Demonstration that limonene is the first cyclic intermediate in the biosynthesis of oxygenated p-menthane monoterpenes in Mentha piperita and other Mentha species

ARCHIVES OF BIOCHEMISTRY Vol. 220, No. 1, January,

AND BIOPHYSICS pp. 79-89, 1983

Demonstration that Limonene is the First Cyclic Intermediate Biosynthesis of Oxygenated p-Menthane Monoterpenes Mentha piper& and Other Mentha Species’12 ROBERT Institute

of Biological

Chemistry,

KJONAAS

AND

RODNEY

Biochemistry/Biophysics Pullman, Washington Received

June

in the in

CROTEAU3

Program, 99164

Washington

State

University.

29, 1982

The volatile oil of mature Me&ha pipe&a (peppermint) leaves contains as major components the oxygenated p-menthane monoterpenes l-menthol (47% ) and l-menthone (24%) as well as very low levels of the monoterpene olefins limonene (1%) and terpinolene (O.l%), which are considered to be probable precursors of the oxygenated derivatives. Immature leaves, which are actively synthesizing monoterpenes, produce an oil with comparatively higher levels of limonene (-3%), and isolation of the pure olefin showed this compound to consist of -80% of the I-(4S)-enantiomer and -20% of the d-(4R)-enantiomer. The time course of incorporation of [U-14C]sucrose into the monoterpenes of M. pipe&a shoot tips was consistent with the initial formation of limonene and its subsequent conversion to menthone via pulegone. d,Z-[9-3H]Limonene and [9,103H 1terpinolene were prepared and tested directly as precursors of oxygenated p-menthane monoterpenes in M. piperita shoot tips. Limonene was readily incorporated into pulegone, menthone, and other oxygenated derivatives, whereas terpinolene was not appreciably incorporated into these compounds. Similarly, d,Z-[9-3H]limonene was specifically incorporated into pulegone in Mentha pulegium and into the C-2-oxygenated derivative carvone in Mentha spicata, confirming the role of this olefin as the essential precursor of oxygenated p-menthane monoterpenes. Soluble enzyme preparations from the epidermis of immature M. piperita leaves converted the acyclic terpenoid precursor [l-3H]geranyl pyrophosphate to limonene as the major cyclic product, providing a further indication that this olefin plays a central role in the formation of oxygenated monoterpenes in Mentha. No free intermediates were detected in the cyclization of geranyl pyrophosphate to limonene, suggesting that the olefin is the first cyclic intermediate to arise in the pathway, and resolution of the biosynthetic limonene, by crystallization of the derived d- and I-carvoximes, indicated an enantiomer mixture nearly identical to that isolated from the leaf oil, Piperitenone is considered to be an essential intermediate in the biosynthesis of Z-menthone, Z-menthol, and other C-~-OXygenated p-menthane monoterpenes in

Mentha

species (Fig. 1). This supposition is based on many lines of evidence, ineluding both radiotracer and genetic studies, which have been reviewed in detail

i This is Scientific Paper No. 6275, Project 0268, College of Agriculture Research Center, Washington State University, Pullman, Wash. 99164. This investigation was supported in part by Department of Energy Contract DE-AM06-76RL02221, Agreement DEAT06-82ER12027, and by grants from the Washington Mint Commission and Mint Industry Research Council. The Hewlett-Packard 5840A-5985B GLC-MS system used in this work was purchased through the

assistance of the National Science Foundation (PCM 8100068). *We have utilized here the common nomenclature based on numbering of the p-menthane system (e.g., limonene = p-mentha-1,8(g)-diene, mentbone = pmenthan-3-one, carvone = p-mentha-6,8(g)-dien2-one). ‘Author to whom correspondence should be addressed. 79

0003-9861/83/010079-11$03.00/o Copyright All rights

0 1983 by Academic Press, Inc. of reproduction in any form reserved

80

KJONAAS

AND

CROTEAU

/

FIG. 1. Proposed terpenes in Mentha

pathway piperita.

for the conversion

elsewhere (l-8). Recently, all of the steps in the conversion of piperitenone to the menthol stereoisomers have been demonstrated in cell-free preparations from Me&a pipe&a (peppermint) leaves, and each of the NADPH-dependent reductases has been subjected to preliminary characterization (9-11). The origin of piperitenone, however, has remained uncertain, and two different biogenetic schemes, based largely on chemical and genetic considerations, have been proposed. One scheme posits the allylic oxygenation of the olefin terpinolene (p-mentha-1,4(8)diene), presumably via piperitenol, whereas the other invokes allylic oxygenation of the olefin limonene (p-mentha-1,8(9)diene), presumably via isoperitenol and isopiperitenone (Fig. 1). Circumstantial evidence for both proposals has been described (l-8), but no definitive evidence to distinguish the two alternatives is presently available. The cyclic dienes are generally minor constituents of Me&ha oils (e.g., limonene constitutes -1% and terpinolene constitutes -0.1% of a commercial peppermint oil sample (12)) and are considered to arise by independent cyclizations of the acyclic precursor geranyl pyrophosphate; reactions which have been

of geranyl

pyrophosphate

d-Neomenthol

to C-3-oxygenated

mono-

demonstrated in cell-free preparations from species other than Me&ha (13-14). Thus, while the origin of the presumptive precursors and the sequence of reactions from piperitenone to other products appear to be well established, there is no direct biochemical support for the proposed intervening steps. In this communication we provide evidence, from both in vivo and in vitro studies, that limonene, rather than terpinolene, is the key cyclic intermediate in the biosynthesis of C-3oxygenated p-menthane monoterpenes in in AI. piperita and of related compounds other Mentha species. EXPERIMENTAL

