Isolation and characterization of peroxisomes from the liver of normal untreated rats

Transcription

1 Eur. J. Biochem. 149, (1985) 0 FEBS 1985 Isolation and characterization of peroxisomes from the liver of normal untreated rats Alfred VOLKL and H. Dariush FAHIMI Department of Anatomy, 11. Division, University of Heidelberg (Received January 18/February 25, 1985) - EJB The classic method of Leighton et al. [(1968) J. Cell Biol. 37, for the isolation of peroxisomes from rat liver involves the use of Triton WR-1339 which alters the biochemical properties of this organelle and requires the specialized type Beaufay-rotor which is not easily available. We have employed Metrizamide as the gradient medium and a commercial type vertical rotor to obtain highly purified and structurally well-preserved peroxisomes from normal untreated animals. The livers were homogenized in buffered 0.25 M sucrose and a slightly modified light mitochondria1 fraction was prepared by differential centrifugation. This was loaded on top of a linear Metrizamide gradient ( g/cm3) and subjected to an integrated force of x lo6 x (g x min) using a Beckman VTi 50 vertical rotor. Peroxisomes banded at the density of g/cm3. In the isolated fraction 95% of the protein was contributed by peroxisomes, which exhibited a strong activity for cyanide-insensitive lipid p-oxidation. The purity of fractions was also confirmed by morphometry, which revealed that 98% of isolated particles consisted of peroxisomes. The latency for catalase was about 90% indicating a high degree of peroxisomal integrity. This corresponded to the low level of extraction of catalase in 3,3 -diaminobenzidine-~tained filter preparations. The entire procedure took about five hours. Highly purified and structurally well preserved peroxisomes should be useful in further elucidation of the function of this organelle and especially in studies of peroxisomal enzymes with multiple intracellular localizations. A simple and reproducible procedure for the isolation of an organelle in highly purified and morphologically well preserved form facilitates its detailed investigation. As far as the peroxisomes are concerned there are three main obstacles : (a) their relative paucity (in liver they make up 2% of the total protein); (b) their density in sucrose, which is very close to that of mitochondria and lysosomes; and (c) their extreme fragility. In the past fifteen years, several methods for the isolation of this organelle have been described [l Each of these procedures has certain limitations and only in a few studies has a comprehensive analysis, including electron microscopy and cytochemistry of isolated peroxisomes, been reported. For example, the classic procedure of Leighton et al. [5] involves the pretreatment of animals with the detergent Triton WR-1339, which affects the activity of peroxisomal enzymes [8] and the morphology of the liver [9]. In addition, it requires the specialized type Beaufay-rotor, which is not easily available. Wattiaux et al. [lo] described the advantages of Metrizamide as a new medium for subfractioning liver homogenates and mentioned its feasibility in obtaining peroxisomes from untreated animals. Similar observations have been reported by others [2, 111. The swing-out rotors Correspondence to A. Volkl, Anatomisches Institut, Universitat Heidelberg, Im Neuenheimer Feld 307, D-6900 Heidelberg, Federal Republic of Germany Enzymes. Hydrogen peroxide : hydrogenperoxide oxidoreductase (EC ); glyco1late:oxygen oxidoreductase (EC ); carboxylic-ester hydrolase (EC 3.1.l. 1); NADPH :cytochrome c oxidoreductase (EC ); cytochrome c:oxygen oxidoreductase (EC ); L-glutamate:NAD+ oxidoreductase (EC ); D-lactate: NADf oxidoreductase (EC I.I.1.28); orthophosphoric-monoester phosphorylase (EC ); urate: oxygen oxidoreductase (EC ); P-D-glucuronide glucuronosohydrolase (EC ). employed by these investigators, however, require prolonged centrifugation times which are deleterious to the integrity of peroxisomes. Vertical rotors having a shorter path length provide a more gentle tool for quick separation of fragile organelles. Neat et al. [7] and Appelkvist et al. [l] used this type of rotor in conjunction with a Percoll-gradient for isolating peroxisomes. However, Percoll is not the ideal gradient medium, because particles of colloidal silica tend to stick along the wall in a vertical rotor [12] and Percoll interferes