- Email: [email protected]
A comparison of extraction methods for the isolation of phospholipids from biological sources
ANALYTICAL
BIOCHEMISTRY
156.244-250 ( 1986)
A Comparison of Extraction Methods for the Isolation of Phospholipids from Biological Sources LADISLAV KOLAROVICANDNESTORC.FOURNIER NesllP Reseurch
Departmeni. Received
Ncsisfc~cLid..
CH-1800
December
10. 1985
L’evq*, Swirzrland
Four classical methods, as well as a method presented in this paper, were compared as to their efficiency in extracting phosphohpids from animal tissue. After the extractions, total lipids were separated quantitatively by DEAE-Sephadex chromatography into their acidic and nonacidic fractions. The two fractions were then further analyzed by gradient saturation high-performance thin-layer chromatography (HPTLC) combined with scanning photodensitometry after coloration with copper acetate. Of the five methods compared, the present and Christiansen’s methods based upon single-phase solvent systems proved to be more efficient than biphasic extraction procedures. The undesirable discriminatory effect exhibited by biphasic solvent systems toward acidic phospholipids which were partly retained in the aqueous phase was confirmed by statistical evaluation of the HPTLC results. Total chromogenic response of acidic phospholipids extracted using biphasic solvent systems was shown to be lower by 10-35s in comparison to the single-phase method of Christiansen. The suitability of the present method for studies involving phospholipid synthesis was confirmed by monitoring the elimination of water-soluble compounds from the single-phase extracts using a classical phospholipid precursor. 2-[3H]glycerol-3-phosphate. The labeled compound was eliminated (99.3-100s) from the single-phase postcentrifugation supematant. followed by DEAE-Sephadex chromatography. o 1~86 Academic press. 1”~. KEY WORDS: phospholipids; extraction: thin-layer chromatography, lipids; microsomes; mitochondria.
Quantitative analyses of biological materials such as heart, lung, liver or kidney tissues, frequently require a particular extraction methodology suitable for rapid isolation of lipids from a large number of samples. In the case of subcellular tissue fractions (e.g., mitochondria, microsomes, or membrane-bound lipid particles) of rat, hamster, or guinea pig organs. the sample amount available for lipid isolation is reduced to milligram or submilligram quantities, depending upon the number of animals sacrificed. Such samples are difficult to extract without losses, particularly when biphasic solvent systems are used. Folch et al. (1) defined a biphasic solvent system procedure for extraction of brain lipids 0003-2697186
$3.00
Copyright 0 1986 by Academic Press, Inc. All rights of reproduction in any form reserved.
244
briefly as follows: 1 vol of tissue is homogenized with an excess of chloroform-methanol (CM)’ 2: 1 (v/v) so that the final sample volume corresponds exactly to 20 vol of the original tissue. Then the homogenate is mixed with 0.2 times its volume of a defined aqueous salt solution whereby a biphasic system is obtained with lipids in the lower phase. Bligh and Dyer (3) found that the above procedure had the disadvantage of employing large volumes of solvent and consequently tisI Abbreviations used: HIP. hexane-isopropanol: CM. chloroform-methanol; HPTLC, high-performance thinlayer chromatography; PH. peak height: PC, phosphatidylcholine: PE, phosphatidylethanolamine.
EXTRACTION
OF
sue and CHC13-MeOH as well as water volumes were redefined. Thus their solvent system separated into practically 100% chloroform and nearly all methanol-water phases. Then Daae and Bremer (3) as well as Bjerve rt ul. (4) found a biphasic n-butanol-water system superior to CHCl,-MeOH systems (1.2) as to its efficiency to extract lipids from rat liver homogenates. The solvent system was also described as being able to penetrate the membrane structure and solubilize the lipids while the proteins remained in an apparently undenatured form (4). However. Christiansen (5) demonstrated for the biphasic extraction procedures of Folch et al. ( 1) as well as that of Bligh and Dyer (2) that the methanol-water (washing) phase retained up to 3 1% of the total radioactivity ( I ‘“C-labeled fatty acids) added to incubation mixtures before extraction. Consequently, Christiansen described the use of single-phase CHC13-MeOH solvent system to obtain an unwashed lipid extract which, after concentration to dryness, was converted into a less polar chloroform extract and a more polar methanol-based extract, thus recovering all radioactive compounds including acyl-CoA formed or glycerides synthesized from I -14Clabeled fatty acids. Unfortunately. the use of two extracts of different polarity for analyzing one sample still presents specitic inconveniences in combination with liquid-column preseparation techniques, sample derivatization, HPLC or capillary gas chromatography, and others. Hara and Radin (6) suggested that lipids can be successfully extracted from rat or mouse brains using a single-phase hexane-isopropano1 (HIP) 3:2 (v/v) solvent system. These authors found the system attractive for several reasons, e.g.. reduced tendency to form interfacial fluff. extraction of only that portion of proteolipid protein which resolubilized in CHC13-MeOH after concentration, avoidance of washing procedures in most cases. freedom from reactive impurities such as found in CHC13 (i.e., HCl and phosgene), ultraviolet
PHOSPHOLIPIDS
245
transparency suitable for subsequent column chromatography and monitoring (7). solvent density suitable for sample centrifugation, an increased lipid-to-proteolipid protein ratio in unwashed extracts. reduced toxicity, and others. From a biochemical point of view, the use of isopropanol for lipid extraction was suggested to inhibit the action of lipases which are responsible for accumulation of phosphatidic acid in plant tissue extracts as mentioned by Christie (8). A method which is a modification of the procedure described by Hara and Radin (6) and which is presented in this paper is based upon a single-phase extraction of lipids from 200 ~1 rat heart microsomal suspension without any washing procedure, using IO ml HIP as the solvent. Thus. in this present work, four previous methods (l-3.5) as well as the one described presently were compared as to their etficiency to extract lipids from rat heart microsomal suspension. A broad-spectrum composition of the substrate offered the possibility ofdetecting any discriminatory tendencies (see Abstract) caused by the extraction procedures or solvents toward apolar. complex, and acidic as well as nonacidic lipid constituents. The subsequent extracts were analyzed by gradient saturation high-performance thin-layer chromatography (HPTLC) combined with scanning photodensitometry as described previously in detail (9,lO). MATERIALS
AND
METHODS
Matcvkls. Phospholipid standards, listed in Table I, were obtained from Supelco (Bellefonte, Pa.), silica gel HPTLC plates from Merck (No. 5628). glycerol 3-phosphate from Merck (No. 4 168). and 2-[3H]glycerol 3-phosphate (disodium salt) from New England Nuclear. All solvents were of analytical grade. Isolrtiott
~1’ tnicr0s0ttw.s
~fiwtn raf hearts.
