Synthetic studies of microtubule stabilizing agent peloruside A: an asymmetric synthesis of C₁₀–C₂₄ segment

Abstract

An asymmetric synthesis of the C₁₀–C₂₄ fragment of the potent antitumor macrolide, peloruside A is described. All three stereogenic centers have been enantioselectively constructed utilizing Evans alkylation, Brown asymmetric allylboration, and a substrate controlled epoxide formation. Other key reactions involved Grubbs’s ring-closing olefin metathesis and Ando’s Z-selective olefination reaction.

Peloruside A, a 16-membered macrolide, was recently isolated from the New Zealand marine sponge Mycale hentscheli.¹ It exhibited potent cytotoxicity against P388 murine leukemia cells with an IC₅₀ value of 10 ng/mL.¹ Like paclitaxel, peloruside A has shown microtubule-stabilizing activity and arrests cells in the G₂-M phase.² It has structural resemblance to epothilones which are undergoing clinical trials.³ Important antitumor activities of peloruside A along with its unique structural features have stimulated immense interest in the synthesis and structure–function studies. Peloruside A’s initial structure and relative stereochemistry were established by NMR studies.¹ However, its absolute stereochemistry was conclusively established only recently after the first total synthesis⁴ followed by its chemical and biological correlation with the natural peloruside A. To date, De Brabander and co-workers have reported the only total synthesis. Synthetic approaches toward fragments of peloruside A have been reported by Paterson et al.⁵ We recently reported an enantioselective synthesis of the C₁–C₉ segment of peloruside A.⁶ In continuation of our on-going studies, we now report a stereocontrolled route to the C₁₀–C₂₄ segment (3) of peloruside A where all three stereocenters have been created by asymmetric synthesis.

As shown in Figure 1, our convergent synthetic plan for peloruside A involves the assembly of C₁–C₉ segment (2) and C₁₀–C₂₄ (3) by an aldol reaction followed by a macrolactonization between the C₁₅-hydroxyl group and the C₁-carboxylic acid. We plan to synthesize the fragment 3 by a nucleophilic addition of isopropylmagnesium halide to the functionalized δ -lactone derived from acrylate ester 4. The corresponding homoallyl alcohol will be derived from protected alcohol 5 by oxidation, olefination and subsequent asymmetric allylboration of the resulting aldehyde utilizing Brown’s protocol. Protected alcohol 5 will be readily available by Evans’ asymmetric alkylation reaction.⁷

Figure 1.

As depicted in Scheme 1, asymmetric alkylation of chiral imide 6 was carried out with benzyloxymethyl chloride using Evans’ protocol^7a to provide the alkylated product 7 as a single isomer in 80% yield (from benzyloxazolidinone). Reduction of imide 7 by LiBH₄ in THF–MeOH at 23°C afforded alcohol 8. Swern oxidation of 8 and subsequent Horner–Emmons reaction of the resulting aldehyde with sodium enolate of (o-cresol)₂P=O(CH₃)CHCO₂Et as described by Ando⁸ furnished tri-substituted Z-olefin 9 in 90% yield over two steps. The Z-selectivity was >99:1 as revealed by ¹H and ¹³C NMR analysis. Our next synthetic plan was to install an α,β-unsaturated δ-lactone with appropriate stereochemistry and then elaborate the syn-1,3-diol functionality by a substrate-controlled diastereoselective epoxidation followed by reductive opening of the resulting epoxide. For the synthesis of α,β-unsaturated δ-lactone, ester 9 was first reduced by Dibal-H to deliver the corresponding alcohol. Dess-Martin periodinane oxidation provided the aldehyde which was subjected to the Brown asymmetric allylboration protocol⁹ with allyldiisopinocampheylborane to afford homoallylic alcohol 10 as the major diastereomer (dr=93:7 by ¹H and ¹³C NMR). Alcohol 10 was obtained in 65% yield over two steps after silica gel chromatography.