with the measurement of peroxisomal enzyme activities [l]. In this respect, Metrizamide was found by our group [13] and others [2, 4, 141 to be a more suitable medium than Percoll. So far, however, no detailed biochemical and morphological data have been published on peroxisomes isolated by this procedure. In the present study we have subfractionated crude preparations of liver cell organelles by means of a linear Metrizamide gradient in combination with a commercially available vertical rotor. The results demonstrate that highly purified and structurally well preserved peroxisomes can be isolated from livers of untreated rats within short timeperiods. MATERIALS AND METHODS Preparative procedures Female Sprague-Dawley rats weighing g were fasted overnight. Under light ether anaesthesia the livers were perfused with physiological saline via the portal vein for 2-3 min, excised and minced in chilled homogenization buffer (buffer A: 5 mm Mops ph 7.4, 250 mm sucrose, 1 mm EDTA, 0.1% ethanol). They were homogenized for 2 min at 1000 rpm in approximately 20 ml of the same buffer, using a

2 258 Supernatant C Crude hornogenate A I? : Suoernatant Pellet B I Pellet D I:, Fig. 1. Diagrammatic representation of the fractionation of rat liver homogenate by differential (steps I, 11, IV) and density gradient centrifugation (step Ill). Conditions used: (I) fixed angle rotor JA-20 (Beckman), 1900 x g (av.), 10 min, 4 C; (11) fixed angle rotor JA-20 (Beckman), 25300xg (av.), 20 min, 4 C; (111) vertical rotor VTi 50 (Beckman), x g (av.), x 106g x min, (IV) fixed angle rotor Ti 45 (Beckman), 1.25 x lo5 xg (av.), 45 min, 4 C. Fractions: (B) heavy mitochondrial fraction; (D) light mitochondrial fraction; (F) final supernatant; (G) microsomal fraction. The composition of gradients: (1) gradient buffer: 5 ml; 5 mm Mops, 1 mm EDTA, 0.1% (v/v) ethanol, 2% (w/v) Dextran T10; (2) suspension of D in homogenization buffer: 5 ml; 250 mm sucrose, 5 mm Mops, 1 mm EDTA, 0.1% (v/v) ethanol, ph 7.4; (3) linear gradient ( g/ cm3) of Metrizamide in gradient buffer; (4) cushion: 2 ml; Metrizamide in gradient buffer, 1.27 g/cm3 Potter-Elvehjem homogenizer held in an ice-water-bath, and only a single stroke was applied for each liver. The homogenate was filtered through two layers of cheesecloth and adjusted to 5 ml/g liver with ice-cold buffer A. Differential-pelleting of the initial homogenate (A) and subfractionation of the crude peroxisomal preparation (D) by density-dependent banding was performed as depicted in Fig. 1. Briefly, two consecutive steps of differential centrifugation of fraction A resulted in the pelleting of fraction D. Each of the pellets obtained was washed once by resuspension in an appropriate volume of buffer A using a glas-rod and recentrifugation under the same conditions. 5 ml of fraction D corresponding approximately to one liver were top-loaded on preformed, linear Metrizamide-gradients ( g/cm3) resting on a 2 ml cushion with the density of 1.27 g/cm3. Metrizamide solutions were prepared in gradient buffer composed of 5 mm Mops ph 7.4, 1 mm EDTA, 0.1% ethanol, 2% (w/v) Dextran T10. The tubes were subjected to an integrated force of x lo6 x (g x min) corresponding to a maximal relative centrifugal force of x g and a total centrifugation time of 45 rnin (acceleration and deceleration included) using a Beckman VTi 50 vertical rotor in a Beckman L5-65B ultracentrifuge (Beckman Instruments, Munich, FRG). Fourteen fractions (2-5 ml) from bottom to top were collected by means of a glas-cannula connected via plastic tubing to a peristaltic pump. They were either diluted for biochemical assays with a hypotonic solution containing 1 mm Na2C03 ph 7.6, 1 mm EDTA, 0.1% ethanol and 0.01 YO Triton X-100 or processed for morphological studies. Aliquots of the main fractions A - G were treated similarly. All manual operations were carried out in the cold-room (5 C). The entire procedure from the anaesthesia to collecting the last gradient fractions takes approximately five hours. Biochemical analyses Assays were run with a recording Beckman spectrophotometer (model 24, Beckman Instruments, Munich, FRG). Following standard procedures were employed for determining enzymatic activities: catalase and a-hydroxyacid oxidase 1151; peroxisomal /3-oxidation [16]; esterase [17]; NADPH-cytochrome c reductase [18] ; cytochrome c oxidase [19]; glutamate dehydrogenase [20, 211. Lactate