Stereocontrolled synthesis of the aconitine D ring from D-glucose

Abstract

The synthesis of a fully oxygenated aconitine D ring precursor from (D)-(+)-glucose is described. The route features a highly diastereoselective alkynyl Grignard ketone addition and a base-mediated enelactone to 1,3-diketone rearrangement.

---

Introduction

Aconitine 1, a C₁₉ norditerpenoid alkaloid isolated from plants of the genus Aconitum,¹ is a potent neurotoxin that exerts its principal pharmacological effect by binding to and activating voltage-gated Na⁺ channels.² Its poisonous properties have gained a certain literary notoriety, featuring in the works of authors as diverse as Ovid, Shakespeare, and Arthur Conan Doyle.³ Related diterpenoid alkaloids display significant antagonist activity at the nicotinic acetylcholine receptor⁴ and cytotoxicity.⁵ The densely functionalized cage structure, containing 15 stereogenic centres, has attracted interest from the synthetic chemistry community.⁶ Despite landmark syntheses of simpler related compounds,⁷ no total synthesis of aconitine has yet been reported, perhaps a reflection of additional stereochemical complexity present throughout the molecule, particularly concentrated in ring D where all six ring atoms are stereogenic centres.⁸ A retrosynthetic strategy predicated upon disconnections across the C(7)–C(8) and C(10)–C(11) bonds back to a CD subunit (Scheme 1) is appealingly convergent, and has been elegantly showcased by Gin^7b and Reisman⁹ in their respective syntheses of neofinaconitine and talatisamine. Noting that five of the six aconitine D ring stereocentres bear oxygen substituents, we envisioned assembly of this key subunit in enantiomerically pure form from a carbohydrate precursor, by exploiting the potential for Ferrier carbocyclisation¹⁰ to convert a pyranose-derived enol ether such as 5 to the corresponding cyclohexanone 4. The additional two carbons necessary to construct the C-ring could then be supplied by addition of an alkynyl metal to ketone 6, setting the stage for ring C closure via Conia-ene intramolecular diketone alkynylation.¹¹ Given that an alkynyl iodide (cf.3, X = I) has been shown to be a competent reactant in this type of cyclisation,^11a the potential to access vinyl halide 2 (or an organometallic derivative thereof) and elaborate it towards the target via cross-coupling is apparent. Benzylidene acetal 6 (R = Me) is an attractive substrate given the rigidity imparted by the trans-fused bicyclic scaffold; moreover, selective reduction of the acetal¹² would maintain a protecting group on the 2° alcohol whilst revealing the 1° alcohol en route to enol 5.

Scheme 1 Retrosynthetic plan.

Towards the realization of this concept, we report here a stereocontrolled synthesis of diketone 3 (X = SiⁱPr₃; P = Bn) from (D)-(+)-glucose as a platform for further synthetic studies towards 1.

Results and discussion

From the outset we recognized that the configuration at the anomeric centre of 6 would influence the stereochemical outcome of nucleophilic attack at the adjacent ketone. Treatment of stannylene acetal β-8 (from benzylidene acetal β-7)¹³ with methyl iodide gave high selectivity for 2° alcohol β-9, whereupon oxidation under Albright–Goldman conditions delivered the desired 3-OMe isomer β-6 (Scheme 2). On the other hand, methylation of α-8 took place selectively at the 2-position (giving α-10), requiring temporary protection of the 2-hydroxyl group as a benzoate ester to access α-6via α-11 as described by Fuchs.¹⁴

Scheme 2 Preparation of ketoglycosides α-6 and β-6.

With both α-6 and β-6 in hand we turned our attention to the installation of the alkyne moiety at C(2). Mikami and Shin reported that whereas addition of methyl magnesium iodide to analogous ketone 12 proceeded via exclusive axial attack to give 13, nitromethane gave the product of equatorial attack 14 (Scheme 3).¹⁵ As the latter outcome would set the correct configuration and install a group capable of manipulation to the alkynyl group of 4, we reacted α-6 with nitromethane under the same conditions but obtained an inseparable mixture of diastereomeric products enriched in the undesired (2R)-isomer 15 (Scheme 4). Under the same conditions, β-6 did give the desired product of equatorial attack 16 as a single stereoisomer, but we were unable to elaborate this intermediate further towards the target.

Scheme 3 Stereochemical outcomes of nucleophilic additions to ketone 12 (Mikami & Shin).¹⁴

Scheme 4 Addition of nitromethane to α-6 and β-6.

