366

Biochimica et Biophysics Acta, 424 (1976) 366-375 @ Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands

BRA 56738

TRIACYLGLYCEROL 6ACCHAROMYCES

LIPASE ACTIVITY

IN BAKER’S

YEAST

CEREVISIAE)

INGER SCHOIJSBOE Department of Biochemistry C, University of Copenhagen, DK-2200 Copenhagen N (Denmark)

Panum Institute, Blegdamsvej 3,

(Received September 12th, 1975)

An investigation of intracellular triacylglycerol lipase activity in baker’s yeast (Saccharomyces cerevisiue) has been performed using emulsified triolein as substrate. Bovine serum has been used as emulsifier since it was found superior to gum arabic and albumin with respect to reproducibility of both triacylglycerol concentration in the assay mixture and specific lipase activity. No ex~acell~~ activity could be detected neither with whole cells nor with water or detergent “extracts” of intact cells as enzyme source. With disrupted cells the level of triacylglycerol lipase activity at a triacylglycerol concentration of 9.6 mM, at pH 7.5, and 30°C was 190 pmol free fatty acids per h per g disrupted cells, Fractionation of a cytoplasmic extract of disrupted cells revealed that about 70% of the activity was associated with membrane fractions and 60% of this activity was present in the mitochondrial fraction. Purification of this fraction was followed by an increase in specific lipase activity which parallels the increase in specific activity of the cytochrome c oxidase.

Lipid particles consisting of equal amounts of triacylglycerols and sterol esters have been isolated from baker’s yeast (Saccharomyces cereuisiae) [l] . It is possible that the ~acyl~ycerol of these lipid particles serves as an energy store from which fatty acids can be mobilized by hydrolysis. Very little is, however, known about intracellular triacylglycerol catabolism in yeast, except that autolyzed baker’s and brewer’s yeast are able to hydrolyze olive oil [Z] . The present study was undertaken in order to obtain more info~ation about the enzymatic activity in baker’s yeast required for the utilization of chemical energy stored as fatty acids in triacylglycerols. The results indicate that more than half

367

of the membrane-bound mitochondria.

intracellular

triacylglycerol

lipase is localized

in the

Materials and Methods Isolation of subcellular fractions. The yeast (S. cerevisiae) was a commercial brand of baker’s yeast produced by De Danske Spritfabrikker A/S, delivered as packed cells 3-6 days after harvesting, stored at 4” C and used within 6 days of delivery. The cells were washed twice with glass distilled water, once with a mannitol solution containing 0.25 M mannitol, 50 mM Tris, 0.1 mM EDTA and

HzS04 to pH 7.4, and stored as packed cells until the next day where the cells were suspended in mannitol solution and homogenized. The homogenate was obtained by mechanical disruption of the cells using glass beads as previously described [l] . pH was readjusted with 1 M KOH immediately after the disruption. The glass beads were removed by sedimentation and washed several times with mannitol solution, which afterwards was added to the homogenate after resedimentation of the glass beads. A cytoplasmic extract containing cytosol plus intact and disrupted organelles [ 31 was obtained as the supernatant after centrifugation of the homogenate (10 min at 1600 X g),

This cytoplasmic extract was fractionated by differential centrifugation into the following three fractions: precipitate after centrifugation for 15 min at 12 000 x g (mitochondrial fraction); precipitate after centrifugation for 60 min at 100 000 X g (microsomal fraction); and final supernatant fluid (soluble fraction) .

Purification of mitochondrial fraction. The mitochondrial fraction was resuspended in mannitol solution and centrifuged at low g value (7 min at 1500 X g).

The mitochondria in the supernatant were sedimented by centrifugation (15 min at 12 000 Xg). The precipitate was separated into a light mitochondrial fraction which could be removed from the heavy mitochondrial fraction simply by decantation after removal of the clear supernatant. After resuspension of the two mitochondrial fractions in mannitol solution the light mitochondrial fraction was centrifuged for 30 mm at 48 000 Xg, resuspended in mannitol solution and centrifuged for 15 min at 12 000 Xg. The resuspended precipitate is referred to as precipitate A. In order to remove contaminating heavy membrane fragments still present, the heavy mitochondrial fraction was layered on top of one-fifth volume of 50% (w/v) sucrose in mannitol solution and centrifuged for 30 min at 48 000 Xg. The floating layer of mitochondria was removed from the centrifuge tubes by suction into a Pasteur pipette, diluted with mannitol solution and sedimented by centrifugation (15 min at 12 000 Xg). The resuspended precipitate is referred to as precipitate B. To account for the protein and enzyme activities during purification, all of the fluids which had been used in washing of the mitochondrial fraction were pooled and centrifuged for 60 min at 100 000 X g. The ensuing precipitate and supematant are referred to as sediment and supematant of washings, respectively. For electron microscopy a pellet of purified mitochondria was fixed in 3% (w/v) glutamldehyde in 0.1 M cacodylate buffer, pH 6.8, and 1% (w/v) 0~0~