dehydrogenase and acid phosphatase were assayed using commercially available test-kits (Sigma: Techn. Bull. no. 340 UV and no. 104). In the assay for urate oxidase a 50 mm borate buffer ph 9.0 was used [22]. The assay for b-glucuronidase was performed at 56 C [23]. Protein was determined according to Lowry et al. [24], after precipitation with trichloroacetic acid. Results expressed in bovine serum albumin equivalents are given in mg/g liver or mg/ml. Enzyme activities are in units/g liver and U/ml respectively: 1U = 1 pmole substrate turned over or product formed per minute with the exception of catalase and cytochrome c oxidase which are defined according to Baudhuin et al. [15] and Cooperstein and Lazarow 1191 respectively. Conversion of the raw data to distribution histograms was done as described by Beaufay et al. [25]. The proportion of protein contributed to the peroxisomal fraction by the different membranes was calculated according to Leighton et al. [5]. Morphological studies The filtration apparatus developed by Baudhuin et al. [26] was used and the method of Leighton et al. [5] with some modifications applied. Briefly, aliquots of fractions from the Metrizamide gradient were mixed with one half their volume of an ice-cold fixative containing 7.5% glutaraldehyde in the same buffer used for the gradients to which 0.48 g/ml of Metrizamide was added. The use of gradient buffer and the addition of Metrizamide to the fixative solution markedly improved the general preservation of peroxisomes and diminished the extraction of matrix. After the fixation for 30 rnin at 4 C the fractions were diluted with 2.5% glutaraldehyde in gradient buffer to a final concentration of 150 pg protein/ ml and 1 ml was filtered through Millipore membranes (Millipore Corp. Bedford, MA, USA) with a pore size of pm and a diameter of 13 mm at 1.0 bar. The filtration took approximately 60 min and subsequently the preparations were either postfixed with the reduced osmium procedure of Karnovsky 1271 or were incubated for 30 rnin in the alkaline DAB medium for cytochemical localization of catalase [28] followed by postfixation in 2% aqueous osmium tetroxide. All material was dehydrated in graded ethanol solutions and embedded in Epon. Prior to dehydration filters were covered with a layer of 5% Agar to prevent the loss of particles. Semithin sections cut with glass knives were stained with toluidine blue and examined by a light microscope. From selected regions showing a uniform layer of particles ultrathin sections were cut with a diamond knife on an ultramicrotome (LKB, Ultratome IV, LKB Instruments, Bromma, Sweden) contrasted with uranyl acetate and lead citrate and examined in a Philips 301 electron microscope. The purity of fractions was assessed by counting the peroxisomes, mitochondria and lysosomes in electron micrographs taken at original

3 259 magnifications of and enlarged 2.5 times. Micrographs including the entire thickness of a pellicle and from widely spaced regions of it were examined and particles were counted. The results on different sections of the same pellicle were very close, thus confirming that the filtration technique used [26] provided adequate random sampling for the isolated peroxisomes. The mean diameter of peroxisomes was determined in electron micrographs of pellicles taken at original magnification of using an Apple graphic tablet attached to an Apple IIe microcomputer. For calibration of magnification a carbon grating replica with 2160 lines per mm (Fullam, Schnectedy, NY, USA) was used, which was photographed at the same magnification as pellicles and further processed in identical fashion as micrographs of peroxisomes. A total of 304 particles exhibiting a well defined and sharp limiting membrane were analyzed by this method. x.