Enzymatic alkylenation of phospholipid fatty acid chains by extracts of Mycobacterium phlei.

Abstract The enzymatic synthesis of tuberculostearic acid (10-methylstearic acid) was catalyzed by extracts of Mycobacterium phlei. This process involved two reactions of the olefinic fatty acid chain of phospholipids. The chain was first alkylenated at the 10-carbon to give a methylene group, which was subsequently reduced to a methyl group. The first reaction could be measured by using S-adenosylmethionine-methyl-14C. The enzyme was found in the supernatant fraction when extracts of cells broken by sonic oscillation were subjected to centrifugation at 100,000 x g. S-Adenosyl-l-methionine was the only effective donor of the 1-carbon unit. Phosphatidylglycerol, phosphatidylinositol, and phosphatidylethanolamine were substrates for the reaction, and both 16- and 18-carbon chains were alkylenated although only the Δ9-olefinic chains appeared to be converted. The enzyme acted upon chains at either position 1 or 2 of the glyceride molecule. Several detergents had little effect on the rate of the reaction.

Tuberculostearic acid (IO-methylstearic acid) is a characteristic component of the membrane lipids of the mycobacteria (1,2).
The synthesis of this compound was first studied in viva by Lennarz, Scheuerbrandt, and Bloch (3), who established that the precursors were oleic acid and the methyl group of methionine. Later experiments by Jaurhguiberry et al. (4,5) showed that the transfer of the methionine methyl group occurred with loss of a proton.
The reaction was also accompanied by a shift of the olefinic proton at the lo-carbon of oleate to the g-carbon (6). These facts led to the formulation of a logical mechanism: Further evidence for this scheme was produced by Jaurbguiberry et al. (7), who isolated labeled lo-methylenestearic acid and labeled tuberculostearic acid from a cell homogenate of Mycobacterium phlei which had been incubated with methioninemethyl-% and subsequently saponified. The details of the reaction, including the nature of the endogenous lipid acceptor, the actual methyl donor, and the hydrogen donor for the reduction of the lo-methylene group to a methyl group, remained to be specified.
EXPERIMENTAL PROCEDURE Materials M. phlei ATCC 354 cells were grown at 30" on Sauton medium (9), harvested at middle or late log phase by filtration, and washed with distilled water.
Washed cells were stored at -15'. S-Adenosyl-L-methionine-methyV4C was purchased from New England Nuclear.
Unlabeled X-adenosyl+methionine was purchased from Calbiochem.
Fresh yeast was purchased from Fleishmann Division of Standard Brands, Inc., Chicago, Illinois.
The venom of Ancistrodon piscivorus piscivorus was obtained from Calbiochem.
Crude lipids of Axotobacter a&is were kindly provided by Dr. A. E. Chung, University of Pittsburgh.
All of the detergents used here are commercially available, and the sources were listed in a previous paper (10). Methyl lo-methylene stearate was synthesized by Dr. Auste Paliokas, University of Chicago. Starting with lo-hydroxystearic acid, a gift of Dr. George Schroepfer, University of Illinois, the methyl lo-ketostearate was easily prepared by chromic acid oxidation (11) and converted to the IO-methylene stearate by a modified Wittig reaction (12).

Preparation of Cell-free Extracts
Washed cells (0.2 g per ml) were suspended in 0.1 M phosphate buffer (pH 7.0) and disrupted at O-4" for 10 to 15 min in a sonic disintegrator (Branson Instruments, Inc., Stanford, Connecticut). The disintegrated material was centrifuged at 10,800 X g for 10 min. The supernatant was removed and recentrifuged, first at 10,500 x g for 10 min and then twice at 20,000 x g for 30 min. The resulting supernatant, designated 20-S was further centrifuged at 105,000 X g for 60 min in the Spinco model L preparative centrifuge. This supernatant was adjusted to 10 mg of protein per ml with 0.1 M phosphate buffer (pH 7.0), and designated 100-S. This crude enzyme preparation retained full activity after storage at -10" for 6 months.
The precipitate was resuspended in phosphate buffer and designated 100-P.

