The Tracing of the Pathway of Mevalonate's Metabolism to Other Than Sterols*

Specifically "4C-labeled mevalonic acids were administered to rats in diabetic ketosis, and the distribution of 14C was determined in the hydroxybutyric acid each rat excreted. Also, the distributions of "4C were deter- mined in hydroxybutyric acid formed by slices of livers and kidneys from rats in diabetic ketosis and incubated with the specifically labeled mevalonic acids. The dis- tributions found are in accord with the conversion of mevalonate to hydroxymethylglutaryl-CoA by the shunt pathway proposed by J. Edmond and G. Popjtik ((1974) J. Biol. Chem. 249, 66-71). That is, carbon 5 of mevalonate was metabolized to form the carboxyl of acetyl-CoA and carbons 2 and 3 of mevalonate were converted in large measure to hydroxybutyric acid without acetyl-CoA as an intermediate, e. the bond between carbon 2 and 3 was not cleaved, while the bond between 1 and 2, traced with [1,2-14C]mevalonate, was cleaved. Similar distributions of 14 C were found in hydroxybutyric acid excreted by rats in diabetic ketosis administered specifically 14C-labeled isovaleric acids, isovaleric acid having in its metabolism intermediates common to those in the shunt pathway. Edmond and Popjk reported that in 9-day-old rats 1 4C from [2-'4C]mevalonate was rapidly incorporated into fatty acids. They (1,

* This research was supported by Grant AM-14507 from the National Institutes of Health and by a grant from the Diabetes Association of Greater Cleveland. 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.
: Supported by Training Grant AM-07319 from the National Institutes of Health.
with CoASH, and (b) reversal of the HMG-CoA reductasecatalyzed reaction.
Estimations of the portions of mevalonate metabolized by the nonsterol-and sterol-forming pathways have been made from yields of 14CO2 and incorporations of 1'4C into sterols from [2-14C]mevalonate or [5-' 4 C]mevalonate (6). With the assumption that the shunt pathway is the nonsterol-forming pathway active in kidney and using the yields of 4C02 from the '4Clabeled mevalonates as the measure of that pathway, the kidney has been concluded to metabolize mevalonate by the pathway to a much greater extent than do other tissues (6)(7)(8)(9).
The purpose of the present study was to determine whether the metabolism of mevalonate in kidney as well as liver is consistent with its conversion to HMG-CoA and, if so, then to determine to what extent in mevalonate's conversion to HMG-CoA its carboxyl carbon is retained, since carbon 1 of mevalonate is cleaved in the shunt pathway, but not in the alternate pathways considered by Edmond and Popjik (1) EXPERIMENTAL PROCEDURES Materials-Female rats of the Sprague-Dawley strain, weighing 200-300 g, were used. They were fed ad libitum. Diabetes was induced in them by intravenous injection of streptozotocin after they had been fasted for 24 h, and then they were maintained with insulin for 9 to 11 days (10). Those used were the ones that developed marked ketosis along with glucosuria after the insulin was discontinued. At killing, the concentrations of glucose (11) and fi-hydroxybutyrate (12) in their blood were determined. Glucose averaged 25.4 mM and hydroxybutyrate 3.4 mM.
R,S-[1-' 4 C]Mevalonolactone (6 Ci//tmol) was purchased from the Amersham Corp., and sodium R,S-[5-' 4 C]mevalonate (15 /tCi/#tmol) was from Research Products International Corp., Mount Prospect, IL. R,S-[2-' 4 C]Mevalonolactone (17 #tCi//tmol) and R-[3-14C]mevalonolactone (52 ILCi/mol) were purchased from New England Nuclear. In addition to the evidence for their purity provided by their manufacturers, the lactones were subjected to thin layer chromatography on silica gel plates using an absolute ethanol:toluene (1:4 by volume) solvent system. Each gave a single spot containing more than 99% of the '4C applied to its plate (13,14). Oxidation of a sample of the sodium [3-'4C]mevalonate by the Kuhn-Roth procedure (10) gave acetic acid with ' 4 C localized to its carboxyl carbon while oxidation of a sample of the sodium [2-'4C]mevalonate gave acetic acid devoid of ' 4 C. The carboxyl carbon of the acetic acid should be derived from carbon 3 and the methyl carbon from carbon 6 of mevalonate.
