Synthase - The Journal of Biological Chemistry

Vol. 268, No. 16,Issue of June 5, pp. 12129-12135,1993 Printed in U S A .

THEJOURNAL OF BIOLOGICAL CHEMISTRY @ 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Avian 3~Hydroxy-3-methy~glutaryl-~oA Synthase CHARACTERIZATION OF A RECOMBINANTCHOLESTEROGENIC ISOZYME AND DEMONSTRATION OF THE REQUIREMENT FOR A SULFHYDRYL FUNCTIONALITYIN FORMATION OF THE ACETYLENZYME REACTION INTERMEDIATE* (Received for publication, December 8,1992, and in revised form, February 17,1993)

Ila Misra, Chakravarthy Narasimhan, and HenryM. MiziorkoS From the Department

of 3 i o c h e m ~ tMedical ~, Colkge of Wisconsin, ~ i ~ ~ a u kWisconsin ee, 53226

cDNA encoding avian liver hydroxymethylglutarylCoA synthase has been cloned into a PET vector, and theresultingexpression plasmid has been used to transform Escherichia coli BL2 l(DE3). Heterologous expression of hydroxymethylglutaryl-CoA synthase occurs upon growth of this bacterial strain in presthe ence of isopropyl-1-thio-@-D-galactopyranoside,with the target enzyme representing over 20% of total cellularprotein. Recombinant enzyme is soluble and stable in crude E. coli extracts, facilitatingits isolation in homogeneousform. With respect to specific activity, acylation stoichiometry, K-,A~-C.,A, and binding of a spin-labeled substrate analog, the recombinant enzyme is equivalent to avianenzyme, suggesting its utility for mechanistic and structuralstudies. Our earlier prediction that this avian cDNA ,encodesthe cholesterogenic cytosolic isozyme is supported by a series of experimental observations. Upon SDS-polyacrylamide gel electrophoresis, the recombinant synthase exhibits mobility in agreement with the 57.6-kDa deduced molecular mass, which exceeds the 53-kDa estimate and experimental observation for the ketogenicmitochondrial isozyme. Activity of the recombinant synthaseis stimulated by Mg“’,as predicted for the cholesterogenic cytosolic isozyme and in contrast to the inhibition observed for themitochondrial isozyme. Although antibody prepared against avian m i t ~ h o n d r i asynthase ~ effectively detects both avian m i t ~ h o n d r and i ~ recombinant synthase8 on Western blots, antibody prepared against rodent cytosolic synthase discriminates between the two proteins, sensitively detecting recombinant enzyme while reacting poorly with authentic mitochondrial enzyme. Directed mutagenesis of the recombinant synthase has been performed to produce a C129S variant, in which the sulfhydryl previously implicated in formation of the acetyl-S-enzyme reaction intermediate is replaced by a hydroxyl group. EPR measurements on the binaryCl29S-spin-labeled acylCoA complex demonstrate that themutant’s substrate binding site is unperturbed in comparison with wildtype protein. These data illustrate the utility of spinlabeled substrate analogs as tools to stringently evaluate the structural integrity of engineered proteins. C129S is catalytically inactive (10”-fold decrease in k.*)despite retaining the ability to formnoncovalent

* This work wassupported inpart by National Institutes of Health Grant DK 21491. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertkement”in accordance with 18 U.S.C. Section 1734 solelyto indicate this fact. $ T o whom correspondence should be addressed: Biochemistry Dept., Medical College of Wisconsin, 8701 Watertown Plank Rd., Milwaukee, WI 53226. Tel.: 414-257-8437; Fax: 414-257-2008,

complexes with acetyl-CoA or a spin-labeled acetylCoA analog. The demonstrated failure of C129S to forma covalent acyl-0-enzyme species accounts for these observations; data derivedfromexperiments performed with a C129G mutant confirm this conclusion. These results distinguish hydroxymethylglutaryl-CoA synthase from @-ketoacyl thiolase. Although thiolase catalyzes a mechanistically related reaction, it does not significantly discriminatebetween sulfhydryl and hydroxyl functionalities in forming a covalent acyl-enzyme intermediate (Thompson, S., Mayerl, F., Peoples, 0. P., Masamune, S., Sinskey, A. J., and Walsh, C. T. (1989)Biochemistry 28,5735-5742).

