Sequential Sonagashira and Larock indole synthesis reactions in a general strategy to prepare biologically active β-carboline-containing alkaloids.

A general synthetic approach to β-carboline-containing alkaloids was developed. Two consecutive palladium-mediated processes, a Sonagashira coupling and a Larock indole annulation reaction, are central to the method. The scope of the approach was investigated and found to be amenable for constructing a variety of biologically significant natural products and also for preparing substituted analogues for optimization and analysis of their biological properties.

Efficient palladium-mediated reactions allow easy access to highly substituted heterocycles, including indoles, one of the most significant and biologically active core structures in natural products.[1] We sought to apply Larock’s method of indole synthesis[2] to prepare 2-pyridyl-indoles 1 by the retrosynthesis shown in Scheme 1. The required Larock substrate 3 would arise from a Sonagashira reaction.[3] An expected advantage of this approach is that each of these consecutive coupling processes would likely tolerate significant structural diversity, a benefit for the optimization of biological activity.

Scheme 1

Pyridyl Indole Retrosynthesis

Under optimized Sonagashira conditions butyne-1-ol 6 and 2-bromopyridine 7 give alkyne 8 in nearly quantitative yield (Scheme 2). For the Larock indole annulation reaction the conditions of Senanayake et al.,[4] using bromoanilines rather than iodoanilines and also using only a small excess (1.2 equiv) of the alkyne, were generally suitable. A slight variation gave optimal results: 2.5 mol % Pd(OAc)2, 5 mol % 1,1′-bis(diphenylphosphino) ferrocene (dppf), 1 equiv of alkyne, KHCO3, DMF, 110 °C, 4 h.[5] Under these conditions the desired indole 10 was obtained in 95% yield with high regioselectivity.[6]

Scheme 2

Synthesis of Indolopyridocoline Triflate

The hydroxyethyl chain of compound 10 was originally intended to be used for installing diverse groups in the indole C3 side chain. Upon conversion of the alcohol 10 to a tosylate or triflate, a yellow precipitate, tetracycle 11 (Scheme 2), formed instantly. In retrospect this is not surprising, as Gribble and Johnson reported an analogous cyclization of a bromide,[7] while Fürstner et al. also described a similar cyclization of an aldehyde.[8] Optimization of conditions gave compound 11 in 94% yield. The efficiency and ease of operation for this three-step route to compound 11 (88% overall yield with purification by simple filtration) was intriguing, as is the structural similarity of product 11 to a number of biologically significant alkaloids. 11 is the dihydro analog of indolopyridocoline (12),[8] a natural product made in 90% yield from 11 through a DDQ-promoted oxidation.

Alkaloid 11 belongs to a very large family of diverse tetracyclic and pentacyclic natural products (Figure 1) including norketoyobyrine, rutaecarpine, deplancheine, isonauclefidine, javacarboline, and dihydroflavopereirine, structures that have attracted significant synthetic interest.[9]

Figure 1

Related β-carboline-containing natural products.

While each step in Scheme 2 is well-precedented, the sequential application of such high-yielding reactions to access members of this class of biologically active alkaloids in a general fashion is noteworthy, especially because it may facilitate rapid structure–activity relationship (SAR) studies in natural product scaffolds. Here, this approach to several members of this family of alkaloids and their substituted analogues is outlined.

A synthesis of the pyridone-containing natural product norketoyobyrine (Scheme 3) tested the generality of the approach. A Sonagashira reaction using isoquinolin-3-yl triflate 19(10) gave alkyne 20 in nearly quantitative yield. Larock indole synthesis cleanly gave indole isoquinoline 21 and then, upon cyclization, pentacycle 22. Norketoyobyrine (13) was then prepared using a mild oxidation procedure.[11] The intermediate pyridinium salt 22 was also reduced to the known semisaturated pentacycle 23.[11]

Scheme 3

Syntheses of Pentacycles Norketoyobyrine and Demethoxycarbonyldihydrogambirtannine

The synthesis of rutaecarpine tested whether a chloroquinazoline is a suitable substrate for the Sonagashira reaction and also presented the key issue of regioselectivity in the cyclization to a pentacycle (Scheme 4).