Preparation of Acetone-treated Enzyme
Crude enzyme preparations generally contained endogenous lipid substrates.
For removal of endogenous lipids we tried acetone treatment, butanol extraction, alcohol-ether extraction, and treatment with ammonium sulfate followed by Sephadex column chromatography.
Of these, only acetone treatment yielded an enzyme preparation which was significantly free of endogenous lipid substrates and was also active.
The crude extract (10 mg of protein per ml, 10 ml) was added dropwise to 70 ml of cold acetone ( -10 to -15") with stirring.
After about 15 min, this suspension was centrifuged at 12,000 x g for 10 min and the supernatant was discarded. The pellet was partially dried by subjecting it to water pump vacuum for 20 min in the cold. The wet acetone powder was ground in a chilled mortar to remove the residual acetone and then mixed with 10 ml of 0.1 M phosphate buffer (pH 7.0). The resulting suspension was centrifuged at 27,000 x g for 30 min and the supernatant was removed.
The supernatant thus obtained was again added to cold acetone and all of the procedures were repeated. Finally, the second acetone precipitate was dissolved in 10 ml of 0.1 M phosphate buffer (pH 7.0), and a clear yellow supernatant was obtained by centrifuging at 37,000 X g for 60 min. This acetone-treated enzyme preparation, when stored at -lo", retained full activity for at least 1) months.

Assay of Enzyme Activity
Method l-Crude extracts or acetone-treated enzyme were incubated with X-adenosylmethionine-methyl-14C and other additions as indicated in tables or figures.
Most incubations were carried out at 30" for 1 hour.
Reactions were stopped by the addition of an equal volume of methanol and 0.2 volume of 85% KOH.
These solutions were heated in a boiling water bath for 30 min. After cooling, methanol was removed with a nitrogen stream.
The alkaline solutions were acidified to pH 1 by the addition of 6 N HCl.
The acidified solutions were extracted three times with 3 ml of ether and the combined ether extracts were washed twice with 2 ml of 2% KCl.
Aliquots were taken for radioactivity measurement and for examinations by chromatography on a thin layer of silicic acid. Thin layer chromatography solvent systems were chloroform-methanolwater (75:15: 1) or petroleum ether (30-60")-ether-acetic acid (60:40: 1) and the spots corresponding to free fatty acids were detected by radioautography and scraped from the plates into scintillation vials for radioactivity measurement.
The methyl esters of the extracted fatty acids were prepared by treatment with diazomethane (13). In order to distinguish methylene fatty acid esters from methyl branched fatty acid esters, 25% silver nitrate-impregnated silicic acid thin layer chromatography was used with the solvent system petroleum ether (30-60% ether (9: 1). Gas-liquid chromatography was performed on a 4-or g-foot column, operated at 190-200' with a helium flow rate of 25 to 50 ml per min. Radioactive fatty acid esters were trapped at the outlet of the column in a small capillary tube. For counting, fatty acid esters were rinsed with toluene scintillator fluid from the collection tube into vials.
Method Z-After incubation for appropriate times, reactions were terminated by the addition of 30 ml of chloroform-methanol (2: 1) to the incubation mixture.
The phases were allowed to separate and the aqueous layer was discarded.
The lower chloroform layer was passed through a sintered glass filter in order to remove the coagulated protein, and the filtrate was evaporated to dryness. The residue was taken up in chloroform-methanol and applied to a silicic acid thin layer chromatography plate.
The solvent system was chloroform-methanol-water (65 : 25 : 4). Radioactive compounds were detected by scanning with a Nuclear-Chicago Actigraph III, model 1002 strip scanner with model 1006 thin layer chromatography plate scanner attachment, or radioautography. Spots were visualized by exposure to iodine vapors, and the areas corresponding to various lipids were scraped into scintillation vials for counting.
In order to confirm whether the radioactivity in phospholipid spots was truly associated with fatty acyl groups, the compounds were eluted with methanol and hydrolyzed, and the fatty acids were examined by thin layer chromatography and gas-liquid chromatography.
Radioautography was performed as described earlier (14). Scintillation counting used a toluene fluid described by Hildebrand and Law (15)  This is essentially a measure of the first enzymatic step in which the X-adenosylmethionine methyl group is transformed into a fatty acid methylene group.
In most cases the second step, reduction to a methyl group, was ignored, although we occasionally report the distribution of radioactivity into both methylene (unsaturated) and methyl (saturated) fatty acids (e.g. Table III).

Catalytic Hydrogenation of Radioactive IlO-Methylene Stearate
Radioactive IO-methylene stearate was collected from the gas-liquid chromatography column and mixed with 1 mg of carrier lo-methylene stearate. This was dissolved in 1 ml of methanol and 2 mg of PtOz were added. Hydrogenation was carried out overnight with shaking at 30-psi HP pressure.
The reaction mixture was filtered through a sintered glass filter and the catalyst was washed with methanol and ether. The combined filtrate was reduced in volume and subjected to gas-liquid chromatography.

