Plant Physiol. (1987) 84, 762-765 0032-0889/87/84/0762/04/$0 1.00/0 Acyl Coenzyme A Preference of Diacylglycerol Acyltransferase from the Maturing Seeds of Cuphea, Maize, Rapeseed, and Canola' Received for publication November 30, 1986 and in revised form February 9, 1987

YI-ZHI CAO AND ANTHONY H. C. HUANG* Department ofBiology, University ofSouth Carolina, Columbia, South Carolina 29208

ABSTRACr it is assumed that the in vivo pool sizes ofthe acyl CoAs determine the acyl specificity in the 3-position of the triacylglycerol. In their seed triacylglycerols, Cuphea carthagenensis contains 62% In seed oil biotechnology, one major goal is to alter the chain lauric acid; maize possesses 50% linoleic acid and 30% oleic acid; length of the fatty acyl moiety of the triacylglycerols (5, 12, 15). rapeseed (Brassica napus L. var Dwarf Essex) has 40% erucic acid; and Researchers are using different approaches to manipulate the Canola (Brassica napus L. var Tower) holds 60% oleic acid and 23% genes controlling enzymes for the elongation of fatty acids. This linoleic acid. Diacylglycerol acyltransferase (EC 2.3.1.20) in the micro- manipulation appears to be theoretically workable. However, somal preparations fron maturing seeds of the above species were tested what is not known is whether the newly designed fatty acid can for their preference in using different forms of acyl coenzyme A (CoA). be accommodated by the other components ofthe triacylglycerol Lauroyl CoA, oleoyl CoA, and erucoyl CoA individually or in equimolar synthesis machinery. The acyl preference ofall the three glycerol mixtures at increasing concentrations were added to the assay mixture acyltransferases in maturing seeds has been studied only with containing diolein, and the formation of trinacylglycerols from the acyl physiological substrates of palmitoyl CoA and C-18 acyl CoAs groups at 24, 32, and 40°C was analyzed. The Cuphea enzyme preferred having 0-3 double bonds (3, 6, 8, 16). Their activities on shorter lauroyl CoA to oleoyl CoA, and was inactive on erucoyl CoA. The maize or longer nonphysiological acyl CoAs are unknown. enzyme had about equal activities on oleoyl CoA and lauroyl CoA, and In an attempt to resolve part of the above unknown, we have was inactive on erucoyl CoA. Enzymes from both rapeseed and Canola studied the acyl CoA preference of seed diacylglycerol acyltrans- had the same pattern of acyl CoA preference, with highest activities on ferase. We also would like to see if the enzyme exerts preference lauroyl CoA. The two enzymes were more active on oleoyl CoA than on on specific acyl CoA, such that this preference reflects the acyl erucoyl CoA at high acyl CoA concentrations (10 and 20 micromolar) at moiety in the triacylglycerol. The enzyme was obtained from 24°C, but were more active on erucoyl CoA than on oleoyl CoA at low four selected oil seeds which have unique and very contrasting acyl CoA concentrations (1.36 micromolar or less) at 32 and 40°C. These fatty acyl moieties in the triacylglycerols. findings are discussed in terms of the contribution of the enzyme to the acyl specificity in storage triacylglycerols and the implication in seed oil biotechnology. MATERIALS AND METHODS Plant Materials. Wild plants of Cuphea carthagenensis con- taining maturing seeds were collected from a local swamp (cour- tesy of Cynthia Smith, Curator of the Biology Department Her- barium). Plants ofinbred maize (Zea mays L., Mo 17), rapeseed (Brassica napus L. var DwarfEssex), and Canola (Brassica napus In oil seeds, the fatty acid composition of the storage triacyl- L. var Tower, courtesy of Dr. W. D. Beversdorfofthe University glycerols is species- and variety-specific, and environmental fac- of Guelph) were grown in a local garden. The oil storage