Stereoselective synthesis of the C1–C8 and C9–C16 fragments of revised structure of (−)-lyngbouilloside

Abstract

The synthesis of C1–C8 and C9–C16 fragments of revised structure of (−)-lyngbouilloside is described starting from geraniol and D-malic acid.

---

In 2002, Gerwick and co-workers¹ reported the isolation of a novel glycosidic macrolide, lyngbouilloside (1a) from the marine cyanobacterium Lyngbya bouillonii collected from Papua New Guinea and later from the North coast of New Britain. On the basis of spectroscopic analysis, the structure of lyngbouilloside was reported as a 14-membered macrolactone embedded with trisubstituted tetrahydropyran ring hemiketal, linked glycosidically through C5 to a 2,4-di-O-methylrhamnoside moiety, having unusual tertiary methyl carbinol at C13, and an E,E-conjugated diene system in the side chain (1a, Fig. 1).

Fig. 1 The original lyngbouilloside structure (1a), the correct structure of (−)-lyngbouilloside (1b).

Although coupling constant analysis, chemical derivatization, and NOE experiments permitted determination of the relative stereochemistry, the lack of available material prevented assignment of the absolute stereochemistry. Lyngbouilloside was found to possess moderate cytotoxic activity toward neuroblastoma and KB cells with IC₅₀ value of 17 M.

Recently, the first enantioselective synthesis of the originally assigned structure of lyngbouilloside aglycon was reported by Cossy et al.² that culminates in a stereochemical reassignment of the C11 stereogenic center and establishes the absolute stereochemistry of the natural product and revised the structure as 1b. To date, only two fragment syntheses³ of the original structure of lyngbouilloside (1a) have been reported. However, there is no report existing on the fragment synthesis of the revised structure of lyngbouilloside (1b). We were attracted to lyngbouilloside as a synthetic target because of its structural complexity and bioactivity. Herein we report the stereoselective synthesis of the C1–C8 and C9–C16 fragments of (−)-lyngbouilloside (1b).

Our retrosynthetic strategy (Scheme 1) began with disconnection of the C8–C9 bond and the lactone carbonyl leading to the fragments 2 and 3 respectively. Fragment 2 could be realized from geraniol whereas the fragment 3 could be generated from D-malic acid.

Scheme 1 Retrosynthetic analysis.

Synthetic strategy for C1–C8 fragment (3)

Synthesis of the C1–C8 fragment (3) of lyngbouilloside began from the alcohol 4 prepared from D-malic acid following the procedure reported⁴ for its enantiomer (Scheme 2). Primary alcohol 4 was oxidized and the resulting aldehyde was subjected to Keck allylation⁵ using allyltributyl tin in the presence of (S)-BINOL and Ti(OⁱPr)₄ to afford 5 in 90% yield. The optical purity of 5 was found to be 97% de by chiral HPLC. The absolute configuration of the newly formed stereogenic centre at C5 was assigned by transforming compound 5 into the corresponding (R)- and (S)- Mosher esters,⁶ which were achieved using (R)- and (S)- MTPA acids with DCC as coupling reagent respectively. From the ¹H NMR of these two esters, positive Δδ_S–R values were observed for the protons H-7 and H-8, while negative Δδ_S–R values were observed for the protons H-2 and H-3. These data established the absolute configuration at C5 as S in compound 5. The secondary hydroxyl group was protected as its benzyl ether to furnish 6. Oxidative cleavage of 6 with NaIO₄ afforded the aldehyde, which without isolation was converted into β-keto ester 7 by reaction with ethyl diazoacetate in the presence of a catalytic amount of tin(II) chloride.⁷

To access the functionalized tetrahydropran ring, 7 was treated with PTSA in MeOH at 0 °C to provide 8 as the sole product in 85% yield by removal of the cyclohexylidene group and concomitant cyclization in one-pot. The structure of compound 8 was established by ¹H NMR (500 MHz, CDCl₃) data and assignments were made with the aid of TOCSY and NOESY experiments (Fig. 2). The characteristic NOE between C7H/C3-OMe suggested that both the proton and OMe are on the same face of the structure. This was further supported by NOE correlation between C5H/C7H, confirming the structure of the compound. The energy minimized structure as shown in Fig. 3 is also in agreement with the assigned structure from NMR data.

