Total synthesis of des-thiomethyllooekeyolide A

Abstract

The first asymmetric total synthesis and validation of the structural assignment of des-thiomethyllooekeyolide A (3) is described, featuring a Shiina macrolactonization and a late-stage pyran−hemiketal formation. The eight stereogenic centers of the C16-polyketide chain were installed by sequential aldol and crotylation reactions.

Graphical Abstract

Coral diseases have significantly influenced coral reef ecosystems around the world in the past decades,¹ and black band disease (BBD) was the first coral disease observed on reefs of Belize in the western Caribbean by Antonius in 1973.² BBD is one of the most pervasive and devastating coral diseases,³ and it consists of a cyanobacteria-dominated, sulfide-rich microbial mat. The mat migrates through coral colonies and degrades coral tissue,⁴ leaving behind the bare coral skeleton. Research indicates that the predominant cyanobacterium in the group that causes BBD disease is Roseofilum reptotaenium.⁵

Paul and coworkers in 2019 examined secondary metabolites present in the black band layer in situ during BBD from various geographic regions of Caribbean and Pacific coral reefs in an effort to find chemical signals or cytotoxins that could be crucial in the proliferation of R. reptotaenium and the BBD layer.⁶ From more than a dozen field-collected BBD samples, they obtained looekeyolide A (1) and looekeyolide B (2) (Figure 1) as significant components (∼1 mg g⁻¹ dry wt). Further research indicates that looekeyolide B (2) is an autoxidation product of looekeyolide A (1) and may play a biological role in reducing H₂O₂ and other reactive oxygen species. These reactive species may occur in the BBD layer as it overgrows and destroys coral tissue.⁶

Figure 1.

Structures of looekeyolides A and B, and des-thiomethyllooekeyolide A

The cyanobacterium probably produces looekeyolide A (1), which was only stable when exposed to helium gas..⁶ The stable looekeyolide B (2), on the other hand, exists as a mixture of R and S sulfoxide diastereomers. In order to establish the structure of looekeyolides A and B, looekeyolide B (2) underwent semi-synthetic desulfurization using Raney-Ni to produce. des-thiomethyllooekeyolide A (3). The absolute configuration of all stereogenic centers of des-thiomethyllooekeyolide A (3) was determined using X-ray crystallography and HPLC analysis of its hydrolysate⁶ (Figure 1).

Structurally, des-thiomethyllooekeyolide A (3) possesses a 20-membered macrocycle that is composed of a highly modified peptidic unit containing a D-leucine, (S)-2-hydroxybutanoic acid, and a C16-polyketide substructure.. The highly oxygenated and methylated C16-polyketide possesses an unstable pyran−hemiketal with eight stereogenic centers and one of which is quaternary. Having a long interest in assignment/revision of marine natural products by total synthesis,⁷ we initiated the synthetic endeavor of this novel macrocyclic natural product to confirm its structural configuration. Herein, we disclose the first asymmetric total synthesis of des-thiomethyllooekeyolide A (3), enabling the confirmation of its structure assignment.

Our retrosynthetic analysis for des-thiomethyllooekeyolide A (3) is illustrated in Scheme 1. We elected to disconnect the target molecule at two ester linkages to afford amide 4 and polypropionate segment 5. Because of the labile nature of the pyran−hemiketal function, we planned to construct this moiety at the late stage of the synthesis following the Shiina macrolactonization.⁸ Thus, the C1−C16 segment 5 would be convergently accessible from the aldol coupling of methyl ketone 6 and aldehyde 7. The C1−C10 segment 6 composed of five stereogenic centers would be constructed through a sequence of stereoselective aldol and crotylation reactions. In addition, the C11−C16 subunit 7 could be easily obtained by the Titanium enolate anti-aldol reaction developed by Ghosh et al.⁹

Scheme 1.