Twenty-four hearts were removed from 2.50g Sprague-Dawley male rats and then washed, minced with scissors,and stabilized for 30 min in the following buffer: 1.98 X IO-’ M Tris,
267 128 716 217 x2 195 1305 1032 2910
Phosphatidylsetine Unknown 3 Phosphatidylinositol Phosphatidylglycerol Phosphattdic acid IJnknown 4 Cardiolipm Free fatty acids CPHh
16.4 12.8 52.3 7. I 6.8 7.2 46.4 X2.6
7.5 41.‘) I.0 35.6 39.0 1.5 17.4 2.6 34.5 78.0
SD’
t 5)
100 100 100 100 100 100 100 100 100
100 100 100 100 100 IO0 100 100 100 100 100
‘kR
280 I27 750 223 84 220 1356 1115 3040
41 1533 136 494 I824 130 1772 52 860 1335 4356
PH
17.0 12.9 5x.4 15.1 4.3 13.0 41.5 96.9
2.4 61.9 2.6 8.7 50.5 8.9 21.8 2.7 27.4 57.5
SD
I05 99 I05 103 102 113 104 I08 I05
33 77 107 94 100 IOX 100 104 95 96 100
Y
265 76 645 215 53 158 II73 788 2585
21.7 9.3 49.9 18.0 3.0 10.4 15.6 9.4
3.1 30.3 22.9 6.9 22.6 2.2 26.9 57.8 -
105 465 1801 94 1755 47 857 133x 4220
SD 15.6
PH
EXTRACTS
YY 53 90 YY 65 81 90 76 89
83 x9 98 7x 99 94 95 96 96
I40
%>R
Folch el al ’ CHCI,-MeOH (I )
LIPID
I75
OF MICROSOMAL
Present work” Hexane-t-PrOH
EVALUATION
I
131 38 290 137 39 90 II31 903 1856
60 51 II7 492 IX04 I I4 1770 54 896 1425 4297
PH
9.4 3.4 24.2 8.9 3.4 9.0 23.8 20.5 -
8.2 2.7 4.5 37.5 37.8 6.3 30.2 3.1 9.3 91.1
SD
Bligh and CHCIJ-MeOH
Dyer” (2)
49 30 41 63 48 46 87 88 64
48 3 92 94 99 95 100 I08 99 IO2 98
WR
194 72 534 163 51 67 952 921 2033
I02 401 1759 100 1685 43 803 1321 4047
71
PH
Daae
13.7 6.9 4x. I 5.0 1.Y 6.3 2x.3 65.9
9.8 39.3 16.7 5.3 16.0 4.9 15.0 97.7
7.7 -
SD
and Bremerb n-BuOH (3)
73 56 75 75 62 34 73 90 70
80 76 96 x3 95 86 x9 95 92
57
RR
Nor? The average peak hetght (PH). standard deviation. and relative percentage recovery (%R) values of the nonacidic (upper sectton) and acidic (lower section) lipid constituents extracted by five dilTerent methods from the rat heart microsomal suspension as obtamed by HPTLC combined with scanning photodensitometry. The average PH values were obtained by a twofold scanning of three HPTLC plates. An arbitrary value of 100 was attrtbuted to the average PH of each lipid component extracted by the method of Christiansen (5) to calculate %R. a Monophasic extraction. ’ Biphasic extraction. ’ PH = peak height units x 0.00 I. “.f for n = 6. ‘Standard devtation. ’ Also start of the chromatogram. p Without unknowns 2 and 3. monoglycerides, cholesterol. and triglycerides. h Without free fatty acids.
I’5 I999 127 525 I830 120 1777 50 901 I398 4379
PH ‘XI
Unknown I’ Unknown 2 Lysophosphatidylcholine Sphingomyelin Phosphatidylcholine Lysophosphatidylethanolamme Phosphatidylethanolamine Monoglycende Cholesterol Tnglycetides XPHP
Lipids
Christiansen” CHCI,-MeOH
HPTLC
TABLE
EXTRACTION
OF
0.25 M sucrose, 2 X 10e3 M EDTA, and 0.1% fatty acid-free bovine albumin. Final pH was adjusted to 7.4 with 5 M HCI. After homogenization ( 15 strokes in a Potter system) of the hearts in the above buffer and 10 min centrifugation at 4°C (320 g), the supernatant was centrifuged for 10 min at 3 100 g. The resulting supernatant was again centrifuged for 10 min at 3 100 g and then 60 min at 105,OOOg. The microsomal pellet was resuspended in the following buffer: 7.9 X lop3 M sodium phosphate (pH 7.45), 3.97 X lop4 M EDTA. 6.7 X lo-’ M KCI, and 0.1 M sucrose. The protein content (55 mg/ml) was determined using bovine serum albumin as a standard. One milliliter of the microsomal suspension contained 18 mg lipids. IImzuw-i-PrOfl E.uraction. A 20-ml Pyrex test tube (Corning, Avon, France) fitted with a polytetrafluoroethylene-lined screw cap, containing 200 ~1 rat heart microsomal suspension (3.6 mg proteins) and 200 ~1 distilled water. was placed in a 50°C water-bath sonicator (100 W). As soon as the sample was dispersed (l-5 min), 10 ml HIP (3:2: v/v) was pipetted into the test tube during sonication. Sample sonication was stopped I min later. After centrifugation for 1 min at 25OOg, the clear liquid phase was decanted and then concentrated to dryness using a Speed Vat vacuum evaporator (Savant Instruments Inc.. Hicksville, N.Y.). CHC13-h~c~Oll E.~truction. Three CM procedures as described by Folch et al. (1), Bligh and Dyer (2) and Christiansen (5), respectively, were adapted for 200~~1 samples of the microsomal suspension. To decrease any losses of lipids incidental to the washing procedures. adequate salt solutions were used ( 1), the samples were centrifuged for 5 min at 3500~. and then the upper phase was removed by aspiration. The last portion of the upper phase (300-500 ~1) was aspirated using a 500~~1 syringe (Camag, Muttenz, Switzerland) fitted with a thin needle (o.d. 0.45 mm). Pasteur pipets plugged with 25-30 mg defatted cotton were used for filtration procedures.