- * > c u.i 13 LL a Materials Female SD-rats (230 g) were obtained from the Zentrale Versuchstieranlage, University of Heidelberg. Metrizamide (Nyegaard & Co., Oslo, Norway) was purchased from Dr Molter GmbH, Heidelberg, FRG. Dextran T 10 (Pharmacia Fine Chemicals, Uppsala, Sweden) was from Deutsche Pharmacia GmbH, Freiburg, FRG; titanium oxysulfate from Riedel-de Haen, Seelze, FRG and glycolic acid from Merck AG, Darmstadt, FRG. Cytochrome c (grade HI), o-nitrophenyl acetate, p-nitrophenyl-/i-glucuronide, palmitoyl-coa, L-glutamate and bovine serum albumin (fraction V) were obtained from Sigma GmbH, Munich, FRG. All other chemicals were from Merck AG, Darmstddt and of the purest analytical grade available. m RESULTS Properties of the initial fractions The compositon of the initial fractions B, D, F and G, reflected by the percentage distribution and the relative specific activity of marker enzymes is shown in Table 1. The bulk of mitochondria is concentrated in fraction B, together with about half of the initial activity of lysosomal enzymes and 40% of the total protein. Thus, fraction B is quite similar to the combined N and M fractions of De Duve et al. [29]. Fraction D represents a crude preparation of peroxisomes corresponding to a slightly modified L-fraction of De Duve et al. [29] and I+-fraction of Leighton et al. [5]. Fraction G is mostly made up of microsomes being comparable to the P fraction of De Duve et al. [29]. The supernatant fraction F (corresponding to fraction S of De Duve et al. [29]) contains approximately 40% of the total catalase activity and the bulk of the lactate dehydrogenase. In Fig. 2 the mean values of the relative specific activity of the marker enzymes for various cell organelles are plotted, demonstrating that fraction D is enriched 4-5 times in peroxisomes, 3 times in lysosomes and about 2 times in mitochondria. +I +I +I +I +I +I +I +I +I +I Properties of fractions obtainedjrom the Metrizamide gradient The distribution and enrichment of cell organelles after the density gradient centrifugation is shown in Fig. 3. Peroxisomes are recovered in fractions 2-4 with the mean equilibrium density of g/cm3. Mitochondria mostly band in a density range of about 1.I 53 g/cm3 corresponding

4 260 rt -Hydroxyacidoxniase Acyl-COA oxdase g f Cytochrome C 34 Oxidase 1.0 Cytochrome C p oxdase 0:5 1.0 I 8 ID1 F I G I Relative protein content Ap(z Ap)- Relative volume JV(XJV)- 1.0 Fig. 2. Distribution of marker enzyme activities in the initial fractions B, D, F and G. For explanation see legend to Fig. 1 and text. Abbreviations used are: U, totals units of an enzyme found in a single fraction;.z U, total units found in all fractions; d p, total protein content of a single fractions; Z A p, total protein content of all fractions. Z U/g liver of marker enzymes determined and Z d pig liver are summarized in Table 1 Fig. 3. Distribution of protein and enzyme activities following density gradient centrifugation offraction D. Peroxisomes are concentrated mainly in the bottom-fractions 2-4 with a mean density of g/ cm3. d V represents the volume of a single gradient fraction, U its corresponding enzyme activity, Z d V the total gradient volume and Z U the total units recovered from all gradient fractions to the bulk of protein. Microsomes predominantly equilibrate in the penultimate top-fraction (density g/cm3), however a tailing down to heavier fractions is also observed. Lysosomes remain in the uppermost part of the gradient being well separated from peroxisomes. Some catalase and a-hydroxyacid oxidase activities are also found in the uppermost gradient fraction, however, no capacity for lipid p-oxidation could be detected in this part of the gradient. Properties of the purified peroxisomal fraction The main biochemical properties of the peroxisomal fraction are listed in Table 2. Estimated by the specific peroxisomal reference enzymes it is purified about 38-fold over the original homogenate. It constitutes approximately 10% of the total peroxisomal enzyme activity and 0.28% of the protein content of the whole liver. By applying the method

5 ~ ~~ ~ ~ ~~ ~ ~ 26 1 Table 2. Properties of the purified peroxisomal fraction from density dependent subfractionation of D Values given are means f standard deviation Enzyme No. of ex- Composition periments of fractions Protein Catalase a-hydroxyacid oxidase Lipid /I-oxidation b-glucuronidase Acid phosphatase Esterase NADPH-cytochrome c reductase Cytochrome c oxidase L-Glutamate dehydrogenase Lactate dehydrogenase % whole liver 0.28 f f f f * Relative specific activity of fractions with respect to whole liver 1.oo f f & f fraction D ~ ~ 1.oo 8.32 f f & f f f of Leighton et al. [5] more than 95% of the total protein content of this fraction is contributed by peroxisomes with mitochondria and microsomes accounting for about 2% each and lysosomes for less than 1 %. The specific activity for lipid P-oxidation in this fraction (130 nmol NAD x min- x mg- ) is comparable to that of peroxisomes characterized as highly