CHAPTER 1
Biosynthesis of C5 — C20Terpenoid Compounds
BY J. R. HANSON
1 Introduction
This chapter for 1972 follows the pattern oflast year's Report. During the year a number of reviews have appeared which discuss different aspects of terpenoid biosynthesis.
2 Mevalonic Acid
The enzyme 3-hydroxy-3-methylglutaryl coenzyme A reductase catalyses the two-step reduction of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) to mevalonic acid using NADPH. The hemithioacetal addition compound of mevaldic acid and coenzyme A is a possible intermediate. Both steps involve hydrogen transfer from the 4a (4R) position of NADPH. The mevaldic acid-coenzyme A hemithioacetal addition compound is a good substrate for HMG-CoA reductase. The hydrogen atom which is transferred in this step has been shown to appear at the 5-pro-S-position in the resulting mevalonic acid. This is in contrast to the stereochemistry of mevaldate reductase in which a 5-pro-R hydrogen atom is introduced. (see Scheme 1)
Cell-free extracts and acetone powder preparations from Agave americana have been shown to phosphorylate mevalonic acid to give phosphomevalonic acid and thence pyrophosphomevalonic acid at an optimum pH of 7.0. Glutathione and mercaptoethanol enhance the activity of these preparations. Tracer studies with tissue cultures derived from Tanacetum vulgare have revealed 8 the formation of phosphomevalonic acid, pyrophosphomevalonic acid, isopentenyl pyrophosphate, dimethylallyl pyrophosphate, and the incorporation of mevalonic acid into monoterpenes. However, the mono-terpene components of the tissue culture differed from those of the whole plant, sabinene being formed rather than isothujone.
Artificial substrates have been studied in systems that form squalene and sterols. Incubation of trans-3-methyl[1,1-3H2]pent-2-enyl pyrophosphate and [1-14C]isopentenyl pyrophosphate with rat liver homogenates gave 1-methyl-squalene and 1,24-dimethylsqualene. However, the sterol fraction contained only 27-methyl-lanosterol and 27-methylcholesterol, indicating greater selectivity in either the epoxidation or cyclization stages of this biosynthesis. The substrate specificity of farnesyl pyrophosphate synthetase from pumpkin fruit has also been studied 10 with artificial allylie pyrophosphates. trans-3-Methyl-undec-2-enyl-, trans-3-methyldodec-2-enyl-, and trans-3-methyltetradec-2-enyl-pyrophosphates were assayed in the enzymatic reaction with isopentenyl pyrophosphate. The tetradecenyl pyrophosphate was inactive. Replacement of the methyl group by ethyl, as in trans-3-ethylhept-2-enyl-, trans-3-ethyloct-2-enyl-, and trans-3-ethyldec-2-enyl-pyrophosphates, gave reactive substrates. However, chain branching in the alkyl residues gave inactive substrates. 3-Ethylbut-3-enyl pyrophosphate also acts 11 as a substrate for farnesyl pyrophosphate synthetase with dimethylallyl pyrophosphate or geranyl pyrophosphate as the starter unit to afford homologues of farnesyl pyrophosphate.
3 Hemiterpenoids
Tryptophan, alanine, and mevalonic acid have been established in earlier work as biosynthetic precursors of echinulin (1). Cyclo-L-alanyl-L-trypto-phanyl (2) is a further intermediate whose isoprenylation has been studied. A cell-free system has been prepared from Aspergillus amstelodami which catalyses the transfer of one isoprene unit from dimethylallyl pyrophosphate to the cyclo-L-alanyl-L-tryptophanyl moiety. An enzyme system has been partially purified from cell-free extracts of E. coli which catalyses the synthesis of N6-(3,3-dimethylallyl)adenosine in transfer RNA.
