Synthesis of Cerebronic Acid from Lignoceric Acid by Rat Brain Preparation

SUMMARY The conversion of [I-Wllignoceric acid (tetracosanoic acid) into cerebronic acid (Z-hydroxytetracosanoic acid) by a cell-free preparation of rat brain was investigated. The a-hydroxylating enzyme required molecular oxygen, Mg++, pyridine nucleotides, and a heat stable water soluble cofactor. The activity had pH optimum of ‘7.8, and the K, value for lignoceric acid was found to be 4.2 pM. The presence of CO did not inhibit the activity. Except for Mgf+, all heavy divalent metal ions were strongly inhibitory. EDTA was also strongly inhibitory. The pyridine nucleotides DPNH and TPNH were equally effective in producing full enzyme activity; their oxidized forms were less effective. The product cerebronic acid was detected only as a component of ceramide or cerebroside and not as free acid or CoA ester. There was no detectable activity in rat brain during the first 6 days after birth, but a rapid increase was noted between the ages of 8 and 21 days. Activity in the brain attained a rats 21 to 28 then gradually

The activity had pH optimum of '7.8, and the K, value for lignoceric acid was found to be 4.2 pM. The presence of CO did not inhibit the activity.
Except for Mgf+, all heavy divalent metal ions were strongly inhibitory.
EDTA was also strongly inhibitory.
The pyridine nucleotides DPNH and TPNH were equally effective in producing full enzyme activity; their oxidized forms were less effective.
The product cerebronic acid was detected only as a component of ceramide or cerebroside and not as free acid or CoA ester. There was no detectable activity in rat brain during the first 6 days after birth, but a rapid increase was noted between the ages of 8 and 21 days. Activity in the brain attained a maximum in rats 21 to 28 days old and then gradually decreased until an age of 85 days was reached.
The cerebellum was found to have the greatest activity.
No ol-hydroxylating activity was detected in any of the extraneural tissues studied.
2-Hydroxy fatty acids, found in small amounts in certain yeasts, bacteria, and mammalian extraneural tissues (l), are uniquely abundant in the mammalian nervous system. The 2hpdroxy acids found in the nervous system consist mainly of both saturated or monounsaturated compounds with a chain length of 22 to 26 carbon atoms.
They contain an unusually high concentration of odd-numbered carbon acids. They were detected exclusively in cerebrosides and sulfatides, which are considered characteristic myelin lipids. Their concentration in the brain increase very rapidly during the period of active myelination and more slowly during the later life of the animal (2).

* This investigation
was supported by Grants HD-05515 and NS-10741 from the National Institutes of Health, United States Public Health Service.
3 Present address, Institute of Biology, University of Tokyo, Meguroku, Tokyo, Japan.
In vivo experiments (3)(4)(5) indicate that the cerebral 2-hydroxy fatty acids are derived directly from the corresponding 1101~ hydroxy fatty acids. The 2-hydroxy acid is subsequently oxidatively decarboxylated to the fatty acid with one less carbon at,om. This sequence of in vivo degradation of fatty acids is referred to as one carbon degradation or a-oxidation. Available evidence suggests that cerebronic acid, a major 2-hydroxy acid of brain sphingolipids, is derived directly from lignoceric acid, a major nonhydroxy acid of these lipids (6).
The cerebral enzymes which convert 2-hydroxy fatty acids to nonhydroxy acids with one less carbon atom have been partially characterized (7)(8)(9). The enzyme responsible for 2-hydroxylation of the fatty acids has not yet been characterized, however. In this communication we describe some of the properties of an enzyme system from rat brain which converts lignoceric to cerebronic acid. The activity of this enzyme, which appears to be limited to the brain, was determined in various areas of the brain at different stages of development. sphingosine sulfate and psychosine sulfate, respectively (12). Mild alkaline methanalysis (13) was used to remove the benzoyl group from the 3-O-benzoyl ceramide in the ceramide synthesis.