purified by Lazarow and De Duve [16]. The free activity of catalase, as defined by Baudhuin [30] is approximately 8-10% implying the high degree of integrity of isolated peroxisomes. Morphological characteristics of isolated peroxisomes Fig. 4 is a low-power survey electron micrograph of the peroxisome fraction. The section is perpendicular to the surface of the pellicle, showing the sample over its whole depth. The fraction consists almost exclusively of peroxisomes which appear well preserved with only a rare mitochondrion and lysosome. The peroxisomes make out 98-99% of the fraction with approximately 1 % mitochondria and less than 1 YO lysosomes (3300 particles counted). The diameter of the isolated peroxisomes ranges between 0.05 and 1.03 lm with a mean diameter of Fm. A distinct and well preserved limiting membrane with no disruptions surrounds the electron opaque matrix of most peroxisomes. In a few instances short tail-like extensions of the membrane are noted (Fig. 6). Although the majority of peroxisomes contain the typical electron dense crystalline inclusions of uricase (Fig. 6), some free cores are also found between the peroxisomes (Fig. 4). In filter preparations incubated in the alkaline 3,3 - diaminobenzidine medium for visualization of catalase (Figs 5 and 7), a uniform electron-dense reaction product is seen over the matrix of the majority of peroxisomes with a few exhibiting different degrees of extraction. The influence of the gradient medium, if any, on extraction of catalase was assessed by comparing the peroxisomes in the D fraction prior to separation in Metrizamide, with those after the gradient centrifugation. As shown in Figs 8 and 9, they do not differ significantly. In both preparations there are many peroxisomes with a uniformly stained matrix comparable to those in well fixed and incubated sections of rat liver [31], next to a few with slight to moderate extraction. The speed of centrifugation in Metrizamide was quite important in obtaining intact peroxisomes. When the gradient centrifugation was performed at x g for 60 min, as indicated by Hajara and Bishop [4], a separate band adjacent to peroxisomes was obtained which consisted exclusively of free peroxisomal cores (Fig. 10). The composition of the fixative was important in preventing the extraction of catalase from peroxisomes. When a 7.5% glutaraldehyde solution buffered with 0.1 M cacodylate ph 7.2 was used, many peroxisomes were partially extracted showing an electron lucent matrix with the 3,3 -diaminobenzidine reaction product localized to the inner aspect of their membranes (Fig. 11). When instead of cacodylate the gradient buffer containing Metrizamide was used the number of such peroxisomes was reduced (Fig. 12). The frequency of damaged particles, however, remained unchanged under both conditions. DISCUSSION Properties of the pur fied peroxisome fraction The peroxisomes obtained by the procedure detailed in this study were: (a) highly purified being practically free of contamination by other cell organelles, and (b) exhibiting a high degree of integrity by biochemical, ultrastructural and cytochemical criteria. Since they were obtained from normal animals without any pretreatment with a detergent or peroxisome proliferating agent, they can be considered as true normal peroxisomes obtained from rat liver. Applying the relationship given by Leighton et al. [5], 95% of the protein in isolated fractions was contributed by peroxisomes. This high degree of purity was confirmed independently by morphometric analysis of filter preparations which revealed that 98% of isolated particles consisted of peroxisomes. Furthermore, this purity is also reflected in the high activity for lipid P-oxidation which matches the values reported by Lazarow and De Duve [16] for highly purified peroxisome preparations. Such fractions are required in studies of enzymes with multiple intracellular locations. Indeed, we have found in mouse liver peroxisomes, using the present isolation method, high levels of epoxide hydrolase [32], which was previously considered to be localized exclusively in microsomes. Furthermore, preliminary studies indicate that rat liver and kidney peroxisomes contain hydroxymethylglutaryl-coa reductase [33], which is also generally considered a microsomal enzyme [34]. This finding confirms the recent immunoelectronmicroscopic observations of Keller et al. [35].

