|Home | About | Journals | Submit | Contact Us | Français|
Acortatarins A and B have been synthesized via stereoselective spirocyclizations of glycals. Mercury-mediated spirocyclization of a pyrrole monoalcohol sidechain leads to acortatarin A. Glycal epoxidation and reductive spirocyclization of a pyrrole dialdehyde sidechain leads to acortatarin B. Acid equilibration and crystallographic analysis indicate that acortatarin B is a contrathermodynamic spiroketal with distinct ring conformations compared to acortatarin A.
Acortatarins A and B are novel spiroketal pyrrole alkaloids from the roots of Acorus tatarinowii (Figure 1).1 Structurally related pollenopyrrosides A and B were isolated contemporaneously from the pollen of Brassica campestris.2 Notably, acortatarins A and B exhibited significant antioxidant activity in a renal cell model for hyperglycemia-induced production of reactive oxygen species (ROS).1 Thus, these natural products are potential starting points for the development of new therapeutics to treat diabetic complications, cancer, and other conditions in which ROS are implicated.3 However, due to low isolation yields from the natural sources,4 efficient synthetic routes are needed to enable detailed biological evaluation. Herein, we report concise, modular syntheses of acortatarins A and B via stereoselective spirocyclizations of glycals. The thermodynamic preferences of both spiroketal natural products and the crystal structure of acortatarin B are also described.
In the original isolation paper, the relative configuration of acortatarin A was established by crystallography and an unnatural absolute l-configuration was assigned based on Mosher analysis.1 An α-ribo relative configuration was assigned to acortatarin B based on ROESY analysis and the l-configuration assumed by analogy. Notably, pollenopyrroside B was separately assigned the enantiomeric d-configuration of acortatarin A based on crystallographic analysis of its pyranose congener pollenopyrroside A (not shown).2
Subsequently, Sudhakar reported the first total syntheses of acortatarins A and B from 2-deoxy-d-ribose and d-arabinose, respectively, leading to structural revisions of both absolute configurations as well as the relative configuration of acortatarin B (Figure 1).5 Thus, acortatarin A and pollenopyrroside B are now recognized to be identical. A second synthesis of acortatarin A from d-mannitol was also reported recently by Brimble.6 These reports provide the first synthetic access to the acortatarins, but their practical utility is limited by low overall yields and reliance upon classical acid-catalyzed spiroketalization reactions that afford low or even undesired diastereoselectivity.7
Our laboratory has a long-standing interest in the stereocontrolled synthesis of spiroketals from glycals,8, 9, 10 and we envisioned that both acortatarins A and B could be synthesized by spirocyclizations of glycals 1 (Figure 2). Direct spirocyclization would provide acortatarin A while epoxidation–spirocyclization would lead to acortatarin B. In the latter case, we recognized that the oxidation state of the pyrrole substituents would be important for enabling chemoselective epoxidation of the glycal. These key intermediates 1 would originate from coupling of appropriate pyrroles 2 with ribal derivative 3, accessed via nucleobase elimination of thymidine.11 At the outset of our studies, the revised structures of the acortatarins had not been reported but, recognizing that both enantiomers of thymidine are commercially available, initial work was carried out with the less expensive, natural d-congener.
Thus, TIPS-protected12 ribal 611 underwent C1-formylation13 and reduction to provide hydroxymethyl ribal 7, which was then converted to iodide 8 (Figure 3).14 The pyrrole dicarboxaldehyde 915, 16 was then coupled under biphasic conditions17 to afford the key pyrrologlycal 10.
To access acortatarin A, we initially attempted reductive spirocyclization of dialdehyde 10 (TFA, Et3SiH), envisioning cyclization of an aldehyde carbonyl followed by in situ reduction of the resulting spirocyclic oxocarbenium intermediate, but this led to a furan side product via Ferrier-type elimination (Figure S1).16 Similarly, stepwise reduction to monoalcohol 11 (Figure 4) followed by treatment with dichloroacetic acid led to a 1:1 mixture of a 2,3-dehydro-α-spiroketal (cf. 12) via Ferrier rearrangement and the undesired β-spiroketal 13.16
Thus, we next investigated oxidative spirocyclizations of pyrrole monoalcohol 11 that would yield spiroketals having a removable C2-substituent, and were delighted to find that treatment with Hg(II) salts afforded the desired 2-mercurial spiroketals, which were then reduced with NaBH4 to afford the diastereomeric spiroketals 12 and 13.9d Initial reactions with Hg(OAc)2 or Hg(TFA)2 led to modest stereoselectivity favoring the desired α-spiroketal 12 (Table 1, entries 1–5). Notably, Hg(TFA)2 resulted in 30% formation of the same Ferrier rearrangement-derived 2,3-dehydro-α-spiroketal observed above (entry 2).
