Proc. Natl. Acad. Sci. USA Vol. 74, No. 9, pp. 3730-3734, September 1977
Biochemistry
Animal cell mutants defective in sterol metabolism: A specific selection procedure and partial characterization of defects (sterol biosynthesis/polyene antibiotic/membrane biogenesis)
Y. SAITO, S. M. CHOU, AND D. F. SILBERT Department of Biological Chemistry, Division of Biology and Biomedical Sciences, Washington University, St. Louis, Missouri 63110
Communicated by P. Roy Vagelos, June 20, 1977
By using a chemically defined medium, a ABSTRACT general and highly specific procedure was devised to select for mutant cells with less abundant or structurally altered sterol in their surface membranes. Within a certain concentration range, the polyene antibiotic filipin was shown to kill only cells with normal (as opposed to decreased) membrane sterol levels. Sterol-requiring derivatives of LM cells were isolated by chemical mutagenesis, filipin treatment, and cloning followed by replica plating in soft agar. Mutants (SI and S2) are described which, when compared to normal cells, show decreased synthesis of desmosterol in vivo from acetate and mevalonate relative to cell number or to fatty acid synthesis. When exogenous sterol is supplied, mutants SI-and S2 grow normally in suspension culture. However, when deprived of sterol supplement, mutant SI grows slower than wild type cells and mutant S2 lyses within one to two generations. Gas/liquid chromatography revealed that the mutants contained a normal spectrum of fatty acids including ubsaturated fatty acyl groups but, unlike wildtype cells, they have less abundant (mutant SI) or no (mutant S2) desmosterol in either the presence or absence of exogenous cholesterol. In vitro experiments with mevalonate as the substrate suggest that the defect in both mutants is in a demethylation reaction subsequent to lanosterol synthesis. The selection method developed here may permit the isolation of mutants with defective membrane incorporation of sterols and other polyisoprenoids as well as defective synthesis of these compounds. Sterols are present in membranes of virtually all eukaryotic cells. An essential role for sterols in the functional integrity of such membranes is indicated by the fact that eukaryotic cells lyse after a generation or two in medium devoid of sterol when sterol synthesis is impaired by oxygenated cholesterol analogues (1, 2) or by genetic mutation (3). Studies with model as well as natural membranes suggest that sterol molecules modulate fatty acyl chain interactions (4, 5). The consequence of this modulation on interactions between phospholipid molecules and between phospholipids and proteins is not fully understood. Furthermore, the preferential concentration of sterols in surface membranes and the mechanism for translocation of sterols from their site of synthesis to the surface membrane have not been
explained.
In this paper, we describe a general and highly selective procedure for isolation of animal cell mutants with membranes containing decreased levels of or structurally altered sterols, and we characterize two mutants with decreased sterol synthesis. Many asts of sterol and polyisoprenoid metabolism and of membrane biochemistry may be probed by this approach.
MATERIALS AND METHODS Materials. All the labeled compounds and sterols were purchased from New England Nuclear and Steraloids, reThe costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.
spectively. Filipin was obtained from G. B. Whitfield, Upjohn Co., via S. C. Kinsky. Lipoprotein-deficient serum and low density lipoprotein were provided by S. Weidman. Cell Growth. LM cells (mouse fibroblasts) were purchased from the American Type Culture Collection (CCL 1.2) and were propagated in suspension culture or in monolayers. Minimal growth medium was Higuchi's medium (6) modified by the addition of 20 mM N-2-hydroxyethylpiperazine-N'2-ethanesulfonic acid (Hepes) buffer (final pH 7.4), 0.5% bovine serum albumin, and 0.015% Darvan no. 2 (R. T. Vanderbilt Co.) (to minimize cell clumping or attachment to the walls of the culture bottle). The suspension culture (25 ml in a 125-ml Wheaton screw-cap bottle) was agitated on a gyratory shaker (140 rpm) at 370. Growth was monitored by measuring cell number in a hemocytometer, and cell densities ranged generally from 0.5 to 3 X 106 cells per ml. Generation time in the minimal medium was about 24 hr. When cells were grown as monolayers or cloned in soft agar (7), a modified Eagle's medium (Flow Laboratories) supplemented with penicillin (50 units/ml) and streptomycin (50,ug/ml) was used and the cultures were grown in a CO2 incubator. Isolation of Mutants with Decreased Membrane Desmosterol Levels. LM cells were mutagenized by incubating cells with N-methyl-N'-nitro-N-nitrosoguanidine (20 ,uM) for 30 min at 370 in minimal medium supplemented with 10% calf serum. Based on the number of clones formed in soft agar before and after treatment, about 10% of the cells survived. The mutagen was washed out and the cells were incubated for several generations in fresh medium to allow for segregation of chromosomes and expression of mutant phenotypes. At this point the cells were transferred to minimal medium and grown for 1 to 1 'k generations so that the membrane sterol of sterol synthesis mutants would decrease below that of normal cells. Filipin in dimethylformamide was diluted into minimal medium and then into the cell suspension to give a final concentration of 5-6 The cells were incubated for 30 min at 370, washed, jgg/ml. resuspended in fresh minimal medium containing 10% calf serum, and grown for several generations. The antibiotic treatment was repeated once; after growth had recovered, the cells were cloned in soft agar containing 10% calf serum. The resultant clones were replica plated by transferring cells with a glass rod onto a set of soft agar plates, each set containing a plate supplemented with 0.5% bovine serum albumin, one supplemented with bovine serum albumin plus cholesterol (20 Clones ,ug/ml), and one supplemented with 10% calf serum.taken as that grew well only with the sterol supplement were putative sterol-requiring mutants. These derivatives were propagated in monolayer cultures and transferred by trypsinization to suspension cultures for biochemical studies and additional growth experiments. After confirmation of their phenotype by these in vivo measurements of sterol synthesis, sterol mutants were recloned before more detailed characterization of the specific biochemical defect. 3730
Biochemistry:
Saito et al.
Proc. Natl. Acad. Sci. USA 74 (1977)
Table 1. Regimen for selection of sterol-requiring mutants -48 to 0 hr 0 to 24 hr 30 min Afterward
4:100
0
Control
+C`
Cholesterol, 1 Mg/mI 2
3
4
5
Filipin,
ag/mI
6
7
8
9
of culture
FIG.
1. Development of resistance to filipin accompanying inhibition of sterol synthesis in growing cultures. In one experiment, LM cells in the minimal medium were preincubated with 25-hy-
(M),
droxycholesterol (1 gg/ml) for 0 (x), 5 (0), 10 and 24 (o) hr and treated with filipin at the concentrations shown. In another experiment (broken lines) control cells and cells with decreased sterol
(-)
level (@) were prepared as described in Table
min.
and treated with filipin
The incorporation of thymidine (dThd) into DNA was
then monitored by incubating cells with [3H]thymidine (1 mM, 100 Ci/mol) at 370 for 1 hr without washing out the filipin.
