THE JOURNAL OF BIOLOGICAL CHEMISTRY Val. No. S. … ·  · 2001-07-10Cytoplasmic 3-Hydroxy-3-methylglutaryl Coenzyme A Synthase from the Hamster ... including 3-hydroxy-3-methylglutaryl

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1986 by The American Society of Biological Chemists, Inc.

Val. 261, No. 8, Issue of March 15, pp. 3710-3716 1986 Printed in l? S. A.

Cytoplasmic 3-Hydroxy-3-methylglutaryl Coenzyme A Synthase from the Hamster I. ISOLATION AND SEQUENCING OF A FULL-LENGTH cDNA*

(Received for publication, September 11,1985)

Gregorio Gil$S, Joseph L. Goldstein$, Clive A. SlaughterlI, and Michael S. Brown$ From the $Departments of Molecular Genetics and Internal Medicine, University of Texas Health Science Center, Southwestern Medical School, Dallas, Texas 75235 and the Wold Spring Harbor Laboratory, Cold Spring Harbor, New York 11 724

We here report the isolation and nucleotide sequenc- ing of a full-length 3.3-kilobase cDNA for the cyto- plasmic form of 3-hydroxy-3-methylglutaryl coen- zyme A (HMG-CoA) synthase, a regulated enzyme in the cholesterol biosynthetic pathway. The cDNA was isolated from UT-1 cells, a compactin-resistant line of Chinese hamster ovary cells. UT- 1 cells produce large amounts of mRNA for HMG-CoA synthase and the next enzyme in the pathway, HMG-CoA reductase, as a result of growth in the presence of compactin, a com- petitive inhibitor of the reductase. The identity of the cDNA for HMG-CoA synthase was confirmed through comparison of the NHz-terminal amino acid sequence predicted from the cDNA with that determined chem- ically from the purified enzyme. Anti-peptide antibod- ies directed against the amino acid sequence predicted from the cDNA precipitated HMG-CoA synthase activ- ity from liver cytoplasm. The feeding of cholesterol to hamsters led to a decrease of more than 85% in the amount of mRNA for HMG-CoA synthase and HMG- CoA reductase in hamster liver. These data indicate that the mRNAs for cytoplasmic HMG-CoA synthase and for HMG-CoA reductase, two sequential enzymes in the cholesterol biosynthetic pathway, are coordi- nately regulated by cholesterol.

The synthesis of cholesterol from acetyl coenzyme A re- quires more than 20 enzymatic reactions. The activity of this pathway is subject to feedback suppression by the sterol end- products of the pathway. This suppression can be observed in livers of animals that are fed cholesterol; it can also be observed in cultured cells that are incubated with exogenous sources of cholesterol, such as plasma low density lipoprotein, which delivers cholesterol to cells by receptor-mediated en- docytosis. Oxygenated sterols, such as 25-hydroxycholestero1, also suppress cholesterol synthesis when added to the culture medium in solvents (reviewed in Ref. 1).

Current evidence indicates that sterols suppress the choles- terol biosynthetic pathway through a reduction in the activi- ties of several enzymes, including 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA1) reductase ( l ) , HMG-CoA synthase

* This work was supported by Research Grants HL 20948 and CA 13106 from the National Institutes of Health. The 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. Section 1734 solely to indicate this fact.

§Recipient of a postdoctoral fellowship from the Juan March Foundation.

The abbreviation used is: HMG-CoA, 3-hydroxy-3-methylglutaryl coenzyme A.

(2-5), squalene synthetase (6, 7) , and enzymes mediating the reactions that convert squalene to cholesterol (8, 9). Quanti- tatively, the most significant point of suppression is prior to the formation of mevalonate. When rats are fed cholesterol or when cultured fibroblasts are incubated with low density lipoprotein-cholesterol, the incorporation of acetate into cho- lesterol declines at a time when the incorporation of mevalon- ate (the product of the HMG-CoA reductase reaction) has not yet fallen (1, 10).

Among the enzymes that are necessary to convert acetyl- CoA to mevalonate, two enzymes have been clearly demon- strated to be under feedback regulation by cholesterol. One is microsomal HMG-CoA reductase. The other is cytoplasmic HMG-CoA synthase, which condenses acetyl-coA with ace- toacetyl-CoA to form HMG-CoA, the substrate for the HMG- CoA reductase reaction (2, 11).

In liver, two distinct forms of HMG-CoA synthase exist, a mitochondrial form and a cytoplasmic form (2, 11, 12). The majority of the enzyme is located in mitochondria. The HMG- CoA produced by the mitochondrial enzyme is largely con- verted by HMG-CoA lyase to ketone bodies (acetoacetate and P-hydroxybutyrate), which are secreted by liver cells under conditions of rapid fatty acid oxidation. The HMG-CoA pro- duced by the cytoplasmic enzyme is thought to act as a substrate for HMG-CoA reductase, with ultimate conversion to cholesterol and other polyisoprenoid products. In liver, the activity of cytoplasmic HMG-CoA synthase is suppressed by cholesterol feeding and increased by treatment with bile acid- binding resins (2). The extensive studies of Lane and co- workers ( 2 , 11, 12) suggested that the cytoplasmic and mito- chondrial forms of the enzyme have distinct biochemical properties, as indicated by different isoelectric points, immu- nologic reactivity, and response to magnesium ion. The cyto- plasmic form of the enzyme has been purified from avian liver (2), and the mitochondrial form has been purified from both avian liver (12, 13, 15) and ox liver (16).