tissues tors such as temperature exert some modifying effects (13, 16). (scutella of maize and the cotyledons of other species) of matur- Within a seed species, the fatty acid composition in each of the ing seeds at a stage when the fresh weights were approximately three positions of a triacylglycerol is also largely inherited (16). halfof the values in the mature seeds were used. In oil seeds as well as in mammalian tissues, triacylglycerols Preparation of Microsomes. All operations were performed at are synthesized from acyl CoA and glycerol-P via three different 0 to 4C. The tissues were chopped with a razor blade in grinding acyltransferases (1, 13, 16, 17). The first two acyltransferases medium (4 ml/g tissue) in a Petri dish, and then homogenized (glycerol-P acyltransferase and lysophosphatidic acid acyltrans- gently with a mortar and pestle. The grinding medium contained ferase) possess some acyl CoA preference, such that their speci- 1 mM EDTA, 0.6 M sucrose, 10 mM KCI, 1 mM MgCl2, 2 mM ficities as well as the in vivo pool sizes of acyl CoAs produce the DTT, and 0.15 M Tricine-KOH buffer (pH 7.5). The homogenate observed positional acyl specificity in the triacylglycerol (1, 17). was filtered through a Nitex cloth (20 x 20 Mm). The filtrate was The third acyltransferase, diacylglycerol acyltransferase (EC centrifuged at l0,OOOg for 15 min, and the supernatant was 2.3.1.20), is the only known enzyme unique to the long triacyl- recentrifuged at l00,OOOg for 100 min. The pellet was resus- glycerol biosynthetic pathway, since the diacylglycerol produced pended in a small volume ofgrinding medium. This microsomal from the first two acyltransferases could be used to produce preparation was either used immediately or frozen at -7O°C until phospholipids or galactolipids. Diacylglycerol acyltransferase is use. supposed (1, 17) to be relatively less specific on acyl CoAs, and Acyl CoAs. Nonradioactive lauroyl CoA, oleoyl CoA, erucoyl CoA, and CoA were obtained from Sigma Corp. Radioactive ' Supported by the National Science Foundation grant DMB 85- [1-'4C]oleoyl CoA was purchased from New England Nuclear 15556. Corp. Radioactive lauroyl CoA and erucoyl CoA were synthe- 762 DIACYLGLYCEROL ACYLTRANSFERASE IN SEEDS 763 sized chemically from CoA and [l--4C]lauric acid and [14-14CJ acid and 30% oleic acid (7). Brassica napus L. var Dwarf Essex erucic acid (Research Products International Corp., Mount Pros- contains about 40% erucic acid, 15% linoleic acid, and 15% pect, IL), respectively, using the acyl chloride method (2). oleic acid in the seed triacylglycerols (7, 15). Seeds of a special Assay of Diacylglycerol Acyltransferase Activity. The enzyme variety of Brassica napus, Tower, which had been bred to elim- activity was assayed by a procedure modified from those reported inate most of the erucic acid, contain about 60% oleic acid and elsewhere (3, 11, 17). The assay was performed at 24, 32, or 23% linoleic acid in the triacylglycerols (15). 40°C. The reaction mixture contained, in a final volume of 250 The microsomal fractions from the cotyledons of Cuphea and 1d, 40 mm Tris-HCI buffer (pH 7.0), 6 mM MgCl2, 4 mM DTT, Brassica, and from the scutella of maize of maturing seeds were 12 mm sucrose, acyl CoA (0.02-0.1 gCi, depending on the used as the source of diacylglycerol acyltransferase. The enzyme experiment) at a concentration indicated in the figures, 0.4 mm preparation was supplied with 1,2-diolein, and with radioactive 1,2-diolein in 0.04% Tween-20, and enzyme. Diolein (2 mM) lauroyl CoA, oleoyl CoA, and erucoyl CoA separately (testing was emulsified with 10 mg Tween-20 in 5 ml of water, and 50 'specificity') or in an equimolar mixture (testing 'selectivity') at Ml of the emulsified preparation were added to the reaction