- Email: [email protected]
Preparation of the proteolipid apoprotein from bovine heart, liver and kidney
Biochimica et Biophysica Acta, 427 (1976) 410-427
(c) Elsevier Scientific Publishing Company, Amsterdam--- Printed in The Netherlands BBA 37312 P R E P A R A T I O N OF T H E P R O T E O L I P I D APOPROTE1N HEART, LIVER AND KIDNEY
FROM
BOVINE
J. FOLCH-PI* and J. D. SAKURA Biological Research Laboratory, McLean Hoapital, Belmont, Mass. 02178, and Department ol"Biological Chemistry, Harvard Medical School, Boston, Mass. 02115 (U.S.A.)
(Received September 2nd, 1975)
SUMMARY Proteolipid apoproteins have been prepared from heart, kidney, and liver by dialysis in chloroform/methanol against chloroform/methanol, acidified chloroform/ methanol, and chloroform/methanol in succession. They are free of lipids ( - 0.05 71, P; < 0.1~o carbohydrate). They show a high content of non-polar amino acids, methionine, and tryptophan and contain little or no half-cystine. They differ from neural proteolipid apoproteins by absence of half-cystine, and of covalently bound fatty acids. As recovered from chloroform/methanol solutions, they are soluble in chloroform/methanol and insoluble in water, but a water-soluble form can be prepared by changing the solvent from chloroform/methanol to water in a stream of nitrogen. The chloroform/methanol-soluble form and the water-soluble form tire interconvertible. O R D and CD spectra of all proteolipid apoproteins indicate 60-70 !'0
INTRODUCTION The proteolipids are a group of protein-lipid complexes characterized by solubility in chloroform/methanol mixtures, and insolubility in water and aqueous solutions. They were first recognized in 1951 by Folch and Lees [1] who found them in chloroform/methanol extracts of bovine and human central nervous systems. Abbreviation: TEMED, N,N,N',N'-tetramethylethylene diamine. " Correspondence should be addressed to : Dr. Jordi Folch-Pi, Biological Research Laboratory, McLean Hospital, 115 Mill Street, Belmont, Mass. 02178, U.S.A.
411 Subsequent work in this and other laboratories found them to be specially abundant in mammalian central white matter, followed by gray matter, heart muscle, kidney and liver in order of decreasing concentration and to be otherwise widely distributed among animal and vegetable tissues usually in relation to membranous structures (see ref. 2 for a review). Most of the later work on proteolipids was carried out on mammalian central nervous white matter where they were soon traced mainly to the myelin sheath. The white matter proteolipids were partially separated from l ipids by a number of different procedures [2]. In 1966 Tenenbaum and Folch-Pi [3] obtained them free of lipids detectable by thin-layer chromatography by dialysis in succession against neutral chloroform/methanol, chloroform/methanol acidified with HC1 to 0.04 M concentration, and a succession of mixtures in which chloroform/methanol was replaced gradually by water until pure water was reached. The final result was an aqueous solution of proteolipid apoprotein. The apoprotein recovered by evaporation of water was still soluble in chloroform/methanol. The solution was otherwise very unstable and after a few days, the apoprotein precipitated out of solution yielding a precipitate that was insoluble in all organic solvent mixtures and aqueous solvents that were used. Zand [4] by optical rotatory dispersion measurements, found that the starting chloroform/methanol-soluble proteolipid protein had a highly a-helical conformation and that, in the gradual shift from chloroform/methanol to water in the course of dialysis, this conformation changed to one without detectable a-helix. An improved procedure [5] yielded preparations of lipid-free proteolipid apoprotein which were freely soluble in chloroform/methanol and in water, forming solutions that remained stable indefinitely. Sherman and Folch-Pi [6] found that the two solubilities corresponded to two conformers, a highly a-helical hydrophobic conformer, and a hydrophilic conformer in which the a-helix content is reduced by about one-half. The two conformers are interconvertible, the hydropbilic one becoming hydrophobic on simple exposure to chloroform/methanol with a return to the higher a-helix content of the hydrophobic starting material. Little work has been carried out on proteolipids from non-neural sources. It was shown early [7] that most of the heart proteolipids were present in the mitochondria. Murakami et al. [8] prepared heart proteolipids by dialysis in organic solvents, and traced most of them to mitochondria [9]. Wolfgram [10] showed the heart proteolipids to be devoid of half-cystine, a finding later confirmed by Eichberg [I 1]. None of these studies succeeded in preparing heart proteolipid apoprotein devoid of detectable lipids. This paper describes the isolation of proteolipid apoproteins from bovine heart, kidney and liver tissues, and the study of some of their chemical and physical properties. A preliminary abstract of the results obtained has appeared [12]. MATERIALS Hen egg white lysozyme and a-chymotrypsin, were obtained from Worthington Biochemical Corp. (Freehold, N.J.). Adrenocorticotropic hormone, cytochrome c, type VI, sperm whale myoglobin, glyceraldehyde-3-phosphate dehydrogenase, glucagon and dithiothreitol were purchased from Sigma Chemical Co. (St. Louis, Mo.).
412 Ovalbumin was supplied in a calibration kit from Pharmacia Fine Chemicals (Piscataway, N.J.). Tris, urea, and sodium dodecyl sulfate (lot No. 3057 and No. 1737) were ultra pure quality from Schwartz-Mann (Orangeburg, N.Y.). Acrylamide, N.N'methylene bisacrylamide, ammonium persulfate, N,N,N',N'-tetramethylethylene diamine (TEMED) and Coomassie Brilliant Blue were obtained from Bio-Rad Laboratory (Richmond, Calif.). All other chemicals were reagent grade or better. Cyanogen bromide peptides from cytochrome c were prepared according to Chu and Yasunobu [13]. METHODS
Analytical polyaco'lamide gel electrophoresis Polyacrylamide gel electrophoresis was carried out in 0.1'), sodium dodecyl sulfate following an adaptation of the general procedure of Weber and Osborn [14] as modified by Agrawal et al. [15]. A stock buffer (pH 7,2) contained 1.95 g NaHzPO4" H20, 9.66 g NazHPO4.7H20, and 2 g sodium dodecyl sulfate per liter, A gel solution was prepared by dissolving 3.35 g acrylamide monomer, 90 mg N,N'-methylene bisacrylamide in water, the final volume adjusted to 15 ml, and diluted with 16.5 ml of stock buffer. Polymerization of the gel solution was initiated by adding 1.5 ml ammonium persulfate ,10.,,~...../ow/v) and 5 0 / d TEMED. The gel solution was poured into glass tubes (5 mm 120 ram) to a height of 90 mm. To obtain a flat gel surface, isobutyl alcohol was layered on top of the gels. The electrode buffer was prepared by dilution of the stock buffer with an equal volume of water. Polyacrylamide gel electrophoresis on 0.1 '?,;, sodium dodecyt sulfate and 8 M urea, pH 6,9 was carried out according to Swank and Munkres [16]. For 30 ml gel solution, 3.62 g acrylamide, 0.13 g N,N'-methylene bisacrylamide, 13.07 g urea, 30 mg sodium dodecyl sulfate, 205/~1 H3PO4, 420 mg Tris and 11 mg EDTA were dissolved in distilled water to a final volume of 28.5 ml. The pH of the solution was adjusted to 6.9 2 0.02 with solid Tris. After deaeraticn with a water aspirator, 1.5 ml ammonium sulfate solution (0.2 ~,;), and 20 t*1 T E M E D was added. The gels were cast into 150 mm glass tubing and overlayed with isobutyl alcohol. The electrode buffer was freshly prepared. It contained 0.1 ~'0 sodium dodecyl sulfate, 0.1 M H~PO4, and 2 mM EDTA. Its pH was adjusted to 6.9 ! 0.02 with solid Tris. 40 to 100 /¢g water-soluble proteolipid apoprotein in water was mixed with 50 to 100 pl of an aqueous soluticn of 100 mg sodium dodecyl sulfate, 150 mg dithiothreitol, and 800 mg sucro,'~e made in 10 ml (referred to as medium A [15]), and heated in boiling water for approximately I min. The same procedure was followed to apply 1-5 t~g of standard proteins. Electrophoresis m sodium dodecyl sulfate or sodium dodecyl sulfate/urea was conducted at I mA per tube for 1 h followed by an increase to 3 mA per tube for approximately 16 h. The gels were then removed Dom the glass tube, and a wire was inserted to mark the dye front. The gels were stained in a 0.25~; (w/v) Coomassie Brilliant Blue according to the procedure described by Agrawal et al. [15].
ORD and CD measurements Optical rotatory dispersion and circular dichroism measurements were carried
413 out with a Cary 60 spectropolarimeter equipped with a 6001 circular dichroism attachment. All measurements were made at room temperature (,~ 23 °C). Overlapping spectra were obtained using a 0.1 and a 1.0 cm cell. Reduced mean residue rotation [m'] and ellipticities [0] were calculated using a value of 108 for the mean residue weight.
Microestimation of total P The procedure followed is a microadaptation of the method of Sperry [17]. A sample containing 0.2 to 2.5 #g of P is weighed or measured into a 10 mm × 70 mm medium heavy glass tube. The sample may be dissolved in up to 0.5 ml of water. Any organic solvent present is removed by evaporation. To the sample is now added 0.04 ml of 60 ~ HC104 and one or two small chips of alundum. The digestion is conveniently run in sets of up to 15 tubes using a beaker narrow and low enough to hold the tubes at an angle so that they protrude about 1 cm above the beaker rim. The beaker is filled with concentrated HzSO 4 up to 1 cm below the rim, and a thermometer is immersed in it to monitor the temperature. The beaker is heated gradually to insure smooth boiling. If marked browning develops, heating is stopped and one or two small drops of 30 ~o Hz02 are added to each brown digest by delivering them against the inner wall of the tube, from a fine tipped pipette calibrated to deliver droplets of about 0.01 ml volume. Browning clears immediately or upon further heating. If it recurs, the addition of HzO2 is repeated. Heating is continued until white fumes of perchloric acid appear, and then for 15 rain longer. The temperature of the bath should then be about 200 °C. Temperatures much above 200 °C might result in losses of P. At the end of the digestion, the tubes are removed from the bath, and cooled and rinsed free of HzSO4 with running water. To each tube is added 0.15 ml of water and the tubes are cooled anew by placing in a water, or ice bath. To each tube is then added 0.1 ml 25 9/00ammonium molyl~date solution and 0.2 ml of the aminonaphthol sulfonic acid reagent. The tubes are buzzed to insure mixing. After allowing 45 min for color development, the absorbance of the samples is measured against an appropriate blank at 660 nm wavelength in 10-mm light path with a spectrophotometer having the necessary microattachments. Standard curves with 0.5, 1.0 and 2.0/zg of P are included in each run. The blank values are very constant usually amounting to 0.05 absorbance units when read against air. Neither blanks nor standards have to be run routinely through the digestion. The accuracy of the method depends on the size of the sample. In the middle of the range it approaches ± 1 ~. The aminonaphthol sulfonic acid used is reagent grade, recrystallized according to Fiske and SubbaRow [18]. 16 mg is dissolved in 25 ml of 15 °/o sodium bisulfite 0.5 ~ sodium sulfite solution. This stock solution keeps for at least two weeks in the dark in a cool place. 1 ml diluted to 100 ml with water constitutes the working solution which must be prepared daily. Alundum chips (No. 14, Norton Co. or similar) are boiled in sulfuric acid containing 2 0 ~ nitric acid, cooled, rinsed with distilled water and dried. They are kept in a glass stoppered flask.
414
Amino acids Amino acids were determined by standard procedures on a Beckman Model 120 amino acid analyzer after 6 M HC1 hydrolysis in vacuo at 110 ' C for 24 h. No correction was made for hydrolysis losses.
