Biomimetically inspired asymmetric total synthesis of (+)-19-dehydroxyl arisandilactone A

Complex natural products are a proven and rich source of disease-modulating drugs and of efficient tools for the study of chemical biology and drug discovery. The architectures of complex natural products are generally considered to represent significant barriers to efficient chemical synthesis. Here we describe a concise and efficient asymmetric synthesis of 19-dehydroxyl arisandilactone A—which belongs to a family of architecturally unique, highly oxygenated nortriterpenoids isolated from the medicinal plant Schisandra arisanensis. This synthesis takes place by means of a homo-Michael reaction, a tandem retro-Michael/Michael reaction, and Cu-catalysed intramolecular cyclopropanation as key steps. The proposed mechanisms for the homo-Michael and tandem retro-Michael/Michael reactions are supported by density functional theory (DFT) calculation. The developed chemistry may find application for the synthesis of its other family members of Schisandraceae nortriterpenoids.

For the geometry optimization the B3LYP functional is chosen, though the energi es are obtained with M06 functional. Use of different methods for energies and structures is a popular and often a proper compromise, especially if the method used for the energies is too expensive for more demanding structure calculations (e.g. MP2//DFT combination is popular). M06 and B3LYP are both hybrid DFT functionals of the comparable computational cost. Is there a special reason for changing of the DFT approach after the optimization?
The B3LYP functional neglects the dispersion effects, and the M06 takes them only partly into account. Therefore, I would recommend to apply D3-correction in order to check the influence of the dispersion interactions on the computational results.
FOR ALL COMPOUNDS The supplementary information does not contain any experimental information with regards to the data collection for the crystal data submitted e.g. experimental information, diffractometer information etc, nor are there tables reporting the crystallographic refinement information for the submitted crystal structures. A section containing these details must be included. It is recommended that the authors use the latest version of Shelx (ShelX-20XX) to perform the final refinements. This would provide more information about the refinement (instruction file and HKL listing). Embedded refinement details are now required when submitting CIFs.
Compound 10: •The completeness of the data is poor, despite the crystallographic symmetry being monoclinic. No reasoning for this is given either in the CIF or the supplementary information. This should be addressed either by collecting more data to achieve at least 98% completeness to 0.8Å or by stating a valid reason for the lack of data in either the supplementary information or the CIF. •Information regarding the crystal _exptl_crystal_description and _exptl_crystal_colour both missing from the CIF. This information should be included.
Compound 16: •Information regarding the crystal _exptl_crystal_description and _exptl_crystal_colour both missing from the CIF. This information should be included. •Ratio Observed / Unique Reflections (too) Low (25%). This implies that the data is weak or non-existant and that only 25% of the collected data are statistically 2 sigma above the background noise of the experiment. This is unacceptable. This is not addressed in either the supplementary information or the CIF. Perhaps recollection for a longer exposure or collection of the data at a synchrotron would yield better data. •Structure contains some questionable bond distances e.g. Small Average Phenyl C-C Distances and a C21 -C22 Sp3 carbon bond of 1.615(8) Å. These should be addressed wither through restraints or through refinement against a better data set.
•Large thermal parameters on main molecule, again indicative of a poor data set or lack of suitable restraints. •Information regarding cell indexation missing. Check. •Temperature of cell measurement and data collection reported as 293K. Check. •Information regarding the crystal also missing e.g. description and colour Compound 21: •Much experimental information missing from the CIF (device type, cell measurement information, crystal description/colour, absorption correction information). This information must be included.
•Relatively large residual peak in the Fourier difference map. This could be indicative of accounted for disorder or twinning, an error in the absorption correction or bad reflections in the refinement. •Temperature of cell measurement and data collection reported as 293K. Check. •Low completeness of data collection, despite the monoclinic crystallographic symmetry. More reflections should be collected to achieve 97% completeness to 0.83Å when using a CuKα source.
Compound 26: The authors don't appear to have even attempted to ensure that this CIF is competed to publication standards. This CIF is purely a direct output CIF from Shelx97 without inclusion/combination with other CIFs regarding the experiment. •Basic crystallographic information is missing from the CIF _symmetry_cell_setting •All experimental information missing from the CIF (device type/method , cell measurement information, crystal description/colour/dimensions, absorption correction information, data collection/reduction software). These MUST be included. •Absolute parameter shift to su on final refinement much greater than 0.2 (9.308). This indicates the refinement did not converge. Additional refinement steps must be performed to achieve convergence. If convergence is not achieved then this indicates an error within the model or the data. •Large residual peak in the Fourier difference map. This could be indicative of accounted for disorder or twinning, an error in the absorption correction or bad reflections in the refinement. •Structure contains solvent accessible voids. These are not accounted for in the supplementary information or CIF. Assignment of additional peaks in the difference map or use of SQUEEZE could be employed to account for this void space. •Ratio Observed / Unique Reflections (too) Low (46%). This implies that the data is weak or non-existant and that only 46% of the collected data are statistically 2 sigma above the background noise of the experiment. This is unacceptable. This is not addre ssed in either the supplementary information or the CIF. Perhaps recollection for a longer exposure or collection of the data at a synchrotron would yield better data. •R1 value is high. This is not accounted for in the supplementary information or CIF. Collection of better data or use of a correct model should be used in the refinement. •Low completeness of data collection, despite the monoclinic crystallographic symmetry. More reflections should be collected to achieve 98% completeness to 0.8Å when usin g a MoKα source. •Temperature of cell measurement and data collection reported as 293K. Check.