3-Hydroxy-3-methylglutaryl-CoA(HMG-COA)~synthase (EC 4.1.3.5) catalyzes the formation of a key intermediate in the cholesterogenic and ketogenic pathways in the following three-step process (Miziorko and Lane, 1977). ESH

+ acetyl-coAacetyl-SE

+ CoASH

(Eq. 1)

Acetoacetyl-CoA + acetyl-SE + ES-HMG-CoA

(Eq. 2)

+ HMG-COA

(Eq. 3)

ES-HMG-COA + HZ0 + ESH

Distinct hepatic isozymes catalyze the synthesis of choles~rogenic2 andketogenic i n t e ~ e d i a t e sclinke en beard et al., 1975). As anticipated for the enzyme that cataiyzes the first irreversible step inthese metabolic pathways, HMG-CoA synthase hasbeen implicated as a control point (Smithet al., 1988; Casals et al., 1992; Quant et al., 1989). These observations account for the recent interest in developing anti-steroidogenic agents that selectively target this enzyme (Omura et ai., 1987). Elements of the active site of this important enzyme have been identified (Miziorko and Behnke, 1985a, 1985b; Vollmer et al., 1988, Miziorko et al., 1990a) in studies that relied on protein prepared from conventional sources. Additional investigation would befacilitated by application of recombinant DNA me tho do lo^ to atlow more convenient production of the enzyme and engineered variants. We previously docuThe abbreviations used are: HMG-CoA, 3-hydroxy-3-methylglutaryl-Cok IPTG,isopropyl-1-thio-0-D-galactopyranoside; R‘, 3-carboxy-2,2,5,5-tetramethyl-1-pyrrolidinyloxyl; PCR, polymerase chain reaction; kb, kilobase pair; bp, base pair; DTT, dithiothreitol. ‘Previously we reported that, on the basis of isotopic labeling patterns (Miziorko et al., 1990), the cholesterogenic isozyme was involved in metabolite channeling. Subsequent analysis of those data, with correction for the maximal isotopic dilution attributable to endogenous metabolites, indicates that channeling appearsto remain a viable explanation for the brain metabolite labeling but need not be invoked to explain the liver data.

12129

12130

A Thioester Is Required for an HMG-CoA Synthase Intermediate

mented the isolation of full-length cDNA encodingavian liver HMG-CoA synthase (Kattar-Cooley et al., 1990) and now report the isolation and characterization of the enzyme that is expressed in ~ s c coli.~The ~utility ~ of this c recombi~ ~ nant enzyme is evaluated by c o m p ~ i s o nof its prope~ieswith the authentic avian liver enzyme. Such a comparison also provides a test of our earlier hypothesis that theisolated avian cDNA encodes the cholesterogenicisozyme. Additionally, the availability of a system for convenient expression and isolation of homogeneous recombinant enzyme facilitates mutagenesis experiments. The utility of this systemis further documented by our report of protein engineering studies aimed at testing structure/function hypotheses and addressing mechanistic aspects of the HMG-CoA synthase reaction.

EcoRI

NcoI

EXPER~MENTAL PROCEDURES

mate^ NCOI

E. coli BL2l(DE3) and PET-3d vector (Studier et al., 1990) were purchased from Novagen (Madison, WI). E. coli DH5a were obtained

