The chemist's ability to make molecules of utmost complexity1 must not hide the fact that the practicability of many such syntheses is still low. The arithmetic demon inherent to any linear sequence constitutes one of the major hurdles in this regard. To overcome this obstacle new methodology and improved retrosynthetic logic are called for which allow more than one bond-making event to be integrated into a single synthetic operation.2 The approaches to the odoriferous alkaloid (R)-(+)-muscopyridine (1), derived from the animal kingdom, and its naturally occurring nor-analogue 2 outlined below tackle this theme and illustrate how priority can be given to the “economy of steps”3 by a highly orchestrated catalysis-based process.
Following its isolation by Ruzicka and Prelog,4 the unusual meta-pyridinophane derivative 1 has been repeatedly targeted.5–7 Despite its rather simple structure, however, none of the reported syntheses is fully satisfactory, being either unduly lengthy and/or poor yielding.8
Our approach to the alkaloid 1 takes advantage of the favorable application profile of an iron-catalyzed alkyl–aryl cross-coupling reaction recently developed in our laboratory as a powerful alternative to established organopalladium chemistry; in the present case the iron–salen complex 3
was applied.9, 10 The method not only allows one to replace expensive precious metal complexes by cheap iron salts, but it is also distinguished by unprecedentedly high reaction rates even at or below room temperature. While aryl chlorides as well as aryl sulfonates qualify as the substrates, the latter are usually distinctly more reactive. This chemoselectivity pattern constitutes a major design element of the synthesis depicted in Scheme 1.
Specifically, reaction of the commercially available bromo alcohol 4 with H2CCHMgBr in the presence of catalytic amounts of CuI affords the rather volatile alkene (X=OH),11 which is immediately converted without further purification into tosylate 5 to minimize losses during workup (65 % over both steps). Subsequent treatment with LiBr in acetone affords the corresponding bromide 6. The enantiomerically pure Grignard reagent derived therefrom is then added by syringe to a solution of the difunctional pyridine derivative 7 and the iron–salen complex 3 (5 mol %) in THF/NMP at 0 °C. Monitoring of the reaction by TLC reveals that substrate 7 is completely consumed after 20 min; compound 9 forms selectively by reaction at the triflate site, together with small amounts of the dialkylation product 10. However, rather than taking recourse to workup and purification at this point, a solution of 6-heptenylmagnesium bromide and additional 3 (5 mol %) were introduced and stirring was continued for 30 min at 0 °C. The ensuing second iron-catalyzed cross-coupling reaction affords the desired diene 11, which is contaminated with compound 10 formed in the first step (80 % combined yield, 11:10≈4:1). Again, no attempt was made to separate these products that are indistinguishable by TLC; instead, a much more convenient chemical “self-clearance” occurs as an integral part of the subsequent metathesis cascade.
Conversion of the product mixture of 11 and 10 to the corresponding hydrochloride salts12 followed by treatment with the readily available ruthenium–indenylidene complex 13
(10 mol %)13, 14 under high dilution conditions leads to the selective cyclization of diene 11 to the meta-pyridinophane 12 (E,Z≈1.8:1), while the hydrochloride of compound 10 remains unchanged. This reflects the greater ease of formation of the 13-membered ring as compared to the kinetically and thermodynamically handicapped 11-membered analogue that would derive from 10 by ring-closing metathesis (RCM).15, 16 After most of the solvent has been distilled off, however, residual 10 is forced to polymerize by acyclic diene metathesis (ADMET)17 and thereby completely disappears from the GC trace while the macrocycle persists. Evidently, the indenylidene catalyst 13 does not lead to any noticeable reopening of the cyclic olefin 12 once it is formed, although RCM is a priori a reversible transformation.18 Therefore this transformation constitutes a noteworthy illustration for selectivity by catalyst tuning which allows one to rigorously distinguish between terminal and disubstituted double bonds.
At this point, workup is further postponed and advantage is taken from the multifaceted behavior of the ruthenium catalyst 13. Thus, the crude mixture containing the cyclic monomer 12, the ADMET-polymer, and the still intact metathesis catalyst is transferred into an autoclave and is stirred under H2 (50 atm) overnight. Thereby, the carbene complex converts into a ruthenium hydride species, which acts as an efficient hydrogenation catalyst.19 Passing the mixture through a short pad of silica suffices to give analytically pure (R)-(+)-muscopyridine 1 in 57 % yield, while the polymeric by-product is retained on top of the column. This streamlined approach to enantiopure 1 together with the most convenient workup outperforms all previous approaches to this natural musk by virtue of its brevity and highly integrated character.
The adaptation of this concept to the synthesis of the achiral nor-analogue 2 is straightforward (Scheme 2). Treatment of substrate 7 with an excess of 5-hexenylmagnesium bromide20 in the presence of catalytic amounts of complex 3 as the precatalyst provides the dialkylation product 14 in 75 % yield after a reaction time of only 20 min. Exposure of the corresponding hydrochloride to the ruthenium–indenylidene complex 13 in a dilute CH2Cl2 solution delivers cycloalkene 15. Subsequent hydrogenation (50 atm H2) of the crude product affords the targeted 10-[2,6]-pyridinophane 2 in 68 % overall yield.
In summary, these syntheses feature several notable aspects: They show the exceptional performance of the emerging iron-catalyzed alkyl–aryl cross-coupling method, which enables the construction of polysubstituted arenes by serial “three-component coupling” in one pot at unusually high reaction rates. Moreover, the syntheses of 1 and 2 illustrate several ways in which one can benefit from the subtleties of olefin metathesis chemistry: not only is it possible to rigorously distinguish between medium-sized rings and macrocycles as well as between terminal and internal double bonds, but it is also possible to perform “tandem catalysis” events (RCM/ADMET + hydrogenation) with a single ruthenium–carbene complex as the catalytically competent precursor.19 Finally, the muscopyridine case study exemplifies an unprecedented “chemical” self-clearance routine that exploits the complementarity between RCM and ADMET. Further work aimed at a rational use of these and related aspects is underway.