PROCEDURES

Plant materials, substrates, and reagents. M. piperita L. cv Black Mitcham (peppermint) plants were grown from stolons under controlled conditions described previously (15). The shoot apex and first leaf pair of actively growing plants (stems bearing 8-12 leaf pairs) were used in in viva experiments with [%]sucrose, whereas newly emerged leaves (-1.0 cm in length) were used in all other in viva experiments and to prepare cell-free epidermis extracts. GLC-MS4 4 Abbreviations raphy; MS, mass matography.

used: GLC, spectrometry;

gas-liquid chromatogTLC, thin-layer chro-

BIOSYNTHESIS

OF

p-MENTHANE

analysis was carried out on an oil sample that was steam-distilled from only shoot apices and first leaf pairs. A pure sample (99 + W) of limonene for determination of optical rotation was obtained from the oil by combination of fractional distillation and argentation column chromatography (Mallinckrodt SilicAR CC-7 containing 8% AgN03, with redistilled hexane). M. pulegium (pennyroyal) and M. spicata (spearmint) were grown under the same conditions employed for M piperita. [U-i4C]Sucrose (3.6 Ci/mol; radiochemical purity >99%) was obtained from New England Nuclear Corporation, and [l-3Hz]geranyl pyrophosphate (100 Ci/mol), [1-aH2]neryl pyrophosphate (100 CVmol), d,l-[l-3Hz]linaloyl pyrophosphate (200 Ci/mol), and d,l-[3-3Hz]a-terpineol (84 Ci/mol) were prepared and purified as described previously (13). d,l-[9-3Hz]Limonene (42.5 Ci/mol) was prepared from d,l-limonene (i.e., technical dipentene (75%; remainder p-cymene) from Aldrich Chemical Co.). Hydroboration of the olefin (in twofold excess) followed by oxidation with alkaline hydrogen peroxide (16) afforded d,l-p-menthl-en-g-o], which was oxidized with CrOs:pyridine complex (1’7) to the corresponding aldehyde and purified by TLC (silica gel G with hexane:ethyl acetate (19:1, v/v) (system A), followed by silica gel G impregnated with 8% AgNOa with hexane:ethyl acetate (9:1, v/v) (system B)). Reduction of the aldehyde with NaBsHI (New England Nuclear) in methanokwater (9:1, v/v) yielded the original alcohol now labeled at C-9 (42.5 Ci/mol). The labeled alcohol (0.4 mmol) was dissolved in 0.6 ml dimethylformamide and converted to the corresponding bromide by dropwise addition to a test tube containing triphenylphosphine dibromide (generated in 0.8 ml dimethylformamide at 0°C using 0.5 mmol triphenylphosphine and 0.4 mmol Br,) (18). This reaction mixture was sealed under Nz, warmed to room temperature, and then heated at 70°C in an oil bath for 12 h. Elimination of the bromide was carried out by addition of 0.5 ml of 1,8diazabicyclo[5.4.0]undec-7-ene and heating at 90°C (19). The reaction was monitored by GLC and was complete within 2 h. The product was extracted with pentane, and after backwashing with dilute HCl and water, the pentane extract was eluted through a 1.5 X lo-cm column of silicic acid (Mallinckrodt SilicAR CC-7) to afford pure d,l-[9-“H,]limonene. The overall yield of limonene, starting from the NaB3H, reduction step, was about 15%. [9,10-aH]Terpinolene (42.5 Ci/mol) was prepared from d,l-[9-3Hz]limonene via d,l-[9,10-3H]u-terpineol obtained by hydroxymercuration (1 equivalent mercuric acetate/mm01 limonene in 1.0 ml tetrahydrofuran:water (l:l, v/v), followed by addition of 3.0 N NaOH and excess NaBH* and subsequent ether extraction of the product (20)). The labeled alcohol (in pyridine at 0°C) was converted to a mixture of [9,103H]terpinolene and [9,10-3H]limonene by dropwise

MONOTERPENES

81

addition of a slight excess of PBr3 in pyridine. The olelins were isolated from the reaction mixture by partitioning between 2-methylbutane and dilute HCl and were separated by elution through a 0.5 X 5-cm column of silicic acid impregnated with 20% AgN03. For use as substrates, the monoterpenes were suspended in water with the aid of Tween 20 (20 pg/ pmol for alcohols and 50 Fg/pmol for olefins) and sonication. Authentic standards of l-menthone, d-isomenthone, I-menthol, d-neomenthol, d-isomenthol, d-neoisomenthol, d-pulegone, d-piperitone, d-menthofuran, piperitenone, and terpinolene (all 97 + %), and a mixture of cis- and trots-isopiperitenol were provided through the generosity of K. Bauer and R. Hopp, Haarmann and Reimer GmbH, Holzminden, West Germany, and R. Carrington, I. P. Callison and Sons, Chehalis, Washington. l-Carvone (98% optically pure), d-carvone (99% optically pure), I-limonene (84% optically pure), and d-limonene (93% optically pure) were from Aldrich Chemical Company. d,l-Isopiperitenone was prepared by allylic oxidation of d,l-limonene with CrO,-pyridine complex (21). The reaction mixture was poured onto a column of magnesium silicate (Florisil, J. T. Baker) which was eluted with ether. Evaporation of solvents provided an oil that was separated by TLC (silica gel G with hexane:ethyl acetate (4:1, v/v) (system C)) to afford d,lisopiperitenone (R, = 0.41) and the coproduct d,l-carvone (R, = 0.58). Polyvinylpyrrolidone (GAF Corp.) and Amberlite XAD-4 polystyrene resin (Rohm and Haas Corp.) were purified by standard procedures for use as adsorbents (22,23). Unless otherwise specified, all other reagents and biochemicals were obtained from Sigma Chemical Company or Aldrich Chemical Company. In viva exptiments. For experiments with [Ui4C]sucrose, matched peppermint cuttings (shoot tip and first leaf pair) were placed in vials and administered an aqueous solution containing 1.39 pmol of [i4C]sucrose (5.0 j&i in 0.1 ml) through the cut stems. Incubation was at 30°C in the light (21,000 lx). After the uptake of labeled material (l-2 h), the vials were kept filled with distilled water and, after periods of 7, 14, and 21 h, cuttings were transferred to a Ten Broeck homogenizer and thoroughly extracted with 15 ml diethyl ether. Authentic volatile oil (20 ~1) or 3 mg each of the appropriate individual standards (limonene, terpinolene, menthone, isomenthone, piperitenone, isopiperitenone, pulegone, menthofuran, piperitone, piperitenol, cis- and trans-isopiperitenol, and the menthol isomers) was added to the extract, which was concentrated to 5.0 ml and subjected to exhaustive microscale steam distillation (13). The aqueous phase of the distillate was saturated with NaCl and the volatile organic products recovered by ether extraction. Aliquots were taken for determination of radioactivity and for radio-GLC analysis.