from Bethesda Research Laboratory (Gaithersburg,MD). Restriction enzymes, T4 DNA ligase and vent DNA polymerase, were purchased from New England Biolahs (Beverly, MA). Sequenase and IPTG (isopropyl-1-thio-P-D-galactopyranoside) were provided by United States Biochemical Corp. R’ (3-carboxy-2,2,5,5-tetramethyl-l-pyrrolidinyloxyl; 3-carboxy-PROXYL) was obtained from Aldrich. R ’ CoA thioester was synthesized using the mixed anhydride, prepared by activation of the free acid using the method of Bernert and Sprecher (1977). Deoxyoligonucleotideswere synthesized by the ProSCHEME 1. Construction of the expression plasmid (pACS) teinjNucleic Acid Facility at the Medical College of Wisconsin. encoding avian cytosolic HMG-CoA synthase. Details are proGeneclean is a product of Bio 101 Inc. (Vista, CA). All other reagents vided under “Methods.” The ATG codon within the NcoI recognition were purchased from Sigma, Pharmacia LKB Biotechnolo~Inc., or site is in frame and functions as the translation initiation codon, Bio-Rad. Antiserum against rat cytosolic synthase was kindly pro- encoding a methionine as residue 1. In this expression construct, vided by Dr. Michael Greenspan (Merck). alanine replaces a nonconserved proline as residue 2. The open, filled, and stippled segments in the diagram represent the coding sequence, Methods untranslated region, and ampicillin resistance gene, respectively. Construction of the Expression Vector-Isolation of full-length Linkers are represented by the striped regions. cDNA encoding avian HMG-CoA synthase from a Xgtll library has been reported (Kattar-Cooley et aL, 1990). The cDNA encoding synthase was derived from EcoRI insert of the h clone NC9, subcloned into pUC13 (pKC5). After endonucleolytic digestion with SmaI and a m EcoRI (Scheme l), the 3”terminus of SmaI-EcoRI insert was filled in with the Klenow fragment of DNA polymerase. Decameric linkers e+-1 f“ 1 bearing an NcoI recognition sequence were ligated to theblunt-ended X insert. The resulting fragment was purified and restricted with NcoI endonuclease to facilitate ligation (Maniatis et at., 1982) into the ”-+“e+ NcoI cloning site of the expression vector PET-3d (Scheme 1).The SCHEME 2. Constructionof mutant alleles. The figure shows a Iigation mix was used to transform DH5a competent cells (Hanahan et aL, 1991). Clones containing the insert in the desired orientation full 1.8-kb NcoI insert and therestriction sites utilized in cloning. X (as verified by restriction mapping) were isolated (Birnboim and represents cysteine 129, the site of mutation. The flanking primers Doly, 1979) and the DNA sequence encoding the N terminus of the and themutagenic primers are indicated by the arrows. PCR overlap extension techniques were employed to generate and amplifl a muexpression target was confirmed by dideoxy chaintermination method (Sanger et ab, 1977). E. coli BL21(DE3) cells were trans- tagenic fragment corresponding to bases 6-813. The mutagenic duplex formed with the purified plasmid pACS for HMG-CoA synthase was digested with BstXI and SphI restriction enzymes; the resulting fragment was incorporated into the expression plasmid as described expression. Construction of Mutant HMG-CoA Synthase-Point mutations under “Methods.” (G385C8;T384G) were engineered in HMG-CoA synthase encoding cDNA by utilizing the overlap extension PCR technique (Ho et al., purification procedures. The ligation mixture was used to transform 1989; Scheme 2). Flanking deoxyoligonucleotide primers and muta- E. coli DH5a cells. Plasmid DNA was isolated from selected transgenic primers encoding serine or glycine substitutions for cysteine formants and analyzed by restriction mapping. Purified plasmids 129were used to generate mutagenic fragments corresponding to were used to transform E. coli BL21(DE3) prior to expression. That bases 6-813. Mutagenic fragments were digested with BstXI andSphI the mutatedBstXI-SphIfragments encoded changes at no other restriction enzymes and purified using the Geneclean kit andprotocol. residue than Cys’* was verified by sequencing both mutagenic DNA As this SphI site isnot unique, three component ligations were strands. performed which involve 560-bp BStXI-SphI (containing the mutaBacterial Growth and Purif~catio~ of HMG-CoASynthase-A single tion), 654-bp SphI-NsiI, and 5.2-kb Nsil-BstXI fragments. The 654- colony of E. coli BLZl(DE3) harboringthe recombinant plasmid was bp SphI-NsiI and 5.2-kb NsiI-BstXI fragments were isolated from grown to stationary phase in LBmedia (Miller, 1972) containing 200 wild-type recombinant plasmid by appropriate restriction and gel hg/ml ampicillin. The culture was diluted (1:lOO) with 3 liters of media of the same composition and grown at 30 “C on a gyrotory The base numbering convention is based on assignment of posi- shaker. Expression was induced by addition of 1 mM IPTG toculture tion 1to the first base of the initiation codon of the avian synthase- at anoptical density of 0.6-0.8. The cells were subsequently harvested encoding cDNA. Thus, position 385 in the coding sequence corre- by low speed centrifugation afterthe culture reached an optical sponds to base 435 in the 1.8-kb sequence reported by Kattar-Cooley density of2.2. The pellet was resuspended in lysis buffer (20 mM sodium phosphate, pH 7.0,1 mM EDTA, 1mM DTT, 0.1 mM phenylet al. (1990).

