groups of mice containing six individuals in each group treated with vehicle, low dose clozapine, high dose clozapine and haloperidol.

Figure 14. Reconstructed ion chromatogram for the [M+H]+ ions in a standard phosphatidylcholine and sphingomyelin mixture in single ion monitoring mode. Approximately 4 pmol of the PC and SM were injected into the system.

Peak identification1. SM 16:02. PC 16:0/18:23. SM 18:04. PC 16:0/18:15. PC 18:0/18:26. SM 20:07. PC 18:0/18:18. SM 24:1

0

250000

5 10 15 20 25 30 35

Time (min)

Inten

sity

(cps

)

1

2

3

6

4

57

8

Peak identification1. SM 16:02. PC 16:0/18:23. SM 18:04. PC 16:0/18:15. PC 18:0/18:26. SM 20:07. PC 18:0/18:18. SM 24:1

0

250000

5 10 15 20 25 30 35

Time (min)

Inten

sity

(cps

)

1

2

3

6

4

57

8

0

250000

5 10 15 20 25 30 35

Time (min)

Inten

sity

(cps

)

1

2

3

6

4

57

8

1

2

3

6

4

57

8

38

5. Data Handling

Nowadays, modern analytical instrumentations make it possible to record very large data sets in short period of time in chemical laboratories and there is a need for efficient data handling. Chemometrics involves experimental design and multivariate analysis methods that include mathematical and statistical methods for modeling and data handling.

Multivariate methods such as principal component analysis (PCA) and projections to latent structure (PLS) make up a set of mathematical tools that may be applied to the analysis and interpretation of spectroscopic data. The data structure can be represented in 1- and 2-dimentional plots 117, 118. A more detail descriptions of these methods can be found elsewhere 119, 120.PCA is a technique used to reduce the number of variables needed to de-scribe the relevant variations. It is useful for data compression and informa-tion extraction. The results must be presented so that the desired information can be easily understood (e.g. in a graph form, like in Paper III) or expressed in a few information rich parameters. In Paper III, PCA and PLS approach was used for the lipid concentration and behavioral variables to elucidate the variations between the various treatment groups.

Experimental design is useful to find out the optimal operating conditions and factors that give the best recovery with a small number of experiments 121. In Paper I, an optimization study was carried out to investigate the opti-mum values of the different variables (temperature, time and extraction sol-vent) expected to have the largest effect on the extraction efficiency of total lipid content from cod muscle.

39

6. Conclusions and Future Aspects

This thesis covers a wide range of analytical method development for lipid analysis in complex biological samples; from sample preparation using PFEand separation with RP-LC to detection by ESI/MS and tandem MS. It shows that modern analytical methods can provide new insights in the ex-traction and analysis of lipids from complex biological samples.

In Papers I, II and III, PFE was used as an alternative extraction technique in comparison to different traditional liquid extraction methods. Recoveries and precision are equivalent or better than those achieved by traditional ex-traction methods. The PFE technique has the additional advantage in terms of small solvent volume consumption, rapid extraction times and automa-tion. Using PFE, both the extraction time and solvent consumption was sub-stantially reduced by 90% (Paper I) and 50% (Paper II) compared to the traditional extraction methods. The main drawback of the PFE technique is the lack of selectivity and hence further sample clean-up steps were found to be needed (Papers I-III).

To use environmentally friendly solvents such as water, PFE in combina-tion with pressurized hot water as an extraction solvent is suggested to be a favored approach in the future. When heated above its boiling point in a closed container, water becomes a more powerful solvent 122. Ease of cou-pling to other analytical instrumentation could be an additional advantage offered by this technique in the future. Further automation and integration of extraction, clean-up and detection steps in single reliable automated process appears to be the next challenge for PFE. However it seems unlikely that future generations of the PFE technique will incorporate higher temperature capabilities (> 200oC) because most reported work lies in the normal operat-ing range of 75-150oC. This is limited by reported problems with analyte and matrix stability 1.