increasing concentrations. The enzymic formation ofradioactive mixture. The reaction was initiated by the successive addition of triacylglycerols was measured. The kinetics ofthe reaction cata- acyl CoA and enzyme. After 2 min the reaction was terminated lyzed by the maize enzyme had been described (3). by adding 1.5 ml heptane/isopropanol/0.5 M H2SO4 (10:40:1, v/ Acyl CoA Specificity of Diacylglycerol Acyltransferase. The v/v). After 30 min, 1 ml 0.1 M NaHCO3, and 1 ml heptane specificity of the enzymes on the three acyl CoAs supplied containing 50 nmol triolein (as carrier) were added, and the separately at increasing concentrations at 24C is shown in Figure mixture was shaken vigorously. The mixture was allowed to 1. The Cuphea enzyme was slightly more active on lauroyl CoA settle for 5 min. The upper phase was retained, and the lower than on oleoyl CoA, and was completely inactive on erucoyl phase was reextracted with 1 ml heptane in the same manner. CoA. The maize enzyme was about 30% more active on oleoyl The combined upper phase was washed with 1.5 ml 1 M NaCl, CoA than on lauroyl CoA, and was also totally inactive on and then evaporated to dryness under a stream ofN2. The residue erucoyl CoA. The two Brassica enzymes were equally active on was dissolved in heptane, and applied to a TLC plate (E. Merck, lauroyl CoA and oleoyl CoA, and were also active on erucoyl Silica Gel 60), and chromatographed in hexane/ether/acetic acid CoA, with activity about two-thirds ofthat on the other two acyl (50:50:1, v/v/v). The silica spot corresponding to standard tri- CoAs. olein was visualized with iodine and scraped. The silica gel was Acyl CoA Selectivity of Diacylglycerol Acyltransferase. The either counted directly for 14C in a scintillation counter, or selectivity of the enzymes on the three acyl CoAs present in subjected to further analyses of the content of different '4C fatty equimolar mixtures each at 1.63, 10, and 20 Mm was studied at acids in the triacylglycerols. In the latter case, the silica gel was 24C (Fig. 2). The Cuphea enzyme was slightly more active on incubated with 1 ml 5% KOH in 95% ethanol at 40°C for 1 h. lauroyl CoA than on oleoyl CoA, and again was completely After hydrolysis, 1 ml of water was added, and the solution was inactive on erucoyl CoA. A similar pattern of activities on the adjusted to pH 1 to 2 with concentrated HCI. Fatty acids were three acyl CoAs was exhibited by the maize enzyme. The two extracted with 1 ml petroleum ether twice. The resulting fatty Brassica enzymes were most active on lauroyl CoA, and their acid extract was dried under N2 and applied to a reverse phase activities on oleoyl CoA and erucoyl CoA were about 50% and TLC plate (Whatman KC-18). The plate was developed with 10 to 20%, respectively, ofthat on lauroyl CoA. acetonitrile/acetic acid/water (90:1:9, v/v/v). The three fatty When the temperature of the reaction was increased from 24 acids were separated from one another cleanly. They were visu- to 32C and 40°C, the total enzyme activities in general increased alized by autoradiography or after spraying with phosphomo- (see legend to Fig. 2). The increase in temperature altered the lybdic acid. The fatty acid spots were scraped into a scintillation acyl CoA selectivity ofthe four enzymes. The alterations may be vial and counted for 14C. due to differential changes in the affinities of the enzymes for Solubility of Acyl CoAs in the Assay Mixture. A recent paper the acyl CoAs as well as in the solubility ofthe acyl CoAs which reports that under certain conditions, Mg2e precipitates acyl CoA have their critical micelle concentration (CMC) at the ,molar from a reaction mixture and thus may interfere with acyltrans- range (14). These factors are difficult to pinpoint in