Rat Hepatic Microsomal Acetoacetyl-CaA Reductase A P-KETOACYL-COA REDUCTASE DISTINCT FROM THE LONG CHAIN B-KETOACYL-COA REDUCTASE COMPONENT OF THE MICROSOMAL FATTY ACID CHAIN ELONGATION SYSTEM*

The present study provides evidence for a new rat liver microsomal enzyme, a short chain &ketoacyl (acetoacetyl)-CoA reductase, which is separate from the long chain 8-ketoacyl-CoA reductase component of the microsomal fatty acid chain elongation system. This microsomal reductase converts acetoacetyl-CoA to B-hydroxybutyryl-CoA at a rate of 70 nmolJminJmg of protein; the enzyme has a specific requirement for NADH and appears to obtain electrons directly from the reduced pyridine nucleotide without the interven- tion of cytochrome ba and its flavoprotein reductase. The apparent K , of the enzyme of the acetoacetyl-CoA was 21 PM and for the cofactor, 18 PM. The pH optimum was broad, ranging from 6.5 to 8.0. The product formed is the D-isomer of 8-hydroxybutyryl-CoA. High carbohydrate fat-free diet resulted in a small but sig- nificant (35%) increase in microsomal acetoacetyl-CoA reductase activity. The cytosol also contains this en- zyme activity, measuring approximately 57% of that found in the microsomes. The mitochondrial activity which is 20-25% higher than the microsomal activity appears to be due to x,-@-hydroxyacyl-CoA in no product formation, while the presence of both NADH and NADPH led to butyryl-CoA formation. Hepatic Microsomal Condensation and Elongation of Palmitoyi- CoA-The rate of hepatic microsomal condensation and elongation of palmitoyl-CoA was measured as described previously (26). To determine the effects of short chain acyl-CoA on the fatty acid elongation system, 100 pM acetoacetyl-CoA or 100 p~ @-hydroxyhex- anoyl-CoA was included in the reaction mixture.

The present study provides evidence for a new rat liver microsomal enzyme, a short chain &ketoacyl (acetoacetyl)-CoA reductase, which is separate from the long chain 8-ketoacyl-CoA reductase component of the microsomal fatty acid chain elongation system. This microsomal reductase converts acetoacetyl-CoA to B-hydroxybutyryl-CoA at a rate of 70 nmolJminJmg of protein; the enzyme has a specific requirement for NADH and appears to obtain electrons directly from the reduced pyridine nucleotide without the intervention of cytochrome ba and its flavoprotein reductase.
The apparent K , of the enzyme of the acetoacetyl-CoA was 2 1 PM and for the cofactor, 18 PM. The pH optimum was broad, ranging from 6. 5 to 8.0. The product formed is the D-isomer of 8-hydroxybutyryl-CoA. High carbohydrate fat-free diet resulted in a small but significant (35%) increase in microsomal acetoacetyl-CoA reductase activity. The cytosol also contains this enzyme activity, measuring approximately 57% of that found in the microsomes. The mitochondrial activity which is 20-25% higher than the microsomal activity appears to be due to x,-@-hydroxyacyl-CoA dehydrogenase which converts acetoacetyl-CoA to L-8-hydroxybutyryl-CoA. The microsomal acetoacetyl-CoA reductase activity was extracted from the microsomal membrane by 0.4 M KCl, resulting in an 8-to 10-fold purification; in addition, the long chain fatty acid elongation system was unaffected by this extraction procedure. Employing @-hydroxyhexanoyl-CoA as a substrate, evidence is also provided for a separate dehydratase which acts on short chain substrates. Lastly, the liver microsomes had no detectable acetoacetyl-CoA synthetase or acetyl-coA acetyltransferase activities. Hence, the possible involvement of the rat hepatic microsomal short chain 8-ketoacyl-CoA reductase, short chain 8-hydroxyacyl-CoA dehydratase, and the previously reported short chain trans-2-enoyl-CoA reductase in the hepatic utilization of acetoacetyl-CoA and in the synthesis of butyryl-CoA for hepatic lipogenesis is discussed.
The liver is considered to be the major site of production of the ketone bodies, acetoacetate, and B-hydroxybutyrate (11, while peripheral tissues like developing brain (2)(3)(4)(5)(6), lactating mammary gland (i'), adipose tissue (8), and developing lungs (9) have been shown to readily utilize acetoacetate for energy * This study was supported by United States Public Health Service Grant AM21633. 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.
_ _ " _ _ I _ _ _ _ and lipogenesis. The utilization of acetoacetate in these tissues appears to require activation to its thiol ester, either via cytoplasmic acetoacetyl-CoA synthetase or mitochondrial 3oxoacid-CoA transferase, followed by thioIase cIeavage to acetyl-coA (3,5-7). In 1971, Stern (10) reported the existence of an acetoacetyl-CoA synthetase in liver cytoplasm and mitochondria, and although its presence in cytoplasm was confirmed (11, 12), it was generally concluded that the enzyme activity was too low to be of physiolo~ca~ importance in providing acetyl-coA for hepatic lipid synthesis. A reinvestigation (13-15) of the cytoplasmic enzyme activity under optimal conditions revealed 4-14 times higher activity than previously reported (10, l l ) , which would be sufficient to make a substantial contribution to lipogenesis. Furthermore, in an isolated perfused rat liver study, it was shown that ketone bodies could contribute up to 22% of carbon units incorporated into fatty acids (14).