Fig. 2 Expansion of the NOESY spectrum showing the characteristic NOE correlations between C7H/C3-OMe and C5H/C7H of compound 8.

Fig. 3 Chemical structure and energy-minimized structure of 8.

IBX oxidation of the derived alcohol in 8 followed by one-carbon Wittig olefination delivered the desired C1–C8 fragment 3 in 60% yield over two steps.

Synthetic strategy for C9–C16 fragment (2)

Preparation of the C9–C16 fragment (2) of lyngbouilloside began from commercially available geraniol. Starting with geraniol, 9 is synthesized as described in the literature.⁸ Oxidative cleavage of the diol from 9 with NaIO₄ followed by reduction with NaBH₄ furnished alcohol 10 (89% yield over three steps), which on benzyl protection (BnBr/NaH/THF/0 °C/4 h) afforded 11 (92%). The subsequent regioselective PMB (p-methoxybenzyl) acetal cleavage with DIBAL-H at 0 °C in CH₂Cl₂ afforded 12 in 90% yield. Further, the alcohol, on oxidation under Swern conditions, furnished an aldehyde that after aldol reaction with the chlorotitanium enolate of N-propionyl thiazolidine-2-thione 13 at −40 °C gave non-Evans syn-aldol⁹ adduct 14 as a diastereomeric mixture (dr 95 : 5, determined by ¹H NMR). The required major isomer 14 was easily separated by silica gel column chromatography (Scheme 3).

Scheme 2 Reagents: (a) (i) (COCl)₂, DMSO, Et₃N, CH₂Cl₂, −78 °C, 3 h; (ii) S-BINOL, Ti(OⁱPr)₄, allyltributyl tin, CH₂Cl₂, −78 °C, 48 h, 90%; (b) NaH, BnBr, TBAI, THF, 0 °C-rt, 4 h, 93%; (c) (i) OsO₄ (cat), NMO, acetone–H₂O (1 : 1), rt, 24 h; (ii) NaIO₄, THF–H₂O (2 : 1), 0 °C, 30 min; (iii) SnCl₂, N₂CHCOOEt, CH₂Cl₂, 0 °C-rt, 3 h, 80% over 3 steps; (d) PTSA, MeOH, 0 °C, 2 h, 85%; (e) (i) IBX, EtOAc, reflux, 3 h. (ii) Ph₃P⁺CH₃I⁻, n-BuLi, THF, −40 °C-rt, 5 h, 60% over two steps.

Scheme 3 Reagents: (a) (i) OsO₄ (cat), NMO, acetone–H₂O (1 : 1), rt, 48 h; (ii) NaIO₄, THF–H₂O (2 : 1), 0 °C, 30 min; (iii) NaBH₄, MeOH, 0 °C, 30 min, 89% over 3 steps; (b) NaH, BnBr, THF, 0 °C-rt, 4 h, 92%; (c) DIBAL-H, CH₂Cl₂, 0 °C, 1 h, 90%; (d) (i) (COCl)₂, DMSO, Et₃N, CH₂Cl₂, −78 °C, 3 h; (ii) thione 13, TiCl₄, DIPEA, CH₂Cl₂, −40 °C, 4 h, 82%; (e) TBSOTf, 2,6-lutidine, CH₂Cl₂, 0 °C, 30 min, 91%; (f) DIBAL-H, CH₂Cl₂, −78 °C, 10 min; (g) Ph₃P⁺CH₃I⁻, n-BuLi, THF, −40–0 °C, 5 h, 85% over 2 steps; (h) DDQ, CH₂Cl₂–H₂O (9 : 1), 0 °C, 30 min, 89%.