PHOSPHOLIPIDS
247
n-BuOH Extraction. A 200~~1 portion of microsomal suspension was mixed with 800 ~1 HZ0 and then extracted with 1.5 ml IIBuOH as described by Daae and Bremer (3). Both the extraction and washing steps were followed by sample centrifugation for 15 min at 3500~ to accelerate separation or clarification of the phases. The BuOH phase was removed quantitatively by aspiration as already described and then concentrated to dryness with a stream of nitrogen at 50°C. HPTLC und .scunning Qhototr’c’nsitorl~Jtr~.. The lipid extracts were redissolved in I ml CHC13--MeOH-HI0 (30:60:8; v/v) and then rapidly separated into the acidic and nonacidic fractions on a disposable minicolumn (ca. 1 ml bed volume) by ion-exchange DEAE-Sephadex chromatography as described previously (9.1 1). All details on the HPTLC analysis of polar lipids were as already published (9,10). The acidic lipid fractions were dissolved by sonication in 1 ml CHClj without desalting. the nonacidic fraction in 1 ml toluene-MeOH (2:1: v/v) or as stated in the parent extraction methods. Four microliters of each fraction was deposited on the HPTLC plate using a Linomat 111apparatus (CAMAG). After elution on boric acid-impregnated plates and charring [(CH3COO)ZCu/H3P04]. the chromatograms were analyzed in a reflectance mode using the Camag-TLC Scanner I interfaced with a Spectra-physics 4100 computing integrator. Both peak height (PH: 1 PH unit = 0.94 pV) and peak area were measured using the valley-tovalley mode of integration which sets baseline points at the low level between peaks. Afoniloring the ,&Up cf u,atcr-solllhit~ pho.~pholipitl prtwrr.sor.s. Rat heart mitochondria were isolated as described previously ( 12) and then re:suspended in the same buffer as indicated for microsomes. To each of three Pyrex test tubes containing both glycerol 3-phosphate (4.33 mg) and 2-[3H]glycerol 3-phosphate (2.27 FCi) in 300 ~1 of the above buffer. 200~~1 volumes of mitochondrial suspension (200 pg proteins) were added, followed by IO-
248
KOLAROVIC
AND
ml volumes of HIP as already described (see Hexane-isopropanol extraction). Aliquots ( 100 ~1) were taken from each sample for radioactivity measurements after the following steps described previously: sonication (a). centrifugation (b), DEAE-Sephadex chromatography, i.e., nonacidic (c) and acidic (d) eluates. After evaporation to dryness, radioactivity of each aliquot was then measured in 15 ml of Insta-Gel II from Packard. RESULTS
AND
DISCUSSION
Figure 1 shows the HPTLC separation of lipids isolated from the rat heart microsomal suspension by the five different extraction
TG CH
MG PE . LPE El
FIG. I. HPTLC of the acidic (left) and nonacidic lipids present in rat heart microsomal suspension. The chromatogram illustrates a comparison of the following biphasic (A. B. and D) and monophasic (C and E) extraction procedures: (A) Folch e/ ul. [CHCI,-MeOH: Ref. (I)]; (B) Bligh and Dyer [CHCI1-MeOH: Ref. (2)]; (C) Christiansen [CHCI,-MeOH: Ref. (5)]: (D) Daae and Bremer [n-BuOHH,O; Ref. (3)];(E) hexane-i-PrOH (present work). HPTLC conditions: elution solvent. CHCI,-MeOH-triethylamineHZ0 (30:35:34:8: v/v): sorbent impregnation agent. 2% boric acid in absolute ethanol: boric acid zone as measured from the top of the HPTLC plate, 70 mm; gradient saturation solvent. 2 X 0.06 ml triethylamine/min: elution time, 28 mitt; elution temperature, 24°C. For other details, see Materials and Methods or Refs. (9,lO). Start of the chromatogram was situated ca. 7.5 mm from the lower end of the HPTLC plate.