During our studies on enzymes of rat hepatic microsomal elongation we discovered a new NADPH-specific short chain trans-2-enoyl-CoA reductase which could reduce crotonyl-CoA to butyryl-CoA and trans-2-hexenoyl-CoA to hexanoyl-CoA, respectively (16). Recent kinetic evidence indicated that this short chain reductase was different from the enoyl-CoA reductase which is involved in the hepatic microsomal fatty acid chain elongation of palmitoyl-CoA to stearoyl-CoA (17). In addition, we also reported the ability of hepatic microsomes to convert acetoacetyl-CoA to butyryl-CoA at a rate of 3-5 nmol/min/mg of protein (16). These results suggested the presence of a microsomal @-ketoreductase, which can reduce acetoacety1-CoA to @-hydroxybutyryl-CoA, and a dehydratase to convert the latter compound to crotonyl-CoA in addition to the previously reported short chain trans-2-enoyl-CoA reductase.
In the present study, evidence is provided that the rat hepatic microsomal short chain p-ketoacyl-CoA reductase (acetoacetyl-CoA reductase') is distinct, from the long chain P-ketoacyl-CoA reductase component of the microsomal fatty acid chain elongation system. Several biochemical properties of the acetoacetyl-CoA reductase, including cofactor specificity, product identification, apparent V,,, and K,,, values for acetoacetyl-CoA and NADH, protein concentration, and pH optimum are presented. In addition, the transfer of reducing equivalents from NADH to the short chain B-ke~acyl-CoA reductase does not require cytochrome bs in contrast to the long chain system. High carbohydrate fat-free diet marginally stimulates the short chain ketoreductase activity, while the Throughout the text the name acetoacetyl-GOA reductase is identical to short chain P-ketoacyl-CoA reductase; we do not wish to imply that the reductase is specific for acetoacetyl-GOA, as at the present time we do not know whether other 0-keto-CoA derivatives are substrates for the enzyme.

7460
long chain enzyme activity is markedly enhanced. Unlike the long chain @-ketoacyl-CoA reductase, the short chain reductase is extracted from the microsomal membrane by high salt concentration, suggesting surface topography. The presence of a separate dehydratase which can act on short chain substrates is also reported. Furthermore, acetoacetyl-CoA reductase activity was also present in the cytosolic and mitochondrial fractions, the latter ostensibly due to the 3-hydroxyacyl-CoA dehydrogenase of the @-oxidation system. Finally, the possible involvement of the rat hepatic microsomal short chain @-ketoacyl-CoA reductase, the short chain @-hydroxyacyl-CoA dehydratase, and the short chain trans-2-enoyl-CoA reductase in the hepatic utilization of acetoacetyl-CoA and in the synthesis of butyryl-CoA for hepatic lipogenesis is discussed.
Zsolutwn of Mitochondria and Microsomes-Male Sprague-Dawley rats, 175-225 g (5-6 weeks old) were given access OCE libitum to Purina Rat Chow or high carbohydrate diet ("fat-free test diet," Nut~tional Biochemicals), following starvation for 24 h. Forty-eight h after refeeding, the animals were killed by decapitation and the preparation of hepatic mitochondrial and microsomal fractions was carried out as described earlier (X), with one modification; following the initial 1 0 0 , O O O X g centrifugation the microsomal pellet was washed and resuspended in 20 mM Tris buffer, pH 7.4, and centrifuged a second time at 100,000 X g; the final pellet was resuspended in the 20 mM Tris buffer. It should be pointed out that the results in this study were obtained with a mitochondrial fraction that was isolated at 6000 X g; in addition, to greatly reduce mitochondrial contamination of the microsomal fraction, the postmitochondrial supernatant was centrifuged at 17,400 X g for 10 min and the pellet (16) discarded. The microsomes were then isolated from the remaining supernatant. For the extraction studies, 1 M KC1 was added to the microsomal suspension (10 mg of protein/ml) to give a final KC1 concentration of 0.4 M, and the supernatant separated from the pellet by centrifugation at 100,000 x g for 45 min; the pellet was again resuspended in 20 mM Tris buffer. The protein of all the subcellular fractions was determined by either the biuret procedure (18) or by the method of Lowry et al. (19).
Chemical Synthesis of DL-@-Hydroxypalmitoyl-CoA and DL-@-HYdroxyhexanoyl-CoA and Enzymatic Synthesis of @-Ketopalmitoyl-CoA-The procedure described by Stoffef and Pruss (20) was used to synthesize %hydroxy fatty acids. Briefly, the fatty acid was reacted with oxalyl chloride to obtain the acyl chloride, which is then reacted with the sodium enolate of methyl acetoacetate and sodium methoxide to yield the &keto ester. The @-ketopalmitoyl or hexanoate ester was then reduced by sodium borohydride to form the @-hydroxy acid methyl ester. The coenzyme A derivatives were synthesized as described by Fong and Schulz (21). @-Hydroxyhexanoyl-CoA was used without further purification, while 6-hydroxypalmitoyl-CoA was purified as described by Al-Arif and Bleacher (22).
The procedure used to obtain /3-ketopalmitoyl-CoA is a modification of the method of Seubert et al. (23), as previously described (24).
