THE JOURNAL 0 1986 by The American Society of Biological Chemists,

OF BIOLOGICAL CHEMISTRY , Inc

Vol. 261, No. 1, Issue of January 5, pp. 194-199, 1986 Printed in U.S.A.

Cocoa Butter Biosynthesis PURIFICATION AND CHARACTERIZATION OF A SOLUBLE sn-GLYCEROL-3-PHOSPHATE ACYLTRANSFERASE FROM COCOA SEEDS*

(Received for publication, November 8, 1984)

Paul J. Fritz, John M. Kauffman, Charles A. Robertson, and Melinda R. Wilson From the Department of Phurmacobm, The Milton S. Hershey Medical Center, The Pennsylvania State University,

” . Hershey, Pennslvania 17033

Glycerol-3-phosphate acyltransferase has been pu- rified from the post-microsomal supernatant of cocoa seeds using differential ammonium sulfate solubility along with anion exchange and gel filtration chroma- tography. Chromatofocusing and isoelectric focusing revealed a series of proteins with acyltransferase ac- tivity having isoelectric points close to 5.2.

Gel filtration on Sephacryl S-300 in 500 mM NaCl, along with polyacrylamide gel electrophoresis (dena- turing and non-denaturing) and immunochemical anal- ysis, gave evidence that the native enzyme has a mo- lecular weight of 2 x 10’ and consists of an aggregate of 10 M. 20,000 subunits.

The highly purified enzyme carries an acyl donor, probably acyl-CoA, although this is not firmly estab- lished. The hydrophobic nature of the purified enzyme was demonstrated by its firm binding to octyl-Sepha- rose.

Mass spectrometric analysis of reaction products re- vealed the presence of both palmitic and stearic acids. Considering that 1) the fatty acids were derived from the purified enzyme; 2) they were found exclusively in the 1-position of glycerol 3-phosphate; 3) the fatty acid positioning and composition is consistent with that found in cocoa butter, the major storage product of cocoa seeds; and 4) the enzyme is found in the post- microsomal supernatant, it seems reasonable to con- clude that the first step in cocoa butter biosynthesis is catalyzed by glycerol-3-phosphate acyltransferase in the cytoplasm of cocoa cotyledon cells.

The unique physical properties of cocoa butter, the major storage product of cocoa seeds, stem from the fact that it is a mixture of triacylglycerols composed of 85% oleic acid in the 2-position, whereas the 1- and 3-positions are occupied by palmitic and stearic acids (1, 2). Lipids make up over 50% of the dry weight of mature cocoa seeds, whereas triacylglycerols make up about 96% of the lipids (3).

Plant triacylglycerol biosynthesis proceeds through the complete glycerol 3-phosphate pathway from l-monoacylglyc- erol 3-phosphate to 1,2-diacylglycerol 3-phosphate, to 1,2- diacylglycerol, to triacylglycerol (4, 5). This laboratory is engaged in a systematic study of the enzymes responsible for the biosynthesis of cocoa butter, and the present study is concerned with the purification and characterization from

* This work was supported in part by awards from the Hershey Foods Corporation. 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.

cocoa seeds of glycerol-P acyltransferase, the enzyme catalyz- ing the first step in the pathway. A preliminary report of this work has been published (6).

EXPERIMENTAL PROCEDURES’

RESULTS AND DISCUSSION

Preliminary Purification Although several studies concern the occurrence and bio-

synthesis of triacylglycerol in seeds, information on plant glycerol-3-phosphate acyltransferase is limited to reports on the enzyme from spinach and pea chloroplasts (7-9). In early experiments on the enzymes of cocoa butter biosynthesis, we discovered the majority of glycerophosphate acyltransferase’ activity in the post-microsomal supernatant fraction of cocoa seed extracts. The enzyme has now been extensively purified following very closely the procedures described by Bertrams and Heinz (8). Summary of a typical purification is given in Table I. In some preparations, where initial specific activities were low due to uncertainties of protein measurements caused by the presence of large amounts of pigments, enrichments exceeding 1000-fold were achieved. The specific activities of our pure preparations are strikingly similar to those reported for the pea chloroplast enzyme (8). The enzyme can be pre- pared either from fresh beans, or from beans kept at -80 “C for several months, though yields and specific activities are lower for the latter. Similar results were obtained using ma- ture (180 days after pollination) or immature (120 days after pollination) seeds.

