The Biosynthesis of C5–C20 Terpenoid Compounds

BY J. R. HANSON

1 Introduction

This chapter, covering 1973, follows the pattern of the previous Reports. A number of reviews on general aspects of terpenoid biosynthesis have appeared. Significant advances have been recorded in the establishment of cell-free systems that mediate stages of terpenoid biosynthesis and in the use of mutants in the investigation of biosynthetic sequences.

2 Mevalonic Acid

There are two pathways leading to the formation of acetoacetate. In the first acetoacetyl co-enzyme A is formed directly from two moles of acetyl co-enzyme A. It is this pathway which is implicated in HMG CoA biosynthesis. In the second, malonyl CoA is involved leading to an enzyme-bound acetoacetate as in fatty acid biosynthesis. Malonate has been shown to be incorporated into HMG CoA and into ergosterol, but the labelling patterns were consistent with decarboxylation prior to incorporation. The incorporation into mevalonate by rat liver preparations was related to the malonyl CoA decarboxylase activity present in this preparation. Mevalonate has not hitherto been identified in the invertebrates. However, its biosynthesis has now been demonstrated in a number of tissues from the fly, Sarcophaga bullata. The block in sterol biosynthesis appears to exist in the condensation of farnesyl pyrophosphate to squalene. Some attempts to distinguish different forms of mevalonate kinase from green leaves and etiolated cotyledons of the French bean, Phaseolus vulgaris, have been described.

3 Hemiterpenoids

The formation of the isoprenoid portion of the furanocoumarins has continued to receive attention. The positions of 7-demethylsuberosin and osthenol in furanocoumarin biosynthesis have been examined. 7-Demethylsuberosin (1), which is formed by prenylation of umbelliferone, was detected by trapping experiments and shown to be a precursor of the linear furanocoumarins such as psoralen (2) and bergapten (3) in Conium maculatum and Heracleum lanatum. Similar results had been obtained previously with Angelica archangelica. Osthenol (4) was a precursor of the angular furanocoumarins such as angelicin, isobergapten, and sphondin. The biosynthetic routes to psoralen (2), bergapten (3), and xanthotoxin (5) have been studied in cell-cultures obtained from Ruta graveolens. 7-Demethylsuberosin (1) and marmesin (6) were shown to be good precursors of these substances in this system confirming the previous results. A similar system derived from Ruta graveolens has been used to study the formation of edulinine and the furoquinoline alkaloids from quinoline derivatives.

Two different pathways for anthraquinone biosynthesis are known. One is entirely based on acetate whilst the other involves shikimic acid and glutamic acid, which afford o-succinoylbenzoic acid which is then converted into a naphthalene and prenylated. Labelling results suggest that ring C of alizarin (7), produced by Rubia tinctorum, and morindone (8), produced by Morinda citrifolia, are derived from mevalonate via dimethylallyl pyrophosphate. Isoprene has been reported as light-dependent natural plant emission from leaf discs of Hamamelis and its 'action spectrum' has been described.

4 Monoterpenoids

With a few exceptions, the incorporation of mevalonate into monoterpenes is poor and in a number of instances extensive randomization of the label has been observed. In several cases mevalonoid labels have been found predominantly in those parts of the molecule derived directly from isopentenyl pyrophosphate with virtually no label in the starter unit derived from dimethylallyl pyrophosphate. The non-uniform labelling of geraniol biosynthesized from 14C carbon dioxide in Perlargoniuum graveolens has been studied. The geraniol was isolated up to 24 h after an initial exposure of 2 h. Degradation showed a greater turn-over of label in the isopentenyl pyrophosphate portion than in the dimethylallyl pyrophosphate portion of geraniol. At first the label was approximately equally divided between the two portions but after 12 h the proportion of the label associated with the isopentenyl pyrophosphate derived half increased to 78%. Subsequently the preferential labelling decreased and approached an equal distribution. These results suggest that the geranyl pyrophosphate which was isolated at first arose as a result of leakage from sites of higher terpenoid biosynthesis, whereas the labelling of the later material was affected by the existence of a dimethylallyl pyrophosphate pool and compartmentalization of isoprenoid biosynthesis.

A cell-free system derived from Mentha piperita, has been established which mediates the cyclization of neryl pyrophosphate (9) to [apha]-terpineol (10). Similar systems were also obtained from Mentha spicata and Daucus carota. Competing phosphatase activity was inhibited by sodium fluoride. Changes in the monoterpene composition of the essential oil of Mentha aquatica have been brought about 18 by genetic substitution from a high limonene yielding strain of Mentha citrata extending studies reported previously with pulegone metabolism. It has been shown that one of the first steps in the formation of the iridoid monoterpenes and the monoterpene portion of the indole alkaloids, is the hydroxylation of geraniol and nerol at C-10. A microsomal mixed-function oxidase, capable of mediating this step, has now been isolated from Vinca rosea. Model systems have been studied 20 for the biological oxidation of monoterpene hydrocarbons based on the photo-oxygenation of α-thujene (11) and sabinene (12) in the presence of chloroplast preparations from Tanacetum vulgare or Juni perus sabina and various dyes.

trans-Verbenol (14) is the principal aggregating pheromone of the bark beetle, Dendroctonus ponderosa. Exposure to the oleoresin from the host-tree, Pinus monticola or to α-pinene (13) increases the production of this pheromone.

