Properties of Purified Rat Hepatic 3-Hydroxy-3-methylglutaryl ...

THEJOURNAL OF BIOLOGICAL CHEMIsTRY Vol. 255, No. 8, Issue of April 25. pp. 3715-3725, 1980 Printed in U.S.A.

Properties of Purified Rat Hepatic 3-Hydroxy-3-methylglutaryl Coenzyme A Reductase and Regulation of Enzyme Activity* (Received for publication, August 6, 1979, and in revised form, November 16, 1979)

Peter A. Edwards,+ Donna Lemongello, John Kane,@Ishaiahu Shechter.7 and Alan M. Fogelman )I From the Division of Cardiology, Department OfMedicine, University of California, Los Angeles, Los Angeles, California 90024 and the §Department OfMedicine, University of California, Sun Francisco, Sun Francisco, California 9414.3

3-Hydroxy-3-methylglutarylcoenzyme A (HMG-CoA) reductase from rat liver microsomes has been purified to apparent homogeneity with recoveries of approximately 50%. The enzyme obtained from rats fed a diet supplemented with cholestyramine had speciiic activities of approximately 21,600 nmol of NADPH oxidized/ min/mg of protein. After amino acid analysis a specific activity of 31,000 nmol of NADPH oxidized/min/mc of amino acyl mass was obtained. The 8 2 0 . ~for HMG-I ‘oA reductase was 6.14 S and the Stokes radius was .39 nm. The molecular weight of the enzyme was 104,JOO and the enzyme subunit after sodium dodecyl sulfatepolyacrylamide gel electrophoresis was 52,000. Antibodies prepared against the homogeneous enzyme specifically precipitated HMG-CoA reductase from crude and purefractions of the enzyme. Incubation of rat hepatocytes for 3 h in the presence of lecithin dispersions, compactin, or rat serum resulted in significant increases in the specific activity of the microsomal bound reductase. Immunotitrations indicated that in all cases these increases were associated with an activated form of the reductase. However activation of the enzyme accounted for only a small percentage of the total increase in enzyme activity; the vast majority of the increasewas apparently due to an increase in the number of enzyme molecules. In contrast, when hepatocytes were incubated with mevalonolactone the lower enzyme activity which resulted was primarily due to inactivation of the enzyme with little change in thenumber of enzyme molecules. Immunotitrations of microsomes obtained from rats killed at the nadir or peak of the diurnal rhythm of 3hydroxy-3-methylglutaryl-CoA reductase indicated that the rhythm results both from enzyme activation and anincreased number of reductase molecules.

The regulation of rat hepatic microsomal 3-hydroxy-3methylglutaryl-CoA reductase, the rate-limiting enzyme of cholesterol biosynthesis (l), has been studied extensively by measuring changes in activity of the enzyme in intact rats (18 ) and in isolated hepatocytes (6, 8-10). Such measurements do not indicate whether changes in * This work wassupported in part by United States Public Health Service Grants HL 19063, HL 22476, and HL 20807 and by a grant from the Edna and George Castera Fund. 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. $ An Established Investigator of the American Heart Association. 1On leave from the Department of Biochemistry, Tel Aviv University, Israel. 11 The recipient of United States Public Health Service Research Career Development Award HL-00426.

enzyme activity result from alterations in the rates of synthesis and/or degradation of HMG-CoA reductase’ and/or from alterations in the catalytic activity of the enzyme. Detailed studies on the regulation of this enzyme are dependent on the availability of homogeneous preparations of HMG-CoA reductase and on antibody specific for the enzyme. A number of investigators have reported purification of the ratliver enzyme (4, 11-16); however, the specific activities of the pure enzyme varied from 6.0 to 19,600 nmol of NADPH oxidized/min/mg of protein and the molecular weightof the subunitis reported to be between 47,000 and 120,000. We report here asimple method of purifying rat liver HMGCoA reductase and the amino acid composition of the enzyme. We have prepared antibody in rabbits against the pure enzyme and have used immunotitrations to determine whether different catalytic forms of the enzyme can exist in the microsomal membrane of isolated rat hepatocytes. Using these techniques we have also re-investigated the nature of the circadian rhythm of HMG-CoA reductase in intact rats. EXPERIMENTALPROCEDURES

Materials-Chemicals were obtained from the sources indicated HMG-CoA, NADPH, CoASH, dithiothreitol, Triton X-100,phenylmethylsulfonyl fluoride, yeast glucose-6-phosphate dehydrogenase, Sigma; agarose-HMG-CoA (V), P-L Biochemicals. Compactin (ML236-B) was a generous gift from Dr. R. Fears (Beechams, United Kingdom) and Dr. A. Endo (Tokyo N6k5 University, Japan). Compactin was converted to the anion as described by Endo et al. (17) and stored at pH 12.0 at -20°C. Cholestyramine (Questran), Mead Johnson; sodium dodecyl sulfate, acrylamide, Bio-Rad; Ouchterlony plates, Hyland; formic acid, Fluka, A.G., Sepharose 6B, Pharmacia. The sources of all other materials have been previously given (4). AnimaZs-Rats were housed under areverse illumination cycle (8), had free access to food and water, and were, where noted, fed powdered rat food supplemented with 5%cholestyramine for 4 days before they were killed.Rat hepatocytes were prepared as previously described (8). Solubilization and Assay of HMG-GOAReductase-Unless otherwise stated, the activity of the microsomal bound reductase and the solubilized enzyme were determined by the radioassay and spectrophotometric assay, respectively, as previously described (18). Microsomes were routinely prepared by two centrifugations at 100,OOO X g (19) and were routinely preincubated for 20 min a t 37°C before assay. Under these conditions we found no evidence of the I4C compound that has been reported by Ness and Moffler (20) to eluteasa contaminant with mevalonolactone from the AGl-X8 formate columns. Background values were the same for blanks assayed in the absence of either cofactors, glucose-&phosphate dehydrogenase, or enzyme. Spectrophotometric assays were carried out in Buffer A (0.2 M KCl, 0.16 M potassium phosphate, 0.004 M EDTA, and 0.01 M dithiothreitol, pH 6.8) together with 0.2 m~ NADPH and 0.1 mM RSHMG-CoA. The solubilization of the enzyme was as previously described (18), except that the microsomes werehomogenized with

