biochemicalcharacterization of a spearmint mutant that ...one ofthe scotch spearmint selections...

Plant Physiol. (1991) 96, 744-7520032-0889/91/96/0744/09/$01 .00/0

Received for publication September 10, 1990Accepted February 27, 1991

Biochemical Characterization of a Spearmint Mutant ThatResembles Peppermint in Monoterpene Content1

Rodney Croteau*, Frank Karp, Kurt C. Wagschal, D. Michael Satterwhite, David C. Hyatt2, andCalvin B. Skotland

Institute of Biological Chemistry, Washington State University, Pullman, Washington 99164-6340 (R.C., F.K.,K.C. W., D.M.S., D.C.H.); and Irrigated Agriculture Research and Extension Center, Washington State University,

Prosser, Washington 99350-9687 (C.B.S.)

ABSTRACT

A radiation-induced mutant of Scotch spearmint (Mentha xgraclis) was shown to produce an essential oil containing prin-cipally C3-oxygenated p-menthane monoterpenes that are typicalof peppermint, instead of the CS-oxygenated monoterpene familycharacteristic of spearmint. In vitro measurement of all of theenzymes responsible for the production of both the C3-oxygen-ated and CS-oxygenated families of monoterpenes from the com-mon precursor (-)-limonene indicated that a virtually identicalcomplement of enzymes was present in wild type and mutant,with the exception of the microsomal, cytochrome P-450-depend-ent (-)-limonene hydroxylase; the C6-hydroxylase producing (-)-trans-carveol in the wild type had been replaced by a C3-hydrox-ylase producing (--trans-isopiperitenol in the mutant. Addi-tionally, the mutant, but not the wild type, could carry out thecytochrome P-450-dependent epoxidation of the ai-unsaturatedbond of the ketones formed via C3-hydroxylation. Although pres-ent in the wild type, the enzymes of the C3-pathway that converttrans-isopiperitenol to menthol isomers are synthetically inactivebecause of the absence of the key C3-oxygenated intermediategenerated by hydroxylation of limonene. These results, whichclarify the origins of the C3- and CS-oxygenation patterns, alsoallow correction of a number of earlier biogenetic proposals forthe formation of monoterpenes in Mentha.

The monoterpene constituents of the essential oils of pep-permint (Mentha piperita L.) and spearmint (native = Menthaspicata L.; Scotch = Mentha x gracilis [28]) are distinguishedby the position of oxygenation on the p-menthane ring3 (18,25). Peppermint produces almost exclusively monoterpenesbearing an oxygen function at C3, such as menthone andmenthol, whereas spearmint species produce almost exclu-sively monoterpenes bearing an oxygen function at C6, typi-

'This investigation was supported in part by a National ScienceFoundation grant (DCB 8803504), by a grant from the WashingtonMint Commission/Mint Industry Research Council, and by Projects0268 and 0281 from the Agricultural Research Center, WashingtonState University, Pullman, WA 99164.

2 Present address: Department of Biochemistry, University of Ari-zona, Tucson, AZ 85721.

3 The numbering system employed here is based on limonene (Fig.1), the parent p-menthane monoterpene of Mentha.

fied by carvone (Fig. 1). Production of either the C3-oxygen-ated family or the C6-oxygenated type is considered to beunder the control of a single gene (25). This regiospecificityof oxygenation is established very early in the monoterpenebiosynthetic sequence where (-)-limonene, the first cyclicolefin to arise from the ubiquitous isoprenoid precursor ge-ranyl pyrophosphate, is hydroxylated exclusively at C3 to (-)-trans-isopiperitenol (in peppermint) or at C6 to (-)-trans-carveol (in spearmint) (13). The Cyt P-450-dependent mixedfunction oxygenases responsible for these transformationshave been isolated from peppermint and spearmint and par-tially characterized (12). These enzyme systems are highlyselective for (-)-limonene (and closely related analogs) andare regiospecific with respect to C3-hydroxylation (pepper-mint) or C6-hydroxylation (spearmint), a finding that hasclarified, at the enzyme level, earlier genetic evaluation of theorigins of the C3- versus C6-oxygenation pattern (6, 25). Inaddition to strict regiospecificity, these two hydroxylases arealso readily distinguished based on differential inhibition bysubstituted imidazoles (12). The remaining enzymatic stepsin the conversion of the first oxygenated products, (-)-trans-carveol in spearmint and (-)-trans-isopiperitenol in pepper-mint, to carvone and menthol isomers, respectively (Fig. 1),have also been previously described (4, 8, 14, 15).Peppermint and both native and Scotch spearmints are

sterile and are necessarily vegetatively propagated. SeriousVerticillium wilt problems have arisen with these clonal spe-cies in most of the major growing areas prompting radiationmutation programs for the production of wilt-resistant types(11). One of the Scotch spearmint selections (mutant 643)arising from this program (11) was shown, on organolepticevaluation, to produce an oil reminiscent of peppermint.Subsequent chemical analysis of this "mutant" spearmint oilrevealed the abnormal presence of largely C3-oxygenatedterpenoids, several of which also bore an epoxide function atC1-C2. In the present paper, we describe the biochemicalcharacterization of mutant 643 with reference to the Scotchspearmint wild type and provide evidence that the metabolicalteration resides at the level ofthe (-)-limonene hydroxylase.Specifically, the C6-hydroxylase in the wild type is replacedby a C3-hydroxylase (and an epoxidase) in the mutant, leadingto a greatly altered pattern of monoterpene products thatmore closely resembles peppermint.

744

www.plantphysiol.orgon April 12, 2020 - Published by Downloaded from Copyright © 1991 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org

BIOCHEMICAL CHARACTERIZATION OF A SPEARMINT MUTANT

t

H-)C_-Piperitone oxide

(-) -trans-Piperitone oxide

)-Piperitenone Piperoxide

6 0

C-) -cie-Ieopiperitenone oxide

+ 9b.

H-Piperitone l-)c18-Piperitone oxide

6

2i2!