[F4C]Lignoceroyl chloride was obtained by treating the free acid with thionyl chloride (12), and psychosine sulfate was prepared from beef brain cerebrosides according to a published procedure (14). Nonradioactive ceramide and cerebrosides were similarly pre-4124 pared with lignoceric acid or uL-cerebronic acid. The hydroxyl group of cerebronic acid was protected by acetylation prior to its use in the synthesis (15 were washed with ice-cold 0.9% NaCl and then homogenized in 2 volumes of ice-cold 0.15 111 NaCl using a Potter-Elvehjem homogenizer fitted with a Teflon pestle. The homogenates were centrifuged at 900 x g for 10 min, and the supernatant solutions (postnuclear fraction) served as the source of the enzyme described in these studies. Protein was determined according to the Lowry method as modified by Hess and Lewin (16).
Determination of Enzyme Activity-Enzyme preparations were assayed for their ability to convert [l-14C]lignoceric acid to cerebronic acid. In the standard assay the following components were incubated with vigorous shaking at 37" for 45 min in a final volume of 0.5 ml: 10 pmoles of Bicine buffer at pH 7.8; 0.5 pmole of MgC&; 0.2 pmole of DPNH; 2.06 nmoles of [l-'4C]lignoceric acid (2 x 1Oj cpm) coated on 10 mg of Celite (17); and 0.3 ml of enzyme preparation (4.5 to 5.0 mg of protein). The reaction was terminated by addition of 8 ml of CHCl,-CH,OH 1: 1. -4fter mixing and centrifugation, the clear supernatant solution was removed and the residue was washed with an additional 2 ml of the same solvent. The combined CHCI,-CHIOH extracts were mixed with 5 ml of CHC13, washed with 2.5 ml of water, and then dried under a stream of nitrogen.
Cerebronic acid was isolated from the crude lipid residue according t.o a published procedure (18) with minor modification. The lipid residue was treated with 2.5 ml of propyleneglycol containing 2.5 mmoles of KOH and 1.25 mg of cerebronic acid carrier for 30 min at 190-200".
The mixture was then cooled to room temperature, and most of the cholesterol was removed by extraction with 3 ml of hexane. The remaining propyleneglycol layer was acidified with 4 ml of 1 N HCI and then extracted with 8 ml of CHCls.
The chloroform layer, containing the free fatty acids, was washed with 6 ml of water, evaporated under a nitrogen stream, and then dried thoroughly over Pz05 under reduced pressure.
The dried residue was dissolved in a small volume of benzeneether 99: 1 and applied to a column containing 0.5 g of Unisil.
The column was first rinsed with 25 ml of benzene-ether 99: 1 to remove nearly all unreacted lignoceric acid and was then eluted with 15 ml of benzene-ether 9:l to obtain the 2-hydroxy fatty acids. The residue from the latter eluate was dissolved in 0.3 ml of CHC&-ethanol 2: 1, diluted with 1.1 ml of ethanol, and treated with 0.2 ml of 1 M cupric oleate in CHCl,, at 4" for 30 min. The insoluble copper chelates of the 2-hydroxy acids were collected by centrifugation, washed twice with 2 ml of ethanol-CHCl, 3:1, and then dissolved in 0.2 ml of 0.5 N methanolic HCl by warming.
The clear, yellowish solution was transferred to a counting vial with three 5-ml portions of a toluene-based scintillation mixture containing Hyamine hydroxide 10-X (8 ml of 1 M methanolic Hyamine hydoxide dissolved in a liter of the scintillation mixture). Identification of Product-The isolation procedure described above is specific for 2-hydroxy fatty acids (18). However, further confirmation of the identity of the product was accomplished by chemical degradation (a) and by thin layer chromatography (b). (a) The enzymic reaction products were recovered as free acids from the copper chelates by treatment with 1 N HCI (18). The fatty acids were then oxidized with permanganate (19) to CO2 and to the nonhydroxy fatty acids with one less carbon atom. 2-Hydroxy fatty acids are oxidized nearly quantitatively under these conditions.
The CO* was collected in a solution of Hpamine hydroxide and the amount of radioactive label determined.
The nonhydroxy fatty acids and small amounts of unreacted 2-hydroxy acids remaining after oxidation were extracted with ether from the reaction mixture.
The amount of radioactive label in these compounds was determined after fractionation by Unisil column chromatography. (b) The copper chelate was dissolved with warming in 1 ml of 0.5 N methanolic HCI. The methyl ester thus formed was extracted with hexane-ether 99: 1, analyzed by thin layer chromatography on Silica Gel G plate with hexane-ether 6:4 as the developing solvent, and subjected to radioautography.