Preparation of Lipid Substrates
Lipids from Log Phase Cells of M. phlei-Total lipids were extracted with chloroform-methanol (2:l) from M. phlei cells harvested in the log phase of growth.
Extracts were washed by the procedure of Folch, Lees, and Sloane Stanley (16). The lipid phase was evaporated to dryness and the residue was extracted with acetone several times. The residue was suspended by the dialysis technique of Fleischer and Klouwen (17) with the slight modification that the insoluble material was removed from the but,anol phase by centrifugation before beginning the dialysis.
Micellar dispersions of M. phlei lipids prepared by the dialysis technique were subject to thin layer chromatography. Four phospholipid spots appeared on staining with iodine: two strong spots corresponding to phosphatidylinositol oligomannosides, one to cardiolipin, and one weak ninhydrin-positive spot corresponding to phosphatidylethanolamine. Standards for these compounds were prepared by Akamatsu and Nojima (18).
Phosphatidylethanolamine and Phosphatidylglycerol from A. a&is-Crude A. agilis lipids were chromatographed by the method of Es&n and Law (19) to give pure phosphatidylethanolamine and crude phosphatidylglycerol.
Pure phosphatidylglycerol was obtained from the crude phosphatidylglycerol fraction by silicic acid column chromatography.
Phosphatidylinositol and Cardiolipira from Yeast-Bakers yeast phosphatidylinositol and cardiolipin were prepared by the method of Trevelyan (20) and purified by silicic acid column chromatography.
Phosphatidylinositol was obtained in pure form, while cardiolipin had a trace of phosphatidylinositol as determined by thin layer chromatography.
Aqueous dispersion of phospholipid preparations were obtained by the dialysis technique of Fleischer and Klouwen (17) or by sonically dispersing the lipids in 0.1 M phosphate buffer (pH 7.0).
Phospholipase AZ Action on Radioactive PhospholipidsRadioactive lipids were extracted from the reaction mixture as described in Method 2 and separated by preparative thin layer chromatography.
The lipids were hydrolyzed with phospholipase At from A. piscivorus piscivorus by the methods of Hildebrand and Law (15). After thin layer chromatography, free fatty acids were eluted with chloroform and lyso phospholipid was eluted with methanol.
The recoveries of both fractions were almost quantitative.
The fractions were counted and the fatty acid composition of each fraction was determined by gasliquid chromatography.

Analytical Methods
Protein concentration was determined by the method of Lowry et al. (21) with bovine serum albumin as a standard.
Phosphorus was determined by the method of Allen (22).