FOURNIER
procedures. Biphasic lipid extracts (i.e.. A. B, and D) differ visually from the monophasic extracts (C and E) by almost complete absence of the polar materials (U2) migrating close to the start of the chromatogram and which remain to be identified. Generally. ceramides with more than five carbohydrate residuesare partitioned into the upper (washing) phase along with a major part of gangliosides and minor amounts of phospholipids (13). In the present case, more acidic (e.g., phosphatidylserine. phosphatidylinositol, or cardiolipin) than nonacidic phospholipids were either lost in the upper phase or not extracted from the microsomal suspension. The tendency was confirmed by a twofold scanning of three HPTLC plates as shown in Table 1. The results demonstrated that significant lossesof acidic phospholipidsoccur when biphasic solvent systems are used. For instance, the sumsof peak height values (S PH) for biphasic extracts (l-3) were by lo-35%’ lower in comparison to the single-phaseextracts prepared by following the procedure of Christiansen (5). Both the mono- and biphasic extraction systemsexhibited only a minimum discriminatory effect toward phosphatidylcholine (PC) and phosphatidylethanolamine (PE), which were the major phospholipids present in the nonacidic fraction. However. the level of phospholipids more hydrophilic (e.g.. lysophosphatidylcholine, lysophosphatidylethanolamine, or sphingomyelin) than PC or PE was comparatively lower in biphasic extracts. The single-phasesystem composed of HIP and Hz0 produced slightly higher phospholipid recoveries than the CM procedure of Christiansen (Table 1) which converts the single-phaseextract, after concentrations to dryness.to two extracts of different polarity (see Introduction). Such differences are possibly associatedwith the concentration step which renders a part of the single-phaseextract (usually a proteolipid protein) (6) insoluble in CHC13or MeOH, thus preventing a complete reextraction of the phospholipids with water-
EXTRACTION
OF
lesssolvents. Also, an increasednumber of extraction steps may contribute to the minor lossesof phospholipids for various reasons.On the other hand. the work of Bligh and Dyer (2) demonstrated that ca. 6% lipids remained in tissue previously extracted with their CM system. Unfortunately, these workers did not provide a phospholipid composition of these residual lipids. In this context it is worth noting that reproducible chromatograms were obtained by gradient saturation HPTLC (9.10) since elution temperature. chamber saturation process, and other conditions were strictly controlled. After elution. the HPTLC plates were exposed to 180* 0S”C for 6 & 0.1 min and then stored over silica gel in order to maintain a constant chromogenic response of the lipids. The HPTLC plates were used without predevelopment in MeOH (a cleanup procedure), which is time-consuming. For the same reason. desalting of the acidic lipid fractions on a Sephadex G-25 minicolumn (14) or Sep-Pak Cl8 cartridges (1.5) was avoided since all samples were analyzed by HPTLC in lessthan 5 h after ion-exchange chromatography. A complete sample dispersion in water before the HIP extraction solvent is added represents an essential step for quantitative recovery of phospholipids by the single-phase extraction procedure. The useof a hot (50°C) water-bath is optional: it promotes a rapid dispersion of certain samples(e.g.. milk fat globule membrane materials) especially rich in high melting point triglycerides or glycolipids. In other cases,sonication at temperatures in the range of 20-30°C is equally efficient and is practical for samplesof volumes lessthan 50 ~1. Also. a complete sample dispersion can be achieved in lessthan 1 min by vortexing the aqueous mixture (400 ~1)with quartz sand (50-100 mg), followed by pipetting the HIP extraction solvent into the test tube at a reduced vortexing speed. The definition of biphasic extraction systems is generally attractive for studies involving synthesis of lipids or phospholipids from
249
PHOSPHOLIPIDS
labeled precursors (e.g., glycerol 3-phosphate) which need to be retained in the aqueous phase. On the other hand. a successfuluse of single-phasesystems in such studies depends primarily upon the degree of insolubilization of the water-soluble precursors and their elimination by filtration or centrifugation. As monitored in the present work, the initial radioactivity resulting from the presence of 2[3H]glycerol 3-phosphate dropped from 100 to 2.7-3% after centrifugation and it was further decreased to the 0.5-0.70/u (nonacidic fraction) and O%Ilevels (acidic fraction), respectively, after DEAE-Sephadex chromatography. The above resultssuggestthat the present methodology can also be attractive for studies involving the use of labeled phospholipid precursors. In summary, this present work demonstrates that single-phase extraction methods produce phospholipid recoveries superior to those obtained by biphasic solvents. The lipid extraction based upon a single-phasesolvent of low density such as HIP is faster and more accurate than other methods. Since filtration stepsare replaced by centrifugation, washing procedures are avoided and the whole operation proceeds exclusively in one test tube. A series of samples can be extracted simultaneously. The volume of integral or subcellular tissue homogenatesneeded for one extraction may vary in the range of 0.02 to 2 ml, provided that other volumes are scaledup or down proportionally. ACKNOWLEDGMENTS
Mr.
The technical assistance of Miss Corine HervtJ Cousin is appreciated.
Monnard
and
REFERENCES I
Folch. J.. Lees. M.. and Sloane-Stanley, G. H. ( 1957) J. Bid. C1wn. 226, 497-509. 2. Bligh. E. G.. and Dyer, W. J. ( 1959) Cbnad .I. Biochn. F~f~~iOl. 37, 9 1 I-9 17. 3. Daae. L. N. W.. and Bremer, J. ( 1970) B/oc,him. BiopllJ:c :icw 210, Y3- 104.
250
KOLAROVIC
4. Bjerve,
K. S., Daae, L. N. W., and Bremer,
AND
J. (1974)
And. Biochetn. 58, 238-245. 5. Christiansen, K. (1975) .4nal. Biochetn. 66. 93-99. 6. Hara, A., and Radin. N. S. (1978) .4nal. Biochem. 90. 420-426. 7. Radin, N. S. (1978) J. Lipid Res. 19, 922-924. 8. Christie. W. W. (1973) in Lipid Analysis (Christie, W. W.. ed.), pp. 30-41, Pergamon. Oxford. 9. Kolarovic. L., and Traitler. H. ( 1985) J. High Resolzrt. Chromatogr. Chromatogr. Commun. 8, 34 l-346. 10. Kolarovic, L., Cousin. H., and Traitler. H. (1985) J.
High Resolni. Chromatogr. Chromatqgr. Comtmttz. 8.838-842.
FOURNIER Il.
Ledeen.
R. W.. Yu, R. K., and Eng. L. F. (1973)
J
Neurochetn. 21. 829-839. 12. Fournier. N. C., and Rahim. M. (1985) Biochemisfrv 24.2387-2396. 13. Hakomori. S. (I 983) in Sphingolipid Biochemistry (Hanahan, D. J.. ed.), Vol. 3. pp. I-165, Plenum. New York. 14. Fishman. P. H.. Quarles. R. H., and Max, S. R. (1979) in Densitometry in Thin Layer Chromatography (Touchstone. J. C., and Sherma. J.. eds.). pp. 3 15327, Wiley. New York. 15. Williams. M. A., and McCluer, R. H. (1980) J. Nell-
rochetn. 35. 266-269.