Soluble plant glycerol-3-phosphate acyltransferases have been found in avocado mesocarp (4), in spinach chloroplasts (7, 8), and in pea chloroplasts (8).

Chromatofocusing and Isoelectric Focusing

Following preliminary purification, the enzyme preparation was passed through a chromatofocusing column. The results of this experiment are shown in Fig. 1. It is seen that a small amount of glycerophosphate acyltransferase activity elutes at an isoelectric point of 6.2 but the majority of the enzyme

Portions of this paper (including “Experimental Procedures” and Figs. 1s-5s) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Doc- ument No. 84M-3424, cite the authors, and include a check or money order for $3.60 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.

The abbreviations used are: glycerophosphate acyltransferase, glycerol-3-phosphate acyltransferase; ACP, acyl carrier protein.

194

Cocoa Soluble Glycerol-3-phosphate Acyltransferase 195 TABLE I

Purification of glycerol-P acyltransferase from 350 g of cocoa beans

Step Total activity

X total volume nmollrninlml

Total protein Yield Specific activity

nrnol/rnin/mg protein

Post-microsomal supernatant 19,200 16.6 g 1.2 Ammonium sulfate 25-50% saturated 12,000 3.7 g 63 3.3 2.8 DEAE step gradient 7,700 315 mg 40 24.5 20 DEAE linear gradient 1,708 14.4 mg 8.9 119 99 Sephacryl S-300 1 1,450 3.2 mg 7.6 453 378 Sephacryl S-300 2 1,245 1.8 mg 6.5 692 577

Enrichment

%

450 2.4 - 5.8

I 375

0 k- 5.4 I 2.0 3- 5.6 2 e:

2300 0 7

c 225 1.2 - 5.2 z ' 150

I-

0.8 - 5.0

75 0.4 - 4.8

0 n n - - d ~

20 40 60 80 IC4 120 140 160 FRACTION

-.- -.-

FIG. 1. Chromatofocusing chromatography of cocoa seed glycerophosphate acyltransferase. A 1 X 45-cm column was packed with Pharmacia polybuffer exchanger (PBE 94) pH range 6- 4, and equilibrated with 0.025 M histidine-HC1, pH 6.2. Post-S-300 glycerophosphate acyltransferase was equilibrated with histidine buffer by passage over a Sephadex G-50 column then applied to the column. The column was eluted with Pharmacia Polybuffer 74, pH 4.0. 1-ml fractions were collected.

elutes in two peaks at isoelectric points 5.2 and 5.3. Two smaller peaks are seen at isoelectric points 4.9 and 4.8.

Isoelectric focusing using vertical tube polyacrylamide gels confirmed the results of the chromatofocusing experiment (Fig. 2). Thus, we conclude that cocoa seed glycerophosphate acyltransferase exists in several molecular forms distinguish- able by small charge differences.

Isoelectric focusing of purified pea chloroplast acyltransfer- ase activity demonstrated the existence of two forms of the enzyme with apparent isoelectric points of 6.3 and 6.6 (8). The two forms co-purified through all other purification steps. In contrast, spinach chloroplast acyltransferase was found to exist in a single molecular form with isoelectric point 5.2 (8).

Hydrophobic Chromatography

Detection of enzyme activity in all preliminary purification steps, as well as in the chromatofocusing and isoelectric focusing experiments, was achieved without addition of a fatty acid donor to reaction mixtures. In addition, it was determined that, even with highly purified glycerophosphate acyltransfer- ase preparations, addition of either acyl-CoA or acyl-ACP had no effect on enzyme activity. We reasoned that if the enzyme was the source of the fatty acid donor, it should be possible to demonstrate this by hydrophobic chromatography. Post-S- 300 glycerophosphate acyltransferase was subjected to hydro- phobic chromatography on an octyl-Sepharose column as described by Rock and Garwin (11). The procedure resulted

ISOELECTRIC FOCUSING

1- TOP (

1400

1200

1000

1 800

600

400

200

8

6

4 pH

2

2 4 6 8 10 12 14 Centimeters from Bottom of Gel

FIG. 2. Isoelectric focusing of cocoa seed glycerophosphate acyltransferase. Vertical tube polyacrylamide gel electrophoresis was performed as described by O'Farrel(l0). The ratio of ampholytes was 4:l (pH 5-7: pH 3-10). Thirty pl (51 pg) of Sephacryl S-300 purified enzyme (specific activity, 0.69 pmollminlmg of protein) were used. After electrophoresis, one gel was sliced, the segments soaked in 0.5 ml of H20 for 1 h to remove urea and for pH determination, then placed in 50 pl of enzyme reaction mixture, incubated for 90 min at 25 "C, mixed with 100 pl of glycerol-P saturated n-butyl alcohol, and counted in a liquid scintillation counter. A second gel was stained with Coomassie Brilliant Blue G-250 (15 min in 3.5% HCIO,, 20% methanol; 15 min in 0.08% Coomassie Brilliant Blue G- 250 in HC104-methanol; destained in 5% acetic acid, 20% methanol).