In extrahepatic cells where ketone body formation is not a major reaction, the most abundant form of HMG-CoA syn- thase is the cytoplasmic enzyme (2-5,11,12). Balasubraman- iam et al. (3) demonstrated that in one extrahepatic tissue (the adrenal cortex), the activities of HMG-CoA synthase and HMG-CoA reductase are both induced when the animals are rendered lipoprotein deficient and are both repressed when plasma cholesterol levels are restored by intravenous injection of LDL. Chang and Limanek (4) and Schnitzer-Polokoff et al. (5) demonstrated that the activities of HMG-CoA synthase and HMG-CoA reductase were suppressed in parallel by the addition of sterols to cultured cells.

In recent years a great deal has been learned about the mechanism of regulation of HMG-CoA reductase, but little is

3710

cDNA for Cytoplasmic HMG-CoA Synthase 3711

known about the regulation of HMG-CoA synthase. Knowl- edge of the reductase emerged as a result of the cDNA cloning of the enzyme, followed by studies of the regulation of the mRNA and protein. The cDNA for HMG-CoA reductase was isolated from UT-1 cells, a clone of Chinese hamster ovary cells that were adapted to growth in the presence of compac- tin, a competitive inhibitor of HMG-CoA reductase (17, 18). In response to growth in the presence of this inhibitor, the UT-1 cells amplified the gene for HMG-CoA reductase 15- fold they also transcribe each copy at an elevated rate (19). While cloning UT-1 cell mRNA, we isolated a partial cDNA that encoded a cholesterol-regulated protein that was distinct from HMG-CoA reductase. In hybrid-sel~ted translation ex- periments, this cDNA hybridized to an mRNA that produced a protein with a molecular weight of approximately 53,000 (20).

By two-dimensional gel electrophoresis, the 53-kilodalton (kDa) protein was shown to be present in increased amounts in the cytoplasm of UT-1 cells as compared to the parental Chinese hamster ovary cells (20). Moreover, the amount of the mRNA for the 53-kDa protein was induced when UT-1 cells were grown in the absence of sterols, and its concentra- tion was repressed by sterols (19,20). We postulated that this 53-kDa protein was a subunit of HMG-CoA synthase. The size and isoelectric point of the 53-kDa protein in UT-1 cells (20) were similar to the reported subunit molecular weight and isoelectric point of the cytoplasm-ic HMG-CoA synthase in avian liver (2). These findings implied that the UT-1 cells were ove~roducing the substrate, HMG-CoA, in response to growth in the presence of an inhibitor that competes with HMG-CoA for the reductase enzyme (20).

We now present data confirming the identity of the 53-kDa protein as the cytoplasmic form of HMG-CoA synthase. We have isolated a full-length cDNA for this enzyme, and from its nucleotide sequence we have deduced the amino acid sequence of the hamster enzyme. In this paper we report this sequence and present immunological and biochemical evi- dence to support the contention that this enzyme is in fact HMG-CoA synthase. In an accompanying paper, we report the isolation of the genomic sequences of hamster HMG-CoA synthase and the characterization of its unusual 5' untrans- lated region (21).

EXPERIMENTAL PROCEDURES

~ a t e r ~ ~ 32P-labeled nucleotides were purchased from New England Nuclear

and ICN. [l-14C]Acetyl-CoA (53 mCi/mmol) was purchased from New England Nuclear. Acetoacetyl-CoA, acetyl-coA, and Blue Seph- arose CL-GB were obtained from Pharmacia P-L Biochemicals. Re- striction enzymes and enzymes used in cDNA cloning and DNA sequencing were obtained from commercial sources. S1 nuclease was purchased from Miles Laboratories, Inc. Oligo~ucleotides were syn- thesized by the phosphoramidite method (22) on an Applied Biosys- tems model 380A DNA synthesizer. Compactin was kindly provided by Akira Endo, Tokyo Noko University, Tokyo, Japan. Mevinolin was kindly pror;ided by Alfred Alberts of Merck, Sharp, and Dohme, Rahway, NJ. Cholestyramine (Questran) was obtained from Mead Johnson and Co., Evansville, IN. Male golden Syrian hamsters were purchased from Sasco, Inc., Omaha, NE. Synthetic peptides corre- sponding to amino acid sequences of HMG-CoA synthase were pur- chased from Peninsula Laboratories, Belmont, CA. Emulphogene BC- 720 was obtained from Research Plus Laboratories, Denville, NJ. Dulbecco's phosphate-buffered saline (catalog no. 310-4190) was pur- chased from Gibco. DEAE-cellulose DE52 and cellulose phosphate P11 were purchased from Whatman. Hydroxylapatite and Pansorbin (Staphylococcus aureus cells) were obtained from Calbiochem-Behr- ing.

General Methods Preparation of plasmid DNA, restriction enzyme digestions,

screening of cDNA libraries, agarose gel electrophoresis, and DNA blotting and hybridization were carried out by standard procedures as described by Maniatis et al. (23). DNA probes were labeled with 32P by nick translation (23) or by random hexanucleotide priming (24). DNA sequencing was done by the chemical method of Maxam and Gilbert (25) and by the dideoxy chain termination method of Sanger et al. (26) with either the M13 universal sequencing primer or specific oligonucleotides after subcloning into bacteriophage M13 vectors as described by Messing (27). Total cellular RNA was isolated from hamster liver by guanidinium thiocyanate extraction (28), fol- lowed by centrifugation through a cesium chloride cushion (29).

Buffers Buffer A consists of 20 mM sodium phosphate, 0.1 mM EDTA, and

0.5 mM dithiothreitol at pH 7. Buffer B consists of 10 mM sodium phosphate, 15% (v/v) glycerol, 0.1 mM EDTA, and 0.5 mM dithio- threitol at pH 7. Buffer C consists of 10 mM sodium phosphate, 10% (v/v) glycerol, 0.5 mM dithiothreitol, and 0.1 mM EDTA at pH 6. Buffer D is the same as Buffer C except that the pH is 7. Buffer E consists of Dulbecco's phospha~-buffered saline (Ca" and M$+- free) containing 1% (v/v) Emulphogene BC-720 and 0.1% (w/v) bovine serum alhumin. Protein was measured by a modification of the Lowry procedure (30).