z

I

A Thioester Required Is

for a n HMG-CoA Synthase Intermediate

methylsulfonyl fluoride, 10% glycerol) and lysed in a French pressure cell a t 16,000 p.s.i. Supernatant was recovered from the crude extract by centrifugation a t 46,000 X g for 45 min. HMG-CoA synthase was purified from the supernatant by an adaptation of the procedure of Clinkenbeard et al. (1975). An ammonium sulfate fraction (30-45% saturation), prepared from the supernatant, was dissolved in 20 mM sodium phosphate buffer, pH 6.5, 0.1 mM EDTA, 0.1 mM DTT and dialyzed overnight against20 liters of buffer of the samecomposition. The dialysate was loaded onto a 2.5 X 64-cm DEAE-cellulose column equilibrated with 20 mM sodium phosphate buffer, pH 6.5, containing 0.1 mM EDTA, and 0.1 mM DTT. The column was washed with 2 column volumes of the equilibrationbuffer and eluted using1.6 liters of a 20-160 mM linear gradient of sodium phosphate buffer, pH 6.5, containing 0.1 mM EDTA and0.1 mM DTT. The fractions containing HMG-CoA synthase activity were pooled and concentrated using an Amicon ultrafiltration cell. The same procedurewasemployed to isolate HMG-CoA synthase variants to apparent homogeneity. 100150 mg of enzyme are isolated from a 3-liter culture. Western Blotting-Methodology reported by HaasandBright (1985) was employed for immunochemical detection of HMG-CoA synthase. Crude E. coli extract waselectrophoresed on an SDSpolyacrylamide gel. The proteins were transferred to a nitrocellulose membrane, followed by incubation witha 1:lOOO dilution of antiserum raised against avian mitochondrial synthase or rat cytosolic synthase and subsequent washing to eliminate nonspecific binding. The resulting complex was visualized by incubation with a solution of ''1protein A, followed by autoradiography. Characterization of the Recombinant Enzyme-For measurement of the overall condensation reaction (Miziorko et al., 1975), 200 p M acetyl-coA was added to a reaction mixture (30 "C) containing 100 mM Tris-HCI, pH 8.2, 100 p~ EDTA, 50 p M acetoacetyl-CoA, and appropriately diluted HMG-CoA synthase (approximately 6 pg wildtype enzyme in 1.0-ml final volume). The reaction rate was monitored by acetyl-CoA-dependent loss of absorbance a t 300 nm, due to condensation with the enolateof acetoacetyl-CoA. The enzyme's specific activity was calculated as micromoles/min/mg. Apparent K,,, measurement was done in the presence of20 p~ acetoacetyl-CoA. T o achieveimproved sensitivity,the overallreaction rate of mutant synthase was performed using the equivalent radioisotopic assay of Clinkenbeard et al. (1975). The reaction mixture included 100 mM Tris-HCI, pH 8.2, 100 p~ EDTA, 200 p~ ["Clacetyl-CoA (10,000 dpm/nmol), 50p~ acetoacetyl-CoA, and 2.0 mg C129S in 400 pl. The reaction was initiated by addition of radiolabeled acetyl-coA to the assay mixture containing the rest of the components a t 30 "C. At specified time intervals, 40-pl aliquots were removed from the incubation mixture and acidified with 6 N HCI. The mixture was heated to dryness, and acid-stable radioactivity due to [L4C]HMG-CoA was measured by liquid scintillation counting. Dataanalysisandan estimate of kinetic parameterswere performed using nonlinearregression analysis (Marquardt, 1963). The enzymes' acetyl-coA hydrolase activity was determined (Miziorko et al., 1975) by measuring the time-dependent depletion of ["Clacetyl-CoA, after its conversion to acid-stable citrate upon reaction with excess citrate synthase and oxaloacetate. The reaction mixture contained 100 mM potassium phosphate, pH 8.0, and 60 pg of HMG-CoA synthase in 300-pl total volume a t 30 "C. The reaction was initiated by addition of 100 K M [14C]acetyl-CoA (10,000 dpm/ nmol). At specific time intervals, 20-pl aliquots were removed and added rapidly to a mixture containing 500 milliunits of citrate synthase and400 PM oxaloacetate in100 mM potassium phosphatebuffer, pH 8.0 (100 p1 final volume). The resulting mix was acidified with 100 pl of 6 N HCI and heated to dryness. The acid-stable radioactivity which is measured is due tounhydrolyzed ["Clacetyl-CoA. The stoichiometryof acetyl-coA bindingwas determined according to the procedure of Vollmer et al. (1988). After 30 "C incubation of the enzyme (120 pg)in 100 mM sodium phosphate, pH 7.5, the mixture was placed on ice. ["ClAcetyl-CoA (10,000 dpm/nmol) was added to bring the 100-pl incubation mixture to a final concentration of 200 p ~ Unbound . acetyl-coA was removed using a G-50 centrifugal column equilibrated with 10 mM sodium acetate, pH 5.0, a t 4 "C. Protein in the recovered samples was estimated by the Bradford assay (1976), and radioactivity wasdetermined by liquid scintillation counting. Stoichiometry of covalent acetylationwas determined according to Miziorko et al. (1975). The incubationmixture,containing ["C] acetyl-coA and 50 pg of wild-type enzymeor 300 pg of mutant enzyme in 100 pl, was treated with 900 pl of ice-cold 10% trichloroacetic acid. The denatured protein was recovered by centrifugation. The pellet