Identification of Lipids Containing [14C]Cerebronic Acid--Lipids extracted from the enzymic incubation mixture were analyzed by thin layer chromatography-radioautography either directly or after fractionation by Unisil column chromatography. The radioactive spots were separately scraped and subjected to methanolysis (20). The fatty acid methyl esters thus obtained were examined by thin layer chromatography-radioautography.
Determination of Radioactivity-The radioactivity was determined wit,h a Tri-Carb liquid scintillation spectrometer model 3820. Eastman Kodak RP/R54 or NS-54T film was used for the detection of radioactive compounds on thin layer chromatograms.

Validation of Assay Method
[l-14C]Cerebronic acid, 10,000 cpm (0.45 mg), was added to a standard incubation mixture containing heat-inactivated enzyme in the absence of substrate and immediately processed. Most of the original radioactivity (86%) was recovered in the copper chelate of the 2-hydroxy acid. The greatest loss of radioactivity occurred during the saponification procedure.
Treatment of [l-14C]lignoceric acid in a similar manner resulted in recovery of 0.04% of original radioactivity in the copper chelate. When Unisil column chromatography was omitted from the procedure, 1 y0 of t,he radioactivity from [l-14C]lignoceric acid contaminated the &elate. Authentic [14C]kerasine was also ceramide or [Wlkerasine was incubat,ed with the brain preparation, no radioactivity was found in this fraction. The lipids in the lower phase of the Folch procedure from a standard assay mixture were fractionated on a Unisil column. Elution with ether produced free fatty acids and cholesterol, and additional elution with CHCl&HsOH 1:4 produced polar lipids. These fractions were then examined by combined thin layer chromatography-radioautography (Fig. 2). In addition to nonhydroxy fatty acids (unchanged substrate), many spots corresponding to complex lipids, including ceramides and cerebrosides, were found to be radioactive.
The spot corresponding to ceramides cont.aining nonhydroxy fatty acids was much more radioactive than the one containing ceramides with 2-hydroxy fatty acids as shown in Fig. 2. On the other hand, cerebrosides with nonhydroxy fatty acids were only slightly radioactive, while those with hydroxy fatty acids were quite radioactive. Each radioactive spot was scraped and subjected to methanolysis (22). The methyl esters formed were analyzed by thin layer chromatography-radioautography (Fig. 3). Radioactive spots corresponding to methyl cerebronate were obtained from ceramides and cerebrosides which contained 2-hydroxy fatty acids. All others gave only one radioactive spot corresponding to methyl lignocerate.
The configuration of the hydroxyl group of the cerebronic acid produced by the enzymic system was determined to be D from the following observations. The lipids extracted from the cnzymic reaction mixture were first subjected to a mild alkaline methanolysis procedure to remove most glycerolipids (13). The alkali-stable lipids were applied to a Unisil column which was eluted with CHC13-CH30H 98:2 to obtain ceramides and then with CHC&-CH30H 3 : 1 to recover cerebrosides and sulfatides (13). The ceramides were then examined by thin layer chromatography-radioautography.
The solvent system, CHC& CH30H 95:5, was used to separate ceramides containing D-and L-cerebronic acid as well as those containing lignoceric acid (23). The results indicated that the enzymic product did contain radioactive ceramides with both lignoceric acid and n-cerebronic acid. Chromatographic examination of the cerebrosides required five subsequent developments with CHCl&H~OH 9: 1 to effect such separation.
Like the ceramides, the cerebrosides with n-cerebronic acid or lignoceric acid were found to be radioactive.
The radioactive ceramide and cerebroside spots which contained n-cerebronic acid were recovered from the plates and subjected to methanolysis.
The liberated methyl esters were then examined by thin layer chromatography-radioautography. carried through this procedure, and it was found that complete cleavage of the amide linkage of cerebrosides occurred.