Determinutwn of the M$+-Enohte Complex of @-Ketopalmitoyl-CoA or Acetoacetyl-CoA-The assay mixture in the sample cuvette contained 100 mM Tris-HC1, pH 8.3, 100 pl of the enzymatically synthesized /3-ketopalmitoyl-CoA (150 p~, final concentration) or acetoacetyl-CoA (200 pM), and 100 mM M&12 in a total volume of 2.5 ml; in the reference cuvette, MgClz was omitted. Formation of Mg+-enolate was determined in an Aminco DW-2 W/VIS spectrophotometer in the split beam mode, and a peak at approximately 306 nm for @-ketopalmitoyl-CoA and 303 nm for acetoacetyl-CoA repre-senting the MP-enolate complex was observed. The concentrations of the @-ketopalmitoyl-CoA and the acetoacetyl-CoA were determined by using the extinction coefficients of 9.9 mM" em" and 21 mM" cm-I, respectively (25).

Measurement of Hepatic Microsomal Cytochrome bg Reoxidatwn
Rates-The rate of reoxidation of NADH-reduced microsomal cytochrome br, was determined by measuring the difference in a~o r b a n c e between 426 and 410 nm in the Aminco DW-2 W/VIS spectrophotometer in the dual wavelength mode as described earlier (24,26). The assay mixture containing 100 mM Tris-HC1, pH 7.4, 5 pM rotenone, and I mg/ml of liver microsomal protein in a total volume of 2.5 mi was placed in a s~t r o p~o t o m e t r i c cuvette. After equilibration at 37 "C, the reduction of cytochrome bs was initiated with 4 pM NADH (final concentration) following its addition to the cup of a plunger which was placed into the cuvette. Acetoacetyl-CoA or @ketopalmitoyl-~A at concentrations indicated in fgure legends or tables was included with NADH in the plunger or placed directly into the microsomal suspension just immediately prior to initiation of the reaction.
Acetoacetyl-CoA Reductase Assay-The enzyme activity was measured by following the rate of oxidation of NADH at 340 minus 400 nm using Aminco DW-2 UV/VIS spectrophotometer. The assay mixture in the cuvette, after all additions, had a total volume of 2.5 ml and contained in final concentration 100 mM Tris-HC1, pH 7.4, 1 mg/ml of microsomal, mi~hondrial, or cytosolic protein, and 5 pM rotenone. Following equilibration at 37 "C, 50 pM NADH (final concentration) was added directly to the cuvette, mixed rapidly, and a plunger, containing 200 p M acetoacetyl-CoA (final concentration) in a micro-cup, was immediately positioned into the same cuvette and a basal NADH oxidation rate was recorded. The plunger was then rapidly depressed and the substrate-stimulated rate of oxidation of NADH was recorded. The enzyme activity is expressed as nanomoles of NADH oxidized/min/mg of protein using extinction coefficient of 6.2 mM" cm".
Product Zdent~ficatwn-Verification of the conversion of acetoacetyl-CoA to @-hydroxybutyryl-CoA was determined by method of Williamson and Mellanby (27). Microsomes were incubated for 7 min with 500 p~ acetoacetyl-CoA and 500 p M NADH. Following incubation, the reaction was terminated, the contents saponified as described previously (26), and following centrifugation, varying aliquots were used for the assay (27). The enzyme employed in the procedure, @hydroxybutyrate dehydrogenase, catalyzes only the conversion of the D-isomer of P-hydroxybutyrate to acetoacetate.
When the product was tested with L-@-hydroxyacyl-CoA dehydrogenase, microsomes were preincubated as indicated above. The reaction was terminated with HClOd (5% final concentration), centrifuged to remove the protein, and then neutralized with 1.0 M NaOH. Varying aliquots of the s u p e~a~n t were added to a reaction mixture containing 5 units of the dehydrogenase and 500 p~ NAD+. In some experiments, the L-@-hydroxyacyl-CoA dehydrogenase was replaced with our purified acetoacetyl-CoA reductase preparation.* @-Hydroxyhexunoyl-CoA De~ydmtase Activity (Short Chain &Hy-d~x y~l -C o A Dehydratase~-The short chain dehydratase activity was assayed by measuring the formation of hexanoyl-CoA by gasliquid chromatography. The assay mixture (total volume of 1.0 ml) contained 100 mM Tris-HC1, pH 7.4, 150 p~ 8-hydroxyhexanoyl-CoA, and 200 p~ NADPH, and the reaction was initiated with 1 mg of microsomal protein. After 1-3 min of incubation at 37 "C, the reaction was terminated by the addition of 0.5 ml of 15% methanolic KOH. Fifty nmol of heptanoic acid dissolved in ethanol were added as the internal standard. After saponification at 65 "C for 40 min, 0.5 mi of 4 N HC1 was added, and fatty acids were extracted into 2 X 2.5 ml of HPLC3 grade hexane. The pooled hexane phase was evaporated to approximately 25 pl by a stream of nitrogen, and a 3-pl aliquot was then analyzed by gas-liquid chromatography as described previously (16,28).
Conversion of A~e t~e t y l -C o A to Crotonyt-CoA and Butyryl-GOA-The procedure employed to determine the formation of crotonyl-CoA and butyryl-CoA was previously described (16). Hepatic microsomal when NADH was used. Replacement of NADH by NADPH resulted conversion of acetoacetyl-CoA to crotonyl-CoA was observed only M. R. Prasad, L. Cook, R. Vieth, and D. L. Cinti, manuscript in preparation.
___. .~ in no product formation, while the presence of both NADH and NADPH led to butyryl-CoA formation.
Hepatic Microsomal Condensation and Elongation of Palmitoyi-CoA-The rate of hepatic microsomal condensation and elongation of palmitoyl-CoA was measured as described previously (26). To determine the effects of short chain acyl-CoA on the fatty acid elongation system, 100 pM acetoacetyl-CoA or 100 p~ @-hydroxyhexanoyl-CoA was included in the reaction mixture.