in considerable loss of enzyme activity but the major recover- able activity was found in the fraction (peak C) eluted from the column with 25% isopropanol (Fig. 3). Peak C was 10 times more active than peak B and, in common with all prior samples in the purification, did not require addition of an exogenous fatty acid donor to the reaction mixture. That is to say, the enzyme preparation itself was the source of the fatty acid transferred to glycerol 3-phosphate. Protein present in peak A , on the other hand, had essentially no enzyme activity when palmitoyl-CoA was not added to the reaction mixture. However, a small amount of activity was seen when palmitoyl-CoA was added.

The hydrophobic chromatography experiment clearly dem- onstrates that long chain fatty acids are associated with the

8 16

Fraction number

24

FIG. 3. Octyl-Sepharose chromatography of purified cocoa seed glycerophosphate acyltransferase. Purified enzyme from the S-300 Sephacryl column was applied to a 1.5 X 5-cm octyl- Sepharose column that had been equilibrated with 20 mM Tris-HC1, pH 7.4. The column was eluted with the same buffer until the 280 nm absorbance was zero, then with 25% (w/v) 2-propanol. Fraction volumes were 0.7 ml. Peaks A, B, and C were combined separately and assayed for glycerol-P acyltransferase activity.

purified enzyme. Rock and Garwin (11) used model com- pounds to show that ACP does not bind to octyl-Sepharose, whereas acyl-ACP binds strongly but can be eluted with solutions of 2-propanol. They showed that acyl-ACP binds to the hydrophobic gel via the acyl chain, strength of binding being proportional to fatty acid chain length. Although our experiment strongly suggests that purified cocoa glycerophos- phate acyltransfera~ has long chain fatty acids associated, it does not prove that the fatty acid derives from acyl-ACP since it is known that acyl-CoA binds strongly to some proteins (12). The soluble glycerophosphate acyltransferase from pea and spinach chloroplasts could use either acyl-CoA or acyl- ACP as a fatty acid donor with preference for the latter (9). It is possible that both potential fatty acid donors are present in the purified cocoa seed glycerophosphate acyltransferase preparations.

Presence of CoA in purified glycerophosphate acyltransfer- ase preparations is suggested from the results of experiments where post-S-300 glycerophosphate acyltransferase was in- cubated with [“Cjpalmitate and Aerobacter aerogenes acyl- CoA synthetase. Polyacrylamide gel electrophoresis of reac- tion mixtures followed by gel drying and fluorography re- vealed a band in the low molecular weight region where acyl- CoA migrates. The results of this experiment imply that free CoA as well as acyl-CoA is associated with the purified en- zyme.

Kinetic Data Typical Michaelis-Menten kinetics were obtained when a

glycerol 3-phosphate concentration curve was run in the ab- sence of added fatty acid donors. A K,,, of 4.2 mM was found, a value similar to that reported for the spinach chloroplast enzyme (8). Reaction velocity was linear for 10 min, then slowed, and finally stopped after 2-3 h.

Purified soluble cocoa seed glycerophosphate acyltransfer- ase could be made responsive to added acyl-CoA by first incubating the enzyme with glycerol 3-phosphate, then pass- ing the incubation mixture over a Sephadex G-50 column. The enzyme recovered after this treatment gives a typical Michaelis-Menten substrate concentration curve with a K , for stearoyl-CoA of about 4 p ~ , again similar to values for the spinach and pea chloroplast enzymes.

Analysis of Reaction Products We conducted experiments to determine the chemical na-

ture of the products formed when glycerol 3-phosphate was incubated with cocoa seed extracts. Peak C from the octyl- Sepharose column (Fig. 3) was the enzyme source used for these experiments. Detailed protocols for the analyses are given under “Experimental Procedures.”

