Bioinspired Asymmetric Total Synthesis of Emeriones A-C.

We report asymmetric bioinspired total syntheses of the fungal metabolites emeriones A-C via stereoselective oxidations of two bicyclo[4.2.0]octadiene diastereomers. The central bicyclic scaffolds are prepared in an 8π/6π electrocyclization cascade of a stereodefined pentaene, which contains the fully assembled side chains of the emeriones. The anti-aldol side chain is made using a Paterson-aldol addition, and the epoxide of the dioxabicyclo[3.1.0]hexane side chain via ring-closure onto an oxidized acetal. Our work has enabled the structural revision of emerione C, and resulted in the synthesis of a "missing" family member, which we call emerione D. DFT calculations identified two methyl groups that govern torquoselectivity in the 8π/6π cascade.

Natural products derived from polyenes that undergo cyclization/isomerization cascades initiated by an 8π electrocyclization have intrigued chemists for decades. The emeriones (Figure 1), one such family of natural products that were isolated from the fungus E. nidulans, display oxidized bicyclo[4.2.0]octadiene cores (red) flanked by a seven carbon aldol fragment (blue) and a propenyl‐substituted dioxabicyclo[3.1.0]hexane system (black). The two side chains (blue and black) of emerione A (1) and B (2) share the same absolute configurations, while the bicyclo[4.2.0]octadieneoxide central scaffolds are enantiomeric with respect to each other. Emerione C has a bridging endoperoxide on the central core, and its proposed structure has a stereochemical configuration similar to emerione B.

Figure 1

Structures of the emeriones and related natural products.

Related substances like shimalactone A (3) and ocellapyrone B (4)[ , ] have been synthesized, but the emeriones are arguably the most complex examples of such natural products, each containing twelve stereocenters, eight of which are contiguous, and two quaternary. Moreover, the dioxabicyclo[3.1.0]hexane system, also found in natural products like verrucosidin (5), is a considerable synthetic challenge alongside the oxidized bicyclo[4.2.0]octadiene scaffolds. Emerione A inhibits NO production in lipopolysaccharide‐induced RAW264.7 cells as well as NDM‐1 at low micromolar concentrations, but the emeriones appear not to have been tested in other assays. Motivated both by their striking structures and potentially undiscovered bioactivities, we chose to target the emeriones for synthesis. We describe herein the successful completion of the syntheses, the structural revision of emerione C, and the synthesis of the originally proposed structure of emerione C, which we name emerione D.

It is plausible that the emeriones are biosynthetically derived from the unsaturated polyketide 6, which after two oxidations gives a diastereomer (7) of the natural product emecorrugatin B (8) (Figure 2, top).[5][6] Two double‐bond isomerizations then generate (E,E,Z,Z,E)‐pentaene 9, which is poised to undergo an 8π/6π electrocyclization cascade. This provides bicyclo[4.2.0]octadienes 10 and 11, which are oxidized to the emeriones. In our retrosynthesis (Figure 2, bottom), we modeled the late stages of our approach on the proposed biosynthesis. Therefore, emeriones A and B would be derived from 10 and 11, respectively, via mono‐epoxidations, and emerione C would be traced back to 11 via [4+2] cycloaddition with 1O2. Intermediates 10 and 11 would arise from pentaene 9 through an 8π/6π electrocyclization cascade, which would form only two of the four Woodward–Hoffmann compatible stereoisomers. Pentaene 9 would be constructed convergently, in a Stille coupling of iodide 12 and stannane 13. Stannane 13 could be derived from iodide 14, which would be prepared in a Paterson anti‐aldol of aldehyde 16 and ketone 15. Iodide 12 can be traced back to aldehyde 17 through a series of olefinations. The trisubstituted epoxide of 17 would be formed via oxidation of para‐methoxyphenyl acetal 18, which would be derived from triol 19. Sequential asymmetric oxidations would generate 19 from (Z,Z)‐dienol 20.

Figure 2

Proposed biosynthesis (top) and retrosynthetic plan (bottom).

