`Biochemistry and Physiology
`
`Issued by THE NATIONAL RESEARCH COUNCIL OF CANADA
`
`VOI,UME 37
`
`AUGUST 1959
`
`NUMBER 8
`
`A RAPID METHOD OF TOTAL LIPID EXTRACTION
`AND PURIFICATION1
`
`Abstract
`Lipid decomposition studies in frozen fish have led t o the development of a
`simple and rapid method for the extraction and purification of lipids from
`biological materials. The entire procedure can be carried out in approximately
`10 minutes; it is efficient, reproducible, and free from deleterious manipulations.
`'The wet tissue is homogenized with a mixture of chloroform and methanol in
`such proportions that a miscible system is formed with the water in the tissue.
`Dilution with chloroform and water separates the homogenate into two layers,
`the chloroform layer containing all the lipids and the methanolic layer containing
`all the non-lipids. A purified lipid extract is obtained merely by isolating the
`chloroform layer. The method has been applied t o fish muscle and may easily
`be adapted t o use with other tissues.
`
`Introduction
`
`In the course of investigations concerning the deterioration of lipids in
`frozen fish, the need arose for an efficient and rapid method of total lipid
`extraction and purification. Furthermore, due to the highly unsaturated nature
`of fish lipids, the method had t o involve only mild treatment so as t o minimize
`oxidative decomposition and the production of artifacts. Several existing
`methods were considered but none was entirely satisfactory. The methods of
`Dambergs (1) and Folch et al. (2) were too time-consuming for routine investiga-
`tions and since the former method entailed heating and evaporation it was
`considered unsuitable for lipid composition studies. The recent method of
`Folch et al. (3), which was published while this study was in progress, was more
`rapid than their previous method but still had the disadvantage of employing
`large and inconvenient volumes of solvent. The method used by Dyer and
`Morton (4) was rapid but extracted only a fraction of the total lipid.
`The present paper describes a method whereby the lipids of biological
`materials can be extracted and purified in a single operation.
`Mixtures of chloroform and methanol have had wide use a s lipid extractants
`and examination of the chloroform-methanol-water
`phase diagram (Fig. 1)
`led t o the following hypothesis. Optimum lipid extraction should result when
`'Manuscript received February 27, 1959.
`Contribution from the Fisheries Research Board of Canada, Technological Station, Halifax,
`Nova Scotia.
`Can. J. Biochem. Physiol. Vol. 37 (1959)
`RIMFROST EXHIBIT 1165 Page 0001
`
`Can. J. Biochem. Physiol. Downloaded from www.nrcresearchpress.com by 167.206.200.52 on 05/11/17
`
`For personal use only.
`
`
`
`912
`
`CANADIAN JOURNAL O F BIOCHEMISTRY AND PHYSIOLOGY. VOL. 37. 1959
`
`the tissue is homogenized with a mixture of chloroform and methanol which,
`when mixed with the water in the tissue, would yield a monophasic solution.
`The resulting homogenate could then be diluted with water and/or chloroform
`t o produce a biphasic system, the chloroform layer of which should contain
`the lipids and the methanol-water
`layer the non-lipids. Hence, a purified
`lipid extract should be obtained when the chloroform layer is isolated.
`
`t
`
`Procedure
`
`Reagents
`Methanol, absolute, analytical reagent ; chloroform, analytical reagent.
`
`Lipid Extraction and Purification
`The following procedure applies t o tissues like cod muscle that contain
`80+ 1% water and about lYo lipid. Each 100-g sample of the fresh or frozen
`tissue is homogenized in a Waring Blendor for 2 minutes with a mixture of
`100 ml chloroform and 200 ml methanol. T o the mixture is then added 100 ml
`chloroform and after blending for 30 seconds, 100 ml distilled water is added
`and blending continued for another 30 seconds. T h e homogenate is filtered
`through Whatman No. 1 filter paper on a Coors No. 3 Buchner funnel with
`slight suction. Filtration is normally quite rapid and when the residue becomes
`dry, pressure is applied with the bottom of a beaker t o ensure maximum re-
`covery of solvent. T h e filtrate is transferred t o a 500-ml graduated cylinder,
`and, after allowing a few minutes for complete separation and clarification, the
`volume of the chloroform layer (at least 150 ml) is recorded and the alcoholic
`layer removed by aspiration. A small volume of the chloroform layer is also
`removed to ensure complete removal of the top layer. T h e chloroform layer
`contains the purified lipid.
`For quantitative lipid extraction the lipid withheld in the tissue residue is
`recovered by blending the residue and filter paper with 100 ml chloroform.
`T h e mixture is filtered through the original Buchner funnel and the blendor
`jar and residue are rinsed with a total of 50 ml chloroform. This filtrate is
`mixed with the original filtrate prior to removal of the alcoholic layer.