Scheme 4

Synthesis of Rutaecarpine

Methyl ether 24,[12] readily prepared in two steps, under Sonogashira conditions smoothly gave alkyne 25. Larock indolization gave a regioisomeric mixture favoring the expected 2-heteroaryl indole product 26. KHCO3 as a base gave an 8:1 preference for 26 in high isolated yield (82%). Tosylate formation prompted pentacycle formation. Interestingly, cyclization proceeded by the attack of the N1 rather than N3 nitrogen atom of the quinazoline ring, as established by an NOE study of 28. Chloride-promoted demethylation then quantitatively gave the pentacyclic product 29, an unknown isomer of rutaecarpine. Cyclization of 26 using HCl/n-butanol exhibited preference for rutaecarpine (16:1, 81% yield).

Tetra- and pentacyclic alkaloids with a nonaromatic ring annulated to the indole comprise a significant portion of this extended family of natural products. Often the ring distal to the indole is functionalized, as in deplancheine (Figure 1), which by our methods requires a suitably substituted pyridine as a Sonagashira substrate. As a model, the methoxybromopyridine 30 was found to undergo Sonagashira coupling to give alkyne 31 (Scheme 5). Larock annulation followed by cyclization gave the tetracycle 33 in 88% yield for the three steps. The related benzyl ether 34 has been converted to deplancheine (15).[13]

Scheme 5

Synthesis of Precursors to Deplancheine

Certain alkaloids require that electron-withdrawing groups be present in the Sonagashira substrate. This is accommodated; tert-butyl ester substituted chloropyridine 35 and butyne-1-ol gave alkyne 36 (Scheme 6). A Larock reaction and cyclization gave tetracycle 38, which was treated with acid to give isonauclefidine triflate (39).[14] The conversion of the intermediate tetracycle 38 to norepiisogeissoschizoate (40) has also been described.[15]

Scheme 6

Isonauclefidine and Norepiisogeissoschizoate

Substitution in the sp3 chain that undergoes cyclization is tolerated. Tetracycle 44, an analog of javacarboline (17),[16] was prepared in three steps and 60% overall yield from the alkyne 41(17) through hydroxyester 43 (Scheme 7).

Scheme 7

Synthesis of an Analogue of Javacarboline

Acetyl bromopyridine 45 (Scheme 8) is also a suitable substrate, giving in three steps a 63% yield of the tetracycle 48, a known precursor[18] to enone 49, a building block for heteroyohimbine type alkaloids.[19] The corresponding alcohol 50 has been converted to 6,7-dihydroflavopereirine and flavopereirine.[18] Presumably alcohol 50 would be easily accessed by Larock annulation and cyclization using alcohol 52 or by the reduction of tetracycle 48. The alcohol 52 indeed undergoes an efficient Larock indole synthesis reaction, as shown by its reaction with the electron-rich aniline 53 to give tetracycle 55, a precursor in four steps to mitragynine (56),[20] a natural product of interest due to its antinociceptive properties (Scheme 9).[21]

Scheme 8

Route to Flavopereirines

Scheme 9

Formal Synthesis of Mitragynine

Herein is described a general strategy to access tetra- and pentacyclic β-carboline-containing alkaloids, a large family of natural products of great interest in organic synthesis and intriguing for their biological properties. Five natural products were synthesized: indolopyridocoline, norketoyobyrine, demethoxycarbonyl dihydrogambirtannine, rutaecarpine, and isonauclefidine. Also shown are formal syntheses of norepiisogeissoschizoate and mitragynine as well as methods useful for the synthesis of deplancheine, javacarboline, 6,7-dihydroflavopereirine, and flavopereirine. Key features are sequential Pd-catalyzed coupling reactions, each with significant substrate tolerances. Six substituted pyridine chlorides, triflates, and bromides were suitable Sonagashira reaction substrates (7, 19, 24, 30, 35, and 45). Two bromoanilines were suitable Larock annulation substrates (9 and 53). Substitution in the central ring is allowed (44). The wide substrate tolerances suggest that using these methods for the synthesis of arrays of natural product analogs for the optimization of their biological properties and study of their mechanisms of action is feasible and such studies will be the subject of future reports from our laboratory.