Our synthesis began with iodide 22, which can be prepared in four steps from propargyl alcohol (21) (Scheme 1A). Aldehyde 23, synthesized by MnO2 oxidation of 22, is prone to isomerization/decomposition. It was therefore used immediately in a Paterson aldol with the E‐configured boron enolate of ketone 24 to give 25 in >95 : 5 diastereomeric ratio (dr). The relative and absolute configuration of 25 was confirmed via X‐ray crystallography. Silyl protection of the hydroxyl group gave 26, followed by reductive removal of the chiral auxiliary. The resulting ethyl ketone (27) was converted to isopropyl ketone 28 via kinetic enolate formation and trapping with methyl iodide. [10] Removal of the silyl protecting group to give 29 could only be realized with HF⋅pyridine; other fluoride sources resulted in significant retro‐aldol reaction, and deprotection was sluggish under acidic conditions. Stille reaction of 29 with Me6Sn2 gave stannane 13.

Scheme 1

Synthesis of 12 and 13. Reagents and conditions: 1. MnO2 (21 equiv), CH2Cl2, rt, 30 min; 2. Cy2BCl (1.8 equiv), Et3N (2.2 equiv), 24 (1.6 equiv), Et2O, −78 °C→0 °C then 23 (1.0 equiv), −78 °C→−20 °C, 51 % (2 steps); 3. TBSOTf (3.1 equiv), 2,6‐lutidine (4.3 equiv), CH2Cl2 −78 °C, 4.5 h, 93 %; 4. SmI2 (4.0 equiv), THF/MeOH, 0 °C, 1 h, 92 %; 5. LiHMDS (2.0 equiv), THF, −78 °C, then MeI (3.0 equiv), 1.5 h, 96 %; 6. HF⋅py/THF (1 : 4), 0 °C→rt, 18 h, 98 %; 7. Pd(PPh3)4 (5 mol %), Sn2Me6 (1.2 equiv), THF, 80 °C, 5 h, 68 %; 8. LiAlH4 (2.5 equiv), THF, 0 °C→rt, 2 h, 93 %; 9. MnO2 (16.5 equiv), CH2Cl2, rt, 18 h; 10. 31 (1.1 equiv), KHMDS (1.1 equiv), 18‐crown‐6 (3.0 equiv), THF, −78 °C, 1 h, then aldehyde (1.0 equiv), −78 °C, 1 h, 76 % (2 steps); 11. DIBAL (2.7 equiv), CH2Cl2, 0 °C, 1 h, 78 %; 12. Ti(Oi‐Pr)4 (0.23 equiv), (−)‐DET (0.27 equiv), 4 Å MS, CH2Cl2, −25 °C, 0.5 h, then TBHP (2.2 equiv), −25 °C, 0.5 h, then 20 (1.0 equiv), −40 °C, 24 h, 97 %, 81 % ee; 13. AD‐mix β (10 mass equiv), MeSO2NH2 (1.0 equiv), t‐BuOH/H2O (1 : 1), 0 °C, 18 h, 68 %, 86 % ee; 14. CSA (0.1 equiv), CH2Cl2, 0 °C, 20 h, then 37 (1.5 equiv), 0 °C→rt, 4 h, 59 %, 96 % ee (recrystallized); 15. 35 (1.0 equiv), p‐TsOH (0.2 equiv), HC(OMe)3 (1.1 equiv), THF; 16. DDQ (1.3 equiv), 4 Å MS, DCE, 80 °C, 2 h, quant.; 17. K2CO3 (6.0 equiv), MeOH, 0 °C→rt, 2 h, 89 %; 18. TPAP (0.05 equiv), NMO (1.5 equiv), 4 Å MS, CH2Cl2, rt, 1.5 h, 77 %; 19. 41 (1.04 equiv), THF, 100 °C (μ‐wave), 2 d, 67 %; 20. 43 (1.2 equiv), LiOt‐Bu (1.2 equiv), THF, 0 °C→rt, 1 h, then 42 (1.0 equiv), THF, rt, 3 h, >95 : 5 dr; 21. DIBAL (3.5 equiv), CH2Cl2, 0 °C, 3 h, 78 % (2 steps); 22. MnO2 (25 equiv), CH2Cl2, rt, 2.5 h, 98 %; 23. Ph3PEt+I− (4.0 equiv), n‐BuLi, (4.0 equiv), THF, 0 °C→rt, 30 min, then I2 (4.0 equiv), THF, −78 °C, 10 min, then NaHMDS (3.8 equiv), THF, −78 °C, 10 min, then 45 (1.0 equiv), THF, −78 °C, 2 h, 87 %, >95 : 5 dr. Ellipsoids of 25 and 36 are depicted at a 50 % probability level. Color code: C, grey; O, red; I, purple, Br, gold.

