Preparation of Lipid Extracts Tissues

The following was first published by W.W. Christie, in Advances in Lipid Methodology - Two, pp. 195-213 (1993) (Ed. W.W. Christie, Oily Press, Dundee) and it is reproduced here by kind permission of P.J. Barnes & Associates (The Oily Press), who retain the copyright. Much of the information is still relevant. A recent list of references to combined extraction/methylation procedures can be accessed here. (This is not discussed below but rather in our methylation review).

Introduction

Quantitative isolation of lipids from tissues in their native state and free of non-lipid contaminants must be accomplished before analysis is attempted. Although this first step can be tedious, time-consuming and relatively uninteresting, any slackness will almost certainly result in the loss of specific components and in the production of artefacts. When appreciable amounts of free fatty acids, diacylglycerols, phosphatidic acid and certain other lipids are detected in extracts, for example, faulty storage of the tissues or of lipid isolates or inappropriate extraction conditions are generally indicated. Many solvents or solvent combinations can be used to extract lipids from tissues, but care must be taken to ensure that lipolytic and other enzymes are deactivated and that the recovery is complete. Any non-lipid contaminants that are co-extracted must then be eliminated from the recovered lipids by washing or other solvent partition procedures, before the sample can be subjected to detailed analysis. At every stage, precautions must be taken to minimize the risk of hydrolysis of lipids or of autoxidation of unsaturated fatty acids.

The choice of extraction procedure will depend on the nature of the tissue matrix, and for example, whether the sample is of animal, plant or microbial origin. Another factor is the amount of information required from the sample; many simple extraction procedures can be used for triacylglycerol-rich tissues such as adipose tissue or oil seeds if the main lipid class only is required for analysis. On the other hand, if a detailed knowledge of every minor lipid class is required, no short cuts are possible. Particular care is necessary for the more polar near-water-soluble lipids such as gangliosides or polyphosphoinositides. This review describes the principles and good practice of tissue handling, lipid extraction and elimination of non-lipid contaminants and the many potential pitfalls. The topic has been reviewed in general terms elsewhere [17,47,73-76,85,123,124], and in addition there exist many specialist reviews that deal with the extraction of specific lipid classes such as gangliosides or inositides (see the appropriate sections below).

Storage of Tissues and Preliminary Treatments prior to Extraction

All tissues, whatever their origin, should ideally be extracted immediately after removal from the living organism, so that there is little opportunity for changes to occur to the lipid components. It is of course essential that plasma or tissue samples be taken with the minimum of stress or trauma, otherwise lipolysis will occur in vivo. Discussion of appropriate surgical or collection procedures is outwith the scope of this review, but it has been dealt with in relation to plasma elsewhere [73,74]. With plant and heart or brain tissues, say, where tissue enzymes are especially active, rapid extraction is essential. When this is not feasible, the tissue should be frozen as rapidly as possible, for example with dry ice or liquid nitrogen, and stored in sealed glass containers at −20ºC in an atmosphere of nitrogen. A storage temperature of as low as −60ºC has been recommended for plasma samples of clinical origin [73], and even lower temperatures may be required in some circumstances (see below). The process of freezing tissues will damage them irreversibly, because the osmotic shock together with formation of ice crystals disrupts the cell membranes. When the original environment of the tissue lipids is altered in this way, they encounter enzymes from which they are normally protected. Especially troublesome are the lipolytic enzymes, which can hydrolyse lipids on prolonged standing, even at −20ºC, and contact with organic solvents can facilitate the process. For example, rapid de-acylation of phospholipids was observed in a bacterium frozen solid at −16ºC, with the rate at −10ºC being greater than that at 39ºC in some circumstances [37]. Appreciable lipolysis also occurred in adipose tissue stored at -8ºC [29], and some hydrolysis of phosphatidylethanolamine was detected in brain tissue stored for two hours at 0ºC [70]. Similar results have been presented for plasma [66,71,72,87] and shrimp [92]. Slow thawing can have a devastating effect on the lipids of tissues.

Lipid peroxidation can also be troublesome in tissues stored at −20ºC and even at −70ºC, and it has been recommended that samples for free radical assay be stored at −196ºC [115].

Phenomena of the same kind have been observed in many plant and animal tissues on storage. The presence of large amounts of unesterified fatty acids, diacylglycerols, phosphatidic acid or lysophospholipids in lipid extracts must be an indication that some permanent damage to the tissues and thence to the lipids has occurred. Such lipids are powerful surfactants and have been found to have disturbing effects on enzymes and membrane functions in vitro and in vivo, so the high concentrations sometimes reported in the literature are clearly incorrect. In plant tissues in particular, the enzyme phospholipase D is released and can attack phospholipids, so that there is an appreciable accumulation of phosphatidic acid and related compounds. For example, phosphatidylmethanol was found to be produced by phospholipase D-catalysed transphosphatidylation during extraction of developing soybean seeds with chloroform-methanol [89]. All the acyl lipids in potatoes were hydrolysed in minutes upon homogenization [28]. Other alterations to lipids can occur that are subtler and so are discerned less easily. For example, losses of galactolipids can take place without any obvious accumulation of partially hydrolysed intermediates [28,93]. Lipoxygenases can cause artefactual formation of oxygenated fatty acids, and autoxidation can be troublesome. Problems in stabilizing plant membranes for lipid compositional analyses have been reviewed [14].

Often these changes are marginal in their overall importance, since alterations to the main lipid components may be small. On the other hand, they can make a crucial difference to the concentrations of some important lipid metabolites. The precise free fatty acid and 1,2-diacylglycerol concentrations of tissues are recognized to be key metabolic parameters. In an especially thorough study, Kramer and Hulan [49] observed that the free fatty acid concentrations obtained when heart tissue was frozen rapidly and pulverized at dry ice temperatures before extraction were only about 15% of the values when similar tissues were extracted by more widely used techniques, i.e. by extracting directly with a homogeniser of the rotating blade type at 0ºC. With the latter, autolysis was presumed to occur during extraction. The levels of diacylglycerols were also threefold higher when the latter technique was used, a finding later confirmed by others [1]. Similarly, lysophosphatidylcholine, which had earlier been reported to be a major constituent of chromaffin granules in the adrenal gland, was found to be absent when the tissues were frozen in liquid nitrogen immediately after dissection [4]. Increases in the concentration of this lipid were found in heart tissues stored at 0ºC or −20ºC [63], but experiments involving studies of the concentration of lysophosphatidylcholine in intact tissues by means of  31P-nuclear magnetic resonance spectroscopy confirmed that it was essentially absent [62,65].