6 262 Fig. 4. A low power survey electron micrograph of the peroxisomalfraction,fiwed with glutaraldehyde followed by the reduced osmium procedure. The entire thickness of the pellicle is shown here which consists almost exclusively of peroxisomes with only a rare mitochondria (M). Section contrasted with lead citrate. Bar = 1 wm Fig. 5. A preparation comparable to that in Fig. 4, but incubated in the alkaline 3,3'-diaminobenzidine for localization of catalase. Most peroxisomes contain a homogenous electron dense reaction product over their matrix. A few damaged particles with partial or complete extraction of catalase are also present. Contrasted with lead citrate. Bar = 1 pm Fig. 6. A higher power view of the isolated peroxisomes showing a distinct and well preserved limiting membrane with no disruptions. The peroxisome matrix appears finely granular and most particles contain the typical crystalline cores. Contrasted with lead citrate. Bar = 1 pm Fig. 7. Higher power view of isolatedperoxisomes incubated in the ulkuline 3,3'-diaminobenzidine. Most pcroxisomes show a uniform staining of matrix with very little extraction. Contrasted with lead citrate. Bar = 1 pm

7 263 Fig. 8. The appearance of peroxisomes and the extraction of catalase in the crude peroxisomal fraction (fraction 0). (*) Well preserved peroxisomes without any extraction of catalase. MITO, mitochondria; LYS, Iysosomes. Contrasted with lead citrate. Bar = 1 pm Fig. 9. The appearance ofperoxisomes and the extraction of catalase in the fraction afier the Metrizamidegradient. (*) Well preserved peroxisomes without any extraction ofcatalase. Contrasted with lead citrate. Bar = 1 pm Fig. 10. A pure corefraction obtained by performing the gradient centrifugution at x g for 60 min as suggested by Hujara und Bishop [4]. The extrusion of cores can be minimized by lowering the speed of centrifugation (see text). Bar = 1 pn Fig. 11. Fixation in cacodylate-bufferedglutaraldehyde showing marked extraction qf catalase (EXT) with 3,3'-diaminobenzidine reaction product localized to the inner aspect of membrane of severalperoxisomes (+). Bar = 1 pm Fig. 12. Fixation in glutaraldehyde in the presence of Metrizamide showing little extraction of catalase. (*) Peroxisomes uniformly stained for catalase. Bar = 1 p By electron microscopy, the isolated peroxisomes showed a distinct limiting membrane with a finely granular matrix, the features known from well-fixed and processed liver tissue. More than half of all peroxisomes contained the crystalline nucleoids of uricase, which is comparable to the frequency seen in sections of rat liver [36, 371. It should be noted, however, that a few free cores were also found between the isolated peroxisomes and that their number multiplied when the speed

8 264 of centrifugation in Metrizamide was increased. Thus, when the gradient was spun at x g for 60 min, as suggested by Hajra and Bishop [4] a separate band consisting exclusively of free cores was obtained next to the peroxisomal fractions (Fig. 10). This corresponds to the distinct peak in distribution of uricase in peroxisomal fractions obtained by these authors [4]. It is therefore important to adjust the centrifugation speed in Metrizamide to a level which permits optimal separation of peroxisomes in a reasonable time without causing the extrusion of cores and damage to the particles. Extraction of matrical proteins In cytochemical preparations the reaction product of catalase was distributed uniformly over the matrix of most peroxisomes, thus confirming the low level of extraction of matrical proteins. This was also reflected in the latency of catalase determined biochemically which is an indicator for the integrity of isolated peroxisomes [30]. It should be added that peroxisomes after the isolation in Metrizamide are still highly susceptible