H f 4 5 8 Dihydrocarvones

Dlihydrocerveols

-) -trens- C-)-CarvoneCerveol

(t) -Pulegone

7b4

0

(e-) -Ieoeenthone

3[i 8a * X~0

(-) -enthone C-) -Menthol

8be

W+ --fomenthmol

Figure 1. Pathways of monoterpene biosynthesis in M. x graceis wild type and mutant 643. Enzymatic activities observed only in wild type(dashed light arrows), only in mutant 643 (light arrows), and in both wild type and mutant (heavy arrows) are indicated. The numbered enzymaticsteps are described in Table II.

MATERIALS AND METHODS

Plant Materials, Substrates, and Standards

Scotch spearmint (Mentha x gracilis) and mutant number643-10-74 (produced by 10 Krad y-irradiation [60Co] to 50%rhizome mortality [11]) were obtained from C. E. Homer,Oregon State University, and were maintained under fieldconditions at the Irrigated Agriculture Research and Exten-sion Center, Prosser, WA. Plants for biosynthetic studies werepropagated from stolons under controlled conditions as pre-viously described (4). [1-3H]Geranyl pyrophosphate (90 Ci/mol) was synthesized and purified by literature procedures(5), and unlabeled substrates and all monoterpene standardswere from our own collection. The 1,2-epoxides of (-)-car-vone, (-)-isopiperitenone, piperitenone, and (+)- and (-)-piperitone were prepared by base-catalyzed epoxidation (16),and the cis-isomer and trans-isomer (major product) werepurified by TLC on silica gel G (ether:hexane, 1:1 [v/v]) fol-lowed by HPLC on a Zorbax Sil column (ether:hexane, 1 :9[v/v]). Polystyrene resin (Amberlite XAD-4) for use as an adsor-bent was obtained from Rohm and Haas, Inc. All other

reagents were purchased from Aldrich or Sigma Chemical Co.unless otherwise noted.

Oil Analysis

Steam distilled oil from field grown plants was diluted 1000-fold in hexane and analyzed by combined GLC-MS. Identi-fication was based on coincidence of retention time on twofused-silica capillary columns of widely differing polarity(RSL 150 and Superox FA) and agreement in mass spectrumwith that of the authentic standard or (for the epoxides) witha literature compilation (30).For monitoring oil composition changes under long-day

conditions (4), leaves of 0.5 to 1, 2, 3, 4, and >6 cm in lengthwere collected from the corresponding positions on the stemsof developmentally synchronized plants. Twenty leaves of thesame size were pooled and extracted with pentane:ether (2:1,v/v) in the presence ofp-cymene (0.5 mg/g tissue) as internalstandard. The extracts were dried over MgSO4, decolorizedwith activated charcoal, concentrated, and subjected to com-bined GLC-MS as before for qualitative and quantitative

745



Plant Physiol. Vol. 96, 1991

analysis (with reference to the internal standard). The analysiswas repeated three times without significant difference inanalytical results. For stem and flower oil analysis, matureplants were employed, and extracts ofthe relevant tissues werecompared to that of midstem leaves.The stereochemical resolution of (±)-piperitone and (+)-

cis-piperitone oxide was accomplished by isolation of eachcomponent from the steam-distilled oil by a combination ofTLC and HPLC (or preparative GLC) as before, and separa-tion by GLC on a chiral phase capillary column (Chirasil-L-valine) (27) with reference to the authentic enantiomers pre-pared as described above. The resolution of limonene andcarvone was accomplished by TLC isolation, the preparationof diastereomeric derivatives, and separation on a normalphase capillary column (Superox FA) as described previously(26).

Enzyme Extracts

The first two leaf pairs of developmentally similar Scotchspearmint and mutant 643 plants (4-6 leaf stage) were usedin all experiments, as the rate of terpene production was mostrapid in this tissue. Extracts of glandular trichomes (the siteof monoterpene biosynthesis [8]) were prepared by a mecha-nized procedure in which the leafsurfaces were gently abradedwith small glass beads (7). This procedure resulted in removalof more than 99% of the glandular trichomes as determinedmicroscopically and is notable for the use of XAD-4 polysty-rene resin (at 1 g/g leaf tissue) as an adsorbent of endogenousterpenoids, reducing these materials to negligible levels in theextracts (7, 8, 12). Following homogenization and filtrationof the extracts, sequential centrifugation at 3,000g, 1 8,000g,and 195,000g was employed to separate light membranesfrom the soluble protein fraction. The resulting 195,000gpellet was first washed and then resuspended in the appropri-ate buffer for assay. Depending on the assay, the solubleenzyme fraction was either concentrated by ultrafiltration(Amicon PM-30) and adjusted to assay conditions either bypassage through a desalting column (Sephadex G-50) or bydialysis (termed crude extracts), or was concentrated by(NH4)2SO4 precipitation (0-60%) and fractionated by gelpermeation chromatography on a column (Sephadex G-100or Sephacryl S-200) equilibrated with the appropriate assaybuffer (termed a partially purified preparation).

Enzyme Assays

Each of the assays outlined below, with the exception ofthe epoxidase assay, has been described in detail elsewhere (4,8, 12, 14, 15). The assay for geranyl pyrophosphate:(-)-limonene cyclase depends upon the conversion of 1-3H-labeled precursor to the olefinic product, with TLC purifica-tion following the addition of authentic carrier and verifica-tion of product identity by radio-GLC (3, 27). Nonenzymaticactivity was negligible.

All other assays utilized unlabeled substrates and employedextractive isolation of the substrate/product mixture, withcapillary GLC for separation and quantitation (using (+)-camphor as internal standard). Product verification was byGLC-MS as before. For assays with unlabeled substrates,

boiled controls and zero-time controls were included, as werecontrols without substrate to allow for background correction,if necessary. Nonenzymatic activity was negligible for allreactions except the very facile terpenone isomerization, forwhich the necessary adjustments were made. These assays(15) were run at pH 6.0 to minimize nonenzymatic activity.Although assays were run with a great excess of cofactor andwith substrate at 5 and 10 times the presumptive Km value toensure linearity over the course of the assay, it should benoted that the assays were developed from studies with pep-permint and native spearmint and were not necessarily opti-mum for Scotch spearmint. As such, activity levels may beunderestimated for Scotch spearmint and mutant 643. Sincean assay for terpenone epoxidase activity had not been pre-viously developed, it seemed prudent to adopt the assayconditions employed for the microsomal Cyt P-450-depend-ent limonene hydroxylase (12).Each assay was run at least thrice with independent prepa-

rations from developmentally similar tissue, and assays withboth wild type and mutant were run concurrently. Activitydeterminations, expressed as rates in pkatals (pmol * s-'), wereall within ±20% of the mean and are provided as simpleaverages of the number of assays run.