Fatty acids To the sample in a test-tube was added an inner standard in 1 ml solution in methanol and 2 ml methanol/5 ~,; HCI; the tube was cooled in an ice bath, s.ealed and heated to 100 °C for 16 h. The tube was then cooled, opened and the methanolizate filtered through glass wool in a funnel or through a small glass sintered filter, and the filtrate and washings combined in a 50-ml flask and evaporated almo~x to dryness. The concentrate was extracted with successive portions of hexane which were combined in a test-tube and washed with water; the washed hexane was evaporated to dryness, and the residue transferred to a small tube with hexane, the hexane removed again by evaporation, and the residue dissolved in 20 IH of toluene. The fatty acids were determined on portions of the toluene solution, usually 4 /~1, by gas-liquid chromatography at 190 °C on an 8 feet × 1/4 inch column packed with 10~o LAO 728 (DEGS) on Diatoport S using an F and M 609 chromatograph with hydrogen flame detector and equipped with a Disc chart integrator (Disc Instrument Co.). Quantitation was done by comparison with the known added amounts of internal standards. The crucial step in this procedure is the filtration of the methanolizate through glass wool or a sintered glass filter. This is because in the course of concentrating the methanolizate to dryness, the presence of protein may result in the capture of methyl esters by the protein in a form that withstands extraction by hexane. The extent of this capture appears to vary from run to run, and it is not the same for the different methyl esters in the methanolizate. The result is the invalidation of comparison with inner standards. Filtering the methanolizate, even when no protein residue appears to be present, eliminates this source of error. Proof of the foregoing was obtained in a series of experiments aimed at determining the reason why we were unable to obtain identical results from parallel analyses of one single preparation. The systematic study of the various steps in the procedure traced the source of variation to the concentration of the methanolizate. In the scrutiny of this step we found that filtration of the methanolizate eliminated the variations among the results. Without filtration, the variations in the results reappeared. Finally, the protein residue, remaining after the hexane extraction of the concentrate of an unfiltered methanolizate, contained fatty acids that could be recovered by extraction with ethyl ether.
Preparation of the proteolipid apoprotein The procedure followed is an adaptation of the one used for the preparation of proteolipid apoprotein from central nervous tissue [2]. It consists in extracting the tissue in a biphasic chloroform/methanol/water mixture. The lower phase of the homogenate corresponds to the washed extract of the classical method [19]. This lower phase is then submitted to dialysis in succession against neutral chloroform/ methanol to remove neutral free lipids, against chloroform/methanol/0.04 M HC1, to remove the acidic lipids combined electrostatically with the protein, and finally against
415 enough changes of neutral chloroform/methanol to eliminate the acid. The final retentate is a solution of proteolipid apoprotein free of complex lipids.
Extraction of tissue Bovine hearts, liver or kidney were obtained from the slaughterhouse (Granite State Packing Co., Manchester, N.H.) packed in ice and transported to the laboratory. The organs were freed of extraneous tissue by dissection and either processed immediately or cut into pieces of the desired size which were packed in aluminum foil and stored at --50 °C until use. To process the frozen material, the samples were placed at room temperature until the aluminum packing could be removed. The still partially frozen tissue was cut into 1-2-g pieces and extracted in the blender. Proteolipids from both neural and non-neural tissues were prepared by a modification of the classical chloroform/methanol extraction procedure [19]. In this modification one volume of tissue is homogenized with 5 volumes of chloroform/ methanol (1 :I, v/v) and 0.5 volumes of aqueous 2 M KC1. The result is a biphasic system which by centrifugation is resolved into two phases with a pad of insoluble material floating at the interface. This procedure can be carried out at any otherwise feasible scale provided that the ratio of tissue to extractants is maintained. As carried out routinely in this work, 110 g of tissue was homogenized in a Waring blendor for 2 min with 550 ml of chloroform/methanol (1:1, v/v) and 55 ml 2 M aqueous KC1. The resulting homogenate was placed in four 250-ml centrifuge tubes and centrifuged for 15 min at 900 x g (2000 rpm in an International Centrifuge No. 2). Of the two phases obtained the upper phase represented about 43 ~o and the lower phase about 57 ~ of the total liquid volume. The lower phase which corresponded to the washed extract in the classical method contained all the proteolipids. It was collected by siphoning, and filtered through a Whatman No. 24 filter to prevent accidental contamination by the interfacial material. 260 to 290 ml of a clear filtrate was then obtained. When compared to the classical chloroform/methanol extraction procedure [19] the present modification has the advantage of using only about 1/4 as much solvent mixture and of combining the extraction of the tissue and the washing of the extract into a single operation.
Separation of proteolipid apoprotein Dialysis of the washed tissue extracts was carried out as described by Lees et al. [20] with minor modifications. Cellulose tubing was cut into pieces of desired length which were rinsed in water, tied at one end, tested for leaks and shaken in the biphasic mixture chloroform/methanol/water (8:4:3, v/v/v) in a glass stoppered round bottom flask for 2 h, each time with 2 changes of solvent mixture. Alternatively, the water-rinsed pieces were let stand in contact with the washing mixture for several days with occasional shaking. After either treatment, the cellulose tubing did not yield to the solvents used in this study any detectable solutes nor any ultraviolet light absorbing material. The washed pieces were kept submerged in the washing mixture until used. In this study, 50 cm lengths of dialysis tubing of size 25 mm fiat were washed.
416 Each length was filled with 50-60 ml of extract, tied closed, and the resulting dialysis bag dialyzed against about 10-fold its volume of chloroform/methanol (2:1, v/v). In one routine run 240 ml of lower phase containing a tot~l of 3 g solutes was distributed among four dialysis bags, and the four bags dialyzed in a 2-1 glass cylinder against 2 1 i t e r s o f c h l o r o f o r m / m e t h a n o l ( 2 : l , v / v ) a t 4 8 C in the dark with magnetic stirring. After 5 days, 2 g (67'~,i of total starting solutes) of solutes were present in the outer liquid. The outer solvent was replaced. After 5 days, 288 mg of solutes had diffused out (9.7 ~; of starting total solutes). The outer liquid was replaced by its own volume of chloroform/methanol/conc. HCI (200:[00:1, v/v/v) and the dialysis was continued in the dark at 10 'C. After 7 days, the outer liquid contained 124 nag (5'Ii, starting solutes) of total solutes. It was replaced by chloroform/methanol (2:1, v.v). Alter 7 days of dialysis in the dark at 10 'C, only 12 nag had dialyzed out. The outer fluid was then replaced four times in succession at two-day intervals. The amounts of solids in the successive dialysates were respectively 4 mg, 3 mg, 2.4 mg and 2 nag. The retentate was collected. It amounted to 208 ml of solution containing 1.50 mg of solids per ml, i.e. a total of 312 nag solids. The sum total of solids in the successive dialysates and in the tinal retentate amounted to 2749 mg leaving 251 rag, i.e. 8,,I,, to be accounted for as losses. Proteolipid apoprotein could be recovered from the retentate by evaporation of the solvent, either in a stream of nitrogen, or in vacuo. The dry residue had a glass-like appearance. It was completely soluble in chloroform: methanol mixtures and insoluble in water.
Preparation qf the water soluble con[ormer o[ proteolipid apoprotein Proteolipid apoprotein recovered from the tinal retentate fronl the dialysis is insoluble in water. To render it ~:oluble it is necessary to replace the chloroform and methanol in its solution by water. This is done by the procedure described by Folch-Pi and Sherman [5, 6, 2] m which the solvents are removed by means of a stream of nitrogen, with concomitant addition of water to the point of immiscibility. The procedure can be run in any number of ways, in a test-tube, in a shallow glass vessel, or in an Erlenmeyer. In this project, a routine run was as follows: a pre-weighed 15-ml test-tube was placed in a beaker containing water at room temperature. The water in the beaker acted as a heat reserw)ir that avoided excessive cooling of the retentate because of the evaporation of the solvents. Therefore, the temperature of the water in the beaker was kept fairly constant by replacing the water that had cooled off. l0 ml of retentate were measured into the tube and a stream of nitrogen was blown from a Pasteur pipette onto the surface of the tube content. In 15 minutes the \olume of retentate was reduced to 2.2 tnl. A second portion of l0 ml of retentate was added into the tube, and the stream of nitrogen continued. In another 15 minutes the volume of the tube contents had been reduced to 3 ml. The stream of nitrogen was cut to a very small rate, the tip of the pipette immersed in the liquid and water added dropwise until after adding 1.7 ml turbidity developed, showing that the point of immiscibility had been reached. The tip of the Pasteur pipette was withdrawn from the liquid, and the rate of delivery of nitrogen was increased. The turbidity faded rapidly. 0.3 ml water was now added along the tube wall, and the tube shaken to insure mixing. Turbidity reappeared. It faded fast. The stream of nitrogen was continued and after another 15 minutes the volume of liquid in the tube was only 1.3 ml. One ml of water
417 was added into the tube, mixing being insured by proper maneuvering of the Pasteur pipette. In 20 minutes, the volume was again reduced to 1.3 ml. Another 1 ml of water was added, and mixed with the contents. In 15 minutes the tube contents had again been reduced to 1.3 ml. A final ml of water was added and the tube stoppered and stored at 4 °C. The solution weighed 2.67 g and contained 27.6 mg of proteolipid apoprotein, i.e., a concentration of 1.04 ~. Portions of the solution were dried in a vacuum desiccator. The residue obtained had a glass-like appearance. It was soluble in water and in chloroform/methanol. The whole procedure took 100 minutes because it was desired to concentrate the retentate so as to prepare a 1 ~ aqueous solution of proteolipid apoprotein. Without this requirement, the process would have taken less than 1 hour. RESULTS
Preparation of proteolipid apoproteins from heart, kidney and liver bovine tissues The procedure of preparation followed had been developed for the central nervous system [2]. It has proved to be applicable to heart, kidney and liver tissues without modification. The extraction of the tissue runs exactly as for neural tissues. Dialysis, if anything, runs more smoothly with the non-neural tissues than with neural tissues, in the sense that the dialysis in acid can be carried out at --10 °C, thus minimizing any effect of the acid. With neural tissues, this step is handicapped by the formation of precipitates inside the dialysis bags, an event that requires moving the dialysis to a higher temperature at which the precipitate will redissolve. This lengthens the procedure. Finally, the preparation of water-soluble proteolipid apoprotein from non-neural tissues is undistinguishable from the process applied to neural tissues. The yields per gram fresh tissue have been approximately: heart, 3 mg; kidney, 2 mg; liver, 1.6 mg. A point deserving emphasis is that with non-neural tissues there is the same diffusion of protein material during the first period of dialysis, that had been observed with neural extracts [20, 21]. That dialyzable protein material had been concentrated and found to be identical in chemical composition to the undialyzable protein [21]. The partial dialyzability of proteolipid apoprotein is compatible with the presence of fractions of proteolipid apoprotein with molecular weights below 12 000 (see below).
Properties of non-neural proteolipid apoproteins The proteolipid apoproteins from heart, kidney and liver removed from the retentate by evaporation of the solvents have a characteristic glass-like appearance. They are freely soluble in chloroform/methanol mixtures, and completely insoluble in water. The proteolipid apoproteins recovered from the water solutions have the same appearance. They are soluble in water and also in chloroform/methanol mixtures, although in the latter case it may be necessary to add 1 or 2 ~ water to the solvent mixture. Solutions of proteolipid apoproteins in organic solvents or in water are usually very clear and they exhibit clean absorbance spectra in the near ultraviolet range, with a maxima at 278 nm and a trough at 255. ~-ay~ ~ l c m (278 nm) = 13.5 ~ 0.15 with a E278 nm/E255 nm ratio of 1.95 to 2.0. These are the exact absorbance values given by neural proteolipid apoproteins.