The synthesis of iodide 12 began with conversion of methyl angelate (30) into angelic aldehyde, which was found to be configurationally labile (Scheme 1B). Therefore, angelic aldehyde was immediately used in a Still–Gennari olefination with 31 to give dienoate 32, which was then reduced to give allylic alcohol 20.[ , ] Sharpless asymmetric epoxidation of 20 proceeded in excellent yield to give 33, but with a modest 81 % ee, as previously observed with Z‐configured allylic alcohols.

While Upjohn oxidation of epoxide 33 to give triol 19 was moderately diastereoselective (72 : 28 dr), Sharpless asymmetric dihydroxylation (SAD) proceeded with an improved dr of 86 : 14. Moreover, due to reagent control in the SAD reaction, 19 was isolated with 86 % ee (Scheme S1). Acid‐catalyzed isomerization of triol 19 proceeded with inversion of stereochemistry at C5 to give tetrahydrofuran 34, which contains the appropriate vicinal anti‐diol configuration for epoxide formation. After numerous attempts to advance 34 to aldehyde 17 (Scheme S2), we hypothesized that the epoxide in 17 could be formed via oxidation of an acetal like 18. Acetal formation was facile: treatment of triol 34 with aldehyde 35 under acidic conditions gave 36, whose absolute and relative configuration was confirmed via X‐ray crystallography. Noting that the preceding reaction is acid‐catalyzed, we developed a one‐pot procedure from triol 19 to acetal 18. In the event, after completion of the CSA‐catalyzed isomerization of 19 to 34, addition of acetal 37 produced 18, which could be crystallized to 96 % ee.

Pleasingly, oxidation of 18 with DDQ produced epoxide 39, presumably through the intermediacy of oxonium 38. To the best of our knowledge, this is the first synthesis of an epoxide from a 1,2‐diol using this approach.[ , ] Methanolysis of 39 gave alcohol 40, which oxidized to aldehyde 17 under Ley–Griffith conditions (TPAP/NMO). Wittig homologation of 17 produced aldehyde 42, which underwent Horner–Wadsworth–Emmons olefination and reduction to obtain alcohol 44. Manganese dioxide oxidation gave verrucosal (45), which was olefinated using Stork–Zhao conditions to produce iodide 12.

To complete the synthesis of the emeriones, 12 and 13 were combined in a Stille coupling to give pentaene 9 (Scheme 2). Stille conditions using Pd2dba3/P(2‐furyl)3/CuI or the Liebeskind variant (CuTC/Pd(PPh3)4) both successfully delivered product. Interestingly, 9 could be purified via chromatography and fully characterized with no apparent isomerization or decomposition. Upon heating in toluene at 55 °C, 9 slowly (3 d) and cleanly isomerized into a roughly equimolar mixture of 10 and 11, as estimated by 1H‐NMR. This outcome must arise via conrotatory 8π electrocyclization of 9 proceeding with essentially no induced diastereocontrol to produce cyclooctatrienes 47 and 48. These diastereomers then each undergo highly torquoselective 6π disrotatory electrocyclization to 10 and 11, respectively. Pleasingly, 10 and 11 were chemo‐ and stereoselectively epoxidized with m‐CPBA at the least hindered of their three double bonds to give (−)‐emerione A (1) and (−)‐emerione B (2), respectively. Spectroscopic and optical rotation data were consistent with the values reported by the isolationists (Tables S1, S2).