such a ferase assays (4). We tested this possibility in our assay mixture heterogenous system involving insoluble enzymes and substrates. minus microsomes, using the reported procedure (4) and detect- At 32 and 40°C, Cuphea enzyme became increasingly more ing the radioactive acyl CoA in the supernatant and the pellet active on lauroyl CoA than on oleoyl CoA (Fig. 2). However, after centrifugation. At acyl CoA concentrations of 1.63 and 10 the maize enzyme became slightly more active on oleoyl CoA Mm (results at both concentrations were very similar), more than 90% of lauroyl CoA and oleoyl CoA, and 80% of erucoyl CoA, remained in the supernatant. Apparently, most ofthe acyl CoAs Cuphea Maize Rapeseed Canola were in a soluble condition in our assay. As a control, when 18: 18:12 diolein and Tween 20 were omitted from the reaction mixture -1 (a condition mimicking that in the report [4]), only 2 to 30% of 0 the acyl CoAs (depending on the concentration and acyl chain tv length) remained in the supernatant. 4) co Protein Assay. Protein was assayed by the Lowry method (10). 0 ir 22:1 RESULTS 0 10 20 0 10 20 Experimental Design. Seeds of four species were selected be- Acyl-CoA concentration ( pJM cause of their content of unique and diverse fatty acids in the FIG. 1. Specificity of diacylglycerol acyltransferase on acyl CoAs. storage triacylglycerols. Seed triacylglycerols of Cuphea carthag- Microsomal enzymes from C. carthagenensis (26 gg protein per assay, enensis contain 62% lauric acid, 13% myristic acid, and 8% 100% = 5.2 pmol/min), maize (I150 ug protein, 64.7 pmol/min), rapeseed capric acid (5). We selected this particular Cuphea species partly (450 gg protein, 54.2 pmol/min), and Canola (380 jAg protein, 71.2 because the wild plants were available locally. Our own gas pmol/min) were used. Lauroyl CoA (12: 0), oleoyl CoA (18: 1), and chromatographic analysis confirmed the above fatty acid com- erucoyl CoA (22: 1) were supplied separately. The assays were performed position. Maize kernel triacylglycerols possess about 50% linoleic at 24"C. 764 CAO AND HUANG Plant Physiol. Vol. 84, 1987

Cuphea Maize Rapeseed Canola ular preference for lauroyl CoA, and the maize enzyme shows 12 ~ * some preference for oleoyl CoA. It was shown earlier that the 12.0 enzyme from maize (having about 50% linoleic acid and 30% oleic acid in seed triacylglycerols) is twice as active on linoleoyl 24°C 50 CoA than on oleoyl CoA (3). The Brassica enzymes are more 118:18:1- active on erucoyl CoA than on oleoyl CoA at acyl CoA concen- 22:1 trations less than 1.36 AM and at 32 and 40°C. Overall, the acyl CoA preferences of the four enzymes are of physiological rele- vance, since they reflect the dominant acyl moieties in the 100 * respective seed triacylglycerols. An earlier study shows that Elm americana lipase from post- 12:0 germinative seeds, a seed species containing 61% of capric acid, 323C ' 50 is highly specific on tricaprin and has little activity on triolein _/22:1 (9). In view of the Cuphea seeds containing 62% lauric acid and only 5% each ofoleic acid and linoleic acid in the triacylglycerols CD (5), one might expect its diacylglycerol acyltransferase to be quite specific on lauroyl CoA. However, the Cuphea enzyme also 100 works quite well on oleoyl CoA. One possible explanation for

/ 12:0 181z .12:0 this relatively low specificity of the Cuphea enzyme is that the Cuphea species used, C. carthagenensis, is considered to be quite 40'C 5C primitive in the Cuphea genus (5). During the evolution of 18 1 18 1 Cuphea species, different species were evolved to have their fatty