418
Chemical composition The proteolipid apoproteins studied show no lipid by thin-layer c h r o m a t o graphy. Their P c o n t e n t is from 0.05 to 0.15 }/o (Table II) indicating minimal a i n o u n t s of phospholipids, if any, and their total c a r b o h y d r a t e by the orcinol method [22] a m o v n t s to 0.1 ~o or less. Briefly, the n o n - n e u r a l proteolipid apoproteins appear to be protein material essentially devoid of lipids or carbohydrates. TABLE I AMINO ACID COMPOSITION OF PROTEOLIPID APOPROTEIN FROM DIFFERENT TISSUES Mol per 100 tool recovered in hydrolysate Tissue
Central whitc matter
PLA prepn.* Arginine Histidine Lysine Aspartic Glutamie Half-cystine Methionine Serine Threonine Proline Glycine Alanine Valine Leucine lsoleucine Tyrosine Phenylalanine
2.6 1.8 3.8 4.0 5.8 4.0 1.9 8.5 8.5 2.8 10.3 12.5 6.9 11.1 4.9 4.6 7.9
Central gray matter
Heart
71-VII 72-V1
72-111
72-VI1
72-V
72-1X
72-1V
2.41 1.91 3.31 6.29 5.22
326 1.61 3.26 6.78 5.69 0.36 4.65 6.29 6.67 4.96 9.11 10.28 6.11 13.09 6.60 3.67 7.15
3.34 1.73 3.90 6.12 6.03 0.27 3.90 6.03 5.85 5.48 9.09 10.02 6.87 14.20 6 40 3.53 7.24
3. t 5 1.48 3.08 6.57 5.70 0.80 3.68 6.7(, (~.30 5.43 9.44 10.38 6.69 13.60 6.43 3.47 7.03
2.99 1.82 3.80 5.98 5 69
2.39 1.65 3.86 4.78 6.07 2.39 2.39 5.61 7.54 3.86 10.58 12.06 7.27 I 1.60 5.71 4.69 7.54
2.62 1.90 3.70 4.84 5.76 1.74 2.78 5.70 7.36 4.52 10.23 12.04 7.15 I 1.88 5.60 4.58 7.55
5.35 6.03 6.69 5.29 8.90 10,05 5.82 14.59 7.50 3.41 7.23
Liver
Kidney
3.94 (~.21 5.98 5.25 9.71 10.80 ¢~.58 14.09 6.86 3.43 6.86
* PLA, proteolipid apoprotein. The a m i n o acid c o m p o s i t i o n is given in Table 1, which also contains the a m i n o acid c o m p o s i t i o n of neural proteolipid apoproteins for comparison. In c o m m o n with n e u r a l proteolipid apoproteins, heart, kidney and liver proteolipid apoproteins show a high p r o p o r t i o n of n o n - p o l a r a m i n o acids, with similar polarity indices (see Table V); a relative a b u n d a n c e of sulfur a m i n o acids; a low c o n c e n t r a t i o n of acidic and basic a m i n o acids, and the same a b s o r b a n c e value in the ultraviolet, indicating similar c o n c e n t r a t i o n s of t r y p t o p h a n . They are different from neural proteolipid apoprotein in lacking half-cystine, with m e t h i o n i n e apparently replacing cystine as major sulfur a m i n o acid. There are other less p r o n o u n c e d but yet real differences: n o n - n e u r a l proteolipid apoproteins have higher c o n c e n t r a t i o n s of alanine, glycine and tyrosine and smaller c o n c e n t r a t i o n s of aspartic acid, proline, leucine and isoleucine than are f o u n d in n e u r a l proteolipid apoproteins. We are indebted to Dr. R. M a r t e n s o n of the N I M H for a n analysis for methyl arginine, on two samples of heart proteolipid apoprotein. N o m e t h y l arginine was found. I n analyses of end-groups on the n o n - n e u r a l
419 proteolipid apoproteins Whikehart and Lees [23] f o u n d aspartic acid as only Nterminal, and lysine as only C-terminal amino acids. The most important chemical difference between the two groups of proteolipid apoproteins is f o u n d in the fatty acids covalently b o u n d (Table II). It can be seen that heart and kidney proteolipid TABLE II FATTY ACIDS COMBINED IN PROTEOLIPID APOPROTEIN FROM DIFFERENT TISSUES All the used preparations contained about 0.1 ~ carbohydrate, as galactose. As galactolipids, this would account for about 0.15 70 fatty acids but this is unlikely because of the absence of fatty acids 20:0 or longer. Source and preparation number of proteolipid apoproteins White Matter average Gray Matter 71-VII 72-V1 Heart* 72-III 72-VII Kidney* 72-1V Liver 72-V 72-IX
Total fatty acids as ~ of preparation (a)
Composition of fatty acid mixture as of values in Column a 14:0 16:0
16:1 18:0 18:1 20:0
2.0-4.0
--
-
3.88 3.30
15.1 49.2 4.4 54.0
60.0
3.1 5.3
9.5
25.5 n.d.
14.1 18.5 n.d. 11.1 2 5 . 2 n.d.
P in preparation
0.01-0.04 0.077 0.042
0.40 0.52
0.033 0.042
0.56
0.017
2.45 2.17
6.7 6.6
26.2 32.9 5.9
37.3 26.6 n.d. 20.2 24.6 n.d.
0.015 --
* The amounts of fatty acids present in heart and in kidney proteolipid apoprotein can readily be attributed to phospholipids, since in a diacylphosphoglyceride, fatty acids amount to 18-fold the concentration of P, on a weight basis. apoproteins contain only 0.5 ~ or less of fatty acids, an a m o u n t that can be accounted for as phospholipids or galactolipids. By comparison, only one-fourth o f the 2 to 4 combined fatty acids f o u n d in neural proteolipid apoproteins can be accounted for as complex lipids [24]. Liver proteolipid apoprotein m a y be different from those from heart and kidney, but this point will require further work before a final conclusion can be arrived at.
Reversible conformational changes of proteolipM apoproteins Proteolipid apoproteins f r o m heart, kidney and liver were identical to neural proteolipid apoproteins in their ability to change c o n f o r m a t i o n reversibly with changes in polarity o f the medium. Both the type o f conformational change and its reversibility is shown in Table Ili. It can be seen that all proteolipid apoproteins studied offer a relatively high content o f a-helix in chloroform/methanol (2:1, v/v), a low content in solution in water, and the return to the original a-helix content u p o n simple dilution o f the aqueous solution with chloroform/methanol. It is likely that this change in c o n f o r m a t i o n is the basis o f the dual solubility o f proteolipid apoprotein in c h l o r o f o r m / m e t h a n o l and in water.
420 T A B L E Ill C H A N G E S IN ~ - H E L I X C O N T E N T O F P R O T E O L I P I D A P O P R O T E I N F R O M D I F F E R E N T BOVINE TISSUES ACCORDING TO THE POLARITY OF THE SOLVENT MEDIUM: REVERSIBILITY OF SUCH CHANGES ORD/CD Experiment No.
Source of proteolipid apoprotein
116 131 129 130 127
W h i t e matter 71-VI-2 G r a y matter Heart Kidney Liver
~-Helix content of proteolipid apoprotein (%) In C M 2:1 Ioriginal dialysis retentate~*
In water (prepared front the retentate)**
66 58 68 58 61
37 37 26 22 22
In ( ' M (by dilution of the a q u e o u s solutionl* 66 (80) ~'~t) 54
* F r o m O R D m e a s u r e m e n t s Ira'] 233 nm. C M , c h l o r o f o r m m e t h a n o l mixture. ** F r o m C D m e a s u r e m e n t s [0] 208 nm.
4¸¸¸¸¸¸
i!!i~
'),
?
i:i
H
K
:ii(!i~!ii¸¸¸¸~)i
LI/¸!
Fig. 1. Polyacrylamide gel electrophoresis of proteolipid apoproteins from heart (H), kidney (K), a n d liver (L) in 0 . 1 % s o d i u m dodecyl sulfate. T h e n u m b e r s s h o w n are the calculated molecular weights in t h o u s a n d s . Experimental details in text.
421
Polyaerylamide gel electrophoresis in presence of O.l% sodium dodeeyl sulfate Proteolipid apoproteins precipitated from aqueous solution upon addition of dilute buffers or 6 M guanidine hydrochloride. When aqueous solutions of the apoproteins were added to medium A, a transient precipitate appeared which readily dissolved upon brief stirring. These solutions were subjected to polyacrylamide gel electrophoresis in 0.1 ~ sodium dodecyl sulfate. The results of a typical experiment are shown in Fig. 1. The heart proteolipid apoprotein yielded a major diffuse band with an apparent molecular weight of 34 000. A less intense band, which also corresponded to 34 000, was observed for kidney and liver proteolipid apoproteins. With all samples, a broad diffuse band centered near 12 000 was seen. A considerable portion of the material did not penetrate the gel, thus suggesting that in the presence of sodium dodecyl sulfate, only partial disaggregation had occurred.
Molecular weight determination in sodium dodecyl sulfate-urea gels Non-neural proteolipid apoproteins were submitted to polyacrylamide gel electrophoresis in 8 M urea and 0.1 ~ sodium dodecyl sulfate. The results are shown in Fig. 2. Compared to the pattern obtained with sodium dodecyl sulfate alone there
3o--i b 12
H
K
L
Fig. 2. Polyacrylamide gel electrophoresis of proteolipid apoproteins from heart (H), kidney (K) and liver (L) in 0.1% sodium dodecyl sulfate/8 M urea. The numbers shown are the calculated molecular weights in thousands. D, the dye front (Bromophenol Blue). Experimental details in text.
422 was a striking increase in the number of bands seen with all samples. Furthermore, the bands were more sharply defined than those observed in 0.1 ~/,', sodium dodecyl sulfate alone. The gel patterns in sodium dodecyl sulfate/urea of heart, kidney and liver proteolipid apoproteins were remarkably similar, each displaying 8 to 10 apparently identical bands. The apparent molecular weight of each band was calculated from a calibration curve using standard proteins or cytochrome c peptides of known molecular weight (Fig. 3). They are given in Table IV. Each apoprotein contained a major band of 30 000, which will be referred to as the major non-neural apoprotein. Three identical bands corresponding to very low molecular weight ( < 5000) were also found in each preparation. To test the possibility of further disaggregation, some experiments were conducted with 8 M urea gels containing varying amounts of acrylamide (7.5 to 15°")~oand sodium dodecyl sulfate (0.1-1.0 ~o). No further changes in gel patterns for non-neural proteolipids were observed under these conditions. 60 50 4030 ~ l b u m i n 2o '~ 0 *< IO ~
T ~
hym°trypsinOgen
~e Cytochrome c e\\
Lysozyme
\
Cytochrome c T • k
t,
\
~u
ACTH
L
rE
j
Cytochrome c ]~ ~
i I !