Analytical Procedures. Lipids were extracted by the method of Bligh and Dyer (8) from cells washed thoroughly with
phosphate-buffered saline (9). Polar and nonpolar lipids in the extracts were separated by Unisil silica gel column chromatography, and nonpolar lipids were subdivided into a saponifiable and a nonsaponifiable fraction. The latter lipids are largely sterols. Gas/liquid chromatography was used for the quantitative analysis of sterols and fatty acids. Sterols were
chromatographed with coprostanol as the internal standard on a 3% OV 17 column at
250°;
Filipin Calf serum, 10%
With Cholesterol, 25-OH-Cho- Filipin Calf serum, 10% decreased 20 ug/ml + 25-OHlesterol, sterol 1 ,g/ml + 25-OHCholesterol, 1
.C
for 30
Cholesterol, No sterol 20 Ag/ml
\
60
0
3731
fatty acids were separated on a
10% diethylene glycol succinate column at 1650
with pentadecanoic acid as the internal standard. Protein was determined by the method of Lowry et al. (10).
RESULTS Enrichment Method. A polyene antibiotic, filipin, is known to interact with certain kinds of sterols in biological as well as artificial membranes and to alter membrane properties (11, 12). In fact, the antibiotic is cytotoxic to cells and this effect depends on the presence of sterols in the cell membranes. Prokaryotic cells that do not contain sterols are not affected by filipin. If cells with decreased membrane sterol content could be shown to be
differentially resistant, resistance to filipin could be used to select mutants defective in sterol synthesis and in other mechanisms required for maintaining normal membrane sterol levels. To explore this approach we prepared cells with decreased membrane sterol content by utilizing oxygenated derivatives of cholesterol, such as 25-hydroxycholesterol, which are known to inhibit sterol synthesis specifically but not to substitute for the normal membrane sterol (13). Growth of LM cells with 25-hydroxycholesterol (1 for 24 hr led to a 5-fold de-
,ug/inl)
crease in the incorporation of radioactive acetate into sterols relative to that observed with cells grown without inhibitor but did not affect the amount of radioactivity incorporated from acetate into phospholipids (data not shown). By using appropriate analytical methods for determining mass, the amount
of cellular sterol relative to protein was found to decrease progressively during growth in the presence of the sterol analogue whereas total phospholipid versus protein remained unchanged (data not shown). Inhibition of sterol synthesis leads to a re-
sg/ml
duction in protein and nucleic acid synthesis after one to two generations of growth and thereafter to cell death. These long-term effects of treatment with 25-hydroxycholesterol were prevented by supplying cholesterol (in the form of calf serum or bovine serum albumin complex) to the cells before they had been cultured for more than one generation with the oxygenated sterol (data not shown). Fig. 1 shows that decreased membrane sterol content results in relative resistance of the cells to the action of filipin. The incorporation of thymidine into DNA by whole cells is a convenient way to quantitate the action of filipin (one can also monitor this process by measuring the incorporation of leucine into protein or by following morphological changes of the cells with an inverted phase microscope). When LM cells were grown with 25-hydroxycholesterol, the amount of cellular sterol [i.e., desmosterol (14)] decreased to 14.0, 10.9, 9.6, and 6.8 ,ug/mg of protein after 0, 5, 10,Iand 24 hr (slightly more than one generation), respectively. The concentration of filipin required for 50% inhibition of thymidine incorporation increased from 4.5 to more than 9 ,g/ml from the beginning to the end of this growth period. Next, we subjected LM cells grown without or with the inhibitor of sterol synthesis to conditions identical to those that would be used in the isolation of sterol-requiring mutants (Table 1). The two populations of cells were grown in a medium containing cholesterol supplement for at least two generations, shifted to medium devoid of sterol and propagated for another generation, treated briefly with filipin, and returned to a medium supplemented with calf serum. Under the conditions used in this experiment, the cells grown with 25-hydroxycholesterol were highly resistant to filipin at 5 ,g/ml whereas the control populations showed little or no resistance. This conclusion is based not only on the effect of filipin on thymidine incorporation determined immediately after treatment (Fig. 1, broken lines) but, more significantly, also on the survival of the cells (Table 2). The number of cells in the population containing normal membrane sterol content fell more than 100-fold by 1 day after treatment and did not recover over the succeeding Table 2. Recovery of cell growth after exposure to filipin Time after Cell count X 10-4 per ml exposure to With filipin (5,ig/ml), hr Control decreased sterol 0 127 140 24 1 84 48 0 114 72 0 123 The two populations of cells treated as shown in Table 1 were treated with filipin as described in Materials and Methods and placed in fresh medium containing 10% calf serum with or without 25-hydroxycholesterol (1 sg/ml) to restore growth. The cell counts are means of at least two determinations that varied by no more than 109.
3732
Biochemistry:
Saito et al.
Proc. Natl. Acad. Sci. USA 74 (1977)
Table 3. In vivo synthesis of sterol measured by incorporation of acetate and mevalonate into nonsaponifiable lipids Mevalonate Acetate incorporation incorporation Into polar lipids Into sterol, into sterol, dpm X 103 dpm X 104 per cell dpm X 104 per cell Cell* per cell 38 77 123 2.1 68 136 Si 87 0.2 10 S2 The cells maintained in medium containing 10% calf serum were washed and transferred to the minimal medium and grown for 24 hr. They were then incubated for an additional 24 hr with [14C]acetate (0.13 mM, 2.45 Ci/mol) and [3H]mevalonate (0.1 mM, 5.05 Ci/mol). The cells were washed thoroughly with phosphate-buffered saline. The lipids were extracted and fractionated. * WT, wild type; Si and S2, mutants. WT
2 days but that of cells with decreased sterol content decreased only 40% and growth of the survivors was apparent over the next 2 days. Hence, if one chose the appropriate concentration of filipin and selected the optimal time after shifting from permissive to nonpermissive (i.e., no sterol supplement) conditions for treatment with the polyene antibiotic, one could use this approach to select effectively for mutant cells with reduced membrane sterol levels. Isolation of Strains with Decreased Sterol Synthesis. Putative sterol-requiring derivatives of LM cells were isolated by using nitrosoguanidine mutagenesis, the filipin enrichment procedure, and the cloning and replica plating techniques described above. In the experiment shown in Table 3, radioactive acetate was used to measure sterol and fatty acid synthesis in wild-type cells and in two mutants presumably defective in sterol synthesis, one generation after shifting the cells from sterol-supplemented to minimal medium. Because acetate derived from a common pool is used for sterol and fatty acid synthesis, the data in Table 3 (columns 2 and 3) suggest that sterol synthesis is impaired but fatty acid synthesis is stimulated in the mutants relative to the wild-type cells. If sterol synthesis is compared to fatty acid synthesis, the defects in the mutants are actually magnified. The incorporation of acetate into the sterol fraction in mutant S2 was much less than that in mutant Si. In order to quantitate the incorporation of acetate into desmosterol in vivo, a separate experiment was carried out in which [3Hldesmosterol was added to the nonsaponifiable lipids (labeled with "4C) and this fraction was further resolved by preparative gas/liquid chromatography. The amount of radioactivity in desmosterol derived from mutants S1 and S2 was 41% and 0.1% of the wild-type level, respectively. Radioactive mevalonate incorporation into sterols was also markedly decreased in the mutants (Table 3, column 4), confirming the results with radioactive acetate and, furthermore, demonstrating that mutants S1 and S2 were defective in the sterol synthesis pathway distal to the hydroxymethylgluctaryl-CoA reductase step. The decrease of sterol synthesis in the mutants as measured by the incorporation of mevalonate appeared to be greater than that observed by acetate labeling. This finding suggests that either these mutants do not take up mevalonate from the medium as efficiently as wild-type cells do or they have increased levels of the reductase which results in greater synthesis of mevalonate and, in this experiment, a lower specific radioactivity of the cellular mevalonate pool. This latter possibility might be expected as a consequence of the lowered level of sterols in the mutants, and it is currently under investigation.
