Biosynthesis

BY J. STAUNTON

The main emphasis of research on alkaloid biosynthesis has shown a significant move away from recently fashionable problems. The terpenoid indole family of alkaloids no longer dominates the scene and the benzylisoquinoline and ergot alkaloids have also lost their prominent position.

It seems unlikely that any new area will achieve in the future the dominant role enjoyed in the recent past by the indole alkaloid family, and the major advances this year have been spread over several unrelated fields. For example the middle and late stages in the biosynthesis of the Ipecac and Cinchona alkaloids have been largely elucidated although important gaps remain to be filled. The interesting work of several groups on the role of lysine in alkaloid biosynthesis continues apace and much further work will be necessary before one can be sure of understanding the overall picture in this complex field. A major breakthrough on the peyote cactus alkaloids has solved a longstanding puzzle in isoquinoline biosynthesis and should stimulate fresh interest in the early stages of isoquinoline formation in other systems.

As the body of knowledge accumulates, it becomes increasingly unlikely that many pathways remain to be discovered. Thus, it is hardly surprising that many of the compounds under investigation this year have been found to arise by modification of already established pathways. However, there is promising evidence that the mesembrine alkaloids are biosynthesized by an unexpected, novel route and the past year has produced its full quota of surprises and puzzling results from both new and established pathways.

Terpenoid Indole Alkaloids. — A surprising development has taken place since the comprehensive review was published in last year's Report. The result came not from the biosynthetic study but from further investigation of the configuration at C(3) of vincoside (3). The stereochemistry at this position was uncertain because past research gave conflicting results. However, it has now been convincingly established that in vincoside the hydrogen at C(3) has the β-orientation.

This result is unexpected from the point of view of the biosynthetic chemist because the configuration at C(3) of vincoside is now opposite to that at the corresponding carbon of the next established intermediate, geissoschizine (4). It is not yet clear how this centre becomes epimerized in the biosynthesis but the following experimental facts have to be accounted for: (a) isovincoside [the C(3) epimer of vincoside] is not biologically active, and (b) the hydrogen at C(5) of loganin (1) [C(5) of loganin corresponds to C(3) of vincoside] is completely retained in the biosynthesis of the three main classes of indole alkaloid, represented in Scheme 1 by vindoline (6) [Aspidosperrna], catharanthine (7) [Iboga], and ajmalicine (8) [Corynanthe].

The retention of the hydrogen from C(5) of loganin implies that the epimerization takes place without cleavage of a carbon–hydrogen bond and thus provides a valuable clue to the mechanism of this intriguing process. In view of the importance of this result, it is essential to show that the hydrogen has not undergone an unexpected migration prior to the epimerization step. The predicted destination of the hydrogen from C(5) of loganin in the absence of a migration is marked by an asterisk in the structures in Scheme 1. The earlier experiment was repeated by administering [5-3H,O-methyl-3H]loganin to Vinca rosea plants. The three alkaloids (6), (7), and (8) were each found to retain both tritium labels without change in isotopic ratio and it was proved by degradation that, for each alkaloid, the activity incorporated into the skeleton resided specifically at the expected position. Thus, it is very probable that the epimerization process involves cleavage of a carbon–carbon or carbon–nitrogen bond. A similar process seems to operate in the biosynthesis of the Ipecac alkaloids and the problem will be discussed further in that section.

Eburnamine–Vincamine Alkaloids. — So far most of the effort on indole alkaloid biosynthesis has been concentrated on the Corynanthe, Aspidosperma, and Iboga systems. It is welcome, therefore, to see the preliminary results of an investigation of the biosynthesis of vincamine (10). Comparable incorporations were observed for [ar-3H]tryptophan, [ar-3H]stemmadenine (5), and [ar-3H]tabersonine (9). These results support the proposal that vincamine is a transformation product of the Aspidosperma system, and it will be interesting to see if further work supports the detailed mechanism proposed for the transformation.

Cinchona Alkaloids. — The biosynthetic route (Scheme 2) to the Cinchona alkaloids quinine (22), cinchonidine (21), and cinchoninine (23) is of exceptional complexity. Earlier work established that quinine is derived by combination of an indolic unit derived from tryptophan with a monoterpenoid unit derived from geraniol, via loganin. The close relationship to indole alkaloid biosynthesis implied by these results has been confirmed by the demonstration that vincoside succeeds loganin in the biosynthesis. In feeding experiments with C. ledgeriana shoots [ar-3H]vincoside was incorporated into each of the Cinchona bases (21), (22), and (23). By analogy with indole alkaloid biosynthesis it is likely that secologanin (2) is an intermediate between loganin and vincoside.