`
`Adaptation to Other Materials
`The above procedure can be applied directly t o any 3 00-g sample containing
`80 g water. Many alterations of the procedure are permissible but it is impera-
`tive t h a t the volumes of chloroform, methanol, and water, before and after
`dilution, be kept in the proportions 3 :2 :0.8 and 2 :2 :1.8, respectively. These
`ratios represent the total volumes present in the ternary systems, including the
`water present in the sample. Thus, in the extraction of materials that d o not
`contain 80Yo water, either the size of the samples can be adjusted so that they
`contain 80 g water, or 100-g samples can be used and the volumes of chloroform
`and methanol changed to give the correct proportions. In cases where the mois-
`ture content is much less than 80% (e.g. fish meal), it is necessary t o add distil-
`led water. When material containing a large amount of lipid is used, or where
`the supply of material is limited, the size of sample can be reduced and the
`above solvent quantities scaled down t o meet requirements.
`
`RIMFROST EXHIBIT 1165 Page 0002
`
`Can. J. Biochem. Physiol. Downloaded from www.nrcresearchpress.com by 167.206.200.52 on 05/11/17
`
`For personal use only.
`
`
`
`BLIGH AND DYER: LIPID EXTRACTION A N D PURIFICATION
`
`Determination of Lipid Content
`A portion of the lipid extract containing 100-200 mg lipid is evaporated t o
`dryness in a tared flask and the weight of the lipid residue determined.
`Evaporation, facilitated by a stream of nitrogen, is carried out in a water bath
`a t 40-50" C and the residue is dried over phosphoric anhydride in a vacuum
`desiccator. After weighing, a small volume of chloroforrn is added t o each
`flask to detect the presence of non-lipid material (insoluble). If non-lipids are
`present, the chloroform is carefully decanted and the flask rinsed three times
`with chloroform. The dry weight of the residue is determined and subtracted
`from the initial weight. The lipid content of the sample is calculated as follows:
`weight of lipid in aliquot X volume of chloroform layer
`Total lipid =
`volume of aliquot
`
`Experimental
`Optimum Conditions .for Lipid Extraction
`The first objective was t o find which of various mixtures of chloroform and
`methanol would yield quantitative extraction. The solvents were mixed in
`
`phase diagram, % iw/w) a t 20' C
`FIG. 1. Chloroform-methanol-water
`(5).
`. . . . . "maximum chloroform tie-line" estimated from data for 0 C (6). - - - dilutions.
`
`RIMFROST EXHIBIT 1165 Page 0003
`
`Can. J. Biochem. Physiol. Downloaded from www.nrcresearchpress.com by 167.206.200.52 on 05/11/17
`
`For personal use only.
`
`
`
`914
`
`CANADIAN JOURNAL O F BIOCHEMISTRY A N D PHYSIOLOGY. VOL. 37, 1959
`
`such proportions that, with the water of the samples, ternary systems were
`formed having compositions as marked on the phase diagram (Fig. 1, points
`A-J). Samples of minced cod muscle (100 g each) were homogenized with t h e
`extractants (Table I) and filtered. T o the monophasic extracts, sufficient water
`
`TABLE I
`Lipid extracted from 100 g cod muscle by ternary mixtures of chloroform-methanol-water
`
`Initial extraction mixture
`
`Dilution solvents
`
`Ref. pt. in
`phase diagram
`(Fig. 1)
`
`Chloro-
`form.
`ml
`
`Methanol.
`ml
`
`Water.*
`ml
`
`Chloro-
`form.
`ml
`
`Water,
`ml
`
`Total
`volume,
`ml
`
`Lipid in
`chloroform
`layer, g
`
`*Including 80 ml from tissue.
`tNo additional solvents required to render system biphasic.
`
`and chloroform were added t o render the systems biphasic. T h e final composi-
`tions of these ternary mixtures are given by the points A', B', C', and D'E' in
`Fig. 1. T h e lipid contents of all chloroform layers were determined a s above.
`The results (Table I) confirmed the initial hypothesis in t h a t more lipid was
`extracted by mixtures of the monophasic area of the diagram than by those of
`the biphasic area. Mixtures C, D , and E gave the highest values; however,
`the volumes of chloroform required for D and E were large and inconvenient.
`Therefore, the area around point C was considered most favorable.
`A tie-line has been drawn in Fig. 1 which for the pilrposes of this paper has
`been called "the maximum chloroform tie-line". This tie-line was particularly
`significant in this study since the lower layers of ternary systems having com-
`positions on or below it are practically 100yo chloroform (6). Systems having
`compositions above this line have chloroform layers contaminated with
`methanol and water. This explains why the chloroform layer of the system
`represented by point F in Fig. 1 was the only one found t o contain non-lipid.