Functional lipidomics is an emerging field of life science and is fast be-coming important in the understanding of lipid related diseases. In Papers III-V a targeted lipidomics approach was presented to characterize and pro-file PC, SM and sulfatide molecular species using LC/MS and MS/MS. A successful separation of PC and SM molecular species was obtained with a minimum detectable quantity in the low fmol range injected on column (Pa-pers III and IV). Although some of the need for LC separation is eliminated by using advanced MS instruments which can utilize different scan modes for separation and also handle lower detection limits, the coupling of LC is

40

still the most promising approach since it can handle total lipid extracts of lower concentration and of more complex nature as an added advantage. Moreover, mass spectrometric injection through LC will often reduce ion suppression effects that are introduced by other components in the crude lipid extract or competition between the analytes (lipids). In Paper V, a tan-dem MS in PIS mode was used to analyze the developmental profile of molecu-lar species brain sulfatide accumulation in ASA deficient (ASA -/-) and wild type control (ASA +/+) mice.

The term “lipidomics” has just emerged in the literature due to recent technological advances in LC/MS and MS/MS and rising concerns over lipid associated disorders. One strategy to combat these diseases is to understand how the composition of cellular lipids and lipid membrane changes. How-ever analysis and quantification of these changes is complicated by the struc-tural diversity, instability, the wide range of lipid concentration and the spe-cialized biochemical systems for their biosynthesis and metabolism 123.Thus, it seems that lipidomics might benefit from a more targeted approach that will focus on a specific group of lipids. Combining data from the analy-sis of various lipids is likely to produce a complete lipidomic map, possibly with the help of bioinformatics methodology that will be developed in the future for lipid profiling.

41

7. Acknowledgements

I would like to express my sincere gratitude to:

My supervisors Assoc. Prof. Jonas Bergquist, Prof. Karin Markides and Monica Waldebäck (senior lecturer) for their limitless effort and guidance, helpful and interesting discussions, encouragement and making me see the positive side of the world throughout my graduate study period. Karin Markides for always making me concentrate on my studies whenever there have been ups and downs in the source of funding. Jonas Bergquist for in-troducing me to neuroscience and lipid mass spectrometry and for always being there to listen to me even if I knock at his office door late after 18:00. Monica Waldebäck for introducing me to the research world in general and PFE technique in particular and for all the information and social life she provided me during my first arrival to Sweden.

Dr. Dan Bylund, Assoc. Prof. Jan-Eric Månsson and Assoc. Prof. Anders Fredriksson for being generous with their time, knowledge and valuable discussions in all lipid/MS projects. Prof. Göran Odham and Ulla Eriksson from Stockholm University for all valuable discussions and meetings in lipid extraction. Assoc. Prof. Rolf Danielson for helpful discussions about chemometrics and data handling. Andreas Dahlin for his help with the graphics. All my co-authors for their interesting and fruitful discussions.

Dr. Ardeshir Amirkhani and Sara Bergström for introducing me to the prac-tical lab work in mass spectrometry and column packing respectively. Daniel Bäckström for his kind assistance in all computer related problems.

All staff members, present and past graduate students at the department of Analytical Chemistry who always remained positive to give any assistance and for providing a friendly and comfortable atmosphere to work in. I really enjoyed the social life and extracurricular activities at the department such as Friday evening Spray Bar, innebandy club (even if I am not as good as….), summer soccer game, hiking in the north of Sweden and the recently intro-duced “Film Kväll”.

Barbro Nelson and Yngve Öst for all administrative and technical assistance.

42

My home University, Asmara University in Eritrea for giving me the oppor-tunity to study in Sweden. SIDA/SAREC for the financial assistance up to my Licentiate degree and Uppsala University for finding financial means to complete my studies.

All staff members of the International Science Programme (ISP) for their assistance in all technical and administrative matters in Sweden.

All my friends, especially the SAUA-group and all Eritrean friends residing in Sweden for their continuous encouragement, advice, motivation, social life and for making me not miss home. Especially Asmerom with his wife Selamawit for the hospitality and warm reception whenever I stay at their home in Stockholm, Husby. All my best friends back home in Asmara, espe-cially Zekarias E., Semere Tekle and Semere T/mariam.

My Swedish family Mr. Börje Nordh for introducing me to the Swedish culture, for always being there to drive me anywhere at my convenience and for the wonderful “Fika” company and refreshment at times my head was overwhelmed with a lot of Chemistry. It was fun every year helping you decorate your Christmas tree. One thing though, would you please stop beat-ing me in “Mini Golf” and let me win only for a single time!!