82

KJONAAS

AND

For experiments with labeled olefins, intact leaves (1.5 g) were placed in a 25-ml screw-cap vial to which 0.2 ml of an aqueous suspension of the appropriate substrate was added (0.175 rmol of d,l-[9-3H]limonene or [9,10-aH]terpinolene). The sealed vial was then shaken to distribute the substrate and incubated at 30°C in the dark or in the light (21,000 lx) for 12 h. In some experiments, the vial was flushed with Nz, to lower the Oz level, before the addition of substrate and sealing for incubation in the light. After incubation, the vial was chilled in ice and the contents transferred to a Ten Broeck homogenizer and thoroughly extracted with 15 ml diethyl ether. Internal standards were added, and the extract was distilled and analyzed as before by radio-GLC and TLC (system C). Control experiments were carried out with tissue that had been steamed for 15 min before incubation. Preparation of epidermis extracts. Epidermis extracts were prepared by a technique described previously (24) which consists of gently brushing off epidermal fragments from individual leaves immersed in cold 0.1 M sodium phosphate buffer (pH 6.5) containing 0.25 M sucrose, 50 mM sodium ascorbate, 50 mM NazSzOs, 5.0 mM MgClz, 1.0 mM dithioerythritol, 1.0 mM EDTA, and insoluble polyvinylpyrrolidone (0.3 g Polyclar AT in 4 ml buffer/g tissue). The buffer containing the epidermal fragments was then transferred to a Ten Broeck homogenizer and homogenized, slurried with an equal tissue weight of XAD4 resin, filtered through cheesecloth, and centrifuged at 105,OOOg for 90 min (pellet discarded) to afford the soluble enzyme fraction. This preparation was brought to assay conditions by dialysis against 0.05 M sodium phosphate buffer (pH 6.5) containing 1 mM ammonium vanadate (V), 0.5 mM dithioerythritol, and 5% sorbitol. Throughout the course of the investigation, these preparations consistently provided limonene cyclase activity in the range of 60 to 120 nmol product per hour per milligram protein. For preparing a crude particulate fraction, the same extraction buffer was used, except that soluble polyvinylpyrrolidone (Plasdone K-90) was substituted for the insoluble polymer and that the XAD4 treatment and filtration were omitted. The particulate fraction obtained by centrifugation of the homogenate (a 500g supernatant was pelleted at 105,OOOg) was washed (by resuspension with the assay buffer and recentrifugation) and then dialyzed to assay conditions as before. Cyclase assay. For measurement of monoterpene cyclase activity, a l-ml aliquot of the enzyme preparation (-5 pg protein), at the buffer conditions described above and containing 10 mM MgClz or 1.0 mM MnCl,, was added to a Teflon-sealed screw-cap vial. The reaction was started by the addition of labeled substrate (50 PM), and 1 ml of pentane was carefully overlaid to trap volatile monoterpene products. Fol-

CROTEAU

lowing incubation for 1 h at 3O”C, the reaction was stopped by vigorous mixing and the pentane layer removed. The aqueous phase was reextracted with an additional 1.5 ml of pentane, and the combined extract eluted through a column (0.5 X 5 cm) of silicic acid (Mallinkrodt SilicAR CC-7) in a Pasteur pipet, followed by washing with another 0.5 ml of pentane. The hydrocarbon fraction thus obtained is free of oxygenated products, which remain adsorbed on the silicic acid. Internal standards (5 mg each of the appropriate olefins) were then added, an aliquot was removed for determination of radioactivity, and the remainder was concentrated under a stream of Nz at 0°C for analysis by radio-GLC and TLC (silica gel G with redistilled hexane (system D)). The oxygenated monoterpene products were recovered by reextracting the original aqueous layer of the incubation mixture with ether (1.5 ml), followed by elution of this extract through the original pentane-washed silicic acid column. Appropriate internal standards (3 mg each) were added to the eluant, an aliquot was taken for tritium determination, and the remainder was concentrated as before for radio-GLC and TLC analysis (silica gel G with hexane:ethyl acetate (7:3, v/v) (system E)). TLC-purified products to be examined further were eluted from the silica gel with ether. In assays with [3-3H]a-terpineol (50 FM) as the substrate, the aforementioned pentane overlay was omitted, and after incubation the pentane was added and the products were analyzed as above. Appropriate boiled controls were included in each experiment, and in all cases nonenzymatic formation of the terpenes of interest was negligible. Protein was routinely estimated by the dye-binding assay (Bio-Rad Laboratories). Preparation of derivatives. Standard literature procedures (25) were employed to convert terpenoid ketones to semicarbazones (semicarbazide hydrochloride:sodium acetate:ketone (1:l:l molar ratio) in 20% ethanol heated on a steam bath for 1 h). The semicarbazones (prepared on the 250-500 mg scale) were extracted from the reaction mixture with ether and purified by preparative TLC (silica gel G with diethyl ether (system F)) before crystallization from aqueous ethanol. Limonene was converted to a mixture of a- and fl-nitrosochlorides with nitrosyl chloride generated from nitrosylsulfuric acid and NaCl/ HCl (26). The crystalline mass of OI- and p-isomers was then heated briefly in pyridine (2’7) to afford carvoxime. (The original endocyclic double bond is functionalized and a new endocyclic double generated in the adjacent position by elimination of HCI, thereby resulting in the conversion of l-(4S)-limonene to d(4R)-carvoxime and vice versa.) The product, after dilution with pure oxime standards (25), was isolated by ether extraction from dilute HCl, purified by preparative TLC (system E), and crystallized from aqueous ethanol. Isopiperitenone was heated with an