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

3715

Regulation

3716

of HMG-CoA

Buffer B (0.1 M KCl, 0.08 M potassium phosphate, 2 EDTA, 10 mM dithiothreitol, pH 7.2) diluted with an equal volume of glycerol. Subsequent dilutions were as previously described (18), except that Buffer B was used as the diuent. Purification of HMG-CoA Reductase-The method is modified from that previously reported by Edwards et al. (4). However, the present method required only one affinity column and gave yields of approximately 50% and enzyme of specific activity similar to those already reported by Edwards et al. (4, 18) and KIeinsek et al. (13). The reductase was purified at room temperature except where otherwise noted. The solubilized enzyme was fractionated with solid ammonium sulfate and the protein precipitating between 35% and 50% ammonium sulfate was dissolved in Buffer B (pH 6.81, containing 30% glycerol and 1.0 M KC1 at a protein concentration between 8 and 12 mg/ml. Aliquots (4 ml) were heated for 20 min at 65”C, diluted with an equal volume of Buffer B (pH 6.8), and centrifuged at 100,000 x g for 30 min. The supernatant was removed and the protein precipitating between 0% and 60% ammonium sulfate collected by centrifugation. The pellet was dissolved in a small volume of Buffer B (pH 6.8) and was routinely stored overnight at 4°C under nitrogen with no loss of activity. The enzyme was warmed at 37°C for 30 min, centrifuged at 100,000 X g for 45 min, and the supernatant containing all the enzyme activity was removed and diluted l.l-fold with 5 mu dithiothreitol. This solution was applied to an Agarose-Hexane-HMG-CoA affinity column (0.25 to 1.5 ml), the column washed with Buffer C (0.05 M KCl, 0.04 M potassium phosphate, 4 mu EDTA, 5 mM dithiothreitol, pH 6.8) until the AZBO = 0 (approximately 12 column volumes), and the enzyme eluted in 5 column volumes of Buffer C containing 1 MM CoASH. The agarose-HMG-CoA resin has been used repeatedly for over 1 year with no loss of affinity for HMG-CoA reductase. Amino Acid Analysis-To remove the buffer salts and CoA the protein was precipitated from solution with 7 volumes of 20% trichloroacetic acid (Mallinckrodt). The protein was centrifuged at 5000 rpm in a Sorvall desk top centrifuge and the precipitate washed once with trichloroacetic acid, and twice with 7 volumes of acetone. This method effectively removed all the CoA because /3-alanine, a normal product of CoA, was absent from the amino acid analysis. Hydrolysis of the reductase in the presence of CoA gave artifacts in the region of serine, glycine, cysteic acid, and in addition gave a discrete peak of /3-alanine. After removal of the CoA the protein was hydrolyzed for 22 h at IlO’C in sealed ampules with 6 N HCl (Ultrex, Baker) after being degassed under vacuum. A crystal of phenol was added to the hydrolysate before hydrolysis to protect tyrosine. The amino acid content of the hydrolysate was determined using a Beckman model 121M amino acid analyzer (21). The amino acyl mass of the enzyme was calculated from the amino acid content of the hydrolysate. Antibody Production-Rat liver HMG-CoA reductase was purified to apparent homogeneity by the method described above. A total of 0.2 mg of purified protein was injected into each of two New Zealand rabbits; the first intradermal and subcutaneous injections (0.07 mg of protein) were given as an emulsion with complete Freund’s adjuvant and subsequent injections at P-week intervals were given in incomplete Freund’s adjuvant. Isolation of the serum immunoglobulin fraction was as previously described (4). Zmmunotitrations-Unless otherwise stated, immunotitrations were carried out using constant amounts of enzyme activity and increasing amounts of antibody in the presence of 0.5% Triton X-104). This mixture was preincubated at 37°C for 30 min and the residual reductase activity was determined after addition of cofactors and [3‘?]HMG-CoA (18). The antibody used in the immunotitration studies was prepared against enzyme purified by a previously published method (4). One enzyme unit is defined as the biosynthesis of 1 nmo1 of mevalonate/min. The equivalence point is defined as the number of enzyme units inactivated by 1.0 pl of antibody. Each 1.0 ml of crude antibody contained 4.8 mg of protein. Polyacrylamide gel electrophoresis and protein determinations were performed as previously described (4). Determination of SZO.~ and Stokes Radius of HMG-CoA Reductase-The S~,W was determined as described by Martin and Ames (22). using sucrose gradients of 5 to 20% in Buffer B or Buffer D (0.05 M KCI, 0.04 M KH2P04, 0.03 M EDTA, 5 mu dithiothreitol, pH 7.2). The Stokes radius was determined by the’ method of Ackers (23) on Sepharose 6B in the presence of Buffer E (Buffer D containing 0.1 M sucrose). RESULTS

Purification

of HMG-CoA

Reductase-Purification

of

Reductase

HMG-CoA reductase by the method described is absolutely dependent on both the time and the concentration of the protein during the 65°C heat treatment (Fig. 1). If the 20-min heat treatment was carried out at protein concentrations between 8 and 12 mg of protein/ml, the enzyme in the supernatant could be purified to apparent homogeneity on an HMG-CoA affinity column (Table I; Fig. 2). The recovery of purified enzyme from the affinity column was 53 + 3% (n = 8). This method of purification routinely gave enzyme of high specific activity (21,500 f 2,800 nmol of NADPH oxidized/ min/mg of protein, n = 5) and yields of approximately 50% from

the

microsomal

fraction.

Purified enzyme was eluted from the HMG-CoA affinity column in a buffer containing 1.0 mM CoASH (Table I) and it was possible that the enzyme activity was partially inhibited by CoASH since inclusion of 1 rnM CoASH in the assay inhibited enzyme activity by approximately 55% (Fig. 3). However, 5 to 10 (~1 of eluted pure enzyme were routinely assayed spectrophotometrically in a total volume of 0.5 ml. Consequently, the CoASH concentration in the assay was 0.01 to 0.02 mM. Such concentrations of CoASH resulted in less than 5% inhibition of the reductase (Fig. 3). In contrast, inclusion of 10 pM compactin in the assay inhibited activity by greater than 98% (Fig. 3). Mevalonic acid, the natural product of HMG-CoA reductase, or its lactone derivitive had no significant

high as 2

effect mM

on reductase

active

(Fig. 3) even though

at concentrations

compactin

as

is structurally

r

IWO-

I O

I

5 Time

15 at 65”

25 (min)

FIG. 1. Effect of heat on HMG-CoA reductase. HMG-CoA reductase was solubilized from microsomal membranes obtained from cholestyramine-fed rata, purified through the 35 to 50% ammonium sulfate step, resuspended in Buffer B, pH 6.8, containing 30% glycerol and 1.0 M KC1 at 1.17 (A), 4.07 (Cl), 8.14 (O), and 12.14 (0) mg of protein/ml and placed in a water bath at 65V. At the indicated times aliquots were removed, diluted with an equal volume of Buffer B, centrifuged at 100,000 X g for 30 rnin, and the specific activity of the soluble enzyme determined. The specific activity is given in nanomoles of NADPH oxidized/min/mg of protein. In each case the recovery of active enzyme was between 90 and 100%. Subsequent purification of the reductase to apparent homogeneity on HMG-CoA affinity columns was dependent on the protein concentration at the time of heating; when enzyme was heated at 4.07 mg of protein/ml, it was not possible to purify the enzyme on HMG-CoA affinity columns as judged by a low enzyme specific activity and multiple protein bands after sodium dodecyl sulfate-polyacrylamide gel electrophoresis.’ Heat treatment at 70°C of samples initially containing 8.14 mg of protein/ml resulted in no higher enzyme activities than obtained at 65’C but in loss of 30% of enzyme activity within 8 min and a 60% loss at 20 min.’

Regulation of HMG-CoA Reductase

3717

TABLE I Purification of HMG-CoA reductase from cholestyramine-fed rats killed at the mid-dark point Hepatic microsomes were prepared from 13 rats fed 5%cholestyramine for 4 days and killed at the midpoint of the 12-h dark cycle. The solubilized reductase activity was measured spectrophotometrically. The microsomal enzyme activity was determinedby a radioassay and gave 3.4 nmol of mevalonic acidsynthesized/min/mg of protein. This latter value was multiplied by 2 to give the theoretical value 6.8 nmol of NADPH oxidized/min/me of orotein. Purificationstep

~ o t aprol

2:zactivity

kin

nmol NADPH

mg

0x1-

dized/ min

Microsomal suspen- 2,728 sion extract Soluble 24,480520 3540% (NHn),SOr 26,539 138.5 Heat 65°C 23,62513.95 Agarose-HMG-CoA 0.483

Reductase specitic activity

Yield

nmol NADPH oxidized /min/mg protein

%

18,500

60

40 20

0

purification

6.8

100

132 10.037

00

47 191 1,683 20.778

143 127 54

1

6.9 28 247 3,056

0

I O 20 CONCENTRATION (mM)

