Biosynthesis

1 Introduction

For ease of access to previous Reports in this series the practice introduced last year of listing them as the first references is continued as is the reference to the earlier comprehensive reviews. (As the volume of published work on alkaloid biosynthesis grows so too does the need for a new comprehensive treatise.)

The use of incorporation efficiencies as a measure of the relative importance of various precursors within a biosynthetic pathway must often be subject to considerable uncertainty since it is difficult to achieve identical conditions in successive feeding experiments. This problem is overcome if the precursors are fed together to the same plant (or culture), distinction between them being made by use of different isotopic labels e.g.14C and 3H. A problem with this approach is that, particularly, tritium (or deuterium) may be lost from one of the precursors, leading to a false result. An ingenious solution to this difficulty is to run a second experiment in which only the substrate which bears the tritium label in the first experiment is fed, but labelled now with 14C and 3H. The 14C:3H ratio observed in the metabolite then allows one to make adjustment for any tritium loss in the first experiment; one is in effect then comparing two sets of 14C incorporation data. [This method has been successfully applied in a study of pyrrolizidine biosynthesis where there was a quite spectacular variation in relative incorporation efficiencies but a fairly constant isotope ratio between the substrates examined (see p. 18)]. Further difficulties which these approaches do not, of course, overcome are uncertainties associated with the absorption and transport of a substrate to the site of biosynthesis. Normally the specific labelling of a metabolite by a precursor is taken as a firm indication that the substrate is involved in the biosynthesis of the metabolite. However, particularly with complex molecules, the specific labelling observed may be the result of fragmentation of the precursor and subsequent incorporation of a labelled fragment. This problem may be overcome very successfully if doubly labelled precursors are used (cf. ref. 12). It is worth noting that in experiments with precursors labelled with both 14C and 3H (normally made by mixing two samples) the 3H: 14C ratio found for the metabolite may be higher (up to 20% has been observed) than that expected because of preferential utilization of molecules labelled with 14C along other pathways (which may include faster conversion of 14C-labelled material through to the alkaloid and subsequent metabolism, so that at the end of the feeding experiment the more slowly metabolized 3H-labelled material is present in greater amount).

2 Piperidine, Pyridine, and Pyrrolidine Alkaloids

Securinine. — The eight carbon unit (normal bonding) of the Securinega alkaloid, securinine (3), is derived from tyrosine. The remaining C5N unit (heavy bonding) can be formed from labelled cadaverine in the expected manner. Lysine (1) was also found to be incorporated but the derived alkaloid was not degraded to determine whether securinine falls into the group of alkaloids which are derived from lysine in unsymmetrical fashion or those which have an obligatory genesis through a symmetrical intermediate (cadaverine). Recent work has established, however, that securinine falls into the former group and that Δ1-piperideine (2) is also implicated in its biosynthesis.

DL-[2-14C]Lysine [as (1)] and [2-14C]-Δ-piperideine [as (2)] afforded securinine (3) in which the label was confined essentially to the asterisked carbon atom. Further, [RS-6-3H; 6-14C]-DL-lysine [as (1)] gave securinine without loss of tritium. Consequently C-6 of lysine does not undergo oxidation in the course of securinine biosynthesis and so the ε-amino-group of (1) must be retained whilst the α-amino-group and carboxy-function are lost. The combined results are consistent with the hypothetical route to securinine shown in Scheme 1. This pathway will now gain more validity if alkaloids with structures similar to those of the proposed intermediates can be found in Securinega or related plants.

It is interesting to note that securinine (3) is the first among those alkaloids with a nitrogen atom common to two rings which is known to avoid a symmetrical intermediate (cf. ref. 17).

Lycopodine. — Lycopodine (4) is formed in Lycopodium tristachyum from lysine [via Δ1-piperideine (2)] and acetic acid. The labelling pattern observed with these precursors suggests that lycopodine (4) could be formed simply from two molecules of pelletierine (5) but in this case 'Occam's Razor' does not apply since radioactive pelletierine (5) labels only one of these 'pelletierine' units in lycopodine (4). The hypothetical pathway proposed to take account of these results has been tested in further experiments. Piperidine-2-acetic acid (6), as its CoA ester, is one of the hypothetical intermediates but the evidence obtained most strongly points to it not being involved in lycopodine biosynthesis, [carboxy-14C]Piperidine-2-acetic acid fed in admixture with DL-[4-3H]lysine [as (1)] gave lycopodine with a 3H: 14C ratio indicating that lysine is ca. 17 times more efficient as a precursor than (6) thus casting doubt on a biosynthetic role for (6). (This experiment is a further application of the useful procedure described on p. 1.) As expected [2-14C]malonic acid is incorporated in the same way as [2-14C]acetic acid (and ca. 14 times more efficiently than lysine, measured in a similar way to the above). Attempted trapping of piperidine-2-acetic acid (6) during metabolism of a mixture of [3H] lysine and [14C]malonate by also feeding inactive (6) gave material containing only tritium whereas the lycopodine was labelled by both precursors. [If (6) were an intermediate then it should have shown the same 3H:14C ratio as that observed for the lycopodine.] Finally an attempt to dilute the incorporation of [2-14C]malonate into C-6 and C-14 of lycopodine (4) with inactive (6) [in the hypothetical pathway these two carbons are derived from C-2 of malonate-acetate via the side-chain of (6)] also failed. It is a consequence of these results that the hypothetical pathway to lycopodine (4) must now be modified.