Further studies have been reported on the biosynthesis of the hemiterpenoid furanocoumarins. A coumarin with a dimethylallyl group adjacent to a hydroxy-group figures in most of the biogenetic speculation concerning the origin of the furan ring in furanocoumarins. Demethyl[l'-14C]suberosin (3) is incorporated by the fruits of Angelica archangelica into the linear furanocoumarins such as bergapten (6), imperatorin (7), and isoimperatorin (8). Degradation of the bergapten established that incorporation occurred without randomization. Just as marmesin (4) acts as a precursor of psoralen (5), bergapten (6), and xanthotoxin (9), so it also acts as a precursor of rutaretin (10) in Ruta graveolens. 7-Hydroxyumbelliferone is the most effective general precursor of the coumarin portion of these furanocoumarins.
A thorough review has appeared on the biosynthesis of the ergot and associated alkaloids.
4 Monoterpenoids
A comprehensive and critical review, covering the literature to April 1971, on the biosynthesis and metabolism of the monoterpenes has appeared.
Mevalonic acid is often incorporated selectively into the second isoprene unit of the monoterpenes. This is attributed to compartmentalization effects in the biosynthesis of dimethylallyl pyrophosphate as a starter unit. 3,3-Dimethyl-acrylic acid has been suggested as an alternative precursor. However, a study of its incorporation into (+)-pulegone (11) by Mentha pulegium indicated extensive randomization of the label and hence decomposition and resynthesis. Further evidence for compartmentalization effects and the presence of an endogenous dimethylallyl pyrophosphate pool participating in monoterpene biosynthesis in vegetative tissue has come from the study of the incorporation of [14C]carbon dioxide and [14C]glucose into the monoterpenes of peppermint, Mentha piperita. The pulegone (11), derived from 14CO2 after different time intervals, was degraded and 90% of the radioactivity was found in the second isoprene unit (i.e. derived directly from IPP). Sucrose co-administered with [2-14C]mevalonic acid to peppermint cuttings enhances the incorporation of mevalonate into monoterpenes, indicative of energy requirements for this biosynthesis.
The isolation and properties of a monoterpene reductase from rose petals have been described. Geraniol and nerol were reduced by a solubilized enzyme preparation to give citronellol. The co-factor requirement was filled only by NADPH and the system had an optimum pH of 8. It was inhibited by p-chloromercuriophenylsulphonic acid, suggesting the presence of an S — H group near the active site.
Salvia officinalis has been shown to specifically incorporate [2-14C]geraniol into (-)-camphor (12) and (-)-borneol (13) (0.5 × 10-3 % and 3.3 × 10-3%) such that the tracer is incorporated into C-2 of both monoterpenes. This work contains a salutary discussion of the dangers of relying solely on g.l.c. purification of terpenes for radioactive tracer studies.
Monoterpenes are probably metabolically labile in higher plants. Feeding of [14C]-labelled p-menth-1-en-8-ol (α-terpineol) (14) or trans-thujan-3-one (15) to Tanacetum vulgare or geraniol to Artemisia annua led to a significant uptake into the carotenoids with a labelling pattern that suggested an incorporation of possibly undegraded C10 units or dimethylallyl pyrophosphate fragments.
Changes in the monoterpene composition of Mentha aquatica have been produced by gene substitution. The oxidation of pulegone (11) to mentho-furan (16) is controlled by a single gene that is not completely dominant, whereas the reduction to menthone involves a different set. Changes in favour of low menthofuran composition have been produced in Mentha aquatica by gene substitution.
The cyclopropane monoterpenoid chrysanthemic acid (17) has an unusual linkage of isoprene units and there has been considerable speculation about a possible biogenetic relationship with irregular terpenes involving cleavage of this ring. The mevalonate labelling pattern of artemisia ketone (18) by Artemisia annua has been described with results that suggested the intervention of a chrysanthemyl ion or its biogenetic equivalent. However, sodium [carboxy-14C]chrysanthemate and the corresponding chrysanthemyl phosphates were not incorporated into artemisia ketone by Santolina chamae-cyparissus. A number of model studies on the formation of this ion have been reported to afford substances such as artemisia alcohol and yomogi alcohol.
The incorporation of 14CO2 into nepetalactone at different time intervals has been reported. A cell-free preparation has been made from Vinca rosea that will methylate loganic acid, which is an intermediate in the biosynthesis of the secoiridoids and the indole alkaloids (see Chapter 4).