12131

wasresuspended in 10% trichloroaceticacid andtransferredto a glass fiber filter. The filters were washed extensively with ice-cold 10% trichloroacetic acid and 50 mM sodium pyrophosphate in 500 mM HCI and once with cold absolute ethanol. Filterswere dried, and radioactivity was determined by liquid scintillation counting. Measurement of R'CoA Binding by EPR-Conventional X-band EPR spectra were recorded usingaVarian Century-Line 9-GHz spectrometer. The samplescontained variable concentrations of HMG-CoA synthase sites (0-930 p ~ in) 50 mM sodium phosphate buffer, pH 7.0, and 50 or 150 p~ R'CoA. R'CoA bound to HMGCoA synthase was calculated by comparing the amplitudes of high, center, or low field lines of sample spectra to thecorresponding lines observed with asolution containing an equal concentration of R' CoA in buffer. The datawere analyzed (Miziorko et al., 1979) by Scatchard plot using linear regression analysis, except in the case of C129G, where weak binding precluded as extensive an analysis. In this case, K d was calculated on the basis of measured free and bound species determined in three separate experiments. The spectra were recorded a t ambient temperature with a modulation amplitude of 1 G, modulation frequency of 100 KHz,and microwave power 5 mW. Field sweep was 100 G, and time constantwas 0.5 s. The EPR spectrumof bound R ' CoA was obtained at 5 G modulationamplitude andvariable gain. Rotational correlation time of the bound spin label was determined using a spectral simulation algorithm (Freed, 1976; Schneider and Freed, 1989). RESULTS