FIG.3. I n vitm modulation of HMG-CoA reductase activity. HMG-CoA reductase obtained from rats fed the cholestyramine diet was purified through the heat step (open symbols) or through the aftinity column (closed symbols). Standard spectrophotometric enzyme assays were carried out in the presence of CoASH (0, .) 3- , hydroxy-3-methylglutaricacid mevalonic acid (V), mevalonolactone (A), or compactin m). The 10-pl samples of purified enzyme assayed in this study contained10 nmol of CoASH in the absence of any additions. The compactin concentration was 10 p ~ .

a),

1

7

I

t F

* 4 E

Q

GEL SLICE

k

Fic. 4. Polyacrylamide gel electrophoresis of purified HMGCoA reductase. The reductase was purified from animals fed the cholestyramine diet to a final specific activity of 24,750 nmol of

- -\

A

8

F

"

C

FIG.2. Analysis of HMG-CoA reductase purified from rat liver microsomes fromanimals fed the cholestyramine-supplemented diet. Fractions were analyzed on 5%acrylamide gels in the presence of sodium dodecyl sulfate: A, proteins solubilized from the microsomes; B , proteins applied to the agarose-HMG-CoA affinity column; C, purifiedHMG-CoA reductase. The bromophenol blue front is shown (F).

similar to mevalonolactone. Homogeneity of Purified Reductase-The purified protein gave a single band after sodium dodecyl sulfate-electrophoresis with an apparent molecular weight of 52,000 (Fig. 2). Purified HMG-CoA reductase was analyzed for homogeneity on 5% acrylamide gels in the absence orpresence of stacking gels. Two discrete protein bandswere observed after electro-

NADPH oxidized/min/mg of protein and 40 pI (4.8 pg of protein) were mixed with 40 pl of 5 mM dithiothreitol and applied to a 5% acrylamide gel plus stacking gel as described by Maurer (24) for gel system No. 6 except that the buffers contained 25 RIM P-mercaptoethanol and thegels were polymerized with riboflavin. After electrophoresis one gel was stained withCoomassie Blue R-250 and its companion gel cut into 0.3-cm lengths and assayed for HMG-CoA reductase activity with [3-"C]HMG-CoA as previously described (4) except that the assaywas for 2.5 min. F represents the bromophenol blue front of the gel. The stacking gel is not shown. Two protein bandsand two activitypeaks were also observedwitha second purified enzyme preparation of specific activity 22,100 nmol of NADPH oxidized/min/mg of protein. MVA, mevalonate.

phoresis on gels containing stacking gels and each protein band was associated with HMG-CoA reductase activity (Fig. 4). In the absence of a stacking gel, the protein bands were more diffuse although enzyme activity was again associated with each band. If the amount of protein applied to these latter gels was low, only the major protein band, with the higher RF, wasobserved, although analysis of the gel for reductase activity showed two bands of enzyme activity.2 Separation of purified preparations of HMG-CoA reductase into two discretepopulations has not been previously reported. It is not known whether the two bands represent P. A. Edwards, D. Lemongello, and A. M. Fogelman, unpublished results.


3718

aggregation of the 104,000-dalton form of the enzyme to a larger, but still active enzyme form, or to the presence of isoenzymes. s20.w, Stokes Radius, Partial Specific Volume, and Molecular Weight of HMG-CoAReductase-Sedimentation velocity experiments were performed at pH 6.8 using impure and purified preparations of enzyme obtained from animals fed either a normal diet or a diet supplemented with cholestyramine; the szo.rr. value was 6.14 S for each preparation (Fig. 5) and corresponded to an apparent molecular weight of 104,000 for the nondissociated enzyme (Fig. SA). Analysis of enzyme solubilized either by the method of Heller and Shrewsbury (12) or by the method of Edwards et al. (18), except that in the latter method the stepinvolving treatment at 37°C for60 , ~6.14, indicating that the min was omitted, also gave a s z ~of method of solubilization and the treatment at 37°C did not signifkantly affect the s20,m value or the apparent molecular weight of the active enzyme. In addition, the spo,wvalue was unchanged when BufferB in the sucrose gradient was replaced by a buffer of lower ionicstrength (Buffer D), or if the enzyme was solubilizedin the presence of 2 m~ phenylmethylsulfonyl fluoride. Taken together, these results demonstrate that under a variety of conditions of enzyme solubilization and for sucrose gradient centrifugation the spo,u, value remained unchanged at 6.14 S and corresponded to an apparent molecular weight of 104,000. The Stokes radius of HMG-CoA reductase was determined from chromatography on Sepharose 6 B by the method of

I/ “OO~

,

10 R 3$

,

,I

$

15

Number

Ackers (23). When yeast glucose-6-phosphate dehydrogenase was added to the sample of reductase applied to thecolumn, the enzyme activities co-eluted (Fig. 6). The Stokes radius of yeast glucose-6-phosphatedehydrogenase was calculated from the data of Yue et al. (25, 26) and Cohn and Edsall (27) and shown to be 3.39 nm. The Stokes radius of HMG-CoA reductase calculated by the methods of Ackers (23) and Fish et al. (28) was 3.39 nm. The apparent molecular weight of HMGCoA reductase calculated from the calibrated Sepharose 6B column was 104,000 (Fig. 6B). The partial specific volume(4of HMG-CoA reductase was calculated from the U for the individual amino acids as described by Cohn and E d s d (27); the partial specific volume for HMG-CoA reductase was 0.735. The molecular weight of HMG-CoA reductase was calculated from the classical equation; M = 6 q N a s / l - 17p, where M is the molecular weight; v, the viscosity of the medium (0.0106 poise); s, the sedimentation coefficient (6.14 X s); a, the Stokes radius 3.39 nm; N, Avogadro’s number; U, the partial specific volume (0.735);and p, the density of the buffer (1.02431).The molecular weight of HMG-CoA reductase was calculated to be 101,400. Amino Acid Composition-The amino acid composition of homogeneous HMG-CoAreductase is given in Table 11. Cholesterol Content of Purified HMG-CoA ReductaseAnalysis of the lipid content of purified reductase (19,000 nmol of NADPH oxidized/min/mg of protein) obtained from cholestyramine-fed animals showed the presence of 0.05 pg of cholesterol/mg of protein. No increase in reductase activity was observedafter preincubation of purified or crude fractions of the reductase obtained from cholestyramine-fed rats with dispersions of phosphatidylcholine, phosphatidylserine, or phosphatidylglycerol at concentrations below 400 pg of phospholipid/0.6 ml assay (data not shown). Inhibition of HMG-CoA Reductase by Nucleotides-The reaction catalyzed byHMG-CoA reductase resultsinthe release of CoASH, NADP, and mevalonic acid. Neither mevalonic acid nor mevalanolactone were inhibitors of reductase activity (Fig. 3). However, CoASH and NADP inhibited enzyme activity at concentrations as low as 40 p~ with KI of 175 p~ and 260 p ~ respectively , (Fig. 7). In addition to the data