The high selectivity for equatorial attack on β-6 prompted us to examine addition of alkynyl metals to these substrates. We were delighted to observe complete selectivity for axial attack of bromomagnesium acetylide upon α-6, giving 17 in 84% yield; moreover, the desired product of equatorial attack was obtained as a single diastereomer by exposing β-6 to the same reagent, giving 18 in 85% yield (Scheme 5). These results parallel those of Miljković for sodium borohydride-mediated reduction of these substrates and may be rationalised by minimisation of dipole–dipole interactions in an early transition state.¹⁶

Scheme 5 Addition of alkynyl Grignards to α-6 and β-6.

The relative configuration of these compounds was tentatively assigned with the aid of NOE studies‡ and confirmed for compound 18 by single crystal X-ray diffractometry (Fig. 1). Unsurprisingly, attempted reductive cleavage of the benzylidene acetal of 18 using di-isobutyl aluminium hydride (DIBALH) resulted in hydroalumination of the alkyne, necessitating installation of protecting groups at both the alcohol and alkyne terminus. This was accomplished by treating β-6 with triisopropylsilylalkynylmagnesium chloride and benzylating the resultant 3° alcohol 19 to give 20 (Scheme 5). Treatment of 20 with DIBALH in toluene¹² afforded 1° alcohol 21 with complete regioselectivity and no evidence of alkyne reduction. Base-induced elimination of iodide 22 (obtained from alcohol 21via an Appel reaction) gave enol ether 5 (Scheme 6). Gratifyingly, upon exposure to mercury(II) chloride, 5 underwent Ferrier carbocyclisation to give cyclohexanone 4 in 98% yield as a single diastereomer.§ The stereoselectivity observed is in line with that observed by Machado and rationalized by a chelated chair-like transition state.¹⁷ A wide range of oxidising agents and conditions was surveyed in an effort to convert 4 to diketone 3, but without success. We therefore considered the possibility that the correct oxidation state at this carbon could be set earlier in the synthesis, envisioning that diketone 3 might be derived directly by rearrangement of enelactone 23,¹⁸ accessible in turn from 22via selective O-demethylation and oxidation of lactol 24 (Scheme 7).

Fig. 1 ORTEP representation of compound 18 (thermal ellipsoids at 30% probability).

Scheme 6 Preparation of β-hydroxyketone 4 from glucopyranoside 20 via Ferrier carbocyclisation.

Scheme 7 Revised retrosynthesis.

Unfortunately, we were unable to identify conditions to selectively 1-O-demethylate either 21 or 22. A second-generation route was therefore developed, where the refractory methyl group was replaced with an allyl group. However, this apparently minor structural change resulted in a reversal in O-methylation regioselectivity (cf.Scheme 2), so a more fundamental route redesign was carried out. To circumvent reliance on the capricious stannylene acetal chemistry, 1,2,5,6-di-O-isopropylidene-α-D-glucofuranose 25 (preparable from D-glucose in a single step)¹⁹ was chosen as a starting point. Methylation of the single exposed hydroxyl group gave 26,²⁰ which was converted to tetraacetate 27, obtained as an inconsequential 85 : 15 mixture of anomers in 63% overall yield. Selective de-acetylation of 27 at the anomeric position with hydrazine,²¹ followed by treatment with base and allyl bromide²² generated an 87 : 13 (β : α) ratio of epimeric allyl glycosides, from which the desired β-anomer 28 was isolated in 44% unoptimised yield following methanolysis of the remaining acetate groups. Benzylidene acetal formation gave 2° alcohol 29. Oxidation, alkynyl Grignard addition, and 3° alcohol protection paralleled the methyl glycoside series, giving 30; pleasingly, complete stereoselectivity for equatorial addition to the ketone was also observed (Scheme 8). Following DIBALH reduction of the benzylidene acetal, alcohol 32 was converted to iodide 33a. Palladium-catalyzed deallylation²³ and hemiacetal oxidation proceeded smoothly to give lactone 34a in 62% yield; however, base-induced elimination of HI was unsuccessful. Alcohol 32 was therefore converted to selenide 33b using Grieco's protocol²⁴ and processed along the same lines to give lactone 34b. Oxidation of the selenide²⁵ triggered spontaneous elimination to give the desired enelactone 23, which upon treatment with sodium methoxide rearranged cleanly to diketone 3, isolated as a mixture of diketone and ketoenol isomers in 67% yield (Scheme 8).

Scheme 8 Synthesis of aconitine D ring precursor 3 from D-glucose.

