Development of the Vinylogous Pictet-Spengler Cyclization and Total Synthesis of (±)-Lundurine A

Abstract

A novel vinylogous Pictet-Spengler cyclization has been developed for the generation of indole annulated medium sized rings. The method enables synthesis of tetrahydro-azocinoindoles with a fully substituted carbon center, which is prevalent in many biologically active alkaloids. The strategy has been applied to the total synthesis of (±)-lundurine A.

Graphical Abstract

The Pictet-Spengler cyclization of tryptamine-based substrates is arguably one of the most powerful strategies for assembling indole-annulated polycyclic structures (Scheme 1a).^1,2In the presence of Lewis or Brønsted acids, reactive iminium ion intermediates undergo a cyclization to furnish tetrahydro-β-carbolines, in some cases with a fully-substituted carbon center (C1). The utility of this venerable transformation is highlighted in several total syntheses of indole alkaloids and other biologically active compounds.³

Scheme 1.

Pictet-Spengler Cyclizations for Natural Product Synthesis.

Despite the prevalence of the tetrahydro-azocinoindole core in several natural products, examples of indole vinylogous Pictet-Spengler cyclizations for the synthesis of larger ring systems are rare (Scheme 1b).⁴ In particular, most alkaloids with this substructure display a fully substituted carbon center adjacent to the nitrogen (C2).⁵ Synthetic approaches to tetrahydro-azocinoindole containing natural products have relied on the assembly of the 8-membered ring and fully substituted carbon center in separate steps.⁶ We envisioned that a vinylogous Pictet-Spengler cyclization could serve as a more general and efficient approach by generating these two challenging structural features in a single transformation.

In this Communication, we report a TMSCl mediated vinylogous Pictet-Spengler cyclization for the generation of tetrahydro-azocinoindoles and demonstrate the broad substrate scope of this novel transformation.⁷ We also apply the vinylogous Pictet-Spengler cyclization to the total synthesis of (±)-lundurine A,⁸ an indole alkaloid that exhibits moderate cytotoxicity in drug-resistant human oral epidermoid carcinoma cells⁹ and has attracted the attention of other synthetic chemistry groups.¹⁰

Our retrosynthetic analysis for lundurine A (1) was based on the hypothesis that the tetrahydro-azocinoindole core with a fully-substituted carbon center (2) could arise from a vinylogous Pictet-Spengler cyclization of N-acyliminium ion 3 (Scheme 1c). This reactive intermediate could be accessed from hydroxylactam 4.

To initiate our studies in the vinylogous Pictet-Spengler cyclization, we explored a series of Lewis acids for the activation of tryptamine-based substrate 5, due to the plethora of examples of traditional Pictet-Spengler cyclizations that are mediated by this mode of activation (Table 1).² In the presence of 1.25 equivalents of TiCl₄, SnCl₄, BCl₃, or AlCl₃, hydroxylactam 5 was transformed into the desired tetrahydro-azocinoindole 6 in low to moderate yields (entries 1–4). When Me₃SiCl was utilized as the Lewis acid, the desired product was isolated in 82% yield (entry 5). The use of a catalytic amount of Me₃SiCl resulted in a decrease in yield (entry 6).

Table 1.

Discovery and Proposed Mechanism of TMSCl Mediated Vinylogous Pictet-Spengler Cyclization.

Reaction conditions: Hydroxylactam 5, Lewis acid (1.25 equiv), CH₂Cl₂ (0.1 M), 0 °C to 23 °C, 18 h. [a] NMR yield, 1,4-dimethoxybenzene as internal standard. [b] Isolated yield, 0.15 mmol scale. [c] 0.5 equiv Me₃SiCl.

For the mechanism of this novel transformation, we hypothesize that upon treatment with a Lewis acid, γ-hydroxylactam 5 is converted to reactive N-acyliminium ion 7,^11,12,13 which may furnish the observed product through two paths. Nucleophilic attack of the vinyl group of indole 7 onto the N-acyliminium ion would lead to tetracycle 8, followed by deprotonation to generate tetrahydro-azocinoindole 6 (path a). Alternatively, nucleophilic attack of indole 7 at the C3-position would lead to spirocycle 9, which could undergo a 1,4-shift to yield the tetrahydro-azocinoindole product (path b).