Fractlon

FIG. 5. Analysis of HMG-CoA reductase on sucrose gradients. Rats were fed a normal diet or a diet supplemented with cholestyramine and killed at the peak of the diurnal rhythm. The reductase was solubilized from the microsomes (18) and either a 0 to 50% ammonium sulfate fraction collected and the pellet dissolved in a minimal volume of Buffer B (pH 6.8) or the enzyme was purified to homogeneity by the method described in Table I. Enzyme (100 pl) was applied to a 4.2 ml sucrose gradient (5 to 20%)in Buffer B (pH 6.8) and centrifuged for 12.5 h ( A )or 17 h ( B ) at 42,500 rpm (235,500 X g) at 20°C in a SW 56 rotor. Aliquots (98 pl) were removed after piercing thetubeand assayed for reductase activity or protein. Aldolase (M, = 161,000), glucose-6-phosphate dehydrogenase (M, = 102,000),bovine serum albumin (BSA) (M, = 68,000), and ovalbumin (Mr = 43,500) were used as standards. SZO., for glucose-6-phosphate dehydrogenase was 6.14 S (25,26) and the value SZO., for HMG-CoA reductase was determined by the method of Martin and Ames (22). The datafrom the 12.5-hcentrifugation has been plotted as described by Martin andAmes (22). R represents the ratio of distance travelled from meniscus by standarddistance travelled from meniscus by unknown. We have arbitrarily set the ratio for glucose-6-phosphate dehydrogenase ( I t 3 ” ) to 1.0. The sample of reductase (0 to 50% (NH&S04 fraction of enzyme solubilized from cholestyramine-fed rats) had a molecular weight of 104,000. B shows the activity peaks obtained after a 17-h centrifugation of HMG-CoA reductase. 0 , 0 to 50% (NH4)2S04fraction of enzyme from cholestyramine-fed rats; 0, purified HMG-CoA reductase from cholestyramine-fed rats; X, 0 to 50% (NH&S04 fraction obtained from rats fed a normal diet. The arrows indicate the position of the protein standards. Tube 1 represents the fraction at the top of the tube (meniscus).

c 2 0 ~

50

va 1

opoferrrlm 1

70

90

I{

;

110

cytochrome

1 130

150

c

vr

1 ’

- 40

170

; 0”

u)

0

F r o c l r o n Number

FIG. 6. Chromatography of HMG-CoA reductase on Sepharose 6B.Rats were fed the diet supplemented with 5%cholestyramine for 4 days, the hepatic microsomes prepared, and the enzyme solubilized as previously described (18). A 0 to 50% ammonium sulfate fraction was prepared, redissolved in a smallvolume of Buffer E, and after addition of yeast glucose-6-phosphate dehydrogenase a sample was applied to Sepharose 6B (91 X 1.6 cm) and 0.98-d fractions collected. The V, and V, values were determined with phage and (2“Clacetate, respectively. The elution of standard proteins is given (A). The ratio of V,/Vo, where VOis the void volume and V, the volume of elution for a specific protein, is given ( E ) .The apparent molecular weight of HMG-CoA reductase was determined to be 104,000.ZgG, y-globulin; BSA, bovine serum albumin.

Regulation of HMG-CoA Reductase shown in Fig. 7, CoASH was shown to be a noncompetitive ) NADP inhibitor with respect to NADPH (KI= 1.05 m ~ and was a noncompetitive inhibitor with respect to HMG-CoA ( K , = 1.72 mM). Preliminary results also indicated that NADH at 200 p~ was a noncompetitive inhibitor with respect to NADPH with a KI of approximately 2.05 m ~ Reductase . ~ activity was not affected by addition of200 PM NAD (data not shown). HiU plots, obtained from enzyme assays in the absence or presence of 200 PM NADP, gave a slope of 1, indicating the absence of negative cooperativity (data not shown). Specificity ofAntibody to HMG-CoAReductase-Antibody raised in rabbits against purified HMG-CoA reductase gave one line on Ouchterlony double immunodiffusion against both impure and homogeneous enzyme preparations (Fig. 8A). No precipitin lines were observed against rat serum. Immunoelectrophoresis at pH 8.6 gave one precipitin line (Fig. 823). Analysis of the antigen-antibody precipitate on sodium dodecyl sulfate-electrophoresis showed the presence of two bands staining for protein, a heavier staining band at approximately 52,000 daltons and a lighter staining band at 25,000 daltons (Fig. 9). The latter band corresponds to the light TABLEI1 Amino acid composition of HMG-CoAreductase HMG-CoA reductase was purified from rats 5% fedcholestyramine for 4days to a final specific activity of 21,209nmolof NADPH oxidized/min/mg of protein. Duplicate hydrolysates and amino acid analyses were carried out.

acid acid

Amino acid

Residues/1000 residues

Lysine Histidine Arginine Aspartic Threonine Serine Glutamic Proline Glycine Alanine Half-cystine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Tryptophan

54.3 89.8 68.0

111.4 7.4 64.7 49.7 101.1 16.4

ND"

ND, not determined.

3719

chains of the immunoglobulin and the band migrating with an apparent molecular weight of 52,000 corresponds to both the heavy chains of immunoglobulin and the subunit of HMGCoA reductase. In other preliminary experiments, isolated rat hepatocytes were incubated in the presence of [3H]leucine and the microsomes isolated. HMG-CoA reductase was solubilized, precipitated with antibody, and the precipitate analyzed on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Analysis of the gel forradioactivity indicated a single radioactive peak at approximately 52,000 (data not shown). Determination of the rates of synthesis and degradation of the reductase under different experimental conditions require specific antibody that will precipitate only the enzyme under investigation. We are currently utilizing the antibody characterized in this study to investigate the turnover of the reductase. Effect of Triton X-100 on Enzyme Activity and Imnunotitration-Interpretation of results obtained from immunotitrations of the microsomal bound enzyme could be complicated if changes in membrane conformation affected the availability of the reductase for antibody. Such effects would be minimized if all the reductase enzyme were solubilized. However, routine solubilization of all the reductase is not presently possible; maximal solubilization occurs only after a concentrated suspension of microsomes are frozen and thawed and then warmed at 37°C for 60 min in a buffer containing 50% glycerol (18). Such a method was considered impractical for solubilizing all the enzyme present in the microsomal fraction obtained from 1.1 X lo7 hepatocytes. We have therefore determined conditions in which most of the other microsomal proteins are solubilized, sincesuch conditions should minimize steric constraints on the antigen-antibody interaction. Exposure of hepatic microsomes to a buffer containing 0.1%, 0.5%, or 1.0% Triton X-100 resulted in solubilization of up to 73% of the microsomal protein and less than 1.5%of the reductase (Table 111).The activity of the microsomal enzyme assayed in the absence or presence of antibody was not affected by Triton X-100 at concentrations below 1.0% (Fig. 10; Table 111).In other experiments HMG-CoA reductase was solubilized, after exposing microsomes to a buffer containing 50% glycerol (18). This latter method solubilizes approximately 10% of the membrane proteins and a high percentage of the reductase (18). Immunotitration of the glycerol-solubilized enzyme was not affected by inclusion of 0.5% Triton X100 in the assay (data notshown). We concluded that both the reductase activity and the reaction of the antibody with the antigen were not affected by

VNADPH (mM)-

FIG.7. Inhibition of HMG-CoA reductase by CoA ( A ) or NADP (B). HMG-CoA reductase was solubilized from microsomes obtained from rats fed the cholestyramine diet and killed at the peak of the circadian rhythm. The soluble enzyme was assayed spectrophotometrically without furthertreatment. In A the CoA concentrations were zero (A), 40 pM (A), 100 pM (e), or 200 pM (0). The K , for

HMG-CoA was 3.0 p ~ The . KT(175 p ~ was ) obtained by replotting the slope of the lines uersus inhibitor concentration.In B the NADP concentration were zero (A),35 pM (A), 87.5 pM (e),or 175 pM (0). The K , for NADPH was 73 pM. The K , (260 phi) was obtained by replotting the reciprocalof the intercept for each line uersus the concentration of NADP.