Conclusions

The synthesis of diketone 3 from 1,2,5,6-di-O-isopropylidene-α-D-glucofuranose 25 (one step from glucose) was accomplished in 14 steps and 2.5% overall yield (3.9% BRSM). The route features a highly stereoselective alkynyl Grignard addition to a ketopyranose controlled by the configuration at the anomeric centre, and an unusual enelactone to diketone rearrangement. Compound 3 represents a promising staging post in the proposed approach to aconitine, given that β-dicarbonyl moieties have been profitably employed by Toste as the nucleophilic component in gold-catalysed Conia-ene cyclisations (including a 5-endo-dig process to furnish a [3.2.1]-bridged product),^11a and replacement of the triisopropylsilyl group with a more tractable iodine substituent is feasible with either alkyne (3) or alkene (2) substrates.²⁶ Moreover, the product of such a reaction would contain carbonyl groups of different reactivity: the ketone embedded in the more strained 5-membered ring would be expected to exhibit greater reactivity towards reducing agents, rendering compound 2 (Scheme 1) potentially accessible via this strategy. Efforts in our laboratory are currently focused on applying these transformations to generate enantiopure CD building blocks as the foundation of a convergent approach to aconitine itself.

Author contributions

JD and IAP carried out all synthetic experimental work under the supervision of DMG and JBS. X-ray crystallography was carried out and analysed by CRR. DMG conceived the ideas and wrote the manuscript. DMG, JD and IAP conceived and designed the reactions. All authors edited and approved the final manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors thank the Leverhulme Trust (Research Grant RPG-2014-115) and the University of Huddersfield for support of this project.

References

1. (a) A. M. A. Qasem, Z. Zang, M. G. Rowan and I. S. Blagborough, Nat. Prod. Rep., 2022, 39, 460–473; (b) F.-P. Wang, Q.-H. Chen and X.-Y. Liu, Nat. Prod. Rep., 2010, 27, 529–570.
2. (a) A. Ameri, Prog. Neurobiol., 1998, 56, 211–235; (b) M. H. Benn and J. M. Jacyno, in Alkaloids Chemical & Biological Perspectives, ed. S. W. Pelletier, J. Wiley & Sons, Inc., New York, 1983, vol. 1, pp. 153–220.
3. (a) P. Ovidius Naso, Metamorphoses, I, 147, Penguin Classics, London, 2004; (b) W. Shakespeare, 2 Henry IV 4.4.49–50, OUP, Oxford, 2008; (c) A. K. Chatterjee, Can. J. Health History, 2022, 39, 281–310.
4. M. A. Turabekova, B. F. Rasulev, F. N. Dzhakhangirov, D. Leszczynska and J. Leszczynski, Eur. J. Med. Chem., 2010, 45, 3885–3894.
5. (a) M.-Y. Ren, Q.-T. Yu, C.-Y. Shi and J.-B. Luo, Molecules, 2017, 22, 267–281; (b) K. Wada and H. Yamashita, Molecules, 2019, 24, 2317–2338.
6. (a) R. M. Conrad and J. DuBois, Org. Lett., 2007, 9, 5465–5468; (b) C. Dank, R. Sanichar, K.-L. Choo, M. Olsen and M. Lautens, Synthesis, 2019, 51, 3915–3946.
7. (a) K. Wiesner, T. Y. R. Tsai, K. Huber, S. E. Bolton and R. Vlahov, J. Am. Chem. Soc., 1974, 96, 4990–4992; (b) Y. Shi, J. T. Wilmot, L. U. Nordstrom, D. S. Tan and D. Y. Gin, J. Am. Chem. Soc., 2013, 135, 14313–14320; (c) Y. Nishiyama, Y. H. S. Yokoshima and T. Fukuyama, J. Am. Chem. Soc., 2014, 136, 6598–6601; (d) C. J. Marth, G. M. Gallego, J. C. Lee, T. P. Lebold, S. Kulyk, K. G. M. Kou, J. Qin, R. Lilien and R. Sarpong, Nature, 2015, 528, 493–498; (e) K. G. M. Kou, B. X. Li, J. C. Lee, G. M. Gallego, T. P. Lebold, A. G. DiPasquale and R. Sarpong, J. Am. Chem. Soc., 2016, 138, 10830–10833; (f) D. Kamakura, H. Todoroki, D. Urabe, K. Hagiwara and M. Inoue, Angew. Chem., Int. Ed., 2020, 59, 479–486.
8. For alternative approaches to construction of the aconitine D ring, see: (a) H. Cheng, L. Xu, D.-L. Chen, Q.-H. Chen and F.-P. Wang, Tetrahedron, 2012, 68, 1171–1176; (b) X.-H. Zhou, Y. Liu, R.-J. Zhou, H. Song, X.-Y. Liu and Y. Qin, Chem. Commun., 2018, 54, 12258–12261; (c) G. Yang, M. Zhu, X. Zhao, J. Lei and L. Xu, Tetrahedron Lett., 2021, 69, 152975.
9. A. R. Wong, N. J. Fastuca, V. W. Mak, J. K. Kerkovius, S. M. Stevenson and S. A. Reisman, ACS Cent. Sci., 2021, 7, 1311–1316.
10. (a) M. E. Jung and S. W. T. Choe, J. Org. Chem., 1995, 60, 3280–3281; (b) R. J. Ferrier and S. Middleton, Chem. Rev., 1993, 93, 2779–2831.
11. (a) S. T. Staben, J. J. Kennedy-Smith and F. D. Toste, Angew. Chem., Int. Ed., 2004, 43, 5350–5352; (b) D. Hack, M. Blümel, P. Chauhan, A. R. Philipps and D. Enders, Chem. Soc. Rev., 2015, 44, 6059–6093.
12. N. Tanaka, I. Ogawa, S. Yoshigase and J. Nokami, Carbohydr. Res., 2008, 343, 2675–2679.
13. (a) P. V. Murphy, J. L. O'Brien, L. J. Gorey-Feret and A. B. Smith III, Tetrahedron, 2003, 59, 2259–2271; (b) H. Qin and T. B. Grindley, J. Carbohydr. Chem., 1996, 15, 95–108.
14. M. Srinivasarao, T. Park, Y. Chen and P. L. Fuchs, Chem. Commun., 2011, 47, 5858–5860.
15. J. Yoshimura, K. Mikami, K.-I. Sato and C.-G. Shin, Bull. Chem. Soc. Jpn., 1976, 49, 1686–1689.
16. M. Miljković, M. Gligorijević and D. Miljković, J. Org. Chem., 1974, 39, 2118–2120.
17. A. S. Machado, D. Dubreuil, J. Cleophax, S. D. Gero and N. F. Thomas, Carbohydr. Res., 1992, 233, C5–C8.
18. (a) N. P. Shusherina, T. Kh. Gladysheva and R. Ya. Levina, Zh. Org. Khim., 1966, 2, 1801–1805; (b) H. Schneider and C. Lüthy, Process for the preparation of cyclic diketones, Eur. Pat. Appl., 2003, 1352901; H. Tomioka, K. Oshima and H. Nozaki, Tetrahedron Lett., 1982, 23, 99–100.
19. R. N. Mitra, K. Show, D. Barman, S. Sarkar and D. K. Maiti, J. Org. Chem., 2019, 84, 42–52.
20. S. Yan and D. Klemm, Tetrahedron, 2002, 58, 10065–11071.
21. G. Excoffier, D. Gagnare and J. P. Utille, Carbohydr. Res., 1975, 39, 368–373.
22. (a) A. S. Henderson, J. F. Bower and M. C. Galen, Org. Biomol. Chem., 2014, 12, 9180–9183; (b) A. Lubineau, S. Escher, J. Alais and D. Bonnaffé, Tetrahedron Lett., 1997, 38, 4087–4090.
23. J. Zeng, S. Vedachalam, S. Xiang and X.-W. Liu, Org. Lett., 2011, 13, 42–45.
24. P. A. Grieco, S. Gilman and M. Nishizawa, J. Org. Chem., 1976, 41, 1485–1486.
25. H. Yamada, M. Adachi and T. Nishikawa, Chem. Commun., 2013, 49, 11221–11223.
26. D. L. Usanov, M. Naodovic, M. Brasholz and H. Yamamoto, Helv. Chim. Acta, 2012, 95, 1773–1789; N. Kerisit, R. Ligny, E. S. Gauthier, J.-P. Guégan, L. Toupet, J.-C. Guillemin and Y. Trolez, Helv. Chim. Acta, 2019, 102, e1800232.