Analytical Procedures

Detailed protocols for TLC, radio-GLC, combined GLC-MS, and scintillation spectrometry have been provided (3, 4,15, 27). The Chirasil-Val III GLC column (Alltech) was 0.32mm i.d. x 50 m fused silica.

RESULTS

Essential Oil Analysis

Organoleptic evaluation of the essential oil of the wilt-tolerant 643 selection of Scotch spearmint (M. x gracilis)indicated that this mutant produced an atypical spearmint oilwhich was reminiscent of, but quite distinct from, the oil ofpeppermint (M. x piperita). Comparable plots of 2-year-oldScotch spearmint and mutant 643, grown on wilt-free soil inProsser, WA, were harvested on the same day (midsummer)and steam distilled. Hay weight per plot was comparable forboth mints, but the distilled oil on a per area or kilogram oftissue basis was about 20% less for the mutant. Analysis ofthe essential oil from both mints by GLC-flame ionizationdetection and GLC-MS revealed the presence of nearly 50components, ofwhich about a dozen were present at or above0.5% of the oil (Table I). The oil of Scotch spearmint wastrue to type in containing approximately 15% (-)-limoneneand 70% (-)-carvone (18); the optical purities of these prod-ucts were shown to be in excess of 97% by chromatographicanalysis of diastereomeric derivatives (26). Although organo-leptic evaluation of the oil of mutant 643 had suggested anabnormal composition, the analytical results were neverthe-less surprising in that over 85% of the oil was composed ofC3-oxygenated monoterpenes (Table I), of which well overhalf were 3-keto-1,2-oxides (Fig. 1). Piperitone oxides andpiperitenone oxide have been previously reported to occur inthe oils of other Mentha species, including M. sylvestris, M.longifolia, and M. rotundifolia, and certain M. arvensis and

746 CROTEAU ET AL.




Table I. Essential Oil Analysis of Scotch Spearmint Wild Type andMutant 643

The oil from field-grown plants was steam distilled and analyzedby GLC-MS as described under "Materials and Methods." Minoramounts of nonterpenoid compounds, such as octanol, were alsonoted, but are not tabulated.

Component Wild Type Mutant 643mg/g tissue (% of total oil)

Pinenes and other olefins 0.34 (2.5) 0.21 (2.0)Limonene 2.12 (15.6) 0.26 (2.5)1,8-Cineole 0.24 (1.8) 0.10 (1.0)Dihydrocarvones 0.15 (1.1) tr.8Carveols and dihydrocarveols 0.34 (2.5) tr.

(and acetates)Carvone 9.33 (68.6) 0.16 (1.5)Menthone 0.14 (1.0) 0.64 (6.2)Isomenthone tr. 0.15 (1.4)Menthol tr. tr.Neomenthol tr. 0.12 (1.2)Pulegone tr. 1.50 (14.4)Piperitone tr. 0.24 (2.3)cis-Pipentone oxide tr. 0.20 (1.9)trans-Piperitone oxide tr. 0.89 (8.6)Piperitenone oxide tr. 5.52 (53.1)Sesquiterpenes 0.34 (2.5) 0.23 (2.2)Total 13.60 (96.6) 10.40 (97.3)

a Trace level observed.

M. cordifolia chemotypes (10, 18), but the presence of thesecompounds and their C3-oxygenated congeners in spearminttypes is unusual (17, 22).A developmental study of oil compositional change as a

function of leaf maturity of plants grown under controlledconditions indicated that the expansion of Scotch spearmintleaves from 0.5 cm to over 5 cm was accompanied by a

gradual decrease in limonene content from over 20% to about15%, with a gradual increase in carvone content from 50% toabout 70%. During this interval, oil content per leaf increasedroughly 10-fold. Compositional changes in the essential oil ofmutant 643 as a function of leaf development were moredramatic. As the leaves expanded, the oil content per leafincreased about 8-fold, and was accompanied by an increasein both piperitenone oxide and pulegone (often to greaterthan 65% and 20%, respectively) and a corresponding de-crease in all other C3-oxygenated types (piperitone oxides,menthone, etc.). The limonene level was highest in youngleaves (--10%) and decreased to under 2.5% in mature tissue.At this stage, C3-oxygenated derivatives of limonene com-prised well over 90% of the oil and 1,2-oxides were veryprevalent. Oil yields produced by the mutant propagatedunder controlled conditions were less than those observed inthe wild type over the course of development; often yields of50% less on a per fresh weight or per leaf basis were notedthroughout leaf maturity, and until the time of yield decline(leaves >6 cm). Stem oil of the mutant, although of loweryield than leaf oil, was essentially identical in composition.The oil extracted from flowers was qualitatively similar to leafoil, but contained somewhat more pulegone and less terpeneoxides than the oil of mature leaves.

The composition of distilled oil from mutant 643 grown inthe field, and the compositional alterations noted in theextracted oil from plants propagated under controlled condi-tions, have remained essentially unchanged for the 4 yearsover which this mutant has been observed. On the basis ofthese analytical studies, it thus appears that the biosynthesisof essential oil terpenes in the 643 mutant has been drasticallyaltered relative to the wild-type Scotch spearmint.

