An Expeditious Total Synthesis of both Diastereoisomeric Lipid Dihydroxytetrahydrofurans from Notheia anomala

Abstract

Short, high yielding syntheses of both diastereomers of the naturally occurring oxylipids 1 and 2 using a combination of organocatalytic hydroxylation of an aldehyde, alkene cross metathesis and palladium(0) catalyzed cyclization chemistry (6-step process) are reported. Furthermore, the influence of the catalyst on the cross metathesis reaction of the homoallylic 1,2-diol has been studied in detail.

As new and improved synthetic transformations are discovered, the synthesis of complex organic molecules has become more efficient. The application of new enantio- and chemo-selective reactions and new methods of carbon-carbon and carbon-heteroatom bond formation can greatly reduce the number of steps required for a given synthesis.

The nematocidal oxylipids 1 and 2 (Figure 1), isolated from the Australian brown algae Notheia anomala,¹ have been targets for synthesis for the past three decades. The biological activity and the challenging 2,5-disubstituted-3-oxygenated tetrahydrofuranyl (thf) motif makes them appealing candidates for total synthesis. Perhaps more importantly, the 2,5-disubstituted-3-oxygenated thf motif is an important structural feature found in many biologically active natural products.²

Figure 1.

Oxylipids Isolated from Southern Australian Algae Notheia anomala.

Due to the 30 year history, syntheses of the oxy lipids 1 and 2 have evolved with improvements in synthetic methodology. Williams et al. reported the first total synthesis of racemic 1 in 1984.³ The Williams synthesis required 11 steps with an overall yield of 17% from an already advanced intermediate. In addition to a racemic biomimetic synthesis of 1 and 2,⁴ there have been nine unique enantioselective syntheses of 1 with overall yields between 2–26% and five enantioselective syntheses of 2 with overall yields between 2–37%.^5,6,7,8 Only three of these syntheses are able to deliver both naturally occuring diastereoisomers.^7,8 This is not unusual since methods for the synthesis of 2,5-disubstituted tetrahydrofurans are often optimized for either the cis or trans isomer, but typically not both. Perhaps the most efficient enantioselective synthesis was reported in a 2009 paper by Britton et al., wherein the natural diastereoisomers 1 and 2 were prepared in 6 steps with 26% and 37% overall yields, respectively.⁸

We recently reported a stereospecific method for the formation of cyclic ethers employing a combination of alkene cross metathesis and Pd(0)-catalyzed cyclization.⁹ This method was applied to the synthesis of 2,5-trans thf-containing fragments of amphidinolides C and F.¹⁰ Since the cross metathesis and Pd(0)-catalyzed cyclization combination can deliver both the cis and trans cyclic ethers, it appeared to be ideal for the syntheses of the 2,5-trans and 2,5-cis oxygenated tetrahydrofuran rings in oxylipids 1 and 2 (Scheme 1). Actually, by careful choice of the coupling partners (e.g. 3 and 4) in the cross metathesis reaction, any of the 2,5-disubstituted-3-hydroxy thf isomers can be prepared. However, a short enantioselective synthesis of the syn-diol 3 was critical to the success of the proposed chemistry.

Scheme 1.

A Retro Synthetic Analysis for the OxyLipids

A rapid synthesis of the syn-diol 3 was envisaged which employed recent advances in organocatalysis.¹¹ D-proline catalyzed nitrosoaldol condensation of heptaldehyde 7 gave the α-aminoxy aldehyde 8, which was reacted directly with allylmagnesium chloride to furnish the syn-diol 3 in 75% isolated yield (Scheme 2) and 5–10% of undesired anti-diol. The diol diastereomers were easily separated by column chromatography.

Scheme 2.

Synthesis of Syn-diol 3

Alternatively, syn-diol 3 was formed via the D-proline catalyzed nitrosoaldol of 4-pentenal 9 with 2-nitrosotoluene, followed by direct addition of pentylmagnesium bromide to the aminoxy aldehyde 10. However, in this route the isolated yield of the syn-diol 3 was considerably lower (40%).

The cross metathesis reaction between syn-diol 3 and the (S)-carbonate 4 (>95% ee) using Grubbs second generation catalyst and CuI as a co-catalyst did not proceed as anticipated (Scheme 3).^9,10,12 Unfortunately, dienal 12 was the major isolated product and the desired cross metathesis product 11 was only a minor component.

Scheme 3.