It has also been demonstrated that enzymic oxidation can cause losses not only of unsaturated fatty acids, but apparently of intact lipids [81]. Hydroperoxide groups of oxidized lipids apparently reacted to form covalent bonds with the proteins of membranes, from which they were released only on treatment with bacterial proteases. Presumably, similar effects would be seen with autoxidized lipids in tissues.

Various pre-treatments have been suggested for de-activating enzymes so that tissues can be stored for longer periods. The lipases in small samples of plant [36] or animal [29] origin have been denatured by plunging into boiling water for short periods, and the shelf life of samples treated in this way was reportedly prolonged to a considerable extent. Boiling with dilute acetic acid solution appeared to have a similar effect [82,83]. However, there is a need for practical re-evaluation of these procedures in the light of modern knowledge before they can be recommended. Boiling plant tissues before extraction certainly de-activated lipoxygenases and increased the recovery of linoleic and linolenic acids [107]. Conventional freeze-drying and perchloric acid pre-extraction of tissues were found to produce artefacts from phospholipids; acetone desiccation did not, but this would cause losses of simple lipids [62].

The use of appropriate solvents for extraction per se can also limit enzymic degradation of plant lipids (see Section C and Section F.4 below).

While it has sometimes been recommended that tissues be stored in saline solution, it is probably better to keep them dry in an atmosphere of nitrogen in all-glass containers or in bottles with Teflon™-lined caps at the low temperatures discussed above. Endogenous tissue antioxidants generally provide sufficient protection against oxidation under these conditions, although this may not be true for serum [71,91]. For example, appreciable decreases in the content of polyunsaturated fatty acids in plasma were observed on prolonged storage at −20ºC, and these were inversely correlated with vitamin A (antioxidant) levels [91]. Plastic bags, vials or other containers should be avoided scrupulously for storage purposes. As soon as possible, tissues should be homogenized and extracted with solvent at the lowest temperature practicable and certainly without being allowed to thaw. Safe storage of plasma lipoproteins, where both the protein and lipid components are potentially unstable, is a major topic in its own right and cannot be discussed here.

Similar precautions must be taken for the storage of lipids, after they have been extracted from tissues. In this circumstance, the principle danger is a loss of unsaturated fatty acid components through autoxidation. For example, rapid autoxidation of cardiolipin was found to occur on storage in chloroform [80]. Lipid extracts should therefore be stored at the lowest practical temperature, in an inert atmosphere, in an apolar solvent and in the presence of antioxidants [17].

Extraction of Tissues with Solvents

In addition to storage considerations, there are two main facets to any practical procedure for extracting lipids from tissues, i.e. firstly, exhaustive extraction and solubilization of the lipids in organic solvents and secondly, removal of non-lipid contaminants from the extracts. They are discussed in this and the next section respectively.

Many different solvents will dissolve pure single lipid classes, but they are only suitable for extracting lipids from tissues if they can overcome the strong forces of association between tissue lipids and other cellular constituents, such as proteins and polysaccharides. However, even polar complex lipids, which do not normally dissolve easily in non-polar solvents, can sometimes be extracted with these when they are in the presence of large amounts of simple lipids such as triacylglycerols. Therefore, the behaviour of a given solvent as a lipid extractant for a specific tissue cannot always be predicted. In order to release all lipids from their association with cell membranes or with lipoproteins, the ideal solvent or solvent mixture must be fairly polar. Yet, it must not be so polar that it reacts chemically with the lipids nor that triacylglycerols and other non-polar simple lipids do not dissolve and are left adhering to the tissues. If chosen carefully, the extracting solvent may have a function in preventing any enzymatic hydrolysis, but vice versa it should not stimulate any side reactions.

There is an increasing awareness of the potential toxicity of solvents to analysts, and this is another factor that must be taken into consideration when selecting a solvent mixture, especially if the laboratory is not adequately equipped with fume hoods or other ventilation. No solvent is completely safe.

Those factors affecting the extractability of lipids by solvents have been reviewed comprehensively by Zahler and Niggli [123]. The two main structural features of lipids controlling their solubility in organic solvents are the hydrophobic hydrocarbon chains of the fatty acid or other aliphatic moieties and any polar functional groups, such as phosphate or sugar residues, which are markedly hydrophilic. Any lipids lacking polar groups, for example triacylglycerols or cholesterol esters, are very soluble in hydrocarbons such as hexane, toluene or cyclohexane and also in moderately polar solvents such as diethyl ether or chloroform. In contrast, they are rather insoluble in a polar solvent such as methanol. The solubility of such lipids in alcohols increases with the chain length of the hydrocarbon moiety of the alcohol, so they tend to be more soluble in ethanol and completely soluble in butan-1-ol. Similarly, lipids with fatty acyl residues of shorter chain length tend to be more soluble in more polar solvents; tripalmitin is virtually insoluble in methanol but tributyrin dissolves readily. Unless solubilized by the presence of other lipids, polar lipids, such as phospholipids and glycosphingolipids, are only slightly soluble in hydrocarbons, but they dissolve readily in more polar solvents like methanol, ethanol or chloroform. Such solvents with high dielectric constants and polarity are required to overcome ion-dipole interactions and hydrogen bonding. Tabulated data on the solubilities of a limited range of "typical" lipids are available [98].

Analysts should be aware that water is also a solvent for lipids, contrary to some definitions of the term, and water in tissues or that used to wash lipid extracts, for example, can alter the properties of organic solvents markedly. Most complex lipids are slightly soluble in water and at least form micellar solutions, and lipids such as gangliosides, polyphosphoinositides, lysophospholipids, acyl-carnitines and coenzyme A esters are especially soluble (see Section G).

Lipids exist in tissues in many different physical forms. The simple lipids are often part of large aggregates in storage tissues, such as oil bodies or adipose tissue, from which they are extracted with relative ease. In contrast, complex lipids are usually constituents of membranes, where they occur in a close association with such compounds as proteins and polysaccharides, with which they interact, and they are not extracted so readily. These interactions are only very rarely through covalent bonds, and in general weak hydrophobic or van der Waals forces, hydrogen bonds and ionic bonds are involved. For example, the hydrophobic aliphatic moieties of lipids interact with the non-polar regions of the amino acid constituents of proteins, such as valine, leucine and isoleucine, to form weak associations. Hydroxyl, carboxyl and amino groups in lipid molecules, on the other hand, can interact more strongly with biopolymers via hydrogen bonds. Lipids such as the polyphosphoinositides are most likely bound to other cellular biopolymers by ionic bonds, and these are not easily disrupted by simple solvation with organic solvents. It is usually necessary to adjust the pH of the extraction medium to effect quantitative extraction in this instance. In addition, purely mechanical factors can limit the extractability of lipids. The helical starch (amylose) molecules in cereals form inclusion complexes with lysophosphatidylcholine, for example, limiting its accessibility to solvents. Also, cell walls in some microorganisms are rather impermeable to solvents, especially in the absence of water, which must be added to cause swelling of cellular polysaccharides. Some fatty acid or other alkyl moieties may indeed be linked directly to proteins or polysaccharides by covalent bonds, and then the optimum isolation procedure is likely to be one more suited to the analysis of the biopolymer rather than of the lipid.

