The Biosynthesis of Lycopodine

In an effort to investigate the generality of the classical scheme for the generation of pelletierine-type compounds according to Path C1 in Scheme 8,we investigated the incorporation of acetate into lycopodine (27). Based on extensive work with radiotracers performed in the 1960s and 1970s predominantly by Spenser and

MacLean, a model had been developed which envisaged the carbon skeleton of lycopodine to arise via the dimerization of pelletierine (28) (Scheme 13 top), which was indeed proven to be a precursor for lycopodine [52]. Curiously, however, it served in this role only for the "left" half of the molecule C-9 to C-16. A rational explanation was subsequently presented which envisaged the inter-mediacy of 4-(2-piperidyl)-3-oxobutanoate (29), a molecule possibly biogeneti-cally related to pelletierine, as the precursor for the "right" half of lycopodine, C-1 to C-8 [53] (Scheme 13 bottom).

When we initiated our work using stable isotope methodology, we postulated a variety of labeling patterns from [ 1,2-13C2]acetate in (27) which might be expected based on the models under discussion until then for the biosynthesis of the acetate derived fragments of tropane- and pelletierine-type alkaloids. Thus, for instance, the "left" half of (27) might be expected to show labeling according to the Path C whereas the "right" half might arise by Path D via (29). Alternatively, the pelletierine incorporated into the "left" half might arise via decarboxylation of (29) and thus to show a Path D labeling pattern (Scheme 14). An experiment with [1,2-13C2]acetate was deemed to be capable of deciding between these possibilities.

29 27

Scheme 13. Biosynthesis of lycopodine 27; pelletierine 28 is only incorporated into the left half

29 27

Scheme 13. Biosynthesis of lycopodine 27; pelletierine 28 is only incorporated into the left half

On a practical level it required three years of effort to obtain the first usable incorporation result. This was for the most part due to the sharply reduced sensitivity of the stable isotope method in comparison to the radiotracer methodology employed in the earlier investigations, even if intramolecularly doubly labeled precursors are applied. Secondly, this work had to be done in the field as, to the best of our knowledge, nobody has succeeded in the culturing of any member of the family Lycopodiaceae in the greenhouse. Furthermore, the application of tracers to cuttings of plants, often a successful method, did not lead to any detectable incorporation in this instance. Thus, our work was confined to the months of July and August when blackflies and mosquitoes in the bush of

Scheme 14. Possible labelling patterns in lycopodine derived from [13C2]acetate; in fact all four labelling patterns were observed

Northern Ontario were tolerable, night temperatures did not fall below freezing and fresh growth at the tips of shoots was visible. Thirdly, the proper choice of plant material was crucial as well. Successful incorporations were observed only if young shoots at the ends of the above-ground rhizomes were used which showed a relatively high proportion of fresh growth. By using this selected plant material for the experiment, the small amount of labeled alkaloid formed during an experiment of one week's duration was not diluted by too large an amount of the endogenous material from previous growing seasons. In this way specific incorporations of 0.3-1.0% above natural abundance could be realized routinely.

In the event, we observed an entirely unexpected labeling pattern after incorporation of [1,2-13C2]acetate [54]. Analysis of the 13C NMR spectra suggested that the patterns expected from Paths C and D were superimposed on each other in both acetate derived portions of the alkaloid. The two patterns were observed in a ratio of 1:1, which did not change over several repetitions of the experiment. In a separate experiment it was found that [1,2,3,4-13C4]ace-toacetate was not incorporated intact but was instead first cleaved to acetate and then incorporated, resulting in an identical labeling pattern to that which had been observed when [1,2-13C2]acetate was fed. The results of these two experiments on (27) were thus qualitatively identical to our observations during the investigation of 6^-hydroxytropine (15) biosynthesis in D. stramonium [22].

In the interpretation of these experiments in Lycopodium tristachyum the argument could be made, as had been in the case of (15), that the preference for cleavage of the C4 unit over intact incorporation was a consequence of higher reaction rates for the former rather than the latter process. We had rejected this explanation in the case of (15) on the basis of the argument that if an intact C4 unit was a precursor for the acetate derived C3 fragment, it would surely not all have been cleaved and at least some of the 13C4 precursor should have survived. In the case of lycopodine, however, we did see an opportunity to probe the status of acetoacetate as a precursor which did not require the feeding of a C4

precursor.We settled on an experiment with [1,2-13C2,2-2H3]acetate and decided to focus our attention on the acetate derived C3 unit C-16,C-15,C-14 which originates from pelletierine [52,53]. We reasoned that the labeling pattern in this portion of (27) suggested that half of the molecules were labeled according to Path C in Scheme 8 and half according to Path D. We expected that lycopodine molecules in which this portion is formed by Path C should retain at least some deuterium from [1,2-13C2,2-2H3]acetate at the C-16 methyl group. Specifically, this methyl group is derived from the acetate starter unit used in the putative Claisen condensation of acetyl CoA with malonyl CoA to yield acetoacetyl CoA. Retention of deuterium at C-2 of acetate starter units of polyketides is well pre-cedented. In the event, NMR analysis of a sample of lycopodine (27) from this experiment showed that it carried only 13C and no trace of 2H was observed. The 13C NMR spectra were qualitatively identical to those obtained after feeding of [1,2-13C2]acetate. This unexpected result could not be reconciled with Path C1 and we focused our attention on finding an entirely different interpretation of our observations.