to further manipulations. Thus, fixation in the routinely used 0.1 M cacodylate buffered glutaraldehyde caused some extraction of catalase. This was prevented by using the gradient buffer and adding 0.48 g/ml Metrizamide to the fixative. The rationale behind this approach was to add the glutaraldehyde to the same medium in which the peroxisomes were isolated in order to prevent any osmotic damage and extraction of matrix proteins. The influence of carrier-buffers and their osmolality in glutaraldehyde based fixatives upon the ultrastructure of cells is well known in electron microscopy of tissues and cells [38, 391, but has received little attention in processing of subcellular fractions. It is of interest that the mean diameter of isolated peroxisomes in the present study was about 0.4 pm which is very close to the values reported for this organelle in situ in liver of normal untreated rats [37, 40, 411. This is probably due to the gentle isolation and processing of peroxisomes in our procedure. Another approach for the preservation of highly fragile peroxisomes was used by Appelkvist et al. [l], who prefixed the crude homogenates of rat liver with a low concentration of glutaraldehyde prior to the final separation on a Percoll gradient. Although this method provided morphologically well preserved peroxisomes, the prefixation with glutaraldehyde affected the activity of some enzymes, especially the P-oxidation system of fatty acids. In addition, their peroxisome fractions displayed a contamination of 13-15% by other cell organelles. Centrifugation procedure The isolation of peroxisomes described here is a two-step procedure involving first the enrichment of the particles by differential pelleting followed by the subfractionation of this crude preparation (D) by density-dependent banding on a linear gradient. Although the recovery rate of this method is relatively low, it was chosen because in our hands it provided the peroxisome fractions with the highest degree of purity. Other approaches such as the direct application of the total homogenate to Metrizamide gradients resulted in increased contamination of peroxisomal fractions as did the subfractionation of preparations obtained by differential banding on step gradients of sucrose or Metrizamide (data not shown). The employment of zonal rotors in combination with Metrizamide as proposed by Connock and Temple [42] may be the method of choice rendering possible even the immediate application of the total homogenates. However, because of the large volume of samples used in zonal rotors and the high price of Metrizamide this approach may prove impracticable. Although our processing of the total homogenate is similar to the classic approach of Leighton et al. [5] our crude fractions, designated B - G, are not identical. The main differences concern the percentage distribution of marker enzymes of peroxisomes (Table 1) and, less significantly, of other organelles. The pretreatment of rats with Triton WR 1339 [5] and the modified centrifugation conditions account probably for these observed differences. In summary, a procedure for the isolation of highly purified and structurally well preserved peroxisomes from the liver of normal untreated rats is described. Such fractions which are virtually free of contamination by other cell organelles should be useful in studies of enzymes with multiple intracellular localization and in the elucidation of the question of biogenesis of peroxisomes. We thank Drs F. Leighton and M. Bronfam (Santiago, Chile) for helpful suggestions in the use of Metrizamide. One of us (A.V.) is grateful to Dr P. Lazarow for his hospitality at the Rockefeller University during a visiting fellowship which was supported by the Deutsche Forschungsgemeinschaft (Vo 317/1-1). We also thank Dr C. Reyero for some of the