There has been considerable interest and speculation on the formation of the irregular monoterpenes. The various theories have been reviewed and a unified hypothesis based on an analogy with presqualene alcohol, has been proposed (Scheme 1). The initial step involves the head-to-head dimerization of two moles of dimethylallyl pyrophosphate to form a chrysanthemyl pyrophosphate. However, studies on the biosynthesis of chrysanthemum monocarboxylic acid from [4(R)-4-3H] mevalonic acid have shown that the label is unequally distributed between the two C5 moieties. The suggestion was made that the C5 isopentane unit leading to the cyclopropane part of the molecule is not derived directly from mevalonic acid. Chrysanthemum dicarboxylic acid is formed from chrysanthemum monocarboxylic acid.

5 Sesquiterpenoids

A review of the biogenetic proposals for the sesquiterpenes has appeared. The biogenetic relationships amongst the sesquiterpene lactones of the Compositae have been discussed.

Considerable interest has centred on the formation of 2-cis-farnesol, required by many of the cyclic sesquiterpenes. A soluble enzyme system which forms the two isomeric farnesols has been isolated from orange flavedo. When [1-3H]geranyl pyrophosphate and isopentenyl pyrophosphate were used as a substrate in the presence of NAD+ the order of appearance of the sesquiterpenes was 2-trans-farnesol, 2-trans-farnesa, 2-cis-farnesa, and finally 2-cis-farnesol. A redox system was proposed to account for the possible isomerization of 2- trans- to 2-cis-farnesol. Aldehydes have been proposed previously as intermediates in monoterpene isomerizations. The biosynthesis of 2-trans-6-trans and 2-cis-6-trans-farnesols have been studied 27 with a cell-free system obtained from tissue-cultures of Andrographis paniculata. This system was shown to incorporate six atoms from [5-3H2] mevalonate into the trans-trans isomer but only five atoms from [5-3H2] mevalonate into the cis-trans isomer. On the other hand, the [4(R)-4-3H] label which occurs on the 2-double bond was retained in both isomers. 2-cis-6-trans-farnesol (15) and trichodiene (16), the hydrocarbon precursor of the trichothecanes, have been shown to be formed by a cell-free preparation from Trichothecium roseum. When all trans-farnesyl pyrophosphate was used as the substrate, there was a loss of one [5-3H2]-mevalonoid label from the C-1 position of the farnesol. Although the farnesals were barely detectable in this system, the co-factor requirements were again indicative of a redox system. A similar system had previously been isolated from Helminthos porium sativum.

2-cis-6-cis-Farnesol pyrophosphate ( 17) is a potential precursor for the dimeric sesquiterpenoid gossypol (18). A protein fraction has been isolated from Gossipium hirsutum which is capable of converting neryl pyrophosphate and isopentenyl pyrophosphate into 2-cis-6-cis-farnesyl pyrophosphate and 2- trans-6-cis-farnesyl pyrophosphate. The same preparation converts geranyl pyrophosphate and isopentenyl pyrophosphate into 2-cis-6-trans-farnesyl pyrophosphate and 2-trans-6-trans-farnesyl pyrophosphate. It does, however, appear to have a different specificity for the two substrates.

The biosynthesis and metabolism of the insect juvenile hormones continue to attract attention. A series of reviews on these hormones has appeared including chapters on their metabolism. An interesting development has been the study of the incorporation of propionate and acetate using organ-culture techniques. The labelling pattern has led to the conclusion that the homo-isoprenoid units arise from one propionate and two acetates and that homo-mevalonate may be a precursor of these substances.

In an investigation on the balance between di- and tri-terpene biosynthesis in which farnesyl pyrophosphate forms a branch point, farnesyl pyrophosphate was shown to be converted into the triphosphate by Gibberella fujikuroi.

Although the epoxy-acid (19) is a precursor of abscisic acid (21), it cannot be detected in avocado by cold-trap experiments. It is, however, also converted into (-)-1',2'-epi-2-cis-xanthoxin acid (20). Phaseic acid (22) and 4-dihydrophaseic acid have been isolated as metabolites of abscisic acid in Phaseolus vulgaris. The metabolism of abscisic acid in Hordeum vulgare has been studied but the products have not been identified.

The cochlioquinones (23) are metabolites of Cochliobolus miyabeanus. Mevalonate feeding experiments have established that they are formed from a farnesyl unit which is attached to an aromatic precursor. The secondary methyl groups on the side-chain arise from methionine.