---

Footnotes

† Electronic supplementary information (ESI) available: Full experimental details and characterisation data for novel compounds, X-ray crystallographic data and NMR spectra. CCDC 2343856. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4ob00561a

‡ The NOESY spectrum for compound 17 displayed a cross-peak corresponding to an interaction between 2-OH and axial 3-H, whereas for compound 18, a cross-peak for 2-OH and axial 4-H was observed. For compound 16, a cross-peak for 2-OH and axial 4-H was also observed, along with one between axial 3-H and one CH₂NO₂ proton. NOESY spectra of compound 20 and other 2-OBn compounds evidenced interaction between the CH₂Ph protons and the –SiⁱPr₃ group but not 4-H, suggesting that these compounds adopt a conformation where the bulky benzyl group is oriented away from the tetrahydropyran ring.

§ The ¹H NMR coupling constants for 2-H [δ_H 4.31 (d, J = 12.0 Hz)] and 3-H [δ_H 3.85 (d, J = 12.0 Hz)] of compound 4 suggest that both these protons are axially disposed, whereas that of 5-H [δ_H 4.10 (t, J = 2.8 Hz)] represents coupling to 6-H_eq [δ_H 2.49 (dd, J = 14.5, 2.8 Hz)] and 6-H_ax [δ_H 2.85 (dd, J = 14.5, 2.8 Hz)] with a J value typical of an equatorial proton.

---