DL-[4,5-3H2, 6-14C]Lysine [as (1)] has been found to give lycopodine (4) with 22% loss of tritium. This was rationalized in terms of a route in which half the tritium is lost from C-5, and confirmed when it was shown that DL-[4-3H, 6-14C]lysine was incorporated without tritium loss. The results of further experiments with doubly labelled samples of lysine (DL-[4,5-3H2]-/D-[6-14C]-lysine and L-[4,5-3 H2]-DL-[6-14C]-lysine) have demonstrated that in L. trystachyum lycopodine (4) is derived from L-lysine whereas pipecolic acid (7) is derived largely, if not exclusively, from D-lysine. These results are analogous to those obtained in other plant species for piperidine alkaloids on the one hand and pipecolic acid on the other.

Pinidine and Coniine. — Pinidine (8) and the hemlock alkaloid coniine (9) are unusual among simple piperidine alkaloids in being derived exclusively from acetate units. The combination is in each case a simple linear one and proceeds either via polyketide intermediates or, arguably in the case of coniine, via the fatty acid (10). Strong evidence from tracer and enzyme studies points to (11) as an intermediate in coniine biosynthesis. Further enzyme studies relate to the N-methylation of coniine (9) and most recently to the conversion of γ-coniceine (12) into coniine (9). An enzyme which effects this step, γ-coniceine reductase, has been isolated from Conium maculatum leaves. It was found that it has an absolute requirement for NADPH and that hydride transfer occurs from the β (pro-S) face of the dihydropyridine nucleotide.

[1-14C]Acetate labels the alternate carbon atoms, C-2, C-4, C-6, and C-9 of pinidine (8). This, by analogy with the biosynthesis of coniine, leads to (13) or (14) as likely intermediates in the formation of this alkaloid. However, neither [10-14C]-5,9-dioxodecanoic acid [as(13)] nor [1-14C]nonadione [as(15)], the decarboxylation product of (14), labelled pinidine significantly. The in vivo transformation of octanoic acid into coniine suggested that decanoic acid might be a precursor for pinidine but [9-14C] and [1014C]-labelled materials were not significantly incorporated.

More positive information was obtained, however, when diethyl [1-14 C]malonate was fed to Pinus jeffreyi together with inactive sodium acetate. The activity at C-2 (9%) in the derived pinidine was much lower than that at C-9 (30%) from which it follows that C-2 and C-7 of pinidine (8) represent the 'starter' acetate unit and the carboxy-function lost in the course of biosynthesis must therefore be sited as in (13) rather than as in (14). It was suggested that the failure of (13) to act as a precursor might indicate that biosynthesis only proceeds via a Δ-derivative of this acid.

Coccinelline. — Coccinelline (16) is an alkaloid isolated from the defensive secretion of the Coccinellidae. [1-14C}Acetate has been found to be incorporated into this arthropod alkaloid with 16% of the activity confined to C-2 plus C-10. This is consistent with derivation via the polyketide (17) (or the alternative with the carboxy-group at what becomes C-10 of coccineiline).

Tropane Alkaloids. — It is clearly established that the tropic acid moiety (19), found in alkaloids such as hyoscyamine (20), is derived from all the carbon atoms of phenylalanine (for reviews see refs. 35 and 36). Important proof that the conversion of phenylalanine (18) into tropic acid (19) involves an intramolecular 1,2-shift of the carboxy-group of (18) or derivative, has come from the results of feeding [1,3-13C2]phenylalanine [as (18)], containing 81% of doubly labelled species, to Datura innoxia. (A 14C label was also included to facilitate the measurement of incoproation.) The scopolamine (21) and hyoscyamine (20) isolated both showed satellite resonances, corresponding to C-1 and C-2, with 13C — 13C coupling. This coupling demands the presence of two contiguous 13C atoms and can only have arisen as the result of intra-molecular rearrangement in the course of the transformation of phenylalanine into tropic acid. (Because of the inevitable dilution of isotopically labelled precursors in the plant with unlabelled endogenous material, inter-molecular rearrangement would have given tropic acid with essentially only one 13C label per molecule — at either C-1 or C-2.)