0~~~ 0
E
10
2 3 4 5 6 Days of growth FIG. 2. The growth of the wild-type and the two mutant cells was followed in the presence of lipoprotein-deficient serum (30 gg of protein and 0.04 jig of cholesterol per ml of culture) (solid line, open symbols) or in the presence of low density lipoprotein (30 ;g of protein and 34 ,ug of cholesterol per ml of culture) (broken line, solid symbols). O and *, Wild-type cells; 0 and *, mutants Si; A and *, S2. The cell counts are means of at least two determinations that varied by no more than 10%.
Growth Properties and Sterol Composition of Mutants S1 and S2. Growth of wild-type LM cells and of mutants SI and S2 was followed in suspension cultures with minimal medium or minimal medium supplemented with cholesterol, lipoprotein-deficient serum, or low density lipoprotein (Fig. 2). In minimal medium with or without lipoprotein-deficient serum, mutant Si grew slower than wild-type cells (generation times 24 hr versus 37 hr) and mutant S2 lysed after about one generation (much like wild-type cells grown with 25-hydroxycholesterol). Addition of cholesterol or low density lipoprotein to minimal medium fully restored the growth of both mutants to that shown by wild-type cells. For reasons not fully understood at present, the growth rate of wild-type and mutant cells in minimal medium supplemented with cholesterol was slightly decreased relative to that of wild-type cells in minimal medium (data not shown). In vivo sterol synthesis measured by acetate or mevalonate incorporation (Table 3) suggested that the decrease in synthesis was more severe in mutant S2 than in mutant Si. The residual synthesis in mutant Si, but not that in mutant S2, was just enough to sustain growth in the absence of sterol supplement (Fig. 2). The amounts of sterols from cells grown with and without cholesterol are shown in Table 4. When subcultured from a medium containing cholesterol to one without sterol supplement, mutants S1 and S2 were found to have 71% and 0%, respectively, of the desmosterol content observed in wildtype cells when measured relative to fatty acids in polar lipids or 63% and 0%, respectively, when measured relative to cellular protein. Under these growth conditions, the total sterol level in the mutants shifted to lower levels than that of the wild-type cells, which is consistent with their survival during the filipin enrichment procedure and with their reduced growth rate (mutant Si) or eventual lysis (mutant S2) shown in Fig. 2. When the different cell types were grown with cholesterol supplement, the total sterol content of the mutants was approximately
Biochemistry:
Saito et al.
Proc. Natl. Acad. Sci. USA 74 (1977) Table 5. In vitro sterol synthesis from mevalonate
Table 4. Amounts of cellular sterols
Growth condition and cell* No sterol WT Si S2 Sterol WT Si S2
ng/,hg of fatty acidt
3733
Relative amounts of products synthesized
ug/mg of proteint
Desmosterol
Total sterol
Desmosterol
Total sterol
165(100)§ 117(71) 0(0)
173(100) 128(74) 78(45)
13.1(100) 8.3(63) 0 (0)
14.0(100)
9.2(66) 5.1(36)
92(100) 199(100) 7.3(100) 18.3(100) 54(59) 208(105) 2.6(36) 16.1(88) 0 (0) 0(0) 249(125) 19.0(104) * The cells were maintained in medium containing 10% calf serum, washed, and transferred to the minimal medium with or without cholesterol (20 ug/ml). In the absence of cholesterol, wild-type cells were grown for 4.9 generations, mutant S1 for 3.8 generations, and mutant S2 for 1.5 generations (because it lyses after 1 to 2 generations in minimal medium). In the presence of the sterol, the growth periods were 3.3, 3.9, and 3.4 generations, respectively. t Numbers shown represent results of a single experiment; fatty acid for this determination was derived from polar lipids. I Averages from two experiments that varied less than 10%. § Numbers in parentheses are % of wild-type values.
the same as that of wild-type cells. Despite the presence of exogenous sterol supplement, wild-type cells continued to synthesize a substantial amount of desmosterol (92 ng/,gg of fatty acid or 7.3 ,gg/mg of protein); mutants SI and S2 produced little or none of their own sterol. Sterol Synthesis from Mevalonate Catalyzed by Cell Homogenates of Mutants SI and S2. When the incorporation of radioactive mevalonate into sterols was studied with cell homogenates from normal LM cells, desmosterol accounted for about 20% and farnesol and squalene intermediates each for about 35% of the reaction product (Table 5). Synthesis of desmosterol by homogenates from mutants SI and S2 was markedly decreased relative to that observed with homogenates from the wild-type cells. On the other hand, synthesis of farnesol, squalene, and lanosterol increased 3- to 5-fold. These results indicate that the defect in sterol synthesis in both mutants lies between lanosterol and desmosterol, possibly in the demethylation reactions because no demethylated precursor of desmosterol accumulated. The fact that in vitro desmosterol synthesis in mutant Si was very low whereas in vivo measurements of sterol content and synthesis showed a more modest change from the normal suggests that the defective component of the biosynthetic mechanism in mutant Si is particularly unstable in vitro. Preliminary experiments in which homogenates of mutants Si and S2 were mixed either with one another or with a homogenate derived from wild-type cells demonstrated that no inhibitor was present in either mutant homogenate but suggested that the defects in the two mutants were in different functions because their homogenates appeared to complement one another. It is interesting to note that intermediates of sterol synthesis that accumulated during incubation with mutant homogenates were not found in the cells even when the cells were grown without sterol supplements in minimal medium (containing bovine serum albumin) or in medium supplemented with lipoprotein-deficient serum. Although sterol intermediates might be excreted and accumulated in certain growth media (e.g., see refs. 16 and 17), the total nonsaponifiable lipid found in the minimal medium was less than 1% of that present in the mutants or wild-type cells. DISCUSSION Because growth and membrane synthesis continue in the ab-
Source of
homogenate Farnesol Squalene WT S1 S2
1.7 5.0 7.7
1.9 5.6 8.0
Lanosterol 0.3 1.5 1.6
X* 0.3 0.4 0.3
Desmosterol 1.0 0.0 0.2
Cells were maintained in medium with 10% calf serum, transferred to the minimal medium, incubated for 24 hr, washed twice with phosphate-buffered saline, and resuspended in 0.1 M Tris, pH 7.5/ 5 mM MgCI2/2 mM MnCl2/30 mM nicotinamide. The suspension was homogenized with a tight-fitting Dounce glass homogenizer (55 strokes) and centrifuged at 3,000 X g for 5 min; the supernatant was used as the cell homogenate. The reaction was run for 30 min at 370 in 0.5 ml of the above buffer containing 1.5 ismol of ATP, 1.5 Amol of NAD, 5 nmol of FAD, 0.5 ismol of NADP, 1.5 A of glucose 6-phosphate, and 2 units of glucose-6-phosphate dehydrogenase. The substrate was Na [5-3H]mevalonate (0.5-1.0 Mmol, 7.5 Ci/mol). The reaction was stopped by the addition of 1 ml of 1.8 M NaOH in ethanol. The reaction products were extracted into n-pentane which was then washed with H20. Squalene was separated from the other reaction products by alumina column chromatography (15). The remaining fraction was further separated into farnesol, lanosterol, X, and desmosterol by chromatography on a 10% AgNO3-impregnated silica gel thin-layer plate with chloroform/acetone, 95:5 (vol/vol) as the solvent. Fractions from the plates were collected with ether and assayed for radioactivity. The numbers shown are relative to dpm/mg of protein for desmosterol in wild-type homogenates and are averages of at least three determinations that varied no more than 15%. * X was found between lanosterol and desmosterol on the AgNO3/silica gel plates and migrated identically on rechromatography.