Finally, my mother Emuna as well as my brothers and sisters for their con-tinuous motivation, inspiration and moral support throughout my graduate study period.

43

8. Summary in Swedish

Lipider är inblandade i många vitala biologiska processer i djur, växter och mikroorganismer. Störningar i lipidmetabolismen kan medföra en mängd olika problem och därför är utvecklandet av metoder för extraktion och ana-lys av lipider avgörande för vår förståelse av dessa system.

Många av de traditionella provextraktionsmetoderna har blivit den sva-gaste länken i modern kemisk analysmetodik. Traditionella metoder kan vara tillförlitliga men långsamma, kräva stora mängder organiska lösningsmedel, vara svåra att automatisera och därför mycket arbetskrävande. Under de senaste decennierna har drivande krafter, såsom behovet av snabbare analy-ser, ökad produktivitet, hänsynstagande till kostnader för analyser, säkerhet för laboratoriepersonal och reduktion av miljöbelastande utsläpp, uppmunt-rat kemister att utveckla nya extraktionstekniker. Superkritisk vätskeextrak-tion, SFE, mikrovågsextraktion, MAE, och trycksatt lösningsmedelsextrak-tion (PFE) är moderna tekniker, där tryck och temperatur kontrolleras och på så sätt förändrar diffusionsegenskaperna hos lösningsmedlen och massöver-föringen under själva extraktionen. Dessa tekniker ger överlägset snabbare extraktioner och system som är enklare att automatisera. De ger dessutom bättre eller likvärdiga utbyten, möjlighet till selektivitet samt eliminerar eller kraftigt minskar volymen av hälsovådliga och miljöbelastande lösningsme-del. I denna avhandling studeras tillförlitlighet och effektivitet av PFE för extraktion av totalt lipidinnehåll från torsk- och sillmuskel (Artikel I), pesti-cider i rapsolja (Artikel II), och fosfatidylkolin (PC) och sphingomyelin (SM) från human hjärnvävnad (Artikel III).

Ur ett ekologisk perspektiv är det mer relevant att uttrycka den kvarva-rande mängden lipofila miljögifter i mängden fett snarare än som total våt vikt. Åtskilliga jämförande laboratorier har visat att den ursprungliga Jensen-extraktionsmetoden för totala lipider och organiska föroreningar ger till-fredsställande utbyten när den används för fettrik vävnad från marina orga-nismer. I mager fiskmuskel (där fettinnehållet är <1%) är däremot utbytet med denna metod ca 25% under det egentliga värdet, vilket i sin tur leder till att mängden miljögifter per lipidinnehåll blir falskt förhöjt. En PFE-metod har utvecklats, där olika extraktionsvariabler (temperatur, tid och samman-sättning av en blandning av lösningsmedel) har undersökts i ett flerfaktor försök. Den utvecklade PFE-metoden har beskrivits (Artikel I) och resultaten har jämförts med olika traditionella extraktionsmetoder. Den utvecklade

44

metoden sparade tid (50%) och lösningsmedelsvolymen minskade med, 50%.

I Artikel II beskrivs en multimetod för att extrahera pesticider, allt från polära till helt opolära, i rapsfrön. Svårigheten är att extrahera de opolära pesticiderna ur de fettrika fröna utan att få med fett från rapsfröna, då detta fett stör den slutliga analysen. Den utvecklade metoden har jämförts med den metod Livsmedelsverket använder idag. Utbytena är likvärdiga men den utvecklade metoden är väsentligt snabbare samt minskar lösningsmedelsan-vändningen med 90 %.

Vidare har i Artikel III en metod utvecklats för separation och karakteri-sering av PC och SM genom användandet av reversed phase kapillärvätskek-romatografi (RP-LC) kopplat till elektrosprayjonisationsmasspektrometri (ESI/MS). Därutöver jämfördes extraktionseffektiviteten av traditionell Folch-metod och PFE för human hjärnvävnad.

I Artikel IV användes den utvecklade metoden för att undersöka lipid-sammansättningen i hjärnvävnad från möss som behandlats med placebo eller neuroleptika (låg eller hög dos clozapine eller haloperidol). Effekten av postnatal järnbehandling på lipidsammansättning och beteende undersöktes i djuren. Därefter studerades om dessa effekter påverkades av subkronisk neu-roleptikabehandling.