3720


-

68,000

ab FIG.8 (left). Immunodiffusion and immunoelectrophoresis of HMGCoA reductase. A, Ouchterlony double immunodiffusion of the reductase. The center well contained 20 pl of anti-reductase IgG. Enzyme was purified from rats fed cholestyramine as described under “Experimental Procedures” and 2O-pl aliquots were added to each well. Well I, glycerol-solubilized enzyme, 3.1 units, well 2, 35 to 50% (NH4),S04fraction, 87 units; well 3.0 to 60% (NHdSO, fraction, 58 units; well 4 and 5, pure HMG-CoA reductase, 32 and 16 units, respectively. One unit is defined here as the oxidation of 1 nmol of NADPH/min. B , Immunoelectrophoresis of pure HMG-CoA reductase (well I) and enzyme in the 35 to 50% (NH&SO, fraction (well 2). Immunoelectrophoresis was performed at pH 8.6 by the method of Garvey et al. (29). The position of the cathode was at the right of the slide. Triton X-100 under conditions where up to 60 to 75% of the membrane proteins were solubilized. Consequently, all subsequent immunotitrations of the microsomal-bound enzyme were performed in the presence of 0.5% Triton X-100. Effect of Hepatocyte Incubation Conditions OR the Cutalytic Activity of HMG-CoAReductase-Incubation of rat hepatocytes for 3 heither under standardconditions or in the presence of compactin, lecithin dispersions, or rat serum resulted in increased specific activities of microsomal HMGCoA reductase(Table IV, Fig. 11) (30, 31). The effect of compactin was maximal a t 1.66 PM and at this concentration the microsomal reductase specific activity increased &fold after a 3-h cell incubation, compared to a 2.6-fold increase in controls (Fig. 11). Inother experiments, preincubation of hepatocytes for 2% h with 0.1 to 20 PM compactin inhibited the incorporation of [2-I4C]acetate into nonsaponifiable lipids during a subsequent 30-min incubation by over95% (data not shown). In agreement with earlier studies (8), the microsomal enzyme specific activity decreased after cells were incubated with high concentrations of mevalonolactone (Table IV). In order to determine whether these changes in enzyme specific activity resulted from changes in either the amount and/or catalytic activity of the reductase, quantitative immunotitrations were conducted with microsomes isolated from the cells. The data indicate that the equivalence point (units of enzyme activity inactivated/l.O pl of antibody) increased after cells were incubated for 3 h under standard conditions or in the presence of compactin, lecithin dispersions, or rat serum (Table IV, Fig. 12).

45,000

4~ -25,700

-F -12,400

A FIG.9 (right). Analysis of the antigen-antibody complex. An impure fraction of solubilized reductase (45 nmol of NADPH oxidized/min/mg of protein) containing200 units of enzyme activity was precipitated with a 1.5-fold excess of antibody in the presence of 1% Triton X-100 and the washed precipitate was analyzed on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The relative positions and molecular weights of protein standards and the bromphenol blue from ( F ) are shown. We have previously demonstrated that immunotitrations of glycerol-solubilized reductase resulted in a linear decay of enzyme activity during inactivation of approximately 90% of the enzyme (4). However, addition of increasing amounts of antibodyto enzymebound to the microsomal membrane resulted in a linear decay of activity until approximately 50% of the enzyme had been inactivated (Fig. 12). The nonlinear decay after approximately 50%inhibition may be due to steric hindrance of enzyme activity by antibody which does not bind to the enzyme’s active site. No evidence for two different decay curves was observed and the results are therefore not consistent withtwo or morepopulations of enzymewith different affinities for the antibody in the same microsomal membrane a t one time. An increased equivalence point is consistent with the presence of an activated form of HMG-CoA reductase. However, in no incubation could the enzyme activation accountfor the increase in enzyme specific activity; for example, when cells were incubated in the presence of compactin, microsomal reductase specific activity increased 3490% but the enzyme was activated only 157%(Table IV). Hence the major increase

Regulation of HMG-CoA Reductase TABLE I11 Effect of Triton X-100on microsomal HMG-CoA reductase and microsomal proteins Hepatic microsomes were prepared from animals fed a normal diet and killed at noon. The reductase was assayed in the presence of the indicated concentration of Triton X-100 and the equivalence point determined with antibody to the reductase as described under “Experimental Procedures.” Duplicate samples were incubated with Triton X-100 for 30 min a t 37OC in the absence of antibody, centrifuged at 100,ooO X g for 70 min, and theamount of microsomal protein and HMG-CoA reductase solubilized determined. The reductase specific activity (nanomoles of mevalonate synthesized/min/mg of protein) is given as a percentage of controls which were not treated with Triton

x-100.

Microsomal

Reduc-

Treatment of microsomes

% : ~$~~~ : : - pr;nv;s tivity

TABLE IV Effect of incubation conditions on the equivalence point for HMGCoA reductase Rat hepatocytes were incubated for 3 h with the indicated compounds, the microsomes isolated, and the reductase specific activity and equivalence point determined as described under “Experimental Procedures.” The reductase specific activity and the equivalence point are also given as a percentage of the values obtained with freshly isolated, nonincubated cells, S and E, respectively. The ratio S/E represents the theoretical increase in the number of reductase molecules which would account for the observed increase in reductase specific activity in the microsomes. The calculated increase in the number of enzyme molecules takes intoaccount the “activated” state of the enzyme. Reduc-

Equiva-

Conditions

lized

B

0.1%Triton X-100 0.5%Triton X-100 1.0%Triton X-100 2.5%Triton X-100

e e l:

%

100

0.4

105 105 102 91

1.0

1.0 1.5 ND“

3721

Reductase specific activity”

tFcs::~~~~~~~

57

73 ND ND

EauivaLnce point as

0.14 0.14 0.14 0.14

ND, not determined.

1

Zero time 3h 3 h + compactin (6.8 p M ) 3 h + serum (40%) 3 h + lecithin (0.8 mg/ml) 3 h + mevalonate (1.6 m)

~

~

(E) . ,

~~I

0.055 0.93 1.92

100 1690 3490

0.14 1 100 0.20 13.1129 0.22 157

3.99 1.13

7970 2060

0.46 0.27

307 10.7193

29

0.05

45

0.016

S/E

control

trol IS)

%

t 2 29

Equivalence point*

22.2 26.0 0.89

Nanomoles of mevalonate synthesized/min/mg of microsomal protein. The number of enzyme units inactivatedby 1.0 pl of antibody.

.. 0.6jf

novo

COMPACTIN CONCENTRATION, JIM

FIG. 11. Stimulation of reductase activity by compactin.Iso-

z

c

\

\ \

Antibody (PO FIG. 10. Inactivation of HMG-CoA reductase b y antibody.

lated rat hepatocytes (1.1 X 10’ cells) were incubated for 3 h in the The specific presence of the indicated concentration of compactin (0). activity of the reductase (nanomoles of mevalonate synthesized/min/ mgof protein) in the microsomal fraction was determined before or after the incubation (0).The structure incubation of the cells (0) of the lactone form of compactin is shown.

with mevalonolactone (Table IV).This result is in agreement with results obtained in uiuo (4) and is consistent with the presence of a partially inactivated (less active) form of the reductase after mevalonolactone treatment. Effect of SodiumFluoride on the Zmmunotitration of HMG-CoA Reductase-Nordstrom et al. have recently rein microsomal reductase specific activity is presumed to result ported that the activity of HMG-CoA reductase in hepatic from an increased number of reductase molecules. In the microsomes isolatedin the presence of sodium fluorideis only experiment with compactin we have calculated that a 22.2- 15%of that observed in controls and they have proposed that fold increase in the number of reductase molecules would this lower levelof enzyme activity may be the physiologically S/E). activeform of the enzyme (3). It is not known whether account forthe experimental results (Table IV, Column We conclude that under all conditions studied the increased reductase kinase, which is reported to inactivate HMG-CoA enzyme activityin isolated liver cells was primarily due to an reductase (3, 32, 33), is inactive during the isolation of the increase in the number of enzyme molecules (Table IV). microsomes in the presence of sodium fluoride and EDTA. However, mevalonolactone treatment appeared to decrease The nature of the physiologically active enzyme has y e t to be enzyme activity with little calculated changein the number of determined. We have previously demonstrated that chaotropic agents, enzyme molecules (Table IV). The equivalence point decreased after cells were incubated such as sodium fluoride, increase the solubilization of the

Microsomal HMG-CoA reductase activitywas determined after preincubation of the microsomes at 37°C in the presence of the indicated or presence of Triton Xamount of antibody and in the absence (0) 100 at 0.1%,).( 0.5%(A),or 1.0% (0).MVA, mevalonate.