Compound 29: •Information regarding the crystal _exptl_crystal_description and _exptl_crystal_colour both missing from the CIF. This information should be included. •Flack parameter >0.5. This could be indicative of miss assigned chirality and the structure being inverted •Large variations in thermal parameters for non-solvent atoms. Should be addressed with appropriate restraints Compound 31: •O1 refined to 0.285(14) occupancy. Is this partial occupancy justified? Combined with short O163 -O1 D -A distance, is the atom correctly assigned? •Flack parameter meaningless (su>>value). This should be revised or addressed in the supplimenary information or CIF. Method for determining absolute configuration has not been stated in the CIF •Information regarding the crystal _exptl_crystal_description, _exptl_crystal_colour and dimensions missing from the CIF. This information should be included. •Large thermal parameter discrepancy and C41 large ADP max/min ratio. These should be refined using suitable restraints to generate a realistic model. •Centre of gravity of residues not within the unit cell. Residues should reside mainly within the unit cell.
Compound 32: •Flack parameter meaningless (su>>value). This should be revised or addressed in the supplimenary information or CIF. •Ratio of maximum/minimum residual density could be an indication of missed disorder/twinning or the inclusion of bad reflections in the refinement •Information regarding the crystal _exptl_crystal_description and _exptl_crystal_colour both missing from the CIF. This information should be included.
Question 1: Clearly state in the manuscript that 19-dehydroxyl arisandilactone A is not a natural compound and modify Figure 1 correspondingly (currently it misleads the reader as it states "Naturally occurring nortriterpenoids" and includes the structure of 1a). Answer: We changed the title of

Question 3:
The computational studies should be applied to the system with the substrates having C19 OMe group to demonstrate that their conclusions actually support biomimetic proposal. Answer ： Thanks for this comment. As described in the revised manuscript, the 19-dehydroxyl arisandilactone A, which has C19 hydrogen atom, was synthesized in experimental part. The theoretical calculation thus employed 19-dehydroxyl arisandilactone A as the substrate. Our computational results showed that the functional groups on C19 position would not participate in the hydroboration step, the homo-Michael step, or the retro-oxo Michael reaction. Therefore, using the substrates having C19 OMe group in DFT studies will not affect the conclusion. In addition, from the comparison of NMR spectrum of the natural arisandilactone A with our synthetic 19-dehydroxyl arisandilactone A, the deviation for the chemical shifts of the 1 H-NMR associated with the natural arisandilactone A and our synthetic 19-dehydroxyl arisandilactone A are quite small, except the proton at C19 (see page 74 in SI), indicating both natural arisandilactone A and our synthetic 19-dehydroxyl arisandilactone A adopt similar conformation, and our biomimetic proposal could be applied to the biomimetic synthesis of natural arisandilactone A.
Question 2. In references 9 and 10 (p.15), the authors should list the whole authors including the corresponding author Y.C. Shen. Answer: According to the guidelines of Nature Commun, when the numbers of authors is over 5, only the first author name to be cited. We therefore don't make the changes as suggested.
Question 3. In the beginning, we observe a great vision and detail discussion for the target. But, the references for the inhibitory activity toward tumors and hepatitis were not cited in the References. Answer: We added a sentence at end of our manuscript: The biological investigation of the synthesized (+)-19-dehydroxyl arisandilactone A (1a), as well as some advanced intermediates in this total synthesis is currently underway in our laboratory. See the Track Change Version at page 21, first paragraph, line 3.