With the identification of Me₃SiCl as an efficient Lewis acid for the vinylogous Pictet-Spengler cyclization, we examined the scope of this transformation (Table 2). A broad range of functional groups could be incorporated into the side chain of hydroxylactam 10 without affecting the efficiency of forming tetrahydro-azocinoindole 12 (Table 2a). For example, alkyl chains (12b), phenyl rings (12a, 12e), functionalized aromatic rings (12f, 12g), heteroaromatic rings (12h), ethers (12i), and acetals (12j) were all compatible with the reaction conditions. In addition to forming vinylogous Pictet-Spengler cyclization products with a fully substituted carbon, we also generated the desired product with a secondary carbinamine (12d).

Table 2.

Substrate Scope of TMSCl Mediated Vinylogous Pictet-Spengler Cyclization.

Reaction conditions: Hydroxylactam 10, Me₃SiCl (1.25 equiv), CH₂Cl₂ (0.1 M), 0 °C to 23 °C, 18 h.^a 40 °C, 48 h.

We also explored substitution around the indole ring of hydroxylactam 10 (Table 2b). We observed product formation with methyl substitution (12k, 12l) as well as ether substitution (12m, 12n). Substitution around the unsaturated functional group at the C2-indole position resulted in the formation of a vinylogous Pictet-Spengler cyclization product with a trisubstituted alkene (12o). The cyclization was compatible with a range of protecting groups on the indole nitrogen, including p-methoxybenzyl (12o), methyl (12p), and allyl (12q). With an electron-withdrawing Boc-protecting group, the vinylogous Pictet-Spengler cyclization required mild heating to yield the desired product (12r).¹⁴

Once we demonstrated the substrate scope of the vinylogous Pictet-Spengler cyclization, we applied this novel transformation for the synthesis of the indole alkaloid (±)-lundurine A (Scheme 2). Tryptamine derivative 13 was accessed from 5-methoxytryptophol.¹⁵ Coupling of this intermediate with Grignard reagent 14 led to the formation of the substrate for the vinylogous Pictet-Spengler cyclizaion (15). Gratifyingly, in the presence of Me₃SiCl, hydroxylactam 15 was converted to tetrahydro-azocinoindole 16 in 70% yield. In a single chemical transformation, we assembled the desired 8-membered ring and fully substituted carbon center of lundurine A.

Scheme 2.

Total Synthesis of (±)-Lundurine A via a Vinylogous Pictet-Spengler Cyclization.

With the core structure of lundurine A in hand, we devised a strategy to complete the total synthesis of the indole alkaloid. A two-step global hydrogenation of tetracycle 16 followed by TBS-protection of the primary alcohol resulted in the formation of silyl ether 17. Carboxymethylation of the indole nitrogen with imidazole carbamate 18¹⁶ followed by TBAF mediated removal of the TBS group unveiled alcohol 19. Swern oxidation and treatment with Et₃SiCl under soft enolization conditions furnished silyl enol ether 20 as an inconsequential mixture of double bond isomers. Next, we performed a two-step tungsten catalyzed dehomologation reaction of silyl enol ether 20 to formally remove a methylene group in the side chain, which yielded aldehyde 22.¹⁷ This transformation presumably proceeds through formation of silyloxy epoxide 21 followed by oxidative cleavage. Conversion of aldehyde 22 to a hydrazone and TMSCl mediated cyclopropanation assembled the strained polycycle 23.^10h Finally, amide 23 was transformed into an α-phenylsulfonyl ketone followed by syn elimination to install the required α,β-unsaturation in the natural product. The spectroscopic data obtained for our synthetic sample of (±)-lundurine A (1) were identical with the data reported by Kam and co-workers in the original isolation paper.⁸

In conclusion, we developed a TMSCl mediated vinylogous Pictet-Spengler cyclization for the construction of tetrahydro-azocinoindoles. This method was successfully applied to the synthesis of (±)-lundurine A (1). The application of this novel reaction to the synthesis of other tetrahydro-azocinoindole containing alkaloids and the development of a catalytic enantioselective vinylogous Pictet-Spengler cyclization are on-going interests in our group.