Scheme 2

Completion of the synthesis of the emeriones. Reagents and conditions: 1. 12 (1.0 equiv), 13 (1.5 equiv), Pd2(dba)3 (0.12 equiv), P(2‐furyl)3 (0.48 equiv), CuI (2.1 equiv), NMP, rt, 20 h, 54 % OR 12 (1.0 equiv), 13 (1.5 equiv), Pd(PPh3)4 (0.10 equiv), CuTC (1.1 equiv), DMF, rt, 1 h, 53 %; 2. PhMe, 55 °C, 3 d, 10: 30 %, 11: 28 %; 3. m‐CPBA (1.0 equiv), NaHCO3 (22 equiv), CH2Cl2/H2O (2 : 1), 0 °C→rt, 45 min, 1: 31 %, 2: 94 %; 4. O2, methylene blue (0.03 equiv), hν, DCE, 10 min, 49: 65 %, 50: 82 %.

When an O2‐saturated dichloroethane solution of 11 with triplet sensitizer was irradiated (400 W, white halogen lamp), a single endoperoxide adduct (50) was formed. We expected 50 to be (−)‐emerione C; however, comparison of NMR spectra of 50 and literature data for emerione C (Figure 3A, Table S4) made clear that the two substances are different. We therefore treated 10 under identical 1O2‐producing conditions to cleanly give endoperoxide 49. This compound had NMR spectra identical to those reported for emerione C (Figure 3B, Table S3). To unambiguously clarify the chemical structures, we solved the structure of 50 by X‐ray crystallography (Figure 3C), and found that it has the originally proposed structure of emerione C. We, therefore, reassign the structure of emerione C (49) as it is depicted in Scheme 2 and name compound 50, which may also be a natural product, (+)‐emerione D.

Figure 3

A) Comparison of emerione C and 50 13C shifts. B) Comparison of emerione C and 49 13C shifts. C) Experimental structure of emerione D (50). Ellipsoids depicted at a 50 % probability level. Color code: C, grey; O, red.

To gain insight into the stereochemical outcome of the electrocyclization cascade, we employed density functional theory (DFT) calculations at the SMD(toluene)‐M06‐2X/Def2‐TZVP//M06‐2X/Def2‐SVP level of theory. The calculations reveal that the two transition states (TS1 and TS2) leading from 9 to 47 and 48, respectively, are nearly isoenergetic as are 47 and 48 (Figure 4). Therefore, the rates of formation and thermodynamic stabilities of 47 and 48 are nearly equal, consistent with experimental observations. The subsequent 6π electrocyclizations of 47 to 10 and 48 to 11, were found to proceed via TS3 and TS4, respectively. These are 1.9 kcal mol−1 and 1.8 kcal mol−1 lower in Gibbs free energy than the diastereomeric transition states TS3′ and TS4′, respectively (Figure S4). The two methyl groups in 47 and 48, which end up on the cyclobutane rings of 10 and 11, were found to have opposing influences on the torquoselectivity of the 6π electrocyclization. Replacing the bridgehead (green) methyl (Figure 4) with a proton results in a reversal of both 6π‐electrocyclization torquoselectivities (Figure S5), indicating that the purple methyl prefers to reside on the convex face of the bicyclo[4.2.0]octadiene. This is consistent with previous calculations. Replacing the purple methyl with a proton had little effect on the torquoselectivity (Figure S6), suggesting a strong and dominant steric penalty when the bridgehead methyl is syn to the vinyl dioxabicyclo[3.1.0]hexane system. Removing both methyl groups resulted in an almost complete loss of diastereoselectivity (Figure S7).

Figure 4

Reaction coordinate diagram and DFT‐calculated transition states of the electrocyclization cascade. The purple and green methyl groups have opposing and unequal effects on the torquoselectivity of the 6π electrocyclization (see Supporting Information).

In conclusion, we have completed an asymmetric bioinspired synthesis of all three emeriones, each with a longest linear sequence of 17 steps. Our synthesis has resulted in the reassignment of the structure of emerione C and the proposal of an additional family member, emerione D. As biological data of the emeriones is limited, current efforts in our lab aim to discover biological activities of these fascinating substances.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Click here for additional data file.

Supporting Information

Click here for additional data file.