Retrosynthetic Analysis of 3

The construction of des-thiomethyllooekeyolide A (3), commenced with the stereoselective construction of the primary synthetic subtarget (6) (Scheme 2). We elected the titanium-mediated anti-aldol reaction of lactates bearing SuperQuats chiral oxazolidine-2-ones (9), developed by Kobayashi et al¹⁰ as a powerful, stereocontrolled entry to the masked chiral diol intermediate. As a result,, the commercial TBS-protected aldehyde 8 was reacted with the titanium enolate produced by treating the SuperQuats derivative 9 with LDA followed by the addition of TiCl(O-i-Pr)₃, to afford the aldol adduct 10 in 89% yield with dr >10:1. Subsequent conversion to the fully protected tetraol 11 was accomplished by sequential reductive cleavage of the chiral auxiliary with NaBH₄ in THF-H₂O, transformation of the primary hydroxyl to its TBDPS ether, and O-methylation of the secondary alcohol (73% overall yield). The subsequent chemoselective removal of the primary TBS ether from 11 to furnish the corresponding primary alcohol was accomplished using catalytic PPTS in ethanol. Treatment of this alcohol with catalytic TEMPO and an excess of trichloroisocyanuric acid (TCCA)¹¹ gave rise to an aldehyde and then immediately subjected to asymmetric Brown crotylation¹² to afford the corresponding homoallylic alcohol in 78% yield as a chromatographically separable 4:1 mixture of diastereomers, and the major diastereomer 12 was easily isolated and characterized. Treatment of 12 with PMB-Br and NaH in the presence of TBAI resulted in the detection of just a minute quantity of the unwanted product 13, despite prolonged reaction durations or elevated temperatures. Pleasingly, the PMB ether 13 could be obtained in 74% yield by treatment with KHMDS and PMBBr in the presence of triethylamine at −50 °C, followed by gradual warming to room temperature.¹³

Scheme 2.

Synthesis of the C1−C10 Segment 6

With intermediate 13 in hand, the next stage was set for the installation of the chiral hydroxyl group at the C3 position. Oxidative cleavage of the terminal olefin 13 with OsO₄ and NaIO₄¹⁴ followed by Brown allylation¹⁵ delivered the desired product 15 in moderate overall yield with low selectivity (dr < 2:1). We therefore considered alternatives, and to our delight, we found that the diastereoselectivity of the formation of homoallylic alcohol 15 could be increased to 10:1 through a four-step sequence involving oxidative cleavage of the terminal olefin, Grignard addition, Dess-Martin periodinane oxidation,¹⁶ and NaBH₄ reduction. The relative configuration of the newly generated hydroxyl group at C-3 was later confirmed by the nOe experiments of the PMP acetal 16, which was obtained from 15 by treatment with DDQ in anhydrous DCM solution. Methylation of 15 followed by deprotection of the TBDPS group gave rise to the primary alcohol 17 in 93% yield over two steps. Oxidation of 17 with Dess-Martin periodinane followed by Pinnick-Lindgren-Kraus oxidation¹⁷ to the corresponding acid and subsequent coupling to Weinreb’s amine gave 18 in 67% overall yield over three steps. DDQ oxidation effected the removal of the PMB ether, and the resultant alcohol was then reprotected as its TBS ether, and this was followed by addition of methyl lithium to furnish segment 6 in 91% yield (Scheme 2). It is worth noting that to protect 12 with a TBS group at an early stage proved to be completely unworkable in this case as the TBS group underwent 1,3-migration to the C3 hydroxyl group under the subsequent NaH/MeI methylation condition. All other methylation conditions that were tested either did not affect the starting material or resulted in decomposition of the substrate.¹⁸