The proposal that the next stages of the pathway would involve intermediates of the Corynanthe type has received experimental support. Corynanthealdehyde (12) was not incorporated but the closely related corynantheal (13) did serve as an efficient precursor for all three Cinchona bases. Thus, the close parallel between the early stages of quinine biosynthesis and the corresponding stages (Scheme 1) of indole alkaloid biosynthesis is established.

From corynantheal onwards the pathways diverge completely. The generation of the skeleton of the Cinchona bases requires not only a further reorganization of the terpenoid moiety but also a fundamental rearrangement of the indolic portion of the molecule to generate the quinoline residue. The currently favoured working hypothesis for this transformation is shown in Scheme 2: (13) [right arrow] (14) [right arrow] (15) [right arrow] (16) [right arrow] (17) and (18). With the required skeleton in hand, only relatively trivial biochemical reactions are required to produce the known Cinchona alkaloids (21), (22), and (23).

The very late stages from (17) and (18) onwards have been intensively investigated. When [11-3H]cinchonidinone (17) was administered to C. ledgeriana shoots, activity was incorporated efficiently into the corresponding alkaloids cinchonidine (21) and cinchonine (23). The incorporation was shown to be specific for cinchonine, by degradation. A low but possibly significant incorporation of (17) into quinine (22) was also observed. Thus, it is possible that the methoxy-group of quinine is introduced after generation of the quinoline system in (17).

The presence of the keto base (17) in Cinchona plants was confirmed by dilution analysis. C. ledgeriana shoots were allowed to metabolize [1-14C] tryptamine and were then worked up for alkaloids after the addition of inactive (17). The recovered carrier was radioactive, which confirms that the keto-base (17) is a natural product of the Cinchona plants and is, therefore, a true intermediate in the biosynthesis of the alkaloids. The presence of the ketone (20) was also demonstrated by dilution analysis of Cinchona plants which had metabolized [ar-3H]vincoside.

The reversibility of the reduction step (17) [right arrow] (21) was examined by feeding [11-3H]cinchonidine (21) to C. ledgeriana shoots and isolating the corresponding ketone (17) after addition of a carrier. The recovered cinchonidinone was radioactive, which proves that reversal does occur [ie. (21) [right arrow] (17)]. In view of this result, it is reasonable to suppose that all the late steps are reversible after generation of the quinoline system in (17) and (18) but obviously many more tracer experiments would be required to prove every detail implied in the scheme.

Thus, the early and late stages of Cinchona alkaloid biosynthesis are well worked out. However, the really unique steps in the pathway lie between (13) and (17) or (18); it is this stage of the biosynthesis that sees the profound skeletal reorganization of the indolic unit to a quinoline. The sequence of intermediates proposed in Scheme 2 is intellectually appealing but, at this stage, remains purely speculative.

Indirect evidence to support the intermediacy of the aldehyde (14) has come from experiments which rule out the corresponding alcohol and the corresponding acid, respectively. The alcohol cinchonamine [(14) but CH2OH instead of CHO] was tested directly by administering the tritiated compound to C. ledgeriana shoots. None of the Cinchona bases showed a significant incorporation of activity. The carboxylic acid corresponding to (14) was not tested directly but was ruled out by an incorporation experiment with [1-14C,1,1-3H2]tryptamine (11). Each of the alkaloids showed an isotopic ratio corresponding to the retention of half the tritium. Thus, one of the two hydrogens residing at C(1) of tryptamine survives the biosynthesis, which rules out the acid corresponding to (14) as an intermediate. The elimination of the alcohol and the acid increases the probability that the aldehyde (14) is involved in the biosynthesis, but it must be emphasized that there is no direct support for this proposal.

The elucidation of the steps between (13) and (17) represents a formidable challenge to both chemical skill and biosynthetic intuition and it can be expected that this area will provide more exciting results in the future.

Ipecac Alkaloids. — Earlier tracer results on the alkaloids emetine (29), cephaeline (28), and ipecoside (27) of Cephaelis ipecacuanha have established that a C9—10 unit (thickened bonds) is derived from geraniol via loganin. The suggested pathway (Scheme 3) shows a parallel, in the early stages, to indole alkaloid biosynthesis. Secologanin (2) derived in the usual way from loganin, is condensed with a unit of dopamine to furnish one of the epimeric isoquinolines desacetyli-pecoside (26) or desacetylisoipecoside (25). Structural reorganization of the terpenoid unit is followed by condensation with a second dopamine unit to generate the alkaloids cephaeline (28) and emetine (29).

The role of secologanin was studied by feeding [O-methyl-3H,6, 6-3H2]secologanin to C. ipecacuanha plants; activity was incorporated into all three alkaloids and in the case of ipecoside the distribution of activity was shown by degradation to correspond to an intact incorporation.