The '4C-labeled lactones were hydrolyzed with NaOH to the sodium salts of their acids (15) just before use. There is ample evidence that the R isomer, but not the S isomer, of mevalonate is metabolized · (16), and mevalonolactone has been reported to be better utilized by liver than its acid (17), although apparently in vivo mevalonolactone is quickly hydrolyzed (15,18 3-methylglutaconyl CoA buffer of pH 7.6 (20). These acids gave single spots with the mobility of isovaleric acid when chromatographed on silica gel plates using a chloroform:butanol (92.5:7.5, by volume) solvent system.
In Vivo Experiments-The labeled compounds were injected intraperitoneally into the rats in diabetic ketosis. Urine collection was begun following the injection in 0.2 ml of 0.154 M NaCl, at the specific activity provided by the manufacturer, of one of the mevalonates (5 tCi), a mixture of [1- 1 4 C]mevalonate and [2-' 4 C]mevalonate (5 /Ci of each), or one of the isovaleric acids (5 pCi). A repeat injection was made 18 h later, and the collection of urine was terminated 18 h after that. The urines from each rat were acidified and extracted continuously with ether. Each ether extract was neutralized with NaOH and evaporated to dryness. The resulting crude sodium hydroxybutyrate was also purified by partition chromatography (21,22). It was acidified with H 2 SO 4 and mixed with a small amount of Celite 535, and this mixture was transferred to the top of a column of Celite prepared by pouring a slurry of the Celite in chloroform. The column was then developed through successive additions of chloroform, chloroform plus 5% and plus 10% (by volume) of 1-butanol equilibrated with 2 N H 2 S0 4 . Mevalonic acid and/or its lactone, extracted by ether from the acidified urines, eluted in fractions of the 100% chloroform. Hydroxybutyric acid eluted in fractions of the 90% chloroform:10% butanol.
The hydroxybutyric acid, containing between 10,000 and 75,000 dpm when mevalonate was injected and between 46,000 and 800,000 dpm when isovaleric acid was injected, was then degraded (23). It was converted to crotonic acid by distillation from sulfuric acid, and the crotonic acid was purified by sublimation and catalytically reduced to butyric acid. Lactic acid, if present in the acidified urine, would have been extracted by ether and eluted in the fractions containing hydroxybutyrate (24), but as shown by a tracing with ' 4 C-labeled lactate, it would not contaminate the crotonic acid. An aliquot of the butyric acid was combusted to CO2 (25). The remainder was converted carbon by carbon to CO 2 by successive Schmidt reactions (23,26). The CO 2 s were collected as their BaCO3s, weighed, and assayed for 14 C (23). Partition chromatography of several of the butyric acids (20), done to test whether a further step in purification was necessary, did not alter the results of degradation. Percentage of recovery for each hydroxybutyric acid, as recorded in the tables to follow, is 100 times the sum of the ' 4 C-specific activities of the four carbons found in the portion of the butyric acid that was degraded, divided by four times the specific activity of the 4CO2 from the portion that was combusted (23). A recovery of about 100% then provides further support for the purity of a butyric acid and the adequacy of its degradation.
In Vitro Experiments-Slices, 1.0 g, from kidney cortex and livers of rats in diabetic ketosis, prepared using a Stadie-Riggs slicer, were incubated at 37 C in 10 ml of Krebs-bicarbonate buffer. The buffer contained [1-14 C]mevalonate and [2- 1 4 C]mevalonate in equal amounts (12.5 /Ci of each and at the mevalonate concentration determined by their specific activities) or just [2- 1 4 C]mevalonate, 1 mM palmitate, 2.5% defatted bovine serum albumin, and 0.5 mM carnitine HCI (27).
Ninety-five per cent 02-5% CO 2 was passed through the incubation vessel during the entire incubation period of 90 min in an effort to decrease '4CO2 fixation in the tissue by washing out the 1 4 CO 2 formed.
The acidified medium at the end of the incubation, after addition of 1 mmol of sodium D,L-hydroxybutyrate as carrier, was continuously extracted with ether. The hydroxybutyrates that were isolated contained between 30,000 and 120,000 dpm. Their purification and degradation were as described for the in vivo experiments.
In one experiment, a liver from a rat in diabetic ketosis was perfused (28). The perfusate was Krebs-bicarbonate buffer containing 2.5% defatted bovine serum albumin, sheep red blood cells to a hematocrit of 22%, 4 mM glucose, and 0.1 mM [1,2-' 4 C]mevalonate (25 pCi). After 90 min of perfusion, the perfusate was acidified, carrier hydroxybutyrate was added, and the hydroxybutyrate was isolated and degraded as just described for incubation of slices.