366

in 0.1 M cacodylate buffer, pH 7. After dehydration and embedding thin sections were stained with uranyl acetate and lead citrate and examined in a Philips 300 electron microscope. Extracts of intact yeast cells. Water and detergent extracts were prepared by stirring unwashed cells with extraction medium (about 10” cells/ml) on an ice bath for 1 h. The non-polar detergents used were 0.1% (w/v) Triton X-100 or 0.5% (w/v) Lubrol WX. The polar detergents were 10 mM deoxycholate or 10 mM taurocholate. The intact cells were removed by centrifugation (10 min at 1600 X g) and in order to prevent interference from the detergents on the measurement of lipolytic activity protein extracted from the cells was removed from the supernatant by precipitation with (NH,), SO4 added to saturation (700 g/l), keeping pH at 7-8 with NH,OH, and centrifugation for 20 min at 27 000 X g. The precipitate was redissolved in 30 mM Tris . HCl, pH 7.5, and dialyzed over night against 3000 volumes of the buffer. Triacylglycerol lipase assay. The amount of fatty acid released by the enzymatic hydrolysis of glyceryltri( [ l-14C] oleate) was measured radiochemically as [ 1-14C] oleic acid. The substrate suspension was prepared just before use. 100 ~1 of glyceryltri([1-‘4C]oleate) (specific activity 6 /&i/mmol, 0.915 g/ml) and 3 ml of emulsifier was sonicated on an ice bath for 4 X 1 min with 2-min intervals. The assay mixture contained in a final volume of 500 ~1: 150 /..d of substrate suspension which theoretically should give a final triolein concentration of 10.2 mM, 200 ~1 0.3 M Tris - HCl, pH 7.5, and protein from the fraction in question. The chosen pH was found optimal for the lipolytic activity in the particulate fraction. The reaction was started by addition of yeast protein at concentrations known to give linear correlation between amount of enzyme and enzymatic activity. After incubation at 30°C for 30 min, during which time the reaction rate was constant, the reaction was stopped by addition of 100 /..d 1 M H3P04 and rapid cooling followed by lipid extraction as described by Dole and Meinertz [4]. In order to determine the initial triacylglycerol concentration in the incubation mixture, aliquots were taken from the upper phase of the Dole extract. Free fatty acids were extracted from other lipids as described by Kaplan [ 51. The radioactivity of free fatty acids and triacylglycerol samples were counted in a liquid scintillation spectrometer (Nuclear Chicago Mark II). A recurrent problem when measuring lipolytic activity towards water-insoluble substrates is the choice of emulsifier. Several emulsifiers have been suggested for insoluble triacylglycerols. Among these gum arabic which was introduced by Desnuelle and coworkers [6,7] appears to be the preferred one. Bile salts and other detergents may be used but heir action is complex [8] . In this laboratory bovine serum-stabilized triacylglycerol emulsions have been used to study intracellular triacylglycerol lipase activity in beef heart [9]. Bovine serum, however, contain some triacylglycerols, although in very low concentration (< 0.6 mM). It might also contain factors affecting the tricylglycerol lipase activity as it has been shown for human serum and lipoprotein lipase [lo121, but it does not hydrolyze the triacylglycerol during the conditions used in the assay. For the study of lipase from yeast it might therefore be an advantage if a more inert and better defined emulsifier could be used. The activity in purified mitochondria was therefore measured towards triolein emulsions stabilized with bovine serum, 7% (w/v) albumin or 10% (w/v) gum arabic. Among these