3722

0 07

07-

0 06 005 U

004

8 z

003 E

0

002

?

0 01 0 ANTIBODY ADDED ( p i )

l 0

10

20

30

ANTIBODY, JI

FIG. 12. Lmmunotitratioa of HMG-CoA reductase from rat hepatocytes. Cells were incubated for zero time (0)or 3 h under standard conditions (0)or 3 h in the presence of 6.8 p~ compactin (A), 40% rat serum (A), or 1.6 m~ mevalonolactone (MVA) (M) and immunotitrationsof the microsomal fraction carriedout as described under "Experimental Procedures."The equivalence points were determined by extrapolation of the straight lines as indicated for two titers (- - -).

reductase from the microsomal membrane (4). We have therefore investigated whether the lower enzymeactivity observed in membrane fractions isolated in the presence of sodium fluoride was a result of solubilization, and hence loss, of the reductase, or was due to thepresence in the microsomes of a partially inactivated enzyme. We have also investigated whether this partially inactivated reductase cross-reacts with the antibody to HMG-CoA reductase. The ratio of enzyme specificactivities or equivalence points from membranes isolated in the absence or presence of 50 mM sodium fluoride were similar, 9.0 and 10.2, respectively (Fig. 13). Similar results were obtained for animals killed at the peak or nadir of the rhythm (Table V) implying that the percentage of sodium fluoride-sensitive enzyme under these experimental conditions was unchanged during the normal circadian rhythm. Solubilization and loss of enzyme during isolation of the microsomes would be expected to result in a lower specific activity of the enzyme but in no change in the equivalence point. Hence, we conclude that microsomes prepared in the absence or presence of sodium fluoride contain the same number of enzymemolecules, but that each molecule of enzyme in the sodium fluoride-treated microsomes has only 11%of the activity of controls. Our data also indicate that the partially inactivated enzyme species present after sodium fluoride treatment cross-reacts with the antibody prepared against enzyme purified fromrats fed cholestyramine. When microsomes wereisolated in the presence of sodium fluoride the enzyme specificactivity was, on average, 12.5% of controls (Table V). We hypothesized that, if the immunotitration results were valid,the enzyme purified from these sodium fluoride-treated microsomes should have a specific activity 12.5% of controls, or approximately 2650 nmol of NADPH oxidized/min/mg of protein for enzyme purified from rats fed cholestyramine and killed at the peak of the diurnal reductase rhythm. We have used the purification procedure described

FIG. 13. Effect of sodium fluoride on HMG-CoA reductase. One rat was fed a normaldiet and killed at the peak of the circadian rhythm, the liver divided into two, and microsomes prepared in the of 50 m~ sodium fluoride.The specific absence (0)or presence (0) activity of the microsomal bound enzyme andthe equivalence point determined from the immunotitration were calculated as described under "Experimental Procedures." TABLEV Effect of sodium fluoride on the specific activity and equivalence point of microsomal HMG-CoA reductase Animals were killedat the 4th h of the 12-h light cycle (L-4) or at the 6th h of darkness (D-6) and microsomes isolated in the absence (-NaF) or presence of 50 m~ sodium fluoride (+NaF). The specific activity of the microsomalreductaseandequivalencepointwere determined as described under "Experimental Procedures." Ratio

Rats killed at

Specific Equivalence activity - NaF point - NaF Specific Equivalence activity + NaF point + NaF (A) (B)

A -

B

s D-6 D-6 7.03D-6 D-6 L-4 L-4

8.9 8.8

10.2 10.2

6.59 9.1

10.0

6.96

6.67

7.75

7.44

87 86 94 91 104 104

in this paper to purify HMG-CoA reductase toapparent homogeneityfrom such animals, the pure enzyme had a specif~cactivity of 3100 nmol of NADPH oxidized/min/mg of protein, a value approximating that predicted from immunotitration of the microsomal enzyme. Activation of HMG-CoA Reductase during the Circadian Rhythm-It has been generally accepted that the circadian rhythm of hepatic HMG-CoA reductase activity results from changes in the rate of synthesis of the enzyme (1). We have reinvestigated the nature of the rhythm using immunotitrations of the microsomal-bound enzyme. The findings that absolute changes in the equivalencepoints for HMG-CoA reductase determined with either microsomal or crude solubilized enzyme correlated with similar changes in the specific activities of enzyme purified to homogeneity from animals under a variety of conditions, validates, we believe the results and conclusions we have drawn from immunotitrations of the microsomal-bound enzyme obtained either from liver cells or from animals killed at the nadir or peak of the circadian rhythm. The specific activity of microsomal HMG-CoA reductase varied 3.7-fold between the diurnal low and diurnal high point


