Cyclization of farnesyl pyrophosphate to the sesquiterpene olefins humulene and caryophyllene by an enzyme system from sage (Salvia officinalis)

ARCHIVES

OF BIOCHEMISTRY

AND

BIOPHYSICS

Vol. 233, No. 2, September, pp. 838-841, 1984

COMMUNICATION Cyclization of Farnesyl Pyrophosphate to the Sesquiterpene Olefins Humulene and Caryophyllene by an Enzyme System from Sage (Salvia o~icinalis)’

RODNEY

CROTEAU’

AND AFAF

GUNDY

Institute of Biological Chemistry, and Biochxmistry/Biophyic.s Program, Washington State University, Pullman, Washington 99164-6.940 Received April 30, 1984

A soluble enzyme preparation obtained from sage (Salvia oficinalis) leaves was shown to catalyze the divalent metal-ion dependent cyclization of trams, trans-farnesyl pyrophosphate to the macrocyclic sesquiterpene olefins humulene and caryophyllene. The identities of the biosynthetic products were confirmed by radiochromatographic analysis and by preparation of crystalline derivatives, and the specificity of labeling in the cyclization reaction was established by chemical degradation of the olefins derived enzymatically from [l-3H2]farnesyl pyrophosphate. These results constitute the first report on the cyclization of farnesyl pyrophosphate to humulene and caryophyllene, two of the most common sesquiterpenes in nature, and the first description of a soluble sesquiterpene cyclase to be isolated from leaves of a higher plant.

Most of the nearly 200 skeletal families of sesquiterpenes are thought to arise via the cyclization of farnesyl pyrophosphate (l), yet very few cell-free systems have been described which carry out these key cyclization reactions. Preparations from the fungi Trichothecium r&urn and Clitocybe illudens convert farnesyl pyrophosphate to trichodiene (2) and illudins (3), respectively, while an extract from Streptomgces transforms farnesyl pyrophosphate to pentalenene (4). Similar cell-free systems from Kadsura japonica seeds (5) and Andrographis paniculata tissue cultures (6) convert the acyclic precursor to the respective olefins, germacrene-C and y-bisabolene. These systems have been employed primarily to examine mechanistic features of the cyclizations, but almost no information is available concerning the properties of the cyclases involved.

‘This is Scientific Paper No. 6820, Project 0268, College of Agriculture Research Center, Washington State University, Pullman, Wash. 99164. This investigation was supported in part by National Science Foundation Grant PCM 82-04391, and by a grant from the Haarmann and Reimer Corporation. *Author to whom inquiries should be made. 0003-9861/84 $3.00 Copyright All rights

0 1984 by Academic Press. Inc. of reproduction in any form reserved.

Salvia o&inalis (common sage) produces an essential oil comprised primarily of monoterpenes, but which also contains humulene and caryophyllene as major sesquiterpene components (about 4% each in a commercial oil sample (7)). This species has been employed successfully as a source of several monoterpene cyclases (8), and the enzyme isolation and assay techniques previously utilized for monoterpene cyclases (9) were therefore adapted to search for the relevant sesquiterpene cyclases. In this communication we describe a soluble enzyme preparation from sage leaves which catalyzes the cyclization of tram, trans-farnesyl pyrophosphate to both humulene and caryophyllene (Fig. l), thus making possible direct investigation of the cyclase(s) responsible for generating two relatively simple and very widespread sesquiterpene types. A sample of the steam-distilled oil obtained from immature leaves of S. omnalis plants raised under greenhouse conditions (10) was analyzed by combined GLC-MS, and was shown to contain 4.1% humulene and 2.2% caryophyllene, with minor quantities of other sesquiterpene olefins, thereby confirming the presence of the products of interest [reference samples of all trans-humulene (>98%) and (-)-caryophyllene (purified to >98% by CC on silicic acid with hexane) were obtained from Fluka AG and PCR Research Chem838

FARNESYL

PYROPHOSPHATE

m

FIG. 1. Cyclization (III), and degradation acid (V). The asterisk

of [1-‘Hlfarnesyl pyrophosphate (I) to humulene of the products to 2,2-dimethyl succinic acid (IV) denotes the position of tritium.