o.o
o.5
RelativeMobility
io
Fig. 3. Calibration curve of polyacrylamide gel c]ectrophoresis in 0,1% sodium dodecyl sulfate/8 M urea. Cytochrome c I, c I1, c Ill, cyanogen bromide peptides from eytochrome c; I, linear least squares regression line for the observed relative mobilities of protein standards with molecular weight greater than 11 700 (cytochrome c); II, linear least squares regression line for protein standards less than 14 300 (lysozyme). DISCUSSION
The widespread distribution of proteolipids throughout animal and vegetal tissues made it imperative to scrutinize proteolipids from different sources, not only to gather new information, but to establish differences and similarities of proteolipids
423 TABLE IV MOLECULAR WEIGHT (±S.D.) OF NON-NEURAL PROTEOLIPID APOPROTEINS The molecular weight is determined by polyacrylamide gel electrophoresis in 0.1 ~ sodium dodecyl sulfate, 8 M urea. The number of experiments is given in parentheses. The band numbers are assigned in descending order from the cathode, Band No. 1 2
3 4 5 4 5 6 7 8 9 10
Heart (6) 35 000 29 100 (5:1000) 27 000 -20 400 16 800 -12 300 :k 800 4 600 :d::800 3 500 5:300 2 400 ~ 300
Kidney (3)
Liver (3)
30 000 % 500 27 800 23 400
30 000 ± 600 28 000 -20 000 17 300 14 600 12 200 ± 1000 10 500 4 400 5:600 3 300 ± 700 2 300 i 650
15 500 13 000 :L 800 10 800 4 700:3_ 300 3 500 ± 500 2 300 ± 400
according to their source of origin. The results reported above have established that the proteolipids from bovine heart, liver and kidney show many similarities and a few differences from the proteolipids previously obtained from the gray and white matters of the central nervous system. Thus, all proteolipids are soluble in chloroform/methanol mixtures and extractable from the tissues with these mixtures. As originally obtained, they are completely insoluble in aqueous solutions. They retain these solubility characteristics throughout the process of gradual delipidation; when maximally delipidated the resulting proteolipid apoproteins still retain their solubility in chloroform/methanol and can now be rendered soluble in water by gradually replacing chloroform/methanol with water in the solvent medium [5, 6]. The dual solubilities correspond to different conformers and the change from one conformer to another is reversible and repeatable (Table lII) i.e. a sample of apoprotein can be changed from the chloroform/methanol-soluble lipophilic conformer to the watersoluble hydrophilic conformer and back to the lipophilic form. This process can be repeated several times without apparent change in either conformer. In chemical composition, the non-neural apoproteins show general similarities, and some specific differences from the neural proteolipids. The respective amino acid patterns have in common the predominance of hydrophobic residues; the relative scarcity of acidic and basic amino acids, an abundance of sulfur-containing amino acids, and of tryptophan. A calculation of the polarity index according to Vanderkooi and Capaldi [25] given in Table V shows that all proteolipid apoproteins have a low polarity index (31-32 ~). On the other hand, compared to the neural proteolipid apoproteins, the non-neural proteolipid apoproteins contain little or no half-cystine, a higher proportion of methionine, small concentrations of glycine, alanine and tyrosine and slightly higher aspartic acid, proline, leucine and isoleucine. All proteolipid apoproteins show the same absorbance at 278 nm, indicating little or no difference in their tryptophan content. The most dramatic difference, however, is the virtual
424 TABLE V POLARITY OF PROTEOLIPID APOPROTEINS FROM VARIOUS TISSUES [251 % Polar, mol percent of Asp, G lu, Lys, Arg. !'; Intermediate, tool percent of Ser, Thr, His, G ty, Tyr. % Polarity : % Polar • (% Intermediate 2}. Source
White matter Gray matter Heart Kidney Liver
Amino acid content o, Polar
"{; Intermediate 2
16.2
16.9
17.0 18. I 18.5 18.9
15.0 13.6 13.6 13.4
",, Polarity
33. I 32.0 31.7 32. I 32.3
absence of fatty acids from heart and kidney apoproteins, against the 2 4'~;, fatty acids covalently combined in central white and gray matters apoproteins. Of course, differences in chemical composition among physically heterogeneous mixtures of substances must be interpreted with great caution. In this work, we limit their significance to showing that the total proteolipid apoproteins from heart, kidney and liver are similar among themselves and different from the total proteolipid apoproteins obtained from the central nervous system. The sharp differences in halfcystine and fatty acid content point to differences in major components of the respective apoprotein mixtures. On the other hand, the smaller differences seen in tile rest of the amino acids might be explained assuming that the two sets of mixtures contain different proportions of common components. The successful preparation of the water soluble proteolipids has permitted their analyses by polyacrylamide gel electrophoresis under different conditions. In 0.1)o sodium dodecyl sulfate all samples showed (Fig. I) substantial amounts of presumably highly aggregated material as a band that had not penetrated the gel. The major part of the material that had entered the gel migrated as a single diffuse band with an apparent molecular weight of 34 000. A number of diffuse lower molecular weight bands were also observed. By contrast, when polyacrylamide gel electrophoresis in sodium dodecyl sulfate/urea was used, all proteolipid apoproteins gave a large number of well-defined bands. Up to 8 to 10 identical bands in the 2500 up to 28 000 range were seen (Fig. 2) with a major band at 30 000. The occurrence of variable proportions of material that did not penetrate the gel, and the increase in the number of bands shown when urea is added to sodium dodecyl sulfate suggests that we are dealing with different states of aggregation. This possibility was advanced by Folch-Pi and Stoffyn [2] on the basis of the partial dialyzability of white matter proteolipids, of the computed minimal molecular size of either 6000 or 12 000, and from the simple numerical relationships of some of the major bands observed in polyacrylamide gel electrophoresis under different conditions of operation. In a more specific suggestion Chan and Lees [19] in their study of crude white matter proteolipids suggested that in polyacrylamide gel electrophoresis in sodium dodecyl sulfate and 8 M urea, myelin proteolipid apoprotein migrates as a series of oligomers of a
425 5000 daltons monomer. These possibilities will require further study on the properties of proteolipid apoproteins. The study of the chemical nature of the major band of heart proteolipid apoproteins (mol. wt. 30 000) is currently under investigation. If we consider the possible implications of some of the chemical properties, we obviously conclude that the marked hydrophobic character of all proteolipid apoproteins is basic to their solubility in some organic solvents; it also requires that the water-soluble conformer must have a very special tridimensional structure which will permit maximal exposure of the scarce hydrophilic residues of the proteolipid apoprotein peptide chain. The presence ofcovalently bound fatty acids, clearly increases the hydrophobicity of the proteolipid apoprotein structure; on the other hand the absence of the acyl radicals in heart and kidney proteolipid apoproteins suggests that the constituent fatty acids must subserve a role particular to the central nervous system, and not to proteolipid apoproteins in general. The presence of a high proportion of half-cystine in neural proteolipid apoprotein in contrast with its essential absence and replacement by methionine in non-neural proteolipid apoproteins suggests that the reduction or oxidation of sulfhydryl groups is not a determinant factor in the solubilities of proteolipid apoproteins and on their structural or functional role in tissues. This suggestion is reinforced by the absence of readily demonstrable -SH groups, and the great stability of -S-S-bridges in neural proteolipid apoprotein reported by Lees et al. [26]. Studies in a number of laboratories have shown that some neural and nonneural proteolipids are concentrated in the mitochondrial subfraction [8, 9, 27]. Recent reports indicate that some or all of the products of mitochondrial protein biosynthesis are proteolipids [28-33]. The molecular weight of the principal product, varies from 2000 [29] to 40 000 [31] and depends upon experimental conditions. The results of our study on proteolipid apoproteins from whole tissue suggest the presence of both low and high molecular weight proteins. Swank et al. [34] in a study of the distribution of proteins in a number of mitochondrial preparations demonstrated in the sodium dodecyl sulfate urea gel system a series of oligopeptides with a molecular weight range of 2500 to 10 000. Our electrophoretic results with non-neural proteolipids under identical conditions are remarkably similar to these observations. The relationships, however, between those oligopeptides and non-neural proteolipids remain to be shown. Recently, however, Tzagoloff et al. [35] has identified a subunit of yeast mitochondrial ATPase as a proteolipid. Finally, Capaldi et al. [36] has isolated a major hydrophobic protein (mol. wt. 29 000) from the inner membrane of bovine heart mitochondria. This protein, which is similar in molecular size to our major heart apoprotein, appears to be a component of the oxidative phosphorylation system [37]. The occurrence of two different conformers of preoteolipid apoprotein in vitro and their ready interconvertibility raises the question of the condition of proteolipid apoprotein in the living tissue. We know that one conformer is hydrophobic and the other hydrophilic. The presence of lipids appears to "lock" proteolipid apoprotein in the hydrophobic conformation. Otherwise the two conformers are freely interconvertible, the polarity of the medium being the factor that determines the conformation adopted by proteolipid apoprotein. The exact question is whether in tissue, proteolipid apoprotein occurs only as one conformer, or as two, and in the
426 latter case, whether the two conformers present undergo in vivo tim same changes that they show in vitro. On superficial observation, the d i s t r i b u t i o n of proteolipid apoprotein a m o n g tissue structures rich in lipids, its extractability with c l H o r o f o r m methanol (accompanied by lipids) and inextractability with aqueous solutions suggest that in the living tissue proteolipid a p o p r o t e i n is present as the hydrophobic conformer. This evidence, however, can be grossly misleading because a "native" hydrophilic proteolipid a p o p r o t e i n could easily be extracted as the h y d r o p h o b i c conformer because of the exposure to c h l o r o f o r m / m e t h a n o l in the course of the extraction. A~ for proteolipid a p o p r o t e i n being locked in a h y d r o p h o b i c c o n f o r m a t i o n by the presence of lipids, it is not difficult to imagine conditions in the tissue which would effectively shield proteolipid apoprotein from contact with lipids. The possibility that proteolipid a p o p r o t e i n be present in tissue as both conformers makes it likely that proteolipid a p o p r o t e i n might undergo c o n f o r m a t i o n a l changes in the m e m b r a n e in vivo. If this be the case, n u m e r o u s possibilities arise for participation of proteolipid a p o p r o t e i n in m e m b r a n e f u n c t i o n : thus, b i n d i n g of specific ligands could effect changes in polarity at the macromolecular level and such changes would u n d o u b t e d l y affect greatly the properties of the microareas in which they had taken place. With this type of consideration we step into a field of pure speculation which might possibly be useful for the f o r m u l a t i o n of working hypotheses and for the definitions of frames of reference of probable alternatives. We feel, hm~ever, that at the present time this type of speculation is idle and outside lhe purpose of this paper. AC K NOWLEDG EM ENTS This work was s u p p o r t e d by grants NS 00130 and FR 05484 from the National Institutes of Health, U.S. Public Health Service. REFERENCES 1 2 3 4 5
Folch, J. and Lees, M. (1951) J. Biol. Chem, 191, 807 817 Folch-Pi, J. and Stoffyn, P. J. (1972) Ann. N.Y. Acad. Sci. 195, 86 107 Tenenbaum, D. and Folch-Pi, J. (1966) Biochim. Biophys. Acta 115, 14t 147 Zand, R. (1968) Biopolymers 6, 939 953 Folch-Pi, J. and Shernlan, G. (1969) Abstr. 2nd hat. Meeting Int. Soc. Neurochcm., p. 169. Tamburini, Milano 6 Sherman, G. and Folch-Pi, J. (1970) J. Neurochem. 17, 597 605 7 Joel, C. D., Karnovsky, M. L., Ball, E. G. and Cooper, O. (1958) J. Biol. Chem. 233. 1565 1573 8 Murakami, M., Sekine, H. and Funahashi, S. (1962) J. Biochem. 51, 431 435 9 Murakami, M., Ozawa, Y. and Funahashi, S. (1963) J. Biochem. 54, 166 172 10 Wolfgram, F. (1967) Biochim. Biophys. Acta 147, 383-385 I1 Eichberg, J. (1969) Biochim. Biophys. Acta 187, 533 545 12 Folch-Pi, J. and Sakura, J. D. (1973) Fed. Proc. 32, 624 13 Chu, R. C. L. and Yasunobu, K. T. (1964) Biochim. Biophys. Acta 89, 148 150 14 Weber, K. and Osborn, M. (1969) J. Biol. Chem. 244, 4406 4412 15 Agrawal, H. C., Burton, R. M., Fishman, M. A., Mitchell, R. F. anti Prensky, A. L. (1972) ,I. Neurochem. 19, 2083 2089 16 Swank, R. T. and Munkres, K. D. (1971) Anal. Biochem. 39, 462 477 17 Sperry, W. M. (1942) Ind. Eng. Chem., Anal. Ed. 14, 88 90
427 18 19 20 21
Fiske, C. H. and SubbaRow, Y. (1925) J. Biol. Chem. 66, 375-400 Folch, J., Lees, M. and Sloane Stanley, G. H. (1957) J. Biol. Chem. 226, 497-509 Lees, M. B., Carr, S. and Folch, J. (1964) Biochim. Biophys. Acta 84, 464-466 Folch-Pi, J. (1971) 3rd Int. Meet. Int. Soc. Neurochem. Abstr. 239, Publ. Akademiai Kiado, Budapest, Hungary 22 Sorenson, M. and Haugaard, G. (1933) Biochem. Z. 260, 247-277 23 Whikehart, D. R. and Lees, M. B. (1971) Fed. Proc. 30, 1247 (Abstr.) 24 Stoffyn, P. and Folch-Pi, J. (1971) Biochem. Biophys. Res. Commun. 44, 147-161 25 Vanderkooi, G. and Capaldi, R. A. (1972) Ann. N.Y. Acad. Sci. 195, 135-138 26 Lees, M. B., Leston, J. A. and Marfey, P. (1969) J. Neurochem. 16, 1025-1032 27 Lees, M. B. (1966) J. Neurochem. 13, 1407-1420 28 Kadenbach, B. and Hadvary, P. (1973) Eur. J. Biochem. 32, 343-349 29 Kadenbach, B. (1971) Biochern. Biophys. Res. Commun. 44, 724-730 30 Tzagoloff, A. and Akai, A. (1972) J. Biol. Chem. 247, 6517-6523 31 Burke, J. P. and Beattie, D. S. (1973) Biochem. Biophys. Res. Commun. 51, 349-356 32 Rowe, M. J., Lansman, R. and Woodward, D. O. (1973) Abstr. Fed, Proc. 2407 33 Kuzela, S., Kolarov, J. and Krempasky, V. (1973) Biochem. Biophys. Res. Commun. 54, 9-16 34 Swank, R. T., Sheir, G. 1. and Munkres, K. D. (1971) Biochemistry 10, 3931-3939 35 Tzagoloff, A., Rubin, M. S. and Sierra, M. F. (1973) Biochim. Biophys. Acta 301, 71-104 36 Capaldi, R. A., Komai, H. and Hunter, D. R. (1973) Biochem. Biophys. Res. Commun. 55, 655659 37 Racker, E. and Kandrach, A. (1971) J. Biol. Chem. 246, 7069-7071
(c) Elsevier Scientific Publishing Company, Amsterdam--- Printed in The Netherlands BBA 37312 P R E P A R A T I O N OF T H E P R O T E O L I P I D APOPROTE1N HEART, LIVER AND KIDNEY
FROM
BOVINE
J. FOLCH-PI* and J. D. SAKURA Biological Research Laboratory, McLean Hoapital, Belmont, Mass. 02178, and Department ol"Biological Chemistry, Harvard Medical School, Boston, Mass. 02115 (U.S.A.)