Animals Male Syrian hamsters (80-100 g) were kept under a 12-h light/l2-

h dark cycle for a t least 7 days prior to use. Animals were fed a chow diet (Wayne Research Animal Diets) supplemented with 4% (w/w) cholestyramine and 0.1% (w/w) mevinolin for 5 days or with 2% (w/ w) cholesterol for 14 days and then killed at the end of the dark cycle. Liver cytosol was prepared at 4 "C by homogenization of the livers in 3 volumes of buffer containing 300 mM sucrose, 20 mM sodium phosphate, and 0.1 mM EDTA at pH 7 in a Potter-Elvehjem homog- enizer with 10 strokes of a Teflon pestle. The homogenate was centrifuged at 30,000 X g for 20 min at 4 "C. The s u p e r n a t ~ t fraction was centrifuged at 100,000 x g for 90 min at 4 "C, and the resulting supernatant solution was dialyzed for 24 h at 4 "C against buffer A (see above).

cDNA Libary Screening

Plasmid p53K-3 was isolated as previously described (20). The cDNA insert in this plasmid corresponded to part of the 3' untrans- lated region of the mRNA (Fig. I). A full-length cDNA for HMG- CoA synthase, designated p53K-312, was isolated from a UT-1 cell cDNA library (18) made with the vectors and methods developed by Okayama and Berg (31) and enriched for longer cDNAs by the sublibrary method (18, 31). The library was initially screened with the 3"P-labeled cDNA insert from p53K-3 (Fig. 1). This screen of -lo5 recombinants yielded more than 50 positive clones, 16 of which were further analyzed. The longest clone, p53K-214, extended -1.4 kilobases more 5' than p53K-3, but it did not contain the entire coding sequence of HMG-CoA synthase. A fragment of 200 base pairs from the 5' end of the insert of p53K-214 was then labeled with 32P and used as a probe for further screening of the same sublibrary. This screen of -8 X lo4 recombinants yielded plasmid p53K-312, a full- length cDNA (Fig. 1).

A ~ i ~ d ~ s against S ~ ~ t ~ ~ i e Peptides Peptides corresponding to deduced amino acids 303-317 and 506-

520 of hamster HMG-CoA synthase (Fig. 2) were synthesized by Peninsula Laboratories, Inc. with the use of solid-phase methods developed by Marglin and Merrifield (32). In both peptides a cysteine residue was added to the NHz-terminal end. The compositions of the peptides were confirmed hy amino acid analysis. The peptides were coupled to keyhole limpet hemocyanin with ma~eimidobenzoyl-N- hydroxysuccinimide ester (33). New Zealand White rabhits were injected subcutaneously on days 0 and 14 with one of the synthetic peptides (0.2 mg) coupled to hemocyanin. The peptide-hemocyanin conjugates were emulsified with Freund's complete adjuvant (day 0) or incomplete adjuvant (day 14) in a total of 1 ml. An intraperitoneal injection of synthetic peptide (0.2 mg) coupled to hemocyanin was given on day 28 and 7 days before each bleeding in a total of 1 ml containing 4 mg of Al(OH)a. Rabbits were bled first on day 35 and

3712 cDNA for Cytoplasmic HMG-CoA Synthase then 7 days after each booster injection. IgG fractions were purified on Protein A-Sepharose CL-4B (34). Immunoblotting of HMG-CoA synthase was carried out as previously described (35).

HMG-CoA Synthase Assay A modification (3) of the radiochemical method of Clinkenbeard

et aZ. (11) was used to measure the synthesis of [14C]HMG-CoA from ['4C]acetyl-CoA. The standard reaction contained the following com- ponents in a final volume of 0.2 ml: 0.1 M Tris-chloride, pH 8; 10 p M acetoacetyl-CoA; 0.1 mM EDTA 20 mM MgC12; 36 pg of cytosolic protein; and 0.2 mM [l-'4C]acetyl-CoA (13,000 cpm/nmol). The re- action mixture was preincubated for 2 min at 30 "C, after which the reaction was started by addition of the [l-'4C]acetyl-CoA. Aliquots (50 pl) were removed at intervals of 3, 6, and 9 min and added to a scintillation counting vial that contained 0.3 ml of 6 N HCl. Each counting vial was then heated in an oven at 95 "C for 2 h, and the nonvolatile radioactivity was counted to determine the amount of [14C]HMG-CoA formed. The amount of nonvolatile 14C radioactivity in the absence of protein was subtracted from the values at 3, 6, and 9 min. All values were linear with respect to time and protein concentration.

Purification of Hamster Cytoplasmic HMG-CoA Synthase All operations were carried out a t 0-4 "C. The 100,000 X g fraction

from 58 g of Syrian hamster liver was prepared as described above from eight animals that had been fed with a chow diet supplemented with cholestyramine and mevinolin. HMG-CoA synthase was purified by four chromatographic procedures, which were adopted from pre- viously published procedures (2, 13,16). The purification was carried out sequentially as described below.

DEAE-cellulose Chromatography-The 100,000 X g fraction (2.2 g of protein) was loaded onto a column of DEAE-cellulose (10 X 2.3 cm) equilibrated with buffer A at a flow rate of 50 ml/h. The column was then washed with -700 ml of the same buffer. The column was eluted with a 300-ml linear gradient of 0-200 mM NaCl in buffer A, and fractions of 5 ml were collected. The fractions containing the peak of HMG-CoA synthase activity (eluting at a NaCl concentration of -130 mM) were combined, concentrated by precipitation at 50% saturated (NH4)2S04, and dialyzed overnight against buffer B.