I Artikel V användes tandemmasspektrometri för s.k. precursor ion scan (PIS) för att karakterisera sulfatidinnehållet i transgena möss (knock-out) som saknar arylsulfatas A (ASA). Denna musstam användes som en sjuk-domsmodell för den humana monogena sjukdomen metakromatisk leukody-strofi (MLD) där funktionen av det lysosomala enzymet ASA är nedsatt. Våra analyser påvisar en förändrad sulfatidsammansättning hos bristdjur jämfört med kontrolldjur.

45

9. References

1. Handley, A. J. Extraction Methods in Organic Analysis; Sheffield Academic Press Ltd: Sheffield, 1999.

2. Majors, R. E. LCGC 1999, 17, S8-S13. 3. Bignert, A.; Göthberg, A.; Jensen, S.; Litzen, K.; Odsjö, T.; Olsson, M.; Reu-

tergårdh, L. Sci. Total Environ. 1993, 128, 121-139. 4. Quasimeme ; Quasimeme Project Office, Marine Laboratory: Aberdeen, UK,

1995.5. Piomelli, D. Prostaglandins 2004, In Press.6. Christie, W. W. Lipid Analysis: Isolation, Separation, Identification and Struc-

tural Analysis of Lipids, Third ed.; The Oily Press: Dundee, 2003.7. Han, X.; Gross, R. W. Mass Spectrom. Rev. 2005, 24, 367-412. 8. Han, X.; Gross, R. W. J. Lipid Res. 2003, 44, 1071-1079. 9. Han, X.; Yang, J.; Chang, H.; Ye, H.; Gross, R. W. Anal. Biochem. 2004, 330,

317-331. 10. Spener, F.; Lagarde, M.; Record, M. Eur. J. Lipid Sci. Technol. 2003, 105, 481-

482. 11. Lagarde, M.; Geloen, A.; Record, M.; Vance, D.; Spener, F. Biochim. Biophys.

Acta 2003, 1634, 61-61. 12. Wilson, J. F. The Scientist 2003, 17, 34-36. 13. Prestwich, G. D. Prostaglandins 2004, In Press.14. Martindale, D. The Scientist 2003, 17.15. Singer, S. J.; Nicolson, G. L. Science 1972, 175, 720-731. 16. Ekroos, K. Ph.D. Dissertation, Max Planck Institute of Molecular Cell Biology

and Genetics, Dresden, 2003.17. Serhan, C. N. Prostaglandins 2004, In Press.18. Albert, B.; Bray, D.; Leis, J.; Raff, M.; Robert, K.; Watson, J. D. Molecular

Biology of the Cell, Third ed.; Garland: New York, 1994.19. http://cellbio.utmb.edu/cellbio/membrane_intro.htm . 20. Siegel, G. J.; Agranoff, B. W.; Albert, R. W.; Molinoff, P. B. Basic Neuro-

chemistry Molecular, Cellular and Medical Aspects, Fifth ed.; Raven Press: New York, 1994.

21. Brugger, B.; Erben, G.; Sandhoff, R.; Wieland, F. T.; Lehmann, W. D. Proc.Natl. Acad. Sci. U.S.A. 1997, 94, 2339-2344.

22. Horrobin, D. F.; Glen, A. I. M.; Vaddadi, K. Schizophr. Res. 1994, 13, 195-207. 23. Ishizuka, I. Prog. Lipid Res. 1997, 36, 245-319. 24. Merrill, A. H., Jr.; Schmelz, E.-M.; Dillehay, D. L.; Spiegel, S.; Shayman, J.

A.; Schroeder, J. J.; Riley, R. T.; Voss, K. A.; Wang, E. Toxicol. Appl. Phar-macol. 1997, 142, 208-225.

25. Vos, J. P.; Lopes-Cardozo, M.; Gadella, B. M. Biochim. Biophys. Acta 1994,1211, 125-149.

26. Sastry, P. S. Prog. Lipid Res. 1985, 24, 69-176. 27. Barenholholz, Y.; Thompson, T. E. Chem. Phys. Lipids 1999, 102, 29-34.

46

28. Hevonoja, T.; Pentikäinen, M. O.; Hyvönen, M. T.; Kovanen, P. T.; Ala-Korpela, M. Biochim. Biophys. Acta 2000, 1488, 189-210.