(Received September 2nd, 1975)
SUMMARY Proteolipid apoproteins have been prepared from heart, kidney, and liver by dialysis in chloroform/methanol against chloroform/methanol, acidified chloroform/ methanol, and chloroform/methanol in succession. They are free of lipids ( - 0.05 71, P; < 0.1~o carbohydrate). They show a high content of non-polar amino acids, methionine, and tryptophan and contain little or no half-cystine. They differ from neural proteolipid apoproteins by absence of half-cystine, and of covalently bound fatty acids. As recovered from chloroform/methanol solutions, they are soluble in chloroform/methanol and insoluble in water, but a water-soluble form can be prepared by changing the solvent from chloroform/methanol to water in a stream of nitrogen. The chloroform/methanol-soluble form and the water-soluble form tire interconvertible. O R D and CD spectra of all proteolipid apoproteins indicate 60-70 !'0
INTRODUCTION The proteolipids are a group of protein-lipid complexes characterized by solubility in chloroform/methanol mixtures, and insolubility in water and aqueous solutions. They were first recognized in 1951 by Folch and Lees [1] who found them in chloroform/methanol extracts of bovine and human central nervous systems. Abbreviation: TEMED, N,N,N',N'-tetramethylethylene diamine. " Correspondence should be addressed to : Dr. Jordi Folch-Pi, Biological Research Laboratory, McLean Hospital, 115 Mill Street, Belmont, Mass. 02178, U.S.A.
411 Subsequent work in this and other laboratories found them to be specially abundant in mammalian central white matter, followed by gray matter, heart muscle, kidney and liver in order of decreasing concentration and to be otherwise widely distributed among animal and vegetable tissues usually in relation to membranous structures (see ref. 2 for a review). Most of the later work on proteolipids was carried out on mammalian central nervous white matter where they were soon traced mainly to the myelin sheath. The white matter proteolipids were partially separated from l ipids by a number of different procedures [2]. In 1966 Tenenbaum and Folch-Pi [3] obtained them free of lipids detectable by thin-layer chromatography by dialysis in succession against neutral chloroform/methanol, chloroform/methanol acidified with HC1 to 0.04 M concentration, and a succession of mixtures in which chloroform/methanol was replaced gradually by water until pure water was reached. The final result was an aqueous solution of proteolipid apoprotein. The apoprotein recovered by evaporation of water was still soluble in chloroform/methanol. The solution was otherwise very unstable and after a few days, the apoprotein precipitated out of solution yielding a precipitate that was insoluble in all organic solvent mixtures and aqueous solvents that were used. Zand [4] by optical rotatory dispersion measurements, found that the starting chloroform/methanol-soluble proteolipid protein had a highly a-helical conformation and that, in the gradual shift from chloroform/methanol to water in the course of dialysis, this conformation changed to one without detectable a-helix. An improved procedure [5] yielded preparations of lipid-free proteolipid apoprotein which were freely soluble in chloroform/methanol and in water, forming solutions that remained stable indefinitely. Sherman and Folch-Pi [6] found that the two solubilities corresponded to two conformers, a highly a-helical hydrophobic conformer, and a hydrophilic conformer in which the a-helix content is reduced by about one-half. The two conformers are interconvertible, the hydropbilic one becoming hydrophobic on simple exposure to chloroform/methanol with a return to the higher a-helix content of the hydrophobic starting material. Little work has been carried out on proteolipids from non-neural sources. It was shown early [7] that most of the heart proteolipids were present in the mitochondria. Murakami et al. [8] prepared heart proteolipids by dialysis in organic solvents, and traced most of them to mitochondria [9]. Wolfgram [10] showed the heart proteolipids to be devoid of half-cystine, a finding later confirmed by Eichberg [I 1]. None of these studies succeeded in preparing heart proteolipid apoprotein devoid of detectable lipids. This paper describes the isolation of proteolipid apoproteins from bovine heart, kidney and liver tissues, and the study of some of their chemical and physical properties. A preliminary abstract of the results obtained has appeared [12]. MATERIALS Hen egg white lysozyme and a-chymotrypsin, were obtained from Worthington Biochemical Corp. (Freehold, N.J.). Adrenocorticotropic hormone, cytochrome c, type VI, sperm whale myoglobin, glyceraldehyde-3-phosphate dehydrogenase, glucagon and dithiothreitol were purchased from Sigma Chemical Co. (St. Louis, Mo.).
412 Ovalbumin was supplied in a calibration kit from Pharmacia Fine Chemicals (Piscataway, N.J.). Tris, urea, and sodium dodecyl sulfate (lot No. 3057 and No. 1737) were ultra pure quality from Schwartz-Mann (Orangeburg, N.Y.). Acrylamide, N.N'methylene bisacrylamide, ammonium persulfate, N,N,N',N'-tetramethylethylene diamine (TEMED) and Coomassie Brilliant Blue were obtained from Bio-Rad Laboratory (Richmond, Calif.). All other chemicals were reagent grade or better. Cyanogen bromide peptides from cytochrome c were prepared according to Chu and Yasunobu [13]. METHODS
Analytical polyaco'lamide gel electrophoresis Polyacrylamide gel electrophoresis was carried out in 0.1'), sodium dodecyl sulfate following an adaptation of the general procedure of Weber and Osborn [14] as modified by Agrawal et al. [15]. A stock buffer (pH 7,2) contained 1.95 g NaHzPO4" H20, 9.66 g NazHPO4.7H20, and 2 g sodium dodecyl sulfate per liter, A gel solution was prepared by dissolving 3.35 g acrylamide monomer, 90 mg N,N'-methylene bisacrylamide in water, the final volume adjusted to 15 ml, and diluted with 16.5 ml of stock buffer. Polymerization of the gel solution was initiated by adding 1.5 ml ammonium persulfate ,10.,,~...../ow/v) and 5 0 / d TEMED. The gel solution was poured into glass tubes (5 mm 120 ram) to a height of 90 mm. To obtain a flat gel surface, isobutyl alcohol was layered on top of the gels. The electrode buffer was prepared by dilution of the stock buffer with an equal volume of water. Polyacrylamide gel electrophoresis on 0.1 '?,;, sodium dodecyt sulfate and 8 M urea, pH 6,9 was carried out according to Swank and Munkres [16]. For 30 ml gel solution, 3.62 g acrylamide, 0.13 g N,N'-methylene bisacrylamide, 13.07 g urea, 30 mg sodium dodecyl sulfate, 205/~1 H3PO4, 420 mg Tris and 11 mg EDTA were dissolved in distilled water to a final volume of 28.5 ml. The pH of the solution was adjusted to 6.9 2 0.02 with solid Tris. After deaeraticn with a water aspirator, 1.5 ml ammonium sulfate solution (0.2 ~,;), and 20 t*1 T E M E D was added. The gels were cast into 150 mm glass tubing and overlayed with isobutyl alcohol. The electrode buffer was freshly prepared. It contained 0.1 ~'0 sodium dodecyl sulfate, 0.1 M H~PO4, and 2 mM EDTA. Its pH was adjusted to 6.9 ! 0.02 with solid Tris. 40 to 100 /¢g water-soluble proteolipid apoprotein in water was mixed with 50 to 100 pl of an aqueous soluticn of 100 mg sodium dodecyl sulfate, 150 mg dithiothreitol, and 800 mg sucro,'~e made in 10 ml (referred to as medium A [15]), and heated in boiling water for approximately I min. The same procedure was followed to apply 1-5 t~g of standard proteins. Electrophoresis m sodium dodecyl sulfate or sodium dodecyl sulfate/urea was conducted at I mA per tube for 1 h followed by an increase to 3 mA per tube for approximately 16 h. The gels were then removed Dom the glass tube, and a wire was inserted to mark the dye front. The gels were stained in a 0.25~; (w/v) Coomassie Brilliant Blue according to the procedure described by Agrawal et al. [15].
ORD and CD measurements Optical rotatory dispersion and circular dichroism measurements were carried
413 out with a Cary 60 spectropolarimeter equipped with a 6001 circular dichroism attachment. All measurements were made at room temperature (,~ 23 °C). Overlapping spectra were obtained using a 0.1 and a 1.0 cm cell. Reduced mean residue rotation [m'] and ellipticities [0] were calculated using a value of 108 for the mean residue weight.
Microestimation of total P The procedure followed is a microadaptation of the method of Sperry [17]. A sample containing 0.2 to 2.5 #g of P is weighed or measured into a 10 mm × 70 mm medium heavy glass tube. The sample may be dissolved in up to 0.5 ml of water. Any organic solvent present is removed by evaporation. To the sample is now added 0.04 ml of 60 ~ HC104 and one or two small chips of alundum. The digestion is conveniently run in sets of up to 15 tubes using a beaker narrow and low enough to hold the tubes at an angle so that they protrude about 1 cm above the beaker rim. The beaker is filled with concentrated HzSO 4 up to 1 cm below the rim, and a thermometer is immersed in it to monitor the temperature. The beaker is heated gradually to insure smooth boiling. If marked browning develops, heating is stopped and one or two small drops of 30 ~o Hz02 are added to each brown digest by delivering them against the inner wall of the tube, from a fine tipped pipette calibrated to deliver droplets of about 0.01 ml volume. Browning clears immediately or upon further heating. If it recurs, the addition of HzO2 is repeated. Heating is continued until white fumes of perchloric acid appear, and then for 15 rain longer. The temperature of the bath should then be about 200 °C. Temperatures much above 200 °C might result in losses of P. At the end of the digestion, the tubes are removed from the bath, and cooled and rinsed free of HzSO4 with running water. To each tube is added 0.15 ml of water and the tubes are cooled anew by placing in a water, or ice bath. To each tube is then added 0.1 ml 25 9/00ammonium molyl~date solution and 0.2 ml of the aminonaphthol sulfonic acid reagent. The tubes are buzzed to insure mixing. After allowing 45 min for color development, the absorbance of the samples is measured against an appropriate blank at 660 nm wavelength in 10-mm light path with a spectrophotometer having the necessary microattachments. Standard curves with 0.5, 1.0 and 2.0/zg of P are included in each run. The blank values are very constant usually amounting to 0.05 absorbance units when read against air. Neither blanks nor standards have to be run routinely through the digestion. The accuracy of the method depends on the size of the sample. In the middle of the range it approaches ± 1 ~. The aminonaphthol sulfonic acid used is reagent grade, recrystallized according to Fiske and SubbaRow [18]. 16 mg is dissolved in 25 ml of 15 °/o sodium bisulfite 0.5 ~ sodium sulfite solution. This stock solution keeps for at least two weeks in the dark in a cool place. 1 ml diluted to 100 ml with water constitutes the working solution which must be prepared daily. Alundum chips (No. 14, Norton Co. or similar) are boiled in sulfuric acid containing 2 0 ~ nitric acid, cooled, rinsed with distilled water and dried. They are kept in a glass stoppered flask.
414
Amino acids Amino acids were determined by standard procedures on a Beckman Model 120 amino acid analyzer after 6 M HC1 hydrolysis in vacuo at 110 ' C for 24 h. No correction was made for hydrolysis losses.