Cinnamic acid (24), a metabolite of phenylalanine, has been linked with tropic acid biosynthesis via its epoxide on chemical grounds but neither of these compounds have previously been found to act as a tropic acid precursor. Results with cinnamoyltropine were similarly negative. More recently examination of [2-14C]cinnamic acid [as (24)] as a precursor for the acid moieties of hyoscyamine (20), scopolamine (21), apohyoscine (22), and littorine (23) also gave negative results under conditions when other precursors were incorporated. However, an incorporation of [2-14C]cinnamic acid into atropine (20) at a level comparable with that of phenylalanine, has recently been recorded. Moreover the label was confined to C-3 of the tropic acid moiety as expected if phenylalanine was utilized via cinnamic acid. This is an important result not least in regard to the mechanism of the rearrangement involved in the conversion of (18) into (19), which is still unknown. No doubt workers in this field will re-examine the previous negative results in the light of this strikingly positive one. One point which may be of significance is that the tropic acid moiety of atropine (20) is racemic (the latest experiment) whereas the earlier experiments were with alkaloids containing optically active tropic acid residues.

Fuller details on the incorporation of phenyllactic acid (25) into tropane alkaloids have been published. This acid is a better precursor than phenylalanine for the acid fragments of (20), (21), and (22). The significance of this difference is doubtful; it may be the result of several causes, the simplest being more effective diversion of phenylalanine into other biosynthetic pathways. None the less phenyllactic acid (25) is clearly a precursor for tropic acid (19) and atropic acid [as (22)] since it is specifically incorporated and moreover labels the alkaloids in the same way as phenylalanine does.

It is known from labelling studies that C-1 and C-3 of phenylalanine (18) appear at C-1 and C-3 respectively of littorine (23). Not unexpectedly label from C-2 of (18) appears at C-2 of (25).

Phenylacetic acid has been reported as a precursor of tropic acid (19), but in a recent experiment it was found not to label hyoscyamine (20) or scopolamine (21). Nor was phenylalanine labelled which excludes incorporation of phenylacetic acid through this amino-acid.

The sequence by which the tropane nucleus as in (26) becomes hydroxylated to give alkaloids like meteloidine (27) and 3 α-tigloyloxy-tropane-7β-ol is unknown. The most recent evidence is conflicting. On the one hand 3α-tigloyloxytropane, doubly labelled as shown in (26), was found to be an intact and specific precursor for meteloidine (27) in Datura innoxia, although both the 3α-tigloyloxytropane (26) and the 3α, 6β-ditigloyloxytropane (28) isolated at the end of the experiment had suffered extensive hydrolysis and re-esterification. On the other hand, and in agreement with the observations on (26) and (28), neither 3α-tigloyloxytropane (26) nor valtropine (26; with saturated double bond) (2-methylbutanoic acid is a precursor for tiglic acid in Datura) were incorporated intact into (26), (28) and (29) in D. innoxia under apparently very similar conditions. This was followed by a different approach in which the sequence of hydroxylation and esterification was deduced by observing the relative activities of the tigloyl moieties of (26), (27), (28), and (29) after feeding tiglic acid in a short-term experiment. Interpretation of the results has led to the proposal of a possible sequence of reactions in which di- and tri-hydroxytropanes are formed by different pathways. However, the conclusions reached must be subject to considerable doubt since they depend on the tacit assumption that ester formation is not significantly reversible. Clearly this is not so in longer term experiments at least and a further complication may be that the rates of hydrolysis-esterification vary from compound to compound.

A mutual interconversion between scopolamine (21) and hyoscyamine (20) has been observed in D. innoxia.

Pyrrolizidine Alkaloids. — Both ornithine (30) and putrescine (31) are specific precursors of retronecine (34) and it is thus a reasonable assumption that arginine (32) might be an alternative source for the pyrrolizidine ring system either through hydrolysis to ornithine or by way of agmatine (33) to putrescine. Rigorous comparison of the relative efficiencies of incorporation of ornithine and arginine have shown that the assumption is correct: arginine is a specific precursor for the retronecine (34) moiety of senecionine in Senecio magnificus but it is slightly less efficiently incorporated than ornithine.

Lythraceae Alkaloids. — The rate at which quinolizidine alkaloids of the cryogenine (35) type are synthesized and degraded in Heimia salicifolia has been studied as has their sequence of appearance in growing plants. The results do not yet add to the preliminary data so far obtained on these alkaloids.

Tenellin. — Tenellin (36), produced by the fungi Beauveria tenella and B. bassiana, has its genesis in acetate, methionine, and phenylalanine. Of particular interest is the observation that the incorporation of phenylalanine involves a skeletal re-arrangement similar to the one which affords tropic acid (see above). It has recently been shown that tropic acid arises by an intra-molecular 1,2-shift of C-1 (phenylalanine numbering). In the same way DL-[1, 3-13C2]phenylalanine containing 81% of doubly labelled species afforded tenellin (36) which showed 13C — 13C coupling between C-4 and C-5. Since an inter-molecular rearrangement would have resulted in separation of the two 13C labels present in each molecule, the rearrangement must be an intra-molecular one. The apparently close resemblance of the rearrangements which afford tropic acid in plants and tenellin in fungi argues for a common rearrangement step in phenylalanine metabolism in plants and fungi and one which is not essentially related to the structure of the ultimate product. It is likely that information gathered on the biosynthesis of one of these metabolites will be of value in elucidating the biosynthesis of the other. In this regard one wonders if cinnamic acid is a precursor for tenellin in the light of the conflicting evidence on tropic acid biosynthesis discussed above.