Testing a variety of detergents, including BRIJ's, MYRJ's, Tween's, Triton's, Miranol's, Cutscum, and bile acid salts, revealed none which was able to replace the Celite. The activity of a standard enzyme preparation increased with increasing amounts of Celite up to 5 mg, remained constant until 13 mg were added, and then gradually decreased with larger amounts (Fig. 1). Virtually no a-hydroxylating activity was observed when 100 mg of Celite were used.
CO,, which might have been produced during the enzymic incubation, was collected and was found to be nonradioactive. Apparently, the enzymes for oxidative decarboxylation of cerebronic acid (9) are not active under these experimental conditions, or the cerebronic acid produced is immediately incorporated into an amide bond.

Identijicalion of Product
Approximately 0.5 mg of 2-hydroxy fatty acid containing 1228 cpm was isolated from a standard incubation product and oxidized with permanganate.
As shown in Table I, nearly all of the radioactivity was recovered as CO,. The products of the enzymic reaction were converted to their methyl esters and were analyzed by thin layer chromatography-radioautography.
Radioactivity was detected only at an RF value equal to that of authentic cerebronic acid. These results, together with the specific method of assaying, indicated that the product was [I-%]cerebronic acid. When the lipids from the enzymic reaction mixture were partitioned by the Folch procedure (21), approximately 1% of the radioactivity was recovered in the upper phase. This was a lo-fold increase over the radioactivity obtained with a heat-inactivated enzyme.
lhone of this radioactivity was attributed to the presence of either free 2-hydroxy acids or 2-hydroxy acids attached to lipids.
The radioactivity was most likely due to the formation of a Cob thioester of the substrate.
When ['"Cl- FIG. 2 (left). Autoradiogram of a thin layer chromatographic separation of crude lipids extracted from a standard assay reaction mixture.
Conditions for the incubation are indicated in the legend of Fig. 1. Developing solvent used was CHCls-CHaOH 9:l.
The crude lipid was preliminarily fractionated on a Unisil column. Fraction E was eluted with ether and contained free fatty acids, both nonhydroxy and 2-hydroxy, and cholesterol. Fraction CM was eluted with CHCla-CHaOH 2:l and contained polar lipids. The radioactive spot in Fraction E corresponds to free lignoceric acid. Cerebronic acid would have remained at the origin under these conditions.
In the radioautogram of Frac-All of the radioactivity present in these lipids were found in cerebronic acid.

Properties of Enzyme
Eject of pH--,Maximum activity of the enzyme was reached at a pH of 7.8 with every buffer tested (Fig. 4). Bicine and Tris buffers were the most useful among those tested and phosphate was least effective.
Effects of Metals-Mg++ was required for full activity of the cr-hydroxylating enzyme, and could not be replaced by Ca++ and Mn++, which were somewhat inhibitory.
Most heavy metal ions, including iron and copper, were strongly inhibitory at 0.1 mM, except for nickel, which was inhibitory only at concentration above 4 mM. EDTA strongly inhibited the enzyme while iron chelators, such as dipyridyl and 0-phenanthroline, had no effect. EGTA (ethyleneglycol-bis @-aminoethyl ether)-N, N'tetraacetic acid), which binds calcium in preference to Mg++, also had no effect on the enzyme activity (Table II).
Pyridine Nucleotide Requirement-The results shown in Table  III indicate that pyridine nucleotides are essential for the enzyme activity.
The activity increases linearly with increasing concentration of DPNH up to 0.15 mM, plateaus briefly and then tion CM Spots 1 and .!? correspond to ceramides containing nonhydroxy or 2-hydroxy fatty acids, respectively, and 4 and 6 correspond to cerebrosides containing these fatty acids. FIG. 3 (right). Autoradiogram of a thin laver chromatogram of fatty acid methyl esters obtained by methanolysis of each spot from Fraction CM of Fig. 2. The number of each samnle corresponds to the spot n<mber given in Fig. 2. The row -of spots labeled N on the left side corresponds to methyl lignocerate and the row labeled H corresponds to methyl cerebronate (indicated by arrows).
gradually decline with further increases in the concentration (Fig. 5). TPNH was as effective as DPNH at all concentrations employed.
Aerobic Nature sf Reaction-The hydroxylation reaction apparently requires OS (Table IV) in conjunction with a pyridine nucleotide, a finding which suggests that the conversion may be catalyzed by a "mixed function oxidase." The presence of cytochrome P-450 in this system was not implicated, since CO did not inhibit the activity.