First, it was determined by thin layer chromatography that a single fatty acid was transferred to glycerol 3-phosphate forming monoacylgIycero1 3-phosphate. An important prop- erty of acyltransferase is the positional specificity. To deter- mine whether the fatty acid transferred was in position 1 or 2 of glycerol 3-phosphate, the reaction product, monoacylglyc- erol 3-phosphate, was isolated and dephosphorylated by al- kaline phosphatase, and the resulting monoacylglycerol ana- lyzed for isomeric composition by thin layer chromatography. The results clearly established that acylation occurred exclu- sively in the 1-position.

The pea and spinach chloroplast glycerophosphate acyl- transferase also possess high positional specificity, both di- recting acyl groups into the C-1 position of glycerol with negligible acylation of C-2 (8).

Finally, the fatty acids transferred were shown by mass spectrometric analysis to be palmitic and stearic. All these observations are consistent with the conclusion that the en- zyme purified from the soluble fraction of cocoa seed extracts is the first enzyme in the biosynthetic pathway leading from glycerol 3-phosphate to cocoa butter.

A n t i ~ ~ s to Cocoa Seed ~ l y c e r ~ p ~ ~ p ~ t e A ~ ~ t r a n s f e r ~ e R ~ w i r n r n ~ ~ ~ s a ~ to Demonstrate Presence of Antibod-

ies-Antibodies to cocoa seed glycerophosphate acyltransfer- ase were induced in a New Zealand White rabbit by injecting isoelectric focusing purified enzyme homogenized with poly- acrylamide gel slices (see Fig. 2). Detailed protocols describing the immunological procedures are given under “Experimental

25

20

?! 15

10

5

0 I II 111 10 loo l o 0 0

fig SERUM PROTEIN

FIG. 4. Rabbit-anti-cocoa seed glycerophosphate acyltrans- ferase. A, solid phase radioimmunoassay is shown. Details of the solid phase radioimmunoassay are given under “Experimental Pro- cedures.” Ordinate represents amount of ‘261-labeled protein A bound to antigen-antibody complex in microtiter wells (I1 and 111) compared to amount bound when preimmune serum was added (I). B, specificity. Using 500-p1 Eppendorf tubes, 0.25 pl of st-S-300 2 (Table I) glycerophosphate acyltransferase was mixed with 13.2 ~1 of antiserum or preimmune serum dilutions containing 11.4, 114, and 1140 pg of protein. The mixtures were incubated at room temperature for 30 min, then overnight at 4 “C. The samples were centrifuged for 15 min in a microfuge and the supernatants were assayed for glycerophos- phate acyltransferase activity.

Cocoa Soluble Glycerol-3-phosphate Acyltransferase 197

MW Stds. A E A E

61

43.

20-

14

FIG. 5. Molecular weight determination of cocoa seed glyc- erophosphate acyltransferase: Sodium dodecyl sulfate-poly- acrylamide gel electrophoresis. Purified cocoa seed glycerophos- phate acyltransferase after octyl-Sepharose chromatography ( l a n e A, 100 pg) or after Sephacryl-S-300 chromatography ( l o n e E, 10 pg) was denatured and reduced by heating for 3 min in 1.5-10-fold excess of loading buffer (63 mM Tris-HCI, pH 6.8; 12% glycerol; 1% sodium dodecyl sulfate, 180 mM 2-mercaptoethanol; 0.01% bromphenol blue). Standards of known molecular weight were treated in the same way. Electrophoresis was conducted in 15% polyacrylamide separating gel, using 6% stacking gels in the presence of 1% sodium dodecyl sulfate in a Tris-HCI buffer system (14). Duplicate gels were run; molecular weight standards and lones A and E on the left were stained with Coomassie Brilliant Blue, destained, and photographed. Lanes A and E on the right show the results of a Western blot to unmodified nitrocellulose followed by radiographic detection with anti-glycero- phosphate acyltransferase and '251-labeled protein A.

Procedures." Presence of antibodies in rabbit serum was demonstrated by solid-phase radioimmunoassay (13). The amount of '9-protein A bound by purifie: glycerophosphate acyltransferase antigen-antibody complex was about 9 times greater when antiserum was added to the microtiter wells compared to the amount bound when preimmune serum was added (Fig. 4A).

Antibody Inhibition of Glycerophosphate Acyltransferase Ac- tivity to Demonstrate Specificity-Ability of the antibody to inhibit glycerophosphate acyltransferase activity was dem- onstrated by conducting the usual enzyme assay after incu- bation with varying amounts of antiserum or control (preim- mune) serum. Fig. 4B shows the decreasing amount of enzyme activity observed after incubation with increasing amounts of antiserum protein.