BIOSYNTHESIS

OF

p-MENTHANE

equivalent weight of PtzO in benzene in a sealed tube at 150°C overnight to affect its conversion to thymol (-55% yield), which was purified by TLC (silica gel G with hexane:ethyl acetate (l:l, v/v) (system G)). The two-phase Cr03 oxidation of menthol to menthone has been described (28). Chromatography and determination of radioactivity. Thin-layer chromatography with the developing solvents indicated in the text was done on 1.0.mm layers of silica gel G (with or without impregnation with 8% AgNO,) activated at 110°C for 3 h. For argentation TLC, solvent system B (hexane:ethyl acetate (9:1, v/v)) or D (redistilled hexane) was employed. For silica gel TLC, the solvents were system A, hexane:ethyl acetate (19:1, v/v); system C, hexane:ethyl acetate (4:1, v/v); system E, hexane:ethyl acetate (7:3, v/v); system F, diethyl ether; and system G, hexane:ethyl acetate (l:l, v/v). The developed chromatograms were sprayed with a 0.2% ethanolic solution of 2,7-dichlorofluorescein to locate (under uv light) the appropriate components, which were eluted from the gel with diethyl ether. Radio-gas-liquid chromatography was performed on a Varian gas chromatograph attached to a Model 7357 Nuclear Chicago radioactivity monitor (calibrated externally with [aH]toluene, with integration of radio output). Chromatography columns were 10 ft X 0.125 in. o.d. stainless steel, packed with 25% Carbowax 4000 on 60/80 mesh Gas-Chrome Q (programmed as indicated in the figure legend), and 8 ft x 0.125 in. o.d. stainless steel, packed with 10% SE30 on 80/100 mesh Gas-Chrome Q (operated at 85°C for hydrocarbons and 165°C for oxygenated compounds). Analytical chromatography and GLC-MS were performed on a 25-m fused silica capillary column coated with SE-30. Radioactivity in liquid samples and TLC isolates was determined in a counting solution (15 ml) consisting of 0.3% (w/v) Omnifluor (New England Nuclear) dissolved in 30% ethanol in toluene. The counting efficiency for 3H was 37% and for 14C was 83%, and all assays were conducted with a standard deviation of less than 3%. RESULTS

Oil composition. Mature M. piperita leaves contain primarily oxygenated pmenthane monoterpenes, such as l-menthol (47%) and I-menthone (24%), with very low levels of the monoterpene olefins limonene (-1%) and terpinolene (-0.1%) thought to be possible precursors of the oxygenated derivatives (7). Young expanding leaves are known to be the most active in monoterpene biosynthesis (29-32), and GLC-MS analysis of the oil from this im-

MONOTERPENES

83

mature tissue (0.24% yield fr. wt.) showed it to contain menthone (35%), menthol (14%), 1,8-cineole (13%), pulegone (7.2%), isomenthone (4.5%), neomenthol (3.‘7%), and limonene (2.6%); terpinolene was present in only trace quantities. Examination of the purified limonene indicated the presence of 79.6% of the I-(4S)-enantiomer and 20.4% of the d-(4R)-enantiomer ([cy]? -75.0 (C = 2, EtOH); highest Lit. value [cy]$ f126.8, Ref. (33)), consistent with earlier reports that the limonene of M. pipe&a was levorotatory (7, 34). The comparatively higher level of limonene in the immature tissue suggested that this olefin might be an early intermediate in the pathway to oxygenated compounds, which is subsequently metabolized to these derivatives on further development. To examine this possibility, preliminary examination of the biosynthetic capability of this tissue was carried out. Time course of [ U-‘“C]sucrose incorporation. Sucrose is an effective precursor of monoterpenes in leaf tissue (31, 35), and when administered (at the 5.0-&i level) to immature &I. pipe&a leaves it afforded (after 14 h) a radioactive volatile oil (0.06 &I) that on radio-GLC analysis was shown to consist largely of limonene (0.010 PCi), pulegone (0.024 PCi), and menthone (0.017 PCi), with less than 0.002 &i each of cineole, neomenthol, and menthol (Figs. 2A and B; the resolution of limonene from cineole was achieved at a lower column temperature). Thus, under the conditions of the experiment, limonene was formed from sucrose in higher relative proportion than that in the monoterpene fraction of the tissue at this early stage of development. Conversely, terpinolene, which was a trace constituent of the oil, was not formed in detectable levels from sucrose in young leaves. Administration of [U14C]sucrosefor 7 and 21 h, under the same conditions as the 14-h incorporation, provided data for the kinetics of labeling which were consistent with the passage of tracer from limonene through pulegone to menthone and finally the menthols (Table I). Conversion of [Y-‘H]limonene into oxygenated monoterpenes of Mentha. To fur-

84

KJONAAS

AND

CROTEAU

r

TABLE

I

TIME COURSE OF INCORPORATION OF [UWISUCROSE INTO MONOTERPENES OF Mentha piperita” Product Time (hr) 7 14 21

Limonene 19 10 8

formation

Pulegone 5 24 20

(&i Menthone 8 17 62

X 103) Menthols 1 3 6

a Cuttings were administered 5.0 pCi [U-“C]sucrose, and the steam-volatile products isolated and analyzed as described in the text. Values indicated are the means of three experiments.