369

emulsifiers, only bovine serum and gum arabic produced emulsions with reproducible triolein concentrations, 9.0 + 0.1 [22] and 9.3 rt 0.3 14) expressed as mean It:SE. with the number of preparations in parentheses for serum and gum arabic, respectively, while the actual concentrations of triacylglycerol in albumin suspensions varied from 1.4 to 9.3 in seven preparations. Reproducible triacylglycerol concentrations are of major importance since saturating substrate conditions are seldom obtainable in lipase studies. With serum emulsified suspensions the mitochondrial fraction was found to have approximately the same level of activity in different preparations (293.06 rt 18.01 [ll] nmol fatty acid * h-’ * mg-‘, mean + S.E.), whereas different levels were found with gum arabic-emulsified suspensions (132-253 nmol fatty acid - h-’ - mg-‘). Microscopic examinations of the lipid particles revealed that the serum-stabilized emulsion contained particles with a diameter of maximum 0.5 pm, while gum ~abic-s~b~ized particles had a diameter varying in size from 5 to less than 0.5 pm indicating that the variation in enzyme activity found with gum arabic was caused by the heterodisperse nature of the substrate suspension. Bovine serum was therefore used as emulsifier throughout the experiments. Other methods, ‘Cytochrome c oxidase was assayed spectrophotometrieally at 2’7°C as described by Smith [13], NADPH-cytochrome c reductase activity was measured employing the procedure of Sottocasa et al. [14] except that rotenon was excluded from the assay. tr-Mannosidase activity was assayed spectrophotometrically according to Van der Wilden et al. [ 151. The protein concen~ation was measured by the m~ification of the method of Lowry et al. [16] described by Miller [17]. Crystalline bovine serum albumin was used as the standard. Chemicals. Glyceryltri( [ 1-14C]oleate) was obtained from the Radiochemical Centre, Amersham, England. Cytochrome c, NADPH and p-nitrophenyl-cu-mannoside were purchased from Sigma Chemical Co. (St. Louis, MO.). Unlabelled triolein was obtained from the Hormel Institute (Austin, Mum.). All other chemicals were of purest grade available from regular commercial sources. Results and Discussion Extracellular triacyiglycerol lipase activity

Before the investigation of intracellular lipases was started it was of importance to know if yeast exhibited extracellular lipolytic activity towards triolein emulsions since this would complicate the inte~retation of experiments performed to characterize intracellular lipases. In the yeast family Endomycetaceae to which S. cereuisiae belongs a screening of extracellul~ lipolytic activity towards different Tweens has demonstrated activity only with two Lipomyces species and a Nematospora while in the great majority of species tested including S. cerevisiae no activity could be demonstrated 1181. It has, however, later been reported that intact yeast cells were able to hydrolyze Tween-20 * [19] which suggested the presence of extracellular lipolytic activity. With the triolein suspension as substrate no lipolytic activity could, however, be detected with intact cells or water extract of these as the enzyme source. The pos*

A

polyoxyalk~lene derivative

of sorbitoi monolaurate.

370

sibihty that Tween-20 besides serving as a substrate activity in the cell wall was tested by extraction of ous detergent solutions. Neither polar nor non-polar lease hydrolytic activity towards triolein suspensions

also exposed latent lipase intact yeast cells with varidetergents were able to refrom the cells.

Intracellular triacylglycerol lipase activity Mechanical treatment of the cells disrupted 50% of the yeast cells and uncovered lipolytic activity amounting to 190 pmol of free fatty acid liberated per h per g of disrupted cells. This activity level is about half that calculated from figures for hydrolysis of an 0.5 M unstabilized olive oil emulsion with autolyzed baker’s yeast [2], but up to eight times higher than the level obtained with Tween-20 as substrate for a disrupted cell preparation [19]. A comparison of the level of the triacylglycerol lipase activity in yeast with that in mammalian tissue homogenates shows that epididymal fat, kidney, liver, heart, diaphragma, lung and spleen all hydrolyse gum arabic-stabilized triolein emulsions at a rate 10-20 times lower [20] than that found for yeast using serum-stabilized triolein emulsions. It must, however, be emphasized that care should be taken when rates of hydrolysis of insoluble substrates are compared. The triacylglycerol concentrations and other characteristics are seldom identiCal.