RESULTS
Acetoacetyl-CoA Reductase Actiuity-Recently, our laboratory reported the ability of rat hepatic microsomes to convert acetoacetyl-CoA to butyryl-CoA (16). Since this conversion occurred only in the presence of both NADH and NADPH and the reduction of crotonyl-CoA required specifically NADPH, it became evident that the initial reduction of acetoacetyl-CoA to P-hydroxybutyryl-CoA required the pyridine nucleotide, NADH. Fig. lA shows the increased rate of NADH oxidation following the addition of 200 ELM acetoacetyl-CoA to a suspension of liver microsomes; this represents a rate of approximately 70 nmol of NADH oxidized/min/mg of protein.
When the NADH concentration was decreased 5-fold to 10 ELM (Fig. lB), the rate of oxidation was markedly reduced.
When NADH was replaced by NADPH, acetoacetyl-CoA did not stimulate the endogenous NADPH oxidation rate (Fig.  IC) indicating that the rate of reduction of acetoacetyl-CoA is NADH specific. The microsomal acetoacetyl-CoA reductase activity could only be measured in the presence of the coenzyme A derivative; for example, free acetoacetic acid did not stimulate the microsomal oxidation of NADH. Furthermore The effect of acetoacetyl-CoA, acetoacetate, malonyl-CoA, and acetyl-coA on the oxidation rate of reduced pyridine nucleotides in rat liver microsomes. The assay mixture consisted of 0.1 M Tris, pH 7.4, 5 pM rotenone, either NADH or NADPH, and 2.5 mg of microsomal protein in a total volume of 2.5 ml. A, the reaction was initiated by addition of 50 p~ NADH followed by immediate placement of a plunger containing substrate into the cuvette, and the basal oxidation rate was measured. After about 40 s, 200 p~ acetoacetyl-CoA in the cup of the plunger was added to the cuvette, and the stimulated oxidation rate was recorded. B, same as A except the NADH concentration was 10 p~. C, same as A except 50 p~ NADPH replaced NADH. D, the basal oxidation rate was obtained with 50 pM NADH and at the arrow designated 1, 50 pM acetyl-coA was added, the oxidation rate recorded, followed by addition of 60 p~ malonyl-CoA at the arrow designated 2, and the rate again recorded. All tracings were taken with the Aminco DW-2 spectrophotometer in the wavelength mode at 340 nm minus 400 nm.
All measurements were made at 36 "C.
neither acetyl-coA alone nor in the presence of malonyl-CoA stimulated NADH oxidation (Fig. 1D) suggesting that the microsomes do not have the enzymatic machinery to generate acetoacetyl-CoA via the condensation reaction.
As seen in Fig. 2, acetoacetyl-CoA reduction is linear up to almost 1 mg/ml of microsomal protein. A plot of the initial velocity of the reaction uersus the substrate concentration resulted in a hyperbolic curve which gave an apparent K,,, of 21 ELM and apparent V, . of 74 nmol/min/mg of protein (Fig.  3, inset). The acetoacetyl-CoA reductase activity appeared to reach maximum at a NADH concentration of 50 ~L M (Fig. 4). The apparent K , for the NADH as determined from a Lineweaver-Burk plot was 18 ELM, while the apparent V, , was 80 nmol/min/mg. The pH optimum for the reductase activity was very broad, with little change in activity between pH 6.5 and 8.0; significant loss of activity occurred below pH 6.0 and above pH 8.5.
Identification of Reaction Products of Microsomal Metabolism of Acet~etyl-CoA-Scheme 1 illustrates the hepatic microsomal pathway for the conversion of acetoacetyl-CoA to butyryl-CoA. To determine that the NADH oxidation represented a conversion of ace~acetyl-CoA to P-hydroxybutyryl-CoA, hepatic microsomes were incubated with acetoacetyl-CoA and NADH for 10 min. Following saponification to release the ~-hydroxybutyric acid from the CoA, an aliquot of the supernatant was added to a mixture containing NAD+ and P-hydroxybutyrate dehydrogenase (26). In this reaction, P-hydroxybutyrate (not the CoA form) is converted by the dehydrogenase to acetoacetate with the formation of NADH. This is precisely what we observed with our supernatant. Furthermore, increasing the amount of supernatant to the reaction mixture resulted in a proportionate increase in NADH formation. However, it should be pointed out the stoichiometric quantities of NAD+ and P-hydroxybutyrate (1:l) cannot be obtained at this time because a portion of the generated ~-hy~oxybutyryl-CoA is quickly converted to crotonyl-CoA. Our laboratory is presently working on an HPLC assay which will be used to quantitate acetoacetate, 8-hydrox-ybutyrate, and crotonate. Preliminary studies with HPLC show an absorbance peak (210 nm) with a retention time of 31.7 min, which is identical to the retention time obtained with a pure sample of crotonic acid? Hence, although we cannot quantitate the crotonate peak at this time because of base-line noise, the peak is not present in the zero time samples; however, the peak is observed in those preparations containing only NADH and in those samples containing NADH plus NADPH. We have already reported (16) the microsomal conversion of crotonyl-CoA to butyryl-CoA by the trans-2-enoyl-CoA reductase in the presence of NADPH.
It should be emphasized that stoichiometric measurements of the product of the first reduction reaction must await not only a sensitive assay procedure but also a specific inhibitor of the dehydratase or purification of the @-keto (acetoacetyl-CoA) reductase.
