Efficient Synthesis of the Plant Growth Regulator Ancymidol

The important cytochrome P450 inhibitor ancymidol is used as a plant growth retardant and has potential for various medicinal applications. However its high price sets economic limits to large-scale applications. Here a short and high-yielding synthesis is reported, providing ancymidol in substantial amounts in a costand time-efficient way.


Introduction
P lant growth retardants are applied to agronomic and horticultural crops to reduce unwanted longitudinal shoot growth without lowering plant productivity.Most growth retardants act by inhibiting gibberellin biosynthesis.To date, four different types of such inhibitors are known: (a) Onium compounds, such as chlormequat chloride, mepiquat chloride, chlorphonium, and AMO-1618, which block the cyclases copalyl-diphosphate synthase and ent-kaurene synthase involved in the early steps of gibberellin metabolism; (b) Structural mimics of 2-oxoglutaric acid, which is the co-substrate of dioxygenases that catalyze late steps of gibberellin formation.Acylcyclohexanediones, e.g.prohexadione-Ca and trinexapac-ethyl and daminozide, block particularly 3 beta-hydroxylation, thereby inhibiting the formation of highly active gibberellins from inactive precursors; (c) 16,17-Dihydro-GA(5) and related structures act most likely by mimicking the GA precursor substrate of the same dioxygenases; (d) Compounds with an Ncontaining heterocycle, e.g.ancymidol, flurprimidol, tetcyclacis, paclobutrazol, uniconazole-P, and inabenfide (Rademacher, 2000).The crucial molecular effect of these pyrimidine and triazole derivatives seems to be the inhibition of gibberellins biosynthesis, specifically at the oxidation of ent-kaurene (1) to ent-kaurenoic acid (2, scheme 1) (Coolbaugh et al., 1978).The transformation involves three oxidative steps, which all require NADPH and molecular oxygen, and are catalyzed by cytochrome P450 mono-oxygenases (Murphy and West, 1978).All three steps are inhibited by ancymidol with comparable efficiency (Coulson et al., 1984).
Furthermore, ancymidol was found to inhibit certain oxidations on steroid backbones, which led to extensive pharmaceutical testing of the molecule.On the basis of this lead structure, for example, researchers from the Eli Lilly laboratories tried to develop inhibitors for the aromatase of human estrogen biosynthesis with the goal of providing novel breast cancer therapeutics.Pyrimidine derivatives show a higher activity than their (partially) saturated analogues, and the hydroxyl and cyclopropyl moieties seem to be equally important for most observed effects (Davenport et al., 1968;Taylor et al., 1987).Cyclopropyl derivatives have repeatedly been described as suicide substrates for eukaryotic cytochrome P450s (Testa and Jenner, 1981;Ortiz-De-Montellano, 1988;Ortiz-De-Montellano and Reich, 1986) or were used for biosynthesis investigations (Patzelt and Robinson, 1993).

Synthesis of the P450 inhibitor ancymidol
Although ancymidol has recently become commercially available, the high price makes largescale applications a costly enterprise.The single published, industrial, synthetic approach, on the other hand, does not provide satisfying experimental details (Taylor et al., 1987).It appeared reasonable to establish a short synthesis on the basis of the arylation of p-anisyl cyclopropyl ketone (8).
The Friedel-Crafts acylation of anisole (3) with acryloyl chloride (4) gave the expected panisyl vinyl ketone (5) in only mediocre yields.Apart from polymerisation products, up to 40% of the formal hydrochlorination product 7 were isolated.This chloride 7, however, was available directly from anisole and 3-chloropropanoyl chloride (6) in more than 90% yield.Neither basecatalyzed nor thermal (Santelli and Bertrand, 1973) elimination of HCl from 7 would produce 5 in satisfactory yields, but when 7 was directly treated with trimethyloxosulfonium iodide (Corey and Chaykovsky, 1965) and two equivalents of potassium hydride, the desired p-anisyl cyclopropyl ketone (8) was obtained in up to 80% yield (scheme 2).A proposed one-pot formation of 8 by the reaction of 3 with 4-chlorobutanoyl chloride (Close, 1957) did not give synthetically useful results.
When the volume of solvent (DMSO) in the transformation of 7 into 8 was reduced, a byproduct, which could always be detected on tlc in the reaction mixture, became the main product.It could be purified and characterized, and the unexpected bicyclic structure 9 was assigned.It presumably arises from two molecules of the ketone 7 and one methyl group from the oxosulfonium salt (scheme 3).
For the formation of the cyclopropyl moiety, an in situ elimination of HCl from 7 can be envisaged.The double bond of the resulting unsaturated ketone 5 would then undergo a Michaeltype attack by dimethyloxosulfonium methylide 10, derived from trimethyloxosulfonium iodide by deprotonation.Ring closure can occur via an intramolecular substitution of the oxosulfonium group in 11 by the enolate carbon (Trost and Melvin, 1975;Block, 1981).When the concentration of 5 and trimethyloxosulfonium iodide in the solution becomes too high, the enolate in 11 might be protonated before the cyclization takes place.After deprotonation at the very acidic sulfonium bearing carbon atom C(4), 12 can react with a second molecule of 5 or 7.An intramolecular aldol reaction would reversibly form the six-membered ring of 14, which, as soon as the oxide is in 1,4trans position to the sulfonium group, can cyclize to give 9 (scheme 3).This rationalization, of course, is so far purely speculative.

PATZELT
The concluding nucleophilic addition of 5-lithiopyrimidine (15) onto the carbonyl group of 8 proceeded cleanly as long as the temperature was kept below -100°C.It was possible to generate 15 in situ by the addition of butyl lithium to a solution of freshly sublimed 5-bromopyrimidine ( 16) and 8. Ancymidol (17) was obtained in 73% yield after recrystallization (scheme 4).It should be mentioned that even minor warming of the reaction mixture or minute impurities in the starting material led to considerable reductions in the yield.
Ancymidol can thus be prepared in a very straight-forward and efficient synthesis in 55% overall yield from inexpensive anisole, and is now available in multigram quantities for biological experiments.

General
All solvents and reagents were purchased from Aldrich or Fluka in their highest available quality.Solvents were distilled before use and dried over an appropriate desiccant (Perrin and Armarego, 1988).Reactions under anhydrous conditions were performed in a N 2 -atmosphere using standard Schlenck techniques (Casey et al., 1990).
Scheme 3. Proposed formation of the bicyclic side product 9 during the ancymidol synthesis.