29. Niemeä, P.; Hyvönen, H. T.; Vattulainen, I. Biophys. J. 2004, 87, 2976-2989. 30. Norton, W.; Auttilio, L. J. Neurochem. 1966, 13, 213-222. 31. Hawthorne, S. B.; Miller, D. J. J. Chromatogr. A 1987, 403, 63-76. 32. Bøwadt, S.; Hawthorne, S. B. J. Chromatogr. A 1995, 703, 549-571. 33. Chester, T. L.; Pinkston, J. D.; Raynie, D. E. Anal. Chem. 1992, 64, 153R-

170R. 34. Gazler, K.; Bati, J.; Valko, K. J. Chromatogr. 1986, 371, 299-306. 35. Lopez-Avilla, V.; Young, R.; Kim, R.; Beckert, W. F. J. Chromatogr. Sci.

1995, 33, 481-484. 36. Richer, B. E.; Jones, B. A.; Ezzell, J. L.; Porter, N. L. Anal. Chem. 1996, 68,

1033-1039. 37. Turner, C.; King, J. W.; L., M. J. Chromatogr. A 2001, 936, 215-237. 38. Berg, H.; Mågärd, M.; Johansson, G.; Mathiasson, L. J. Chromatogr. A 1997,

785, 345-352. 39. Johnson, R. B.; Barnett, H. J. Aquaculture 2003, 216, 263-282. 40. Snyder, J. M.; King, J. W.; Jackson, M. A. J. Chromatogr. A 1996, 750, 201-

207. 41. King, J. W.; Johnson, J. H.; Friedrich, J. P. J. Agric. Food Chem. 1989, 37,

951-954. 42. Lembke, P.; Engelhardt, H. Chromatographia 1993, 35, 509-516. 43. Motohashi, N.; Nagashima, H.; Párkányi, C. J. Biochem. Bioph. Meth. 2000,

43, 313-328. 44. Bartle, V. J. D.; Clifford, A. A. J. Chromatogr. A 1993, 642, 283-299. 45. Rissato, S. R.; Galhiane, M. S.; Knoll, F. R. N.; Apon, B. M. J. Chromatogr. A

2004, 1048, 153-159. 46. Hartonen, K.; Bøwadt, S.; Dybdahl, H. P.; Nylund, K.; Sporring, S.; Lund, H.;

Oreld, F. J. Chromatogr. A 2002, 958, 239-248. 47. Berset, J. D.; Holzer, R. J. Chromatogr. A 852, 852, 545-558. 48. Berg, B. E.; Lund, H. S.; Kringstad, A.; Kvernheim, A. L. Chemosphere 1999,

38, 587-599. 49. Bøwadt, S.; Mazeas, L.; Miller, D. J.; Hawthorne, S. B. J. Chromatogr. A 1997,

785, 205-217. 50. Fernández, I.; Dachs, J.; Bayona, J. M. J. Chromatogr. A 1996, 719, 77-85. 51. Tilio, R.; Krishnan, K.; Kapila, S.; Nam, K. S.; Facchetti, S. Chemosphere

1994, 29, 1849-1858. 52. Paré, J. R. J.; Jacqueline, M. R. B.; Stafford, S. S. TrAC-Trend Anal. Chem.

1994, 13, 176-184. 53. Eskilsson, C. S.; Björklund, E. J. Chromatogr. A 2000, 902, 227-250. 54. LeBlanc, G. LCGC 1999, 17, S30-S37. 55. Focant, J. F.; Pirard, C.; Pauw, E. D. Talanta 2004, 63, 1101-1113. 56. Fitzpatrick, L. J.; Zuloaga, O.; Etxebarria, N.; Dean, J. R. Rev. Anal. Chem.

2000, 19, 75-122. 57. Waldebäck, M. Licentiate Thesis, Uppsala University, Uppsala, 1999.58. Richter, B. E. LCGC 1999, 17, S22-S28. 59. Lundstedt, S.; Bavel, B. V.; Haglund, P.; Tysklind, M.; Öberg, L. J. Chroma-

togr. A 2000, 883, 151-162. 60. Schantz, M. M.; Nichols, J. J.; Wise, S. A. Anal. Chem. 1997, 69, 4210-4219. 61. Saim, N.; Dean, J. R.; Abdullah, M. P.; Zakaria, Z. Anal. Chem. 1998, 70, 420-

424. 62. Lou, X.; Janssen, H.-G.; Cramers, C. A. Anal. Chem. 1997, 69, 1598-1603.