Blue Sepharose Chromatography-The dialyzed solution from step 1 (45 mg of protein) was loaded onto a column of Blue Sepharose CL- 6B (8 X 2.3 cm) equilibrated in buffer B. After washing the column with the same buffer, the enzyme was eluted with 100 ml of buffer B supplemented with 250 mM NaC1. The active fractions were dialyzed overnight against buffer C.

Cellulose Phosphate Chromatography-This step was carried out as described by Lowe and Tubbs (16). The dialyzed protein from step 2 (7 mg) was applied to a column of cellulose phosphate (8 X 2.3 cm) that had been equilibrated and washed with buffer C. The enzyme was eluted with 60 ml of 0.2 mM acetoacetyl-CoA dissolved in buffer C. The active fractions were combined and concentrated to about 1.7 ml by ultrafiltration through a Diaflo PM-30 membrane in a stirred Amicon cell and dialyzed against buffer D.

Hydroxyapatite Chromatography-The protein from step 3 (2.4 mg) was applied to a column of hydroxyapatite (2 X 1.5 cm), which was equilibrated and washed with buffer D. The enzyme was eluted with a 100-ml linear gradient of 10-400 mM sodium phosphate in buffer D. The protein (-1 mg) was concentrated, desalted by repeti- tive filtration and dilution with water through a Amicon Centricon 30 microconcentrator, and used directly for determination of amino acid sequence.

Protein Sequence of HMG-CoA Synthase The amino acid sequence of the NH2 terminus of purified Syrian

hamster HMG-CoA synthase was performed on an Applied Biosys- tems model 470A gas phase sequencer using standard programming. The phenylthiohydantoin amino acid derivatives were identified by reverse-phase high performance liquid chromatography on a Waters Associates high performance liquid chromatography system with a Waters NOVA-PAK C18 steel column. Chromatographic conditions were as previously described (36). A single sequenator run was per- formed on 175 pmol of protein. The average repetitive yield for this run was 90.2%.

Measurement of HMG-CoA Reductase and Synthase mRNA by DNA-Excess Solution Hybridization

The content of reductase and synthase mRNA in hamster liver was measured by DNA-excess solution hybridization (37) as previ- ously described (38, 39). For measurement of synthase mRNA, an M13 DNA template of synthase cDNA was prepared by subcloning a 435-base pair PstI-PstI fragment (nucleotides -35 to 401) from the plasmid p53K-312 into the PstI site of the M13mp8 vector. A single- stranded 3ZP-probe was synthesized with a universal sequencing primer 17 nucleotides in length; 0.25 mM dTTP, dATP, and dGTP; 40 p~ [ C L - ~ ~ P I ~ C T P ; and the Klenow fragment of Escherichia coli DNApolymerase I. After incubation for 1 h a t 37 "C, a single-stranded fragment (265 nucleotides) was isolated by EcoRV digestion followed by 7 M urea, 5% polyacrylamide gel electrophoresis and hydroxylapa- tite chromatography (37). The single-stranded [32P]cDNA probe (-2.9 X lo4 cpm/fmol) was used to measure synthase mRNA by the same procedure used to measure reductase mRNA (38, 39). The single-stranded reductase 32P-probe was prepared as described (38, 39) and had a specific activity of 3.5 x lo4 cpm/fmol.

RESULTS

Fig. 1 shows the restriction map of p53K-312, the plasmid that contains a full-length cDNA for HMG-CoA synthase. Also shown is the structure of p53K-3, the previously de- scribed plasmid (20) that was used as a probe to obtain the full-length cDNA. The full-length cDNA corresponds to an mRNA of 3.3 -kilobases in length, of which 1560 base pairs represent the protein-coding region and 1640 base pairs rep- resent an unusually long 3' untranslated region.

Fig. 2 shows the complete nucleotide sequence of the cDNA and the predicted amino acid sequence of the cytoplasmic HMG-CoA synthase protein. The position of the initiator methionine was confirmed by protein sequence experiments (see below). Amino acids 114-134 (solid underline in Fig. 2) show a 19 out of 21 match with the previously determined sequence of the active site of the avian mitochondrial HMG- CoA synthase (15). A computer search for homology between the amino acid sequence of hamster cytoplasmic HMG-CoA synthase and those of proteins stored in a data bank collected by R. F. Doolittle of the University of California, San Diego (version of August 1985) revealed no significant similarities to any sequence in the bank.

To confirm that this cDNA represented cytoplasmic HMG- CoA synthase, we prepared anti-peptide antibodies directed against two predicted peptides, representing amino acids 303- 317 and amino acids 506-520, the latter corresponding to the

- p53K-3

egy for the hamster HMG-CoA synthase cDNAs. The scale at FIG. 1. Restriction endonuclease map and sequencing strat-

the top designates the nucleotide positions (in kilobases) relative to the protein initiation codon. The thick black line represents the 520- amino acid-coding region. The regions encompassed by the cDNA inserts in plasmids p53K-3 and p53k-312 are indicated below the restriction endonuclease map. The arrows indicate the direction and extent of DNA sequence established in a given experiment. Vertical bars at the ends of the arrows indicate that the dideoxy chain termi- nation method was used solid circles indicate that the chemical method was used double slash marks indicate that the restriction sites used for sequencing are located in the vector DNA. Sequencing reactions not coinciding with restriction endonuclease sites employed specific oligonucleotides as primers on M13 subclones.