`Point P in the monophasic area of Fig. 1 was chosen as a convenient starting
`point for further investigation since the composition was equivalent with
`100 ml chloroform, 200 ml methanol, and 80 ml water. T h e composition after
`adding another 100 ml chloroform is represented by point Q in the biphasic
`area of the diagram. Point R, which is below the maximum chloroform tie-line,
`is obtained by subsequent dilution with 100 ml water. This sequence was
`followed with cod muscle. The tissue (100 g) was homogenized with 100 ml
`chloroform and 200 ml methanol, filtered, and the residue rinsed with 100 ml
`chloroform in three portions. The pooled filtrates were mixed with 100 ml
`distilled water and the mixture was allowed t o separate in a 500-ml graduated
`cylinder. Separation of the layers was rather slow but satisfactory. T h e
`
`RIMFROST EXHIBIT 1165 Page 0004
`
`Can. J. Biochem. Physiol. Downloaded from www.nrcresearchpress.com by 167.206.200.52 on 05/11/17
`
`For personal use only.
`
`
`
`BLIGH AND DYER: LIPID EXTRACTION AND PURIFICATION
`
`915
`
`chloroform layer contained 0.70 g lipid, the highest yield obtained for this lot
`of cod muscle (Table I).
`Further experimentation showed that the same result was obtained when
`dilution with chloroform and water preceded filtration. T h e chloroform had
`t o be added before the water and each addition followed by blending. This
`resulted in a rapidity of separation of the layers and a yield of lipid which
`could not be obtained by adding the solvents in a different order.
`The extraction of frozen samples did not significantly lower the final tempera-
`ture of the extract; consequently it was unnecessary t o thaw samples prior t o
`extraction.
`Substitution of glass fiber filter paper for Whatman No. 1 filter paper
`showed that with cod flesh there was no significant loss of lipid with the latter
`due to adsorption. Sintered glass filters also had no advantage over the regular
`filter except perhaps with materials such as fish meal.
`
`Non-Lipid i n Chloroform Layer
`I t was shown by three methods: ( a ) the determination of chloroform-
`insoluble material, (b) purification according t o Folch et al. (2), and (c) purifica-
`tion according t o Shorland et al. (7), that the lipid extracted by the above pro-
`cedure contained no significant amounts of non-lipid material. T h e amounts of
`impurities, if present a t all, were below the sensitivity limit of each of these
`methods.
`
`Lipid in Methanol- Water Layer
`The methanol-water layer was quantitatively removed and evaporated t o
`less than 50 ml by distillation under reduced pressure a t 40-SO0 C. T h e
`coricentrate was quantitatively transferred t o a separatory funnel and extracted
`four times with an equal volume of ethyl ether. Any emulsions were broken
`by centrifugation. The weight of lipid recovered from the pooled extracts was
`8 mg. This loss, being only about 1% of the total lipid, was considered insigni-
`ficant in most applications.
`
`Lipid Remaining in Tissue Residue
`Several procedures were employed t o recover any lipid remaining in the
`extracted tissue. The most effective procedure involved re-blending the tissue
`residue and filter paper with 100 ml chloroform, followed by filtration and
`rinsing of the blendor jar and residue with a total of 50 ml chloroform. T h e
`recovered lipid weighed 40 mg, approximately 6% of the total extracted lipid.
`Re-extraction of the washed residue with chloroform-methanol-water
`did
`not yield any further amount of lipid. Thus, the initial extraction isolated
`approximately 94% of the extractable lipid.
`The washed residue was further treated with 200 ml 6 N hydrochloric acid
`a t 100' C for 90 minutes t o establish whether there was lipid remaining which
`was bound in such a way that chloroform-methanol would not remove it. T h e
`liberated "lipids" were extracted from the digest according t o A.O.A.C.
`specifications (8). The yield was 47 mg (Table 11). At least part of this
`material may have originated from inositides (see (9)), the protein complexes
`RIMFROST EXHIBIT 1165 Page 0005
`
`Can. J. Biochem. Physiol. Downloaded from www.nrcresearchpress.com by 167.206.200.52 on 05/11/17
`
`For personal use only.
`
`
`
`916
`
`CANADIAN JOURNAL OF BIOCHEMISTRY A N D PHYSIOLOGY. VOL. 37, 1959
`
`of which have been found to be resistant to chloroform-methanol extractiorl
`(2). Furthermore, the drastic acid treatment may have produced some fat-like
`material from non-lipid constituents of the muscle.
`
`TABLE I1
`Grams of lipid extracted from 100 g cod muscle by four methods
`
`Method of
`
`Dambergs (1)
`
`A.O.A.C. (9)
`
`Folch et al. (3)
`
`I'resen t
`method
`
`Sample I
`0.65
`0.67
`0.66
`0.68
`
`Sample I1
`0.76
`0.76
`0. 05*
`
`*Lipid isolated from residue after digestion with hydrochloric acid.
