Total Synthesis of Bryostatin 7 via C-C Bond Forming Hydrogenation: Merged Redox-Construction Events for Synthetic Efficiency

Abstract

The marine macrolide bryostatin 7 is prepared in 20 steps (longest linear sequence) and 36 total steps. A total of 5 C-C bonds are formed using hydrogenative methods. The present approach represents the most concise synthesis of any bryostatin reported, to date, setting the stage for practical syntheses of simplified functional analogues.

The bryostatins are a family of 20 marine natural products originally isolated from the bryozoan Bugula neritina¹ that possess a polyacetate backbone and differ largely on the basis of substitution at C₇ and C₂₀ (Figure 1).² The bryostatins display diverse biological effects, including antineoplastic activity, immunopotentiating activity, restoration of apoptotic function, and the ability to act synergistically with other chemotherapeutic agents.³ Neurological effects also are evident, including activity against Alzheimer’s disease,⁴ neural growth and repair and the reversal of stroke damage,⁵ as well as memory enhancement.⁶

Figure 1.

Bryostatins 1–17 and prior total syntheses.^a

^aSee supporting information for a graphical summary of prior syntheses.

As their natural abundance is insufficient to advance clinical and biochemical studies, the bryostatins have emerged as a vibrant testing ground for polyketide construction. To date, total syntheses of bryostatin 1 (Keck, 2011),^7a bryostatin 2 (Evans, 1998),^7b,c bryostatin 3 (Yamamura, 2000),^7d,e bryostatin 7 (Masamune, 1990),^7f bryostatin 9 (Wender, 2011)^7g and bryostatin 16 (Trost, 2008)^7h have been reported. A formal synthesis of bryostatin 7 (Hale, 2006)^8a and total syntheses of C20-epi-bryostatin 7 (Trost, 2010)^8b and C20-deoxybryostatin (Thomas, 2011)^8c have been disclosed. In efforts led by Wender^9,10 and Keck,¹¹ simplified bryostatin analogues that retain high potency have been identified.

Given the longstanding challenges associated with defining concise routes to the bryostatins, these natural products were deemed an ideal vehicle to benchmark the utility of the C-C bond forming hydrogenations developed in our laboratory.¹² Retrosynthetically, a convergent assembly of the bryostatin 7 core from Fragments A and B employing the Keck-Yu pyran annulation^13,14 and Yamaguchi macrolactonization¹⁵ was envisioned. For the synthesis of Fragment A, hydrogen-mediated reductive coupling of conjugated glyoxal 6 and enyne 9 appeared strategic, as the key C20-C21 bond would be formed with control of the C20 carbinol stereochemistry and C21 olefin geometry.¹⁶ The planned synthesis of Fragment B, which incorporates the bryostatin A-ring, takes advantage of three transfer hydrogenative processes: enantioselective double allylation of 1,3-propanediol to furnish the C₂-symmetric diol 11,17a subsequent aldehyde tert-prenylation^17b to establish the C7 carbinol stereochemistry and install the C8 gem-dimethyl moiety and, finally, allylation^17c,d at C9 to introduce the C11 aldehyde. The feasibility of these fragment syntheses has been established in model systems (Scheme 1).^16,17e

Scheme 1.

Retrosynthetic analysis of bryostatin 7 illustrating C-C bonds formed via hydrogenative coupling.

The synthesis of Fragment A begins with the hydroxymethylation of 3-methyl-2-butanone 1 to furnish the aldol product in accordance with the literature procedure.¹⁸ Moffatt-Swern oxidation of the aldol product provides ketoaldehyde 2, which upon Horner-Wadworth-Emmons olefination delivers the α,β-unsaturated ester 3. All compounds up to this point are isolated by vacuum distillation, expediting access to large quantities of material. Conversion of 3 to the enol silane followed by addition of LiAlH₄ to the reaction mixture directly provides the allylic alcohol 4.¹⁹ Treatment of crude allylic alcohol 4 with tert-butyldimethylsilyl chloride followed N-bromosuccinimide provides the α-bromoketone 5 in 84% yield over the two-step sequence from α,β-unsaturated ester 3. Finally, Kornblum oxidation of α-bromoketone 5 delivers the glyoxal 6 (Scheme 2).²⁰

Scheme 2.

Synthesis of Fragment A via hydrogen-mediated reductive coupling of glyoxal 6 and 1,3-enyne 9.^a

^aIndicated yields are of material isolated by silica gel chromatography or distillation. See supporting information for experimental details.

Preparation of the requisite 1,3-enyne 9 takes advantage of the Sharpless asymmetric dihydroxylation of crotononitrile 7, which provides the diol in 86% enantiomeric excess.²¹ The diol is converted to the acetonide and exposed to diisobutylaluminum hydride to provide the aldehyde 8, which is a known compound previously prepared using a 6-step sequence.²² Chelation controlled propargylzinc addition converts aldehyde 8 to the homopropargylic alcohol, which is formed as a 5:1 mixture of diastereomers.²² As described in the supporting information, the minor isomer is easily converted to the desired epimer using a Mitsunobu inversion protocol. The homopropargylic alcohol is converted to the TBDPS ether and subjected to Sonogashira coupling to deliver the 1,3-enyne 9 (Scheme 2).

