Synthesis of the ABC Ring System of Azaspiracid. 2. A Systematic Study into the Effect of C₁₆ and C₁₇ Substitution on Bis-spirocyclization

Abstract

A systematic study into the effect of C₁₆ and C₁₇ substitution on the stereochemical outcome of bis-spirocyclization to form the ABC ring system of azaspiracid is disclosed. Successful construction of the natural 10R,13R bis-spirocyclic stereochemistry has been accomplished on the C₁₆ benzyloxy-containing precursor.

The azaspiracids are an intriguing class of recently isolated natural products that possess a complex structural framework as well as considerable biological activity.^1-3 As discussed in the previous paper,⁴ the D ring appears to exert considerable influence on the bis-spirocyclization. Based on these results, our efforts shifted toward the construction of selected substrates containing substitution at C₁₆ or C₁₇ (Scheme 1). The C₁₇ series appeared more attractive as inspection of the potential chair conformations of bis-spiroketals 2 and 3⁴ revealed that the transoidal and the cisoidal structures were both capable of placing the C₁₇ allyl substituent equatorial on the basis of the proposed conformation for bis-spiroketals 5 and 6.

Scheme 1.

Potential Combinations for Substitution Patterns at C₁₆ and C₁₇

C₁₇ Substitution

Spirocyclization of previously described keto sulfone 10⁴ using our preferred conditions for keto sulfone substrates (CSA, MeCN)⁵ led to a mixture of stereoisomers (Scheme 2). Treatment of the spirocycle 11 with n-BuLi induced β-elimination to yield the elaborate enol ether 12 in 70% yield, along with 10% of the presumed C₁₀ epimer. The strategy allowed for the protection of the C₁₃ carbonyl function while selectively revealing the C₁₇ hydroxyl group. Oxidation at C₁₇ followed by Brown allylation⁶ yielded the homoallylic alcohol 13 in greater than 20:1 d.s. Removal of the sulfone functionality revealed the highly labile enol ether 14, which rapidly underwent spirocyclization under the standard conditions (0.04 M CSA, t-BuOH/PhMe, 14–18 h) to give the bis-spiroketal 6 as a single diastereomer. Unfortunately, NOESY and COSY NMR experiments (CDCl₃) confirmed the cisoidal 10R,13S relationship of the bis-spiroketal. While it is important to point out that the enol ether spirocyclization precursor 14 is structurally different than ketone 4, previous work in our laboratory has shown that the C₁₀ ketal is readily ionized under acidic conditions,⁵ thereby ensuring formation of comparable spirocyclization intermediates from both precursors. It is interesting to note, however, that the sulfone moiety once again provided added stability,⁷ as 13 was indefinitely stable in the freezer. Also, our ketalization/β-elimination/desulfonylation strategy represents, to the best of our knowledge, a unique strategy for the effective construction of highly labile enol ethers (e.g., 14).

Scheme 2.

Bis-spirocyclization of C₁₇ Allyl Substrate^a

^a Key: (i) CSA, MeCN, 90%; (ii) n-BuLi, THF, −78 °C, 70%; (iii) TPAP, NMO, CH₂Cl₂, mol. sieves, 96%; (iv) (+)-Ipc₂Ballyl, Et₂O, pentane, 70%, >20:1 d.s.; (v) 5% Na/Hg, Na₂HPO₄, MeOH, THF, −10 °C; (vi) CSA, t-BuOH, PhMe, 76% (over two steps).

C₁₆ Substitution

The synthesis of the required benzyloxy aldehyde 21 began with the known Evans alkylation product 15⁸ (Scheme 3). Conversion to its benzyl ester 16 followed by Sharpless asymmetric dihydroxylation,⁹ in situ lactonization, and TIPS protection yielded the lactone 17, in overall 66% yield from 16, as a separable 3:1 ratio at C₁₆ favoring the desired stereochemistry.¹⁰ While the diastereomeric selectivity in the dihydroxylation is less than optimum (3:1 d.s.), direct dihydroxylation of the chiral oxazolidinone-containing alkene 15 provided inferior results (1.5:1 d.s.). This observation is consistent with our previous findings with other oxazolidinone-containing alkenes.¹¹ Subsequent reduction using LiBH₄ provided the diol 18. Next, sequential protection at C₁₃ and C₁₆ produced the fragment 19, along with a small amount of impurities derived from migration of the pivaloate and silyl protecting groups. Removal of the TIPS ether under standard conditions allowed for easy purification. The C₁₇ hydroxyl was reprotected as its TES ether 20. Cleavage of the C₁₃ pivaloate protecting group did prove problematic with DIBAL-H;¹² however, the use of LiBH₄, in the presence of a small amount of saturated aqueous NH₄Cl, cleanly provided the desired alcohol in 99% yield. Finally, TPAP oxidation yielded the target aldehyde 21.

Scheme 3.