Enzyme Activity Measurements

To determine the metabolic basis for the altered oil com-position of mutant 643, it was necessary to evaluate differ-ences in the monoterpene biosynthetic pathways of the mu-tant and the Scotch spearmint progenitor. In vivo studies todetermine pathway alternatives can be equivocal, and wechose to examine this problem at the cell-free enzyme levelsince all of the relevant activities, with the exception of thepresumptive epoxidation step(s), have been demonstrated incell-free extracts of spearmint or peppermint and reliableassays have been developed (4, 8, 12, 14, 15). All enzymepreparations were derived from leaf epidermis extracts (7)since the glandular trichomes appear to be the exclusive sitesof monoterpene biosynthesis and the relevant enzymes thusfar examined are restricted to these structures (7, 8). Theresults of these enzyme-level studies are tabulated (Table II),and each enzymatic step is described below in sequence,starting with the formation of (-)-limonene which is theparent olefin for monoterpene biosynthesis in both pepper-mint and spearmint species (13).

Geranyl pyrophosphate:(-)-limonene cyclase (enzyme 1,Fig. 1) was obtained in the 195,000g supernatant of the leafgland extracts (the only other products formed from geranylpyrophosphate in these crude systems were trace levels ofbicyclic olefins and 1,8-cineole, and geraniol formed by theaction ofphosphatases), and each extract was partially purifiedby chromatography on Sephacryl S-200. In preparations fromboth wild type and mutant, only a single olefin cyclase activitywas observed and the major product was limonene (>98%)as determined by radio-GLC. Resolution of this biosyntheticolefin produced by extracts of both Scotch mint and mutant,by the preparation and separation of diastereomeric deriva-tives (26), indicated the presence of >97% (-)-(4S)-limonenein each case. The activity from both Scotch spearmint andmutant preparations eluted in the region corresponding to amol wt of 55,000, which is characteristic of the (-)-limonenecyclase from other Mentha species (M. x piperita, M. spicata)(J Rajaonarivony, R Croteau, unpublished data). Activitylevels for Scotch spearmint and mutant were roughly com-parable, and not unlike those observed in gland extracts fromM. piperita and M. spicata leaves at a similar stage ofdevelopment.The Cyt P-450-dependent (-)-limonene hydroxylase activi-

ties (enzymes 2 and 3) were examined in microsomal prepa-rations from epidermal gland extracts ofboth mint types. The02/NADPH-requiring reaction in Scotch spearmint micro-somes gave rise exclusively to (-)-trans-carveol, and (-)-trans-isopiperitenol was not detected as a product, indicating thepresence of only (-)-limonene-6-hydroxylase (12). Con-versely, microsomes obtained from the mutant, when incu-

747




Table II. Enzymes of Monoterpene Biosynthesis in Extracts ofScotch Spearmint Wild Type and Mutant 643

The preparation and assay of each enzyme is described under"Materials and Methods" and each reaction catalyzed is illustrated inFigure 1.

Activity in GlandEnzyme Enzyme Name ExtractsNumber

Wild type Mutant 643

pkat/g tissue

1 Geranyl pyrophosphate:(--Limonene cyclase 20 17

2 (--Umonene-6-hydroxylase 6 NDa3 (-)Limonene-3-hydroxylase ND 54 Terpenol dehydrogenase with:

(-)trans-Carveol 11 13(--trans-lsopiperitenol 30 35

5 Terpenone-1'2-reductase with:(-)lsopiperitenone 20 17Piperitenone ND ND(±)-Piperitone ND ND(-)-Carvone ND ND

6 Terpenone isomerase with:(-)-isopiperitenone (or oxide) 164 116(+)-cis-lsopulegone 1176 952

7 Terpenone-A4'8-reductase:a (+-Pulegone to (--men- 165 48

thoneb (+)-Pulegone to (+)-isomen- 56 15

thonea Piperitenone to (+-piperitone 16 6b Piperitenone to (--piperitone 15 6a (+-Piperitenone oxide to (-- 87 22

trans-piperitone oxideb (+)Piperitenone oxide to (-)- 34 7

cis-piperitone oxide8 Menthone reductase to:a (-)Menthol 1 1b (+)Neomenthol 16 13

9 Terpenone-A1'2-epoxidase with:(--lsopiperitenone ND 3Piperitenone ND 1(+)Piperitone ND 3(--Piperitone ND 3

a Activity not detectable (<0.1 pkat/g tissue).

bated with (-)-limonene, produced only (-)-trans-isopiperi-tenol, and (-)-trans-carveol was not detected as a product,indicating the exclusive presence of a (-)-limonene-3-hydrox-ylase (12). Pinenes, although present in minor quantities inboth oils, did not serve as hydroxylase substrates in eithercase, as expected (12). Although the measured level of limo-nene-6-hydroxylase in Scotch spearmint appeared to beslightly higher than the level of limonene-3-hydroxylase inthe mutant (Table II), these values should be taken with somecaution since the isolation procedure and assay conditionswere adopted from studies on the corresponding limonenehydroxylases from M. spicata and M. piperita (12), and havenot yet been optimized for M. x gracilis. It is also importantto note that the oil ofM. x gracilis (Table I) contains a minoramount of menthone (a product arising from C3-hydroxyl-

ation and the reactions described below) and that the oil ofthe 643 mutant contains a low level of carvone (a productarising from C6-hydroxylation and oxidation of the resultingalcohol), indicating that the corresponding hydroxylases arelikely present but at levels much too low to be detected bythe assay.

Terpenol dehydrogenase activity (enzyme 4) was obtainedin the soluble supernatants of surface gland extracts (7) ofboth plant types, and was partially purified by a combinationof(NH4)2SO4 precipitation and chromatography on SephacrylS-200. Dehydrogenase activity, measured with both (-)-trans-isopiperitenol and (-)-trans-carveol as substrates (the cis-isomers were not substrates), eluted as a single peak in theregion corresponding to a mol wt of 62,000 for extracts fromboth plant types. The terpenol dehydrogenase from nativespearmint (M. spicata) is of identical mol wt and, althoughappearing to be a single species by gel permeation chromatog-raphy, is known to be comprised of two electrophoreticallydistinguishable proteins (8). Since the activity levels for bothsubstrates were similar for preparations from both plant types(Table II), the question was not pursued further in this in-stance. Both dehydrogenase activities were present at levelsgreater than the activities of the preceding enzymatic steps(the hydroxylases), probably accounting for the lack of signif-icant accumulation of either carveol or isopiperitenol in therespective essential oils.Terpenone isomerase (enzyme 6) has been previously stud-

ied in peppermint leaf extracts (4, 15). The activity is easilyassayed (by GLC) in crude, soluble protein preparations andboth (-)-isopiperitenone and (+)-cis-isopulegone are utilizedas substrates. Oil gland extracts from both M. x gracilis planttypes were shown to contain high levels of isomerase activity,especially with cis-isopulegone as substrate (Table II). Neithercarvone nor trans-isopulegone was a substrate for the isomer-ase from either source.Terpenone-A' 2-reductase (enzyme 5) was next examined.

