The preparative isolation of mitochondria from Chinese hamster ovary cells

ANALYTICAL

BIOCHEMISTRY

163, 350-357

The Preparative

(1987)

Isolation of Mitochondria Hamster Ovary Cells’

EDWARD A. MADDEN

AND BRIAN

from Chinese

STORRIE’

Department of Biochemistry and Nutrition, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061 Received August 7, 1986 A “hybrid” discontinuous gradient consisting of 6% Percoll overlaid on metrizamide separated mitochondria from other organelles in a Chinese hamster ovary cell postnuclear supematant in a single I5-min centrifugation. The mitochondrial preparation contained about 25% of the mitochondrial marker, cytochrome-coxidase. in a form that was about 90% latent. Based on the postnuclear supematant, cytochrome-c oxidase activity was enriched approximately 45fold. Trace amounts of lysosomal, rough endoplasmic reticular, Go@, peroxisomal, plasma membrane, and cytosolic markers were found in the preparation. Electron microscopy revealed that the preparation consisted almost exclusively of mitochondria with only minor amounts of contaminating organelles. Analysis of the mitochondrial preparation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis demonstrated that the mitochondrial preparation had a unique protein profile compared to the postnuclear supernatant and other gradient interfaces. Separation of the mitochondria into membrane and lumenal (matrix) fractions by treatment with 100 mM Na2C03, pH 11.5, also indicated that the mitochondria were intact; they were rich in lumenal proteins. The data indicate that the mitochondria represent maximally about 2.2% of Chinese hamster ovary cell postnuclear supematant protein. These isolated mitochondria should prove useful for problems in molecular cell biology. 0 1987 Academic Press, Inc. KEY WORDS: mitochondria; mitochondrial membrane; subcellular fractionation; cell organelles; metrizamide; Percoll.

Several methods are available for the isolation of mitochondria from animal tissues such as liver. These methods include differential centrifugation and sucrose-densitygradient centrifugation [for a review see (l)]. Although these methods are satisfactory for the purification of mitochondria from animal tissues, mitochondria remain difficult to separate from tissue culture cells. Purification of tissue culture cell mitochondria would simplify many problems in the molecular cell biology of mitochondria. In Percoll gradients mitochondria colocahze with lysosomes (2). In metrizamide gradients

mitochondria can be separated from lysosomes, but the mitochondria are often contaminated with a fair amount of plasma membrane (3). Our strategy has been to combine the ability of Percoll to separate the plasma membrane and microsomal membranes from mitochondria and lysosomes with the ability of mettizamide to separate mitochondria from lysosomes. By overlaying Percoll on a metrizamide step gradient, we have created a “hybrid” gradient in which mitochondria can be rapidly isolated from cultured animal cells.

’ Supported by PHS Grant GM-28188. Dedicated to Professor Guiseppe Attardi on the 20th year of his research on the molecular biology of animal cell mitochondria. * To whom correspondence should be addressed.

Materials. Analytical grade metrizamide was manufactured by Nyegaard (Oslo, Norway) and was purchased from Accurate Chemical Co. Percoll was purchased

0003-2697187

$3.00

Copyright 0 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

MATERIALS

350

AND METHODS

PREPARATIVE

MITOCHONDRIAL

from Pharmacia Fine Chemicals. Reagents for SDS-PAGE3 were of electrophoresis grade and were purchased from Polysciences. 4 - Methylumbelliferyl - 2 - acet amido - 2 - deoxy - 8 - D - glucopyranoside, 4methylumbelliferyl - 0 - D - galactopyranoside, and 4-methylumbelliferyl-a-D-mannopyranoside were manufactured by Koch-Light Laboratories and purchased from Research Products International. All other chemicals were reagent grade or better. Cells and culture. CHO-S (C2) cells were grown in suspension culture in aMEM media supplemented with 10% heat-inactivated fetal calf serum as described by Pool et al. (4). Cell number was determined with a hematocytometer. Labeling conditions. To metabolically label cellular RNA, CHO cells were cultured in the presence of [2-14C]uridine (0.02 &i/ ml, 5 1.2 mCi/mmol) for 2 1 h. Cellular protein was metabolically labeled by exposing CHO cultures to [35S]methionine (6 pCi/ml, I 13 Ci/mmol) for 2 days. Homogenization and isolation of PNS. All the procedures were performed at 4°C. Cells were disrupted by nitrogen cavitation and further homogenized by four strokes with a Potter-Elvehjem homogenizer (5). The homogenate was centrifuged at 13OOg,, for 5 min. The supernatant was decanted and placed on ice. The nuclear pellet was resuspended in 0.25 M sucrose and centrifuged at 13OOg,, for 5 min. This step was repeated and the supernatants from each wash were pooled to give a total PNS.

