Programmable meroterpene synthesis

The bicyclo[3.3.1]nonane architecture is a privileged structural motif found in over 1000 natural products with relevance to neurodegenerative disease, bacterial and parasitic infection, and cancer among others. Despite disparate biosynthetic machinery, alkaloid, terpene, and polyketide-producing organisms have all evolved pathways to incorporate this carbocyclic ring system. Natural products of mixed polyketide/terpenoid origins (meroterpenes) are a particularly rich and important source of biologically active bicyclo[3.3.1]nonane-containing molecules. Herein we detail a fully synthetic strategy toward this broad family of targets based on an abiotic annulation/rearrangement strategy resulting in a 10-step total synthesis of garsubellin A, an enhancer of choline acetyltransferase and member of the large family of polycyclic polyprenylated acylphloroglucinols. This work solidifies a strategy for making multiple, diverse meroterpene chemotypes in a programmable assembly process involving a minimal number of chemical transformations.

31, 32, 33, 34, 35, 37, 38, and 39 are previously reported substrates (identical yields and dr) while 26,27, and 36 appear to be new. If the ratio of THF/Et2O is important, it worth mentioning. It is stated that lithium enolates are ritical in the process. Have the authors tried other enolates such as ones with Na, K, or cations? Additionally, some other ethereal solvents may be worth investigating if interested: MTBE and 2-MeTHF. 4. (Page 3, Paragraph 2) For 1.4:1 dr (substrate 38) and 1.5:1 dr (substrate 39), "predominantly" may be too strong of a term for modest selectivity. 5. (Page 4, Paragraph 2) The differences in reactions needed to elaborate the cyclopentane fragment compared to the hyperforin synthesis are interesting and some good insights are offered here. 6. (Page 4, Paragraph 3) In the Wacker oxidation literature, these reactions can proceed by synor anti-oxypalladation. In Figure 4 structure 43, anti-oxypalladation is shown as a possible mechanistic pathway. Palladium doesn't coordinate as well to isolated highly substituted alkenes, so I wonder if a syn-alkoxypalladation pathway could allow the palladium to coordinate to the alkene while being bound to oxygen of the vinylogous acid. Additionally, in Figure 4 conditions e for the unusual Wacker-type cyclization step, TMSOH is shown but its role is unclear. Is it important for the success of the reaction, and if so, what does it do? 7. (Page 5, Paragraph 1) To me, it seems like it is quite a conceptual jump to go from the discussion of entry 6 to entry 7 and from hypervalent iodine to Pd. I wonder if the discussion could be arranged to talk about the eventual successful solution with Pd Wacker-type cyclization last for a smoother and more progressive discussion. 8. (Page 5, Paragraph 3) It is quite surprising that such small changes can lead to such a big difference in reaction outcome. 9. (Page 5, Paragraph 3) The analysis of PPAR building blocks is a useful one and helps put this work, past work, and related work into context. While the two building most prevalent building blocks 16 and 41 have been transformed into bicyclo[3.3.1]nonanes, the most convincing demonstration of the advertised programmable approach to meroterpenes would include elaboration of 53 and 54 to bicyclic structures as well. However, the cited references do show other methods for synthesis of motifs not covered by the reported building blocks.
Formatting Changes Recommended: 1. (Page 1, Paragraph 3) "…garsubellins (8-9)8-9" --superscript should be a comma for consistency with other formatted references 2. (Page 1, Paragraph 3) "Hundreds of bicycle[3.3.1]nonane-containing PPAPs exists…" --(exists should be changed to exist for sentence agreement) 3. (Page 2, Paragraph 3) "labelling" --fun fact learned during review for this paper: labeling is American English spelling while labelling is British English spelling. 4. (Page 2, Paragraph 4) "isobutyrl" --misspelled and should be isobutyryl as in paragraph 3. 5. (Page 3, Paragraph 4) "disease51,52." --move period before the references.  Figure 5) The mechanism for this ring expansion looks different from identical conditions on a similar substrate in Maimone's previous hyperforin synthesis. Perhaps I am missing something, but has the mechanistic understanding changed since then such that a different mechanism is being proposed? 13. (Experimental Procedures, Supplementary Note 1, Page SI-2). For the chromatography conditions 5% Et2O in hexanes -> 15% EtOAc in hexanes, should Et2O be EtOAc? 14. (Experimental Procedures, Supplementary Note 2, Page SI-3). How is LTMP prepared? Usually this is good to include. Or at least cite a reference to indicate which procedure is being followed. 15. (Experimental Procedures, Supplementary Note 2). As a general question, do all of these compounds exist predominantly in the diketone form with little to no enol? 16. (Experimental Procedures, Supplementary Note 2). As a general comment, it seems that melting point is sometimes listed as "m.p." and other times as "mp." IR reporting format is slightly different between sections of the experimental section with "vmax" included in some places but not others. These are relatively minor. 17 In the present manuscript, Maimone and co-workers report a concise (10 steps) total synthesis of PPAP natural product garsubellin A via a key oxidative isomerization of the 5,6-fused bicycle to the bicyclo[3.3.1]nonane core structure. The tetrahydrofuran moiety was synthesized via an unusual vinylogous acid Wacker-type cyclization. Additionally, a lithium enolate 4-carbon annulation reaction was developed to synthesize the 5,6-fused bicyclic precursor, and a series of 1,3diketones were prepared. The manuscript and the Supporting Information (experimental details and NMR spectra) were well prepared. I am quite supportive of this elegant natural product synthesis work being published in Nature Communications. The authors should address the following minor comments and concerns below before publication: 1. Additives such as HMPA and DMPU are known to increase the reactivity of lithium enolate. In the annulation reaction (