Scheme 1. Reagents and conditions:

(a) TiCl₄, Et₃N, CH₂Cl₂, PhCH₂OCH₂Cl, 0°C, 1.5 h; (b) LiBH₄, MeOH, THF, 23°C, 1 h (94%); (c) (COCl)₂, DMSO, Et₃N, CH₂Cl₂, −60°C, 45 min; (d) (o-cresol)₂^PO(CH₃)CHCO₂Et, NaH, THF, −78 to −20°C, 2 h (90% over two steps); (e) Dibal-H, CH₂Cl₂, −78° to −40°C, 1 h (96%); (f) Dess–Martin periodinane, NaHCO₃, CH₂Cl₂, 23°C, 1.5 h; (g) CH₂CHCH₂B[(+)-Ipc]₂, Et₂O, −80°C, 3 h (65% over two steps); (h) CH₂CHCOCl, Et₃N, CH₂Cl₂, 0°C, 2 h (77%); (i) Cl₂(Pcy)₂RuCHPh, CH₂Cl₂, 40°C, 12 h (83%); (i) H₂O₂, 6N-aq NaOH, MeOH, 1.5 h (74%); (k) NaBH₄, PhSeSePh, AcOH, ⁱPrOH, 0°C, 30 min (quant.); (l) TBDPSCl, imidazole, DMAP, DMF, 23°C, 13 h (quant.).

For synthesis of α,β-unsaturated δ-lactone, we relied upon ring-closing olefin metathesis protocol described in our recent work.¹⁰ Reaction of 10 with acryloyl chloride and triethylamine at 0°C furnished acryloyl ester 11. Exposure of 11 to a catalytic amount (10 mol%) of first generation commercial Grubb’s catalyst¹¹ in CH2Cl2 at reflux for 12 h provided α,β-unsaturated δ-lactone 12 in 83% yield after one recycle of the recovered starting material.¹² The use of titanium tetra-isopropoxide as a co-catalyst (40 mol%) significantly lowered the yield (62%) of δ-lactone 12. Reaction with second generation Grubb’s catalyst did not improve the overall yield (74%) either. For elaboration of the hydroxy γ-lactone with appropriate stereochemistry, δ-lactone 12 was exposed to nucleophilic epoxidation with alkaline hydrogen peroxide in methanol at 0°C to furnish the corresponding epoxide in 74% yield as a single isomer. Treatment of the resulting epoxide with diphenyldiselenide and sodium borohydride in 2-propanol afforded exclusively hydroxylactone 13 in quantitative yield.¹³ Lactone 13 possesses the required syn-1,3-diol stereochemistry necessary for peloruside A synthesis. Protection of alcohol 13 under standard condition provided TBDPS ether 14 in quantitative yield.

For conversion of -lactone 14 to the C₁₀–C₂₄ fragment, we first attempted direct opening of lactone ring with isopropylmagnesium chloride. However, reaction of 14 with excess isopropylmagnesium chloride under a variety of reaction conditions did not provide the desired ketone. In an alternative approach, δ-lactone 14 was transformed into C₁₀–C₂₄ fragment ketone 16 as follows (Scheme 2). Reaction of 14 with N,O-dimethylhydroxylamine hydrochloride in the presence of trimethylaluminum in CH₂Cl₂ according to Weinreb procedure¹⁴ furnished the corresponding hydroxy Weinreb amide. Protection of the resulting alcohol as MEM–ether provided Weinreb amide 15 in 86% yield (from 14). Treatment of 15 with isopropylmagnesium chloride (5 equiv.) in THF at 23°C for 5 h afforded C₁₀–C₂₄ fragment ketone 16¹⁵ in 61% yield. It turned out that both MEM-and TBDPS-protecting groups were critical to form 16. Replacement of the MEM-group with a PMB-group resulted in substantial β-elimination as well as degradation of starting material.

Scheme 2. Reagent and conditions:

(a) AlMe₃, HN(OCH₃)-CH₃·HCl, CH₂Cl₂, 23°C, 2.5 h (93%); (b) MEMCl, DIPEA, 23°C, 9 h (92%); (c) ⁱPrMgCl, THF, 23°C, 5 h (61%).

In summary, a highly stereoselective synthesis of C₁₀–C₂₄ segment of peloruside A has been accomplished. The key steps involved an Evans asymmetric alkylation, Ando’s Z-selective olefination, Brown asymmetric allylboration, Grubbs’s ring-closing olefin metathesis, and substrate controlled stereoselective epoxidation. Work toward the total synthesis of peloruside A is in progress.