In order to extract lipids from tissues, it is necessary to use solvents that not only dissolve the lipids readily but overcome the interactions between the lipids and the tissue matrix, and it is essential to perturb both the hydrophobic and polar interactions at the same time. No single pure solvent appears to be suitable as a general-purpose lipid extractant, although evidence has been presented that ethanol (20 mL/g tissue) under reflux for five minutes will extract essentially all the lipids from liver homogenates [57]. As with much other interesting work, the method was not extended to other tissues, and there is some danger of transesterification occurring as a side reaction under these conditions (see Section E below). While there are limitations to its use and alternatives are frequently suggested, most lipid analysts have accepted that a mixture of chloroform and methanol in the ratio of 2:1 (v/v) will extract lipids more thoroughly from animal, plant or bacterial tissues than other simple solvent combinations (the endogenous water in the tissues should be treated as a ternary component of the system). Since the publication of a classic paper on the subject by Folch, Lees and Stanley [26] in 1957, this has become the standard against which other methods are judged (although there are earlier applications of these solvents [25,106]).

Schmid [96] has addressed the problem of why chloroform-methanol (2:1, v/v) is such a good lipid extractant. The capacity of chloroform to associate with water molecules, presumably by weak hydrogen bonds, is a key property. Provided that the ratio of chloroform-methanol to tissue (assumed to be mainly water) is greater than 17:1, the equivalent of 5.5% water can be solvated and remain in a single phase. In contrast, mixtures of methanol with carbon tetrachloride or tetrachlorethylene, which lack the active proton of chloroform, were only able to solubilize relatively small amounts of water. Many practical methods have been developed for chloroform-methanol extraction and these are discussed in Section F below.

There are disadvantages, however. In addition to the toxicity problem, which is controllable in a well-ordered laboratory, the mixture is a potent irritant to skin. Neither chloroform nor methanol is completely stable and both were found to generate acidic by-products, which could catalyse esterification of free fatty acids or transesterification of lipids (see also Section E below) [99].

Chloroform-propan-2-ol mixtures have been recommended for the extraction of erythrocytes [9,27].

Dichloromethane-methanol (2:1, v/v) was found to give identical results to chloroform-methanol in the extraction of plasma and liver in one study [12], and the lower toxicity of dichloromethane was regarded an advantage.

Propan-2-ol-hexane (3:2, v/v) mixtures have probably been the second most frequently used solvent for the extraction of lipids from animal tissues, partly because of the good extractive properties per se but largely for the much lower toxicity [34,84]. Better recoveries of prostaglandins were reported with this mixture than with the Folch procedure [94], but it does not extract gangliosides quantitatively. Others used propan-2-ol-hexane (20:78, v/v) for extraction as part of a procedure for the isolation of cerebrosides and sulphatides from brain tissues [121]. Methanol-hexane (1:1, v/v) appears to be an odd choice for the extraction of lipids, because of the lack of complete miscibility, but it was reported to give excellent results with leaf tissue, and indeed higher concentrations of sensitive lipids such as phosphatidylinositol were found than with the Folch procedure [105]. Hexane-ethanol (5:2, v/v) was recommended for the extraction of ubiquinone [41,108], and may have wider applications [11]. Similarly, heptane-ethanol with the surfactant sodium dodecyl sulphate added has been recommended for determining vitamin E/lipid ratios in animal tissues [10].

Systematic studies of the solubilities of certain lipids in toluene-ethanol mixtures indicated that this combination might have superior properties to chloroform-methanol, but it does not appear to have been tested adequately with complex lipids or with samples of real biological interest [97]. Benzene-ethanol [61], benzene-methanol [104] and propan-2-ol-benzene-water (2:2:1, v/v) [2] also gave excellent results with some samples, but such potentially toxic mixtures are best avoided.

Butan-1-ol saturated with water has been recommended for the extraction of cereals or wheat-flour [60,68], in which the lipids may be in close association with starch, some in the form of inclusion complexes. The structure of the starch granules appears to be the most important factor, however [67]. This solvent may have wider uses, for example for quantitative extraction of lipids that are relatively soluble in water, such as lysophospholipids [7,31] or acylcarnitines [56,69]; hexan-2-ol has even been recommended for the latter [69]. Butan-1-ol-diisopropyl ether (2:3, v/v) was reported to be suitable for complete extraction of lipids from plasma, without causing denaturation of the proteins [13].

It is well known that diethyl ether or chloroform alone are good solvents for lipids, yet they are poor extractants of lipids from tissues. They can, however, have some practical value for the isolation of the non-polar lipids from triacylglycerol-rich tissues, such as oil seeds or adipose tissue, as they do not extract significant amounts of non-lipid contaminants at the same time. When they are used to extract plant tissues, these solvents also enhance the action of phospholipase D [42] unfortunately, as does butan-1-ol [20]. Propan-1-ol and propan-2-ol strongly inhibit this reaction and the latter, which has the lower boiling point, has been recommended for use with plant tissues, as a preliminary extractant especially [43,77,78].

While simple lipids and glycolipids dissolve readily in acetone, it will not dissolve phospholipids readily and indeed is often used to precipitate them from solution in other solvents, in effect as a crude preparative procedure. The lipid mixture is usually dissolved in diethyl ether and then four volumes of cold anhydrous acetone is added to precipitate the phospholipids [33]. On the other hand, endogenous water and the solubilizing effects of other lipid components may permit acetone to extract more phospholipids from animal or plant tissues than might be predicted from a knowledge of the solubility of lipid standards in the pure solvent. For example, ethyl acetate-acetone-water (2:1:1 by volume) has been recommended for extracting lipids from cultures of human cells [102,103]. Acetone has also been recommended as a preliminary extraction solvent, before conventional chloroform-methanol extraction [59]. A disadvantage is that it can react with certain lipids to produce artefacts (see Section E below). As glycolipids are soluble in acetone, chromatographic solvents containing this solvent are frequently utilized in the separation of glycolipids from phospholipids.

In recent years, supercritical fluids have been evaluated as extractants for lipids (reviewed elsewhere [6,50]). While these appear to hold promise for selected simple lipids, there appears to be little prospect for more general use at the moment.