Most common inhibitors of the energy linked electron transfer system strongly inhibited the enzymic hydroxylation as also shown in Table IV. These results suggest that the hydroxylation system is closely related to the respiratory chain and not to the hydroxylation system which depends upon cytochrome P-450. E$ect of Various Other Cofactors-The flavins, FMN and FAD, were both strongly inhibitory at a concentration of 1 mM, while ATP, CoA, pyruvate, and succinate at the same concentration had no effect on the cu-hydroxylating activity.
Substrate Spec$icity-The brain a-hydroxylating system appears to be highly specific for free lignoceric acid as indicated in    chain on ol-hydroxylating activity The complete system described in the legend of Fig. 1 was incubated under one of the gas phases listed with or without inhibitor addition. slight activity detected for them (Table   V) could be due to hydrolysis (15, 24) and utilization of released lignoceric acid for the hydroxylation.
Other Inhibitors and StimzLlaters-The effect of several lipids, cerebronic acid, 2-keto lignoceric acid, ceramides, and cerebrosides which are metabolically related to the enzymic substrate or product of the a-hydroxylating enzyme were found to be inhibitory (Table VI). Sphingosine, which is an acceptor of 2hydroxy fatty acid CoA esters (25), had no effect, while psychosine was stimulative.
The observed stimulation by tripalmitin is difficult to explain.
Sucrose completely inhibited the enzymic reaction when it was employed to homogenize the brain tissue. The inhibitory effect of sucrose was confirmed by adding various amounts of sucrose to the standard assay system (Fig. 6).
Apparent Michaelis Constant- Fig.  7 indicates the effect of substrate concentrations on the enzymic reaction. The apparent K, was found to be 4.2 PM by using Lineweaver-Burk plots of the data.
Requirement for Unknown Cofactor-Cerebronic acid was not produced until approximately 2 mg of protein were added.  Conditions for assay are described in the legend of Fig. 1.
When various amounts of the postnuclear fraction of rat brain were incubated with /l-14C]lignoceric acid, product formation increased linearly with protein addition until a plateau was reached at approximately 4 mg (Fig. 8). However, the rate of formation of cerebronic acid became a linear function of the amount of enzyme added between 0 and 4 mg of protein, when boiled postnuclear fraction was added so that the total amount of protein was held constant at 5.5 mg in each tube. Bovine serum albumin could not replace boiled 900 x g supernatant.
This suggests that there is a cofact.or which is required above a certain concentration.
No change was noted in the clear supernatant when the boiled enzyme was centrifuged at 1500 rpm. These results indicate that there is an unknown water soluble and heat stable cofactor which is essential for the a-hydroxylation.
Other Properties of Enzyme-The postnuclear fraction of the rat brain completely lost its ol-hydroxylating activity when heat,ed at 60" for 1 min. The amount of cerebronic acid formed increased proportionately with the time of incubation at 37" up to 1 hour as shown in Fig. 9.
Change of Enzymic Activity as Function oj Age-Postnuclear fractions tiere prepared from the brains of rats of various ages, and their activities for ol-hydroxylation were assayed under optimum condibions.
Slight a-hydroxylating act,ivity is found in 7-day-old rat brain, but the specific activity then increases sharply until the animal reaches 21 days, at u hirh time it plateaus for 2 weeks and then declines (Fig. 10).
Distribution of a-Hydroxylating Activity-In 30-day-old rats the postnuclear fraction of the cerebellum had the highest activity (7.23 pmoles per mg of protein per hour).
Thediencephalone (3.18 pmoles), t.he cerebral hemispheres (0.88 pmole), and brain stem (0.57 pmole) all had lower activities. Under the conditions used, no activity for converting lignoceric acid to cerebronic acid was detected in the postnuclear fractions of liver, kidney, spleen, heart, and testes. DISCUSSION All the cerebronic acid formed from lignoceric acid by the brain enzyme system was recovered in ceramides or cerebrosides and not as the free acid or its CoA ester (Figs. 2 and 3). 1 he CoA esters of 2-hydroxy fatty acids are converted to ceramides (25) and then to cerebrosides in the brain (17). In a study of fatty acid chain length specificity of the enzymes catalyzing the synthesis of cerebrosides, Radin (26) suggested that the rate limiting step for the synthesis of cerebrosides containing 2.hydroxy acids is at the step of ceramide formation.