Reviewer #1
1) Comment: The superscript "#" on the authors' name is not defined in the manuscript. Response: This has been added.
2) Comment: It could be interesting if the author can demonstrate the oxidative rearrangement with the diketone 35 as the product is hard to access by other method. Response: This is an excellent suggestion and something we have also considered. To date, however, we have not successful shifted a non-bicyclic substrate.
3) Comment: A newly published total synthesis that could access to a series of endo-type B PPAPs with the bulky geranyl groups should be cited (Org. Lett. 2019, 21, 8075-8079). Response: This work has been added to the citations.

4) Comment:
In the discussion, the author has analyzed the potential synthesis of exo-type A and B PPAPs with current strategy, which concluded that over 90% of the exo-members could be accessed. In the reviewer's opinion, maybe the hypothesis is better to only include the exo-type A as no synthesis or further discussion of the exo-type B was revealed in the manuscript. Response: This is a fair point and we have modified the text to clearly state what we can actually potentially make using the chemistry documented in the manuscript. In addition, we have modified Figure 6 to only show the % of members for the type A skeleton.

Reviewer #2
1) Comment: "What structural features make type A PPARs the most synthetically challenging? Perhaps a brief mention could offer some clarification." Response: This was based on several considerations and observations in the field. First, the type A PPAPs have historically required the most synthetic steps to construct-often by a significant margin. For example, hyperforin initially took 51 steps to synthesize and garsubellin A close to 20 steps. Second, biomimetic cyclizations, which typically offer the shortest routes to PPAPs, often do not favor formation of type A skeletons. This is exacerbated by type A PPAPs which possess a quaternary center in close proximity to a bulky isopropyl ketone (such as found in both garsubellin and hyperforin). To the best of our knowledge, these targets have not been made through direct biomimetic cyclization.
We do however, acknowledge that "synthetic challenge" depends on the strategy being used, thus making our original statement more of an opinion than hard fact. To avoid confusion/controversy, we have therefore omitted "the most synthetically challenging" in the text.
2) Comment: As a general curiosity question, is diketene easy to obtain? Who sells it? Some vendors indicate that this product is continued and no longer for sale. This might be of interest to someone who wants to apply the annulation method in their work. Response: This is an excellent question. Our initial work used Sigma-Aldrich diketene, but we recognize they have since discontinued this product. Several vendors in China sell it, thus offering an easy solution to researchers there. An Org. Synth. Procedure has also been reported to prepare it, but requires quite sophisticated equipment and is not useful for researchers wanting small amounts quickly.
As a service to the synthetic community, we have detailed in the Supplementary information section (new supplementary note 1) a simple, "in-house" protocol to prepare diketene from inexpensive acetyl chloride requiring only a simple Hickman distillation head for isolation. The text has also been modified to note this preparation is include in the SI.
We have evaluated these other etheral solvents and found no improvement. The solubility of each enolate also dictates to some extent the optimal solvent mixture for the transformation. For certain substrates, pure Et 2 O is fine and for others, adding amounts of THF can be beneficial. We have also evaluated Na and K in the form of the bases NaH, NaHMDS, and KHMDS (the K and Na variants of LTMP were not used). These gave messier reactions and we could only detect mainly Oacylation in the crude mixture. This is in line with counterion effects on enolate acylation reactions with Mander's reagent as well. 4) Comment: For 1.4:1 dr (substrate 38) and 1.5:1 dr (substrate 39), "predominantly" may be too strong of a term for modest selectivity. Response: We have removed this descriptor.