FIG. 2. Radio-gas-liquid chromatograms of the steam-distilled fraction of the ether-soluble products obtained when (B) [U-‘%]sucrose or (C) d,l-[9‘Hllimonene was incubated with M. pipe&a leaves, and when d,l-[9-3H]limonene was incubated with leaves of (D) M. spicata and (E) M. pulegium. The smooth bottom tracing (A) represents the response of the flame-ionization detector to coinjected standards of (a) d,Z-limonene; (b) 1,8-cineole; (c) terpinolene; (d) I-menthone; (e) d-isomenthone; (f) d-neomenthol; (g) l-menthol; (h) d-pulegone; and (i) d-piperitone/d,l-isopiperitenol. d-Menthofuran, disomenthol, d-neoisomenthol, piperitenone, and d,lisopiperitenone were also included as standards, but as little radioactivity was associated with these compounds, they are not indicated in the figure. In studies with &f. spicata, only d,l-carvone was employed as an internal standard (position of elution indicated by the arrow in chromatogram D). The gas-liquid chromatographic column (Carbowax 4000) was held at 130°C for 8 min and then programmed at 4”C/min to 175°C and held isothermally. The argon flow rate was held at 120 cm3/min. To resolve limonene from

ther examine the possible role of limonene as a precursor of the C-&oxygenated terpenes, d,l-[9-3H]limonene was prepared and incubated with immature M. pipe&a leaves in the light for 12 h. Extraction of the tissue with ether, followed by steam distillation of the extract, provided a volatile fraction in which about half of the original label (3.4 &i) was recovered. (Considerable loss of the volatile substrate occurred during incubation and handling, and much less than 1 &i of the label was recovered in residual fractions, including the nonvolatile and water-soluble materials.) Radio-gas-liquid chromatographic analysis of the volatile fraction indicated that the substrate had been converted to several products chromatographically coincident with menthone, pulegone, piperitone or isopiperitenol, neomenthol, and menthol (Fig. 2C); however, thin-layer chromatographic separation of the products (system C) allowed the recovery of only menthone (0.15 &i), pulegone (0.13 &i), menthol (0.076 &i), cis/trans-isopiperitenol (0.031 &i), and neomenthol (co.030 PCi). Negligible activity was recovered with authentic piperitone (pmenth-1-en-3-one). Subsequent analysis of the products obtained on incubating [3H]limonene with the thermally inacti-

l&cineole, the column was held at 100°C with a flow rate of 60 cms/min. See text for details of incorporation rates and distribution of radioactivity into the various products.

BIOSYNTHESIS

OF

p-MENTHANE

vated control verified the nonenzymatic formation of two products gas chromatographically coincident with piperitone and neomenthol, respectively. These presently unidentified products are presumed to result from photooxidative processes (36), as they were not formed in detectable amounts when [3H]limonene was incubated with active tissue in the dark (see below). The identities of the major biosynthetic products as I-menthone, d-pulegone, and Z-menthol were verified by dilution of the TLC-purified terpenoids with authentic materials and crystallization of an appropriate derivative to constant specific activity (through a minimum of three crystallizations). GrH]Menthone was converted to the semicarbazone and crystallized to a constant specific activity of 56.2 &i/mol (mp 186-189°C; Lit. 189°C (33)), while d-pulegone semicarbazone was crystallized to a constant specific activity of 44.9 #Xmol (mp 172-174°C; Lit. 174°C (33)). The presumptive l-[3H]menthol was oxidized to menthone, and the semicarbazone was prepared and crystallized to a constant specific activity of 30.0 &i/m01 (mp 184-187°C; the depressed melting point resulting from epimerization to disomenthone on Cr03 oxidation). The recovered levels of neomenthol and isopiperitenol (the cis- and trans-isomers were not readily separated under our chromatographic conditions) were too low to permit the preparation of crystalline derivatives, and so these compounds were oxidized to their respective ketones, menthone and isopiperitenone, and their identities confirmed by coincidence on radio-GLC analysis. The incorporation of [3H]limonene into oxygenated monoterpenes of M. piper&a was examined in five separate experiments, and, in most cases, low but detectable levels (co.03 PCi) of 3H were associated with isopiperitenone on radio-GLC analysis (each experiment ineluded isopiperitenone and piperitenone, and the corresponding alcohols, menthofuran and the four menthol stereomers as internal standards, but for the sake of simplicity not all of these standards are indicated in Fig. 2). As the level of radio-

MONOTERPENES

85

activity was too low to permit the preparation of a crystalline derivative, the identity of isopiperitenone was substantiated by conversion to thymol and radio-GLC analysis. Although circumstantial, the summary of these results suggests a pathway from limonene to piperitenone, and its reduction products, that involves isopiperitenol and isopiperitenone as intermediates (Fig. 1). Steam-inactivated tissue on incubation with [3HJlimonene afforded, on identical analysis, none of the aforementioned authentic terpenoid products, but only the two presumed photooxidation products. Incubation of immature leaves with [3H]limonene in the dark provided a product distribution very similar to that obtained in the light (menthone (0.17 &i), pulegone (0.10 PCi), menthol (0.04 &i), and minor amounts (co.03 &I each) of neomenthol, isopiperitenol, and isopiperitenone), and in this instance no photooxidation products were detected. Thus, photochemical processes were not involved in the conversion of limonene to the authentic oxygenated terpenoid products observed. Incubation of [3H]limonene with tissue under a Nz atmosphere afforded only low levels of oxygenated terpenes (CO.03 PCi), suggesting a requirement for dioxygen in the conversion of limonene to menthone and its congeners. [9,10-3H]Terpinolene was also tested as a precursor of oxygenated monoterpenes in M. piper& under the same conditions as for limonene, and in this instance no incorporation into the appropriate oxygenated monoterpenes was detected (however, low levels of photooxidation products were noted). The result of this direct feeding experiment, therefore, was consistent with the [U-‘4C]sucrose time-course experiment in suggesting that terpinolene was not a precursor of C-3-oxygenated p-menthane monoterpenes in hf. pipe&a. In a final set of in viva experiments, the role of limonene as a precursor of oxygenated p-menthane monoterpenes in both M. spicata and M. pulegium was examined. M. spicata produces an oil containing primarily the C-2-oxygenated ketone l-carvone (p-mentha-6,8(g)-dien-Z-one), and the