sence of sterol synthesis and lead to eventual cell death (1, 2), conventional procedures for selecting auxotrophs would not be very efficient for isolating sterol-requiring mutants. Resistance to filipin is a very selective method for obtaining sterol mutants. Furthermore, because the concentration of polyene antibiotic can be adjusted to detect rather small differences in membrane sterol levels (Fig. 1), one does not have to leave putative mutants without sterol supplement for periods that result in cell death. The polyene antibiotic nystatin has been used to obtain mutants in yeast and fungi that synthesize molecules that are precursors and substitute for the normal membrane sterols but do not complex with the antibiotic (18, 19). However, in the present study, the short exposure to an appropriate polyene antibiotic permitted the isolation of mutants synthesizing less sterol as well as those synthesizing altered sterol. The biochemical defects in mutants Si and S2 are specific to the sterol pathway because growth of both derivatives was restored to that of wild-type cells by sterol supplement (in the form of bovine serum albumin-cholesterol complex or low density lipoprotein). The fact that their fatty acid composition was normal, particularly the unsaturated fatty acid content, rules out a defect in synthesis of cytochromes such as that described for pleiotropic mutants of yeast impaired in the synthesis of both sterol and unsaturated fatty acids (20). Recently, a sterol-requiring mutant of Chinese hamster ovary cells was obtained by using the conventional approach for isolating auxotrophs (3). Interestingly, this mutant, like mutants SI and S2, also appears to be defective in a demethylation reaction. Both mutants (SI and S2) appear to retain their defective characteristics on recloning and after 9 months of serial propagation, as has been reported for the sterol synthesis mutant of Chinese hamster ovary cells (3). The general isolation method described in this paper should permit collection of various mutants which will help elucidate
3734
Biochemistry:
Saito et al.
many aspects of sterol and polyisoprenoid metabolism. With the knowledge that the primary biochemical defect in the mutants characterized in this paper is restricted to a segment of the sterol pathway not required for synthesis of other polyisoprenoid compounds and that secondary effects such as accumulation of sterol precursors does not seem to occur in vivo (see Results), one can now use the mutants to modify membrane lipid composition, to study the interrelationship between sterol content and phospholipid structure, and to correlate physical and functional properties of the plasma membrane. We thank S. C. Kinsky for providing us with filipin and S. Weidman for the low density lipoprotein and lipoprotein-free serum. These studies were supported by American Cancer Society Research Grant BC-198 and U.S. Public Health Service Research Career Development Award 5 K04-GM 70654 to D.F.S. 1. Chen, H. W., Kandutsch, A. A. & Waymouth, C. (1974) Nature 251,419-421. 2. Kandutsch, A. A. & Chen, H. W. (1977) J. Biol. Chem. 252, 409-415. 3. Chang, T. Y., Telakowski, C., Vanden Heuvel, W., Alberts, A. W. & Vagelos, P. R. (1977) Proc. Natl. Acad. Sci. USA 74, 832-836. 4. De Kruyff, B., Demel, R. A. & van Deenen, L. L. M. (1972) Biochim. Biophys. Acta 255,331-347. 5. Rottem, S., Cirillo, V. P., DeKruyff, B., Shinitzky, M. & Razin, S. (1973) Biochim. Biophys. Acta 323, 509-519.
Proc. Natl. Acad. Sci. USA 74 (1977) 6. Higuchi, K. (1970) J. Cell. Physiol. 75,65-72. 7. Kuroki, T. (1975) in Methods in Cell Biology, ed. Prescott, D. M. (Academic Press, New York), Vol. IX, pp. 157-178. 8. Bligh, E. G. & Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37, 911-917. 9. Paul, J. (1970) in Cell and Tissue Culture (Williams and Wilkins Co., Baltimore, MD), 4th ed., p. 91. 10. Lowry, O.H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 11. Kinsky, S. C. (1970) Annu. Rev. Biochem. 10, 119-142. 12. Norman, A. W., Demel, R. A., DeKruyff, B. & van Deenen, L. L. M. (1972) J. Biol. Chem. 247, 1918-1929. 13. Kandutsch, A. A. & Chen, H. H. (1974) J. Biol. Chem. 249, 6057-6061. 14. Rothblat, G. H., Burns, C. H., Conner, R. L. & Landrey, J. R. (1970) Science 169, 180-182. 15. Popjak, G. (1969) in Methods in Enzymology, ed. Clayton, R. B. (Academic Press, New York and London), Vol. XV, pp. 393-454. 16. Edwards, P. A., Fogelman, A. M. & Popjak, G. (1976) Biochem. Biophys. Res. Commun. 68,64-69. 17. Fogelman, A. M., Seager, J., Edwards, P. A. & Popjak, G. (1977) J. Biol. Chem. 252,644-651. 18. Bard, M. (1972) J. Bacterlol. 111, 649-657. 19. Morris, D. C., Safe, N. & Subden, R. E. (1974) Biochem. Genet. 12,459-466. 20. Karst, F. & Lacroute, F. (1973) Biochem. Blophys. Res. Commun. 52,741-747.