Oxidative Cleavage of the Syn-diol 3 by Grubbs II-CuI Catalyst System

Although structurally similar diols undergo successful cross metathesis,¹³ there are reports of subsequent cleavge of the diol.^14,15,16 We use CuI as a co-catalyst to improve the reaction rate of cross metathesis with the phosphono allylic carbonate 4.^9,10 It is thought that CuI acts as a phosphine scavenger to produce a 14 electron ruthenium species, but this also destabilizes the active metal alkylidene species in the process leading to decomposition.^12c,17 The ruthenium species formed as a result of decomposition can, in the presence of an oxidant, cleave the diol.^14b It appeared that the solution would be the use of a more robust catalyst system.

Cross metathesis of syn-diol 3 and phosphono allylic carbonate 4 using Hoveyda-Grubbs II catalyst led to a slower reaction but with significantly improved product distribution. The oxidative cleavage was reduced to <15% (Scheme 3). Perhaps more surprisingly, the CM reaction of 3 and 4 using Grubbs II (and no CuI) resulted in further improvement in the yield of the product 11 to 74% and diminished the oxidative cleavage to <5%. Likewise, the CM reaction with Grubbs II catalyst between syn-diol 3 and (R)-carbonate 13 yielded 74% of the diastereomeric phosphono allylic carbonate 14 (Scheme 3).

Palladium(0)-catalyzed cyclization of 11 proceeded smoothly to give the 2,5-trans-tetrahydrofuranyl-(E)-vinyl phosphonate 5 in 93% isolated yield. Similarly, 2,5-cis-tetrahydrofuranyl-(E)-vinyl phosphonate 15 was formed in 89% isolated yield, along with 5% of the 2,5-trans-tetrahydrofuranyl-(Z)-vinyl phosphonate 16 (Scheme 4), which was separated by column chromatography.

Scheme 4.

Synthesis of the 2,5-Trans and 2,5-Cis Thf Vinyl Phosphonates

The tetrahydrofuranyl vinyl phosphonates were converted into the corresponding tetrahydrofuranyl alcohols for complete characterization (Scheme 5). The alcohols are also potentially useful intermediates. The vinyl phophonates 5 and 15 were subjected to ozonolysis followed by treatment of the ozonide with a large excess of DIBAL-H to furnish the diols 17 and 18 in 70% and 81% isolated yields, respectively. Reduction of the ozonide derived from vinyl phosphonate 5 with an excess of NaBH₄ was faster and higher yielding, but 5% of the C(9) epimer of 17 was also formed.

Scheme 5.

Ozonolysis of the Vinyl Phosphonates

Finally, suitable conditions for the transformation of the vinyl phosphonates to their respective furanyl aldehydes were investigated with the goal of achieving a short and efficient total synthesis of oxylipids 1 and 2. Initially, the vinyl phosphonates were subjected to ozonolysis followed by reduction of the ozonide with various reagents, including Me₂S and polystyrene immobilized PPh₃ (PS-PPh₃). Unfortunately, the results were not satisfactory. Reduction of the ozonide with PS-PPh₃ in CH₂Cl₂/MeOH gave mixture of stable acetals and hemicaetals, which failed to react with 8-nonenylmagnesium bromide. However, reduction of the ozonide with PS-PPh₃ in CH₂Cl₂, followed by treatment with 8-nonenylmagnesium bromide did give the desired product 1, albeit in a low 30% yield.

In an effort to improve the yield of the last step, we switched to oxidative cleavage of the vinyl phosphonate using OsO₄/NaIO₄. The typical one pot procedure using a mixture of OsO₄ and NaIO₄ in dioxane-water was extremely slow and failed to reach completion after 5 days. However, the two step process of dihydroxylation of the vinyl phosphonate 5 with OsO₄ and NMO, followed by glycol cleavage using NaIO₄ in CH₂Cl₂-H₂O furnished the aldehyde 19. The hydroxylated tetrahydrofuranyl aldehydes are highly unstable and therefore aldehyde 19 was immediately subjected to the addition of 8-nonenylmagnesium bromide without further purification (Scheme 6).⁸ The trans-thf containing oxylipid 1 was formed as the major product along with its column separable C(10)-epimer in a 2.5:1 ratio and 63% combined yield. Following the same reaction sequence, C9/C10 epimeric natural product 2 was prepared from vinyl phosphonate 15 as the major product along with its column separable C(10)-epimer in a 4:1 ratio in 68% combined yield. The spectral data of the natural products are in complete agreement with those reported in the literature.

Scheme 6.

Synthesis of Oxylipids 1 and 2

In summary, we report short, high yielding syntheses of the oxylipids 1 and 2. A combination of organocatalytic hydroxylation of an aldehyde, alkene cross metathesis and palladium(0) cyclization provides very efficient syntheses (6-steps) of the natural diastereomeric oxylipids 1 and 2 in 23% and 27% overall yield, respectively.