With a reliable route to useful quantities of segment 6 in hand, we then turned our attention to the aldol coupling of 6 with aldehyde 7¹⁹ (Scheme 3). Surprisingly, the stereoselectivity of this aldol coupling reaction was affected by the reaction conditions. Using LDA or LiHMDS as a base in THF/HMPA resulted in an overwhelming preference for the unwanted diastereomer (20b), with a diastereoselectivity of 20a:20b = 1:10 to 1:15. Literature precedent²⁰ suggested that the selectivity varied depending on the counterion associated with the base employed and was also affected by the chiral substrates. To our delight, when the NaHMDS or KHMDS was employed as a base, the diastereoselectivity was reversed (20a:20b = 2:1 or 4:1, respectively), with the best selectivity for the Felkin diastereomer (20a) being obtained with the potassium enolate. The stereochemical outcomes of the base-promoted aldol coupling reaction upon formation of 20a and 20b, might be interpreted by Felkin-Anh model and Cram-chelate model,^20a respectively. Although the minor isomer could not be separated from the diastereomeric mixture of the aldol reaction by silica gel column chromatography at this stage, they were easily separated after the C-11 hydroxyl was protected as its TBS ether followed by removal of the PMB protecting group to furnish 5 in 58% overall yield over two steps. This compound, however, proved to be relatively unstable upon dissolution in NMR grade deuterated chloroform since it was completely transferred to hemiacetal 21 during ¹³C NMR experiments.²¹ The NMR spectroscopic analysis of hemiacetal 21 revealed its precise stereochemical configurations of the compound, including the NOE correlation between the methyl group and its adjacent proton, as indicated on its structure (see 21, Scheme 3). Additionally, this study also confirmed the stereochemistry of aldol product 20a.

Scheme 3.

Total Synthesis of des-thiomethyllooekeyolide A (3)

Having established the route to C1−C16 segment 5, we next explored the esterification and macrolactonization to complete the total synthesis. Esterification of alcohol 5 with acid 4 was initially attempted using Yamaguchi²² or Keck protocol;²³ However, these reaction conditions also caused formation of the hemiacetal 21 prior to esterification. To circumvent this problem, the more reactive acid chloride 22,²⁴ was employed, and the reaction was performed in the presence of NMM, a catalytic amount of DMAP in dichlorometane, which resulted in 89% yield of ester 23 with minimal hemiacetal formation. Removal of the N-Fmoc group of 23 with diethylamine (DEA) followed by condensation with hydroxyl acid 24 mediated by (7-Azabenzotriazol-1-yloxy)tripyrrolidinophosphonium hexa-fluorophosphate (PyAOP)²⁵ gave rise to peptide 25 in 61% yield. The osmium-mediated dihydroxylation (catalytic in K₂O_SO₄•2H₂O) of the terminal alkene was followed by diol cleavage with silica-gel supported sodium periodate²⁶ to afford the desired aldehyde which was immediately oxidized to the carboxylic acid 26 in 63% overall yield using Pinnick-Lindgren-Kraus oxidation conditions. Macrolactonization of seco-acid 26 was initially tried under the popular Yamaguchi conditions. However, in this case, the yield never exceeded 41%. In contrast, macrolactonization with the more reactive Shiina anhydride (2-Methyl-6-nitrobenzoic anhydride, MNBA) in the presence of DMAP induced lactone formation minimizing the formation of side-products and afforded the desired macrolide in 65% yield. Finally, the first total synthesis of des-thiomethyllooekeyolide A (3) was achieved through a one-pot process involving deprotection of silyl with concomitant hemiacetal formation by exposure of the macrolactone to hydrochloric acid, followed by hydrogenolytic removal of the benzyl ether. The proton and carbon spectra of fully synthetic des-thiomethyllooekeyolide A were identical to those of naturally derived material,⁶ though the spectra of the synthetic material were cleaner. The optical rotation of [α]²⁰_D +17.0 (c 0.20, MeOH) was identical in sign to but slightly lower than the reported value of [α]²⁵_D +29.3 (c 0.05, MeOH), therefore establishing the absolute stereochemistry as shown for 3.⁶

In summary, the first asymmetric total synthesis of des-thiomethyllooekeyolide A (3) was achieved in thirty steps over the longest linear sequence and in a 0.3% overall yield starting from readily available aldehyde 8. The key features of the synthesis include a Shiina macrolactonization and a late-stage hemiacetal formation. The eight stereogenic centers of the C16-polyketide chain were introduced stereoselectively through a sequence of aldol and crotylation reactions. The synthetic route we developed is general, efficient, and will allow for the rapid synthesis of other members of the looekeyolide family..