This enzyme is extremely active in peppermint oil glandextracts, and is responsible for the reduction of (-)-isopiperi-tenone to (+)-cis-isopulegone, which is subsequently isomer-ized to (+)-pulegone in this species (4). Assays were conductedusing oil gland extracts from both M. x gracilis and mutant643, and the isopiperitenone reductase activity, by severalindependent experiments, was shown to be similar in the twoplant types, and relatively low when compared to extracts ofpeppermint (4). However, the relative level of A1'2-reductasecompared with isomerase activity in the mutant (as indicatorsof in vivo activities) is entirely consistent with the morefavorable partitioning of (-)-isopiperitenone toward piperi-tenone and its derivatives (oxides), rather than toward (+)-cis-isopulegone and its derivatives (pulegone and menthone),as evidenced from oil analysis (Table I). Piperitenone andpiperitone were not substrates for the A",2-reductase, as ob-served previously for the peppermint enzyme (4), nor was(-)-carvone detectably reduced, consistent with the low levelsof dihydrocarvone observed in the essential oil of the wildtype.The reduction of the A4'8-double bond of the various con-

jugated terpenones was next evaluated. This NADPH-de-pendent reductase activity in peppermint (1, 4) is thought tobe composed of two distinct enzyme species of differing

CROTEAU ET AL.748




stereospecificity: one for the reduction of the A4'8-bond to the4S-isopropyl function [e.g. to (-)-menthone by enzyme 7a];the other for the reduction of the A4'8-bond to the epimeric4R-isopropyl function [e.g. to (+)-isomenthone by enzyme7b] (20, 24). The two activities were not separated here, butwere assayed simultaneously (4); both reductase activities wereconsistently lower (on a fresh weight basis) in leaves of themutant than of the wild type. The reduction of (+)-pulegoneto (-)-menthone proceeded at roughly three times the rate ofreduction of (+)-pulegone to (+)-isomenthone in both Scotchmint and mutant, consistent with the greater abundance ofmenthone over isomenthone in the oil of selection 643.Piperitenone also served as a substrate for A4'8-reduction topiperitone. Resolution of the naturally occurring product byGLC separation on a chiral phase capillary column (27)demonstrated the presence of a nearly racemic mixture (Fig.2), thus suggesting nearly equivalent rates of 4S- versus 4R-reduction to give (+)-(4S)-piperitone and (-)-(4R)-piperitone,respectively. (+)-(4S)-Piperitone is the enantiomer most com-

o

0

D _l

30 34 38

Time (min)Figure 2. Gas chromatographic resolution of piperitone (A) and cis-piperitone oxide (C) from the oil of mutant 643. The relevant com-pounds were isolated by TLC and HPLC, and separated by capillaryGLC on the Chirasil-Val IlIl column programmed from 450C (5 min) at2°C/min to 1 750C for piperitone, and from 700C (5 min) at 1 0°C/minto 1750C for cis-piperitone oxide. Panels B and D illustrate theseparations of racemic piperitone and racemic cis-piperitone oxide,respectively. Elution order was verified by separate injection of opti-cally pure standards, and the identity of the enantiomers was con-firmed by GLC-MS on the chiral phase column.

monly reported in Mentha oils (18). The A4'8-reductase wasinactive with (-)-carvone or (-)-isopiperitenone as substrate(or with dihydrocarvone, carveol or isopiperitenol which alsobear the isopropenyl (A8'9) rather than the isopropylidene(A4'8) moiety). Indeed, A8'9-reduction was not observed evenin crude (unfractionated) extracts with either NADH orNADPH as reductant.

Since it seemed possible that the cis- and trans-piperitoneoxides could arise by reduction of piperitenone oxide, ratherthan by reduction of piperitenone and epoxidation of theresulting piperitones, the A4'8-reduction of (+)-( lS:2S)-piper-itenone oxide (isolated from the oil of mutant 643) wasevaluated. With these preparations, the reduction of piperi-tenone oxide was quite rapid, giving a mixture of (-)-(lS:2S:4R)-cis- and (-)-(1S5:2S:4S)-trans-piperitone oxides asproducts, with the (-)-trans-isomer predominating by a factorof about three (corresponding roughly to the ratio observedin the oil of the mutant 643; cf. Tables I and II). Resolutionof the naturally occurring cis-piperitone oxide by chromato-graphic separation on the chiral capillary column (the trans-oxide was not resolvable by this method) did verify theessentially exclusive presence of the (-)-cis-enantiomer (Fig.2). By inference, the trans-oxide would also be the (-)-antipode (Fig. 1).Two NADPH-dependent, stereospecific (-)-menthone re-

ductases (dehydrogenases 8a and 8b) have been previouslydescribed in peppermint leaf extracts; one for the productionof (-)-menthol, the other for the production of the epimericalcohol (+)-neomenthol (14). Both were examined simulta-neously in gland extracts of M. x gracilis and mutant 643.The reduction to (-)-menthol was detectable in extracts ofboth plants (Table II), whereas reduction to (+)-neomentholwas observed at a much higher level, thus rationalizing thepresence of neomenthol in the oil of the mutant (Table I).(+)-Isomenthone also serves as a substrate for these reductases(at roughly comparable rates), but the unsaturated ketonesare poor substrates (data not shown). Nevertheless, theseenzymes probably account for the conversion of dihydrocar-vone to dihydrocarveols which are observed as minor com-ponents in the oil of Scotch spearmint.Terpenone epoxidase activity (enzyme 9) has not previously