Removal of unbroken cells, cell debris and nuclei from the homogenate caused a decrease in the activity to 88% of that found in the complete homogenate. Distribution of the lipase activity In an attempt to localize the ~~acellul~ lipase a differenti~ fractionation of the cytoplasmic extract was performed. Lipase activity was found in all fractions (Table I). 45% of the total enzyme activity in the cytoplasmic extract was found in a mitochondrial fraction and 27% in a microsomal fraction while 22% was recovered in the soluble fraction. Since these observations suggested that the membr~e-bound lipase was associated with the mitochond~~ fraction this fraction was purified (cf. Materials and Methods). As shown in Table II the specific lipase activity increased in all the particulate subfractions and the same activity was observed for the lipase in precipitate B and precipitate A fraction as well as in the sediment of washings. TABLE I SUBCELLULAR

DISTRIBUTION

OF PROTEIN AND LIPASE ACTIVITY

IN YEAST

The reIative specific activity of the Iipaae is obtained as the ratio between percentage of total activity and percentage o! total protein. The vaIues represent the mean f S.E. from four preparations. Protein Cytop1asmic extract hfitochondriai fraction Microsomal frdction Soluble fraction Recovm

100 19.2 23.2 48.7 SO.5

f * f a

3.6 6.9 7.1 5.2

Lipase activity

Relative specific activity

100 45.8 24.8 22.2 92.9

1 2.4 + 0.4 1.2 f 0.4 0.5 f 0.1 -

f f f *

1.0 4.6 5.4 8.1

371

TABLE

II

THE SPECIFIC

ACTIVITY OF LIPASE AND MARKER ENZYMES BY PURIFICATION OF THE MITOCHONDRIAL FRACTION

IN SUBFRACTIONS

OBTAINED

The mitochondriaI fraction was purIfled as described in MateriaIs and Methoda. The spacHIc activity of the Iipaae is expressed as nmol fatty acid IIberated * h-i * mg protein-‘, that of cytoc’hrome c oxidase as and that of NADPHcytochrome c reductase nmol reduced eytoehrome c oxidized . mIn_’ . m6 protein’ as nmol cytochrome c reduced * min-’ . ma protein’ . a-Mannoddaae activity is expressed as p-nitropho nyl+nmnnosIde hydrolyzed * mine’ * m6 protein-’ . Specific activity of: Cytochrome c oxidase Soniflcation EDTA in wash medium MitochondriaI fraction Precipitate B Precipitate A Sediment of washings Supematant of washings

NADPHcytochrome c reductase

a-mannosidase

LiPase

+ +

+ +

+ +

+ +

1036 2blO 1982 926 0

58.0 32.0 63.3 74.1 6.2

12.8 1.0 18.6 21.2 6.8

78.9 208.9 132.3 94.5 13.4

+ 79.3 94.7 95.4 91.3 26.7

190.2 323.2 263.2 211.8 22.1

It has been demonstrated by Frazer and Walsh [21] that the rate of hydrolysis of lipid by lipase is dependent of the surface area and inversely proportional to the radius of the lipid particle. Benzonana and Desnuelle [221 found that these parameter8 had an effect on the Km value for the enzyme in that finer emulsions of oil gave lower Km values than did coarser emulsions. These result.8 indicate that the formation of the enzyme-substrate complex takes place at the surface of the triacylglycerol particles. Membrane-bound lipase may, therefore, act only at collisions between lipase carrier particles and substrate particles and the size of both of these particles must be critical for the resulting enzyme activity. In order to obtain almost the same size of lipase carrier particles in the mitochondrial subfractions, these as well as an aliquot of the mitochondrial fraction itself were sonified for 2 X 15 min (MSE, 150 W ultrasonic disintegrator), on an ice bath before assaying the activity of the lipase and the marker enzymes. After sonification the lipase activity increased in all of the particulate subfractions, but not to the same extent. Highest activation was observed in precipitate B. The decrease in lipase activity in supematant of washings may indicate that the sonification affected the lipase activity. This could also explain why no activation wa8 observed in the mitochondrial fraction either but, the matter is complicated by the question of the relative increase of the number of inert particles (non-lipase carrier particles) introduced by sonification. The parallel increase of the lipase and the cytochrome c oxidase activities in the purified fractions containing precipitate B and precipitate A and the decrease in specific activity of NADPHcytochrome c reductase and o-mannosidase confirm the indication that the membrane-bound intracellular lipase in yeast is localized in the mitochondrial membrane. This localization of an intracellular lipase is in contrast to the localization in mammalian tissue where this activity at basic pH has been shown to be associated with the microsomal fraction [9,23,24] and at acidic pH with the lysosomal fraction [25]. NADPHcytochrome c reductase has recently been shown to be present in the microsomal