RESULTS
Distributions of 14 C in the hydroxybutyric acids excreted by the ketotic diabetic rats injected with the 4 C-labeled mevalonic and isovaleric acids are recorded in Table I  b Rat, during collection of the hydroxybutyric acid, was also fed phenylaminobutyric acid (1%) for another purpose (10,29). butyric acid, with one-fifth as much 1 4 C in carbon 1 as carbon 3. In the hydroxybutyric acid formed from [5-' 4 C]mevalonate, 14C was also localized to carbons 1 and 3, but there was 1.1 to 1.4 times as much 14  Distributions of 14 C in hydroxybutyric acids formed when [2-14 C]mevalonate and [1,2-14 C]mevalonate were incubated with liver and kidney slices and when [1,2-14C] mevalonate was perfused through a liver from rats in diabetic ketosis are recorded in Table II. The distributions in the hydroxybutyric acid formed from [1,2-14 C]mevalonate and [2-' 4 C]mevalonate by the liver slices and the perfused liver were similar to those in the hydroxybutyric acid excreted by the ketotic rats (Table  I). While there was still more than twice as much 14 C in carbon 2 as carbon 4, there was much less 14 C in carbons 1 and 3 of hydroxybutyric acid formed by kidney slices than in hydroxybutyric acid formed by the liver preparations.

DISCUSSION
HMG-CoA is an intermediate in all three of the postulated nonsterol-forming pathways (Fig. 1) considered possible explanations for the formation of CO 2 and fatty acids from carbons 2 and 5 of mevalonic acid (1) and their incorporation into hydroxybutyrate (30). Cleavage of the HMG-CoA formed by each pathway would yield acetoacetic acid and hence hydroxybutyric acid, containing carbons 2, 3, and 6 of the mevalonic acid, and acetyl-CoA, containing carbons 4 and 5. Two molecules of acetyl-CoA can condense to form 1 molecule of acetoacetyl-CoA which can also yield hydroxybutyric acid.
Therefore  (29). Greater activity in carbon 1 than carbon 3 can be attributed to the formation of HMG-CoA by the condensation of labeled acetyl-CoA with unlabeled acetoacetyl-CoA formed from the terminal four carbons of fatty acids and/or formed from ketogenic amino acids, e.g. phenylalanine and leucine (29). More than 1.1 to 1.4 as much ' 4 C in carbon 2 as carbon 4 of hydroxybutyric acid when [2-' 4 C] mevalonic acid was injected (1.0/0.47 = 2.1 from Table I Table  I), therefore, indicates that a significant portion of acetoacetic acid formed from mevalonate was formed without acetyl-CoA as an intermediate. This is in keeping with the formation of the acetoacetic acid from the mevalonate via HMG-CoA without cleavage of the bond between carbons 2 and 3. That there is activity in carbon 4 when [2-' 4 C]mevalonic acid and in carbon 1 when [3-' 4 C]mevalonic acid were administered is in accord with some of the acetoacetic acid formed from HMG-CoA being converted to acetoacetyl-CoA (31), with the acetoacetyl-CoA then being cleaved to acetyl-CoA, and the acetyl-CoA then reconverted to acetoacetic acid.
Via reversal of the HMG-CoA reductase-catalyzed reaction or the conversion of mevalonic acid directly to 3-methylglutaconyl-CoA (see Fig. 1) carbon 1 of mevalonic acid would be retained in the formation of HMG-CoA. Therefore, [1,2-14 C] mevalonic acid would yield [1,2-' 4 C]hydroxybutyric acid. Via the shunt pathway, carbon 1 is lost as C0 2 , so [2-1 4 C]hydroxybutyric acid would be formed. Therefore, the formation of [2-14 C]hydroxybutyric acid from [1,2-14 C]mevalonic acid by kidney slices (Table II) is in keeping with the reactions of the shunt pathway and eliminates the other pathways as contributing significantly to mevalonate's metabolism in kidney.
That the distributions of ' 4 C in hydroxybutyric acid excreted by the rats given [1,2-' 4 C]mevalonic acid were similar to those in hydroxybutyric acid formed by liver slices and the perfused liver (Tables I and II) indicates that the hydroxybutyric acid that was excreted was formed primarily in the liver, the principal organ of ketone body formation in the rat. Hydroxybutyric acid formed from [1,2-14 C]mevalonic acid contained 1 4 C in carbon 1 but so did the hydroxybutyric acid formed from [2-4 C]mevalonic acid, although to a lesser extent. Since the specific activities of carbons 1 and 2 of the [1,2-' 4 C] mevalonic acid were the same and since the ratio of incorporation into carbon 1 to carbon 2 of hydroxybutyrate from [2-14 C]mevalonate was 0.29/1.00 (Table I), if carbon 1 of mevalonate had been completely retained in the formation of HMG-CoA, as an approximation, a ratio of 1.29 would have been expected from [1,[2][3][4][5][6][7][8][9][10][11][12][13][14] C]mevalonic acid. That it was only 0.56 means that (1.29 -0.56)100 = 73% of carbon 1 was lost in the formation of hydroxybutyric acid from mevalonate. This is a minimum value in consideration of the possible explanations for the incorporation of 14 C into carbon 1 of hydroxybutyric acid when [2-' 4 C]mevalonic acid was administered.