86

KJONAAS

AND

limonene isolated from this source is also reported to be levorotatory (34, 37). Incubation of immature M. spicata leaves with d,Z-[9-3H]limonene (7.5 yCi) afforded a volatile fraction (3.2 PCi), which on radio-GLC analysis was shown to consist of residual substrate (2.8 PCi) and carvone (0.18 yCi) (Fig. 2D). M. pulegium produces an oil containing d-pulegone as a major constituent, and the limonene isolated from this source is levorotatory (34). Incubation of immature M. pulegium leaves with d,l-[9-3H]limonene (7.5 PCi) afforded a volatile fraction (3.2 &i), which on radio-GLC analysis was shown to consist of residual substrate (2.5 &i) and pulegone (0.30 &i) (Fig. 2E). Steam-inactivated controls of M. spicata and M. pulegium afforded neither carvone nor pulegone, although minor amounts of photooxidation products were again observed. These results support the general role of limonene as the precursor of C-3-oxygenated monoterpenes in M piperita and M. pulegium, and they also confirm earlier proposals regarding the precursor role of limonene in the biosynthesis of the C-2-oxygenated compound carvone in M. spicata (1, 3, 7, 8, 37, 38). Cyclixation of geranyl pyrophosphate to limonene in cell-free extracts. If limonene is a key intermediate in the biosynthesis of oxygenated p-menthanes, then, based on earlier work on the origin of this and other monoterpene olefins (8, 39), limonene should be a major product of the cyclization reaction(s) in M. piperita. This possibility was examined using cell-free extracts prepared from the epidermis of young M. piperita leaves. The epidermal oil glands are the presumed site of monoterpene biosynthesis (1, 31, 40), and epidermis extracts from M piperita have been used previously for studies on monoterpene metabolism (24). Such preparations are not only enriched in the enzymes of monoterpene biosynthesis, but are also relatively free of competing activities (e.g., phosphatases) that can cause interferences in whole-leaf extracts. Incubation of the soluble supernatant obtained from a homogenate of M. pipe&a epidermis with the acyclic precursor [l-

CROTEAU

3H]geranyl pyrophosphate under general conditions developed for other monoterpene cyclases (8), followed by radio-GLC analysis of the isolated hydrocarbon fraction, indicated the presence of a single labeled compound coincident with authentic d,l-limonene. The product was verified as limonene by conversion to carvoxime via the nitrosochloride as described below. Radiochromatographic examination of the oxygenated monoterpene fraction obtained from the incubation mixture revealed that no detectable cyclic products were present; only geraniol (and a small amount of the corresponding aldehyde), derived from the action of endogenous phosphatases, and linalool (3,7-dimethylocta-1,6-dien-3-ol), presumably from the nonenzymatic rearrangement of geranyl pyrophosphate (41), were observed. Boiled controls produced no detectable limonene, geraniol, or geranial from geranyl pyrophosphate, although linalool was formed by the nonenzymatic rearrangment referred to above. Preliminary investigations indicated that the phosphatase/pyrophosphatase inhibitor ammonium vanadate (V) (42,43) materially improved limonene formation when included in the incubation mixture (an approximate doubling of activity at the 1 mM inhibitor level), whereas vanadate (IV), molybdate, and fluoride were less effective. All subsequent assays of limonene cyclase activity were carried out in the presence of 1 mM NHIVOB. A divalent cation is an absolute requirement for all monoterpene cyclases thus far examined (8, 39), and in the present case dialysis of the soluble preparation to remove divalent cations reduced limonene cyclase activity to a level only marginally higher than that of a boiled control (Table II). Addition of 1.0 mM MnClz restored essentially full activity, whereas MgCl, up to the 10 mM level was much less effective in promoting limonene formation. Particulate preparations from M piperita epidermis were relatively inactive in cyclizing geranyl pyrophosphate to limonene; exhibiting, at best, reaction efficiencies less than 10% of those of comparable soluble enzyme preparations (Table II).

BIOSYNTHESIS

OF

p-MENTHANE

TABLE

87

MONOTERPENES

II

CONVERSIONOF ACYCLIC AND CYCLICPRECURSORSTO LIMONENE BY CELL-FREE EXTRACTS OF Mentha piperita EPIDERMIS Substrate Geranyl pyrophosphate Geranyl pyrophosphate Geranyl pyrophosphate Geranyl pyrophosphate Neryl pyrophosphate Neryl pyrophosphate a-Terpineol Geranyl pyrophosphate Neryl pyrophosphate

Enzyme fraction Soluble Soluble Soluble Soluble Soluble Soluble Soluble Particulate Particulate

Additions 1 10 1 1 10 1 1 1

mM mM mM mM mM mM mM mM

None MnC12 MgC& MnC& (boiled) MnCla MgCl, MnCla MnCl, MnCl,

Limonene formed (nmol) 0.01 0.45 0.15 10.01 0.09 0.03 10.01 0.04
’ Epidermis extracts (105,OOOg supernatant or 105,OOOg-5009 pellet) consisting of -5 fig protein in 1 ml of 0.05 M sodium phosphate buffer (pH 6.5) containing 1 mM ammonium vanadate, 0.5 mM dithioerythritol, 50 PM substrate, and 5% sorbitol were incubated at 30°C for 1 h and the products analyzed as described in the text. In the absence of vanadate, maximum product formation was reduced from 0.45 to 0.21 nmol.