The above results suggest that the microsomes convert acetoacetyl-CoA to the D-isomer of ~-hydroxybutyryl-CoA, since P-hydroxybutyrate dehydrogenase is stereospecific for D-~-hydroxybutyra~. This conclusion is supported by the experiment in which L-@-hydroxyacyl-CoA dehydrogenase added to the neutralized microsomal supernatant, obtained following incubation with NADH and acetoacetyl CoA (see "Materials and Methods"), resulted in no NADH formation. However, the addition of our partially purified reductase preparation did result in the reduction of NAD+, suggesting the presence of the D-isomer in the microsomal supernatant.
Are Acetoacetyl-CoA Reduction and P-Ketopalmitoyl-CoA R e d~t~n C~a~y~e d by the Same Enzyme?-The first indication that two separate P-ketoacyl-CoA reductases existed in rat hepatic microsomes was the observation that short chain P-ketoacyl-CoAs such as acetoacetyl-CoA ( Fig. 1) undergo reduction only in the presence of NADH, whereas total fatty acid elongation is supported by either NADPH or NADH, with NADPH being the more effective cofactor (29). Additional evidence for two separate P-ketoreduc~ses was obtained by examining the involvement of cytochrome bs in the reduction of acetoacetyl-CoA and ~-k e~p a l m i~y l -C o A .
We have recently reported (24) that cytochrome b5 participates only in the first reduction step of fatty acid chain elongation. As shown in Fig. 5, although 8-ketopalmitoyl-CoA was capable of stimulating the rate of reoxidation of cytochrome bs, acetoacetyl-CoA had no effect on the reoxidation rate. The stimulation of the rate of reoxidation of cytochrome bS by pketopalmitoyl-CoA also occurred under anaerobic conditions and in the presence of 1 mM KCN indicating that the results cannot be explained by stimulation of the desaturase pathway (24). Thus, unlike the long chain &ketoacyl-CoA reductase, acetoacetyl-CoA reductase can accept reducing equivalents directly from NADH without the intervention of the cytochrome b5 system.
To further establish two separate enzymes, we examined the effect of acetoacetyl-CoA on the NADH-dependent microsomal elongation of palmitoy~-CoA. In the presence of microsomes, palmitoyl-CoA undergoes condensation with malonyl-CoA to yield 8-ketostearoyl-CoA. This intermediate is then converted to 6-hydroxystearoyl-CoA by the microsomal @-ketoacyl-CoA reductase in the presence of NADH; The reaction was initiated with NADH placed into the cup of a plunger. The rate of cytochrome b6 reoxidation was determined by measuring the difference in absorbance between 426 and 410 nm in the Aminco DW-2 spectrophotometer in the dual wavelength mode. A, autooxidation of cytochrome bs following reduction with 4 ELM NADH (no P-ketoacyl-CoA or acetoacetyl-CoA added). B, addition of 6 p~ 6-ketopalmitoyl-CoA to the cuvette immediately before initiation of reduction of cytochrome b,. C, addition of 200 DM acetoacetyl-CoA to the cuvette immediately before the initiation of reduction of cytochrome b,. All reactions were measured at 37 "C.

TABLE I Effect of acetoacetyl-CoA and B-hydroxyhexanoyl-CoA on microsomal condensation and NADH-dependent elongation of palmitoyl-CoA
The assay mixture (1.0 ml total volume) contained 2.0 ELM rotenone, 500 p~ NADH, 25 p~ [2-"C]malonyl-CoA, 15 pM palmitoyl-CoA, 100 mM Tris, pH 7.4, and when present, either 100 p~ acetoacetyl-CoA or 100 p~ @-hydroxyhexanoyl-Coft; the reaction was initiated with microsomes (1 mg/ml, final concentration) from rats on fat-free diet, incubated at 37 "C for 1 and 3 min, and terminated with 0.5 ml of 15% KOH/methanol. Condensation and elongation were determined as described under "Materials and Methods." ultimately stearoyl-CoA would be synthesized. If the 6-ketoacyl-CoA reductase component of the chain elongation system and the acetoacetyl-CoA reductase were identical enzymes, the addition of acetoacetyl-CoA to an assay mixture containing microsomes, palmitoyl-CoA, malonyl-CoA, and NADH should result in competitive inhibition of chain elongation of palmitoyl-CoA. As shown in Table I acetoacetyl-CoA, at a concentration 7-fold greater than that of palmitoyl-CoA, had no effect on the elongation of palmitoyl-CoA or on the initial condensation reaction. Similarly, 100 PM P-hydroxyhexanoyl-CoA did not inhibit the microsomal elongation of palmitoyl-CoA (Table I). However, under these conditions, P-hydroxyhexanoyl-CoA was rapidly converted to hexanoyl-CoA by the microsomes (30 nmol/min/mg of protein) (Fig. 6). These results strongly suggest that the dehydratase which catalyzes the conversion of P-hydroxypalmitoyl-CoA to trans-2-octadecenoyl-CoA differs from the dehydratase that converts 0hydroxyhexanoyl-CoA to trans-2-hexenoyl-CoA. Hence, rat hepatic microsomes appear to contain a separate short chain system consisting of three components: 1) p-ketoacyl-CoA   Values are means f S.D. obtained from 3 animals per group. The value in parentheses represents the per cent of stimulation above the control value.
e Statistically significant differences were tested by Student's t test. p < 0.05 for controls uersus fat-free diet.
reductase; 2) @-hydroxyacyl-CoA dehydratase; and 3) trans-2-enoyl-CoA reductase. Effect of High Carbohydrate FFD-Rats were starved for 24 h followed by refeeding for 48 h, either normal rat chow or the fat-free diet. As seen in Table 11, the acetoacetyl-CoA reductase activity in liver microsomes obtained from rats on the FFD was increased from 71.3 to 97.1 nmol/min/mg of microsomal protein or a 35% stimulation, whereas the long chain P-ketoacyl-CoA reductase activity was markedly stimulated (240%) by the FFD, when P-ketopalmitoyl-CoA was employed as a substrate. Apparently the two reductases respond differently to the high carbohydrate diet; whether longer term feeding of FFD markedly stimulates the acetoacetyl-coA reductase is unknown at the present time.