As an alternative to conventional solvent extraction, the technology of column chromatography has been adapted to the purpose in specific circumstances (see Section G.3 below, for example, and reviewed elsewhere [19]).

Removal of Non-Lipid Contaminants from Extracts

When polar organic solvents are used to extract lipids from tissues, they tend to co-extract appreciable amounts of natural non-lipid materials, such as amino acids, carbohydrates, urea and even salts, as contaminants. A variety of procedures have therefore been developed to eliminate these, ideally without causing losses of lipids. For example, one of the simplest methods consists in evaporating the polar solvents, followed by dissolving the residual lipids in a small volume of a relatively non-polar solvent, such as hexane-chloroform (3:1, v/v), leaving many of the extraneous non-lipid substances behind [57]. Such a clean-up is rarely complete so the procedure is little used, although it should not be overlooked when large numbers of similar samples have to be purified for routine analysis by less demanding techniques, such as thin-layer chromatography (TLC). Other procedures that have been tried, but with limited success only, include dialysis, adsorption and cellulose column chromatography, electrodialysis and electrophoresis.

Much of the contaminating material can be removed from chloroform-methanol (2:1, v/v) mixtures simply by shaking with one-fourth the volume of water, or better a dilute aqueous salt solution (e.g. 0.88% potassium chloride) as described in the classic paper by Folch, Lees and Stanley [26]. The solvents then partition into two layers or phases, the lower consisting of chloroform-methanol-water in the ratio 86:14:1 (by volume) with an upper phase in which the proportions are respectively 3:48:47. The purified lipid is contained in the lower phase, which amounts to about 60% of the total volume, while the upper phase contains the non-lipid contaminants. Unfortunately, any gangliosides that may have been present also partition into the upper layer. These are minor compounds and their analysis is rather specialized, so a simple washing procedure of this kind yields satisfactory lipid samples for most purposes. When they are required for further analysis, gangliosides can be recovered from the Folch upper phase by dialysis followed by lyophilization (see Section G.3 below) [40].

It is not always recognized how important it is that the proportions of chloroform, methanol and water in the combined phases should be as close to 8:4:3 by volume as is practicable. When it is necessary to wash the lower phase again to ensure the elimination of all the non-lipid contaminants, methanol-water (1:1, v/v), i.e. a mixture similar in composition to the upper phase, should be used in order to maintain the correct proportions, otherwise some loss of polar lipids can occur.

In a successful adaptation of the above method that is especially suited to large samples with a high water content, Bligh and Dyer [8] took into account the water already present in the samples when adding further water in the washing step. As this procedure uses smaller volumes of chloroform and methanol, it is both economical and convenient.

A more time-consuming method of removing non-lipid contaminants, if more elegant and complete, consisted in carrying out the washing procedure by liquid/liquid partition chromatography on a column. The aqueous phase (the 'Folch' upper layer) is immobilized on a column of a hydrophilic dextran gel, such as Sephadex G-25™, while the organic lower phase is passed through the column. While this type of lipid purification procedure was first developed by Wells and Dittmer [113], a modification described by Wuthier [120] is simpler and thus more suitable for large numbers of samples. In brief, chloroform, methanol and water in the ratio 8:4:3 by volume were partitioned as in a conventional 'Folch' wash, the column of Sephadex G-25™ was packed in suspension in the upper phase and the crude lipid extract was applied to the column in a small volume of the lower phase. While lipids were eluted rapidly by further lower phase, contaminants of low molecular weight remained on the column. Gangliosides and non-lipids could be recovered from the column by washing with 'Folch' upper phase, and the column regenerated for further use. The technique can also be used to remove acid, alkali and salts from lipid samples. An alternative simple procedure was described by others [117]. In contrast, a more complicated Sephadex™ column procedure was described [101] in which larger amounts of lipids were purified and gangliosides were obtained in a discrete fraction free of non-lipid contaminants. As various bile acids were obtained in distinct fractions from the conventional lipids, the method appeared to be particularly suited to the analysis of bile lipids [90]. However, all of these methods are very time-consuming, and as a result they have never achieved widespread use. Such column procedures are rather different in principle from a solid-phase extraction method described below (Section G.3).

An alternative approach consisted in pre-extraction of tissues with 0.25% acetic acid, which deactivated the enzymes (see Section B above), and simultaneously removed all potential contaminants of the lipid extracts [82,83]. Although this procedure appears only to have been applied to brain tissue and soybeans to date, it might repay further investigation.

Artefacts of Extraction Procedures

The production of free fatty acids, diacylglycerols, phosphatidic acid or lysophospholipids, due to faulty storage of tissues prior to extraction is discussed in Section B above. Extraneous substances can be introduced into lipid extracts from innumerable sources, and these are discussed elsewhere in this web site (see my review of methylation procedures). Plasticisers are especially troublesome.

When chloroform-methanol or, indeed, any alcoholic extracts that contain lipids are heated or stored for long periods in the presence of small amounts of sodium carbonate or bicarbonate of tissue origin, base-catalysed transesterification can occur and appreciable amounts of methyl esters may be found in the extracts [55]. By adjusting the pH of the extraction medium to 4 to 5, the problem can be circumvented [55,109]. Similar findings are often described, and it is possible that both acidic and basic non-lipid contaminants may catalyse the same reaction. Tissue acyl-transferases can also catalyse ethyl ester formation [64]. However, small amounts of methyl and ethyl esters do appear to occur naturally in some tissues, and when they are detected confirmation should be obtained of their natural origin. Artefacts of extraction or storage can be eliminated simply by extracting the tissues with solvents that do not contain any alcohol, such as diethyl ether [23], hexane [44], acetone [46] or acetone-chloroform [54], and repeating the analysis for methyl esters on this material.

While 6-O-acyl-galactosyldiacylglycerols are known to be natural components of some plant tissues, it appears that they may also be formed as artefacts by acyl transfer from other lipids when cells are disrupted; they are found in much smaller amounts when the tissues are homogenized in the presence of the extracting solvent [38]. Similar difficulties may be encountered with homogenates of bacteria [100].

Some rearrangement of plasmalogens may occur when they are stored for long periods in methanol [110].

Acetone should not be used in the analysis of tissues rich in polyphosphoinositides, such as brain lipids, as it causes rapid de-phosphorylation [22,114]. An acetone derivative (imine) of phosphatidyl-ethanolamine was reported to be formed in vitro during extraction of freeze-dried tissue with acetone [3,39].