pala%1Ol3

'wxl8mdardaql

U! ?uasazdare

'x000 OS S~~Oqlua~sds~Il-Uo~l~~la

'~~~uoq~olrur

Pa?wndJo

qd8~o~~!wuoxS~aI3

(H)'XOOOO~

'was

osttrarasmam%ay

amrqmatu

313

fraction from yeast [26] and used as a marker enzyme for this fraction [27] and a-mannosidase activity as a marker enzyme for vacuolar membranes [ 151. The activity of NADPHcytochrome c reductase and cu-mannosidase in precipitate B may indicate that this fraction is still contaminated to a certain extent with microsomal proteins and vacuolar membranes, but the change of a-mannosidase and NADPH-cytochrome c reductase activities with the purification of the mitochondrial fraction appear to be inconsistent with a localization of the lipase in any of these fractions. The purity of precipitate B was also tested electron microscopically. Electron micrographs of this fraction shows the mitochondria to be relatively pure (Fig. 1A). The few electron-dense granules may be ribosomes and the electron-transparent bodies of vacuolar or microsomal origin. The elongated structures are probably debris of plasma membranes, which may still be present despite the washing of the mitochondria in mannitol solution followed by low speed centrifugation and a high speed centrifugation on a discontinuous sucrose gradient (cf. Materials and Methods). Fig. 1B shows that the mitochondria exhibit a morphology normally shown for yeast mitochondria [28-301 and with no cytoplasmic ribosomes attached to the mitochondrial outer membrane. The effect of EDTA on the lipolytic activity EDTA has been included in the homogenization media as recommended for preparation of yeast mitochondria [28,29,31-331. Furthermore it has recently been shown to be necessary for the removal of cytoplasmic ribosomes which in the presence of Mg2+ are attached to the mitochondrial outer membrane [25]. It was, however, observed that further dilution of the final cytoplasmic extract with media containing EDTA prior to assaying the lipase activity led to a decrease in specific activity of the lipase, while dilution with media not containing EDTA had no effect on the activity. Purification of the mitochondrial fraction in media containing EDTA also led to a much lower activity than obtained after purification in EDTA-free media (Table II). Subsequent sonification of these two mitochondrial preparations led to almost the same relative increase in specific activity in precipitate B and it is therefore unlikely that the higher activity measured in precipitate B purified without EDTA was caused by disintegration of the mitochondria following the dilution of the EDTA originally present. The direct effect of EDTA in the assay system was examined. Precipitate B purified without EDTA in the washing media was assayed for lipase activity in the presence of either EDTA or ethyleneglycol-bis-@-aminoethylether)-Nfl’tetraacetic acid (EGTA), which was added to the incubation mixture prior to the addition of the mitochondrial protein. The concentrations of the two chelators in the assay mixture were chosen so as to give at most the same final concentration as if they were added with the mitochondrial protein washed and diluted in EDTA medium. No difference in activities were, however, observed which could possibly reflect the rather high Ca2+ and Mg’+concentration in the assay mixture (from the serum). On the other hand the observation of a decreased activity in the cytoplasmic extract diluted with EDTA medium showed that this effect of EDTA cannot be reversed by Ca’+and Mg2+ in the assay. Lipolytic activity of the mitochondria measured in a Ca2+- and Mg’tdepleted assay mixture was unaffected by addition of Ca2+ or Mg2+ to the