Biochemistry
Animal cell mutants defective in sterol metabolism: A specific selection procedure and partial characterization of defects (sterol biosynthesis/polyene antibiotic/membrane biogenesis)
Y. SAITO, S. M. CHOU, AND D. F. SILBERT Department of Biological Chemistry, Division of Biology and Biomedical Sciences, Washington University, St. Louis, Missouri 63110
Communicated by P. Roy Vagelos, June 20, 1977
By using a chemically defined medium, a ABSTRACT general and highly specific procedure was devised to select for mutant cells with less abundant or structurally altered sterol in their surface membranes. Within a certain concentration range, the polyene antibiotic filipin was shown to kill only cells with normal (as opposed to decreased) membrane sterol levels. Sterol-requiring derivatives of LM cells were isolated by chemical mutagenesis, filipin treatment, and cloning followed by replica plating in soft agar. Mutants (SI and S2) are described which, when compared to normal cells, show decreased synthesis of desmosterol in vivo from acetate and mevalonate relative to cell number or to fatty acid synthesis. When exogenous sterol is supplied, mutants SI-and S2 grow normally in suspension culture. However, when deprived of sterol supplement, mutant SI grows slower than wild type cells and mutant S2 lyses within one to two generations. Gas/liquid chromatography revealed that the mutants contained a normal spectrum of fatty acids including ubsaturated fatty acyl groups but, unlike wildtype cells, they have less abundant (mutant SI) or no (mutant S2) desmosterol in either the presence or absence of exogenous cholesterol. In vitro experiments with mevalonate as the substrate suggest that the defect in both mutants is in a demethylation reaction subsequent to lanosterol synthesis. The selection method developed here may permit the isolation of mutants with defective membrane incorporation of sterols and other polyisoprenoids as well as defective synthesis of these compounds. Sterols are present in membranes of virtually all eukaryotic cells. An essential role for sterols in the functional integrity of such membranes is indicated by the fact that eukaryotic cells lyse after a generation or two in medium devoid of sterol when sterol synthesis is impaired by oxygenated cholesterol analogues (1, 2) or by genetic mutation (3). Studies with model as well as natural membranes suggest that sterol molecules modulate fatty acyl chain interactions (4, 5). The consequence of this modulation on interactions between phospholipid molecules and between phospholipids and proteins is not fully understood. Furthermore, the preferential concentration of sterols in surface membranes and the mechanism for translocation of sterols from their site of synthesis to the surface membrane have not been
explained.
In this paper, we describe a general and highly selective procedure for isolation of animal cell mutants with membranes containing decreased levels of or structurally altered sterols, and we characterize two mutants with decreased sterol synthesis. Many asts of sterol and polyisoprenoid metabolism and of membrane biochemistry may be probed by this approach.
MATERIALS AND METHODS Materials. All the labeled compounds and sterols were purchased from New England Nuclear and Steraloids, reThe costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.
spectively. Filipin was obtained from G. B. Whitfield, Upjohn Co., via S. C. Kinsky. Lipoprotein-deficient serum and low density lipoprotein were provided by S. Weidman. Cell Growth. LM cells (mouse fibroblasts) were purchased from the American Type Culture Collection (CCL 1.2) and were propagated in suspension culture or in monolayers. Minimal growth medium was Higuchi's medium (6) modified by the addition of 20 mM N-2-hydroxyethylpiperazine-N'2-ethanesulfonic acid (Hepes) buffer (final pH 7.4), 0.5% bovine serum albumin, and 0.015% Darvan no. 2 (R. T. Vanderbilt Co.) (to minimize cell clumping or attachment to the walls of the culture bottle). The suspension culture (25 ml in a 125-ml Wheaton screw-cap bottle) was agitated on a gyratory shaker (140 rpm) at 370. Growth was monitored by measuring cell number in a hemocytometer, and cell densities ranged generally from 0.5 to 3 X 106 cells per ml. Generation time in the minimal medium was about 24 hr. When cells were grown as monolayers or cloned in soft agar (7), a modified Eagle's medium (Flow Laboratories) supplemented with penicillin (50 units/ml) and streptomycin (50,ug/ml) was used and the cultures were grown in a CO2 incubator. Isolation of Mutants with Decreased Membrane Desmosterol Levels. LM cells were mutagenized by incubating cells with N-methyl-N'-nitro-N-nitrosoguanidine (20 ,uM) for 30 min at 370 in minimal medium supplemented with 10% calf serum. Based on the number of clones formed in soft agar before and after treatment, about 10% of the cells survived. The mutagen was washed out and the cells were incubated for several generations in fresh medium to allow for segregation of chromosomes and expression of mutant phenotypes. At this point the cells were transferred to minimal medium and grown for 1 to 1 'k generations so that the membrane sterol of sterol synthesis mutants would decrease below that of normal cells. Filipin in dimethylformamide was diluted into minimal medium and then into the cell suspension to give a final concentration of 5-6 The cells were incubated for 30 min at 370, washed, jgg/ml. resuspended in fresh minimal medium containing 10% calf serum, and grown for several generations. The antibiotic treatment was repeated once; after growth had recovered, the cells were cloned in soft agar containing 10% calf serum. The resultant clones were replica plated by transferring cells with a glass rod onto a set of soft agar plates, each set containing a plate supplemented with 0.5% bovine serum albumin, one supplemented with bovine serum albumin plus cholesterol (20 Clones ,ug/ml), and one supplemented with 10% calf serum.taken as that grew well only with the sterol supplement were putative sterol-requiring mutants. These derivatives were propagated in monolayer cultures and transferred by trypsinization to suspension cultures for biochemical studies and additional growth experiments. After confirmation of their phenotype by these in vivo measurements of sterol synthesis, sterol mutants were recloned before more detailed characterization of the specific biochemical defect. 3730
Biochemistry:
Saito et al.
Proc. Natl. Acad. Sci. USA 74 (1977)
Table 1. Regimen for selection of sterol-requiring mutants -48 to 0 hr 0 to 24 hr 30 min Afterward
4:100
0
Control
+C`
Cholesterol, 1 Mg/mI 2
3
4
5
Filipin,
ag/mI
6
7
8
9
of culture
FIG.
1. Development of resistance to filipin accompanying inhibition of sterol synthesis in growing cultures. In one experiment, LM cells in the minimal medium were preincubated with 25-hy-
(M),
droxycholesterol (1 gg/ml) for 0 (x), 5 (0), 10 and 24 (o) hr and treated with filipin at the concentrations shown. In another experiment (broken lines) control cells and cells with decreased sterol
(-)
level (@) were prepared as described in Table
min.
and treated with filipin
The incorporation of thymidine (dThd) into DNA was
then monitored by incubating cells with [3H]thymidine (1 mM, 100 Ci/mol) at 370 for 1 hr without washing out the filipin.