47

63. Björklund, E.; Nilsson, T. Trends Anal. Chem. 2000, 19, 434-445. 64. Björklund, E.; Bøwadt, S.; Nilsson, T.; Mathiasson, L. J. Chromatogr. A 1999,

836, 285-293. 65. Dionex. ASE 200 Accelerated Solvent Extractor Operator´s Manual, Dionex

Corporation: United States of America, 1996.66. Jensen, S.; Häggberg, L.; Jorundsdottir, H.; Odham, G. J. Agric. Food Chem.

2003, 51, 5607-5611. 67. Schäfer, K. Anal. Chim. Acta. 1998, 358, 69-77. 68. Boselli, E.; Velazco, V.; Caboni, M. F.; Lercker, G. J. Chromatogr., A 2001,

917, 239-244. 69. Richer, B. E.; Covino, L. LCGC 2000, 18, 1068-1073. 70. Olsson, N. U.; Salem, N., Jr. J. Chromatogr. B 1997, 692, 245-256. 71. Robins, S. J.; Patton, G. M. J. Lipid Res. 1986, 27, 131-139. 72. Wang, Y.; Krull, I. S.; Liu, C.; Orr, J. D. J. Chromatogr. B 2003, 793, 3-14. 73. Toyo'oka, T. J. Chromatogr. B 1995, 671, 91-112. 74. Wang, F.; Zhhao, Y.; Wang, P. J. Chromatogr. Sci. 2003, 41, 142-144. 75. Christie, W. W.; Urwin, R. A. J. High Resol. Chromatogr. 1995, 18, 97-100. 76. Christie, W. W.; Gill, S.; Nordbäck, J.; Itabashi, Y.; Sanada, S.; Slabas, A. R.

Phytochem. Analysis 1998, 9, 53-57. 77. Kaduce, T. L.; Norton, K. C.; Spector, A. A. J. Lipid Res. 1983, 24, 1398-1403. 78. Dobson, G.; Christie, W. W.; Nikolova-Damyanova, B. J. Chromatogr. B 1995,

671, 197-222. 79. Nikolova-Damyanova, B.; Christie, W. W.; Herslöf, B. J. Chromatogr. A 1996,

749, 47-54. 80. Christie, W. W. Analusis 1998, 26, M34-M40. 81. Christie, W. W. J. Chromatogr. A 1988, 454, 273-284. 82. Christie, W. W. J. Lipid Res. 1994, 33, 9-18. 83. Christie, W. W.; Breckenridge, G. H. M. J. Chromatogr. B 1989, 469, 261-269. 84. Patton, G. M.; Robins, S. J. Method Enzymol. 1990, 187, 195-215. 85. Jungalwala, F. B.; Hayssen, V.; Pasquini, J. M.; McCluer, R. H. J. Lipid Res.

1979, 20, 579-587. 86. Smith, M.; Jungalwala, F. B. J. Lipid Res. 1981, 22, 697-704. 87. Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science

1989, 246, 64-71. 88. Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Mass

Spectrom. Rev. 1990, 9, 37-70. 89. Cech, N. B.; Enke, C. G. Mass Spectrom. Rev. 2001, 20, 362-387. 90. Cole, R. B. Electrospray Ionization Mass Spectrometry; Wiley: New York,

1997. 91. de la Mora, J. F.; van Berkel, G.; Enke, C. G.; Cole, R. B.; Martinez-Sanchez,

M.; Fenn, J. B. J. Mass Spectrom. 2000, 35, 939-952. 92. Griffiths, W. J.; Brown, A.; Reimendal, R.; Yang, Y.; Zhang, J.; Sjövall, J.

Rapid. Commun. Mass. Spectrom. 1996, 10, 1169-1174. 93. Korachi, M.; Blinkhorn, A. S.; Drucker, D. B. Eur. J. Lipid Sci. Technol. 2002,

104, 50-56. 94. Tavana, A. M.; Drucker, D. B.; Hull, P. S.; Boote, V. FEMS Immunol. Med.