icals, respectively]. A 105,OOOg supernatant used as the enzyme source was prepared from such tissue by an extraction method previously employed for the isolation of monoterpene cyclases (9). Thus, 5 g of immature leaves was homogenized with an equal weight of insoluble polyvinylpolypyrrolidone (Polyclar AT) in 10 ml cold 0.1 M sodium phosphate buffer, pH 6.5, containing 150 mM sucrose, 10 mM Na2S205, 10 mM sodium ascorhate, and 5 mM dithioerythritol, and the homogenate was then slurried with 5 g polystyrene resin (XAD-4). Following centrifugation at 2’7,OOOg and 105,OOOg, the soluble enzyme preparation was dialyzed overnight against cold 20 mM Mes (4-morpholineethanesulfonic acid), 5 mM sodium phosphate buffer, pH 6.2, containing 10% (v/v) glycerol and 1 mM dithioerythritol. Aliquots (1 ml, equivalent to 0.5 g fresh wt of the original tissue) were incubated in screw-capped vials in the presence of 15 mM MgClz and 30 PM [l-3H]farnesyl pyrophosphate. The [l3Hlfarnesyl pyrophosphate (84.8 Ci/mol) was prepared by the method of Dixit and co-workers (II), with final purification by chromatography on O-diethylaminoethyl cellulose [40 mM (NH&COJ and verification of the product in the usual manner (10). Following incubation at 30°C for 1 h, each tube was chilled in ice, and the radioactive pentane-soluble products were extracted (2 X 1 ml) and partitioned on a silicic acid column (9) into a sesquiterpene hydrocarbon fraction eluted by pentane (typically 3-6% conversion of suhstrate) and an oxygenated sesquiterpene fraction eluted with ether (typically 20-30% conversion of substrate). The polar, oxygenated sesquiterpene fraction was subsequently demonstrated by radio-TLC and radio-GLC to be comprised largely of farnesol (>95%), released from the substrate by endogenous phosphatases-pyrophosphatases (12). In the presence of 1 mM concentrations of the phosphatase inhibitor ammonium vanadate(V), loss of substrate to competing hydrolytic enzymes could be reduced by half with concomitant, hut minor, increase in hydrocarbon formation. Small amounts of nerolidol, the tertiary isomer of farnesol, were also found in the oxygenated sesquiterpene fraction, and were presumed to he gen-

839

CYCLIZATION

P

(II) and (-)-caryophyllene and trans-norcaryophyllenic

erated by nonenzymatic solvolysis of the substrate under the conditions of the assay (13). The possibility that nerolidyl pyrophosphate might function as an intermediate or alternate substrate in the cyclization (1) has not yet been fully addressed; however, a search of the water-soluble reaction products remaining after enzyme incubation (by phosphatase hydrolysis and radiochromatographic analysis) revealed only traces of this product. Radio-GLC of the hydrocarbon fraction indicated the presence of two labeled components, in a ratio of about one to two, coincident with authentic coinjected standards of caryophyllene and humulene, respectively (Fig. 2). TLC of the hydrocarbon fraction on silica gel-8% AgNOs [with hexane:ether:benzene (40:10:1, v/v/v)] also afforded two radioactive components, one comigrating with carrier humulene (R, = 0.15) and the other with authentic caryophyllene (R, = 0.31). Elution of the individual components from the gel with ether, followed by radio-GLC analysis, showed the [sH]caryophyllene and [3H]humulene to he chemically and radiochemically pure. Boiled enzyme controls did not produce significant levels of sesquiterpene olefins, although nerolidol and farnesol were detected in these incubation mixtures, and were presumed to arise by nonenzymatic solvolysis of the substrate. In the absence of a divalent cation, olefin synthesis from [l-3Hlfarnesyl pyrophosphate was negligible, and of the several cations examined (Mn*+, Co’+, Ni*+, Zn’+, and Mgz+ as the chlorides) Mgz+ at the 15 mM level was most effective in promoting enzymatic cyclization of farnesyl pyrophosphate to caryophyllene and humulene. A divalent cation is a requirement for all known monoterpene and sesquiterpene cyclases (9), and in the present instance the concentration and type of metal ion had no appreciable influence on the distribution of olefins produced (i.e., the ratio of caryophyllene to humulene remained -1:2 throughout). Particulate fractions prepared from the leaf homogenate (soluble polyvinylpolypyrrolidone was substituted for the insoluble polymer in the extraction buffer, and the preparation was centrifuged at SOOOg

CROTEAU

840

TIME

AND GUNDY

(min.)