Fatty acids To the sample in a test-tube was added an inner standard in 1 ml solution in methanol and 2 ml methanol/5 ~,; HCI; the tube was cooled in an ice bath, s.ealed and heated to 100 °C for 16 h. The tube was then cooled, opened and the methanolizate filtered through glass wool in a funnel or through a small glass sintered filter, and the filtrate and washings combined in a 50-ml flask and evaporated almo~x to dryness. The concentrate was extracted with successive portions of hexane which were combined in a test-tube and washed with water; the washed hexane was evaporated to dryness, and the residue transferred to a small tube with hexane, the hexane removed again by evaporation, and the residue dissolved in 20 IH of toluene. The fatty acids were determined on portions of the toluene solution, usually 4 /~1, by gas-liquid chromatography at 190 °C on an 8 feet × 1/4 inch column packed with 10~o LAO 728 (DEGS) on Diatoport S using an F and M 609 chromatograph with hydrogen flame detector and equipped with a Disc chart integrator (Disc Instrument Co.). Quantitation was done by comparison with the known added amounts of internal standards. The crucial step in this procedure is the filtration of the methanolizate through glass wool or a sintered glass filter. This is because in the course of concentrating the methanolizate to dryness, the presence of protein may result in the capture of methyl esters by the protein in a form that withstands extraction by hexane. The extent of this capture appears to vary from run to run, and it is not the same for the different methyl esters in the methanolizate. The result is the invalidation of comparison with inner standards. Filtering the methanolizate, even when no protein residue appears to be present, eliminates this source of error. Proof of the foregoing was obtained in a series of experiments aimed at determining the reason why we were unable to obtain identical results from parallel analyses of one single preparation. The systematic study of the various steps in the procedure traced the source of variation to the concentration of the methanolizate. In the scrutiny of this step we found that filtration of the methanolizate eliminated the variations among the results. Without filtration, the variations in the results reappeared. Finally, the protein residue, remaining after the hexane extraction of the concentrate of an unfiltered methanolizate, contained fatty acids that could be recovered by extraction with ethyl ether.
Preparation of the proteolipid apoprotein The procedure followed is an adaptation of the one used for the preparation of proteolipid apoprotein from central nervous tissue [2]. It consists in extracting the tissue in a biphasic chloroform/methanol/water mixture. The lower phase of the homogenate corresponds to the washed extract of the classical method [19]. This lower phase is then submitted to dialysis in succession against neutral chloroform/ methanol to remove neutral free lipids, against chloroform/methanol/0.04 M HC1, to remove the acidic lipids combined electrostatically with the protein, and finally against
415 enough changes of neutral chloroform/methanol to eliminate the acid. The final retentate is a solution of proteolipid apoprotein free of complex lipids.
Extraction of tissue Bovine hearts, liver or kidney were obtained from the slaughterhouse (Granite State Packing Co., Manchester, N.H.) packed in ice and transported to the laboratory. The organs were freed of extraneous tissue by dissection and either processed immediately or cut into pieces of the desired size which were packed in aluminum foil and stored at --50 °C until use. To process the frozen material, the samples were placed at room temperature until the aluminum packing could be removed. The still partially frozen tissue was cut into 1-2-g pieces and extracted in the blender. Proteolipids from both neural and non-neural tissues were prepared by a modification of the classical chloroform/methanol extraction procedure [19]. In this modification one volume of tissue is homogenized with 5 volumes of chloroform/ methanol (1 :I, v/v) and 0.5 volumes of aqueous 2 M KC1. The result is a biphasic system which by centrifugation is resolved into two phases with a pad of insoluble material floating at the interface. This procedure can be carried out at any otherwise feasible scale provided that the ratio of tissue to extractants is maintained. As carried out routinely in this work, 110 g of tissue was homogenized in a Waring blendor for 2 min with 550 ml of chloroform/methanol (1:1, v/v) and 55 ml 2 M aqueous KC1. The resulting homogenate was placed in four 250-ml centrifuge tubes and centrifuged for 15 min at 900 x g (2000 rpm in an International Centrifuge No. 2). Of the two phases obtained the upper phase represented about 43 ~o and the lower phase about 57 ~ of the total liquid volume. The lower phase which corresponded to the washed extract in the classical method contained all the proteolipids. It was collected by siphoning, and filtered through a Whatman No. 24 filter to prevent accidental contamination by the interfacial material. 260 to 290 ml of a clear filtrate was then obtained. When compared to the classical chloroform/methanol extraction procedure [19] the present modification has the advantage of using only about 1/4 as much solvent mixture and of combining the extraction of the tissue and the washing of the extract into a single operation.
Separation of proteolipid apoprotein Dialysis of the washed tissue extracts was carried out as described by Lees et al. [20] with minor modifications. Cellulose tubing was cut into pieces of desired length which were rinsed in water, tied at one end, tested for leaks and shaken in the biphasic mixture chloroform/methanol/water (8:4:3, v/v/v) in a glass stoppered round bottom flask for 2 h, each time with 2 changes of solvent mixture. Alternatively, the water-rinsed pieces were let stand in contact with the washing mixture for several days with occasional shaking. After either treatment, the cellulose tubing did not yield to the solvents used in this study any detectable solutes nor any ultraviolet light absorbing material. The washed pieces were kept submerged in the washing mixture until used. In this study, 50 cm lengths of dialysis tubing of size 25 mm fiat were washed.
416 Each length was filled with 50-60 ml of extract, tied closed, and the resulting dialysis bag dialyzed against about 10-fold its volume of chloroform/methanol (2:1, v/v). In one routine run 240 ml of lower phase containing a tot~l of 3 g solutes was distributed among four dialysis bags, and the four bags dialyzed in a 2-1 glass cylinder against 2 1 i t e r s o f c h l o r o f o r m / m e t h a n o l ( 2 : l , v / v ) a t 4 8 C in the dark with magnetic stirring. After 5 days, 2 g (67'~,i of total starting solutes) of solutes were present in the outer liquid. The outer solvent was replaced. After 5 days, 288 mg of solutes had diffused out (9.7 ~; of starting total solutes). The outer liquid was replaced by its own volume of chloroform/methanol/conc. HCI (200:[00:1, v/v/v) and the dialysis was continued in the dark at 10 'C. After 7 days, the outer liquid contained 124 nag (5'Ii, starting solutes) of total solutes. It was replaced by chloroform/methanol (2:1, v.v). Alter 7 days of dialysis in the dark at 10 'C, only 12 nag had dialyzed out. The outer fluid was then replaced four times in succession at two-day intervals. The amounts of solids in the successive dialysates were respectively 4 mg, 3 mg, 2.4 mg and 2 nag. The retentate was collected. It amounted to 208 ml of solution containing 1.50 mg of solids per ml, i.e. a total of 312 nag solids. The sum total of solids in the successive dialysates and in the tinal retentate amounted to 2749 mg leaving 251 rag, i.e. 8,,I,, to be accounted for as losses. Proteolipid apoprotein could be recovered from the retentate by evaporation of the solvent, either in a stream of nitrogen, or in vacuo. The dry residue had a glass-like appearance. It was completely soluble in chloroform: methanol mixtures and insoluble in water.
Preparation qf the water soluble con[ormer o[ proteolipid apoprotein Proteolipid apoprotein recovered from the tinal retentate fronl the dialysis is insoluble in water. To render it ~:oluble it is necessary to replace the chloroform and methanol in its solution by water. This is done by the procedure described by Folch-Pi and Sherman [5, 6, 2] m which the solvents are removed by means of a stream of nitrogen, with concomitant addition of water to the point of immiscibility. The procedure can be run in any number of ways, in a test-tube, in a shallow glass vessel, or in an Erlenmeyer. In this project, a routine run was as follows: a pre-weighed 15-ml test-tube was placed in a beaker containing water at room temperature. The water in the beaker acted as a heat reserw)ir that avoided excessive cooling of the retentate because of the evaporation of the solvents. Therefore, the temperature of the water in the beaker was kept fairly constant by replacing the water that had cooled off. l0 ml of retentate were measured into the tube and a stream of nitrogen was blown from a Pasteur pipette onto the surface of the tube content. In 15 minutes the \olume of retentate was reduced to 2.2 tnl. A second portion of l0 ml of retentate was added into the tube, and the stream of nitrogen continued. In another 15 minutes the volume of the tube contents had been reduced to 3 ml. The stream of nitrogen was cut to a very small rate, the tip of the pipette immersed in the liquid and water added dropwise until after adding 1.7 ml turbidity developed, showing that the point of immiscibility had been reached. The tip of the Pasteur pipette was withdrawn from the liquid, and the rate of delivery of nitrogen was increased. The turbidity faded rapidly. 0.3 ml water was now added along the tube wall, and the tube shaken to insure mixing. Turbidity reappeared. It faded fast. The stream of nitrogen was continued and after another 15 minutes the volume of liquid in the tube was only 1.3 ml. One ml of water
417 was added into the tube, mixing being insured by proper maneuvering of the Pasteur pipette. In 20 minutes, the volume was again reduced to 1.3 ml. Another 1 ml of water was added, and mixed with the contents. In 15 minutes the tube contents had again been reduced to 1.3 ml. A final ml of water was added and the tube stoppered and stored at 4 °C. The solution weighed 2.67 g and contained 27.6 mg of proteolipid apoprotein, i.e., a concentration of 1.04 ~. Portions of the solution were dried in a vacuum desiccator. The residue obtained had a glass-like appearance. It was soluble in water and in chloroform/methanol. The whole procedure took 100 minutes because it was desired to concentrate the retentate so as to prepare a 1 ~ aqueous solution of proteolipid apoprotein. Without this requirement, the process would have taken less than 1 hour. RESULTS
Preparation of proteolipid apoproteins from heart, kidney and liver bovine tissues The procedure of preparation followed had been developed for the central nervous system [2]. It has proved to be applicable to heart, kidney and liver tissues without modification. The extraction of the tissue runs exactly as for neural tissues. Dialysis, if anything, runs more smoothly with the non-neural tissues than with neural tissues, in the sense that the dialysis in acid can be carried out at --10 °C, thus minimizing any effect of the acid. With neural tissues, this step is handicapped by the formation of precipitates inside the dialysis bags, an event that requires moving the dialysis to a higher temperature at which the precipitate will redissolve. This lengthens the procedure. Finally, the preparation of water-soluble proteolipid apoprotein from non-neural tissues is undistinguishable from the process applied to neural tissues. The yields per gram fresh tissue have been approximately: heart, 3 mg; kidney, 2 mg; liver, 1.6 mg. A point deserving emphasis is that with non-neural tissues there is the same diffusion of protein material during the first period of dialysis, that had been observed with neural extracts [20, 21]. That dialyzable protein material had been concentrated and found to be identical in chemical composition to the undialyzable protein [21]. The partial dialyzability of proteolipid apoprotein is compatible with the presence of fractions of proteolipid apoprotein with molecular weights below 12 000 (see below).
Properties of non-neural proteolipid apoproteins The proteolipid apoproteins from heart, kidney and liver removed from the retentate by evaporation of the solvents have a characteristic glass-like appearance. They are freely soluble in chloroform/methanol mixtures, and completely insoluble in water. The proteolipid apoproteins recovered from the water solutions have the same appearance. They are soluble in water and also in chloroform/methanol mixtures, although in the latter case it may be necessary to add 1 or 2 ~ water to the solvent mixture. Solutions of proteolipid apoproteins in organic solvents or in water are usually very clear and they exhibit clean absorbance spectra in the near ultraviolet range, with a maxima at 278 nm and a trough at 255. ~-ay~ ~ l c m (278 nm) = 13.5 ~ 0.15 with a E278 nm/E255 nm ratio of 1.95 to 2.0. These are the exact absorbance values given by neural proteolipid apoproteins.