The stereochemistry of 14 was confirmed by ¹³C NMR spectroscopic analysis.¹⁰ The resulting secondary alcohol was protected as tert-butyldimethylsilylether¹¹ to provide 15 (91%). Reductive cleavage of the chiral auxiliary with DIBAL-H afforded aldehyde 16 followed by one-carbon Wittig homologation with triphenylphosphonium methylide produced terminal olefin 17. Finally, DDQ-mediated oxidative cleavage of the PMB (para-methoxybenzyl) ether produced the targeted C9–C16 subunit 2 in 89% yield.

Thus, we have synthesized the C1–C8 and C9–C16 segments 2 and 3 of lyngbouilloside (1b), marine macrolide, with reassigned stereochemical and absolute configuration at C11, and further investigations toward the total synthesis of lyngbouilloside (1b) are in progress.

Acknowledgements

TRR thank UGC, New Delhi for the award of a fellowship.

References

1. L. T. Tan, B. L. Marquez and W. H. Gerwick, J. Nat. Prod., 2002, 65, 925–928.
2. A. El Marrouni, R. Lebeuf, J. Gebauer, M. Heras, S. Arseniyadis and J. Cossy, Org. Lett., 2012, 14(1), 314–317.
3. (a) A. El Marrouni, A. Kolleth, R. Lebeuf, J. Gebaur, S. Prevost, M. Heras, S. Arseniyadis and J. Cossy, Nat. Prod. Commun., 2013, 8, 965–972; (b) J. Gebauer, S. Arseniyadis and J. Cossy, Synlett, 2008, 712–714; (c) D. Webb, H. A. van den, M. Kogl and S. V. Ley, Synlett, 2009, 2320–2324.
4. S. Hanessian, A. Ugolini, D. Dube and A. Glamyan, Can. J. Chem., 1984, 62, 2146.
5. G. E. Keck, K. H. Tarbet and L. S. Geraci, J. Am. Chem. Soc., 1993, 115, 8467.
6. (a) I. Ohtani, J. Kusumi, Y. Kashman and H. Kakisawa, J. Am. Chem. Soc., 1991, 113, 4092; (b) W. Y. Yoshida, P. J. Bryan, B. J. Baker and J. B. McClintock, J. Org. Chem., 1995, 60, 780.
7. C. R. Holmquist and E. J. Roskamp, J. Org. Chem., 1989, 54, 3258–3260.
8. S. Takano, M. Akiyama, S. Sato and K. Ogasawara, Chem. Lett., 1983, 1593.
9. M. T. Crimmins, B. W. King and E. A. Tabet, J. Am. Chem. Soc., 1997, 119, 7883–7884.
10. The stereochemistry of 14 was confirmed on the basis of ¹³C NMR spectra by converting 2 (derived from 14, Scheme 3) into compound 18 through the following sequence of reaction. In compound 2, TBS group was deprotected followed by acetonide protection of 1,3-diol to give the cyclic moiety 18. The appearance of distinctly different methyl resonance peaks at δ 19.1 and 30.3 ppm and the acetal carbon resonating at δ 98.5 ppm (see the ESI†) confirmed the presence of syn-acetonide 18 by using Rychnovsky's method.¹²
    .
11. M. T. Crimmins and E. A. Tabet, J. Org. Chem., 2001, 66, 4012–4018.
12. (a) S. D. Rychnovsky and D. J. Skalitzky, Tetrahedron Lett., 1990, 31, 945–948; (b) D. A. Evans, D. L. Rieger and J. R. Gage, Tetrahedron Lett., 1990, 39, 7099–7100.

---

Footnote

† Electronic supplementary information (ESI) available: Spectral data of selected compounds and copies of ¹H and ¹³C NMR spectra of all compounds. See DOI: 10.1039/c3ra44354j

---