Microbiol. 1998, 21, 57-64. 95. Murray, H. E.; Schulten, H. R. Chem. Phys. Lipids 1981, 29, 11-21. 96. Ma, Y. C.; Kim, H. Y. Anal. Biochem. 1995, 226, 293-301. 97. Kim, H. Y.; Yergey, J. A.; Salem, N., Jr. J. Chromatogr. 1987, 394, 155-170. 98. Kim, H. Y.; Salem, N., Jr. Anal. Chem. 1987, 59, 722-726. 99. Kim, H. Y.; Salem, N., Jr. Anal. Chem. 1986, 58, 9-14.

48

100. Schiller, J.; Süß, R.; Arnhold, J.; Fuchs, B.; Leßig, J.; Müller, M.; Petkovic, M.; Spalteholz, H.; Zschörnig, O.; Arnold, K. Prog. Lipid Res. 2004, 43, 449-488.

101. Jackson, S. N.; Wang, H. J.; Woods, A. S. J. Am. Soc. Mass Spectrom. 2005,16, 133-138.

102. Hoffmann, E. d. J. Mass Spectrom. 1996, 31, 129-137. 103. Kuksis, A.; Myher, J. J. J. Chromatogr. B 1995, 671, 35-70. 104. Hsu, F. F.; Bohrer, A.; Turk, J. J. Am. Soc. Mass Spectrom. 1998, 9, 516-526. 105. Ho, Y.; Huang, P. Rapid Commun. Mass Spectrom. 2002, 16, 1582-1589. 106. Griffiths, W. J. Mass Spectrom. Rev. 2003, 22, 81-152. 107. Hsu, F. F.; Bohrer, A.; Turk, J. Biochim. Biophys. Acta 1998, 1392, 202-216. 108. Marbois, B. N.; Faull, K. F.; Fluharty, A. L.; Raval-Fernandes, S.; Rome, L. H.

Biochim. Biophys. Acta 2000, 1484, 59-70. 109. Whitfield, P. D.; Sharp, P. C.; Johnson, D. W.; Nelson, P.; Meikle, P. J. Mol.

Genet. Metab. 2001, 73, 30-37. 110. Abidi, S. L.; Mounts, T. L. J. Chromatogr. A 1992, 598, 209-218. 111. Patton, G. M.; Fasulo, J. M.; Robins, S. J. J. Lipid Res. 1982, 23, 190-196. 112. McHowat, J.; JH Jones, J. H.; Creer, M. H. J. Lipid Res. 1996, 37, 2450-2460. 113. Brouwers, J. F. H. M.; Gadella, B. M.; van Golde, L. M. G.; Tielens, A. G. M.

J. Lipid Res. 1998, 39, 344-353. 114. Brouwers, J. F.; Vernooij, E. A.; Tielens, A. G.; van Golde, L. M. J. Lipid Res.

1999, 40, 164-169. 115. Christie, W. W. In Advance in Lipid Methodology-One; The Oily Press: Dun-

dee, 1992, pp 239-271. 116. Karlsson, A. A.; Michelsen, P.; Odham, G. J. Mass. Spectrom. 1998, 33, 1192-

1198. 117. Wold, S.; Berglund, A.; Kettaneh, N. J. Chemometrics 2002, 16, 377-386. 118. Jaumot, J.; Vives, M.; Gargallo, R. Anal. Biochem. 2004, 327, 1-13. 119. Esbensen, K. H. Multivariate Data Analysis - in practice, 4th ed.; CAMO ASA:

Oslo, 2000.120. Eriksson, L.; Johansson, E.; Kettaneh-Wold, N.; Wold, S. Introduction to multi-

and megavariate data analysis using projection methods (PCA & PLS);UMETRICS AB: Umeå, 1999.

121. Eriksson, L.; Johansson, E.; Kettaneh-Wold, N.; Wikström, C.; Wold, S. De-sign of Experiments; UMETRICS AB: Umeå, 2000.

122. Majors, R. E. LCGC 2002, 20, 1098-1113. 123. Balazy, M. Prostagladins 2004, 73, 173-180.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 56

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science and Technology, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology. (Prior to January, 2005, the series was published under the title "Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology".)

Distribution: publications.uu.seurn:nbn:se:uu:diva-5810

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2005