Acetoacetyl-CoA-stimulated NADH Oxidation in Rat Liver
Cytosolic and Mitochondrial Fractions-We have recently reported (16) that cytosol contamination of the microsomes was less than 1% based on two different cytosol marker enzymes, lactate dehydrogenase and malate dehydrogenase. Under these same isolation conditions, we compared the acetoacetyl-CoA reduction by NADH in microsomes uersus cytosol. To our surprise rat liver cytosol contained significant amounts of reductase activity, 43.3 k 8.8 nmol of NADH oxidized per min per mg of cytosolic protein obtained from three different cytosolic fractions. Although the cytosolic activity was high, the microsomal fraction obtained from the same three animals averaged 76 nmol/min/mg of microsomal protein or 75% higher activity in the microsomal fraction than in the cytosolic fraction. Interestingly, the acetoacetyl-CoA reductase activity in the cytosol was not enhanced by the %day FFD. In fact, the activity was slightly decreased to 33.4 f 4.3 nmol/min/ mg of cytosolic protein, a reduction of 23%. Certainly, cytosolic contamination could not account for the acetoacetyl-CoA reductase activity observed in the microsomal fraction; rather it appears that the enzymatic activity exists in both fractions.
The mitochondrial fraction was also active in the reduction of acetoacetyl-CoA. Under the same conditions as for the microsomal fraction, the addition of 200 p~ acetoacetyl to the mitochondrial fraction (from untreated rats) resulted in a NADH oxidation rate of 88 f 11 nmol/min/mg of mitochondrial protein for two different mitochondrial preparations measured in triplicate, a rate that is 24% higher than that observed in the microsomes. Since the mitochondria possessed slightly higher activity than the microsomes, the extent of contamination of the latter fraction by the mitochondria was determined by measuring succinate cytochrome c reductase and glutamate dehydrogenase activities (30), mitochondrial markers, in both fractions. While these activities were 295 and 116 nmol/min/mg of protein, respectively, in the mitochondria, the microsomal activities were 5.0 and 4.6 nmol/ min/mg of protein, indicating a 1.7 to 4.0% contamination. Hence, a 4% contamination would contribute a total of only 3.5 nmol/min/mg of protein to the microsomal activity. The mitochondrial activity is most likely attributed to the 8oxidation enzyme, L -8hydroxyacyl-CoA dehydrogenase, which in the presence of NADH can convert acetoacetyl-CoA to 8-hydroxybutyryl-CoA. However, this enzyme generates only the L-isomer. As noted earlier, the product of the microsomal reaction is the D-isomer.
Extraction of Acetoacetyl-CoA Reductase Activity from Rat Liver Microsomes--In our attempt to isolate the short chain P-ketoacyl-CoA reductase from the microsomes, we observed that high concentrations of KC1 were capable of stripping a significant portion of the enzyme activity from the microsomes. As shown in Table 111, homogenization of microsomes in 0.4 M KC1 resulted in significant removal (82 and 84%) of acetoacetyl-CoA reductase activity from liver microsomes from both untreated and FFD animals, respectively. This initial isolation step has resulted in a 8-to 10-fold purification. Interestingly, increasing the KC1 concentration to 0.6 M did not remove the additional 10 to 12% residual reductase activity. However, reducing the KC1 concentration below 0.4 M resulted in a proportionate decrease in reductase activity extracted from the microsomes. It was also observed that under conditions where we extracted 84% of the acetoacetyl-CoA reductase activity, there was absolutely no effect on the microsomal fatty acid chain elongation of palmitoyl-CoA (Table 111). This activity was measured in the FFD animals because this diet induces chain elongation activity. These results also support the existence of at least two @-ketoacyl-CoA reductases.
Attempts to Detect Acetoacetyl-CoA Synthetase and Acetyl-COA Acetyltransferase (Thiolase) Activities in Hepatic Microsomes-Cytosolic acetoacetyl-CoA synthetase is one of the key enzymes in the utilization of ketone bodies in lipogenesis (3,(5)(6)(7). Having discovered the ability of microsomes to convert acetoacetyl-CoA to butyryl-CoA, the site of acetoacetyl-CoA synthesis becomes an obvious question. Of the several pathways that can synthesize acetoacetyl-CoA (15), we examined liver microsomes for the presence of two activities, i.e. acetoacetyl-CoA synthetase and acetyl-coA acetyltransferase (thiolase). The detection or measurement of acetoacetyl-CoA was followed by the absorption of its Me-enolate complex which occurs at 303 nm. The assay mixture contained 100 p~ acetoacetic acid, 100 pM CoASH, 500 pM ATP, 0.1 M Tris-HC1, pH 8.2,50 mM KCI, 1 mg/ml of microsomal protein, and 50 mM MgC12. The spectrum from 250 to 350 nm was continuously recorded from 0 to 10 min. Under these conditions, there was no peak formation at 303 nm. However, when microsomes were replaced by the cytosolic fraction, a significant peak was formed at 303 nm, which decreased with the time of incubation. This disappearance of the 303-nm peak was most likely due to the thiolase or hydrolase activity reported to be present in the cytosol (15).