22 acids in the storage triacylglycerols shortened from C- 18 to C- 10 22:1 and C-8. Even though C. carthagenensis possesses mostly C- 12 20 0 10 20 0 10 20 and C-10 fatty acids in its seed triacylglycerols, it has only a brief Acyl-CoA concentration ( pM ) evolutionary history. It originates from two parents both ofwhich FIG. 2. Selectivity of diacylglycerol acyltransferase (on acyl CoAs. possessed substantial C-18 in seed triacylglycerols. Presumably, Microsomal enzymes from C. carthagenensis (26 usg prc )tein per assay, the diacylglycerol acyltransferase of C. carthagenensis has not 100% = 8.0, 15.3, and 10.3 pmol/min at 24, 32, and 40'( respectively), had time to evolve to become more specific on lauroyl CoA. maize (150 ,ug protein, 104, 127, and 157 pmol/min), raipeseed (450 ,g Although the two Brassica napus varieties examined have very protein, 45.3, 83.8, 189 pmol/min), and Canola (380 ,g g protein, 72.1, diverse fatty acid compositions in their seed triacylglycerols, their 103.8, and 353 pmol/min) were used. Lauroyl CoA (12: 0), oleoyl CoA diacylglycerol acyltransferases possess virtually identical acyl (18: 1), and erucoyl CoA (22: 1) were supplied togethe:r in equimolar CoA preference. This is not surprising, since erucic acid-free mixtures. The assays were performed at 24, 32, and 40C Brassica varieties were obtained by breeding through the alter- ations of a few genes (presumably those coding for fatty acid ive on erucoyl than on lauroyl CoA. Both enzymes were still inacti y elongation enzymes), and thus their diacylglycerol acyltransferase CoA. remains unchanged from the enzymes ofthe parents. The increase in temperature from 24 to 32°C and 4O°C en- hanced the activity ofthe two Brassica enzymes on [erucoyl CoA There is an increasing interest in using both classical breeding to relative to lauroyl CoA or oleoyl CoA (Fig. 2). This 4enhancement and modern biotechnology improve the quality of seed oils may reflect a preferential improvement of the sol ubility of the (5, 12, 15). The outstanding goal is to change the carbon chain least soluble erucoyl CoA. Specifically, at 32 alnd 40°C, the length and the degree of unsaturation of the fatty acids in the Brassica enzymes had higher activities on erucoyl CoA than on triacylglycerols. From the known biosynthetic pathway of fatty oleoyl CoA at concentrations of 1.63 uM (Fig. 2), ).82 and 0.41 acid elongation and desaturation, it seems quite workable by jM (in a separate experiment, data not shown). using classical breeding or genetic engineering to selectively delete, or even add in, a certain gene for a particular enzyme so DISCUSSION that a desirable fatty acid can be synthesized. However, what is not known is whether the newly designed fatty acid can be In the current study of diacylglycerol acyltransferase, two accommodated by the other components of the triacylglycerol major factors are involved in determining acyl CcA preference synthesis machinery. Diacylglycerol acyltransferase is one of under the assay conditions. One factor is the inherited these components. As shown in the current study, the enzyme of the enzyme. The other factor is the availabilit offroperties from diverse species can easily accommodate physiological or which is influenced its concentration tY theCaMc CoA, by (whel nonphysiological fatty acids of shorter carbon chains, such as has been reached), the temperature, and the occurrrence of other lauric acid which is very important in industry. The great success competing acyl CoAs and amphipathic materials. TsTheonobservedenzyme of breeding erucic acid-free rapeseeds attests to the above state- effects of temperature and acyl CoA concentratiorin ment, since the Brassica enzyme works equally well on erucoyl selectivity are best explained in terms of changes the availa bility (solubility) ofthe acyl CoAs (maize and Bras~sica enzymes) CoA and oleoyl CoA. However, the enzyme from diverse species as well as in