Molecular Weight and Subunit Structure The molecular weight of nondenatured glycerophosphate

acyltransferase as determined by gel-filtration chromatogra- phy on Sephacryl S-300 columns in 500 mM NaCl is estimated to be about 200,000. Enzyme assay of gel slices obtained after

polyacrylamide gel electrophoresis under nondenaturing con- ditions confirmed this estimate. However, when gel electro- phoresis was performed in the presence of sodium dodecyl sulfate, the vast majority of protein, estimated by Coomassie Blue staining, representing all proteins present except for minor contaminants, was concentrated in a single area of the gel with electrophoretic mobility corresponding to a molecular weight of about 20,000. Antibodies to cocoa seed glycerophos- phate acyltransferase formed immune complexes only with this polypeptide (Fig. 5 ) , confirming that the M , 20,000 pro- tein was an enzyme subunit. Hence, we conclude that the native molecular weight of cocoa seed glycerophosphate acyl- transferase is 2 X lo5 and that the enzyme is composed of 10 M, 20,000 peptides. Both molecular forms of pea chloroplast glycerophosphate actyltransferase had molecular weights of 42,000, whereas the spinach enzyme was slightly larger (8). No subunit determinations were reported for these enzymes.

Stoichiometry Lamb and Fallon (12) reported that 1 mg of rat microsomal

protein could bind about 68 nmol of palmitoyl-CoA, all of which could be removed by incubation with high concentra- tions of albumin. If we assume an average molecular weight of 100,000 for the rat microsomal protein, it can be calculated that each nanomole of protein binds 6.8 nmol of palmitoyl- CoA. Incubation of highly purified cocoa seed glycerophos- phate acyltransferase under assay conditions for up to 3 h, long enough to allow transfer of all bound fatty acid, resulted in formation of 100 nmol of product for each nanomole of enzyme, or 10 nmol of bound fatty acid donor/subunit. The amount of product formed was a linear function of protein concentration.

REFERENCES 1. Brockerhoff, H., and Yurkowski, M. (1966) J. Lipid Res. 7.62-

2. Mattson, F. H., and Volpenheim, R. A. (1961) J. BioL Chem.

3. Lehrian, D. W., and Keeney, P. G. (1980) J. Am. Oil Chem. Soc.

64

236,1891-1893

57.61-65 4. Barron, E. J., and Stumpf, P. K. (1962) Biochim. Bwphys. Acta

60.329-337 5. Gurr,. M. I., Blades, J., Appleby, R. S., Smith, C. G., Robinson,

M. P., and Nichols, B. W. (1974) Eur. J . Biochem. 4 3 , 281- 290

6. Fritz, P. J., Kauffman, J. M., Robertson, C. A., and Wilson, M. R. (1984) Biochemistry 23,3374

7. Joyard, J., and Douce, R. (1977) Biochim. B+.ohys. Acta 486,

8. Bertrams, N., and Heinz, E. (1981) Plant Physwl. 68,653-657 9. Frentzen, M., Heinz, E., McKeon, T. A., and Stumpf, P. K. (1983)

273-285

Eur. J. Biochem. 129,629-636 10. OFarrell, P. H. (1975) J. Biol. Chem. 250,4007-4021 11. Rock, C. O., and Ganvin, J. L. (1979) J. Bwl. Chem. 254,7123-

7128 12. Lamb, R. G., and Fallon, H. J. (1972) J. Biol. Chem. 2 4 7 , 1281-

1287 13. Stetler, D. A., Rose, K. M., Wenger, M. R., Berlin, C. M., and

Jacob, S. T. (1982) Proc. Natl. Acad. Sci. U. S. A. 7 9 , 7499- 7503

14. Laemmli, U. K. (1970) Nature 227,680-685 15. Monroy, G., Kelker, H. C., and Pullman, M. E. (1973) J. Biol.

16. Schleif, R. F., and Wensink, P. C. (1981) Practical Methods in Chem. 248,2845-2852

Moleculnr Biology, p. 74, Springer-Verlag, New York

Continued on next page.

198 Cocoa Soluble Glycerol-3-phosphate Acyltransferase

Cocoa Soluble Glycerol-3-phosphate Acyltransferase 199

""""""""""""" SOLVENT FRONT

PHOSPHATIDIC ACID

LYSOPHOSPHATIDIC ACID