374

system in concentrations from 0 to 2 mM and 0 to 1 mM, respectively. Also addition of F- in concentrations up to 50 mM had no effect on the lipolytic activity in a complete assay mixture. Thus, it seems that the enzyme has no requirement for free Mg2+ or Ca’+. The relative distribution of the lipase activity in the different subcellular fractions was unaffected by the EDTA effect since all fractions except the soluble one were diluted to give almost identical ratios between concentrations of EDTA and subcellular proteins. Acknowledgements Professor 0. Behnke, Department of Anatomy C, University of Copenhagen, is cordially thanked for performing the electron microscopy. The valuable comments on the manuscript by Dr. P.K. Jensen, Department of Biochemistry C, University of Copenhagen, and the expert technical assistance by Miss Ida Henningsen and Mr. John Svaneborg is gratefully acknowledged. Bovine serum was kindly supplied by Professor P. Havskov S$rensen, The Royal Veterinary and Agricultural University, Copenhagen, and baker’s yeast was donated by De Danske Spritfabrikker A/S. The research was supported by grant no. 511-1838 from the Danish State Science Research Council. References 1 Clausen. M.K.. Cbriatiansch. K., Jensen. P.K. and Behnke. 0. (1974) FEBS Lett. 43.176-179 2 Gorbach, G. and Giintner. H. (1932) Sber. Akad. Wiss. Wien 141. 416428 3 De Duve, C.. Pressman, B.C., Gianetto, R.. Wattiaux. R. and Appelmans, F. (1966) Biochem. J. 60, 604-617 4 Dole, V.P. and Meinertz. H. (1960) J. Biol. Chem. 235, 2595-2599 5 Kaplan, A. (1970) Anal. Biochem. 33. 218-225 6 Sarda. L. and Desnuelle. P. (1958) Biochim. Biophys. Acta 30. 513-521 7 Desnuelle. P.. Constantin. M.J. and Baldy. J. (1955) Bull. Sot. Chim. Biol. 37. 285-290 8 Wills, E.D. (1965) Advances in Lipid Research (Paoletti. R. and Kritchevsky. D.. eds.). Vol. 3. pp. 197-240, Academic Press, New York 9 Schousboe, 1.. Bartels, P.D. and Jensen. P.K. (1973) FEBS Lett. 35. 279-283 10 Ganesan, D., Bradford, R.H., Alaupovic. P. and McConathy. W.J. (1971) FEBS Lett. 15, 205-208 11 Fielding, C.J. (1970) Biochim. Biophys. Acta 206. 118-124 12 Fielding, C.J., Lim, C.T. and Scanu. A.M. (1970) Biochem. Biophys. Res. Commun. 39, 889-894 13 Smith. L. (1956) Methods of Biochemical Analysis (Glick. D.. ed.). Vol. 2. pp. 427434, Wiley Interscience, New York 14 Sottocasa. G.L., Kuylenstierna, B., Emster. L. and Bergstrand. A. (1967) J. Cell Biol. 32. 415-458 15’ Van der Wilden. W., Matile. P.W.. Schellenberg. M.. Meyer. J. and Wiemken. A. (1973) 2. Naturforsch. 28.416-421 16 Lowry. O.H.. Rosebrough. N.J.. Farr. A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275 17 Miller, G.L. (1969) Anal. Chem. 31. 964 18 Werner, H. (1966) Zentralbl. Bakt. Parasit. Kde. 200. Abt. 1.113-124 19 Nurminen. T. and Suomalainen, H. (1970) Biochem. J. 118.759-763 20 Biale. Y., Gorin. E. and Shafrir. E. (1968) Biochbn. Biophys. Acta 162. 28-39 21 Frazer. A.C. and Walsh, V.G. (1933) J. Physiol. Lond. 78. 467474 22 Benzonana, G. and Desnuelle, P. (1966) Biochim. Biophys. Acta 106.121-136 23 Carter. Jr.. J.R. (1967) Biochbn. Biophys. Acta 137, 147-156 24 Colbeau, A., Cuault, F. and Vignais. P.M. (1974) Biochimie 56.276-288 25 Mahadevan. S. and Tappel. A.L. (1968) J. Biol. Chem. 243.2849-2864 26 Yoshida. Y.. Kuma0ka.H. and Sate. R. (1974) J. Biochem. Tokyo 76.1201-1210 27 Cobon. G.S.. Crowfoot. P.D. and Linnane. A.W. (1974) Biochem. J. 144.266-276 28 Balcavage, W.X. and Mattoon. J.R. (1968) Biochim. Biophys. Acta 163. 621-530 29 Duell, E.A., Inoue. S. and Utter, M.F. (1964) J. Bacterial. 88.1762-1773

375 30

Kellems, R.E.. Allison, V.F. and Butow. R.A. (1974)

31 Szabo. A.S. and Avers, C.J. (1969) 32 Janki, R.M.. Aithal, H.N.. 446461 33 Londesborough. J.C. (1975)

J. Cell Biol. 65, l-14

Ann. N.Y. Acad. Sci. 168.302-312

Tustanoff.

E.R.

and Ball, A.J.S.

FEBS Lett. 50.283-288

(1976)

Biocbim.

Biophys.

Acta 376,