Formation of the hybrid Percoll/metrizamide discontinuous gradient. A schematic presentation of the hybrid gradient is given in Fig. 1. All solutions were in 0.25 M sucrose

’ Abbreviations used: SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; CHO, Chinese hamster ovary; aMEM, Eagle’s minimum essential medium, a-modification without ribonucleosides and deoxyribonucleosides; PNS, postnuclear supernatant; TCA, trichloroacetic acid.

ISOLATION

351

and expressed as w/v. Densities were determined by weighing 1 ml aliquots. Gradients were prepared in cellulose nitrate tubes for the Beckman SW 40 rotor. Two milliliters of 35% metrizamide was overlaid with 2 ml of 17% metrizamide followed by 5 ml of 6% Percoll. The tubes were then gently filled to within 1 mm of the top with PNS (about 4.75 ml). Solutions were overlaid with an 18 gauge syringe needle. Centrifugation was standardly performed for 15 min at 50,5OOg,, (20,000 rpm) in a Beckman L5-50B ultracentrifuge. The slowest possible acceleration rate was used and the brake was set to slow.

Assay of cytochrome-c oxidase. Cytochrome-c oxidase was assayed by a modification of the procedure of Attardi et al. (6). The substrate, horse heart cytochrome c (Sigma, type III), was reduced with sodium dithionite to an AS50/A5,5between 6.0 and 9.0. The assay mixture consisted of 0.0 1-O. 1 ml of sample mixed in a cuvette with 1.O ml of 18 PM reduced cytochrome c in 50 IIIM potassium phosphate buffer, pH 6.2. Sucrose (0.25 M) was added, as needed, to bring the volume to 1.1 ml. The detergent Lubrol PX (0.2%, final concentration) was routinely included in the reaction mixture to permeabilize the mitochondria (7). The decrease in Assoat 25°C was measured for 2 min in a Response spectrophotometer (Gilford Systems). The reaction blank contained 0.25 M sucrose instead of sample. The decrease in A&min due to sample settling was determined by omitting cytochrome c from the reaction mixture. The maximum contribution of sample settling to apparent cytochrome-c oxidase activity was 4%. For the determination of the latency of cytochrome-c oxidase, gradient fractions were assayed for enzyme activity in the presence or absence of Lubrol PX in a reaction mixture containing 0.25 M sucrose in 50 mM potassium phosphate buffer, pH 6.2. Latency (%) was defined as [(activity in the presence of Lubrol - activity in the absence of Lu-

352

brol)/activity x 100.

MADDEN

AND STORRIE

in the presence of Lubrol]

Analysis of organellar markers. The following markers were assayed: plasma membrane, alkaline phosphodiesterase I (8); lysosomal, ,&hexosaminidase (4) and ,f3-galactosidase (5); Golgi apparatus, a-mannosidase II, pH 7.4 (4,9); cytosolic, lactate dehydrogenase (10); and peroxisomal, catalase (11). The presence of the rough endoplasmic reticulum was indicated by long-term incorporation of [14C]uridine into acid precipitable material (12). Protein was determined fluorometrically by the fluorescamine procedure (13,14). Percoll had no effect on any of the marker activities tested. The metrizamide concentration in the enzyme assay mixtures was always less than 1.5%. Metrizamide was found to inhibit @-hexosaminidase, P-galactosidase, lactate dehydrogenase, and cr-mannosidase II activities if the final metrizamide concentration was greater than 2%. The recoveries of all the markers, including cytochrome-c oxidase, ranged from 80 to 107% (Table l), and were not effected by freezethawing. Electron microscopy of the mitochondrial preparation. An equal volume of 5% glutaraldehyde in 0.2 M cacodylate, pH 7.4, was added to the material recovered at the 17/35% metrizamide interface. The primary fixation was for 18 h at 4°C. An equal volume of 1% Os04 in 0.1 M cacodylate, pH 7.4, was then added and the preparation was incubated for 20 min at 4°C. The organelles were pelleted at 53,OOOg,, (20,000 rpm) for 20 min in a SW 27 rotor. The pelleted material was then washed 3 times with 0.1 M cacodylate, dehydrated, and embedded in Spurr’s low viscosity resin. The sections were poststained with lead citrate and uranyl acetate and viewed in a Zeiss EM IO-CA electron microscope at an accelerating voltage of 80 kV. SDS-PAGE. The CHO cells, metabolically labeled with [35S]methionine, were pooled with 3 vol of unlabeled cells. Immedi-