Although
we used the procedure reported by Jaureguiberry et al. (7) and Azerad, Bleiler-Hill, and Lederer (23) for preparing homogenates of M. phlei, we used much smaller volumes for incubations and S-adenosylmethionine instead of methionine. We used longer periods of centrifugation at higher speeds (see "Methods") because we found that the preparation used by Jaureguiberry et al. (7) contained many whole bacterial cells. Further centrifugal fractionation showed that most of the enzymatic activity resided in the 100,000 X 9 supernatant fraction ( Table I). Addition of phospholipids to the reactions did not stimulate incorporation into alkylenated fatty acids (Table I). Our first task was to establish the nature of the labeled fatty acid derivatives formed in this system when S-adenosylmethionine served as the substrate.
For this purpose we used crude 100,000 X Q supernatant fraction with S-adenosylmethioninemethyl-14C and endogenous lipid acceptors.
The fatty acid fraction isolated after saponification was treated with diazomethane and the esters were chromatographed on silver nitrate thin layer chromatography.
Most of the radioactive compounds were found in the saturated and unsaturated fatty acid ester fractions (Fig. 1). The unsaturated fraction usually contained most of the activity, although the ratio of the two fractions varied with enzyme preparations and incubation conditions.
The formation of lo-methylene stearic and lo-methylstearic acid bears a strong resemblance to the enzymatic formation of cyclopropane fatty acids, a process also involving reaction of Sadenosylmethionine with olefinic chains of phospholipids (10,14). Since some of the mycobacteria produce very long chain fatty acids with cyclopropane rings (24,25), it was important to  Table  II. After hydrolysis, the acids were extracted and converted to methyl esters. Plates were prepared with 25% AgNOa in Merck Silica Gel G. The solvent system was petroleum ether (30-60')ether (9:l). Spots were located by spraying with 2,4-dichlorofluorescein and viewing under ultraviolet light. Bands were then scraped from the plate for the radioactivity determination. The figure shows the plate at the bottom, and the level of radioactivity in each area of the plate above.
MS, methyl lo-methylene stearate; TS, methyl tuberculostearate. establish that the enzymatic products that we isolated were not cyclopropane products, but indeed lo-methylene stearic and IO-methylstearic acids. A gas-liquid chromatographic system was therefore devised which would separate esters of these two compounds from each other and from cyclopropane derivatives of comparable chain length.
For this purpose, OV 17 served as an appropriate liquid phase (Fig. 2). Note that both methyl tuberculostearate and methyl IO-methylenestearate were labeled. No label was found in the cyclopropane fatty acid ester fraction during these studies. In order to characterize the labeled product which chromatographed with carrier methyl lo-methylenestearate, the material was isolated from the column and reduced with hydrogen and platinum in methanol. These conditions do not reduce cyclopropane fatty acid esters. All of the methylene compound was converted to labeled methyl lo-methyl stearate, as judged by rechromatography (Fig. 2). Early experiments with added exogenous lipids gave disappointing results (Table I), for inhibition usually resulted.
An indirect approach was therefore used to obtain information about the endogenous lipid substrate.
Following incubation with labeled S-adenosylmethionine, the enzymatic reaction mixture was extracted with solvents and the labeled lipid products were chromatographed on thin layer chromatography. The lipid fractions were isolated and saponified, and the fatty acids were extracted and counted.
The results provided a clear The radioactive esters were obtained as described in Fig. 1   indication that both IO-methylene stearic acid and lo-methylstearic acid formed in the enzymatic reaction were bound to phospholipid molecules (8). The results were somewhat complicated by the formation of large amounts of other labeled products, one of which ultimately was proved to be a labeled fatty acid methyl ester formed by alkylation of the fatty acid car-boxy1 group (26).
A more direct approach became feasible when we obtained an acetone-treated preparation which was relatively free of endogenous lipid substrates.
Synthesis of labeled fatty acid derivatives now became dependent upon addition of the extracted lipid fractions or purified phospholipids (Table II). Addition of various phospholipids stimulated the formation of labeled fatty acid derivatives (Table II), but fatty acid methyl esters were also formed (Fig. 3). This was probably the result of the liberation of free fatty acid by the action of phospholipases, followed by carboxyl alkylation catalyzed by fatty acid methyl ester synthetase (26). Different acetone-treated enzyme prepa-rations gave somewhat different results with regard to their response to added phospholipids, to the level of activity which survived acetone treatment, and to the ratio of saturated and unsaturated acid products. Table II shows results with three typical preparations.
Depending upon the nature of the fatty acid chains in the added phospholipid substrates, labeled products of differing chain lengths were synthesized.
Thus, when a phosphatidylethanolamine from A. agilis, which contained a large proportion of palmitoleic acid (Table III) was added, labeled products were formed with retention times on gas-liquid chromatography corresponding to 17-carbon acid esters. Lennarz, Scheuerbrandt, and Bloch (3) had observed the formation of a 17-carbon homologue of tuberculostearic acid in M. phlei. Very little 19-carbon product was formed when phospholipids from A. agilis were used, probably because cis-vaccenic acid rather than oleic acid is present (27). Table III summarizes the products observed with various phospholipids.