BIOCHEMISTRY
156.244-250 ( 1986)
A Comparison of Extraction Methods for the Isolation of Phospholipids from Biological Sources LADISLAV KOLAROVICANDNESTORC.FOURNIER NesllP Reseurch
Departmeni. Received
Ncsisfc~cLid..
CH-1800
December
10. 1985
L’evq*, Swirzrland
Four classical methods, as well as a method presented in this paper, were compared as to their efficiency in extracting phosphohpids from animal tissue. After the extractions, total lipids were separated quantitatively by DEAE-Sephadex chromatography into their acidic and nonacidic fractions. The two fractions were then further analyzed by gradient saturation high-performance thin-layer chromatography (HPTLC) combined with scanning photodensitometry after coloration with copper acetate. Of the five methods compared, the present and Christiansen’s methods based upon single-phase solvent systems proved to be more efficient than biphasic extraction procedures. The undesirable discriminatory effect exhibited by biphasic solvent systems toward acidic phospholipids which were partly retained in the aqueous phase was confirmed by statistical evaluation of the HPTLC results. Total chromogenic response of acidic phospholipids extracted using biphasic solvent systems was shown to be lower by 10-35s in comparison to the single-phase method of Christiansen. The suitability of the present method for studies involving phospholipid synthesis was confirmed by monitoring the elimination of water-soluble compounds from the single-phase extracts using a classical phospholipid precursor. 2-[3H]glycerol-3-phosphate. The labeled compound was eliminated (99.3-100s) from the single-phase postcentrifugation supematant. followed by DEAE-Sephadex chromatography. o 1~86 Academic press. 1”~. KEY WORDS: phospholipids; extraction: thin-layer chromatography, lipids; microsomes; mitochondria.
Quantitative analyses of biological materials such as heart, lung, liver or kidney tissues, frequently require a particular extraction methodology suitable for rapid isolation of lipids from a large number of samples. In the case of subcellular tissue fractions (e.g., mitochondria, microsomes, or membrane-bound lipid particles) of rat, hamster, or guinea pig organs. the sample amount available for lipid isolation is reduced to milligram or submilligram quantities, depending upon the number of animals sacrificed. Such samples are difficult to extract without losses, particularly when biphasic solvent systems are used. Folch et al. (1) defined a biphasic solvent system procedure for extraction of brain lipids 0003-2697186
$3.00
Copyright 0 1986 by Academic Press, Inc. All rights of reproduction in any form reserved.
244
briefly as follows: 1 vol of tissue is homogenized with an excess of chloroform-methanol (CM)’ 2: 1 (v/v) so that the final sample volume corresponds exactly to 20 vol of the original tissue. Then the homogenate is mixed with 0.2 times its volume of a defined aqueous salt solution whereby a biphasic system is obtained with lipids in the lower phase. Bligh and Dyer (3) found that the above procedure had the disadvantage of employing large volumes of solvent and consequently tisI Abbreviations used: HIP. hexane-isopropanol: CM. chloroform-methanol; HPTLC, high-performance thinlayer chromatography; PH. peak height: PC, phosphatidylcholine: PE, phosphatidylethanolamine.
EXTRACTION
OF
sue and CHC13-MeOH as well as water volumes were redefined. Thus their solvent system separated into practically 100% chloroform and nearly all methanol-water phases. Then Daae and Bremer (3) as well as Bjerve rt ul. (4) found a biphasic n-butanol-water system superior to CHCl,-MeOH systems (1.2) as to its efficiency to extract lipids from rat liver homogenates. The solvent system was also described as being able to penetrate the membrane structure and solubilize the lipids while the proteins remained in an apparently undenatured form (4). However. Christiansen (5) demonstrated for the biphasic extraction procedures of Folch et al. ( 1) as well as that of Bligh and Dyer (2) that the methanol-water (washing) phase retained up to 3 1% of the total radioactivity ( I ‘“C-labeled fatty acids) added to incubation mixtures before extraction. Consequently, Christiansen described the use of single-phase CHC13-MeOH solvent system to obtain an unwashed lipid extract which, after concentration to dryness, was converted into a less polar chloroform extract and a more polar methanol-based extract, thus recovering all radioactive compounds including acyl-CoA formed or glycerides synthesized from I -14Clabeled fatty acids. Unfortunately. the use of two extracts of different polarity for analyzing one sample still presents specitic inconveniences in combination with liquid-column preseparation techniques, sample derivatization, HPLC or capillary gas chromatography, and others. Hara and Radin (6) suggested that lipids can be successfully extracted from rat or mouse brains using a single-phase hexane-isopropano1 (HIP) 3:2 (v/v) solvent system. These authors found the system attractive for several reasons, e.g.. reduced tendency to form interfacial fluff. extraction of only that portion of proteolipid protein which resolubilized in CHC13-MeOH after concentration, avoidance of washing procedures in most cases. freedom from reactive impurities such as found in CHC13 (i.e., HCl and phosgene), ultraviolet
PHOSPHOLIPIDS
245
transparency suitable for subsequent column chromatography and monitoring (7). solvent density suitable for sample centrifugation, an increased lipid-to-proteolipid protein ratio in unwashed extracts. reduced toxicity, and others. From a biochemical point of view, the use of isopropanol for lipid extraction was suggested to inhibit the action of lipases which are responsible for accumulation of phosphatidic acid in plant tissue extracts as mentioned by Christie (8). A method which is a modification of the procedure described by Hara and Radin (6) and which is presented in this paper is based upon a single-phase extraction of lipids from 200 ~1 rat heart microsomal suspension without any washing procedure, using IO ml HIP as the solvent. Thus. in this present work, four previous methods (l-3.5) as well as the one described presently were compared as to their etficiency to extract lipids from rat heart microsomal suspension. A broad-spectrum composition of the substrate offered the possibility ofdetecting any discriminatory tendencies (see Abstract) caused by the extraction procedures or solvents toward apolar. complex, and acidic as well as nonacidic lipid constituents. The subsequent extracts were analyzed by gradient saturation high-performance thin-layer chromatography (HPTLC) combined with scanning photodensitometry as described previously in detail (9,lO). MATERIALS
AND
METHODS
Matcvkls. Phospholipid standards, listed in Table I, were obtained from Supelco (Bellefonte, Pa.), silica gel HPTLC plates from Merck (No. 5628). glycerol 3-phosphate from Merck (No. 4 168). and 2-[3H]glycerol 3-phosphate (disodium salt) from New England Nuclear. All solvents were of analytical grade. Isolrtiott
~1’ tnicr0s0ttw.s
~fiwtn raf hearts.