Expression and Isolation of Recombinant HMG-CoA Synthase"T7 polymerase-dependent protein synthesisis induced by addition of IPTG to the E. coli BLZl(DE3) culture that has been transformed with expression plasmid pACS, which contains synthase-encoding cDNA. There is a marked accumulation of protein (Fig. 1, lane 2) which exhibits a subunit molecular mass in excess of that observed for a liver mitochondrial synthase marker (Fig. 1, lune I). No comparable protein appearsupon addition of IPTG tobacteria containing only the parent PET-3d vector (Fig. 1, lanes 5-7). There is a 5-kDa increment in subunit molecularmassfor the avian cytosolic HMG-CoA synthase (Clinkenbeardet al., 1975) over that of the mature mitochondrial isozyme (Reed et al., 1975). Thus, appearanceof this prominent Coomassie-stained band prompted activity assays which demonstrated high levels of HMG-CoA synthase activity in extractsof bacteria that harbor the expression plasmid (Table I) and no activity in Samples from bacteria containing the PET-3d vector. Based on the specific activity of purified synthase (uide infra),approximately 24% of protein in crude bacterial extracts is attrib-

1 2 3 4 5 6

7

I "

FIG. 1. Expression of cytosolic HMG-CoA synthase in E. coli. The figure shows a Coomassie-stained SDS-polyacrylamidegel. Lane 1, avian liver mitochondrial HMG-CoA synthase (2 pg); lanes 2, 3, and 4 , total extract, 46,000 X g supernatant, and 100,000 X g supernatant, respectively, from E. coli harboring theexpression plasmid (12 pg of protein); lanes 5-7, fractions equivalent to those described for lanes 2-4 but derived from bacterial cells carrying the vector plasmid only (no cDNA insert; 12 pg of protein).


12132

TABLE I Purification of recombinant avian cytosolic HMG-CoA synthase from E. coli A three liter culture of E. coli harboring the expression plasmid encodine HMG-CoA svnthase was used for enzvme isolation. step Purification

zzl;

protein units mg

Crude extract High speed supernatant (NH4)zSO4 30-45% fractionation DEAE-cellulose chromatography

1

2

1,820 1,732

unitslmg

-fold

76

0.24 0.23

(1) 0.96

(100) 90

2.0

67

4.1

33

451 407

623 0.48 302 150

151

3

4

Purification Yield

1.00

5

6

7

TABLE I1 Comparison of the characteristics of recornbinant and avian cytosolic HMG-CoA synthase In the overall condensation reaction, K,,,is an apparent value, as the assay is performed in presence of 20 p~ acetoacetyl-CoA, which is slightly inhibitory. Parameters

Liver enzyme" Recombinant enzymeb

0.8-1.0 0.5-1.0 Specific activity (pmol/min/mg) 300 270 K,,,,A~.C~A (overall reaction; p M ) 0.62 0.50-0.75 Acetylation stoichiometry (mol/ mol subunit) Effect of MgCl? on activity 40% increase 45% increase Hydrolase activity (pmol/min/ 1.0 X lo-* 1.8 X IO-* mg) Km,Ac.c.,A (hydrolase reaction; p M ) 14 12 "Data for the liver enzyme are taken from Clinkenbeard et al. (1975) and Miziorko et al. (1975). * Data for recombinant synthase are from this report, employing methodology described in detail under "Experimental Procedures."