418
Chemical composition The proteolipid apoproteins studied show no lipid by thin-layer c h r o m a t o graphy. Their P c o n t e n t is from 0.05 to 0.15 }/o (Table II) indicating minimal a i n o u n t s of phospholipids, if any, and their total c a r b o h y d r a t e by the orcinol method [22] a m o v n t s to 0.1 ~o or less. Briefly, the n o n - n e u r a l proteolipid apoproteins appear to be protein material essentially devoid of lipids or carbohydrates. TABLE I AMINO ACID COMPOSITION OF PROTEOLIPID APOPROTEIN FROM DIFFERENT TISSUES Mol per 100 tool recovered in hydrolysate Tissue
Central whitc matter
PLA prepn.* Arginine Histidine Lysine Aspartic Glutamie Half-cystine Methionine Serine Threonine Proline Glycine Alanine Valine Leucine lsoleucine Tyrosine Phenylalanine
2.6 1.8 3.8 4.0 5.8 4.0 1.9 8.5 8.5 2.8 10.3 12.5 6.9 11.1 4.9 4.6 7.9
Central gray matter
Heart
71-VII 72-V1
72-111
72-VI1
72-V
72-1X
72-1V
2.41 1.91 3.31 6.29 5.22
326 1.61 3.26 6.78 5.69 0.36 4.65 6.29 6.67 4.96 9.11 10.28 6.11 13.09 6.60 3.67 7.15
3.34 1.73 3.90 6.12 6.03 0.27 3.90 6.03 5.85 5.48 9.09 10.02 6.87 14.20 6 40 3.53 7.24
3. t 5 1.48 3.08 6.57 5.70 0.80 3.68 6.7(, (~.30 5.43 9.44 10.38 6.69 13.60 6.43 3.47 7.03
2.99 1.82 3.80 5.98 5 69
2.39 1.65 3.86 4.78 6.07 2.39 2.39 5.61 7.54 3.86 10.58 12.06 7.27 I 1.60 5.71 4.69 7.54
2.62 1.90 3.70 4.84 5.76 1.74 2.78 5.70 7.36 4.52 10.23 12.04 7.15 I 1.88 5.60 4.58 7.55
5.35 6.03 6.69 5.29 8.90 10,05 5.82 14.59 7.50 3.41 7.23
Liver
Kidney
3.94 (~.21 5.98 5.25 9.71 10.80 ¢~.58 14.09 6.86 3.43 6.86
* PLA, proteolipid apoprotein. The a m i n o acid c o m p o s i t i o n is given in Table 1, which also contains the a m i n o acid c o m p o s i t i o n of neural proteolipid apoproteins for comparison. In c o m m o n with n e u r a l proteolipid apoproteins, heart, kidney and liver proteolipid apoproteins show a high p r o p o r t i o n of n o n - p o l a r a m i n o acids, with similar polarity indices (see Table V); a relative a b u n d a n c e of sulfur a m i n o acids; a low c o n c e n t r a t i o n of acidic and basic a m i n o acids, and the same a b s o r b a n c e value in the ultraviolet, indicating similar c o n c e n t r a t i o n s of t r y p t o p h a n . They are different from neural proteolipid apoprotein in lacking half-cystine, with m e t h i o n i n e apparently replacing cystine as major sulfur a m i n o acid. There are other less p r o n o u n c e d but yet real differences: n o n - n e u r a l proteolipid apoproteins have higher c o n c e n t r a t i o n s of alanine, glycine and tyrosine and smaller c o n c e n t r a t i o n s of aspartic acid, proline, leucine and isoleucine than are f o u n d in n e u r a l proteolipid apoproteins. We are indebted to Dr. R. M a r t e n s o n of the N I M H for a n analysis for methyl arginine, on two samples of heart proteolipid apoprotein. N o m e t h y l arginine was found. I n analyses of end-groups on the n o n - n e u r a l
419 proteolipid apoproteins Whikehart and Lees [23] f o u n d aspartic acid as only Nterminal, and lysine as only C-terminal amino acids. The most important chemical difference between the two groups of proteolipid apoproteins is f o u n d in the fatty acids covalently b o u n d (Table II). It can be seen that heart and kidney proteolipid TABLE II FATTY ACIDS COMBINED IN PROTEOLIPID APOPROTEIN FROM DIFFERENT TISSUES All the used preparations contained about 0.1 ~ carbohydrate, as galactose. As galactolipids, this would account for about 0.15 70 fatty acids but this is unlikely because of the absence of fatty acids 20:0 or longer. Source and preparation number of proteolipid apoproteins White Matter average Gray Matter 71-VII 72-V1 Heart* 72-III 72-VII Kidney* 72-1V Liver 72-V 72-IX
Total fatty acids as ~ of preparation (a)
Composition of fatty acid mixture as of values in Column a 14:0 16:0
16:1 18:0 18:1 20:0
2.0-4.0
--
-
3.88 3.30
15.1 49.2 4.4 54.0
60.0
3.1 5.3
9.5
25.5 n.d.
14.1 18.5 n.d. 11.1 2 5 . 2 n.d.
P in preparation
0.01-0.04 0.077 0.042
0.40 0.52
0.033 0.042
0.56
0.017
2.45 2.17
6.7 6.6
26.2 32.9 5.9
37.3 26.6 n.d. 20.2 24.6 n.d.
0.015 --
* The amounts of fatty acids present in heart and in kidney proteolipid apoprotein can readily be attributed to phospholipids, since in a diacylphosphoglyceride, fatty acids amount to 18-fold the concentration of P, on a weight basis. apoproteins contain only 0.5 ~ or less of fatty acids, an a m o u n t that can be accounted for as phospholipids or galactolipids. By comparison, only one-fourth o f the 2 to 4 combined fatty acids f o u n d in neural proteolipid apoproteins can be accounted for as complex lipids [24]. Liver proteolipid apoprotein m a y be different from those from heart and kidney, but this point will require further work before a final conclusion can be arrived at.
Reversible conformational changes of proteolipM apoproteins Proteolipid apoproteins f r o m heart, kidney and liver were identical to neural proteolipid apoproteins in their ability to change c o n f o r m a t i o n reversibly with changes in polarity o f the medium. Both the type o f conformational change and its reversibility is shown in Table Ili. It can be seen that all proteolipid apoproteins studied offer a relatively high content o f a-helix in chloroform/methanol (2:1, v/v), a low content in solution in water, and the return to the original a-helix content u p o n simple dilution o f the aqueous solution with chloroform/methanol. It is likely that this change in c o n f o r m a t i o n is the basis o f the dual solubility o f proteolipid apoprotein in c h l o r o f o r m / m e t h a n o l and in water.
420 T A B L E Ill C H A N G E S IN ~ - H E L I X C O N T E N T O F P R O T E O L I P I D A P O P R O T E I N F R O M D I F F E R E N T BOVINE TISSUES ACCORDING TO THE POLARITY OF THE SOLVENT MEDIUM: REVERSIBILITY OF SUCH CHANGES ORD/CD Experiment No.
Source of proteolipid apoprotein
116 131 129 130 127
W h i t e matter 71-VI-2 G r a y matter Heart Kidney Liver
~-Helix content of proteolipid apoprotein (%) In C M 2:1 Ioriginal dialysis retentate~*
In water (prepared front the retentate)**
66 58 68 58 61
37 37 26 22 22
In ( ' M (by dilution of the a q u e o u s solutionl* 66 (80) ~'~t) 54
* F r o m O R D m e a s u r e m e n t s Ira'] 233 nm. C M , c h l o r o f o r m m e t h a n o l mixture. ** F r o m C D m e a s u r e m e n t s [0] 208 nm.
4¸¸¸¸¸¸
i!!i~
'),
?
i:i
H
K
:ii(!i~!ii¸¸¸¸~)i
LI/¸!
Fig. 1. Polyacrylamide gel electrophoresis of proteolipid apoproteins from heart (H), kidney (K), a n d liver (L) in 0 . 1 % s o d i u m dodecyl sulfate. T h e n u m b e r s s h o w n are the calculated molecular weights in t h o u s a n d s . Experimental details in text.
421
Polyaerylamide gel electrophoresis in presence of O.l% sodium dodeeyl sulfate Proteolipid apoproteins precipitated from aqueous solution upon addition of dilute buffers or 6 M guanidine hydrochloride. When aqueous solutions of the apoproteins were added to medium A, a transient precipitate appeared which readily dissolved upon brief stirring. These solutions were subjected to polyacrylamide gel electrophoresis in 0.1 ~ sodium dodecyl sulfate. The results of a typical experiment are shown in Fig. 1. The heart proteolipid apoprotein yielded a major diffuse band with an apparent molecular weight of 34 000. A less intense band, which also corresponded to 34 000, was observed for kidney and liver proteolipid apoproteins. With all samples, a broad diffuse band centered near 12 000 was seen. A considerable portion of the material did not penetrate the gel, thus suggesting that in the presence of sodium dodecyl sulfate, only partial disaggregation had occurred.
Molecular weight determination in sodium dodecyl sulfate-urea gels Non-neural proteolipid apoproteins were submitted to polyacrylamide gel electrophoresis in 8 M urea and 0.1 ~ sodium dodecyl sulfate. The results are shown in Fig. 2. Compared to the pattern obtained with sodium dodecyl sulfate alone there
3o--i b 12
H
K
L
Fig. 2. Polyacrylamide gel electrophoresis of proteolipid apoproteins from heart (H), kidney (K) and liver (L) in 0.1% sodium dodecyl sulfate/8 M urea. The numbers shown are the calculated molecular weights in thousands. D, the dye front (Bromophenol Blue). Experimental details in text.
422 was a striking increase in the number of bands seen with all samples. Furthermore, the bands were more sharply defined than those observed in 0.1 ~/,', sodium dodecyl sulfate alone. The gel patterns in sodium dodecyl sulfate/urea of heart, kidney and liver proteolipid apoproteins were remarkably similar, each displaying 8 to 10 apparently identical bands. The apparent molecular weight of each band was calculated from a calibration curve using standard proteins or cytochrome c peptides of known molecular weight (Fig. 3). They are given in Table IV. Each apoprotein contained a major band of 30 000, which will be referred to as the major non-neural apoprotein. Three identical bands corresponding to very low molecular weight ( < 5000) were also found in each preparation. To test the possibility of further disaggregation, some experiments were conducted with 8 M urea gels containing varying amounts of acrylamide (7.5 to 15°")~oand sodium dodecyl sulfate (0.1-1.0 ~o). No further changes in gel patterns for non-neural proteolipids were observed under these conditions. 60 50 4030 ~ l b u m i n 2o '~ 0 *< IO ~
T ~
hym°trypsinOgen
~e Cytochrome c e\\
Lysozyme
\
Cytochrome c T • k
t,
\
~u
ACTH
L
rE
j
Cytochrome c ]~ ~
i I !