cDNA for Cytoplasmic HMG-CoA Synthase 3713

FIG. 2. Nucleotide sequence of the cDNA strand corresponding to the hamster HMG-CoA synthase mRNA and the predicted amino acid se- quence of the protein. The nucleotides are numbered on the right-hand side; nucleotide 1 is the A of the ATG codon that encodes the initiator methionine; negative numbers refer to the 5’ un- translated region. The amino acids are numbered underneath the sequence; res- idue 1 is the initiator methionine. In the 5’ untranslated region, the nucleotide at position -99 was a G in the cDNA, whereas a T was obtained in the genomic clone (see accompanying paper (21)). In the coding region, the nucleotide at po- sition 788 was a C in the cDNA, whereas a T was obtained in the genomic clone; this single base affects the amino acid assignment at position 263 (C -+ serine; T -+ phenylalanine). The four polyaden- ylation signals used to position the 3’ end at the mRNA are boxed (see accom- panying paper (21)). The amino acid se- quence homologous to the active site of avian mitochondrial HMG-CoA syn- thase (15) is denoted by the solid under- line. The amino acid sequence corre- sponding to two synthetic peptides that were used to raise anti-peptide antibod- ies (residues 303-317 and 506-520) are indicated by the dashed underlines. 95% of the sequence in the coding region was determined on both strands of the cDNA, and 90% of the coding sequence shown in the figure was confirmed by sequencing genomic DNA (21). The DNA sequence of the 3’ untranslated

the cDNA. region was determined on one strand of

-1

120

240

360

480

600

720

840

960

extreme COOH terminus of the protein. We then prepared a 100,000 x g supernatant fraction from livers of hamsters that had been fed cholestyramine plus mevinolin, a regimen that induces high levels of cytoplasmic HMG-CoA synthase activ- ity. Aliquots of the extract were incubated either with the anti-COOH-terminal peptide IgG or with an IgG fraction from a nonimmunized rabbit. The antibody/protein com- plexes were then adsorbed to Stuphylococcus Protein A. As shown in Fig. 3A, increasing amounts of the anti-peptide IgG progressively removed increasing amounts of the synthase activity from the supernatant. A large proportion of the immunoprecipitated enzymatic activity was recovered in the pellet (Fig. 3B).

Fig. 4 shows that HMG-CoA synthase activity and the protein that reacts with the anti-peptide (506-520) antibody were eluted together on a DEAE-cellulose column. By im- munoblotting after sodium dodecyl sulfate-gel electrophoresis, the antibody stained a protein of 53 kDa, which corresponds to the previously reported molecular weight for cytoplasmic HMG-CoA synthase from the liver of the chicken and the rat (2).

To obtain further confirmation of the identity of the cDNA and HMG-CoA synthase, we purified cytoplasmic HMG-CoA synthase from hamster liver to homogeneity using four chro- matographic steps that were adopted from previous publica- tions (see “Experimental Procedures”). The purified enzyme

showed a major band at 53 kDa and a minor band at 48 kDa after Coomassie staining (Fig. 5, Lune B) . The major band at 53 kDa and the minor band at 48 kDa both reacted with the anti-peptide antibody directed against amino acids 303-317 of HMG-CoA synthase (Fig. 5 , Lune D). However, only the 53-kDa protein reacted with the anti-peptide antibody di- rected against the predicted COOH terminus of synthase (Fig. 5, Lune C). We interpret these data to indicate that the smaller fragment represents a proteolytic degradation product which has lost the COOH terminus.

The purified HMG-CoA synthase was subjected to amino acid sequencing, and the results were compared with the amino acid sequence predicted from the cDNA sequence. As shown in Table I, the protein sequence showed that the initiator methionine had been removed. Thereafter, there was an excellent match with the amino acid sequence predicted from cDNA sequence, with the exception of a threonine (amino acid sequence) uersus alanine (cDNA sequence) at position 9 of the predicted protein sequence. We attribute this difference to a polymorphism between the Syrian hamster, from which the protein sequence was obtained, and the Chinese hamster, from which the cDNA was cloned.

In previous studies we demonstrated that the mRNA for the 53-kDa protein (now identified as HMG-CoA synthase) was under feedback regulation by cholesterol in cultured hamster UT-1 cells (19, 20). Fig. 6 shows that cholesterol

3714 cDNA for Cytoplasmic HMG-CoA Synthase

A. Supernotant Fraction 8. Pellet

100~-0-0-0- 0 -

\ ‘ Nonimmune

-

40 Immune 0

20 -

0 f l l l l l

0 0.3 0.6 0.9 0 0.3 0.6 0.9

Ant i - Peptide IgG (rngltube)

FIG. 3. Immunoprecipitation of HMG-CoA synthase activ- ity from hamster liver cytosol by anti-peptide antibody di- rected against deduced amino acid sequence of HMG-CoA synthase. A 100,000 X g cytosolic fraction was prepared from the livers of hamsters fed a diet supplemented with cholestyramine and mevinolin as described under “Experimental Procedures.” Aliquots of cytosol (1.2 mg of protein in 50 pl of buffer A) were pretreated for 10 min a t room temperature with varying amounts of Pansorbin in a final volume of 0.6 ml of Dulbecco’s phosphate-buffered saline con- taining 1% (v/v) Emulphogene BC-720 and 0.1% (w/v) bovine serum albumin (buffer E). The amount of Pansorbin was varied according to the amount of IgG to be added subsequently (0.5 pl of 20% (w/v) Pansorbin added per pg of IgG). After pretreatment with Pansorbin, the mixture was centrifuged a t 12,500 X g for 5 min a t room temper- ature, and aliquots of the resulting supernatant (0.45 ml) were incu- bated with 0.15 ml of buffer E containing the indicated amount of nonimmune IgG (0) or immune anti-COOH-terminal IgG (0). After incubation for 4 h a t room temperature, varying amounts of Pansor- bin (0.5 pl of 20% Pansorbin/pg of IgG) were again added to each reaction. After incubation for 20 min a t room temperature, the immunoprecipitates (pellet fraction) were isolated by centrifugation (12,500 X g for 5 min) and washed once by vigorous vortexing in the presence of 0.6 ml of buffer E. Each pellet fraction (resuspended in 0.6 ml of buffer E) and the combined supernatant and wash fractions from the original immunoprecipitation (in a volume of 1 ml of buffer E) were assayed for HMG-CoA synthase activity as described under “Experimental Procedures.” Each assay was performed with 20 p1 of the pellet and 30 p1 of the supernatant fraction. The “100% of initial activity” for the unfractionated cytosol was 0.11 nmol/min/pl or 4.6 nmol/min/mg of protein.