FIG. 2. Radio-gas-liquid chromatogram of the hydrocarbon fraction obtained by incubating the soluble enzyme preparation from sage leaves with [laHlfarnesyl pyrophosphate in the presence of 15 rnhf MgC12. The upper tracing is the response of the radioactivity monitor attached to the gas-liquid chromatograph. The lower smooth tracing is the thermal conductivity detector response obtained from coinjetted standards of caryophyllene (a) and humulene (b). A 12 ft X l/8 in. column coated with 15% Carbowax was employed, and was operated at 1’75°C. Further details of the procedure can be found elsewhere (8).

followed by collection of the 2’7,OOOgand 105,OOOgpellets) were washed with the assay buffer and assayed at the same dilution under the identical conditions employed for the dialyzed supernatant fraction. Negligible amounts of sesquiterpene olefms were produced in these assays. Although sesquiterpene biosynthesis in pine needles is reported to be microsomal(14), the present cyclase(s) is operationally soluble, as are most other monoterpene and sesquiterpene cyclases thus far examined (9). In order to confirm the identities of the sesquiterpene olefins produced by the soluble enzyme system from sage, the hydrocarbon fraction isolated from several large-scale incubations with [l-aHlfarnesyl pyrophosphate was, following addition of 25 mg each of carrier, separated by argentation TLC into humulene and caryophyllene. The [SHJhumulene (0.38 &i) was diluted with additional carrier (4 mmol), and a portion (2 mmol) was converted to the AgNO, adduct by standard procedures (15) and crystallized to constant specific activity from 95% ethanol [94.5 & 2.8 &i/mol, mp 175-1’76°C (dec., cap. tube); Lit. 175-175.5”C (dec.) (15)]. [‘HlCaryophyllene (0.21 &i), following similar dilution with 4 mmol carrier, was converted to the blue @-nitrosite (16) and crystallized from 95% ethanol to a constant specific activity of

55.5 -+ 1.5 pCi/mol [mp 113-114°C; Lit. 115°C (IS)], thus confirming the identity of this product. Since the caryophyllene used as carrier was predominantly the (-)-isomer [[ag - 8.2’ (C = 4.4, EtOH); Lit. values for [a]D of -8” to -14.02” have been reported for “pure” (-)-caryophyllene (17)], the observed cocrystallization of radioactivity would suggest the labeled biosynthetic product to be this optical isomer. To locate the position of label in the biosynthetic products derived from [l-SHJfarnesyl pyrophosphate, the olefins were degraded by classical procedures (18) we have employed previously (19). Ozonolysis of the remaining [‘Hlhumulene (2 mmol) afforded labeled 2,2-dimethyl succinic acid (34% yield; see Fig. 1). which was purified by TLC [silica gel G with ether:hexane:acetic acid (65:35:2, v/v/v)] and crystallized from 2 N HCl [mp 140-141°C; Lit. 139-140’ (ZO)]. Treatment with excess 5% sodium methoxide in methanol under reflux exchanged essentially all of the tritium present in the dicarboxylic acid (sp act < 1 &i/mol), confirming the position of the original label at C-l of humulene as predicted (Fig. 1). A portion of the previously purified [‘HJcaryophyllene (2 mmol) was also ozonized, and the crude reaction products were further oxidized with HNOa to afford trans-norcaryophyllenic acid (16% yield; see Fig. l), purified as before by TLC and crystallization from water [mp 126’C, Lit. 125-127°C (21, 22)]. Although the HNOs oxidation step might have been expected to remove exchangeable a-hydrogens, the norcaryophyllenic acid was nevertheless subjected to the above treatment with sodium methoxide/methanol affording, in this instance, purified product of essentially unchanged specific activity (55.9 * 1.8 &i/mol). Thus, the tritium label in norcaryophyllenic acid was inaccessible to exchange, indicating the position within the cyclobutane ring as predicted (Fig. 1). These experiments conclusively demonstrate the enzymatic cyclization of farnesyl pyrophosphate to humulene and caryophyllene in a manner consistent with the scheme illustrated in Fig. 1, in which intramolecular electrophilic attack by the carbon (Cl) bearing the pyrophosphate ester on C-11 of the terminal double bond generates the macrocyclic cation, which is stabilized to form all trots-humulene by deprotonation at C-9 or undergoes secondary closure to the Irons-fused, four-membered ring with deprotonation of the C-3 methyl to yield (-)-caryophyllene. Humulene and caryophyllene are among the most common sesquiterpenes in nature, and the cyclization to humulene would appear to represent one of the simplest of all possible cyclization reactions of farnesyl pyrophosphate. The cyclization to humulene is of additional significance since humulene is considered to be the precursor (by reprotonation of the olefm) of a number of more complex polycyclic sesquiterpenes, including the illudoids, hirsutanoids, and pentalenoids (1).