ately after the discontinuous gradient centrifugation, phenylmethylsulfonyl fluoride was added to a final concentration of 2 mM from a 200 mM stock in isopropanol to the material of each interface and a sample of PNS. A small aliquot of the mitochondrial preparation was saved. The remainder of the mitochondrial preparation was pelleted at 20,000 rpm for 20 min in a SW 27 rotor. The pellet was suspended in 10 ml of 100 mM Na2C03, pH 11.5, and the mitochondrial preparation was carbonate fractionated as described ( 15). The SDS-PAGE formulation of Dreyfuss et al. (16) was followed. The resolving gel was 8% polyacrylamide. Gel lanes were loaded with 4500 cpm/well and were electrophoresed at 30 mA constant current. The gels were processed for fluorography with Amplify (Amersham Corp.) according to the manufacturers specifications. RESULTS

A total PNS (about 60% of total cell protein) was prepared from CHO cells that were disrupted by N2 cavitation. The PNS was applied to a hybrid Percoll/metrizamide gradient. After centrifugation for 15 min at 20,000 rpm in an SW 40 rotor, four distinct bands were seen in the hybrid Percoll/metri-

FIG. 1. Schematic representation of the hybrid Percoll/metrizamide density gradient for the isolation of mitochondria from Chinese hamster ovary cells. The density solutions (expressed as w/v in 0.25 M sucrose) were layered into SW 40 rotor tubes (Beckman) as described under Materials and Methods. The gradients were centrifuged for at least 15 min at 20,000 rpm (50,5OOg,,). For an explanation ofthe interface labels see the Results. Mz, me&amide; S, sucrose; P, Percoll.

PREPARATIVE

MITOCHONDRIAL TABLE

RECOVERY(% OF PNS)OFVARIOUSMARKERS

353

ISOLATION

1

FROMTHEHYBRIDPERCOLL/METRIZAMIDEGRADIENT

Interface Organellar marker

Top

VP

P/17

17135

Cytochrome-c oxidase fl-Hexosaminidase @-Galactosidase Alkaline phosphodiesterase I a-Mannosidase II Catalase Protein RNA Lactate dehydrogenase

ND” 9.7 10.4 18.2 31.3 53.8 74.2 76.1 84.7

9.0 31.9 29.1 72.8 42.2 44.7 21.8 27.4 12.8

48.9 41.8 39.5 5.1 3.5 9.4 4.1 1.5 ND

22.8 0.7 0.8 1.2 ND 0.2 0.5 0.7 ND

n ND, not detected.

zamide discontinuous gradient (Fig. 1): Floating material was observed at the top of the gradient and at the three gradient interfaces; sucrose/Percoll (S/P), Percoll/ 17% metrizamide (P/17%), and 17% metrizamide/35% metrizamide (17/35%). Analysis of the interfaces for marker enzyme activities indicated that approximately 25% of the mitochondrial marker, cytochrome-c oxidase, was recovered at the 17/35% interface of the gradient (Table 1). This interface contained only minor amounts of the marker enzymes for lysosomes and plasma membrane (Table 1). With shorter centrifugation times the yield of mitochondria at the 17/35% interface was decreased. No increase in mitochondrial yield was observed with up to 1.5 h of centrifugation. Longer centrifugations resulted in decreased resolution. In three separate experiments, the recoveries of mitochondria in the 17/35% interface ranged from 23 to 28%. Relative to the PNS, the mitochondrial preparation was enriched from 43.8- to 45.4-fold (Fig. 2) and thus represented maximally about 2.2% of the total PNS protein. Cytochrome-c oxidase was from 84 to 96% latent in the mitochondrial preparation as determined by activity in the presence and absence of Lubrol. Cytochrome-c oxidase activity was about 93%

latent in the PNS, 47% latent at the S/P interface, and 92% latent at the P/l 7% interface. Electron microscopic examination of the mitochondrial preparation indicated that almost all of the organellar profiles were mitochondria (Fig. 3). Most of the mitochondrial profiles appeared to be intact. At higher magnifications most of the mitochondrial profiles contained an intact double membrane (not shown).

FIG. 2. Analysis of the mitochondrial preparation. The specific activities are relative to the PNS. The markers were assayedas described under Materials and Methods.

354

MADDEN

AND

STORRIE

FIG. 3. Morphology of the mitochondria preparation. The preparation was fixed with 2.5% glutaraldehyde in 0.1 M cacodylate, pH 7.4, for I8 h at 4’C, followed by 0.5% OS04 for 20 min at 4°C. The sections were poststained with lead citrate and uranyl acetate. The bar represents 1.0 pm.