Practical Extraction Procedures

1. Some General Comments

All solvents contain small amounts of lipid-like contaminants and only those of highest quality need not be distilled routinely before use. Plastic apparatus and containers (other than those made from Teflon™) should be avoided at all costs, because plasticisers (usually diesters of phthalic acid) are leached out surprisingly easily and may appear as spurious peaks on chromatograms and affect UV spectra, for example. Organic solvents will extract substantial amounts of these compounds, and wet animal tissues alone in contact with plastic can extract small amounts.

Polyunsaturated fatty acids will autoxidize very rapidly if left unprotected in air. Although natural tissue antioxidants such as tocopherols may afford some protection, it is advisable to add an additional antioxidant such as BHT ("butylated hydroxy toluene" or 2,6-di-tert-butyl-p-cresol) at a level of about 25 mg/L to the solvents used for storage purposes. This need not interfere with later chromatographic analyses, as it tends to evaporate on removal of solvents or appears in conveniently empty regions of chromatograms [18]. Whenever practicable, extraction procedures should be carried out in an atmosphere of nitrogen, and both tissues and tissue extracts should be stored at -20ºC under nitrogen. It is helpful to de-aerate solvents by flushing them with nitrogen or helium before use.

If a blender is used to homogenize tissues, it should be one in which the drive to the knives or grinders is from above, so that there is no contact between solvent and any greased seals or bearings. Lyophilized tissues are particularly difficult to extract and it may be necessary to re-hydrate them before extraction to ensure quantitative recovery of lipids.

When the aqueous and organic phases are partitioned in an extraction procedure, it should be noted that centrifugation may be necessary or of assistance in ensuring complete separation of the layers. This also serves to compact the interfacial layer, rendering it easier to recover for further extraction if need be.

Solvents should be removed from lipid extracts under vacuum in a rotary film evaporator at a temperature only a little above ambient and no higher than 40ºC. When a large amount of solvent must be evaporated, the extract should be concentrated to a small volume and then transferred to as small a flask as is convenient so that the lipids do not dry out as a thin film over a large area of glass. While there is no need to bleed nitrogen in continuously during the evaporation process, as the solvent vapours effectively displace any air, the vacuum should be broken eventually with nitrogen. Lipids should not be left in the dry state, but should be taken up in or covered by an inert non-alcoholic solvent such as hexane. The last traces of water may be removed by co-distillation with ethanol or toluene.

As cautioned earlier, it should never be forgotten that chloroform and methanol are highly toxic and should only be used in well-ventilated areas. The single operation most likely to introduce appreciable amounts of solvent vapour into the atmosphere is filtration. Mixtures of chloroform and methanol, especially, are powerful irritants when they come into contact with the skin. (A colleague of the author was taught a painful lesson when he spilt a large volume down his trousers!).

 

2. The "Folch" Procedure

A large number of different extraction procedures, varying in the nature of the solvents, the method of homogenizing, removal of contaminants and many other aspects, have been described in the literature. Unfortunately, a high proportion of these have been tried in only one laboratory or with a limited range of tissues or organisms. Only rarely have they been tested exhaustively and compared rigorously with a widely used procedure, such as that of Folch, Lees and Stanley [26] or Bligh and Dyer [8]. Until this is done, whatever the virtues of alternatives, the latter two procedures are likely to continue in their position of pre-eminence. An extraction procedure must be selected to give the optimum yield of representative lipids in as practical a manner as possible. While a method chosen for routine analysis of the cholesterol content of large numbers of clinical samples, say, must be accurate and reproducible, it need not incorporate modifications to recover every trace of such complex lipids as gangliosides or inositides. If the latter compounds are required for analysis, however, there is no alternative but to use complex and exhaustive extraction procedures.

For consistent results with any method, a strict protocol must be adopted that follows the principles laid down by the originators of the method. With the procedure of Folch et al. [26], it is essential that the ratio of chloroform, methanol and saline solution in the final mixture be close to 8:4:3. On the other hand, some variation in the approach to attaining the optimum concentrations may be possible. The tissue can be homogenized initially in the presence of both solvents, for example, although better results are often obtained if the methanol (10 mL/g tissue) is added first, followed after brief blending by chloroform (20 volumes). More than one extraction may be needed, but with most tissues the lipids are removed almost completely after two or three treatments. Generally, there is no need to heat the solvents with the tissue homogenates, but this may sometimes be necessary, for example with wet bacterial cells [111]. Treatment with acid [79] or proteases [86] may also be of benefit in this instance. However, the extractability of tissue is variable and depends both on the nature of the tissue and of the lipids. For example, the extractability of gangliosides from brain is highly dependent on the concentration of monovalent cations in the tissue. It may be necessary in some instances to employ more stringent extraction procedures; Ways and Hanahan [112] recommended separate re-extraction with chloroform on its own followed by methanol alone, while others [90] advised that a five-stage extraction procedure, using both acidic and basic solvent systems, was required for exhaustive recovery of lipids from tissues.

3. The Bligh and Dyer Method

The Bligh and Dyer method [8] is probably the second most common extraction procedure for lipids, and it is probably the most misunderstood and abused. It was developed as an economical method of extracting the lipids from large volumes of wet tissue, from frozen fish specifically, with the minimum volume of solvent. In essence, the endogenous water in the tissue was considered as a ternary component of the extraction system and sufficient chloroform and methanol were added to give a single-phase system for homogenization. After filtering, the residue was re-homogenized with fresh chloroform to ensure that the simple lipids were extracted completely, and the combined organic layers were added to fresh saline solution that produced a biphasic mixture. The chloroform layer contained the lipid, of course.

If applied correctly to wet tissues, this procedure gives good recoveries of the more important lipid classes. It is rapid and therefore suited to many routine applications. However, too much should not be expected of it and incomplete recovery of the metabolically important but quantitatively minor acidic lipids is inevitable. When it has apparently been found wanting, the author has sometimes noted that the strict protocol laid down by Bligh and Dyer [8] has not always been followed.

 

4. Extraction of Plant Tissues

The lipids of plant material and photosynthetic tissues especially are liable to undergo extensive enzyme-catalysed degradation when extracted with chloroform-methanol mixtures, as discussed in Section E above. The problem is best overcome by means of a preliminary extraction with propan-2-ol [43]. The favoured method [77,78] is to extract the plant tissue with 100 volumes of propan-2-ol, and then to re-extract the solid residue with chloroform-methanol (2:1, v/v). After evaporation of the solvent, the crude extract is taken up in fresh chloroform-methanol and given a "Folch" wash (see Section F.2 above) to eliminate non-lipid contaminants. This appears to be a milder method than extraction with ethanol-trichloroacetic acid (3:1, v/v) as has been suggested [45].