A previous indication (9, 44) that neryl pyrophosphate (3,7-dimethylocta-2(cis),6dienyl pyrophosphate) was enzymatically cyclized to ol-terpineol (p-menth-l-en-8 ol), which, in turn, might be dehydrated to olefins, was based on studies with crude particulate preparations from whole M. pipe&a leaves; however, this observation could not be substantiated with epidermis preparations. Thus, [l-3H]neryl pyrophosphate did not give rise to cu-terpineol in levels detectable higher than those of boiled controls, and this substrate afforded significantly lower rates of limonene formation than did geranyl pyrophosphate (Table II). Furthermore, when d,Z-[3-3H]cu-terpineol was incubated with similar preparations, no limonene or other olefin was detected as a product (Table II). Therefore, as has been shown with other species (13, 14), a-terpineol is not a precursor of limonene in M. pipe&a, and no evidence was obtained to suggest the involvement of any free intermediate in the conversion of the acyclic precursor to the cyclic olefin. Additional studies (data not shown) indicated that the acyclic compound linaloyl pyrophosphate (which can serve as a precursor of cyclic monoterpenes in a number of other systems (8)) could also serve as a precursor of limonene

in the epidermis

preparation from M. pithree times as efficient as geranyl pyrophosphate) and that mesophyll-derived preparations were essentially devoid of limonene cyclase activity even in the presence of vanadate.

perita (roughly

Stereochemistry of the cyclixation product. Because the limonene isolated from M. pipe&a oil was shown to consist of -80%

of the l-isomer and -20% of the d-isomer, it was essential to determine the stereochemical composition of the product derived from [l-3H]geranyl pyrophosphate in the cell-free system. Biosynthetic [3H]limonene was converted to the nitrosochloride and then, by treatment with base, to carvoxime. This sample was divided in half, and one half diluted about a millionfold with pure d-carvoxime (recrystallized to [a]$ t39.8”; mp 73°C) to a specific activity of 110 &i/mol. The remaining half was diluted identically with pure Z-carvoxime (recrystallized to [ag -39.7”; mp 73°C). The d-[“Hlcarvoxime (from Z-[3H]limonene) was crystallized to a constant specific activity of 81.8 &i/mol (mp 73”C), whereas the Z-[3H]carvoxime (from d-[3H]limonene) was crystallized to a constant specific activity of 25.9 &i/m01 (mp 73°C) (Lit. mp 72-73’C (33)). Thus, by radiochemical fractional crystallization,

88

KJONAAS

AND

the biosynthetic product was shown to consist of 76% l-limonene and 24% d-limonene, consistent with the oil analysis. DISCUSSION

The results of this communication provide direct evidence that limonene, rather than terpinolene, is the progenitor of C-3oxygenated monoterpenes in M. pipe&a and that this olefin is synthesized in the leaf epidermis by an operationally soluble cyclase. This work also seems to eliminate a role for a-terpineol in the formation of limonene as implied from earlier studies based on whole-leaf extracts in which a particulate neryl pyrophosphate:a-terpineol cyclase activity (but no soluble cyclase) was reported (44). The inability to demonstrate the soluble limonene cyclase in this earlier work likely resulted from the presence of mesophyll-derived phosphatases that rapidly hydrolyze prenyl pyrophosphates in these whole-leaf extracts. The relevance to monoterpene biosynthesis of the particulate a-terpineol cyclase activity obtained earlier from M. piperita (44), and more recently by Suga and co-workers from M. spicata (45), is now questionable. The data presented also suggest a pathway from limonene to the C-3-oxygenated ketone piperitenone involving cis- or transisopiperitenol and isopiperitenone as intermediates (Fig. l), and they additionally confirm the earlier indications that limonene serves as a precursor of the C-~-OXygenated ketone carvone. Photooxidative processes are unlikely to be involved in these transformations. Finally, supportive evidence is provided that both d- and llimonene are synthesized in M. pipe&a and that both enantiomers may be subsequently oxygenated in this species (i.e., the stereochemical composition of the direct cyclization product is nearly identical to the limonene accumulated in the oil, implying a lack of discrimination in the oxygenation process). This latter possibility may explain recent observations (37, 46) on the apparent lack of stereo/regiospecificity in the formation of the C-3 side chain of oxygenated p-menthanes in Men-

CROTEAU

tha, as determined by labeling patterns obtained when basic precursors such as mevalonate and geraniol are administered to intact plants. Further studies on the cyclization(s) to the enantiomeric limonenes, and on the subsequent oxygenation and reduction steps, are underway in order to examine in more detail the pathway and the stereochemical features of these reactions. ACKNOWLEDGMENTS

We thank R. Hamlin for raising the plants, and R. Hopp and R. Carrington for the generous gifts of monoterpene standards. REFERENCES LOOMIS, W. D. (1967) in Terpenoids in Plants (Pridham, J. B., ed.), pp. 59-82, Academic Press, New York. 2. BANTHORPE, D. V., CHARLWOOD,B. V., AND FRANCIS, M. J. 0. (19’72) Chem. Rev. 72, 115-155. 3. HEFENDEHL, F. W., AND MURRAY, M. J. (1976) 1.

Lloydia 4. 5. 6.

7.

8.

9.

10. 11. 12. 13. 14. 15.

39, 39-52.