We concluded from the outcome of our double label experiment with [1,2-13C2, 2-2H3]acetate that the carbon atom destined to become the C-16 methyl group of lycopodine had to be activated at some time during the biosynthetic process to such an extent that facile deuterium/protium exchange was possible. A derivative which would fulfil such a structural requirement might be malonic acid (or its CoA derivative) or an analogous intermediate. Careful reexamination of the NMR data from all the experiments indicated that the ratio of the two labelling patterns in the acetate derived C3 units of lycopodine was always exactly 1: 1 (within the accuracy of 13C integration). It was a conceptually crucial step when we came to realize that the observed 1: 1 ratio was probably not adventitious. Instead, this ratio may be a necessary consequence of the mechanism by which the acetate derived fragments of (27) were assembled. If the mixed incorporation patterns had been due to competition between two paths, some variation in this ratio would have to be expected over several repetitions of the experiment.

The most straightforward explanation for the two labeling patterns occurring in a 1 : 1 ratio is the presence of a symmetrical intermediate between acetate and the immediate precursor for the C3 unit, before the latter is joined to the imine (14). The structural prerequisites for such a putative intermediate were: firstly, a compound with C2v symmetry capable of delivering a C3 unit equivalent to acetone and, secondly, the two sites destined to become the methyl groups of this acetone equivalent had to be sufficiently activated to undergo ready exchange of the protons a to the keto group. These requirements are fulfilled by acetonedi-carboxylic acid or its bisCoA thioester (30), Scheme 15.

Upon Claisen condensation of two molecules of malonyl CoA with one another, a molecule of acetonedicarboxylic acid CoA ester would be formed which could either be hydrolyzed to the free acid or further activated to its bisCoA thioester. It is important that the derivative (30) have C2v symmetry to explain the labeling pattern in (27). The monoCoA ester would not fulfil this condition! After condensation of (30) with (14), the keto acid (29) would be obtained and would show two distinct labeling patterns depending on which of the methylene

Scheme 15. New proposal for the biosynthesis of pelletierine in Lyopodium tristachyum via acetonedicarboxylic acid

carbon atoms of (30) formed the bond to C-2 of (14). After decarboxylation of (29), or hydrolysis/decarboxylation, pelletierine (28) would be obtained, which also would show two distinct labeling patterns. Combinations of the isotopomers of (28) and (29) with each other according to Scheme 13 would then yield the experimentally observed labeling patterns.

We were exceedingly fortunate that the quintessential proof of this idea, namely the direct incorporation of acetonedicarboxylic acid into lycopodine was indeed observed when [2,3,4-13C3]acetonedicarboxylate was fed to L. tristachyum. In fact, the experiment was so successful when it was first attempted that the resulting data were not easily interpretable. In order to achieve maximum sensitivity for the experiment, the precursor had not been diluted with unlabelled material. This precaution proved to be unnecessary as a 1 % specific incorporation was observed. Careful analysis of the splitting patterns for C-15 and C-8 indicated that about 50% of the isotopically labeled molecules contained two labeled C3 units and thus showed interunit coupling. A repetition of this experiment one year later with precursor properly diluted with unlabelled material, again yielded labeled alkaloid and this time the coupling pattern was readily analyzed and found to be consistent with our expectations.

The current proposal for assembly of the carbon skeleton of lycopodine (27) [55] is as shown in Scheme 13 bottom. 4-(2-Piperidyl)-3-oxobutanoate (29) is the central intermediate in this scheme. This material may suffer decarboxylation yielding pelletierine (28) which is incorporated into the C-9 to C-16 fragment. The keto group at C-15 (lycopodine numbering) of that fragment is then condensed in a Knoevenagel type reaction with the methylene group of the ^-keto acid function of 4-(2-piperidyl)-3-oxobutanoate to form the C-15 to C-8 bond.

In an attempt to rationalize the underlying chemistry depicted in Scheme 13, one is bound to arrive at the same conclusion that was reached when the successful incorporation of (22) into tropane (1) was being considered. Apparently activation of the C-3' methyl group of pelletierine, destined to become C-8 of lycopodine (27), as a methylene placed between two carbonyl groups is required for successful formation of the C-15/C-8 bond.

0 0

Post a comment