Some Special Cases

Problems are most often encountered with lipids that contain highly hydrophilic functional groups. With the common biphasic extraction systems, these lipids may partition into the aqueous layer and can be lost. Improved recoveries can then be obtained by utilizing monophasic extraction media, by using salting-out procedures to force these lipids into the organic phase or by recovering the organic material from the aqueous phase.

1. Acyl-Coenzyme A Esters

A monophasic extraction medium has been recommended for quantitative recovery of acyl-coenzyme A esters by Christiansen [16]. In essence, a conventional chloroform methanol (2:1, v/v) extract was taken to dryness without the washing step, and the residue was divided into chloroform- and methanol-soluble fractions, which were only combined at the time of analysis. A comparable approach has been suggested for phospholipids [48]. On the other hand, it has been demonstrated that butan-1-ol is capable of extracting acyl-CoA esters rather efficiently, without formation of emulsions [5]. Others have used a salting out approach, i.e. a preliminary extraction with propan-2-ol-phosphoric acid, partition with hexane to remove simple lipids, addition of a saturated ammonium sulphate solution and then extraction with chloroform-methanol (1:2, v/v) [58,119].

Rosendahl and Knudsen [88], in contrast, have indicated that the problem may be more difficult than has hitherto been realized. They were able to attain consistent yields of about 55% of the total CoA esters only by combining two-phase extraction with salting out and the addition of acyl-CoA-binding protein.

2. Lysophospholipids and Polyphosphoinositides

Poor and inconsistent yields can also be obtained in the extraction of strongly hydrophilic glycerolipids, for similar reasons to those discussed in the previous section, unless precautions are taken. When lysophospholipids are major components of tissue extracts, it has been recommended that acid (best avoided as it may be harmful to any plasmalogens present) or inorganic salts be added during extraction with chloroform-methanol, or better that water-saturated butan-1-ol be used to extract the lipid [7,31].

With polyphosphoinositides, it is especially important to ensure that tissues are stored in a manner such that enzymatic degradation is minimized. Strategies such as single phase extraction, salting out or acidification are then used to optimize the recoveries, although there does not appear to be a consensus as to which of these is best for the purpose. For example, it has been suggested that tissues should be extracted with chloroform-methanol in the presence of calcium chloride initially and subsequently after acidification [35]. Most groups appear to favour exhaustive extraction procedures involving repeated treatment of the tissue with various chloroform-methanol mixtures together with 1M hydrochloric acid [15,21,32,95]. In an alternative method, an ion-pairing reagent, tetrabutylammonium sulphate, was added to the medium and gave apparently much higher recoveries of polyphosphoinositides [30]. However, this approach does not seem to have been followed by others.

 

3. Gangliosides

Gangliosides are so soluble in water that they remain entirely in the aqueous phase during a 'Folch' extraction. The established procedure for isolation of these compounds has involved dialysis of the 'Folch' upper phase to remove ions and other compounds of low molecular weight followed by lyophilization [40]. As an alternative to the 'Folch' system, partition of lipid extracts between diisopropyl ether, butan-1-ol and 50 mM aqueous sodium chloride has been suggested, with the gangliosides again being retained in the aqueous layer [51,52].

Large-scale column procedures for isolating gangliosides are described in Section D above. Williams and McCluer [118] were among the first to adapt solid-phase extraction methodology to the analysis of lipids, and described the use of Sep-Pak™ ODS cartridges for the isolation of gangliosides from tissue extracts. The method involved simply passing the upper aqueous phase from a Folch extract through a column, packed with a bonded octadecylsilyl phase, which retained the gangliosides for subsequent recovery by elution with chloroform-methanol. This procedure with occasional minor variations appears now to be widely accepted for the purpose [122].

Of course, such preparations do contain other materials, and further purification by chromatographic and chemical means is essential for many purposes [24]. The best method for this is a subject of some debate among the experts [53,116,122].

 

ACKNOWLEDGEMENT: This review was published as part of a programme funded by the Scottish Executive Rural Affairs Department.

 