Similarly, the presence of acetyl-coA acetyltransferase activity in microsomes was examined by incubating 20 or 50 p~ acetyl-coA, 0.1 M Tris-HC1 buffer, pH 8.2,50 mM MgC12, and 1 mg/ml of microsomal protein in a total volume of 2.5 ml. Again, under these conditions no formation of Me-enolate complex occurred. As a control, 1 mg of microsomal or cytosolic protein was added to a Me-enolate complex of acetoacetyl-coA. In 12 min, only 40% of the complex decreased in the presence of microsomal protein, whereas within 3 min the M$+-enolate complex disappeared in the presence of cytosolic fraction. These results indicate that the inability to observe a M$+-enolate complex with the microsomal fraction is not due to the presence of a very active hydrolase or thiolase activity in microsomes, but rather due to an absence in microsomes of acetoacetyl-CoA synthetase and acetyl-coA acetyltransferase activities. It should be noted that the method employed was capable of detecting as little as 1.0 p~ acetoacetyl-CoA. DISCUSSION The present study provides evidence for the existence of a new rat hepatic microsomal enzyme which catalyzes the reduction of acetoacetyl-CoA to @-hydroxybutyryl-CoA in the presence of the cofactor NADH. This enzyme has been designated short chain P-ketoacyl-CoA reductase or specifically acetoacetyl-CoA reductase. It differs from the well known long chain 8-ketoacyl-CoA reductase component of the fatty acid chain elongation system in several ways: 1) the acetoacetyl-CoA reductase has a specific requirement for NADH whereas the long chain reductase utilizes either NADPH or NADH; 2) the acetoacetyl-CoA reductase obtains its electrons directly from NADH without the intervention of cytochrome b~ and its flavoprotein reductase, while the long chain ketoreductase receives its electrons from cytochrome b6 (24); 3) high salt (KCl) concentrations can extract the acetoacetyl-CoA reductase from the microsomal membrane whereas the long chain enzyme is unaffected by such concentrations; and 4) acetoacetyl-CoA, the substrate for the short chain enzyme, had no effect on the microsomal chain elongation of palmi- The microsomal enzymatic activity in the presence of acetoacetyl-CoA is quite high, approximately 70 nmol/min/mg of protein; the apparent K,,, for the substrate was 21 p~, and for the cofactor, 18 PM. Product identification using D-phydroxybutyrate dehydrogenase, L-8-hydroxyacyl-CoA dehydrogenase, and partially purified reductase indicated that acetoacetyl-CoA was reduced to the D-isomer of @-hydroxybutyryl-CoA; indeed, the CoA derivative rather than the free toyl-CoA.  acid was the product since without saponification the D-phydroxybutyrate dehydrogenase was unable to oxidize the product to acetoacetate. Stoichiometric quantitation was, however, impossible since a portion of the 8-hydroxybutyryl-CoA was converted to crotonyl-CoA by the microsomal dehydratase. We can, however, state at this time based on our previous findings that the ketoreduction of acetoacetyl-CoA is not the rate-limiting step since the rate of butyryl-CoA formation from acetoacetyl-CoA in the presence of both NADH and NADPH was only 4-5 nmol/min/mg of microsomal protein. It should be noted that because of the unavailability of higher homologs of acetoacetyl-CoA, no substrate specificity studies were conducted. The NADH-dependent reduction of acetoacetyl-CoA by cytosol was unexpected; the specific activity was 57% of that found in the microsomal fraction. When the activity is expressed in terms of grams of liver, wet weight, the total activities are very similar. For example, based on microsomal protein content/g of liver, which is approximately 55 mg/g (31, 32), the total microsomal acetoacetyl-CoA reductase activity is 55 X 70 or 3850 nmol/min/g of liver, whereas based on cytosolic protein of 75 mg/g of liver measured in our laboratory and from the data of Stern (lo), a value of 75 X 43 or 3225 nmol/min/g of liver was obtained. These results suggest that the enzymatic activity exists in both subcellular fractions, microsomes and cytosol. Whether these fractions contain the identical enzyme is under current investigation.

FFD rat
The activity measured in the mitochondria can most likely be attributed to the L-@-hydroxyacyl-CoA dehydrogenase; although its activity is 20-25% higher than that found in microsomes, it cannot account for the microsomal activity for two reasons: 1) the microsomes had only a 2-4% mitochondrial contamination which would yield a total activity of 3.5 nmol/min/mg of protein and 2) the product generated by the reductase was D-@-hydroxybutyryl-CoA, whereas the dehydrogenase catalyzes the formation of the L-isomer.
The mammalian liver contains at least three enzymatic activities that are responsible for the utilization of acetoacetyl-CoA (15): acetoacetyl-CoA hydrolase (deacylase), 3-ketothiolase, of which four forms or isozymes may exist, and HMG-CoA synthase. One of the enzymes, the hydrolase gen-erates acetoacetate, another enzyme, the thiolase, forms 2 acetyl-coA molecules and the synthase generates HMG-CoA. A fourth enzymatic activity must now be added to this group: the NADH-specific short chain P-ketoacyl-CoA (acetoacetyl) CoA reductase. Three of the enzymes are found in the mitochondrial and cytosolic fractions; the thiolase and synthetase are predominant in mitochondria and the hydrolase is predominant in cytosol. The newly discovered reductase activities are present in the microsomal and cytosolic fractions.