Because of the report of Jaureguiberry et al. (7) in which methionine-methyl-14C was used as a substrate, we decided to test various possible Cl donors for the synthesis of fatty acid products in our enzyme preparations.
The data of Table IV show that only S-adenosylmethionine was an effective substrate for this reaction.
Several parameters for the enzymatic reaction were examined. Although no extensive determination of this effect of pH variation has been made, the reaction had about twice the initial velocity at pH 7 as at pH 8. The reaction rate was linear with protein concentration over a 4-fold range, and with time up to 1 hour under the usual conditions of incubation.
The effect of the following detergents was tested: sodium dodecyl sulfate, Triton X-100, Tween 80, Cutscum, and hexadecyltrimethyl ammonium bromide.
None had any dramatic stimulatory effect on the rate of the reaction.
The question of positional specificity for olefinic chains at position 1 or 2 on the phosphatide was examined by isolating the fatty acids from each position after the enzymatic alkylenation of phosphatidylglycerol and phosphatidylinositol fatty a cis-Vaccenic acid (27). b Oleic acid (20).
acyl chains and phospholipase AZ treatment in the manner of Hildebrand and Law (15). Table V summarizes the results of this experiment. DISCUSSION The transalkylenating enzyme in M. phlei extracts which catalyzed the formation of derivatives of lo-methylene stearic acid was obtained in soluble form when cells were broken by sonic oscillation.
The fact that some activity remained in the membrane particles sedimenting at 100,000 X g may indicate that the enzyme originally resided in the membrane, where its lipid substrate doubtless was localized.
The enzyme catalyzed the reaction between S-adenosyl-L-methionine and olefinic acid chains in phospholipid molecules.
While it was readily established that S-adenosylmethionine was a more effective methylene donor than methionine, serine, or formate, the nature of the fatty acid substrate proved more difficult to establish. The major problem was that crude extracts contained adequate lipid substrate and were not stimulated by the addition of exogenous lipids.
Acetone treatment, while it resulted in some loss of activity, removed a significant, although variable, amount of endogenous lipid substrate.
The best preparations showed almost lOO-fold stimulations when exogenous phospholipids were added, although lo-fold stimulations were much more common.
Free oleic acid gave no stimulation of synthesis of lo-methylene stearate derivatives. Added phospholipids were alkylenated on the fatty acid chains, and the otherwise unaltered phospholipids could be extracted and recovered by thin layer chromatography.
Hydrolysis and gas-liquid chromatography showed that both Cl6 and Cu olefinic acid chains were alkylenated to give the Cir and Crs acid derivatives. The ratio of products reflected to some degree the composition of the substrate phospholipids.
Experiments designed to determine whether the alkylenating enzyme had a preference for olefinic acid at position 1 or 2 of the glyceride gave perplexing results.
Natural phospholipids of different fatty acid compositions were used as substrates.
The alkylenated chains were labeled from S-adenosylmethionine-methyl-14C, and product acids from each position were isolated after treatment with phospholipase A2, which specifically hydrolyzes an acid esterified at position 2 of a phosphatide.
With phosphatidylglycerol and phosphatidylinositol as substrates, the enzyme alkylenated predominantly position 2 of the inositide, but preferred position 1 of the phosphatidylglycerol by a factor of 2: 1. An important difference between these substrates was the absence of oleic acid in the phosphatidylglycerol.
However, appropriate olefinic precursor acids were present at both positions 1 and 2 of both substrates, yet the alkylenation position differed for the two. A definitive solution to the problem of what determines positional specificity for the alkylenating enzyme must await the availability of appropriate phospholipid substrates with two identical olefinic chains.
It is of interest to compare the properties of the Mycobacferium alkylenating enzyme with the cyclopropane synthetase enzymes from Clostridium amd Serratiu (10,14). All three of these enzymes catalyze the alkylenation of olefinic acid chains in phospholipid molecules, with S-adenosylmethionine as a methylene donor.
The products are similar, except that the Mycobacterium enzyme gives a branched methylene group instead of a cyclopropane ring. All three enzymes are soluble when the cells are broken by sonic disruption or similar drastic treatments. The Mycobacterium and Xerratia enzymes contain adequate amounts of endogenous phospholipid substrate (14). The Clostridium enzyme shows a preference for phosphatidylethanolamine and for the fatty acid chain in position 1 (15,28). Phosphatidylglycerol and phosphatidylinositol are better substrates for the Mycobacterium enzyme and the positional specificity is less well defined. Both Mycobacterium and Clostridium show an unusual distribution of fatty acids in the phosphatides, in which unsaturated and branched or cyclopropane acids predominate in position 1 (15,29).
Cyclopropane synthetase from Clostridium is stimulated several-fold when the phosphatidylethanolamine substrate is mixed with an anionic detergent.
The Mycobacterium alkylenating enzyme is not stimulated appreciably by any of several types of detergent.
As in the case of cyclopropane fatty acid synthesis, tuberculostearic acid increases in the stationary phase of growth (3). It seems likely that both processes involve enzymes which alkylenate the membrane phospholipids after these have been deposited by the phospholipid synthetases.
They are the only two presently known processes in which fatty acid alkyl chains of phospholipids undergo enzymatic alteration. The definitive raison d'etre for such enzymatic processes has yet to be offered.