5 Sesquiterpenoids
The predominant isomer of farnesol is the 2,6-trans,trans-isomer, which is well known as a precursor of the steroids. However, the 2-cis-6-trans-isomer has often been suggested as a possible precursor of sesquiterpenoid substances. A soluble enzyme preparation has been reported from Pinus radiata seedlings which condenses isopentenyl pyrophosphate with geraniol to give both the 2-cis- and 2-trans-isomers of farnesol. A fungal system from Helminthosporium sativum has been shown to isomerize the trans-2,3 double bond of epoxy-farnesol and farnesol to a cis double bond with the exchange of one hydrogen atom at C-1. The same fungus converts the epoxide into 10,11-dihydroxy-farnesol, 10,11-dihydroxyfarnesic acid, and 9,10-dihydroxygeranyl acetone.
The biosynthesis of the insect juvenile hormone (19) continues to present incorporation problems. The acid 10-epoxy-7-ethyl-3,11-dimethyltrideca-2,6-dienoic acid (20) acts as a substrate for the hormone in the giant silk moth, Hyalophora cecropia. L-Methionine gave the ester methyl group. However, it did not contribute to the carbon skeleton whilst farnesol, farnesol pyrophosphate, propionate, and mevalonate were apparently not utilized for the biosynthesis of the hormone under the conditions of these experiments. There was a very low incorporation of [2-14C]acetate into juvenile hormone.
The fungal metabolite siccanin (21) contains a sesquiterpenoid fragment and a fragment derived from orsellinic acid. Previous reports had described a cell-free system from Helminthos porium siccans for the synthesis of trans-γ-mono-cyclofarnesol (22). The formation of siccanochromene-A (25) has now been studied by the incubation of cell-free systems from Helminthosporium siccans with mevalonolactone and orsellinic acid, and trans,trans-famesyl pyrophosphate and orsellinic acid. Orsellinic acid was found to be the only suitable aromatic co-substrate for the enzyme preparation. Presiccanochromenic acid (23) and siccanochromenic acid (24) were also shown to be precursors of siccanochromene-A.
Earlier experiments on the biosynthesis of tutin (28) wit h [2-14C]mevalonic acid and [2,2-3H2,2-14C]mevalonic acid had indicated its sesquiterpenoid nature and had suggested that it was formed via copaborneol (27). Specifically tritiated copaborneol has now been shown to be converted into tutin (28) by Coriaria japonica without randomization of the label. The initial step in the cyclization to form copaborneol has been regarded as the cyclization of the electrophilic C-1 of farnesyl pyrophosphate with the distal double bond to form a germacrane cation (26). In order for cyclization to proceed further the reactive centre must then be transferred from C-11 to C-1. One proposal involved a double 1,2-hydride shift from C-10 to C-11 and C-1 to C-10. However, this proposal has been eliminated from tutin biosynthesis by the retention of a (4R)-[4-3H]mevalonoid hydrogen at C-10: this carbon atom becomes C-4 in tutin (28).
The biosynthesis of caryophyllene (29) from [2-14C]mevalonic acid in Mentha piperita has been studied. Degradation of the caryophyllene revealed a preferential labelling of those portions of the molecule derived directly from isopentenyl pyrophosphate. Petasin (30), formed in Petasites hybridus, possesses the rearranged eremophilane skeleton. The proposal has been made that it is formed via a β-germacrene, although there are also alternative suggestions involving a spiro-intermediate. The biosynthesis has been studied with [2-14C]- and (4R)-[4-3H]mevalonic acid. The labelling pattern was consistent with the postulated mode of biosynthesis from trans,trans-farnesyl pyrophos-phate with a methyl migration from C-10 to C-5 and a hydrogen shift from C-5 to C-4α [see (31)].