been reported in mint extracts, but the assumption was madethat this activity would be microsomal and NADPH/02-dependent, as with the limonene hydroxylases (12). Such anNADPH/02-dependent epoxidase activity was, in fact, readilydetected in microsomal preparations from the epidermalglands of mutant 643, but not in extracts from Scotch spear-mint (Table II). Over 70% of the epoxidase activity waslocated in the light membrane fraction (as was the C3-hy-droxylase of the mutant), and the activity was inhibited byCO (-70% inhibition at a CO:02 ratio of 9:1) indicating theinvolvement of cytochrome P450 (29). The soluble enzymefraction of both Scotch spearmint and mutant 643 prepara-tions was devoid of epoxidase activity when assayed underidentical conditions. With (-)-isopiperitenone as a substrate,A1"2-epoxidation to (-)-cis-isopiperitenone oxide (Fig. 1) wasreadily confirmed by MS [m/z 166 (P+, 2%), 96 (100), 67(70)] (Table II). With piperitenone as substrate, the productwas assumed to be the naturally occurring (+)-(15S:2S)-piper-itenone oxide (Fig. 1) based on the stereochemical result with

749




isopiperitenone. This enantiomer is characteristic of Mentha(18), and the absolute configuration of the epoxide functionwas additionally fixed by the previously mentioned resolutionof the naturally occurring cis-piperitone oxide (Fig. 2), whichbore the IS:2S-epoxide. With both (+)-(4S)- and (-)-(4R)-piperitone as substrates, only the corresponding cis-piperitoneoxides were produced (Table II). Because the dominatingepimer in the oil of mutant 643 is (-)-(lS:2S:4S)-trans-piperitone oxide (Table I), the epoxidation of piperitone doesnot offer a viable route to this metabolite, which must arise,predominantly, by the 4S-A4'8-reduction of (+)-(lS:2S)-piper-itenone oxide (Fig. 1). Additionally, and in spite of the factthat the enzymatic epoxidation of (+)- and (-)-piperitone isreadily detectable, the cis-piperitone oxide present in the oilof mutant 643 was shown to be comprised largely of the (-)-cis-(lS:2S:4R)-enantiomer, thus limiting the origin of thisminor metabolite to the lS:2S-epoxidation of (-)-(4R)-piper-itone (or to the 4R-(A4'8)-reduction of (+)-(lS,2S)-piperiten-one oxide).

(-)-Carvone and (+)-pulegone were not substrates for epox-idation, indicating that the A1,2- and A89-bonds of carvoneand the A4'8-bond of pulegone are refractory to this reactiontype in these systems. A minor product, tentatively identifiedas 3-hydroxycarvone by GLC-MS analysis, was observedwhen microsomal preparations from mutant 643 were incu-bated with carvone under the conditions of the hydroxylase/epoxidase assay (NADPH plus 02). Although isopiperitenonewas readily epoxidized by preparations from mutant 643, thecorresponding terpenol, isopiperitenol, was not an epoxidasesubstrate, presumably because of the stereoelectronic influ-ence of the adjacent hydroxyl function.

Based on compositional analysis, stereochemical evalua-tion, and in vitro assay of the relevant activities, the primarypathway for (-)-limonene metabolism in mutant 643 can beformulated as C3-hydroxylation to (-)-trans-isopiperitenoland oxidation to (-)-isopiperitenone (Fig. 1). Reduction ofthis central intermediate at the A",2-position and isomerizationof the A8 9-double bond yields (+)-pulegone, from which (-)-menthone and (+)-isomenthone and their congeners, typicalof peppermint, are formed. However, the major route for theconversion of (-)-isopiperitenone seemingly involves epoxi-dation to isopiperitenone oxide and rapid isomerization topiperitenone oxide, or direct isomerization to piperitenonewhich is then entirely transformed to the 1,2-epoxide. [Therate of enzymatic isomerization of cis-isopiperitenone oxideto piperitenone oxide was difficult to determine accuratelysince the chemical isomerization was so facile. Nevertheless,it is safe to assume that the enzymatic isomerization of (-)-cis-isopiperitenone oxide to (+)-piperitenone oxide is as rapidas the conversion of (-)-isopiperitenone to piperitenone(Table II)]. Reduction of (+)-piperitenone oxide subsequentlyproduces both (-)-trans- and (-)-cis-piperitone oxides, withthe trans-isomer predominating. A relatively small fraction ofpiperitenone (-5%) would appear to be reduced directly to(+)- and (-)-piperitone, in nearly equal amounts, with aportion of the (-)-isomer undergoing subsequent epoxidationto (-)-cis-piperitone oxide. As indicated by the results inTable II, most of the enzymatic machinery required for themetabolism of C3-oxygenated p-menthane monoterpenes ap-pears to be present in wild-type Scotch spearmint, but is

inactive because of the absence of the C3-hydroxylase whichproduces (-)-isopiperitenol.

DISCUSSION

In comparing enzyme activities between M. x gracilis(Scotch spearmint) wild-type and the wilt-tolerant mutant643, the most striking observation is that the (-)-limonene-6-hydroxylase in the wild-type has been replaced almost en-tirely by a (-)-limonene-3-hydroxylase in the mutant thatleads to the exclusive production of C3-oxygenated p-men-thane monoterpenes. It is thought that the oxygenation pat-tern on the p-menthane skeleton is under the control of asingle gene in Mentha species and that the production of C6-oxygenated terpenes is the dominant trait (18, 25). Our datashow that irradiation has disabled C6-hydroxylation in mu-tant 643; however, the mechanism by which defective C6-hydroxylation was apparently "replaced" by the "peppermint-type" C3-hydroxylation system is not known. To determineif the limonene-3-hydroxylases from mutant 643 and pepper-mint were, in fact, similar, substrate and product specificityof the two enzymes, and their relative sensitivities to substi-tuted azole inhibitors, were compared, as these are the bestcriteria now available to distinguish the Mentha hydroxylasesystems (12). The results (Table III) indicate that, based onthese criteria, the C3-hydroxylase in mutant 643 closely re-sembles that ofpeppermint. Thus, the mutation has generateda phenotype similar to that described for a homozygousrecessive that lacks C6-oxygenated products and produces theC3-oxygenation pattern (18, 25). The limonene-6-hydroxyl-ases from native and Scotch spearmint were also compared(Table III) and, not surprisingly, were shown to be very muchalike.The second notable feature of the 643 mutant is the ap-

