The Lycopodium genus is well known for producing alkaloids with novel structures (Fig. 1) [1,2,3,4,5,6,7,8,9,10]. Many of these natural products have important biological activities. For example, lycopodine (1), the first alkaloid isolated from the Lycopodium family [11], triggers apoptosis of cancer cells by modulating 5-lipoxygenase and serves as a potential candidate for anticancer drugs. Huperzine A (2) is an acetylcholinesterase inhibitor that can cross the blood–brain barrier; [12] thus, it is used to treat patients with neurodegenerative diseases. The alkaloid lycojapodine A (6) not only shows inhibitory activity against acetylcholinesterase but also demonstrates anti-HIV activity with EC50 (half-maximal effective concentration) value of 85 µg/mL [13]. Nevertheless, the biological activities of many other Lycopodium alkaloids are not known due to the scarcity or unavailability of the natural products for investigation, such as lycopladine H (7) [14]. The structural diversity of Lycopodium alkaloids led Ayer and Trifonov [4] to categorize these alkaloids into four classes: the lycodine class, the lycopodine class, the fawcettimine class, and the miscellaneous class. The novel structures of Lycopodium alkaloids make these natural products excellent candidates for total synthesis. Indeed, the Lycopodium alkaloids have attracted a number of chemists to explore different synthetic routes. Representative work include Ayer’s [15] and Stork’s [16] syntheses of lycopodine (both in 1968), Heathcock’s (1986) [17] and Lei’s (2012) [18] syntheses of fawcettimine, and Sarpong’s [19] and Siegel’s [20] syntheses of complanadine A (2010).

Fig. 1
figure 1

Selected Lycopodium alkaloids

We envisioned a transannular strategy to access members from different Lycopodium categories. As shown in Scheme 1, lycopodine (1, from the lycopodine class), lycopladine H (7, from the miscellaneous class), and fawcettimine (3, from the fawcettimine class) could be derived from three bicyclic carbonyl compounds (8, 9, and 11, respectively) via transannular Mannich reactions [21]. Compounds 8, 9, and 11 are conceptually different oxidation products of the key intermediate 12, and the olefin functionality in 12 would provide an excellent opportunity for ring-closing metathesis retrosynthetically. Compound 13 should be easily accessed from simple starting material 14 [22]. Therefore, different classes of Lycopodium alkaloids could potentially be synthesized by taking good advantage of the key intermediate 12. Evans and Scheerer [23] reported the synthesis of clavolonine using a similar strategy. Inspired by this transannular Mannich approach, here we report our progress toward lycopodine synthesis using a similar transannular strategy.

Scheme 1
scheme 1

The initial strategy design using transannular Mannich reactions from a common intermediate and the key step in Evans’ synthesis of clavolonine

As shown in Scheme 2, we started our preparation of compound 12 by iodination of commercially available alcohol 15, leading to iodide 16 in 72% yield [24]. N-alkylation of Boc-protected allylamine with 16 gave carbamate 17 in 57% yield. Hydroboration of carbamate 17 on the less hindered alkene was followed by B-alkyl Suzuki coupling with compound 14 to afford enone 18 in 74% yield [25]. Conjugate addition was conducted with the aid of LiCl, forming a silyl enol ether [26]. The crude silyl enol ether was directly deprotected with tetrabutylammonium fluoride, and ketone 19 was formed in 68% yield over two steps. Next, a ring closing metathesis (RCM) reaction was explored to connect the two carbon chains of compound 19. Although the desired product could be obtained using Grubbs’ second-generation catalyst in refluxing dichloromethane as the solvent, this RCM reaction gave a better result when toluene, a solvent with a higher boiling point, was used with only 10 mol% Grubbs’ second-generation catalyst [27]. Thus, bicyclic ketones 20 was formed in 33% yield, together with other isomers in 23% yield. X-ray crystallography confirmed the structure of 20.

Scheme 2
scheme 2

Synthesis of metathesis product 20

Given the fact that the olefin in the macrocycle of 20 is electronically unbiased, we decided to employ the conformational bias of 20 to differentiate the two olefinic carbons. Specifically, we explored the possibility of synthesizing the lycopodine framework using compound 20 through a transannular aza-Prins cyclization via intermediates 23 and 24 (Scheme 3). Therefore, compound 20 was subjected to acidic conditions with elevated temperatures for the transannular cyclization. Nonetheless, none of the conditions we tried led to the formation with a structure like 22 in a detectable amount. These results could result either from unsuccessful formation of iminium 24 or from unsuccessful C–C bond formation in the aza-Prins step. In order to probe the formation of iminium 24, we removed the Boc group of compound 20 under thermal conditions, and the putative compound 23 was subjected to reductive amination conditions. However, we could not detect the formation of compound 25, which indicated that the iminium formation step needed further investigations.

Scheme 3
scheme 3

Attempts for lycopodine synthesis

In conclusion, an advanced intermediate 20 was synthesized in a straightforward way. This intermediate could potentially be useful in the synthesis of lycopodine via transannular cyclization strategies.