Analytical Procedures. Lipids were extracted by the method of Bligh and Dyer (8) from cells washed thoroughly with
phosphate-buffered saline (9). Polar and nonpolar lipids in the extracts were separated by Unisil silica gel column chromatography, and nonpolar lipids were subdivided into a saponifiable and a nonsaponifiable fraction. The latter lipids are largely sterols. Gas/liquid chromatography was used for the quantitative analysis of sterols and fatty acids. Sterols were
chromatographed with coprostanol as the internal standard on a 3% OV 17 column at
250°;
Filipin Calf serum, 10%
With Cholesterol, 25-OH-Cho- Filipin Calf serum, 10% decreased 20 ug/ml + 25-OHlesterol, sterol 1 ,g/ml + 25-OHCholesterol, 1
.C
for 30
Cholesterol, No sterol 20 Ag/ml
\
60
0
3731
fatty acids were separated on a
10% diethylene glycol succinate column at 1650
with pentadecanoic acid as the internal standard. Protein was determined by the method of Lowry et al. (10).
RESULTS Enrichment Method. A polyene antibiotic, filipin, is known to interact with certain kinds of sterols in biological as well as artificial membranes and to alter membrane properties (11, 12). In fact, the antibiotic is cytotoxic to cells and this effect depends on the presence of sterols in the cell membranes. Prokaryotic cells that do not contain sterols are not affected by filipin. If cells with decreased membrane sterol content could be shown to be
differentially resistant, resistance to filipin could be used to select mutants defective in sterol synthesis and in other mechanisms required for maintaining normal membrane sterol levels. To explore this approach we prepared cells with decreased membrane sterol content by utilizing oxygenated derivatives of cholesterol, such as 25-hydroxycholesterol, which are known to inhibit sterol synthesis specifically but not to substitute for the normal membrane sterol (13). Growth of LM cells with 25-hydroxycholesterol (1 for 24 hr led to a 5-fold de-
,ug/inl)
crease in the incorporation of radioactive acetate into sterols relative to that observed with cells grown without inhibitor but did not affect the amount of radioactivity incorporated from acetate into phospholipids (data not shown). By using appropriate analytical methods for determining mass, the amount
of cellular sterol relative to protein was found to decrease progressively during growth in the presence of the sterol analogue whereas total phospholipid versus protein remained unchanged (data not shown). Inhibition of sterol synthesis leads to a re-
sg/ml
duction in protein and nucleic acid synthesis after one to two generations of growth and thereafter to cell death. These long-term effects of treatment with 25-hydroxycholesterol were prevented by supplying cholesterol (in the form of calf serum or bovine serum albumin complex) to the cells before they had been cultured for more than one generation with the oxygenated sterol (data not shown). Fig. 1 shows that decreased membrane sterol content results in relative resistance of the cells to the action of filipin. The incorporation of thymidine into DNA by whole cells is a convenient way to quantitate the action of filipin (one can also monitor this process by measuring the incorporation of leucine into protein or by following morphological changes of the cells with an inverted phase microscope). When LM cells were grown with 25-hydroxycholesterol, the amount of cellular sterol [i.e., desmosterol (14)] decreased to 14.0, 10.9, 9.6, and 6.8 ,ug/mg of protein after 0, 5, 10,Iand 24 hr (slightly more than one generation), respectively. The concentration of filipin required for 50% inhibition of thymidine incorporation increased from 4.5 to more than 9 ,g/ml from the beginning to the end of this growth period. Next, we subjected LM cells grown without or with the inhibitor of sterol synthesis to conditions identical to those that would be used in the isolation of sterol-requiring mutants (Table 1). The two populations of cells were grown in a medium containing cholesterol supplement for at least two generations, shifted to medium devoid of sterol and propagated for another generation, treated briefly with filipin, and returned to a medium supplemented with calf serum. Under the conditions used in this experiment, the cells grown with 25-hydroxycholesterol were highly resistant to filipin at 5 ,g/ml whereas the control populations showed little or no resistance. This conclusion is based not only on the effect of filipin on thymidine incorporation determined immediately after treatment (Fig. 1, broken lines) but, more significantly, also on the survival of the cells (Table 2). The number of cells in the population containing normal membrane sterol content fell more than 100-fold by 1 day after treatment and did not recover over the succeeding Table 2. Recovery of cell growth after exposure to filipin Time after Cell count X 10-4 per ml exposure to With filipin (5,ig/ml), hr Control decreased sterol 0 127 140 24 1 84 48 0 114 72 0 123 The two populations of cells treated as shown in Table 1 were treated with filipin as described in Materials and Methods and placed in fresh medium containing 10% calf serum with or without 25-hydroxycholesterol (1 sg/ml) to restore growth. The cell counts are means of at least two determinations that varied by no more than 109.
3732
Biochemistry:
Saito et al.
Proc. Natl. Acad. Sci. USA 74 (1977)
Table 3. In vivo synthesis of sterol measured by incorporation of acetate and mevalonate into nonsaponifiable lipids Mevalonate Acetate incorporation incorporation Into polar lipids Into sterol, into sterol, dpm X 103 dpm X 104 per cell dpm X 104 per cell Cell* per cell 38 77 123 2.1 68 136 Si 87 0.2 10 S2 The cells maintained in medium containing 10% calf serum were washed and transferred to the minimal medium and grown for 24 hr. They were then incubated for an additional 24 hr with [14C]acetate (0.13 mM, 2.45 Ci/mol) and [3H]mevalonate (0.1 mM, 5.05 Ci/mol). The cells were washed thoroughly with phosphate-buffered saline. The lipids were extracted and fractionated. * WT, wild type; Si and S2, mutants. WT
2 days but that of cells with decreased sterol content decreased only 40% and growth of the survivors was apparent over the next 2 days. Hence, if one chose the appropriate concentration of filipin and selected the optimal time after shifting from permissive to nonpermissive (i.e., no sterol supplement) conditions for treatment with the polyene antibiotic, one could use this approach to select effectively for mutant cells with reduced membrane sterol levels. Isolation of Strains with Decreased Sterol Synthesis. Putative sterol-requiring derivatives of LM cells were isolated by using nitrosoguanidine mutagenesis, the filipin enrichment procedure, and the cloning and replica plating techniques described above. In the experiment shown in Table 3, radioactive acetate was used to measure sterol and fatty acid synthesis in wild-type cells and in two mutants presumably defective in sterol synthesis, one generation after shifting the cells from sterol-supplemented to minimal medium. Because acetate derived from a common pool is used for sterol and fatty acid synthesis, the data in Table 3 (columns 2 and 3) suggest that sterol synthesis is impaired but fatty acid synthesis is stimulated in the mutants relative to the wild-type cells. If sterol synthesis is compared to fatty acid synthesis, the defects in the mutants are actually magnified. The incorporation of acetate into the sterol fraction in mutant S2 was much less than that in mutant Si. In order to quantitate the incorporation of acetate into desmosterol in vivo, a separate experiment was carried out in which [3Hldesmosterol was added to the nonsaponifiable lipids (labeled with "4C) and this fraction was further resolved by preparative gas/liquid chromatography. The amount of radioactivity in desmosterol derived from mutants S1 and S2 was 41% and 0.1% of the wild-type level, respectively. Radioactive mevalonate incorporation into sterols was also markedly decreased in the mutants (Table 3, column 4), confirming the results with radioactive acetate and, furthermore, demonstrating that mutants S1 and S2 were defective in the sterol synthesis pathway distal to the hydroxymethylgluctaryl-CoA reductase step. The decrease of sterol synthesis in the mutants as measured by the incorporation of mevalonate appeared to be greater than that observed by acetate labeling. This finding suggests that either these mutants do not take up mevalonate from the medium as efficiently as wild-type cells do or they have increased levels of the reductase which results in greater synthesis of mevalonate and, in this experiment, a lower specific radioactivity of the cellular mevalonate pool. This latter possibility might be expected as a consequence of the lowered level of sterols in the mutants, and it is currently under investigation.