o.o
o.5
RelativeMobility
io
Fig. 3. Calibration curve of polyacrylamide gel c]ectrophoresis in 0,1% sodium dodecyl sulfate/8 M urea. Cytochrome c I, c I1, c Ill, cyanogen bromide peptides from eytochrome c; I, linear least squares regression line for the observed relative mobilities of protein standards with molecular weight greater than 11 700 (cytochrome c); II, linear least squares regression line for protein standards less than 14 300 (lysozyme). DISCUSSION
The widespread distribution of proteolipids throughout animal and vegetal tissues made it imperative to scrutinize proteolipids from different sources, not only to gather new information, but to establish differences and similarities of proteolipids
423 TABLE IV MOLECULAR WEIGHT (±S.D.) OF NON-NEURAL PROTEOLIPID APOPROTEINS The molecular weight is determined by polyacrylamide gel electrophoresis in 0.1 ~ sodium dodecyl sulfate, 8 M urea. The number of experiments is given in parentheses. The band numbers are assigned in descending order from the cathode, Band No. 1 2
3 4 5 4 5 6 7 8 9 10
Heart (6) 35 000 29 100 (5:1000) 27 000 -20 400 16 800 -12 300 :k 800 4 600 :d::800 3 500 5:300 2 400 ~ 300
Kidney (3)
Liver (3)
30 000 % 500 27 800 23 400
30 000 ± 600 28 000 -20 000 17 300 14 600 12 200 ± 1000 10 500 4 400 5:600 3 300 ± 700 2 300 i 650
15 500 13 000 :L 800 10 800 4 700:3_ 300 3 500 ± 500 2 300 ± 400
according to their source of origin. The results reported above have established that the proteolipids from bovine heart, liver and kidney show many similarities and a few differences from the proteolipids previously obtained from the gray and white matters of the central nervous system. Thus, all proteolipids are soluble in chloroform/methanol mixtures and extractable from the tissues with these mixtures. As originally obtained, they are completely insoluble in aqueous solutions. They retain these solubility characteristics throughout the process of gradual delipidation; when maximally delipidated the resulting proteolipid apoproteins still retain their solubility in chloroform/methanol and can now be rendered soluble in water by gradually replacing chloroform/methanol with water in the solvent medium [5, 6]. The dual solubilities correspond to different conformers and the change from one conformer to another is reversible and repeatable (Table lII) i.e. a sample of apoprotein can be changed from the chloroform/methanol-soluble lipophilic conformer to the watersoluble hydrophilic conformer and back to the lipophilic form. This process can be repeated several times without apparent change in either conformer. In chemical composition, the non-neural apoproteins show general similarities, and some specific differences from the neural proteolipids. The respective amino acid patterns have in common the predominance of hydrophobic residues; the relative scarcity of acidic and basic amino acids, an abundance of sulfur-containing amino acids, and of tryptophan. A calculation of the polarity index according to Vanderkooi and Capaldi [25] given in Table V shows that all proteolipid apoproteins have a low polarity index (31-32 ~). On the other hand, compared to the neural proteolipid apoproteins, the non-neural proteolipid apoproteins contain little or no half-cystine, a higher proportion of methionine, small concentrations of glycine, alanine and tyrosine and slightly higher aspartic acid, proline, leucine and isoleucine. All proteolipid apoproteins show the same absorbance at 278 nm, indicating little or no difference in their tryptophan content. The most dramatic difference, however, is the virtual
424 TABLE V POLARITY OF PROTEOLIPID APOPROTEINS FROM VARIOUS TISSUES [251 % Polar, mol percent of Asp, G lu, Lys, Arg. !'; Intermediate, tool percent of Ser, Thr, His, G ty, Tyr. % Polarity : % Polar • (% Intermediate 2}. Source
White matter Gray matter Heart Kidney Liver
Amino acid content o, Polar
"{; Intermediate 2
16.2
16.9
17.0 18. I 18.5 18.9
15.0 13.6 13.6 13.4
",, Polarity
33. I 32.0 31.7 32. I 32.3
absence of fatty acids from heart and kidney apoproteins, against the 2 4'~;, fatty acids covalently combined in central white and gray matters apoproteins. Of course, differences in chemical composition among physically heterogeneous mixtures of substances must be interpreted with great caution. In this work, we limit their significance to showing that the total proteolipid apoproteins from heart, kidney and liver are similar among themselves and different from the total proteolipid apoproteins obtained from the central nervous system. The sharp differences in halfcystine and fatty acid content point to differences in major components of the respective apoprotein mixtures. On the other hand, the smaller differences seen in tile rest of the amino acids might be explained assuming that the two sets of mixtures contain different proportions of common components. The successful preparation of the water soluble proteolipids has permitted their analyses by polyacrylamide gel electrophoresis under different conditions. In 0.1)o sodium dodecyl sulfate all samples showed (Fig. I) substantial amounts of presumably highly aggregated material as a band that had not penetrated the gel. The major part of the material that had entered the gel migrated as a single diffuse band with an apparent molecular weight of 34 000. A number of diffuse lower molecular weight bands were also observed. By contrast, when polyacrylamide gel electrophoresis in sodium dodecyl sulfate/urea was used, all proteolipid apoproteins gave a large number of well-defined bands. Up to 8 to 10 identical bands in the 2500 up to 28 000 range were seen (Fig. 2) with a major band at 30 000. The occurrence of variable proportions of material that did not penetrate the gel, and the increase in the number of bands shown when urea is added to sodium dodecyl sulfate suggests that we are dealing with different states of aggregation. This possibility was advanced by Folch-Pi and Stoffyn [2] on the basis of the partial dialyzability of white matter proteolipids, of the computed minimal molecular size of either 6000 or 12 000, and from the simple numerical relationships of some of the major bands observed in polyacrylamide gel electrophoresis under different conditions of operation. In a more specific suggestion Chan and Lees [19] in their study of crude white matter proteolipids suggested that in polyacrylamide gel electrophoresis in sodium dodecyl sulfate and 8 M urea, myelin proteolipid apoprotein migrates as a series of oligomers of a
425 5000 daltons monomer. These possibilities will require further study on the properties of proteolipid apoproteins. The study of the chemical nature of the major band of heart proteolipid apoproteins (mol. wt. 30 000) is currently under investigation. If we consider the possible implications of some of the chemical properties, we obviously conclude that the marked hydrophobic character of all proteolipid apoproteins is basic to their solubility in some organic solvents; it also requires that the water-soluble conformer must have a very special tridimensional structure which will permit maximal exposure of the scarce hydrophilic residues of the proteolipid apoprotein peptide chain. The presence ofcovalently bound fatty acids, clearly increases the hydrophobicity of the proteolipid apoprotein structure; on the other hand the absence of the acyl radicals in heart and kidney proteolipid apoproteins suggests that the constituent fatty acids must subserve a role particular to the central nervous system, and not to proteolipid apoproteins in general. The presence of a high proportion of half-cystine in neural proteolipid apoprotein in contrast with its essential absence and replacement by methionine in non-neural proteolipid apoproteins suggests that the reduction or oxidation of sulfhydryl groups is not a determinant factor in the solubilities of proteolipid apoproteins and on their structural or functional role in tissues. This suggestion is reinforced by the absence of readily demonstrable -SH groups, and the great stability of -S-S-bridges in neural proteolipid apoprotein reported by Lees et al. [26]. Studies in a number of laboratories have shown that some neural and nonneural proteolipids are concentrated in the mitochondrial subfraction [8, 9, 27]. Recent reports indicate that some or all of the products of mitochondrial protein biosynthesis are proteolipids [28-33]. The molecular weight of the principal product, varies from 2000 [29] to 40 000 [31] and depends upon experimental conditions. The results of our study on proteolipid apoproteins from whole tissue suggest the presence of both low and high molecular weight proteins. Swank et al. [34] in a study of the distribution of proteins in a number of mitochondrial preparations demonstrated in the sodium dodecyl sulfate urea gel system a series of oligopeptides with a molecular weight range of 2500 to 10 000. Our electrophoretic results with non-neural proteolipids under identical conditions are remarkably similar to these observations. The relationships, however, between those oligopeptides and non-neural proteolipids remain to be shown. Recently, however, Tzagoloff et al. [35] has identified a subunit of yeast mitochondrial ATPase as a proteolipid. Finally, Capaldi et al. [36] has isolated a major hydrophobic protein (mol. wt. 29 000) from the inner membrane of bovine heart mitochondria. This protein, which is similar in molecular size to our major heart apoprotein, appears to be a component of the oxidative phosphorylation system [37]. The occurrence of two different conformers of preoteolipid apoprotein in vitro and their ready interconvertibility raises the question of the condition of proteolipid apoprotein in the living tissue. We know that one conformer is hydrophobic and the other hydrophilic. The presence of lipids appears to "lock" proteolipid apoprotein in the hydrophobic conformation. Otherwise the two conformers are freely interconvertible, the polarity of the medium being the factor that determines the conformation adopted by proteolipid apoprotein. The exact question is whether in tissue, proteolipid apoprotein occurs only as one conformer, or as two, and in the
426 latter case, whether the two conformers present undergo in vivo tim same changes that they show in vitro. On superficial observation, the d i s t r i b u t i o n of proteolipid apoprotein a m o n g tissue structures rich in lipids, its extractability with c l H o r o f o r m methanol (accompanied by lipids) and inextractability with aqueous solutions suggest that in the living tissue proteolipid a p o p r o t e i n is present as the hydrophobic conformer. This evidence, however, can be grossly misleading because a "native" hydrophilic proteolipid a p o p r o t e i n could easily be extracted as the h y d r o p h o b i c conformer because of the exposure to c h l o r o f o r m / m e t h a n o l in the course of the extraction. A~ for proteolipid a p o p r o t e i n being locked in a h y d r o p h o b i c c o n f o r m a t i o n by the presence of lipids, it is not difficult to imagine conditions in the tissue which would effectively shield proteolipid apoprotein from contact with lipids. The possibility that proteolipid a p o p r o t e i n be present in tissue as both conformers makes it likely that proteolipid a p o p r o t e i n might undergo c o n f o r m a t i o n a l changes in the m e m b r a n e in vivo. If this be the case, n u m e r o u s possibilities arise for participation of proteolipid a p o p r o t e i n in m e m b r a n e f u n c t i o n : thus, b i n d i n g of specific ligands could effect changes in polarity at the macromolecular level and such changes would u n d o u b t e d l y affect greatly the properties of the microareas in which they had taken place. With this type of consideration we step into a field of pure speculation which might possibly be useful for the f o r m u l a t i o n of working hypotheses and for the definitions of frames of reference of probable alternatives. We feel, hm~ever, that at the present time this type of speculation is idle and outside lhe purpose of this paper. AC K NOWLEDG EM ENTS This work was s u p p o r t e d by grants NS 00130 and FR 05484 from the National Institutes of Health, U.S. Public Health Service. REFERENCES 1 2 3 4 5
Folch, J. and Lees, M. (1951) J. Biol. Chem, 191, 807 817 Folch-Pi, J. and Stoffyn, P. J. (1972) Ann. N.Y. Acad. Sci. 195, 86 107 Tenenbaum, D. and Folch-Pi, J. (1966) Biochim. Biophys. Acta 115, 14t 147 Zand, R. (1968) Biopolymers 6, 939 953 Folch-Pi, J. and Shernlan, G. (1969) Abstr. 2nd hat. Meeting Int. Soc. Neurochcm., p. 169. Tamburini, Milano 6 Sherman, G. and Folch-Pi, J. (1970) J. Neurochem. 17, 597 605 7 Joel, C. D., Karnovsky, M. L., Ball, E. G. and Cooper, O. (1958) J. Biol. Chem. 233. 1565 1573 8 Murakami, M., Sekine, H. and Funahashi, S. (1962) J. Biochem. 51, 431 435 9 Murakami, M., Ozawa, Y. and Funahashi, S. (1963) J. Biochem. 54, 166 172 10 Wolfgram, F. (1967) Biochim. Biophys. Acta 147, 383-385 I1 Eichberg, J. (1969) Biochim. Biophys. Acta 187, 533 545 12 Folch-Pi, J. and Sakura, J. D. (1973) Fed. Proc. 32, 624 13 Chu, R. C. L. and Yasunobu, K. T. (1964) Biochim. Biophys. Acta 89, 148 150 14 Weber, K. and Osborn, M. (1969) J. Biol. Chem. 244, 4406 4412 15 Agrawal, H. C., Burton, R. M., Fishman, M. A., Mitchell, R. F. anti Prensky, A. L. (1972) ,I. Neurochem. 19, 2083 2089 16 Swank, R. T. and Munkres, K. D. (1971) Anal. Biochem. 39, 462 477 17 Sperry, W. M. (1942) Ind. Eng. Chem., Anal. Ed. 14, 88 90
427 18 19 20 21
Fiske, C. H. and SubbaRow, Y. (1925) J. Biol. Chem. 66, 375-400 Folch, J., Lees, M. and Sloane Stanley, G. H. (1957) J. Biol. Chem. 226, 497-509 Lees, M. B., Carr, S. and Folch, J. (1964) Biochim. Biophys. Acta 84, 464-466 Folch-Pi, J. (1971) 3rd Int. Meet. Int. Soc. Neurochem. Abstr. 239, Publ. Akademiai Kiado, Budapest, Hungary 22 Sorenson, M. and Haugaard, G. (1933) Biochem. Z. 260, 247-277 23 Whikehart, D. R. and Lees, M. B. (1971) Fed. Proc. 30, 1247 (Abstr.) 24 Stoffyn, P. and Folch-Pi, J. (1971) Biochem. Biophys. Res. Commun. 44, 147-161 25 Vanderkooi, G. and Capaldi, R. A. (1972) Ann. N.Y. Acad. Sci. 195, 135-138 26 Lees, M. B., Leston, J. A. and Marfey, P. (1969) J. Neurochem. 16, 1025-1032 27 Lees, M. B. (1966) J. Neurochem. 13, 1407-1420 28 Kadenbach, B. and Hadvary, P. (1973) Eur. J. Biochem. 32, 343-349 29 Kadenbach, B. (1971) Biochern. Biophys. Res. Commun. 44, 724-730 30 Tzagoloff, A. and Akai, A. (1972) J. Biol. Chem. 247, 6517-6523 31 Burke, J. P. and Beattie, D. S. (1973) Biochem. Biophys. Res. Commun. 51, 349-356 32 Rowe, M. J., Lansman, R. and Woodward, D. O. (1973) Abstr. Fed, Proc. 2407 33 Kuzela, S., Kolarov, J. and Krempasky, V. (1973) Biochem. Biophys. Res. Commun. 54, 9-16 34 Swank, R. T., Sheir, G. 1. and Munkres, K. D. (1971) Biochemistry 10, 3931-3939 35 Tzagoloff, A., Rubin, M. S. and Sierra, M. F. (1973) Biochim. Biophys. Acta 301, 71-104 36 Capaldi, R. A., Komai, H. and Hunter, D. R. (1973) Biochem. Biophys. Res. Commun. 55, 655659 37 Racker, E. and Kandrach, A. (1971) J. Biol. Chem. 246, 7069-7071