There have been further reports from a number of laboratories concerning the biosynthesis of the trichothecane group of fungal metabolites. α-Bisabolol, γ-bisabolene, and monocyclofarnesol are not precursors of trichothecin (34). However, the hydrocarbon trichodiene (33), which is probably the key intermediate in this biosynthesis, was converted into trichothecolone (35), 12,13-epoxytrichothec-9-ene, and trichodiol by Trichothecium roseum. Degradation of trichothecolone (35), biosynthesized from (4R)-[4-3H,2-14C]mevalonic acid, had located a (4R)-[4-3H] label at C-10, suggesting that the farnesyl chain was folded in the manner shown in (32). An independent degradation has now been reported, locating a [2-14C] label at C-8 in accord with this scheme. Full papers have now appeared on the formation of the trichothecane nucleus, on the biosynthesis of helicobasidin, and on the biosynthesis of mycophenolic acid.
The stereochemistry of the formation of the double bonds in abscisic acid (36), biosynthesized by avocado fruit, has been studied using (2R)-[2-3H)-, (2S)-[2-3H)-, and (5S)-[5-3H]-mevalonates. The anticipated stereochemistry of the hydrogen atoms derived from C-2 and C-5 of mevalonate is shown in (37). The C-3' and C-4 hydrogen atoms of abscisic acid (36) were derived from a 2-pro-R-mevalonoid hydrogen atom. The hydrogen atom at C-5 of abscisic acid is derived from a 5-pro-S-mevalonoid hydrogen. The presence of some label at positions 3' and 4 when (2S)-[2-3H]mevalonic acid was the precursor was attributed to the action of isopentenyl isomerase.
The acid-catalysed cyclization of nerolidol to the bisabolenes and thence to cedrene has been studied as a possible model for the biosynthesis. The same authors have also discussed the possible intervention of spiranic intermediates in the biosynthesis of cedrene.
6 Diterpenoids
The biosynthesis of diterpenoid compounds has been reviewed. The manool: manoyl oxide group of diterpenoids could arise by the cyclization of a geranyl-linalool in which even the oxide-ring formation might be concerted with the initial cyclization. However, (-)-labda-8,13-dien-15-[3H]ol pyrophosphate (copalol pyrophosphate) (38) was specifically incorporated 59 into (-)-13-epimanoyl oxide (olearyl oxide) (39), indicating that the geraniol:linalool type of allylic isomerization could take place at the bicyclic level. A number of model studies in the cationic rearrangements and cyclizations related to the bio-synthesis of the bicyclic and tricyclic diterpenoids have been reported in the past few years. A biogenetically patterned synthesis of deoxytaondiol methyl ether from manool and toluquinol 4-methyl ether has been described.
The biosynthesis of the virescenosides [D-altropyranosides of diterpenoid aglycones such as virescenol A (40)] from [l-13C]- and [2-13C]-acetate has been studied by 13C n.m.r. spectroscopy. Enrichment of the atoms was observed as shown (41), in accord with currently accepted proposals.
A number of aspects of kaurene and gibberellin biosynthesis have been discussed during the year. The incorporation of [3'-2H3]mevalonic acid into (-)-kaurene (42) by a cell-free system from Gibberella fujikuroi has been studied. The labelling pattern excludes a (-)-pimaradiene (43) from the biosynthesis. The previously reported low, but specific, incorporation of (-)-pimaradiene was therefore a microbiological transformation. Some evidence has been presented for a kaurene-protein complex in extracts from immature pea, Pisum sativum. This renders the lipophilic diterpene accessible to oxidation by mixed function oxidases in the microsomal fraction. Some evidence has also been found for this in Gibberella fujikuroi. The sequence of oxidation on ring B in kaurene :gibberellin biosynthesis leading to 7β-hydroxy-(-)-kaurenoic acid (44) has been described in a full paper. A cell-free system from the endosperm of Cucurbita pepo has been prepared which converts [2-14C)mevalonate into (-)-kaurene (42) and thence into 7β-hydroxy-(-)-kaurenoic acid (44) and the gibberellin A12 aldehyde (45). The separation of µg quantities of enzymatic products can be achieved using g.l.c.-mass spectra. In this instance the high degree of labelling obtained with a cell-free system when dilution of the label by endogenous substrates was avoided, was put to advantage and the 14C incorporation was determined by mass spectrometry.