Synthesis of the C₁₆ Benzyloxy-Substituted Aldehyde^a

^a Key: (i) BnOLi, PMBOH, THF, 99%; (ii) AD mix α, NaHCO₃, t-BuOH, H₂O; (iii) TIPSOTf, 2,6-lutidine, CH₂Cl₂, −78 °C, 66% overall yield from 16 (3:1 d.s.); (iv) LiBH₄, MeOH, THF, 0 °C, 99%; (v) PivCl, Et₃N, DMAP, CH₂Cl₂, 66%; (vi) NaH, BnBr, DMF, −50 °C to −10 °C; (vii) TBAF, THF, 79% (over two steps); (viii) TESCl, DMAP, Et₃N, CH₂Cl₂, 95%; (ix) LiBH₄, saturated aqueous NH₄Cl, THF, 0 °C to rt, 99%; (x) TPAP, NMO, CH₂Cl₂, molecular sieves, 87%.

Lithiation of sulfone A with LDA followed by addition of the aldehyde 21 provided the hydroxy sulfone adduct as a labile mixture¹³ of all four diastereomers (Scheme 4). Immediate oxidation to the ketone sulfone 22 was accomplished with TPAP in a 60% yield over two steps. Desulfurization under standard conditions followed by bis-spirocyclization (0.04 M CSA, t-BuOH/PhMe, 14–18 h) gave two spiroketals in a 3:5 ratio (8/9) in an overall 80% yield. After careful separation of the two isomers, the stereochemistry of the more polar isomer was determined to be the cisoidal ketal 9 (2D NMR, CDCl₃). We were gratified to find, however, that the less polar spirocycle 8 possessed the natural transoidal stereochemistry at C₁₀ and C₁₃ as established by ROESY and COSY correlations (C₆D₅N). While the NOE between the C₉ alkene to C₁₇ hydrogen was previously observed in the transoidal product 2 from the D ring truncated series,⁴ the NOE between the C₆ hydrogen and the methyl at C₁₄ was not present. This lack of signal can be explained by the placement of the C₁₃ furan oxygen and the C₁₄ methyl substituents in axial orientations within the pyran C ring. Further NMR evidence supports this hypothesis: (1) a strong NOE is observed between the C₁₄ methyl and the C₁₆ hydrogen and (2) the coupling constants between the C₁₆ and C₁₇ hydrogens match the predicted data for the proposed C ring, chair conformation using the Karplus correlation.¹⁴ Furthermore, the alternate chair conformer (with the C₁₃ furan oxygen and the C₁₄ methyl substituent in equatorial orientations) is significantly higher in energy (3.4 kcal/mol) using the B3LYP density functional and a 6-31G(d) basis set. It should also be noted that the observed NOE and coupling constant data would not be expected for the alternate non-natural transoidal bis-spirocycle.⁵ Finally, the proposed conformation of transoidal spirocycle 8 is in contrast to the previously discussed transoidal spirocycle 2, which appeared to place the C₁₃ furan oxygen and the C₁₄ methyl substituents in equatorial orientations.

Scheme 4.

Bis-spirocyclization of C₁₆ Benzyloxy Substrate^a

^a Key: (i) LDA, THF, −78 °C then 21; (ii) TPAP, NMO, CH₂Cl₂, molecular sieves 60% (over two steps); (iii) 5% Na/Hg, Na₂HPO₄, MeOH, THF, −10 °C; (iv) CSA, PhMe/t-BuOH, 80%, 3:5 ratio (8/9).

To further study the nature of the bis-spiroketalization, the reaction with ketone 7 was performed at lower temperatures (−10 to +4 °C, 21 h) and reduced molarity of the acid catalyst (0.003 M) under otherwise identical reaction conditions (1:1 t-BuOH/PhMe). We were intrigued to discover that the predominate product was the cisoidal bis-spirocycle 9. Gratifyingly, further warming of the reaction to room temperature for an additional 48 h resulted in the previously observed (3:5 ratio of 8:9) equilibrium mixture.¹⁵ It would appear from these observations that the cisoidal bis-spirocycle 9 is the result of kinetic control while the transoidal bis-spirocycle 8 can be accessed under thermodynamic conditions. With one equilibration cycle, a 50% overall yield of the desired transoidal bis-spirocycle 8 can be obtained. Finally, use of Nicolaou and co-workers’ conditions¹⁶ for equilibration of their cisoidal bis-spirocycle to the natural transoidal species (3 equiv of TFA, CH₂Cl₂) provided inferior results for the conversion of 9 to 8 (approximately 1:3 ratio for 8:9).¹⁷

The first systematic study into the effect of substituents on the bis-spirocyclization of a series of precursors has been presented. The C₁₆ oxygen substitution facilitated formation of a nearly equal mixture of the cisoidal and transoidal bis-spirocycles while C₁₇ allyl substitution provided sole access to the cisoidal species. Our continuing progress toward the total synthesis will be reported in due course.