To complete the synthesis of Fragment A, the glyoxal 6 and 1,3-enyne 9 are subjected to hydrogen-mediated reductive coupling to furnish the α-hydroxyketone in 77% yield as a 7:1 mixture of diastereomers.¹⁶ Notably, although the coupling product incorporates multiple points of unsaturation, over-reduction is not observed under the conditions of hydrogenative coupling. Exposure of α-hydroxyketone to acetic anhydride provides the acetate. Selective deprotection of the allylic TBS-ether in the presence of the TBDPS ether, which is accomplished using HF-pyridine, provides the allylic alcohol. Finally, oxidation of allylic alcohol delivers the enal, Fragment A, in a total of 10 steps from 3-methyl-2-butanone 1 or crotononitrile 7 (Scheme 2).

Efforts toward Fragment B begin with allyl acetate mediated double allylation of 1,3-propanediol 10^17a,e to form C₂-symmetric diol 11. This process employs an iridium catalyst generated in situ from [Ir(cod)Cl]₂, allyl acetate, 4-chloro-3-nitrobenzoic acid and (S)-Cl,MeO-BIPHEP. Because the minor enantiomer of the mono-allylated intermediate is converted to the meso-diastereomer,²³ diol 11 is obtained as a single enantiomer, as determined by chiral stationary phase HPLC analysis. Previously, the mono-TBS ether of diol 11 was prepared in 7 steps from 1,3-propanediol through iterative use of Brown’s reagent for carbonyl allylation.^24a Alternatively, a four step protocol for the preparation diol 11 from acetylacetone is described.^24b Ozonolysis of diol 11 delivers an unstable lactol, which is protected in situ as the bis-TBS ether to provide aldehyde 12 as a single isomer. Transfer hydrogenation of aldehyde 12 in the presence of 1,1-dimethylallene promotes tert-prenylation^17b to form neopentyl alcohol 13. In this process, isopropanol serves as the hydrogen donor and the discrete iridium complex prepared from [Ir(cod)Cl]₂, allyl acetate, m-nitrobenzoic acid and (S)-SEGPHOS²⁵ is used as catalyst. Notably, complete levels of catalyst directed diastereoselectivity are observed.²⁶ Exposure of neopentyl alcohol 13 to acetic anhydride followed by ozonolysis provides β-acetoxy aldehyde 14. Reductive coupling of aldehyde 14 and allyl acetate under transfer hydrogenation conditions results in the formation of homoallylic alcohol 15. As the stereochemistry of this addition is irrelevant, an achiral iridium complex derived from [Ir(cod)Cl]₂, allyl acetate, m-nitrobenzoic acid and BIPHEP is employed as catalyst. Selective removal of the glycosidic silyl ether followed by concomitant Dess-Martin oxidation of the lactol and homoallylic alcohols provides β,γ-enone 16. Remarkably, treatment of a methanolic solution of 16 to pyridinium p-toluenesulfonate triggers sequential lactone ring opening followed by formation of the cyclic ketal 17a. Ozonolysis of 17a provides Fragment B in a total of 10 steps from 1,3-propanediol 10 (Scheme 3).

Scheme 3.

Synthesis of Fragment B employing multiple transfer hydrogenative C-C bond formations.^a

^aIndicated yields are of material isolated by silica gel chromatography. See supporting information for experimental details.

The union of Fragment A and Fragment B is achieved through Keck-Yu annulation to form the B-ring pyran.^13,14 The desired adduct 18a is accompanied by the elimination product 18b, however, both compounds participate in acidic methanolysis to form triol 19b. Chemoselective hydrolysis of the C1 methyl ester in the presence of the C7 and C20 acetates employing trimethyltin hydroxide²⁷ followed by selective TES-protection of the triol reveals a hydroxy acid, which upon Yamaguchi macrolactonization¹⁵ provides tetraene 20. Concomitant Johnson-Lemieux oxidation²⁸ of the olefinic termini of tetraene 20 in the presence of the neopentyl olefin at C16-C17 installs both the Bring ketone and C-ring enal. Whereas Corey-Gilman oxidation of enal failed,²⁹ the corresponding N-heterocyclic carbene promoted process provided the desired methyl ester 21 in good isolated yield.³⁰ Finally, as practiced in prior syntheses,^7a,b,c,d,g asymmetric olefination of the B-ring ketone using Fuji’s chiral phosphonate³¹ followed by global deprotection using HF-pyridine provides bryostatin 7 (Scheme 4).

Scheme 4.

Union of Fragment A and Fragment B and total synthesis of bryostatin 7.^a

^aIndicated yields are of material isolated by silica gel chromatography. See supporting information for experimental details.

The present synthesis of bryostatin 7 is accomplished in 20 linear and 36 total steps,³² representing the most concise sequence to any bryostatin reported, to date. The concise nature of this approach can be attributed to the rapid assembly of key fragments A and B, as availed through application of C-C bond forming hydrogenations developed in our laboratory¹² – a technology that has enabled dramatic simplification in the synthesis of other polyketide natural products.³³ This work serves as a prelude to even shorter routes to the bryostatins and simplified functional analogues. More broadly, the merged redox-construction events central to this study speak to an emerging retrosynthetic paradigm, wherein C-C bond construction is accompanied by withdrawal of hydrogen.^34,35