References

  1. Abe, K. and Kogure, K., J. Neurochem., 47, 577-582 (1986).
  2. Allen, P.C., Anal. Biochem., 45, 253-259 (1972).
  3. Ando, N., Ando, S. and Yamakawa, T., J. Biochem. (Tokyo), 70, 341-348 (1971).
  4. Arthur, G. and Sheltawy, A., Biochem. J., 191, 523-532 (1980).
  5. Banis, R.J., Roberts, C.S., Stokes, G.B and Tove, S.B., Anal. Biochem., 73, 1-8 (1976).
  6. Bartle, K.D. and Clifford, T.A., in Advances in Applied Lipid Research, Vol. 1, pp. 217-264 (1992) (edited by F.B. Padley, JAI Press, London).
  7. Bjerve, K.S., Daae, L.N.W. and Bremer, J., Anal. Biochem., 58, 238-245 (1974).
  8. Bligh, E.G. and Dyer, W.J., Can. J. Physiol., 37, 911-917 (1959).
  9. Broekhuyse, R.M., Clin. Chim. Acta, 51, 341-343 (1974).
  10. Burton, G.W., Webb, A. and Ingold, K.U., Lipids, 20, 29-39 (1985).
  11. Cabrini, L., Landi, L., Stefanelli, C., Barzanti, V. and Sechi, A.M., Comp. Biochem. Physiol., 101B, 383-386 (1992).
  12. Carlson, L.A., Clin. Chim. Acta, 149, 89-93 (1985).
  13. Cham, B.E. and Knowles, B.R., J. Lipid Res., 17, 176-181 (1976).
  14. Chapman, D.J. and Barber, J., Methods Enzymol., 148, 294-319 (1987).
  15. Christensen, S., Biochem. J., 233, 921-924 (1986).
  16. Christiansen, K., Anal. Biochem., 66, 93-99 (1975).
  17. Christie, W.W. and Han, X. Lipid Analysis - Isolation, Separation, Identification and Lipidomic Analysis (4th edition), 446 pages (Oily Press, Bridgwater, U.K.) (2010) - (Now online here.).
  18. Christie, W.W., Gas Chromatography and Lipids, The Oily Press, Ayr (1989) (Now online here..).
  19. Christie, W.W., in Advances in Lipid Methodology - One, pp. 1-17 (1992) (edited by W.W. Christie, Oily Press, Ayr) (now online here..).
  20. Colborne, A.J. and Laidman, D.L., Phytochemistry, 14, 2639-2645 (1975).
  21. Creba, J.A., Downes, C.P., Hawkins, P.T., Brewster, G., Michell, R.H. and Kirk, C.J., Biochem. J., 212, 733-747 (1983).
  22. Dawson, R.M.C. and Eichberg, J., Biochem. J., 96, 634-643 (1965).
  23. Dhopeshwarkar, G.A. and Mead, J.F., Proc. Soc. Exp. Med., 109, 425-429 (1962).
  24. Fidelio, G.D., Ariga, T. and Maggio, B., J. Biochem. (Tokyo), 110, 12-16 (1991).
  25. Folch, J., Ascoli, I., Lees, M., Meath, J.A. and LeBaron, F.N., J. Biol. Chem., 199, 833-841 (1951).
  26. Folch, J., Lees, M. and Stanley, G.H.S., J. Biol. Chem., 226, 497-509 (1957).
  27. Freyburger, G., Heape, A., Gin, H., Boisseau, M. and Cassagne, C., Anal. Biochem., 171, 213-216 (1988).
  28. Galliard, T., Phytochemistry, 9, 1725-1734 (1970).
  29. Geyer ,K.G. and Goodman, H.M., Proc. Soc. Exp. Biol. Med., 133, 404-406 (1970).
  30. Grove, R.I., Fitzpatrick, D. and Schimmel, S.D., Lipids, 16, 691-693 (1981).
  31. Hajra, A.K., Lipids, 9, 502-505 (1974).
  32. Hajra, A.K., Fisher, S.K. and Agranoff, B.W., in Neuromethods 7. Lipids and Related Compounds, pp. 211-225 (1988) (edited by A.A. Boulton, G.B. Baker and L.A. Horrocks, Humana Press, Clifton).
  33. Hanahan, D.J., Turner,M.B. and Jayko, M.E., J. Biol. Chem., 192, 623-628 (1951).
  34. Hara, A. and Radin, N.S., Anal. Biochem., 90, 420-426 (1978).
  35. Hauser, G. and Eichberg, J., Biochim. Biophys. Acta, 326, 201-209 (1973).
  36. Haverkate, F. and van Deenen, L.L.M., Biochim. Biophys. Acta, 106, 78-92 (1965).
  37. Hazlewood, G.P. and Dawson, R.M.C., Biochem. J., 153, 49-53 (1976).
  38. Heinz, E., Biochim. Biophys. Acta, 144, 333-343 (1967).
  39. Helmy, F.M. and Hack, M.H., Lipids, 1, 279-281 (1966).
  40. Kanfer, J.N., Methods Enzymol., 14, 660-664 (1969).
  41. Katayama, K., Takada, M., Yuzuriha, T., Abe, K. and Ikenoya, S., Biochem. Biophys. Res. Commun., 95, 971-977 (1980).
  42. Kates, M., Can. J. Biochem. Physiol., 34, 967-980 (1956).
  43. Kates, M. and Eberhardt, F.M., Can. J. Bot., 35, 895-905 (1957).
  44. Kaufmann, H.P. and Viswanathan, C.V., Fette Seifen Anstrichm., 65, 925-629 (1963).
  45. Kaul, K. and Lester, R.L., Plant Physiol., 55, 120-129 (1975).
  46. Kinnunen, P.M. and Lange, L.G., Anal. Biochem., 140, 567-576 (1984).
  47. Klein, R.A. and Kemp, P., in Methods in Membrane Biology. Vol. 8, pp. 51-217 (1977) (edited by E.D. Korn, Plenum Press, New York).
  48. Kolarovic, L. and Fournier, N.C., Anal. Biochem., 156, 244-250 (1986).
  49. Kramer, J.K.G. and Hulan, H.W., J. Lipid Res., 19, 103-106 (1978).
  50. Laakso, P., in Advances in Lipid Methodology - One, pp. 81-119 (1992) (edited by W.W. Christie, Oily Press, Ayr).
  51. Ladisch, S. and Gillard, B., Anal. Biochem., 146, 220-231 (1985).
  52. Ladisch, C. and Gillard, B., Methods Enzymol., 138, 300-306 (1987).
  53. Ledeen, R.W. and Yu, R.K., Methods Enzymol., 83, 139-191 (1982).
  54. Leikola, E., Nieminen, E. and Solomaa, E., J. Lipid Res., 6, 490-493 (1965).
  55. Lough, A.K., Felinski, L. and Garton, G.A., J. Lipid Res., 3, 478-480 (1962).
  56. Lowes, S. and Rose, M.E., Trends Anal. Chem., 8, 184-187 (1989).
  57. Lucas, C.C. and Ridout, J.H., Prog. Chem. Fats Other Lipids, 10, 1-150 (1970).
  58. Mancha, M., Stokes, G.B. and Stumpf, P.K., Anal. Biochem., 68, 600-608 (1975).
  59. McGrath, L.T. and Elliot, R.J., Anal. Biochem., 187, 273-276 (1990).
  60. Mecham, D.K. and Mohammad, D.K., Cereal Chem., , 32405-415 (1955).
  61. Melton, S.L., Moyers, R.E. and Playford, C.G., J. Am. Oil Chem. Soc., 56, 489-493 (1978).
  62. Meneses, P., Para, P.F. and Glonek ,T., J. Lipid Res., 30, 458-461 (1989).
  63. Mock, T., Pelletier, M.P.J., Man ,R.Y.K. and Choy, P.C., Anal. Biochem., 137, 277-281 (1984).
  64. Mogelson, S. and Lange, L., Fed. Proc., 42, 2046 (1983).