feeding to hamsters also causes a profound decrease in the amount of mRNA for cytoplasmic HMG-CoA synthase in the liver. To perform this assay, increasing amounts of total cellular RNA were incubated with a uniformly 3’P-labeled single-stranded cDNA probe derived from p53K-312. The amount of hybridizable mRNA was quantitated by digestion with Sl nuclease and precipitation with trichloroacetic acid. Cholesterol feeding for 14 days led to a more than 90% reduction in the amount of hybridizable mRNA for HMG- CoA synthase (Fig. 6A). The relative magnitude of fall in mRNA for HMG CoA synthase was greater than the fall in the amount of mRNA for HMG CoA reductase in the same livers (Fig. 6B).

DISCUSSION

In the current paper, we report the isolation and nucleotide sequence of a full-length cDNA for the cytoplasmic form of HMG-CoA synthase from hamster. The identification of the cDNA as HMG-CoA synthase was based on multiple criteria: 1) anti-peptide antibodies directed against the predicted amino acid sequence precipitated cytoplasmic HMG-CoA syn- thase activity from hamster liver extracts; 2) the anti-peptide antibodies reacted with purified hamster liver HMG-CoA synthase on immunoblots; 3) the NH2-terminal amino acid

NaCl Gradient

I

Fraction Number

FIG. 4. Copurification of HMG-CoA synthase activity and anti-peptide immunoreactivity after DEAE-cellulose chro- matography of hamster liver cytosol. The 100,000 X g fraction of cytosol (1100 mg of protein) was loaded onto a DEAE-cellulose column and eluted with a NaCl gradient as described under “Exper- imental Procedures.” Aliquots of each 5-ml fraction (50 pl) were removed for assay of HMG-CoA synthase activity (0). Another 50- p1 aliquot of each fraction was mixed with an equal volume of buffer containing 62.5 mM Tris-chloride, 15% (w/v) sodium dodecyl sulfate, 8 M urea, 10% (w/v) sucrose, and 100 mM dithiothreitol at pH 6.8 and heated a t 90 “C for 5 min. The sample was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, transferred to ni- trocellulose, and incubated with 20 pg/ml anti-peptide (506-520) rabbit IgG. The nitrocellulose was subsequently incubated with lZ5I- labeled goat anti-rabbit IgG (lo6 cpm/ml; lo4 cpm/ng), washed, and exposed to x-ray film for 40 h at -20 “C. Autoradiograms of the immunoblots of fractions 24-45 are shown in the inset. The relative intensity of the 53-kilodalton band was determined by densitometric scans of the radioautographs, and the data were plotted (0).

Coomassie- Immuno- staining blots

mm 200- a

3 97- 91

1 6 8 - m 0

43 - - ~ r ) -m

FIG. 5. Coomassie stain and immunoblotting of purified hamster liver HMG-CoA synthase. Synthase was purified with the four chromatographic steps described under “Experimental Pro- cedures.” Aliquots of the purified enzyme (20 pg of protein) were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophore- sis in lunes B-D. Protein standards of known molecular weight were subjected to electrophoresis in lane A. Lunes A and B were stained with Coomassie Brilliant Blue G-250. In lunes C and D the proteins were transferred to nitrocellulose and incubated with 20 pg/ml of either anti-peptide (506-520) IgG (lune C) or anti-peptide (303-317) IgG (lane D). The nitrocellulose was subsequently incubated with Iz5I- labeled goat anti-rabbit IgG (lo6 cpm/ml; lo4 cpm/ng). The nitrocel- lulose was exposed to x-ray film for 3 h a t -70 “C with intensifying screen.

sequence predicted from the cDNA sequence was almost identical to the amino acid sequence determined with the purified protein; and 4) the predicted amino acid sequence contained a region that was highly homologous to the previ- ously determined sequence of the active site of avian mito-

cDNA for Cytoplasmic HMG-CoA Synthase

TABLE I

3715

Amino acid sequence of the NHz-terminal end of HMG-CoA synthase: comparison of results by direct chemical analysis (Syrian hamster) and cDNA cloning (Chinese hamster)

One amino acid residue that could not be assigned by Edman degradation is denoted by “?”. The asterisk (*) indicates the single amino acid difference observed between the amino acid sequence of the Syrian hamster synthase and the cDNA sequence of the Chinese hamster synthase.