5) Comment:
In the Wacker oxidation literature, these reactions can proceed by syn-or antioxypalladation. In Figure 4 structure 43, anti-oxypalladation is shown as a possible mechanistic pathway. Palladium doesn't coordinate as well to isolated highly substituted alkenes, so I wonder if a syn-alkoxypalladation pathway could allow the palladium to coordinate to the alkene while being bound to oxygen of the vinylogous acid. Additionally, in Figure 4 conditions e for the unusual Wacker-type cyclization step, TMSOH is shown but its role is unclear. Is it important for the success of the reaction, and if so, what does it do? Response: These are excellent questions. Either could be possible, but both form the same product in this case. The usual mechanistic probes are not valid in this setting since a gem-dimethyl group is formed in the intermediate. We do note that in the Wacker literature, the syn oxypalladation is formed in many instances, but these typically involve alcohol nucleophiles. In the case of a phenol (which is closer in pka to a vinylogous acid as used herein) the anti pathway is formed with an LPdCl 2 catalyst and the syn pathway formed with a cationic Pd(II) source; Pd(OAc) 2 is likely not cationic under the neutral conditions we used. We have thus added to the text "For simplicity, only an anti-oxypalladation pathway is depicted" and have added a reference on precisely this issue.
The use of TMSOH came initially from our attempts to trap the Pd(II) intermediate with a second oxygen nucleophile (TMSOH, H 2 O, Allyl alcohol, etc) which could become the garsubellin A tertiary hydroxyl group. Not surprisingly, this never proved possible, but we found that the yield of the alkene was boosted from ~80% yield to 95% yield by applying the TMSOH:DMSO solvent mixture. This could be due to solvation effects since out of the many conditions tried only the TMSOH:DMSO mixture generated a homogenous mixture after the addition of Cu(OAc) 2 . 6) Comment: To me, it seems like it is quite a conceptual jump to go from the discussion of entry 6 to entry 7 and from hypervalent iodine to Pd. I wonder if the discussion could be arranged to talk about the eventual successful solution with Pd Wacker-type cyclization last for a smoother and more progressive discussion. Response: These results are discussed somewhat in the order which they occurred in the project. Entry 6 was nice since it formed some of the tertiary alcohol product directly, but only in low yield (25%), a result which we could not boost further. Given that most of the material (40%) was the alkene, this result necessitated a second Mukaiyama hydration step to convert the mixture of products into the tertiary acohol. Since a second step would be needed, we then focused on just optimizing for the alkene, thus the switch to Palladium which is superior for this type of transformation. 8. Comment: It is quite surprising that such small changes can lead to such a big difference in reaction outcome. Response: We were surprised too. The garsubellin THF ring is quite susceptible to base-mediated hydrolysis. Once the coupling is complete the reaction needs to be stopped. Dioxane is a common, and often optimal, solvent for cross coupling reactions, but can also function as a reducing agent if transmetallation doesn't proceed rapidly. 9. Comment: The analysis of PPAR building blocks is a useful one and helps put this work, past work, and related work into context. While the two building most prevalent building blocks 16 and 41 have been transformed into bicyclo[3.3.1]nonanes, the most convincing demonstration of the advertised programmable approach to meroterpenes would include elaboration of 53 and 54 to bicyclic structures as well. However, the cited references do show other methods for synthesis of motifs not covered by the reported building blocks. Response: As suggested by reviewer #1, we have altered figure 6 and the final discussion to only focus on the type A PPAPs which are the focus of this work. Given this, former substrates 53 and especially 54 (now numbered 62 and 63) form only a very small proportion of type A members (14% and 2% respectively) and we believe of less interest to the readers of this publication. We agree that there are other existing methods to prepare these types of PPAPs. Given we have a route to the products leading to nearly 80% of the relevant members in a minimal number of steps using this chemistry and we have previously demonstrated entry into DMOA members, we believe our advertisement has merit. 10) Comment: (Page 1, Paragraph 3) "…garsubellins (8-9)8-9" --superscript should be a comma for consistency with other formatted references Response: This has been corrected 11) Comment: (Page 1, Paragraph 3) "Hundreds of bicycle[3.3.1]nonane-containing PPAPs exists…" --(exists should be changed to exist for sentence agreement) Response: This has been corrected 12) Comment: (Page 2, Paragraph 3) "labelling" --fun fact learned during review for this paper: labeling is American English spelling while labelling is British English spelling. Response: Interesting 13) Comment: (Page 2, Paragraph 4) "isobutyrl" --misspelled and should be isobutyryl as in paragraph 3.