Another and more likely explanation is "4CO 2 fixation. There are minor pathways by which CO 2 can be fixed to become carbon 1 of acetate and these could become prominent, when the conversion of the carbons of mevalonate to hydroxybutyric acid via HMG-CoA is relatively small. Indeed, only a relatively small amount of mevalonate is metabolized via the shunt in liver (6) and in accord with this, the amounts of '4C incorporated into hydroxybutyric acid in this study were relatively small. The small yields of '4C in hydroxybutyric acid may be a function of the compartment in which HMG-CoA is formed from the mevalonate, as well as the small contribution of the shunt pathway to mevalonate's metabolism. Formation of HMG-CoA occurs both in mitochondria and cytoplasm (32), but formation of 3-methylglutaconyl-CoA has been demonstrated only in mitochondria (33), the location of 3-methylglutaconyl-CoA hydratase has not been defined (34), and HMG-CoA lyase is present primarily in mitochondria (35).
Since 14C02 is formed from every molecule of [1-' 4 Cjmevalonate metabolized by the shunt or sterol-forming pathway, the quantity of 14CO 2 formed from [2-' 4 Clmevalonate in its metabolism must be less than from [1-14 C]mevalonate. Therefore, assuming that '4CO2 fixation is primarily responsible for the incorporation of '4C from [2-' 4 C]mevalonate into carbon 1 of hydroxybutyric acid, the quantity of '4C fixed must then be greater for the same quantity of [1,2-' 4 C]mevalonate as [2-"4C]mevalonate metabolized. This is the reason the estimate that 73% of the mevalonate metabolized via HMG-CoA proceeded with cleavage of the bond between carbon 1 and 2 of the mevalonate is considered a minimum.  (Table I). The greater activity in carbon 2 than carbon 4 is also presumably attributable to condensation of labeled acetyl-CoA with unlabeled acetoacetyl-CoA (29).
The lack of 4C in carbons 1 and 3 of the hydroxybutyric acid when the '4C-labeled isovaleric acids were administered, but not when [2-14 C]mevalonate was administered, is explained if isovaleric acid's conversion to HMG-CoA, and subsequently hydroxybutyric acid, is relatively greater than mevalonate's conversion relative to 14CO2 fixation. One would expect relatively less of a '4CO 2 fixation contribution from [4-'4C]isovaleric acid metabolism than [2-'4C]mevalonic acid, since '4CO 2 would be formed from both in the Krebs cycle, but the overall yield of 14CO2 would be relatively greater from [2-' 4 C]mevalonate. This is so since 14CO 2 from [2-' 4 C]mevalonate would also be formed in the sterol-forming pathway (6).
Accepting the operation of the shunt pathway, some of the '4C found in carbon 1 of hydroxybutyric acid formed from the various labeled mevalonic acids presumably occurred by fixation of CO 2 by isovaleryl-CoA to form 3-methylglutaconyl-CoA. However, 14C must have been converted to [1-' 4 C]acetyl-CoA by another pathway as well, since for example, [2-'4C] mevalonate gave hydroxybutyric acid with a carbon 1 to carbon 2 ratio of 0.29, but a carbon 3 to carbon 4 ratio of 0.22/ 0.47 = 0.46. That is, acetyl-CoA formed from carbon 1 and 2 of acetoacetyl-CoA derived from [2-1 4C]mevalonic acid could not have condensed to give the distribution of 14C found in carbons 3 and 4 of hydroxybutyric acid. That the incorporation of 14C into carbon 1 of hydroxybutyric acid was negligible when [2-4C]mevalonate was incubated with kidney slices, but not liver slices, is in keeping with the shunt being more active in kidney than liver (6) and, therefore, 14CO2 fixation, to any extent it occurred, not being apparent.
These studies have been made under the condition of diabetic ketosis, since our tracing of the pathway depends upon the isolation of hydroxybutyric acid. While the contributions of the shunt pathway to mevalonate's metabolism appear to vary depending on the conditions (36)(37)(38), the reactions comprising the pathway presumably are the same under differing conditions.