3723

The amino acid content of purified reductase differs from that previously reported by Heller and Shrewsbury (12). In the current study we have confirmed our previous finding (4) that thesubunit molecular weight is approximately 52,000, a value significantly different from previous reports of 120,000 (12) or 65,000 (14) but similar to the value of 47,000 reported determined for soluble enzyme obtained after exposing the frozen- by Srikantaiah et al. (16). In addition we have demonstrated thawed microsomes to glycerol (18).All immunotitrations were carried out with the radioassay for HMG-CoA reductase(18).Values in that after electrophoresis of purified enzyme on native acrylamide gels, enzymeactivity and the protein bands co-migrated parentheses referto thenumber of animals. (Fig. 4). We also report a novel observation that enzyme Reductase equivalence point Reductase purified by the present method could bedissociated on native killed specific acDiet gels into two active enzyme fractions. Taken together these at tivity Microsomal Soluble enenzyme zyme data indicate that thepresent studies result in an essentially nmol MVA"/ pure preparation of rat liver HMG-CoA reductase. Enzyme min/mg propurified fromanimals fed a diet supplemented with cholestyrtein amine contained about 0.05 pg of cholesterol/mg of protein. It NDb L-40.22 2 0.08 0.20 -+ 0.09 Normal (5) is not known whether enzyme activity is regulated by enzymeNDb D-6 0.81 f 0.29 0.38 f 0.08 Normal (5) NDb 0.94 f 0.10 0.94f 0.11 bound lipids. Cholestyramine (4) D-6 A number of investigators have reported that theapparent a MVA, mevalonate. molecular weight of impure samples HMG-CoA reductase, ND, not determined. estimated from sieve chromatography on Sephadex G-200 (4, 11, 12) or Agarose 0.5 m (34), was approximately 200,000. In (Table VI). However, the equivalence point increased from agreement with Heller and Shrewsbury (12) we previously 0.20 to 0.38 in animals killed at theperiod of low or maximal found that HMG-CoA reductase behaves in an anomalous activity, respectively (Table VI). Such a change in the equiv- manner on certain batches of Agarose 0.5 m chromatography alence point is consistent with a 1.9-fold activation of the (18); the nonspecific adsorbtion of the protein to certain reductase during the diurnal rhythm with the more active batches of Agarose precludes the correct estimation of the form of the enzyme observed at the peak of the rhythm. We enzyme molecular weight on these batches. In the present propose that the3.7-fold change in enzyme activity which was study we demonstrate unequivocally thatthe molecular observed during the diurnal rhythm results from a 1.9-fold weight of crude or pure HMG-CoA reductase determined activation of the enzyme and a 2-fold increase in the number from sucrose gradient centrifugation, chromatography on of enzyme molecules.We are currently investigating whether Sepharose 6B, or determined from the classical equation M the increased number of molecules results from changes in the = 6m~Nus/(l- Cp) was 104,000. Therefore the active enzyme rate of synthesis and/or degradation of the enzyme. The data appears to be a dimer containing 2 subunits of M , = 52,000. in Table VI also demonstrate that thesame equivalence point We are currently investigating conditions in which enzyme was obtained from microsomal-bound enzyme and for solubi- aggregation may occur, to yield enzyme of molecular weight lized enzymeand provide further evidence that theinteraction approximating 200,000. Previous estimations of the molecular of the antibody with the enzyme is not affected by membrane weight of HMG-CoA reductase which gave values of approxconstitutents at least during inactivation of approximately imately 200,000 (4,11,12,34) were all determined on partially 50% of the enzyme activity. purified preparations of HMG-CoA reductase. It is possible that association of these impure samples of reductase with the DISCUSSION other proteins originally found in the microsomal membrane We report here a rapid method of purification for rat hepatic would lead to the higher apparent molecular weights of the HMG-CoA reductase which yielded enzyme of high specific enzyme that have been observed. Antibody raised against the pure protein gavebne precipitin activity (18,700 to 26,006 m o l of NADPH oxidized/min/mg of protein) and overall recoveries from the liver microsomes line against both pure and crude fractions of the reductase, of approximately 50%. We have previously reported similar after bothOuchterlony immunodiffusion and immunoelectrospecific activities but recoveries of 15 to 25% for enzyme phoresis (Fig. 8). The protein precipitated by the antibody had a subunit of approximately 52,000 daltons (Fig. 9). These purified by an alternative method (4). A number of investigators have reported purification of rat results indicate that the antibody preparation is specific for liver HMG-CoA reductase (4, 11-16) but with the exception HMG-CoA reductase. of one report by Kleinsek et al. (13)the specific activities were Validation of the use of immunotitrations to probe for 0.02 to 25% of those reported here. different activated forms of microsomal-bound reductase was The specific activity values we report here may be under- provided both when cholestyramine feeding was shown to estimations. In one particular sample of purified HMG-CoA activate HMG-CoA reductase as judged by enzyme purificareductase (specific activity 21,209 nmol of NADPH oxidized/ tion (4), immunotitration of solubilized enzyme (4), or immumin/mg of protein), the protein concentration was found to notitration of microsomal bound enzyme (Table VI) and when be 76.7pgof protein/ml. However after trichloroacetic acid addition of NaF to the microsomes resulted in decreased precipitation and amino acid analyses we have calculated the activity of the pure enzyme and corresponding changes in the aminoacyl mass to be 39.88 pg/ml. Analysisof the concentra- immunotitration of crude enzyme.Valid interpretation of tions of aspartic acid, valine, and isoleucine, three amino acids immunotitration data is dependent on two assumptions: that which are not affected by the CoA artifacts, before and after the affinity of different forms of the enzyme for antibody is trichloroacetic acid precipitation showed that 76%of the similar and that different experimental conditions do not protein was precipitated under theseconditions. Consequently result in varying amounts of a totally inactive form of the the corrected specific activity of HMG-CoA reductase would reductase which continues to bind the antibody. The first be increased 1.46-fold to 31,000 nmol of NADPH oxidized/ assumption appears to be met for enzyme from rats fed a min/mg of aminoacyl mass. normal and cholestyramine-supplemented diet (TableIV) (4),

TABLEVI Effect of diet and time of day on reductase specific activity and equivalence point Microsomes were prepared from rats fed a normal diet or a diet supplemented with cholestyramine and the reductase specific activity was also orequivalencepointdetermined.Theequivalencepoint

2

3724


and for enzymeisolated in the absence or presence of sodium fluoride (Fig. 13). With these assumptions in mind the data are consistent with at least four different forms of HMG-CoA reductase: 1) the low equivalence point enzyme found after mevalonolactone administration to hepatocytes or after isolation of microsomes with sodium fluoride. We are currently investigating whether the same mechanism results in these relatively inactive forms of the enzyme. 2) The enzyme found in microsomes prepared from animals killed at the basal period of the rhythm; 3) enzyme in the livers of animals killed at thepeak of the circadian rhythm or found in hepatocytes incubated under conditions which increased enzyme activity; and 4) enzyme found in microsomes isolated from animals fed cholestyramine. As discussed above we cannot exclude the possibility that different experimental conditions result in differing percentages of totally inactive:active reductase and that theinactive enzyme is still capable of cross-reacting with the antibody. An activated form of the reductase was present in the microsomal membrane after hepatocytes hadbeen incubated in the presence of lecithin dispersions, or serum, or compactin (Table IV, Fig. 12). However, the mechanisms by which the latter three agents result in increased reductase activity is thought to be different. Lecithin is reported to promote the efflux of cellular cholesterol (31, 35-37) and rat serum containing high density lipoproteins may act by the same mechanism (30).We have previously reported that efflux of cellular cholesterol leads to increased reductase activity in rat hepatocytes (31) and human leukocytes (35) and such studies are in agreement with those of Burns and Rothblat(36).Compactin is a potent inhibitor of HMG-CoA reductase (17) (Fig. 3) and when added to cells inhibits the conversion of acetate to cholesterol’ (38-40). The compactin-mediated inhibition of the reductase was reported to lead to theproduction of large amounts of latent enzyme which wasinactive within the cell because of the presence of compactin (38, 39). Removal of compactin, which in the present report occurred during the isolation of the microsomes, resulted in conversion of inactive latent reductase to the active enzyme. The presence of a microsomal enzyme of reduced catalytic activity following incubation of hepatocytes with mevalonolactone (Table IV) is in agreement with conclusions drawn from previous studies onthe solubilized enzymeobtained from the intact rat (4). It is not known whether mevalonolactone treatment also increases the rateof reductase degradation or whether partially degraded or dissociated enzyme will crossreact with the antibody. Direct inhibition of the reductase by 2 m~ mevalonic acid or mevalonolactone was not observed (Fig. 3) and these data are consistent with a previous proposal that theinhibition of reductase in intact rats or in isolated rat hepatocytes following administration of mevalonolactone is not a result of direct inhibition of the enzyme, but results from the conversion of mevalonolactone into some as yet unidentified inhibitor (4). Although activation of HMG-CoA reductase in isolated hepatocytes appeared to play a role in increasing enzyme activity under a variety of experimental conditions, we have calculated that themajor increase in activity resulted from an increased number of enzyme molecules (Table IV). We are currently investigating whether this increase is associated with changes in the synthesis or degradation of the enzyme, or both. We report here, for the first time, data indicating that the circadian rhythm of hepatic HMG-CoA reductase results in part from enzyme activation and that enzyme at the peak of the rhythm is more active than that at the nadir (Table VI).