Twenty-four hearts were removed from 2.50g Sprague-Dawley male rats and then washed, minced with scissors,and stabilized for 30 min in the following buffer: 1.98 X IO-’ M Tris,
267 128 716 217 x2 195 1305 1032 2910
Phosphatidylsetine Unknown 3 Phosphatidylinositol Phosphatidylglycerol Phosphattdic acid IJnknown 4 Cardiolipm Free fatty acids CPHh
16.4 12.8 52.3 7. I 6.8 7.2 46.4 X2.6
7.5 41.‘) I.0 35.6 39.0 1.5 17.4 2.6 34.5 78.0
SD’
t 5)
100 100 100 100 100 100 100 100 100
100 100 100 100 100 IO0 100 100 100 100 100
‘kR
280 I27 750 223 84 220 1356 1115 3040
41 1533 136 494 I824 130 1772 52 860 1335 4356
PH
17.0 12.9 5x.4 15.1 4.3 13.0 41.5 96.9
2.4 61.9 2.6 8.7 50.5 8.9 21.8 2.7 27.4 57.5
SD
I05 99 I05 103 102 113 104 I08 I05
33 77 107 94 100 IOX 100 104 95 96 100
Y
265 76 645 215 53 158 II73 788 2585
21.7 9.3 49.9 18.0 3.0 10.4 15.6 9.4
3.1 30.3 22.9 6.9 22.6 2.2 26.9 57.8 -
105 465 1801 94 1755 47 857 133x 4220
SD 15.6
PH
EXTRACTS
YY 53 90 YY 65 81 90 76 89
83 x9 98 7x 99 94 95 96 96
I40
%>R
Folch el al ’ CHCI,-MeOH (I )
LIPID
I75
OF MICROSOMAL
Present work” Hexane-t-PrOH
EVALUATION
I
131 38 290 137 39 90 II31 903 1856
60 51 II7 492 IX04 I I4 1770 54 896 1425 4297
PH
9.4 3.4 24.2 8.9 3.4 9.0 23.8 20.5 -
8.2 2.7 4.5 37.5 37.8 6.3 30.2 3.1 9.3 91.1
SD
Bligh and CHCIJ-MeOH
Dyer” (2)
49 30 41 63 48 46 87 88 64
48 3 92 94 99 95 100 I08 99 IO2 98
WR
194 72 534 163 51 67 952 921 2033
I02 401 1759 100 1685 43 803 1321 4047
71
PH
Daae
13.7 6.9 4x. I 5.0 1.Y 6.3 2x.3 65.9
9.8 39.3 16.7 5.3 16.0 4.9 15.0 97.7
7.7 -
SD
and Bremerb n-BuOH (3)
73 56 75 75 62 34 73 90 70
80 76 96 x3 95 86 x9 95 92
57
RR
Nor? The average peak hetght (PH). standard deviation. and relative percentage recovery (%R) values of the nonacidic (upper sectton) and acidic (lower section) lipid constituents extracted by five dilTerent methods from the rat heart microsomal suspension as obtamed by HPTLC combined with scanning photodensitometry. The average PH values were obtained by a twofold scanning of three HPTLC plates. An arbitrary value of 100 was attrtbuted to the average PH of each lipid component extracted by the method of Christiansen (5) to calculate %R. a Monophasic extraction. ’ Biphasic extraction. ’ PH = peak height units x 0.00 I. “.f for n = 6. ‘Standard devtation. ’ Also start of the chromatogram. p Without unknowns 2 and 3. monoglycerides, cholesterol. and triglycerides. h Without free fatty acids.
I’5 I999 127 525 I830 120 1777 50 901 I398 4379
PH ‘XI
Unknown I’ Unknown 2 Lysophosphatidylcholine Sphingomyelin Phosphatidylcholine Lysophosphatidylethanolamme Phosphatidylethanolamine Monoglycende Cholesterol Tnglycetides XPHP
Lipids
Christiansen” CHCI,-MeOH
HPTLC
TABLE
EXTRACTION
OF
0.25 M sucrose, 2 X 10e3 M EDTA, and 0.1% fatty acid-free bovine albumin. Final pH was adjusted to 7.4 with 5 M HCI. After homogenization ( 15 strokes in a Potter system) of the hearts in the above buffer and 10 min centrifugation at 4°C (320 g), the supernatant was centrifuged for 10 min at 3 100 g. The resulting supernatant was again centrifuged for 10 min at 3 100 g and then 60 min at 105,OOOg. The microsomal pellet was resuspended in the following buffer: 7.9 X lop3 M sodium phosphate (pH 7.45), 3.97 X lop4 M EDTA. 6.7 X lo-’ M KCI, and 0.1 M sucrose. The protein content (55 mg/ml) was determined using bovine serum albumin as a standard. One milliliter of the microsomal suspension contained 18 mg lipids. IImzuw-i-PrOfl E.uraction. A 20-ml Pyrex test tube (Corning, Avon, France) fitted with a polytetrafluoroethylene-lined screw cap, containing 200 ~1 rat heart microsomal suspension (3.6 mg proteins) and 200 ~1 distilled water. was placed in a 50°C water-bath sonicator (100 W). As soon as the sample was dispersed (l-5 min), 10 ml HIP (3:2: v/v) was pipetted into the test tube during sonication. Sample sonication was stopped I min later. After centrifugation for 1 min at 25OOg, the clear liquid phase was decanted and then concentrated to dryness using a Speed Vat vacuum evaporator (Savant Instruments Inc.. Hicksville, N.Y.). CHC13-h~c~Oll E.~truction. Three CM procedures as described by Folch et al. (1), Bligh and Dyer (2) and Christiansen (5), respectively, were adapted for 200~~1 samples of the microsomal suspension. To decrease any losses of lipids incidental to the washing procedures. adequate salt solutions were used ( 1), the samples were centrifuged for 5 min at 3500~. and then the upper phase was removed by aspiration. The last portion of the upper phase (300-500 ~1) was aspirated using a 500~~1 syringe (Camag, Muttenz, Switzerland) fitted with a thin needle (o.d. 0.45 mm). Pasteur pipets plugged with 25-30 mg defatted cotton were used for filtration procedures.