Species Methodof analysis

Aminoacid sequence

Syrian Edmandegra- ProGly Ser LeuProLeuAsnThr GluAla ? TrpProLys AspVal G l y I l e V a l A l a L e u G l u hams te r da t ionof

p u r i f i e d p r o t e i n

*

Chinese Predic t ion 1 10 20 hamster f romnucleo- Met ProGly Ser LeuProLeuAsnAlaGluAlaCysTrpProLys AspVal Gly I leVal AlaLeuGlu

t i d e s e q u e n c e of cDNA

I E 2 0 L

0 25 50 75 100

B. HMG CoA Reductase

4

3

2 +Cholesterol

100 200 300 0

Total Cellular RNA (pg)

FIG. 6. Suppression of mRNA for cytoplasmic HMG-CoA synthase and HMG-CoA reductase in the livers of hamsters fed cholesterol. Hamsters were fed a chow diet alone or supple- mented with 2% cholesterol for 14 days as described under “Experi- mental Procedures.” Total liver RNA was hybridized with a single- stranded 32P-labeled synthase cDNA probe (panel A ) or with a single- stranded 32P-labeled reductase cDNA probe (panel B) , and the hybrids were treated with S1 nuclease followed by trichloroacetic acid precip- itation as described under “Experimental Procedures.” Blank values of 1.6 X lo3 and 0.9 X lo3 cpm (representing the amount of trichlo- roacetic acid-precipitable S1-resistant radioactivity in the absence of RNA) were subtracted from each value for the synthase and reductase reactions, respectively.

chondrial HMG-CoA synthase (15). Several new insights into the biology of HMG-CoA syn-

thase have emerged from the current studies. First, the mRNA for the cytoplasmic form of this enzyme is shown to be regulated in parallel with the mRNA for HMG-CoA reductase, the next enzyme in the pathway. Since the reductase mRNA is regulated, at least in part, by transcriptional control (19, 40), it seems likely that the synthase will be under similar regulation. Indeed, the 5’ ends of both genes show certain homologous features, as described in the accompanying paper (21). Presumably these two genes interact with a common sterol-regulatory protein. It seem likely that other genes in the cholesterol biosynthetic pathway will also be subject to similar repression.

A second insight involves the relation between the mito- chondrial and cytoplasmic forms of HMG-CoA synthase. The cDNA that we have cloned specifies the cytoplasmic enzyme on the basis of several observations: 1) the 53-kDA protein that it encodes was found in the 100,000 X g supernatant of UT-1 cells (20) and hamster liver (Figs. 3-5); and 2) the amount of this mRNA in liver is repressed by cholesterol feeding, a finding that is consistent with the ability of choles-

terol to regulate the cytoplasmic form of HMG-CoA synthase (2, 11).

The active site of the mitochondrial form of HMG-CoA synthase must closely resemble that of the cytoplasmic form. Miziorko and Behnke (14,15) purified the mitochondrial form of the enzyme from chicken liver and used an active site- specific covalent modification reagent to identify the petpide that contained the active site. The sequence of this peptide is nearly identical to amino acids 114-134 predicted for the hamster cytosolic enzyme (Fig. 2 and Ref. 15). Thus, the sequence of the active site is conserved not only across major species barriers but also between the mitochondrial and cy- toplasmic forms of the enzyme. It is likely that the mitochon- drial enzyme is encoded by a different gene than the cyto- plasmic form. Alternatively, it is possible that the mitochon- drial and cytoplasmic forms of the enzyme are synthesized from the same gene, but the mitochondrial form differs from the cytoplasmic form by means of alternative mRNA splicing.

A third insight derived from these studies relates to the general mechanism by which cells may become resistant to competitive inhibitors of essential enzymes. Up to now such resistance has emerged from overproduction of the enzyme, frequently as a result of gene amplification, as is the case with resistance to inhibitors of dihydrofolate reductase and aden- osine deaminase (41). Indeed, UT-1 cells have also used the gene amplification mechanism to overproduce HMG-CoA re- ductase in the presence of its competitive inhibitor, compactin (19). However, in addition to overproducing the reductase, the UT-1 cells also overproduce HMG-CoA synthase. Since compactin competes with HMG CoA for the reductase enzyme (42), the UT-1 cells acquire an increase in their resistance to compactin by overproducing HMG-CoA. The overproduction of HMG-CoA synthase is a result of increased transcription of the mRNA for the enzyme. In contrast to the reductase, the gene for HMG-CoA synthase is not amplified in UT-1 cells (19). It is possible that the half-life of the synthase mRNA is also prolonged in UT-1 cells, but this has not been studied in detail. These findings indicate that cells may be- come resistant to competitive enzyme inhibitors by overpro- ducing the substrate as well as the inhibited enzyme.

In an accompanying paper we isolate and characterize the gene that gives rise to the mRNA for cytoplasmic HMG-CoA synthase (21). Through further studies of this gene, it should be possible to determine the basis for the coordinate regula- tion of HMG-CoA reductase and HMG-CoA synthase.

Acknowledgments-We thank Gloria Brunschede and Tom Fischer for excellent technical assistance; Henry Miziorko for communicating results prior to publication; Kenneth Luskey, David Russell, and Laura Liscum for helpful discussions; Russell Doolittle for performing

3716 cDNA for Cytoplasmic HMG-CoA Synthase the computer search for amino sequence homologies; and Kenneth stein, J. L., and Brown, M. S. (1982) Proc. Natl. Acad. Sci. U. Luskey for critical review of the manuscript. S. A. 7 9 , 6210-6214

21. Gil, G., Brown, M. S., and Goldstein, J. L. (1986) J. Biol. Chem.

1.

2.

3.

4.

5.

6.

7.

8.

9.

10. 11.

12.

13. 14.

15.

16. 17.

18.

19.

20.

REFERENCES Brown, M. S., and Goldstein, J. L. (1980) J. Lipid Res. 21, 505-

Clinkenbeard, K. D., Sugiyama, T., Reed, W. D., and Lane, M.