The hepatic synthesis of acetoacetyl-CoA is most probably derived from two sources, the mitochondrial &oxidation of long chain fatty acids and the cytosolic conversion of acetoacetate to acetoacetyl-CoA in the presence of ATP (10, 11,  13, 14).
The question which must be addressed and to which an answer must be sought is: what is the physiological role of the microsomal short chain p-ketoacyl-CoA reductase? The enzyme appears to be loosely bound to the membrane surface, based upon the ease with which the enzyme can be stripped from the microsomal membrane without the intervention of detergents. This is a reasonable location for the enzyme since the substrate is hydrophilic. Since the cytosol already contains three acetoacetyl-CoA-utilizing enzymes (hydrolase, thiolase, and HMG-CoA synthase) all of which have significant activities (0.41, 6.78, and 0.11 pmol/min/g of liver, wet weight, respectively) what is the need for an additional acetoacetyl-CoA-metabolizing enzyme which appears to be found in both microsomes and cytosol? Even if the rat liver cytosolic acetoacetyl-CoA synthetase (13) were to function optimally, the rate of generation of acetoacetyl-CoA would only approximate 0.3-0.5 pmol/min/g, wet weight. This represents just 10% of the rate of the microsomal NADH acetoacetyl-CoA reductase activity. Aragon and Lowenstein (15) have calculated that livers of fed rats contain a concentration of acetoacetyl-CoA that is in the low nanomolar range; however, Menahan et al. (33) have estimated the concentration to be 1-10 p~. These concentrations are well below the apparent K,,, of the substrate for any of the aforementioned enzymes, for example, an apparent K,,, of 33 pM for the thiolase (341, 25 pM for the hydrolase (15), and 21 p~ for the microsomal @-ketoreductase. Furthermore, there appear to be at least two forms of cytosolic HMG-CoA synthase and both forms have apparent K , values less than 3 p~ (35). With such a low K,, one would expect rapid conversion of any generated acetoacetyl-CoA to HMG-CoA.
All of the values, of course, have been obtained from in vitro conditions; the in vivo situation may be quite different, however. Now regulatory factors, such as hormones, binding proteins, inhibitory products, or intermediates may markedly influence the enzyme activities. Indeed sufficient acetoacetyl-CoA must be formed since Endemann et al. (14) have shown that ketone bodies are incorporated into sterols as well as fatty acids to a significant degree.
Notwithstanding the aforementioned results, one of the functions of the microsomal @-ketoreductase in concert with the dehydratase and enoyl-CoA reductase may be to provide butyryl-CoA for lipogenesis. Lin and Kumar (36) have shown that both rat mammary gland and liver do prefer butyryl-CoA over acetyl-coA as the primer in fatty acid synthesis. In 1969, Nandekar and Kumar (37) reported the presence in the cytosol of lactating rabbit mammary gland, a NADH-dependent @-ketoacyl-CoA reductase, in addition to a cytosolic thiolase and enoyl-CoA hydratase. They found that these three enzymes plus fatty acid synthetase, which provided the enoyl-CoA reductase activity associated with it, synthesized butyryl-CoA from acetyl-coA. Lin and Kumar (36) demonstrated that the rate of incorporation of butyryl-CoA, as primer, during fatty acid synthesis in rat liver is 1.26 nmol/min/mg of protein. We have previously (16) reported that microsomes generated 3.0 nmol of butyryl-CoA/min/mg of microsomal protein and more recently, 4-5 nmol/min/mg of protein. If liver can synthesize 8-14 pmol of palmitate/h/g of tissue (38), then 8-14 pmol of primer is necessary; our microsomal system can provide approximately 17 pmol of butyryl-CoA precursor/ h/g of liver. Hence, our microsomal system may play a role in providing the liver with butyryl-CoA for lipogenesis.
At this time it is difficult to assess the link between hepatic ketone body formation, cytosolic synthesis of acetoacetyl-CoA and @-hydroxybutyryl-CoA, and microsomal reduction and dehydration of these coenzyme A forms. During abnormal intermediary metabolism as a result of a pathophysiological condition, such as diabetes, in which there can arise excess production of ketone bodies, can this newly discovered microsomal system participate in the utilization of the ketone bodies? Is this enzyme system inducible in such pathophysiological conditions as diabetes? Additional work is necessary before providing answers to these questions. If a connection exists between ketone body utilization and the microsomal reduction system, the cytosol must also be involved since it is this site where the ketone bodies are activated. We observed no acetoacetyl-CoA synthetase or acetyl-coA acetyltransferase activities in the liver microsomal fraction.
The reason for the @-ketoacyl-CoA reductase activity in the cytosol is unknown presently. We observed that contrary to the microsomal enzyme, the cytosolic activity was not inducible by the fat-free diet; in fact, there was a small reduction in reductase activity in the cytosol. Since the cytosol was capable of generating only 0.6 nmol of butyryl-CoA/min/mg of protein (161, a role for the enzyme in providing this primer is questionable.
Finally, our results with P-hydroxyhexanoyl-CoA, in which this substrate was dehydrated and converted to hexanoyl-CoA by the microsomes in the presence of NADPH and in which it did not inhibit chain elongation of palmitoyl-CoA strongly suggest that at least two dehydratases exist in the endoplasmic reticulum, the short chain enzyme, and the well known component of the chain elongation system.