PREPARATIVE

MITOCHONDRIAL

As an additional test of the purity of the mitochondrial preparation, the SDS-PAGE pattern of the mitochondrial fraction was compared to that of the other gradient interfaces as well as the PNS (Fig. 4). About 20 bands were visible in the mitochondrial preparation (lane E) with major proteins at 20, 43, 50, 67, and 112 kDa. There was a cluster of prominent bands between 32 and 39 kDa. The resemblance of the total mitochondrial pattern to the P/ 17% (lane D) pattern was probably due to the high amount of mitochondria in this interface (see Table 1). Indeed, electron microscopy indicated that

A BCDEFG

FIG. 4. SDS-PAGE of the PNS (A) and material isofrom the top(B), S/P (C), and P/i 7% (D) interfaces of the hybrid Percoll/metrizamide gradient. The mitochondrial preparation (E) was separated into membrane (F) and soluble (G) proteins by fractionation with 100 mM Na*CO-,, pH Il.5 (15). The resolving gel was prepared from a stock of 33.5% acrylamide and 0.3% N,N’bisacrylamide to a final concentration of 8% acrylamide (16). The cultures were labeled with [3SS]methionine 2 days prior to the isolation of the PNS and gradient centrifugation as described under Materials and Methods. The gels were loaded with 4500 cpm/lane and fluorography was performed with Amplify (Amersham). The M, markers were carbonic anhydrase, ovalbumin, bovine serum albumin, phosphorylase b, Bgalactosidase, and myosin (in order of increasing MJ.

ISOLATION

355

mitochondria represented roughly 50% of the profiles in P/17% preparation (not shown). The resemblance of the top (lane B) and S/P (lane C) interface patterns to the PNS pattern (lane A) was probably due to these interfaces containing 96% of the PNS protein (Table 1). As an additional test of the intactness of the mitochondrial preparation, the mitochondria were separated into membrane and lumenal fractions by pH 11.5 carbonate treatment (15). If the mitochondria are intact, then lumenal proteins should be abundant and may well represent the bulk of the total mitochondrial proteins. As shown in Fig. 4, lanes E-G, the lumenal protein pattern appeared identical to the total protein pattern, and there was little overlap with the membrane fraction. Approximately 30% of the total mitochondrial TCA-precipitated [35S]methionine radioactivity was found in the membranous fraction. The mitochondrial membrane fraction displayed prominant proteins at 20,22,23,26,53,68,98, and 109 kDa. Fujiki et al. utilized Na2C03 to isolate mitochondrial membranes from purified rat liver mitochondria (17). The CHO cell mitochondrial membrane and soluble fractions appeared similar to the same fractions obtained from rat liver mitochondria. In both the rat liver and CHO cell fractionations the soluble protein profile closely resembled the total mitochondrial protein profile.

lated

DISCUSSION

Hybrid Percoll/metrizamide gradient centrifugation provides a rapid procedure for the isolation of mitochondria from a tissue culture cell line. By marker analysis and electron microscopic examination the mitochondrial preparation appeared highly enriched. When the cytochrome-c oxidase specific activity of the mitochondrial preparation was compared to that of the PNS, an enrichment of about 45-fold occurred. In

356

MADDEN

AND STORRIE

CHO cells mitochondria appear to be maximally about 2.2% of the PNS protein. The amount of CHO cell mitochondria was no doubt affected by the fact that the cells were grown in suspension cultures (4). Tissue culture cell lines are often grown in conditions which are more hypoxic than conditions in vivo. Thus it could be expected that cultured cells would contain less mitochondria than tissues. Mitochondria compose 39% of the liver hepatocyte ( 18) and 2 1% of the pancreatic exocrine cell (19) total cell membrane surface area. To the best of our knowledge this is the first report of mitochondrial isolation from a tissue culture cell iine. These isolated mitochondria should prove useful for problems in molecular cell biology. It is noteworthy that the mitochondria from CHO cells can be separated from lysosomes by sedimentation through metrizamide. Rat liver mitochondria cannot be sep arated from lysosomes by simple sedimentation through metrizamide (3). Mitochondria and lysosomes were the only major organelles to transverse the 6% Percoil layer of the hybrid Percollfmetrizamide gradient. Work from other laboratories (5) as well as our own (unpublished observations) has indicated that under prolonged centrifugation conditions in a fixed angle or a vertical rotor peroxisomes also sediment through Percoll. The failure of peroxisomes to sediment through the 6% Percoll layer of our discontinuous gradient is presumably due to their small size. Percoll gradients do not self-form very well in swinging-bucket rotors (20). Thus, under our conditions Percoll separates organelles by sedimentation velocity rather than by density. The yield of mitochondria can be increased by flotation of the P/ 17% interface of the gradient which contains a mixture of mitochondria and lysosomes through a second discontinuous metrizamide gradient to obtain a second mitochondrial fraction. However, the purity of the second mitochondrial fraction is less than that of the mitochondrial