SCH~TTE, H.-R. (1976) Progr. Botany 38,129-147. CHARLWOOD,B. V., AND BANTHORPE, D. V. (1978) Progr. Phytochem. 5, 65-125. CROTEAU, R. (1980) in Fragrance and Flavor Substances (Croteau, R., ed.), pp. 13-36, D & PS Verlag, Pattensen. LAWRENCE, B. M. (1981) in Progress in the Chemistry of Essential Oils (Mookherjee, B. D., and Mussinan, C. J., eds.), pp. 1-81, Allured, Wheaton, III. CROTEAU, R. (1981) in Biosynthesis of Isoprenoid Compounds (Porter, J. W., and Spurgeon, S. L., eds.), Vol. I, pp. 225-282, Wiley, New York. BURBOTT, A. J., CROTEAU, R., SHINE, W. E., AND LOOMIS, W. D. (1974) VIth Int. Essential Oil Congress, Paper No. 17, San Francisco. BURBOTT, A. J., AND LOOMIS, W. D. (1980) Plant Physiol. (Suppl.) 65, 96, (Abst. 522). KJONAAS, R., MARTINKUS-TAYLOR, C., AND CROTEAU, R. (1982) Plant Physiol. 69, 1013-1017. LAWRENCE, B. M., HOGG, 3. W., AND TERHUNE, S. J. (1972) Flavour Znd. 3, 467-472. CROTEAU, R., AND KARP, F. (1976) Arch. Biochem Biophys. 176, ‘734-746. CHAYET, L., ROJAS, C., CARDEMIL, E., JABALQUINTO, A. M., VICUNA, R., AND CORI, 0. (1977) Arch. Biochem. Biophys. 180, 338-327. CROTEAU, R., AND MARTINKUS, C. (1979) Plant Physiol 64, 169-175.

BIOSYNTHESIS

OF

p-MENTHANE

16. ACHARYA, S. P., BROWN, H. C., SUZUKI, A., NoZAWA, S., AND ITOH, M. (1969) J. Org. C&m. 34, 3015-3022. 17. RATCLIFFE, R., AND RODEHORST, R. (1970) J. Org. Chem. 35, 4000-4002. 18. WILEY, G. A., HERSHKOWITZ, R. L., REIN, B. M., AND CHUNG, B. C. (1964) J. Amer. C&m. Sot. 86, 964-965. 19. OEDIGER, H., MILLER, F., AND EITER, K. (1972) Synthesis 591-598. 20. BROWN, H. C., AND GEOGHEGAN, P. J. (1970) J. Org. Ckem. 35, 1844-1850. 21. DAUBEN, W. G., LORBER, M., AND FULLERTON, D. S. (1969) J. Org. Chem. 34, 3587-3592. 22. LOOMIS, W. D. (1974) in Methods in Enzymology (Fleischer, S., and Packer, L., eds.), Vol. 31, pp. 528-544, Academic Press, New York. 23. LOOMIS, W. D., LILE, J. D., SANDSTROM, R. P., AND BURBOTT, A. J. (1979) Phytochemistry 18,10491054. 24. CROTEAU, R., AND WINTERS, J. N. (1982) Plant Physiol. 69, 975-977. 25. STERRETT, F. S. (1975) in The Essential Oils (Guenther, E., ed.), Vol. II (reprinted), pp. 809820, Krieger, Huntington, N. Y. 26. BECKHAM, L. J., FESSLER, W. A., AND KISE, M. A. (1951) Ckem. Rev. 48, 321-335. 27. PUMMERER, R., AND GRASER, F. (1953) Annalen 583, 207-224. 28. BROWN, H. C., AND GARG, C. P. (1961) J. Amer. Chem. Sot. 83, 2952-2953. 29. BURBOTT, A. J., AND LOOMIS, W. D. (1969) Plant Physiol. 44, 173-179. 30. BATTAILE, J., AND LOOMIS, W. D. (1961) B&hem. Bioph.ys. Acta 51, 545-552. 31. CROTEAU, R. (1977) Plant Physiol. 59, 519-520. 32. CROTEAU, R., FELTON, M., KARP, F., AND KJONAAS, R. (1981) Plant Pkysiol. 67, 820-824.

89

MONOTERPENES

33. BARTON, D. H. R., AND HARPER, S. H. (1953) i?~ Chemistry of Carbon Compounds (Rodd, E. H., ed.), Vol. IIB, pp. 507, 515, 526-529, Elsevier, New York. 34. GUENTHER, E. (1974) The Essential Oils, Vol. III (reprinted), pp. 575-683, Krieger, Huntington, N. Y. 35. CROTEAU, R., BURBOTT, A. J., AND LOOMIS, W. D. (1972) Phytochemistyy 11, 2937-2948. 36. SCHENCK, G. O., NEUM~LLER, 0. A., OHLOFF, G., AND SCHROETER, S. (1965) Annalen, 687,26-39. 37. AKHILA, A., BANTHORPE, D. V., AND ROWAN, M. G. (1980) Pkytochemistry 19, 1433-1437. 38. NES, W. R., AND MCKEAN, M. L. (1977) Biochemistry of Steroids and Other Isopentenoids, p. 283, Univ. Park Press, Baltimore. 39. CORI, O., CHAYET, L., DE LA FUENTE, M., FERNANDEZ, L. A., HASHACEN, U., PEREZ, L. M., PORTILLA, G., ROJAS, C., SANCHEZ, G., AND VIAL, M. V. (1980) Mol. Biol. Biochem. Bioph ys. 32, 97-110. 40. LOOMIS, W. D., AND CROTEAU, R. (1973) Rec. Advan. Pkytochem. 6, 147-185. 41. VIAL, M. V., ROJAS, C., PORTILLA, G., CHAYET, L., PEREZ, L. M., CORI, O., AND BUNTON, C. A. (1981) Tetrahedron 37, 2351-2357. 42. LOPEZ, V., STEVENS, T., AND LINDQIJIST, R. N. (1976) Arch. Biochem. Biopkys. 175, 31-38. 43. CROTEAIJ, R., AND KARP, F. (1979) Arch. Biochem. Biophys. 198, 523-532. 44. CROTEAU, R., BURBOTT, A. J., ANDLOOMIS, W. D. (1973) Biochem. Biophys. Res. Commun. 50, 1006-1012. 45. SUGA, T., SHISHIBORI, T., AND MORINAKA, H. (1980) J. CiLem. Sot. Chem. Commun. 167-168. 46. AKHILA, A., AND Pflanzenphysiol.

BANTHORPE, 99, 277-282.

D. V. (1980)

2.