0~~~ 0
E
10
2 3 4 5 6 Days of growth FIG. 2. The growth of the wild-type and the two mutant cells was followed in the presence of lipoprotein-deficient serum (30 gg of protein and 0.04 jig of cholesterol per ml of culture) (solid line, open symbols) or in the presence of low density lipoprotein (30 ;g of protein and 34 ,ug of cholesterol per ml of culture) (broken line, solid symbols). O and *, Wild-type cells; 0 and *, mutants Si; A and *, S2. The cell counts are means of at least two determinations that varied by no more than 10%.
Growth Properties and Sterol Composition of Mutants S1 and S2. Growth of wild-type LM cells and of mutants SI and S2 was followed in suspension cultures with minimal medium or minimal medium supplemented with cholesterol, lipoprotein-deficient serum, or low density lipoprotein (Fig. 2). In minimal medium with or without lipoprotein-deficient serum, mutant Si grew slower than wild-type cells (generation times 24 hr versus 37 hr) and mutant S2 lysed after about one generation (much like wild-type cells grown with 25-hydroxycholesterol). Addition of cholesterol or low density lipoprotein to minimal medium fully restored the growth of both mutants to that shown by wild-type cells. For reasons not fully understood at present, the growth rate of wild-type and mutant cells in minimal medium supplemented with cholesterol was slightly decreased relative to that of wild-type cells in minimal medium (data not shown). In vivo sterol synthesis measured by acetate or mevalonate incorporation (Table 3) suggested that the decrease in synthesis was more severe in mutant S2 than in mutant Si. The residual synthesis in mutant Si, but not that in mutant S2, was just enough to sustain growth in the absence of sterol supplement (Fig. 2). The amounts of sterols from cells grown with and without cholesterol are shown in Table 4. When subcultured from a medium containing cholesterol to one without sterol supplement, mutants S1 and S2 were found to have 71% and 0%, respectively, of the desmosterol content observed in wildtype cells when measured relative to fatty acids in polar lipids or 63% and 0%, respectively, when measured relative to cellular protein. Under these growth conditions, the total sterol level in the mutants shifted to lower levels than that of the wild-type cells, which is consistent with their survival during the filipin enrichment procedure and with their reduced growth rate (mutant Si) or eventual lysis (mutant S2) shown in Fig. 2. When the different cell types were grown with cholesterol supplement, the total sterol content of the mutants was approximately
Biochemistry:
Saito et al.
Proc. Natl. Acad. Sci. USA 74 (1977) Table 5. In vitro sterol synthesis from mevalonate
Table 4. Amounts of cellular sterols
Growth condition and cell* No sterol WT Si S2 Sterol WT Si S2
ng/,hg of fatty acidt
3733
Relative amounts of products synthesized
ug/mg of proteint
Desmosterol
Total sterol
Desmosterol
Total sterol
165(100)§ 117(71) 0(0)
173(100) 128(74) 78(45)
13.1(100) 8.3(63) 0 (0)
14.0(100)
9.2(66) 5.1(36)
92(100) 199(100) 7.3(100) 18.3(100) 54(59) 208(105) 2.6(36) 16.1(88) 0 (0) 0(0) 249(125) 19.0(104) * The cells were maintained in medium containing 10% calf serum, washed, and transferred to the minimal medium with or without cholesterol (20 ug/ml). In the absence of cholesterol, wild-type cells were grown for 4.9 generations, mutant S1 for 3.8 generations, and mutant S2 for 1.5 generations (because it lyses after 1 to 2 generations in minimal medium). In the presence of the sterol, the growth periods were 3.3, 3.9, and 3.4 generations, respectively. t Numbers shown represent results of a single experiment; fatty acid for this determination was derived from polar lipids. I Averages from two experiments that varied less than 10%. § Numbers in parentheses are % of wild-type values.
the same as that of wild-type cells. Despite the presence of exogenous sterol supplement, wild-type cells continued to synthesize a substantial amount of desmosterol (92 ng/,gg of fatty acid or 7.3 ,gg/mg of protein); mutants SI and S2 produced little or none of their own sterol. Sterol Synthesis from Mevalonate Catalyzed by Cell Homogenates of Mutants SI and S2. When the incorporation of radioactive mevalonate into sterols was studied with cell homogenates from normal LM cells, desmosterol accounted for about 20% and farnesol and squalene intermediates each for about 35% of the reaction product (Table 5). Synthesis of desmosterol by homogenates from mutants SI and S2 was markedly decreased relative to that observed with homogenates from the wild-type cells. On the other hand, synthesis of farnesol, squalene, and lanosterol increased 3- to 5-fold. These results indicate that the defect in sterol synthesis in both mutants lies between lanosterol and desmosterol, possibly in the demethylation reactions because no demethylated precursor of desmosterol accumulated. The fact that in vitro desmosterol synthesis in mutant Si was very low whereas in vivo measurements of sterol content and synthesis showed a more modest change from the normal suggests that the defective component of the biosynthetic mechanism in mutant Si is particularly unstable in vitro. Preliminary experiments in which homogenates of mutants Si and S2 were mixed either with one another or with a homogenate derived from wild-type cells demonstrated that no inhibitor was present in either mutant homogenate but suggested that the defects in the two mutants were in different functions because their homogenates appeared to complement one another. It is interesting to note that intermediates of sterol synthesis that accumulated during incubation with mutant homogenates were not found in the cells even when the cells were grown without sterol supplements in minimal medium (containing bovine serum albumin) or in medium supplemented with lipoprotein-deficient serum. Although sterol intermediates might be excreted and accumulated in certain growth media (e.g., see refs. 16 and 17), the total nonsaponifiable lipid found in the minimal medium was less than 1% of that present in the mutants or wild-type cells. DISCUSSION Because growth and membrane synthesis continue in the ab-
Source of
homogenate Farnesol Squalene WT S1 S2
1.7 5.0 7.7
1.9 5.6 8.0
Lanosterol 0.3 1.5 1.6
X* 0.3 0.4 0.3
Desmosterol 1.0 0.0 0.2
Cells were maintained in medium with 10% calf serum, transferred to the minimal medium, incubated for 24 hr, washed twice with phosphate-buffered saline, and resuspended in 0.1 M Tris, pH 7.5/ 5 mM MgCI2/2 mM MnCl2/30 mM nicotinamide. The suspension was homogenized with a tight-fitting Dounce glass homogenizer (55 strokes) and centrifuged at 3,000 X g for 5 min; the supernatant was used as the cell homogenate. The reaction was run for 30 min at 370 in 0.5 ml of the above buffer containing 1.5 ismol of ATP, 1.5 Amol of NAD, 5 nmol of FAD, 0.5 ismol of NADP, 1.5 A of glucose 6-phosphate, and 2 units of glucose-6-phosphate dehydrogenase. The substrate was Na [5-3H]mevalonate (0.5-1.0 Mmol, 7.5 Ci/mol). The reaction was stopped by the addition of 1 ml of 1.8 M NaOH in ethanol. The reaction products were extracted into n-pentane which was then washed with H20. Squalene was separated from the other reaction products by alumina column chromatography (15). The remaining fraction was further separated into farnesol, lanosterol, X, and desmosterol by chromatography on a 10% AgNO3-impregnated silica gel thin-layer plate with chloroform/acetone, 95:5 (vol/vol) as the solvent. Fractions from the plates were collected with ether and assayed for radioactivity. The numbers shown are relative to dpm/mg of protein for desmosterol in wild-type homogenates and are averages of at least three determinations that varied no more than 15%. * X was found between lanosterol and desmosterol on the AgNO3/silica gel plates and migrated identically on rechromatography.