These results are not in agreement with a recent publication in which Hardgrave et al. concluded that enzyme at thepeak of the rhythm was less active than that at the nadir (41). We have previously reported that thecircadian rhythm was abolished following cycloheximidetreatment (19) and it is therefore possible that enzyme activation occurs by a cycloheximide-sensitive step. Higgins et al. (14) have reported that the circadian rhythm of HMG-CoA reductase activity correlated with changes in the rate of synthesis of a protein whose subunit was 65,000 daltons. However, recent studies in this laboratory have indicated that the subunit of homogeneous preparations of HMG-CoA reductase has a molecular weight of approximately 52,000 (Fig. 2) (4). It is not known whether this discrepancy in subunit size results from different methods of enzyme solubilization and subsequent purification or to contaminants in some of the “purified” enzyme preparations. It is pertinent that the szo,u value and apparent molecular weight of the reductase were unaffected by a variety of solubilization techniques and that addition of the protease inhibitor, phenylmethylsulfonyl fluoride during solubilization of the enzyme did not affect the experimental values. We are currently investigating whether the rate of synthesis and/or degradation of HMG-CoA reductase (with a subunit of 52,000 daltons) changes during the diurnal rhythm of enzyme activity. The findings that thereductase activity is inhibited by low (40 p )concentrations of NADP or CoASH may indicate that such inhibition may be important in the physiological regulation of enzyme activity. These studies emphasize that regulation of HMG-CoA reductase is a complex process and that a detailed understanding of these regulatory phenomena is dependent on the availability of homogeneous preparations of enzyme and on specific antibody that reacts only with the enzyme. Acknowledgments-We aregrateful for the excellent technical assistance of Show-FungLan. We express thanks to Dr. M. Haberland for useful discussions andto Dr. V. Schumaker for determinationof the viscosity and density of Buffer D. REFERENCES 1. Rodwell, V. W.,McNamara, D. J., and Shapiro, D. J . (1973) Adu. Enzymol. 38,373-412 2. McNamara, D. J., and Rodwell, V. W . (1972) in Biochemical Regulatory Mechanisms in Eukaryotic Cells (Kun, E., and Grinsolii, S., Eds) pp. 206-243, J. Wiley & Sons, New York 3. Nordstrom,J. L.,Rodwell, V. W., and Mitschelen, J. J. (1977) J. Bwl. Chem. 252,8924-8934 4. Edwards, P. A., Lemongello, D., and Fogelman, A. M. (1979) Biochim. Biophys. Acta 574,123-135 5. Nepokroeff, C . M., Lakshmanan, M. R., Ness, G. C., Dugan, R. E., and Porter,J. W.(1974) Arch. Biochem. Biophys. 160,387393 6. Erickson, S. K., Matsui, S. M., Shrewsbury,M. A., Cooper, A. D., and Gould, R. G. (1978) J. B i d . Chem. 253,4159-4164 7. Goldfarb, S. (1978) J. Lipid Res. 19,489-494 8. Edwards, P. A., Popjak, G., Fogelman, A. M., and Edmond, J. (1977) J. Biol. Chem. 262, 1057-1063 9. Edwards, P. A., Lemongello, D., and Fogelman, A. M. (1979) J. Lipid Res. 20,2-7 10. Gould, R. G . (1977) in Cholesterol Metabolism and Lipolytic Enzymes (Polonovski.. J... ed) DP. 13-38, Maason Publishing, Inc.;New York 11. Kawachi. T.. and Rudnev. H. (1970) Biochemistry 9.1700-1705 12. Heller, R. A:, and Shrewsbury; M. A. (1976) J. Bioi Chem. 251, 3815-3822 13. Kleinsek, D. A., Ranganathan, S., and Porter, J. W. (1977) Proc. Natl. Acad. Sei. U. S. A. 7, 1431-1435 14. Higgins, M. J. P., Brady, D., and Rudney, H. (1974) Arch. Biochem. Bwphys. 163,271-282 15. Tormanen, C. D., Redd, W . L., Srikantaiah, M. V., and Scallen, T. J. (1976) Biochem. Biophys. Res. Commun. 68,754-762 ”

Regulation of HMG-CoA Reductase M.V., Tormanen, C. D., Redd, W. L., Hardgrave, J. E., and Scallen, T. J. (1977) J. Biol. Chem. 252,6145-6150 17. Endo, A., Kuroda, M., and Tanzawa, K. (1976) FEBS Lett. 72, 16. Srikantaiah,

323-326 18. Edwards, P. A., Lemongello, D., and Fogelman, A.M. (1979) J. Lipid Res.2 0 , 4 0 4 6 19. Edwards, P. A., and Gould, R. G. (1972) J. Biol. Chem. 247,15201524 20. Ness, G. C., and Moffler, M.H. (1978) Arch. Biochem. Biophys. 189,221-223 21. Spackman, D. H. (1967) Methods Enzymol. 11,3-15 22. Martin, R. G., and Ames, B. N. (1961) J. Biol. Chem. 236,13721379 23. Ackers, G. K. (1964) Biochemistry 3, 723-730 24. Maurer, H. R. (1971) Disc Electrophoresis and Related Tech-

niques of Polyacrylamide Gel Electrophoresis, 2nd Ed, Walter de Gruyter, Berlin 25. Yue, R. H.,Noltmann, E. A., and Kuby, S. A. (1967) Biochemistry 6,1174-1183 26. Yue, R. H., Noltmann, E.A., and Kuby, S. A. (1969) J. Biol. Chem. 244, 1353-1364 27. Cohn, E. J., and Edsall, J. T. (1943) Proteins, Amino Acids and

Peptides, Reinhold Publishing Corp., New York 28. Fish, W.W., Reynolds, J. A., and Tanford, C. (1970) J. Biol. Chem. 245,5166-5168

3725

29. Garvey, J. S., Cremer, N. E., and Sussdorf,D. H. (1977) Methods in Immunology, pp. 218-321, W.A. Benjamin, Reading, Mass 30. Edwards, P. A. (1975) Biochim. Biophys. Acta 409,39-50 31. Edwards, P. A., Fogelman, A. M., and Popjak, G. (1976) Biochem. Bwphys. Res. Commun. 68,64-69 32. Gibson, D. M., and Ingebritson, T. S. (1978) Life Sei. 23, 26492664 33. Beg, Z. H., Stonik, J. A., and Brewer, H. B. (1978) Proc. Natl. Acad. Sci. U. S. A . 75,3678-3682 34. Brown, M.S., Dana, S. E., Dietschy, J. M., and Siperstein, M. D. (1973) J. Biol. Chem. 248,4731-4738 35. Fogelman, A. M., Seager, J., Edwards, P. A., and Popjak, G. (1977) J. Biol. Chem. 252,644-651 36. Burns, C. H., and Rothblat, G. H. (1968) Biochim. Biophys. Acta 176,616-625 37. Stein, O., and Stein, Y. (1973) Biochim. Biophys. Acta 326,232244 38. Brown, M.S., Faust, J. R., Goldstein, J. L., Kaneko, I., and Endo, A. (1978) J. Biol. Chem. 253, 1121-1128 39. Bensch, W. R., Ingebritsen, T. S., and Diller, E. R. (1978) Biochem. Biophys. Res. Commun. 82,247-254 40. Kaneko, I., Hazama-Shimada, Y., and Endo, A. (1978) Eur. J. Biochem. 87,313-321 41. Hardgrave, J. E., Heller, R. A., Herrera, M. G., and Scallen, T. J. (1979) Proc. Natl. Acad.Sci. U.S. A . 76, 3834-3838