For cerebrosides containing nonhydroxy acids the rate limiting step is the galactosyltransferase reaction which converts ceramides to cerebrosides.
Our finding that all newly synthesized cerebronic acid is immediately incorporated into ceramides ar:d cerebrosides indicates that the cY-hydroxylation of fatty acids may be the rate limiting step for the synthesis of cerebrosides u hich contain 2-hydroxy acids. Both the developmental studies (Fig. 10) and the tissue distribution studies suggest a close relationship between ol-hydroxylating activity and both cerebroside accumulation and myelin formation in brain.
The observation that newly synthesized 2-hydroxy acids are immediately converted into sphingolipids suggests that the 2-hydroxy fatty acids which serve as substrates for the a-oxidation system may be those released from cerebrosides by cerebroside galactosidase (15) and ceramidaee activity (24).

2-Hydroxy
fatty acids in the brain cluster around specific chain lengths.
The most predominant cluster is between carbon atoms 22 to 26, with cerebronic acid as the major component. Another group found in much smaller amounts includes CM and Cl8 (I). Ullman and Radin (25) studied the specificity of brain acyltransferase in the synthesis of ceramides and found that the transferase is not a controlling factor in the distribution of 2-hydroxy acids in cerebrosides.
Our finding that the a-hydroxylating enzyme is specific for lignoceric acid and is not active with stearic acid and other fatty acids (Table V) suggests that this enzyme might control the formation of specific homologs of cerebrosides containing 2-hydroxy acids.
Many biological hydroxylating systems have been shown to involve cytochrome P-450. W-and w-l hydroxylation of fatty acids in mammalian liver (27,28), kidney (29,30), some yeasts (31,32), and bacteria (33,34) have been studied and found to involve a mixed-function mono-oxigenase requiring a reduced pyridine nucleotide and molecular oxygen.
Cytochrome P-450 appears to play a central role in many of these hydroxylating systems. The hydroxylating enzyme of brain does not appear to involve cytochrome P-450 because CO did not inhibit the reaction as indicated in Table IV, but it was sensitive to general inhibitors of the respiratory chain, such as cyanide and antimycin A (Table IV).
The hydroxylating enzyme of brain therefore appears to be linked to the respiratory chain of the brain based on these observations. Enzymes which hydroxylate the cy-position of fatty acids are present in young plant leaves (35,36), germinating peanut (37, 38), bacteria (39), and mammalian liver (40). The liver enzyme is responsible for conversion of phytanic acid to 2-hydroxyphytanic acid. This enzyme is missing in patients of Refsum's disease (41). IJnfortunately, most of these enzymes 9. LIPPEL, K., AND 40), the liver preparation did not hydroxylate lignoceric acid, while the brain preparation was not capable of oxidizing phytanic acid (42). Comparison of the enzyme from brain with that from plants reveals several differences.
The enzyme from peanut requires both a DPN and an HzOz generating system (37, 38), while the young pea leaf system utilizes DPN and O2 for the hydroxylation (35,36). Both enzymes are strongly inhibited by imidazole.
The enzyme from brain differs in that it requires DPNH and O2 and is slightly stimulated by imidazole.
The a-hydroxylating enzyme of brain is very labile, and loss of activity during subfractionation of the postnuclear fraction has been observed.
However, we recently have succeeded in restoring some of the activity after combining a crude mitochondrial fraction with the 100,000 x g supernatant.
Neither fraction exhibited any activity separately, suggesting that both fractions are necessary for the activity.
Heating either fraction resulted in complete loss of activity.
These results, combined with the requirement for a heat-stable, unknown factor found in the postnuclear fraction, suggest that there are three components necessary for the cu-hydroxylating activity; a mitochondrial fraction, a heat-labile, soluble fraction, and a third heat-stable, soluble fraction.
Upon ultrafiltration of the cytosol fraction, the heat-labile factor emerged in fractions corresponding to a molecular weight more than 10,000. Further characterization of these fractions are currently underway in this laboratory.