sence of sterol synthesis and lead to eventual cell death (1, 2), conventional procedures for selecting auxotrophs would not be very efficient for isolating sterol-requiring mutants. Resistance to filipin is a very selective method for obtaining sterol mutants. Furthermore, because the concentration of polyene antibiotic can be adjusted to detect rather small differences in membrane sterol levels (Fig. 1), one does not have to leave putative mutants without sterol supplement for periods that result in cell death. The polyene antibiotic nystatin has been used to obtain mutants in yeast and fungi that synthesize molecules that are precursors and substitute for the normal membrane sterols but do not complex with the antibiotic (18, 19). However, in the present study, the short exposure to an appropriate polyene antibiotic permitted the isolation of mutants synthesizing less sterol as well as those synthesizing altered sterol. The biochemical defects in mutants Si and S2 are specific to the sterol pathway because growth of both derivatives was restored to that of wild-type cells by sterol supplement (in the form of bovine serum albumin-cholesterol complex or low density lipoprotein). The fact that their fatty acid composition was normal, particularly the unsaturated fatty acid content, rules out a defect in synthesis of cytochromes such as that described for pleiotropic mutants of yeast impaired in the synthesis of both sterol and unsaturated fatty acids (20). Recently, a sterol-requiring mutant of Chinese hamster ovary cells was obtained by using the conventional approach for isolating auxotrophs (3). Interestingly, this mutant, like mutants SI and S2, also appears to be defective in a demethylation reaction. Both mutants (SI and S2) appear to retain their defective characteristics on recloning and after 9 months of serial propagation, as has been reported for the sterol synthesis mutant of Chinese hamster ovary cells (3). The general isolation method described in this paper should permit collection of various mutants which will help elucidate
3734
Biochemistry:
Saito et al.
many aspects of sterol and polyisoprenoid metabolism. With the knowledge that the primary biochemical defect in the mutants characterized in this paper is restricted to a segment of the sterol pathway not required for synthesis of other polyisoprenoid compounds and that secondary effects such as accumulation of sterol precursors does not seem to occur in vivo (see Results), one can now use the mutants to modify membrane lipid composition, to study the interrelationship between sterol content and phospholipid structure, and to correlate physical and functional properties of the plasma membrane. We thank S. C. Kinsky for providing us with filipin and S. Weidman for the low density lipoprotein and lipoprotein-free serum. These studies were supported by American Cancer Society Research Grant BC-198 and U.S. Public Health Service Research Career Development Award 5 K04-GM 70654 to D.F.S. 1. Chen, H. W., Kandutsch, A. A. & Waymouth, C. (1974) Nature 251,419-421. 2. Kandutsch, A. A. & Chen, H. W. (1977) J. Biol. Chem. 252, 409-415. 3. Chang, T. Y., Telakowski, C., Vanden Heuvel, W., Alberts, A. W. & Vagelos, P. R. (1977) Proc. Natl. Acad. Sci. USA 74, 832-836. 4. De Kruyff, B., Demel, R. A. & van Deenen, L. L. M. (1972) Biochim. Biophys. Acta 255,331-347. 5. Rottem, S., Cirillo, V. P., DeKruyff, B., Shinitzky, M. & Razin, S. (1973) Biochim. Biophys. Acta 323, 509-519.
Proc. Natl. Acad. Sci. USA 74 (1977) 6. Higuchi, K. (1970) J. Cell. Physiol. 75,65-72. 7. Kuroki, T. (1975) in Methods in Cell Biology, ed. Prescott, D. M. (Academic Press, New York), Vol. IX, pp. 157-178. 8. Bligh, E. G. & Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37, 911-917. 9. Paul, J. (1970) in Cell and Tissue Culture (Williams and Wilkins Co., Baltimore, MD), 4th ed., p. 91. 10. Lowry, O.H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 11. Kinsky, S. C. (1970) Annu. Rev. Biochem. 10, 119-142. 12. Norman, A. W., Demel, R. A., DeKruyff, B. & van Deenen, L. L. M. (1972) J. Biol. Chem. 247, 1918-1929. 13. Kandutsch, A. A. & Chen, H. H. (1974) J. Biol. Chem. 249, 6057-6061. 14. Rothblat, G. H., Burns, C. H., Conner, R. L. & Landrey, J. R. (1970) Science 169, 180-182. 15. Popjak, G. (1969) in Methods in Enzymology, ed. Clayton, R. B. (Academic Press, New York and London), Vol. XV, pp. 393-454. 16. Edwards, P. A., Fogelman, A. M. & Popjak, G. (1976) Biochem. Biophys. Res. Commun. 68,64-69. 17. Fogelman, A. M., Seager, J., Edwards, P. A. & Popjak, G. (1977) J. Biol. Chem. 252,644-651. 18. Bard, M. (1972) J. Bacterlol. 111, 649-657. 19. Morris, D. C., Safe, N. & Subden, R. E. (1974) Biochem. Genet. 12,459-466. 20. Karst, F. & Lacroute, F. (1973) Biochem. Blophys. Res. Commun. 52,741-747.