second substrate, acetoacetyl-CoA, HMG-CoA synthase catalyzes hydrolysis of acetyl-coA (Miziorko et al., 1975). The recombinant enzyme also catalyzes this partial reaction, exhibiting a rate and Michaelis constant (Table 11) that are in good agreement with that earlier report. It had been established previously that avian HMG-CoA synthase specifically binds a spin-labeled substrate analog, R ' CoA, as a competitive inhibitor withrespect to acetyl-coA (Miziorko et al., 1979). When evaluated for interaction with recombinant enzyme, R'CoA was found to bind at a stoichiometry of 0.91 protomer and with a K d =lo2 PM, in agreement with the FIG. 2. SDS-polyacrylamide gel electrophoresis of recom- values reported for the avian enzyme. Moreover, simulation binant HMG-CoA synthase. Samples correspond to wild-type en- of the spectral featuresof the immobilized spin-labeled acylzyme at various stages of purification (lanes 3-6) and purified C129S CoA (Freed, 1976; Schneider and Freed, 1989) suggested a (lane 7) HMG-CoA synthase. Lane 1,molecular mass markers (serum rotational correlation time,T~ = 35 ns, which agrees with the albumin, 66.2 kDa; ovalbumin, 45 kDa; carbonic anhydrase, 31 kDa; range of values estimated for the avianenzyme. This T~ value soybean trypsininhibitor, 21.5 kDa; lysozyme, 14.4 kDa);lane 2, for the bound spin probe predicts, according to the Stokesmitochondrial HMG-CoA synthase isolated from chicken liver; lane Einstein equation, virtually complete immobilization of the 3, total extract of E. coli containing the plasmid encoding wild-type with the HMG-CoA synthase (12 pg); lane 4, supernatant (IO pg) obtained acyl group on a115-kDa protein,inagreement after centrifugation of bacterial extract at 46,000 X @or 45 min; lane dimeric nature of the native avianenzyme. Recombinant HMG-CoA Synthase Represents the Cholester5, 30-45% (NH4)2S04fraction (10 pg); lane 6, DEAE-eluate (10 pg); lane 7, purified C129S (3 pg). ogenic Isozyme-The primary sequence of the protein, as deduced from avian cDNA sequence data, has been assigned utable to the targetof T 7 polymerase-dependent expression. to thecytosolic cholesterogenic HMG-CoA synthase (KattarImportantly, virtually all of the expressed protein is active Cooley et al., 1990). This evaluationwas based on comparison and soluble, as judgedfrom measurementson high speed between the deduced sequenceand theempirically determined supernatants (Fig. 1, lanes 3 and 4; Table I). The level of (Edman degradation)sequence of a series of peptides isolated overexpression facilitates isolation of significant amounts of from the mitochondrialketogenic enzyme. Availability of the homogeneous enzymeby application of rudimentary saltfrac- isolated recombinant enzyme facilitates further tests of this tionation and ion exchange chromatography procedures (Fig. assignment. In addition to the observation of the predicted 2; TableI). Specific activity of the isolated recombinant increment in subunit molecular mass that distinguishes misynthase is in excellent agreement with the highest values tochondrial and recombinantenzymes (Fig. 2), furtherdifferreported for the avian liver cytosolic enzyme (Clinkenbeard ences are apparent uponimmunochemical analysis. In Westet al., 1975). Unlike that enzyme, which was vulnerable to ernblotexperiments,antiserumpreparedagainst isolated proteolysis, expression in E. coli results in a n enzyme that is avian mitochondrial enzyme (Miziorko, 1985) sensitively dequite stable in crude extracts, during isolation and upon long tects this antigen (Fig. 3A) as well as the recombinant synterm storage (50% activity after one year a t 4 "C; negligible thase. However, when antiserum prepared against the rodent loss a t -80 "C). cytosolic cholesterogenic synthase (Mehrabian et al., 1986) is Properties of Recombinant HMG-CoA Synthase-Table I1 tested using an identical blot of the avian cDNA-encoded provides a comparison of the characteristics of recombinant proteins,itdetectstherecombinantsynthase with much and avian liver HMG-CoA synthases. In addition to compa- higher sensitivity than the avianmitochondrial protein (Fig. rable catalytic activityin the overall condensation reaction to 3B). In retrospect, this discrimination, which is consistent form HMG-CoA, the apparentMichaelis constant is inexcel- with the assignmentof the recombinant enzyme as the cytolent agreementwith that estimatedfor the liver enzyme. The solic isozyme, isquite reasonable. The homology between stoichiometry with which the covalentacetyl-S-enzyme inter- cytosolic synthases from differenteukaryotes(ratuersw mediate can be trapped on the recombinant enzyme (Table chicken, 84% identity) is considerably higher than the ho11) is also in good agreement with the range of values observed mology between cytosolic and mitochondrial proteins from using various avian liver preparations. In the absence of the the same species (65% identity for the rat enzymes; Ayte et


B2A B1 I A2

12133

TABLE I11 Kinetic parameters of wild-type and mutant enzymes Parameter Wild type C129S C129G Rate of condensation (pmol/