pearance of an epoxidase activity that is not present in thewild type or in either native spearmint or peppermint. Sinceit seems extremely unlikely that mutant 643 bears two alteredCyt P-450 systems, one for C3-hydroxylation and another forCl,C2-epoxidation, we suggest that the same modified oxy-genase catalyzes both reactions. That hydroxylation and epox-idation are functions of the same oxygenase system is sup-ported by the identical recovery of both activities in the lightmembranes (73 ± 2%), by the identical kinetics of CO inhi-bition (70 ± 5% inhibition at CO:02 of 9:1), and by theidentical inhibition of both activities by clotrimazole (18.5 +2 gM). Additionally, the stereochemistry of oxygen insertionin both reactions is consistent with the same protein catalyst(i.e. oxygen addition to the "backside" of the cyclohexanoidring as illustrated in Fig. 1). There is precedent for multiplemodes of oxygenation by the same Cyt P-450 in hepaticmicrosomal systems (2, 19). Evaluation ofthis novel, possiblybifunctional, Cyt P-450 system, and its relationship to theC3-hydroxylase system of peppermint that lacks epoxidaseactivity, will require solubilization and purification of theenzyme from microsomes of the mutant plants.

Finally, in deciphering the origins of the C3- versus C6-oxygenation pattern, we have confirmed that the criticaldeterminant of the pathway is the hydroxylation step in thetransformation of the common precursor (-)-limonene. Wealso demonstrated that essentially all of the remaining en-

750 CROTEAU ET AL.




Table 1II. Comparison of Substrate Specificity and Inhibitor Sensitivity of Mint HydroxylasesRelative rates for each hydroxylase system are based on the conversion of (-)limonene set at 100. Dihydrolimonene = p-menth-1 -ene.

Terpinolene is the A4,8-isomer of limonene. Protocols for the inhibitor experiments are provided in reference 12.C6-Hydroxylase C3-Hydroxylase

SubstrateNative spearmint Scotch spearmint Mutant 643 Peppermint

product (relative rate)

(--Limonene (-)trans-Carveol (-)trans-Carveol (-)trans-lsopiperitenol (-)trans-lsopiperitenol(100) (100) (100) (100)

(+)Limonene (+)-cis-Carveol (+)-cis-Carveol (+)trans-lsopiperitenol (+)trans-lsopiperitenol(88) (86) (53) (50)

(--8,9-Dihydrolimonene (--trans-Carvotanacetol (--trans-Carvotanacetol (+)trans-Piperitol (+)trans-Piperitol(76) (67) (26) (37)

(+)8,9-Dihydrolimonene (+)-cis-Carvotanacetol (+)-cis-Carvotanacetol (--trans-Piperitol (-)trans-Pipentol(66) (71) (27) (37)

Terpinolene NPDa NPD NPD NPD

Inhibitor ISOJAM

Miconazole 14 10 280 405Clotrimazole 18 10 20 17

a No product detected.

zymes of monoterpene metabolism, following hydroxylation,were present in both peppermint and spearmint, but pheno-typically silent in the latter because of the absence of the C3-hydroxylated intermediate. These findings, along with earlierwork (4, 13, 15), have shown that there are a number of errorsin several biogenetic schemes for the origin ofMentha terpen-oids that were based on chemical analysis of chemotypes andhybrids (see ref. 18 for a summary of this work). A proposal(10) that both C6- and C3-oxygenated p-menthanes are de-rived from a common dioxygenated precursor (a diospheno-lene) is clearly untenable. Similarly, proposals (9, 21, 23, 25)that the C3-oxygenated series is formed by the conversion ofterpinolene to piperitenone are incorrect. The proposal thatthe Mentha A gene is needed for the conversion of piperiten-one to pulegone (20, 24) can now be seen as a requirementfor the reduction of isopiperitenone to isopulegone, followedby isomerization of the latter to pulegone (Fig. 1) (4).

These results with Scotch spearmint and mutant 643, andthe comparison with the enzyme content of peppermint,illustrate the limitations of biogenetic schemes proposed onthe basis of genetic crossing experiments and oil compositionalone. The need for biochemical analysis is particularly crucialwhere key intermediates, such as isopiperitenone and isopule-gone, do not accumulate, and where entire pathways mayappear to be absent due to the lack of a single, early inter-mediate (such as isopiperitenol via C3-hydroxylation). How-ever, considerable caution must be exercised in interpretingsuch in vitro enzyme activity measurements, particularly inattempting to decipher the flux through alternate pathwayswhen enzyme extraction efficiency and optimum assay con-ditions are unknown, and when in situ metabolite (interme-diate) concentrations are sufficiently uncertain to precludeaccurate kinetic evaluation. Nevertheless, when two similartissues are extracted and assayed by the same procedure, validcomparisons can be made, particularly when the differencesin apparent enzyme activity levels are large. This approach

offers the only definitive means of establishing the presenceor absence of a given pathway for the biosynthesis of mono-terpene natural products.

ACKNOWLEDGMENTS

We thank Greg Wicheins for raising the plants, Karen Maertensfor typing the manuscript, and Dr. Donald Roberts, Plant Technol-ogies, Inc., for helpful discussion.