A highly significant result emerged when the epimeric isoquinolines (25) and (26) were tested in vivo. Both compounds are available from the in vitro condensation of dopamine with secologanin. [3'-14C] Desacetylipecoside (26) was found to be an efficient and specific precursor of all three alkaloids (27), (28), and (29) whereas the epimer, desacetylisoipecoside (25), was biologically inactive.

The specific incorporation of (26) rather than its epimer (25) into cephaeline (28) and emetine (29) is surprising when one compares the configuration at C(5) of (25) with that at the corresponding position [C(11b)] of (28) and (29). The absolute configuration of (25) is rigorously established by relation to ipecoside. Thus, the transformation (26) [right arrow] (28) requires inversion of the configuration at C(5)of(25).

The situation is reminiscent of the transformation of vincoside into geissoschizine in indole alkaloid biosynthesis. In the case of the indole alkaloid pathway it was proved by feeding [5-3H]loganin that the C — H bond at the relevant asymmetric centre is not broken in the epimerization step. A similar experiment in C. ipecacuanha with [5-3H]loganin has established that the hydrogen at C(5) of loganin is again retained and that it resides in the alkaloids specifically at the corresponding position, C(11b). Thus, the two biosynthetic pathways show the same puzzling feature: when the heterocyclic ring is formed by condensation of an arylethylamine with an aldehyde, the new asymmetric centre is generated in the 'wrong' configuration; epimerization of this centre is therefore necessary and it is known that the mechanism does not involve cleavage of the C — H bond. It will be interesting to see whether this is a general pattern for biosynthetic pathways involving secologanin; also, the mechanism of the epimerization deserves further investigation.

The sequence of intermediates after (26) is still an open question. Evidence has been obtained that the ethyl group of (28) is not generated directly by reduction of the vinyl group, but that the double bond first migrates to generate an ethylidene group. Thus, when [2,2-3H2,2-14C]geraniol was administered to the plant the tritium was found to be completely lost in the alkaloids (28) and (29).

An earlier report that glycine is incorporated specifically into the terpenoid unit of cephaeline has stimulated similar research in the indole alkaloid series. However, glycine was found not to be a specific precursor of either strychnine or ajmalicine.

Ergot Alkaloids. — The tryptophan and tryptamine derivatives (30) and (31) respectively, labelled in each case with 14C at the carbinol carbon, have been shown to be precursors of the alkaloids agroclavine (32) and elymoclavine (33). These results throw further light on the sequence of intermediates leading from tryptophan to the ergoline system.

Several papers have appeared on the biosynthesis of the peptide side-chains of the various ergot alkaloids. In the case of ergotamine (34), activity was incorporated from [1-14C]alanine into the 2-hydroxyalanyl residue. An investigation of ergotoxine (35) biosynthesis produced no surprises; DL-[2-14C]tryptophan was specifically incorporated into the lysergic acid moiety and L-proline served as precursor of the proline unit of the side-chain. In contrast, the origin of the hydroxyethyl side-chain of N-(α-hydroxyethyl)lysergamide (36) is still obscure. The following precursors have been administered to Claviceps paspali: [1-14C]acetate, [14C] formate, [2-14C] mevalonic acid, [2-14C]indole, DL-[3-14C]tryptophan, DL-[3-14C]serine, DL-[2-14C]alanine, and DL-[2-14C]-pyruvate. Activity was incorporated from each of these compounds but only the last two introduced activity into the hydroxyethyl group. However, the incorporation into this group was non-specific so that none of the compounds tested in this investigation qualified as a direct precursor of the hydroxyethyl side chain.

Miscellaneous Indole Alkaloids. — Echinulin (38) is known to be assembled from tryptophan, mevalonic acid, and alanine. The indole derivative (37) has now been shown to be converted efficiently by Aspergillus amstelodami into echinulin and is, therefore, probably an intermediate.

The fungal metabolite cyclopiazonic acid 1 (40) has been shown to derive from tryptophan, mevalonic acid, and acetate. The corresponding bis-seco-derivative (39) also gave an incorporation when administered to Penicillium cyclopium Westling and is, therefore, probably an intermediate.

Isoquinoline Alkaloids. — This year has seen the solution of a longstanding mystery in alkaloid biosynthesis: the origin of the 'extra' skeletal carbons of the peyote cactus alkaloids, anhalonidine (43) and anhalamine (47). The major portion of the skeleton is derived in each case from tyrosine, by a well established pathway leading to the intermediate phenethylamine (41) but, despite much research, the origin of C(1) of (47) and C(1) + C(9) of (43) remained unsolved.