the enzyme molecule per se such as its relative acyl cannot accommodate fatty acids of a longer carbon chain such CoA affinities (Cuphea and maize enzymes). How(ever, it is also as erucic acid, unless it belongs to those seeds having a high possible that both factors contribute to the obsenved acyl CoA content of that longer fatty acid. Diacylglycerol acyltransferase preference of each of the four enzymes. in mammals is supported to be the least acyl specific enzyme After taking the above considerations into accsount, several among the three glycerol acyltransferases. Whether or not the trends can be depicted from the results. The enzy'mes from all other two seed acyltransferases are more acyl specific remains to four species are active on lauroyl CoA in the in vitr'o assays. This be seen. Notwithstanding, researchers in biotechnology should preference undoubtedly reflects in part the high solubility of keep these factors in mind when dealing with the modification lauroyl CoA. Nevertheless, the Cuphea enzyme shiows a partic- of fatty acids in seed oils. DIACYLGLYCEROL ACYLTRANSFERASE IN SEEDS 765 LITERATURE CITED 698 9. LIN YH, C Yu, AHC HUANG 1986 Substrate specificities of lipases from corn 1. BELL RB, RA COLEMAN 1983 Enzymes of triacylglycerol formation in mam- and other seeds. Arch Biochem Biophys 244: 346-356 mals. In PD Boyer, ed, The Enzymes, Vol 16, Lipid Enzymology. Academic 10. LOWRY OH, NJ ROSEBROUGH, AL FARR, RJ RANDALL 1951 Protein measure- Press, New York, pp 87-1 1 1 ment with folin phenol reagent. J Biol Chem 193: 265-275 2. BISHOP JE, AK HAJRA 1980 A method for the chemical synthesis of `4-C- 11. MARTIN BA, RF WILSON 1983 Properties ofdiacylglycerol acyltransferase from labeled fatty acyl coenzyme A's of high specific activity. Anal Biochem 106: spinach leaves. Lipids 18: 14 344-350 12. RATTRAY JBM 1984 Biotechnology and the fats and oils industry-an over- 3. CAO YZ, AHC HUANG 1986 Diacylglycerol acyltransferase in maturing oil view. J Am Oil Chem Soc 61: 1701-1712 seeds of maize and other species. Plant Physiol 82: 813-820 13. ROUGHAN PG, CR SLACK 1982 Cellular organization of glycerolipid metabo- 4. CONSTANTINIDES PP, JM STEIM 1986 Solubility of palmitoyl-Coenzyme A in lism. Annu Rev Plant Physiol 33: 97-132 acyltransferase assay buffers containing magnesium ions. Arch Biochem 14. SMrrH RH, GL POWELL 1986 The critical micelle concentration of some Biophys 250: 267-270 physiologically important fatty acyl-coenzyme A's as a function of chain 5. GRAHAM SA, F HIRSINGER, G ROBBELEN 1981 Fatty acids of Cuphea (Lyth- length. Arch Biochem Biophys 244: 357-360 raceae) seed lipids and their systematic significance. Am J Bot 68: 908-917 15. STEFANSSON BR 1983 The development of improved rapeseed cultivars. In 6. GRIFFITHs G, AK STOBART, S STYMNE 1985 The acylation of sn-glycerol-3- JKG Kramer, FD Sauer, WJ Pigden, eds, High and Low Erucic Acid Rapeseed Oils. Academic Press, Toronto, pp 143-159 phosphate and the metabolism ofphosphatidate in microsomal preparations 16. STYMNE S, AK STOBART 1986 Triacylglycerol biosynthesis. In PK Stumpf, EE from the developing cotyledons of safflower (Carthamus tinctorius L) seed. Conn, eds, The Biochemistry ofPlants, Vol 10, Lipids. Academic Press, New Biochem J 230: 379-388 York. In press 7. HILDITCH TP, PN WILLIAMS 1964 The Chemical Constitutions of Natural 17. YAMASHITA S, K HOSAKA, Y MIKI, S NUMA 1981 Glycerolipid acyltransferases Fats. John Wiley & Sons, New York from rat liver: l-acylglycerophosphate acyltransferase, I-acylglycerophos- 8. ICHIHARA K 1984 sn-Glycerol-3-phosphate acyltransferase in a particulate phoryl-choline acyltransferase, and diacylglycerol acyltransferase. Methods fraction from maturing safflower seeds. Arch Biochem Biophys 232: 685- Enzymol 71: 528-536