LITERATURE CITED

1. Battaile J, Burbott AJ, Loomis WD (1968) Monoterpene inter-conversions: metabolism ofpulegone by a cell-free system fromMentha piperita. Phytochemistry 7: 1159-1163

2. Capdevila J, Saeki Y, Falck, JR (1984) The mechanistic pluralityof cytochrome P450 and its biological ramifications. Xenobi-otica 14: 105-118

3. Croteau R, Cane DE (1985) Monoterpene and sesquiterpenecyclases. Methods Enzymol 110: 383-405

4. Croteau R, Venkatachalam KV (1986) Metabolism of monoter-penes: demonstration that (+)-cis-isopulegone, not piperiten-one, is the key intermediate in the conversion of (--isopiper-itenone to (+)-pulegone in peppermint (Mentha piperita). ArchBiochem Biophys 249: 306-315

5. Davisson VJ, Woodside AB, Poulter CD (1985) Synthesis ofallylic and homoallylic isoprenoid pyrophosphates. MethodsEnzymol 110: 130-144

6. Fujita Y (1960) Problems in the genus Mentha (III). Koryo 59:41-42

7. Gershenzon J, Duffy MA, Karp F, Croteau R (1987) Mechanizedtechniques for the selective extraction of enzymes from plantepidermal glands. Anal Biochem 163: 159-164

8. Gershenzon J, Maffei M, Croteau R (1989) Biochemical andhistochemical localization of monoterpene biosynthesis in theglandular trichomes of spearmint (Mentha spicata). PlantPhysiol 89: 1351-1357

9. Hefendehl FW, Murray MJ (1972) Changes in monoterpenecomposition in Mentha aquatic produced by gene substitu-tion. Phytochemistry 11: 189-195

10. Hendriks H, Van Os FHL (1976) Essential oil oftwo chemotypesof Mentha suaveolens during ontogenesis. Phytochemistry 15:1127-1130

751




11. Horner CE, Melouk HA (1977) Screening, selection and evalu-ation ofirradiation induced mutants ofspearmint for resistanceto Verticillium wilt. In Induced Mutations Against Plant Dis-eases. International Atomic Energy Proceedings, Vienna, pp253-262

12. Karp F, Mihaliak CA, Harris JL, Croteau R (1989) Monoterpenebiosynthesis: specificity of the hydroxylations of (-)-limoneneby enzyme preparations from peppermint (Mentha piperita),spearmint (Mentha spicata) and perilla (Perilla frutescens)leaves. Arch Biochem Biophys 276: 219-226

13. Kjonaas R, Croteau R (1983) Demonstration that limonene isthe first cyclic intermediate in the biosynthesis of oxygenatedp-menthane monoterpenes in Mentha piperita and other Men-tha species. Arch Biochem Biophys 220: 79-89

14. Kjonaas R, Martinkus-Taylor C, Croteau R (1982) Metabolismof monoterpenes: conversion of l-menthone to i-menthol andd-neomenthol by stereospecific dehydrogenases from pepper-mint (Mentha piperita) leaves. Plant Physiol 69: 1013-1017

15. Kjonaas RB, Venkatachalam KV, Croteau R (1985) Metabolismof monoterpenes: oxidation of isopiperitenol to isopiperiten-one, and subsequent isomerization to piperitenone by solubleenzyme preparations from peppermint (Mentha piperita)leaves. Arch Biochem Biophys 238: 49-60

16. Klein E, OhloffG (1963) Der stereochemische Verlauf der alkal-ischen Epoxydation von a,f3-ungesattigten Carbonylverbin-dungen der cyclischen Monoterpenreihe. Tetrahedron 19:109 1-1099

17. Kokkini S, Vokou D (1989) Mentha spicata (Lamiaceae) chem-otypes growing wild in Greece. Econ Bot 43: 192-202

18. Lawrence BM (1981) Monoterpene interrelationships in theMentha genus: a biosynthetic discussion. In BD Mookheijee,CJ Mussinan, eds, Essential Oils. Allured, Wheaton, IL, pp 1-81

19. Licht HJ, Madyastha KM, Coscia CJ, Krueger RJ (1980) Com-parison of plant and hepatic cytochrome P-450-dependentmonoterpene monoxygenases. In MJ Coon, AH Conney, RW

Estabrook, HV Gelboin, JR Gillette, PH O'Brien, eds, Micro-somes, Drug Oxidations and Chemical Carcinogenesis. Vol 1.Academic Press, New York, pp 211-215

20. Lincoln DE, Murray MJ (1978) Monogenic basis for reductionof (+)-pulegone to (-)-menthone in Mentha oil biogenesis.Phytochemistry 17: 1727-1730

21. Lincoln DE, Murray MJ, Lawrence BM (1986) Chemical com-position and genetic basis for the isopinocamphone chemotypeof Mentha citrata hybrids. Phytochemistry 25: 1857-1863

22. Misra LN, Tyagi BR, Thakur RS (1989) Chemotype variationin indian spearmint. Planta Med 55: 575-576

23. Murray MJ, Hefendehl FW (1973) Changes in monoterpenecomposition ofMentha aquatic produced by gene substitutionfrom a high limonene strain ofM. citrata. Phytochemistry 12:1875-1880

24. Murray MJ, Lincoln DE, Hefendehl FW (1980) Chemogeneticevidence supporting multiple allele control of the biosynthesisof (-)-menthone and (+)-isomenthone stereoisomers in Men-tha species. Phytochemistry 19: 2103-2110

25. Murray MJ, Reitsema RH (1954) The genetic basis of theketones, carvone and menthone in Mentha crispa L. J AmPharm Assoc Sci Ed 43: 612-613

26. Satterwhite DM, Croteau R (1987) Resolution of monoterpeneenantiomers by gas chromatography. J Chromatogr 407: 243-252

27. Satterwhite DM, Croteau R (1988) Applications of gas chroma-tography to the study of terpenoid metabolism. J Chromatogr452: 61-73

28. Tucker AO, Fairbrothers DE (1990) The origin of Mentha xgracilis (Lamiaceae). I. Chromosome numbers, fertility, andthree morphological characters. Econ Bot 42: 183-213

29. West CA (1980) Hydroxylases, monooxygenases, and cyto-chrome P-450. In DD Davies, ed, The Biochemistry of Plants,Vol 2. Academic Press, New York, pp 317-364

30. Yukawa Y, Ito S (1973) Spectral Atlas of Terpenes and RelatedCompounds. Hirokawa Publ. Co., Tokyo

752 CROTEAU ET AL.



biochemicalcharacterization of a spearmint mutant that ...one ofthe scotch spearmint selections...

Documents