`
`Comparison with Existing Methods
`Samples of well-mixed minced cod muscle (100 g each) were extracted using
`the described method and the methods of Dambergs ( I ) , Folch et al. (3), and
`A.O.A.C. (8). The results are shown in Table 11. The average lipid content
`of sample I as determined by the present method was 0.666% (standard
`deviation, 0.013), which was higher than the values obtained using the A.O.A.C.
`method (8). For sample 1 I , which had a higher lipid content than sample I,
`the method of Folch et al. (3) and the present method gave yields which were
`not significantly different; however, the extract of the former method usually
`contained small amounts of non-lipid material. T h e values obtained using the
`method of Dambergs (1) were somewhat lower than those of the other methods.
`T h e flesh of several species of fish (100-g samples) with different fat contents,
`and two lots of fish meal (10-g samples) were analyzed by the described method
`
`TABLE I11
`Grams of lipid extracted from various samples using the present method
`and the method of Dambergs (1)
`
`Present
`method
`
`Method of
`Dambergs (1)
`
`Sample*
`
`Halibut fillets
`Cod fillets (pre-rigor)
`Cod fillets (5 yr at - 10' F)
`Cod fillets (18 mo a t +lo0 F)
`Swordfish steaks
`Rosefish fillets
`Plaice fillets
`Fish meal (A)
`Fish meal (B)
`
`*All samples except the fish meals were 100 g; of the meals 10 g was used per extraction.
`+Tissue residue was not washed.
`
`RIMFROST EXHIBIT 1165 Page 0006
`
`Can. J. Biochem. Physiol. Downloaded from www.nrcresearchpress.com by 167.206.200.52 on 05/11/17
`
`For personal use only.
`
`
`
`BLIGH AND DYER: LIPID EXTRACTION AND PURIFICATION
`
`917
`
`and that of Dambergs (1). The results in Table I11 show that the new method
`for the most part extracted more lipid than the method of Dambergs (1).
`
`Discussion
`When the proper proportions of chloroform and/or water are added to a
`miscible chloroform-methanol-water mixture, the system separates into two
`layers. The proportions may be chosen in such a manner that the lower layer
`will be practically 100yo chloroform and the upper layer nearly all methanol-
`water. This concept was applied to the extraction and purification of tissue
`lipids. I t was found to be very effective; the miscible solvent mixture was an
`efficient lipid extractant which upon the addition of chloroform and water
`separated into two layers and in so doing separated the lipids from the non-
`lipids. The method which was developed is simpler and less time-consuming
`than procedures hitherto reported and is well adapted to routine lipid analyses,
`and also to lipid composition studies since only mild treatment of the samples
`is involved. For practical purposes, total lipid extraction was complete;
`slightly more fatty material could be isolated onl~r after hydrochloric acid
`digestion. The separation of lipids and non-lipids was nearly quantitative.
`The results of lipid determinations on a sample of cod muscle were reproducible
`to within + 2y0 (standard deviation).
`In some applications it may be unnecessary to wash the tissue residue since
`the washings contained only 6y0 of the total lipid and the lipid composition of
`the washings, as shown by free fatty acid and phospholipid determinations,
`did not appear to differ from that of the original extract.
`Some chloroforn~ is evaporated during filtration of the homogenate but this
`is not serious provided the decrease in volume is not too great.
`
`References
`DAMBERGS. N. T. Fisheries Research Board Can. 13, 791 (1956).
`FOLCH, J., ASCOLI, I., LEES, M., MEATH, J. A., and LEBARON, F. N. J. Biol. Chern. 191,
`833 (1951).
`FOLCH, J., LEES, M., and STANLEY, G. H. S. J . Biol. Chern. 226, 497 (1957).
`DYER, W. J. arld MORTON, M. L. J. Fisheries Research Board Can. 13, 129 (1956).
`BANCROFT, W. L). Phys. Iiev. 3, 21 (1895).
`BONNER, W. D. J. Phys. Chern. 14, 738 (1910).
`SHORLAND, F. B., BRUCE, L. W., and JESSOP, A. S. Biochern. J. 52, 400 (1952).
`Official methods of analysis. 8th ed. Assoc. Offic. Agr. Chemists, Washington, D. C.
`1955. p. 311.
`GARCIA, M. D., LOVERN, J. A., and OLLEY, J. Biochern. J. 62, 99 (1956).
`
`RIMFROST EXHIBIT 1165 Page 0007
`
`Can. J. Biochem. Physiol. Downloaded from www.nrcresearchpress.com by 167.206.200.52 on 05/11/17
`
`For personal use only.
`
`