The Hg(OAc)2-derived 2-mercurial spiroketals exhibited a 7.8 Hz C2-H/C3-H coupling constant, consistent with a 2,3-trans relationship arising from β-mercuration (Figure S2).16 Since the expected anti-oxymercuration would then lead to the desired α-spiroketal 12,18 we postulate that the undesired β-spiroketal 13 arises from net syn-oxymercuration via an oxocarbenium intermediate. Thus, to accelerate anti-oxymercuration, pyrrole monoalcohol 11 was pretreated with NaHMDS to form a more reactive alkoxide nucleophile, resulting in increased stereoselectivity for the desired α-spiroketal 12 (entry 7). Surprisingly, however, longer reaction times prior to NaBH4 reduction led to further increased stereo-selectivity, indicative of an unanticipated equilibrium effect in this reaction (entries 6–10). Such equilibration was not observed without base,16 and other bases provided comparable or lower stereoselectivity.16 Desilylation of the mixture of 12 and 13 then provided the separable acortatarin A (14) and C1-epi-acortatarin A (15).16, 19
Next, we pursued an epoxidation–spirocyclization approach to acortatarin B.9l In initial epoxidation studies, pyrrole monoalcohol 11 and its diol congener (not shown) were prone to pyrrole oxidation. In contrast, pyrrole dicarboxaldehyde 10 underwent chemoselective β-epoxidation of the glycal with DMDO to form the putative epoxide 16 (Figures 5 and S3).16 Addition of NaBH4 in MeOH afforded the α-spiroketal methanol adduct 22a (Table 2, entry 1). In contrast, NaBH4 in THF provided the desired β-spiroketal 17 as a single diastereomer, along with a tetracyclic side product 21 (entry 2). Attempted ionic reduction with Et3SiH resulted only in tetracycle 21 (entry 3). Conversely, reductive spirocyclization with acidic NaBH3CN yielded the epimeric α-spiroketal 18 and tetracycle 21 (entry 4). NaBH(OAc)3 led to α-spiroketal acetate adduct 22b (entry 5) while LiEt3BH and L-Selectride yielded complex mixtures (entries 6,7). Finally, Bu4NBH4, aided by its solubility in CH2Cl2, provided the desired β-spiroketal 17 in excellent yield and diastereoselectivity (entries 8, 9). Spiroketals 17 and 18 were seperable and desilylation provided acortatarin B (19) and its C1-epimer (20).16, 20
We next investigated acid-catalyzed equilibration of the natural products and their unnatural C1-anomers, as well as the TIPS-protected congeners (12–15, 17–20).21 In both series, the α-spiroketal was favored by a 65:35 ratio (Figures 4, ,55).16 Notably, this favors the unnatural anomer of acortatarin B. Accordingly, although it is commonly assumed that spiroketal biosynthesis is a spontaneous, thermodynamically-controlled process, acortatarin B is a contrathermodynamic spiroketal whose biosynthesis may be under enzymatic stereocontrol.22, 23
Finally, we obtained a crystal structure of acortatarin B for comparison to the reported structure of acortatarin A (Figure 6).1 Interestingly, acortatarins A and B adopt distinct furanose envelope conformations (E1 vs. E2) and morpholine half-chair conformations (OH1 vs. 1HO) to allow double anomeric stabilization in both systems.
In conclusion, we have developed efficient, stereocontrolled syntheses of acortatarins A and B from a key pyrrologlycal 10. Acortatarin A was synthesized in 9 steps and 30% overall yield from d-thymidine, with 9:1 diastereoselectivity at the spiroketal-forming step and acortatarin B was accessed in 8 steps and 41% overall yield with complete diastereoselectivity. This compares favorably to previous syntheses7 and provides practical access to the natural products and a variety of analogues. Mechanistic analysis of the opposite stereoselectivities observed in these two spirocyclizations and biological studies are ongoing and will be reported in due course.
We thank Prof. Yong-Xian Cheng (Kunming Institute) for providing samples of the acortatarins, Dr. George Sukenick, Dr. Hui Liu, Hui Fang, and Dr. Sylvi Rusli (MSKCC) for expert NMR and mass spectral support, Dr. Kristen Kirschbaum (U. Toledo) for X-ray crystallographic analysis, and the NIH (P41 GM076267, T32 CA062948-Gudas) for financial support.
Supporting Information Available. Detailed experimental procedures and analytical data for all new compounds. This material is available free of charge via the Internet at http://pubs.acs.org.