PHOSPHOLIPIDS
247
n-BuOH Extraction. A 200~~1 portion of microsomal suspension was mixed with 800 ~1 HZ0 and then extracted with 1.5 ml IIBuOH as described by Daae and Bremer (3). Both the extraction and washing steps were followed by sample centrifugation for 15 min at 3500~ to accelerate separation or clarification of the phases. The BuOH phase was removed quantitatively by aspiration as already described and then concentrated to dryness with a stream of nitrogen at 50°C. HPTLC und .scunning Qhototr’c’nsitorl~Jtr~.. The lipid extracts were redissolved in I ml CHC13--MeOH-HI0 (30:60:8; v/v) and then rapidly separated into the acidic and nonacidic fractions on a disposable minicolumn (ca. 1 ml bed volume) by ion-exchange DEAE-Sephadex chromatography as described previously (9.1 1). All details on the HPTLC analysis of polar lipids were as already published (9,10). The acidic lipid fractions were dissolved by sonication in 1 ml CHClj without desalting. the nonacidic fraction in 1 ml toluene-MeOH (2:1: v/v) or as stated in the parent extraction methods. Four microliters of each fraction was deposited on the HPTLC plate using a Linomat 111apparatus (CAMAG). After elution on boric acid-impregnated plates and charring [(CH3COO)ZCu/H3P04]. the chromatograms were analyzed in a reflectance mode using the Camag-TLC Scanner I interfaced with a Spectra-physics 4100 computing integrator. Both peak height (PH: 1 PH unit = 0.94 pV) and peak area were measured using the valley-tovalley mode of integration which sets baseline points at the low level between peaks. Afoniloring the ,&Up cf u,atcr-solllhit~ pho.~pholipitl prtwrr.sor.s. Rat heart mitochondria were isolated as described previously ( 12) and then re:suspended in the same buffer as indicated for microsomes. To each of three Pyrex test tubes containing both glycerol 3-phosphate (4.33 mg) and 2-[3H]glycerol 3-phosphate (2.27 FCi) in 300 ~1 of the above buffer. 200~~1 volumes of mitochondrial suspension (200 pg proteins) were added, followed by IO-
248
KOLAROVIC
AND
ml volumes of HIP as already described (see Hexane-isopropanol extraction). Aliquots ( 100 ~1) were taken from each sample for radioactivity measurements after the following steps described previously: sonication (a). centrifugation (b), DEAE-Sephadex chromatography, i.e., nonacidic (c) and acidic (d) eluates. After evaporation to dryness, radioactivity of each aliquot was then measured in 15 ml of Insta-Gel II from Packard. RESULTS
AND
DISCUSSION
Figure 1 shows the HPTLC separation of lipids isolated from the rat heart microsomal suspension by the five different extraction
TG CH
MG PE . LPE El
FIG. I. HPTLC of the acidic (left) and nonacidic lipids present in rat heart microsomal suspension. The chromatogram illustrates a comparison of the following biphasic (A. B. and D) and monophasic (C and E) extraction procedures: (A) Folch e/ ul. [CHCI,-MeOH: Ref. (I)]; (B) Bligh and Dyer [CHCI1-MeOH: Ref. (2)]; (C) Christiansen [CHCI,-MeOH: Ref. (5)]: (D) Daae and Bremer [n-BuOHH,O; Ref. (3)];(E) hexane-i-PrOH (present work). HPTLC conditions: elution solvent. CHCI,-MeOH-triethylamineHZ0 (30:35:34:8: v/v): sorbent impregnation agent. 2% boric acid in absolute ethanol: boric acid zone as measured from the top of the HPTLC plate, 70 mm; gradient saturation solvent. 2 X 0.06 ml triethylamine/min: elution time, 28 mitt; elution temperature, 24°C. For other details, see Materials and Methods or Refs. (9,lO). Start of the chromatogram was situated ca. 7.5 mm from the lower end of the HPTLC plate.