preparation isolated from the hybrid gradient due to contamination by lysosomes, peroxisomes, and plasma membrane. Centrifuge-grade metrizamide can be substituted for analytical-grade metrizamide for the isolation of mitochondria, but the centrifuge grade is much more difficult to dissolve at higher concentrations in 0.25 M sucrose. Preliminary evidence suggested that Nycodenz can be substituted for metrizamide (B. Storrie and L. Hastings, unpublished observation). Nycodenz costs less and can be autoclaved. It would seem likely that a dense sucrose solution could be substituted for the 35% metrizamide layer of the hybrid gradient. These rapidly isolated mitochondria with their intact membranes are likely to exhibit tight respiratory control. However, this remains to be demonstrated. ACKNOWLEDGMENTS The authors thank Andrea Ferris for her excellent technical assistance and Dr. R. Grayson and Nancy Rechcigl of the Electron Microscopy Laboratory, Callege of Agriculture and Life Sciences for their assistance.

REFERENCES 1. Pedersen, P. L., Greenawalt, J. W., Reynafarje, B., Hullihen, J., Decker, G. L., Soper, J. W., and Bustamente, E. (1978) in Methods in Cell Biology (Prescott, D. M., Ed.), Vol. 20, pp. 41 l-481, Academic Press, New York. 2. Stonie, B., Wirt, J. B., Mah, V. T., and Maurey, K. M. (1986) Cell Biol. ht. Rep. 10,49-53. 3. Wattiaux, R., Wattiaux-De Coninck, S., RonveauxDupal, M., and Dubois, F. (1978) J. Cell Biol. 78, 349-368. 4. Pool, R. R., Jr., Maurey, K. M., and Storrie, B. (1983) Cell Biol. ht. Rep. 7, 361-367. 5. Rome, L. H., Garvin, A. J., Allietta, M. M., and Neufeld, E. F. (1979) Cell 17, 143-153. 6. Attardi, B., Cravioto, B., and Attardi, G. (1969) J. Mol. Biol. 44,47-70. 7. Rafael, J. (1983) in Methods of Enzymatic Analysis (Bergmeyer, H. U., Ed.), Vol. 3, pp. 266-273, Verlag Chemie, Weinheim. 8. Aronson, N. N., Jr., and Touster, 0. (1974) in Methods in Enzymology (Fleischer, S., and

PREPARATIVE

9. 10. 11. 12. 13.

MITGCHONDRIAL

Packer, L., Eds.), Vol. 3 1, pp. 90-102, Academic Press, New York. Tulsiani, D. R. P., Hubbard, S. C., Robbins, P. W., and Touster, 0. (1982) J. Biol. Chem. 257, 3660-3668. Decker, L. A. (1977) Worthington Enzyme Manual, p. 20. Worthington Biochemical Corp., Freehold, NJ. Leighton. F., Poole, B., Beaufay, H., Baudhuin, P., Coffey, J. W., Fowler, S., and De Duve, C. (1968) J. Cell Biol. 37, 482-5 13. Hubbard, A. L.. and Cohn, Z. A. ( 1975) J. Cell Biol. 64.438-460. BBhlen, P., Stein, S.. Dairman, W., and Udenfriend, S. (I 973) Arch. Biochem. Biophys. 155,2 13-220.

ISOLATION

357

14. Udenfriend, S., Stein, S., Biihlen, P., Dairman, W., Leimgruber, W., and Weigele, M. (1972) Science 178,87 l-872. 15. Fujiki, Y ., Hubbard, A. L., Fowler, S., and Lazarow, P. B. (1982) J. Cell Biol. 93, 97-102. 16. Dreyfuss, G., Adam, S. A., and Choi, Y. D. ( 1984) Mol. Cell. Biol. 4. 4 15-423. 17. Fujiki, Y., Fowler, S., Shio, H., Hubbard, A. L., and Lazarow, P. B. (1982) J. Cell Biol. 93, 103-l 10. 18. Weibel, E. R., Sttiubli, W., Gnagi, H. R., and Hess, F. A. ( 1969) J. Cell Biol. 42,68-9 1. 19. Bolender, R. P. (1974) J. Cell Biol. 61,269-287. 20. Pertoft, H., Laurent, T. C., figs, T., and Kagedal, L. ( 1978) Anal. Biochem. 88.27 l-282.