FARNESYL

PYROPHOSPHATE

The present results constitute the first report on the cyclization of farnesyl pyrophosphate to humulene and caryophyllene in a cell-free system, and the first such soluble cyclase to be isolated from leaves of a higher plant. In spite of the fact that the sesquiterpenes comprise but a small fraction of the oil produced by sage (relative to monoterpenes), the sesquiterpene cyclase(s) is remarkably active on a tissue weight basis compared to the relevant monoterpene cyclases (8,9). More detailed studies are underway to determine the basis for this surprising observation, and to examine the number and nature of sesquiterpene cyclases present in the cell-free preparation. ACKNOWLEDGMENTS

We thank S. Combelic for raising the plants, D. Thornton for typing the manuscript, and M. Felton for technical assistance in the early phases of the research. REFERENCES

1. CANE, D. E. (1981) in Biosynthesis of Isoprenoid Compounds (Porter, J. W., and Spurgeon, S. L., eds.), Vol. 1, pp. m-374, Wiley, New York. 2. CANE, D. E., SWANSON, S., AND MURTHY, P. P. N. (1981) J. Amer. Chem Sot 103, 2136-2138. 3. PRICE, M., AND HEINSTEIN, P. (1978) Uoydia 41, 574-577. 4.

CANE, D. E., AND TILLMAN, A. M. (1983) J. Amer. Chem

Sot. 105. 122-124.

5. MORIKAWA, K., HIROSE, Y., AND NOZOE, S. (1971) Tetrahedron Z&t, 1131-1132. 6. ANASTASIS, P., FREER, I., GILMORE, C., MACKIE, H., OVERTON, K., AND SWANSON, S. (1982) J. Chem. Sot. (Chem. Commun), 268-270.

CYCLIZATION

841

7. LAWRENCE, B. M., Hocc, J. W., AND TERHUNE, S. J. (1971) Parf Cosm Suv. France 1,256-259. 8. CROTEAU, R. (1981) in Biosynthesis of Isoprenoid Compounds (Porter, J. W., and Spurgeon, S. L., eds.), Vol. 1, pp. 22.5282, Wiley, New York. 9. CROTEAU, R., AND CANE, D. E. (198x) in Methods in Enzymology (Law, J. H., and Rilling, H. C., eds.), Vol. 110. Academic Press, New York. 10. CROTEAU,R., AND KARP, F. (1976) Arch. Biochem Biophys. 176, 734-746. 11. DIXIT, V. M., LASKOVICS, F. M., NOALL, W. I., AND POULTER, C. D. (1981) J. Org. Chem. 46,19671969. 12. CROTEAU, R., AND KARP, F. (1979) Arch Biochem. Biophys. 198. 523-532. 13. GEORGE-NASCIMENTO, C., PONT-LEZICA, R., AND CORI,0. (1971) Biochem Biophys Res. Cummun 45,119-124. 14. BERNARD-DAGAN, C., PAULY, G., MARPEAU, A., GLEIZES, M., CARDE, J.-P., AND BARADAT, P. (1982) PhysioL V&g. 20, 775-795. 15. HILDEBRAND, R. P., AND SUTHERLAND,M. D. (1961) Au& J. Chem 14, 272-275. 16. DEUSSEN, E., AND LEWINSOHN, A. (1907) Liebigs. Ann Chem 356, l-23. 17. DAMODARAN, N. P., AND DEV, S. (1968) Tetrahedron, 4113-4122. 18. BRYANT, R. (1969) in Rodd’s Chemistry of Carbon Compounds (Coffey, S., ed.), 2nd ed., Vol. 2c, pp. 282-289, Elsevier, Amsterdam/New York. 19. CROTEAU, R., AND LOOMIS, W. D. (1972) Phyte chemistry 11,1055-1066. 20. SUTHERLAND, M. D., AND WATERS, 0. J. (1961) Au& J. Chem. 14, 596-605. 21. EVANS, W. C., RAMAGE, G. R., AND SIMONSEN, J. L. (1934) J. Chem Sot, 1806-1810. 22. RYDON, H. N. (1937) J. Chem SW., 1340-1342.