FOURNIER
procedures. Biphasic lipid extracts (i.e.. A. B, and D) differ visually from the monophasic extracts (C and E) by almost complete absence of the polar materials (U2) migrating close to the start of the chromatogram and which remain to be identified. Generally. ceramides with more than five carbohydrate residuesare partitioned into the upper (washing) phase along with a major part of gangliosides and minor amounts of phospholipids (13). In the present case, more acidic (e.g., phosphatidylserine. phosphatidylinositol, or cardiolipin) than nonacidic phospholipids were either lost in the upper phase or not extracted from the microsomal suspension. The tendency was confirmed by a twofold scanning of three HPTLC plates as shown in Table 1. The results demonstrated that significant lossesof acidic phospholipidsoccur when biphasic solvent systems are used. For instance, the sumsof peak height values (S PH) for biphasic extracts (l-3) were by lo-35%’ lower in comparison to the single-phaseextracts prepared by following the procedure of Christiansen (5). Both the mono- and biphasic extraction systemsexhibited only a minimum discriminatory effect toward phosphatidylcholine (PC) and phosphatidylethanolamine (PE), which were the major phospholipids present in the nonacidic fraction. However. the level of phospholipids more hydrophilic (e.g.. lysophosphatidylcholine, lysophosphatidylethanolamine, or sphingomyelin) than PC or PE was comparatively lower in biphasic extracts. The single-phasesystem composed of HIP and Hz0 produced slightly higher phospholipid recoveries than the CM procedure of Christiansen (Table 1) which converts the single-phaseextract, after concentrations to dryness.to two extracts of different polarity (see Introduction). Such differences are possibly associatedwith the concentration step which renders a part of the single-phaseextract (usually a proteolipid protein) (6) insoluble in CHC13or MeOH, thus preventing a complete reextraction of the phospholipids with water-
EXTRACTION
OF
lesssolvents. Also, an increasednumber of extraction steps may contribute to the minor lossesof phospholipids for various reasons.On the other hand. the work of Bligh and Dyer (2) demonstrated that ca. 6% lipids remained in tissue previously extracted with their CM system. Unfortunately, these workers did not provide a phospholipid composition of these residual lipids. In this context it is worth noting that reproducible chromatograms were obtained by gradient saturation HPTLC (9.10) since elution temperature. chamber saturation process, and other conditions were strictly controlled. After elution. the HPTLC plates were exposed to 180* 0S”C for 6 & 0.1 min and then stored over silica gel in order to maintain a constant chromogenic response of the lipids. The HPTLC plates were used without predevelopment in MeOH (a cleanup procedure), which is time-consuming. For the same reason. desalting of the acidic lipid fractions on a Sephadex G-25 minicolumn (14) or Sep-Pak Cl8 cartridges (1.5) was avoided since all samples were analyzed by HPTLC in lessthan 5 h after ion-exchange chromatography. A complete sample dispersion in water before the HIP extraction solvent is added represents an essential step for quantitative recovery of phospholipids by the single-phase extraction procedure. The useof a hot (50°C) water-bath is optional: it promotes a rapid dispersion of certain samples(e.g.. milk fat globule membrane materials) especially rich in high melting point triglycerides or glycolipids. In other cases,sonication at temperatures in the range of 20-30°C is equally efficient and is practical for samplesof volumes lessthan 50 ~1. Also. a complete sample dispersion can be achieved in lessthan 1 min by vortexing the aqueous mixture (400 ~1)with quartz sand (50-100 mg), followed by pipetting the HIP extraction solvent into the test tube at a reduced vortexing speed. The definition of biphasic extraction systems is generally attractive for studies involving synthesis of lipids or phospholipids from
249
PHOSPHOLIPIDS
labeled precursors (e.g., glycerol 3-phosphate) which need to be retained in the aqueous phase. On the other hand. a successfuluse of single-phasesystems in such studies depends primarily upon the degree of insolubilization of the water-soluble precursors and their elimination by filtration or centrifugation. As monitored in the present work, the initial radioactivity resulting from the presence of 2[3H]glycerol 3-phosphate dropped from 100 to 2.7-3% after centrifugation and it was further decreased to the 0.5-0.70/u (nonacidic fraction) and O%Ilevels (acidic fraction), respectively, after DEAE-Sephadex chromatography. The above resultssuggestthat the present methodology can also be attractive for studies involving the use of labeled phospholipid precursors. In summary, this present work demonstrates that single-phase extraction methods produce phospholipid recoveries superior to those obtained by biphasic solvents. The lipid extraction based upon a single-phasesolvent of low density such as HIP is faster and more accurate than other methods. Since filtration stepsare replaced by centrifugation, washing procedures are avoided and the whole operation proceeds exclusively in one test tube. A series of samples can be extracted simultaneously. The volume of integral or subcellular tissue homogenatesneeded for one extraction may vary in the range of 0.02 to 2 ml, provided that other volumes are scaledup or down proportionally. ACKNOWLEDGMENTS
Mr.
The technical assistance of Miss Corine HervtJ Cousin is appreciated.
Monnard
and
REFERENCES I
Folch. J.. Lees. M.. and Sloane-Stanley, G. H. ( 1957) J. Bid. C1wn. 226, 497-509. 2. Bligh. E. G.. and Dyer, W. J. ( 1959) Cbnad .I. Biochn. F~f~~iOl. 37, 9 1 I-9 17. 3. Daae. L. N. W.. and Bremer, J. ( 1970) B/oc,him. BiopllJ:c :icw 210, Y3- 104.
250
KOLAROVIC
4. Bjerve,
K. S., Daae, L. N. W., and Bremer,
AND
J. (1974)
And. Biochetn. 58, 238-245. 5. Christiansen, K. (1975) .4nal. Biochetn. 66. 93-99. 6. Hara, A., and Radin. N. S. (1978) .4nal. Biochem. 90. 420-426. 7. Radin, N. S. (1978) J. Lipid Res. 19, 922-924. 8. Christie. W. W. (1973) in Lipid Analysis (Christie, W. W.. ed.), pp. 30-41, Pergamon. Oxford. 9. Kolarovic. L., and Traitler. H. ( 1985) J. High Resolzrt. Chromatogr. Chromatogr. Commun. 8, 34 l-346. 10. Kolarovic, L., Cousin. H., and Traitler. H. (1985) J.
High Resolni. Chromatogr. Chromatqgr. Comtmttz. 8.838-842.
FOURNIER Il.
Ledeen.
R. W.. Yu, R. K., and Eng. L. F. (1973)
J
Neurochetn. 21. 829-839. 12. Fournier. N. C., and Rahim. M. (1985) Biochemisfrv 24.2387-2396. 13. Hakomori. S. (I 983) in Sphingolipid Biochemistry (Hanahan, D. J.. ed.), Vol. 3. pp. I-165, Plenum. New York. 14. Fishman. P. H.. Quarles. R. H., and Max, S. R. (1979) in Densitometry in Thin Layer Chromatography (Touchstone. J. C., and Sherma. J.. eds.). pp. 3 15327, Wiley. New York. 15. Williams. M. A., and McCluer, R. H. (1980) J. Nell-
rochetn. 35. 266-269.