Dendrobine (24) is one of the few sesquiterpene alkaloids. It is, however, related to the picrotoxane series. During the biosynthesis of dendrobine a hydrogen atom is transferred from C-1 of farnesol to C-8 of dendrobine.

The sesquiterpenoid quinone, helicobasidin (25), has been the subject of a number of biosynthetic studies. The labelling pattern from [2-13C] mevalonic acid has been established using n.m.r. spectroscopy. These results confirm the labelling pattern which had been determined previously from 14C studies.

The formation of the trichothecanes continues to attract attention. [6-3H]Farnesyl pyrophosphate (26) is incorporated into roridin A and verrucarol (27) with no loss of radioactivity. Degnidation of the verrucarol established the presence of the label at C-2. This result confirms the hydride shift from the central double bond of farnesyl pyrophosphate to C-2 of the trichothecanes (see Scheme 2) and the fact that hydroxylation at C-2 proceeds with retention of configuration.

[2-14C]Mevalonic acid has been incorporated into hirsutic acid C (28) and a related ketone, complicatic acid, both of which are metabolites of the fungus, Stereum complicatum. Hirsutic acid C is a precursor of complicatic acid. The incorporation of [2-3H2]- and [4(R)-4-3H]mevalonoid hydrogens into the sesquiterpenoid fungal metabolite, illudin M (29) has been studied. Although the [4(R)-4-3H] label was at the expected site, the [2-3H2] mevalonoid labelling pattern suggested that the formation of these compounds might not lie through the proposed humulene intermediate.

6 Diterpenoids

The biosynthesis of the gibberellin plant-growth hormones has attracted considerable attention. Ent-kaurene (30) is an important precursor of these substances. A soluble enzyme system for its biosynthesis has been obtained from Pisum sativum shoot tips.

Similar systems mediating kaurene biosynthesis have also been obtained from Echinocystis macrocar pa, Gibberella fujikuroi, and Ricinus communis. The accumulation of ent-kaurene was stimulated by dithiothreitol and inhibited by AMO 1618. Ent-kaurene metabolism has also been studied 44 in Hordeum distichon (barley). However, the microsomal fraction in this case epoxidized and hydroxylated the olefinic double bond rather than oxidizing C-19.

Mutants of the fungus Gibberella fujikuroi have provided a very effective means of studying gibberellin biosynthesis. The mutant Bl-41a has a block between ent-kauren-19-al and ent-kauren-19-oic acid. Gas chromatography combined with mass spectroscopy provided the means of studying biosynthesis in the absence of endogenous substrates. Ent-7α-hydroxykaurenoic acid was formed after ent-kaurenoic acid. Gibberellin A12 aldehyde was metabolized to gibberellin A14 aldehyde (31–32) and thence via gibberellin A14 to gibberellins A4 (37), A7 (38), A1 (39) and A3 (gibberellic acid) (40). Gibberellin A12 alcohol and A14 alcohol were efficient precursors of gibberellic acid. Gibberellin A12 was converted into gibberellins A9 (33), A15 (34), A24 (35), and A 25 (36); i.e. those gibberellins which lack a ring A hydroxygroup. Gibberellins A13, A15, A25, A36, and A37 were not metabolized. Gibberellin A24 (34) was converted into gibberellin A25 (35). Gibberellin A4 (37) was converted into gibberellin A3 gibberellin A7 (38) was converted into gibberellin A3. Gibberellins A1 and A16 were terminal gibberellins in this system. The 3-hydroxylation of gibberellin A12 aldehyde (31) has been studied 46 by g.c.m.s. in the slow-growing mutant of Gibberella fujikuroi REC-193A. The results imply that 3-hydroxylation is the first step in the conversion of gibberellin A12 aldehyde into gibberellins A14, A4, and A7. Another mutant has been isolated (R-9) and this was shown to be blocked for gibberellin A1 and A3 biosynthesis but not for the biosynthesis of gibberellins A4 and A7. Cultures of this mutant convert low concentrations of gibberellin A1 (39) into gibberellic acid (40).

The incorporation of four 5-pro-S mevalonoid hydrogen atoms into gibberellic acid and the location of two of these at C-6 and C-14 has been shown by tritium labelling. The C-14β (exo-H) of gibberellic acid (40) is derived from a 5-pro-S hydrogen atom of mevalonic acid.

Although gibberellins A3 (40) and A7 (38) have been detected in germinating barley ( Hordeum distichon), the barley incorporated mevalonate into terpenoid fractions but not into gibberellic acid. The conversion of gibberellin A5 (41) into gibberellic acid (40) has been shown to occur in dwarf pea seedlings ( Pisum sativum). The metabolism of gibberellin A1 (39) in potato leaves has also been investigated 51 but the polar products were not identified. The conversion of gibberellin A4 (37) into gibberellin A1 (39) has been established in dwarf rice. Gibberellin A1 was then shown 53 to be converted into gibberellin A8 in these seedlings. This conversion had previously been established in Phaseolus vulgaris.