  65. Mogelson,S ., Wilson, G.E. and Sobell, B.E., Biochim. Biophys. Acta, 619, 680-688 (1980).
  66. Moilanen, T. and Nikkari, T., Clin. Chim. Acta, 114, 111-116 (1981).
  67. Morrison, W.R. and Coventry, A.M., Starch, 41, 21-23 (1989).
  68. Morrison, W.R., Tan, S.L. and Hargin, K.D., J. Sci. Food Agric., 31, 329-340 (1980).
  69. Morrow, R.J. and Rose, M.E., Clin. Chim. Acta, 211, 73-81 (1992).
  70. Moscatelli, E.A. and Duff, J.A., Lipids, 13, 294-296 (1978).
  71. Mueller, H.W., Loeffler, E. and Schmandt, W., Clin. Chim. Acta, 124, 343-349 (1982).
  72. Muhlfellner, O., Muhlfellner, G., Zofel, P. and Kaffarnik, H., Z. Klin. Chem. Klin. Biochem., 10, 37-41 (1972).
  73. Naito, H.K. and David, J.A., in Lipid Research Methodology, pp. 1-76 (1984) (edited by J.A.Story, A.R. Liss Inc., New York).
  74. Nelson, G.J., in Blood Lipids and Lipoproteins. Quantitation, Composition and Metabolism, pp. 3-24 (1972) (edited by G.J. Nelson, Wiley & Sons, New York).
  75. Nelson, G.J., in Analysis of Lipids and Lipoproteins, pp. 1-22 (1975) (edited by E.G. Perkins, American Oil Chemists' Soc., Champaign).
  76. Nelson, G.J., in Analyses of Oils and Fats, pp. 20-59 (1991) (edited by E.G. Perkins, American Oil Chemists' Soc., Champaign).
  77. Nichols, B.W., Biochim. Biophys. Acta, 70, 417-422 (1963).
  78. Nichols, B.W., in New Biochemical Separations, pp. 321-337 (1964) (edited by A.T. James and L.J. Morris, Van Norstrand, New York).
  79. Nishihara, M. and Koga, Y., J. Biochem. (Tokyo), 101, 997-1005 (1987).
  80. Parinandi, N.L., Weis, B.K. and Schmid, H.H.O., Chem. Phys. Lipids, 49, 215-220 (1988).
  81. Parinandi, N.L., Weis, B.K., Natarajan, V. and Schmid, H.H.O., Arch. Biochem. Biophys., 280, 45-52 (1990).
  82. Phillips, F.C. and Privett, O.S., Lipids, 14, 590-595 (1979).
  83. Phillips, F.C. and Privett, O.S., Lipids, 14, 949-952 (1979).
  84. Radin, N.S., Methods Enzymol., 72, 5-7 (1981).
  85. Radin, N.S. in Neuromethods 7. Lipids and Related Compounds, pp. 1-61 (1988) (edited by A.A. Boulton, G.B. Baker and L.A. Horrocks, Humana Press, Clifton).
  86. Rizzo, A.F. and Korkeala, H., Biochim. Biophys. Acta, 792, 367-370 (1984).
  87. Rogiers, V., Clin. Chim. Acta, 84, 49-54 (1978).
  88. Rosendahl, J. and Knudsen, J., Anal. Biochem., 207, 63-67 (1992).
  89. Roughan, P.G., Slack, C.R. and Holland, R., Lipids, 13, 497-503 (1978).
  90. Rouser, G., Kritchevsky, G. and Yamamoto, A., in Lipid Chromatographic Analysis. Vol. 1, 99-162 (1967) (edited by G.V. Marinetti, Edward Arnold, London).
  91. Salo, M.K., Gey, F. and Nikkari, T., Internat. J. Vit. Nutr. Res., 56, 231-239 (1986).
  92. Sasaki, G.C. and Capuzzo, J.M., Comp. Biochem. Physiol., 78B, 525-531 (1984).
  93. Sastry, P.S. and Kates, M., Canad. J. Biochem., 3, 1280-1287 (1964).
  94. Saunders, R.D. and Horrocks, L.A., Anal. Biochem., 143, 71-75 (1984).
  95. Schacht, J., Methods Enzymol., 72, 626-631 (1981).
  96. Schmid, P., Physiol. Chem. Phys., 5, 141-150 (1973).
  97. Schmid, P., Calvert, J. and Steiner, R., Physiol. Chem. Phys., 5, 157-166 (1973).
  98. Schmid, P. and Hunter, E., Physiol. Chem. Phys., 3, 98-102 (1971).
  99. Schmid, P., Hunter, E. and Calvert, I., Physiol. Chem. Phys., 5, 151-155 (1973).
  100. Shaw, N., Bacteriol. Revs., 34, 365-377 (1970).
  101. Siakotos, A.N. and Rouser, G., J. Am. Oil Chem. Soc., 42, 913-919 (1965).
  102. Slayback, J.R.B., Campbell, I.M. and Vaughan, M.H., Biochim. Biophys. Acta, 431, 217-224 (1976).
  103. Slayback,J.R.B., Cheung,L.W.Y., and Geyer, R.P., Anal. Biochem., 83, 372-384 (1978).
  104. Sobus, M.T. and Holmlund, C.E., Lipids, 11, 341-348 (1976).
  105. Somersalo, S., Karunen, P. and Aro, E.M., Physiol. Plant., 68, 467-70 (1986).
  106. Sperry, W.M. and Brand, F.C., J. Biol. Chem., 213, 69-76 (1955).
  107. Spreitzer, H., Schmidt, J. and Spiteller, G., Fat Sci. Technol., 91, 108-113 (1989).
  108. Takada, M., Ikenoya, S., Yuzuriha, T. and Katayama, K., Biochim. Biophys. Acta, 679, 308-314 (1982).
  109. Tuchman, M. and Krivit, W., J. Chromatogr., 307, 172-179 (1984).
  110. Viswanathan, C.V., Hoevet, S.P., Lundberg, W.O., White, J.M. and Muccini, G.A., J. Chromatogr., 40, 225-234 (1969).
  111. Vorbeck, M.L. and Marinetti, G.V., J. Lipid Res., 6, 3-6 (1965).
  112. Ways ,P. and Hanahan, D.J., J. Lipid Res., 5, 318-328 (1964).
  113. Wells, M.A. and Dittmer, J.C., Biochemistry, 2, 1259-1263 (1963).
  114. Wells, M.A. and Dittmer, J.C., Biochemistry, 4, 2459-2468 (1965).
  115. Whiteley, G.S.W., Fuller, B.J. and Hobbs, K.E.F., Cryo-Letters, 13, 83-86 (1992).
  116. Wiegandt, H. (editor), Glycolipids. New Comprehensive Biochemistry, Vol. 10, Elsevier, New York (1985).
  117. Williams, J.P. and Merrilees, P.A., Lipids, 5, 367-370 (1970).
  118. Williams, M.A. and McCluer, R.H., J. Neurochem., 35, 266-269 (1980).
  119. Woldegiorgis, G., Spennetta, T., Corkey, B.E., Williamson, J.R. and Shrago, E., Anal. Biochem., 150, 8-12 (1985).
  120. Wuthier, R.E., J. Lipid Res., 7, 558-561 (1966).
  121. Yahara, S., Kawamura, N., Kishimoto, Y., Saida, T. and Tourtellotte, W.W., J. Neurol. Sci., 54, 303-315 (1982).
  122. Yates, A.J., in Neuromethods 7. Lipids and Related Compounds, pp. 265-327 (1988) (edited by A.A. Boulton, G.B. Baker and L.A. Horrocks, Humana Press, Clifton).
  123. Zahler, P. and Niggli, V., in Methods in Membrane Biology, Vol. 8, pp. 1-50 (1977) (edited by E.D. Korn, Plenum Press, New York).
  124. Zhukov, A.V. and Vereshchagin, A.G., Adv. Lipid Res., 18, 247-282 (1981).