morphologic observations in the initial phase of this study, Dr K. Beier for morphometric analysis and Dr P. Baudhuin and his staff in Brussels for advice in preparation of pellicles from cell fractions. Excellent technical assistance of Heribert Mohr, Karin Mertzsch, Uschi Graber, Inge Frommer and Gabie Kramer and the secretarial help of Annemarie Achten is gratefully acknowledged. This study was supported by a grant (Fa 146/1-2) from the Deutsche Forschungsgemeinschaft, Bonn-Bad Godesberg, FRG. REFERENCES 1. Appelkvist, E. L., Brunk, U. & Dallner, G. (1981) J. Biochem. Biophys. Methods 5, Bronfman, M., Leighton, F. & Feytmans, E. (1982) Ann. N.Y. Acad. Sci. 386, Connock, M. J., Kirk, P. R. & Sturdee, A. P. (1974) J. Cell Biol. 61, Hajra, A. K. & Bishop, J. E. (1982) Ann. N. Y. Acad. Sci. 386, Leighton, F., Poole, B., Beaufay, H., Baudhuin, P., Coffey, J. W., Fowler, S. & De Duve, Ch. (1968) J. Cell Biol. 37, Neat, C. E. & Osmundsen, H. (1979) Biochem. J. 180, Neat, C. E., Thomassen, M. S. & Osmundsen, H. (1980) Biochem. J. 186, Ishii, H., Suga, T., Hayashi, H. & Ninobe, S. (1979) Biochim. Biophys. Acta 582, Trout, J. (1978) Virchows Arch. B Cell Pathol. 28, Wattiaux, R., Wattiaux-De Coninck, S., Ronveaux-Dupal, M. F. & Dubois, F. (1978) J. Cell Biol. 78, Bronfman, M., Inestrosa, N. C. & Leighton, F. (1979) Biochem. Biophys. Res. Commun. 88, Rickwood, D. (1982) Anal. Biochem. 122, Reyero, Cr., Volkl, A. & Fahimi, H. D. (1982) Biol. Cell. 45,228. Osmundsen, H. (1983) in Zodinated density gradient media, apractical approach (Rickwood, D., ed.) pp , IRL Press, Oxford. 15. Baudhuin, P., Beaufay, H., Rahman-Li, Y., Sellinger, 0. Z., Wattiaux, R., Jacques, P. & De Duve, C. (1964) Biochem. J. 92, Lazarow, P. B. & De Duve, C. (1976) Proc. Nut1 Acad. Sci. USA 73, Beaufay, H., Amar-Costesec, A,, Feytmans, E., Thines-Sempoux, D., Wibo, M., Robbi, M. & Berthet, J. (1974) J. Cell Biol. 61,

9 Sottocasa, G. L., Kuylenstierna, B., Ernster, L. & Bergstrand, A. (1967) J. Cell Biol. 32, Cooperstein, S. J., Lazarow, A. (1951) J. Biol. Chem. 189, Beaufay, H., Bendall, D. S., Baudhuin, P. & De Duve, C. (1959) Biochem. J. 73, Hogeboom, G. H. & Schneider, W. C. (1953) J. Biol. Chem. 204, Priest, D. G. & Pitts, 0. M. (1972) Anal. Biochem. 50, Meisler, M. & Paigen, K. (1972) Science (Wash. DC) 177, Lowry, 0. H., Rosebrough, N. J., Farrand, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193, Beaufay, H., Jacques, P., Baudhuin, P., Sellinger, 0. Z., Berthet, J. & De Duve, C. (1964) Biochem. 92, Baudhuin, P., Evrard, P. & Berthet, J. (1967) J. Cell Biol. 32, Karnovsky, M. J. (1971) Proc. 11th Annu. Meet. Am. Soc. Cell Biol. 51, 146 (abstract). 28. Angermiiller, S. & Fahimi, H. D. (1981) Histochemistry 71, De Duve, C., Pressman, B. C., Gianetto, R., Wattiaux, R. & Appelmans, F. (1955) Biochem. J. 60, Baudhuin, P. (1969) Ann. N.Y. Acad. Sci. 168, Fahimi, H. D. (1969) J. Cell Biol. 43, Waechter, F., Bentley, P., Bieri, F., Staubli, W., Volkl, A. & Fahimi, H. D. (1983) FEES Lett. 158, Volkl, A,, Zaar, K., Stegmeier, K. & Fahimi, H. D. (1984) Japan Soc. Cell Biol. p Sabine, J. R. (1983) in CRC series in enzyme biology (Sabine, J. R., ed.) pp. 4-10, CRC Press, Boca Raton, Florida. 35. Keller, G. A., Singer, S. I., Barton, M. C. & Shapiro, D. J. (1983) J. Cell Biol. 97, 361a. 36. Flatmark, T., Christiansen, E. N. & Kryri, H. (1981) Eur. J. Cell Biol. 24, Staubli, W., Schweizer, W., Suter, J. & Weibel, E. R. (1977) J. Cell Bid. 74, Fahimi, H. D. & Drochmans, P. (1965) J. Microsc. 4, Hayat, M. A. (1970) Principles and techniques of electron microscopy: biological applications, vol. 1, Van Nostrand Reinhold Comp., New York. 40. Bock, P., Kramar, R. & Pavelka, M. (1980) Cell Biol. Monugr. 7, Hruban, Z. & Rechcigl, M., Jr (1969) Int. Rev. Cytol., Suppl. 1, p Connock, M. J. &Temple, N. J. (1983) Int. J. Biochern. 15,