Balasubramaniam, S., Goldstein, J. L., and Brown, M. S. (1977)

Chang, T-Y., and Limanek, J. S. (1980) J. Biol. Chem. 255 ,

Schnitzer-Polokoff, R., Torget, R., Logel, J., and Sinensky, M. (1983) Arch. Biochem. Biophys. 227 , 71-80

Faust, J. R., Goldstein, J. L., and Brown, M. S. (1979) Proc. Natl. Acad. Sci. U. S. A. 7 6 , 5018-5022

Slakey, L. L., Craig, M. C., Beytia, E., Briedis, A., Feldbruegge, D. H., Dugan, R. E., Qureshi, A. A., Subbarayan, C., and Porter, J. W. (1972) J. Biol. Chem. 247,3014-3022

Gonzalez, R., Carlson, J. P., and Dempsey, M. E. (1979) Arch. Biochem. Biophys. 196, 574-580

Gaylor, J. L. (1981) in Biosynthesis of Isoprenoid Compounds (Porter, J. W., and Spurgeon, S. L., eds) Vol. I, pp. 481-543, John Wiley and Sons, New York

517

D. (1975) J. Biol. Chem. 250 , 3124-3135

Proc. Natl. Acad. Sci. U. S. A. 7 4 , 1421-1425

7787-7795

Siperstein, M. D. (1970) Curr. Top. Cell. Regul. 2 , 65-100 Clinkenbeard, K. D., Reed, W. D., Mooney, R. A., and Lane, M.

Reed, W. D., Clinkenbeard, K. D., and Lane, M. D. (1975) J. Bwl.

Miziorko, H. M. (1985) Methods Enzymol. 110,19-26 Miziorko, H. M., and Behnke, C. E. (1985) Biochemistry 2 4 ,

Miziorko, H. M., and Behnke, C. E. (1985) J. Biol. Chem. 260 ,

Lowe, D. M., and Tubbs, P. K. (1985) Biochem. J. 227,591-599 Chin, D. J., Luskey, K. L., Faust, J. R., MacDonald, R. J., Brown,

M. S., and Goldstein, J. L. (1982) Proc. Natl. Acad. Sci. U. S.

Chin, D. J., Gil, G., Russell, D. W., Liscum, L., Luskey, K. L., Basu, S. K., Okayama, H., Berg, P., Goldstein, J. L., and Brown, M. S. (1984) Nature 308 , 613-617

Luskey, K. L., Faust, J. R., Chin, D. J., Brown, M. S., and Goldstein, J. L. (1983) J. Bid. Chem. 258,8462-8469

Luskey, K. L., Chin, D. J., MacDonald, R. J., Liscum, L., Gold-

D. (1975) J. Biol. Chem. 250,3108-3116

Chem. 250,3117-3123

3174-3179

13513-13516

A. 79,7704-7708

261,3717-3724 22. Caruthers, M. H., Beaucage, S. L., Becker, C., Efcavitch, W.,

Fisher, E. F., Galluppi, G., Goldman, R., deHaseth, P., Martin, F., Matteucci, M., and Stabinsky, Y. (1982) Genetic Engineer- ing: Principles and Methods (Setlow, J. K., and Hollaender, A.) Vol. 6, pp. 1-17, Plenum Publishing Corp., New York

23. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, pp. 1-545, Cold Spring Harbor Laboratory Press, New York

24. Feinberg, A. P., and Vogelstein, B. (1983) Anal. Biochem. 132,

25. Maxam, A. M., and Gilbert, W. (1980) Methods Enzymol. 6 5 , 6-13

499-5~n 26. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl.

27. Messing. J. (1983) Methods Enzvmol. 101. 20-78

"_ " _ Acad. Sci. U. S. A. 74 , 5463-5467

28. Chirgwin, J."., Przybyla, A. E , MacDonald, RI J., and Rutter,

29. Glisin, V., Crkvenjakov, R., and Byus, C. (1974) Biochemistry

30. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J.

31. Okayama, H., and Berg, P. (1983) Mol. Cell. Biol. 3,280-289 32. Marglin, A., and Merrifield, R. B. (1970) Annu. Reu. Biochem.

33. Schneider, W. J., Slaughter, C. J., Goldstein, J. L., Anderson, R. G. W., Capra, J. D., and Brown, M. S. (1983) J. Cell Biol. 9 7 ,

34. Beisiegel, U., Schneider, W. J., Goldstein, J. L., Anderson, R. G. W., and Brown, M. S. (1981) J. Biol. Chem. 256,11923-11931

35. Liscum, L., Luskey, K. L., Chin, D. J., Ho, Y. K., Goldstein, J. L., and Brown, M. S. (1983) J. Biol. Chem. 258 , 8450-8455

36. Ricci, T. M. (ed) (1984) Chromatography Notes (1984) Vol. 11, p. 6, Millipore Waters Chromatography Division, Milford, MA

37. Williams, D. L., Dawson, P. A., Newman, T. C., and Rudel, L. L. (1985) J. Biol. Chem. 260, 2444-2451

38. Chin, D. J., Gil, G., Faust, J. R., Goldstein, J. L., Brown, M. S., and Luskey, K. L. (1985) Mol. Cell. Biol. 5 , 634-641

39. Gil. G.. Faust. J. R.. Chin. D. J.. Goldstein. J. L.. and Brown. M.

W. J. (1979) Biochemistry 18, 5294-5299

13,2633-2637

(1951) J. Biol. Chem. 193,265-275

39,739-866

1635-1640

S . (1985) Cell 41,' 249-258 '

40. Osborne, T. F., Goldstein, J. L., and Brown. M. S. (1985) Cell 42,203-212

. .

41. Schimke, R. T. (1984) Cancer Res. 44,1735-1742 42. Endo, A., Kuroda, M., and Tanzawa, K. (1976) FEBS Lett. 72 ,

323-326