Question 4. The authors should expect for more distinct and precise result and discussion concerning the yields and challenging steps. Answer: We appreciate the reviewer's precise comment, and made some revisions in our text for more distinct and precise result and discussion concerning the yields and challenging steps. Example 1: in the Track Change Version at page 12, second paragraph, line 3: "Exposure of 20 to a solution of N,O-dimethylhydroxylamine hydrochloride and isopropylmagnesium chloride in THF at −20 °C 37 formed a Weinreb amide in 90% yield, which was reacted with methylmagnesium chloride to afford a ketone in 89% yield. This ketone then underwent Peterson olefination 38 by reaction with freshly produced (trimethylsilyl)methylmagnesium chloride in the presence of CeCl 3 39 in THF at 0 °C, and the resultant mixture was worked up by the treatment with silica gel to give an olefin, which, without purification, was then subjected to a desilylation with TBAF to give 21 in 82% yield four the two steps." Example 2: in the Track Change Version at page 14, first paragraph, line 3, the dew description is "Thus, after intensive experimentation, we found that treatment of 24 with BH 3 ·Me 2 S 43 in THF at 0 °C, followed by oxidation with a solution of H 2 O 2 (30%) in the presence of a weak base (Na 2 B 4 O 7 ) 44 , product 26 could be obtained in 65% yield, together with its diastereoisomer (16% yield)." Example 3: in the Track Change Version at page 14, second paragraph, 2 line, the new description is " In the event, 27 was coupled with 28 in the presence of BF 3 ·Et 2 O in CH 2 Cl 2 at -78 o C, and the resultant mixture was then stirred at -35 o C for 1 h, followed by workup to give product 29 in 66% yield as a single diastereoisomer (Figure 8). Example 4: in the Track Change Version at page 116, first paragraph, line 1, the new description is: "We therefore, treated 31 with NaOMe in methanol to invert its stereochemistry at C25, and the resultant product was then subjected to Raney nickel-mediated debenzylation to give product 32 in 73% yield and 85% yield, respectively. Thus, our total synthesis of 19-dehydroxyl arisandilactone A (1a) was eventually achieved in 87% yield by oxidation of 32 with DMP." Question 5. Could the authors propose mechanism and cite application references for the key step Oxa-homo-Michael and retro-Michael/Michael reaction? Answer: The mechanism for the Oxa-homo-Michael has been illustrated in the Figure 3, intermediate B, and the references are cited as references 14-15. For the retro-Michael/Michael reaction, its mechanism has been illustrated in Figure 3 (see page 6), and the reference is cited as references 18.
Question 5. Some differences and mistakes were found between contents and references. Answer: We have made some corrections during the revision of our text. Zhen Yang) presents a synthesis and a quantum chemical study of the reaction pathway for an important bioactive compound from the Schisandraceae family -(+)-19-dehydroxyl arisandilactone. The authors emphasize, that no synthesis of this compound has been reported earlier. The ability to synthesize such a complicated but promising bioactive system seems to be indeed very important. Since the suggested synthesis involves several steps, where the mechanism is not obvious and different processes could happen simultaneously, the quantum chemical modeling can substantially support the experimental evidence.
Several questions can nevertheless arise for the computational part of the study. The choice of the quantum chemical method should be justified.
For the geometry optimization the B3LYP functional is chosen, though the energies are obtained with M06 functional. Use of different methods for energies and structures is a popular and often a proper compromise, especially if the method used for the energies is too expensive for more demanding structure calculations (e.g. MP2//DFT combination is popular). M06 and B3LYP are both hybrid DFT functionals of the comparable computational cost. Is there a special reason for changing of the DFT approach after the optimization? Answer: Thanks for the comments. The suggestion of MP2//DFT for calculation has been accepted. The MP2/6-311+G(d) method was employed for solvation single point energy calculation in THF solvent. All structures in Figure 10, 11, and 12 have been recalculated using MP2. However, as MP2 functional is too expansive, solvation single point energy calculation on some structures could not be achieved. Specifically, for some structures in Figure 11 and 12, which have more than 90 atoms, there would be more than 1216 basis functions and 1994 primitive gaussians in MP2 calculation. Corresponding calculation would be terminated because the RWF file generated during calculation process is too large. Fortunately, all the MP2 calculations towards the structures in Figure 10 are successful. The computational results are shown below as Figure S1. Besides, another suggestion about the dispersion effects was also considered. B3LYP-D3(BJ) method with 6-311+G(d) was also used to recalculate the solvation single point energies of structures in Figure 10. As shown in the following free energy profile, the values in parenthesis are free energies obtained by B3LYP-D3(BJ) and corresponding values in square brackets are free energies obtained by MP2. Comparison between these two sets of data suggests that the energy values calculated at MP2/6-311+G(d) level of theory are very close to that obtained by B3LYP-D3(BJ)/6-311+G(d). On another hand, the overall free energy profile obtained by MP2 is higher in comparison with M06 calculated data. However, the tendencies shown in these two sets of data are all the same, and the conclusion would not be changed. Therefore, the M06 functional could give reliable energy information for this work. The Figure S1 was added in Supporting Information for clarity.
With regards to the combination of M06 with B3LYP, this strategy has been widely used in Houk and other chemists' work. Its reliability and validity have been demonstrated. Specifically, B3LYP is known as a popular method for DFT calculation and has been proven to give accurate structural information. While Prof. Truhlar developed M06 functional, which is more expensive than B3LYP, could afford more accurate energy information. Therefore, the combination of M06 with B3LYP could efficiently gain reliable structural information and energy information in DFT calculation.  Figure S1 Energy profiles for the hydroboration of 24 calculated by B3LYP-D3(BJ)/6-311+G(d) and MP2/6-311+G(d) in THF.

Question 2:
The B3LYP functional neglects the dispersion effects, and the M06 takes them only partly into account. Therefore, I would recommend to apply D3-correction in order to check the influence of the dispersion interactions on the computational results. Answer: This suggestion was also adopted in this work. B3LYP-D3(BJ) method with 6-311+G(d) was used to recalculate the solvation single point energies of structures in Figure 10, 11, and 12. Computational result for Figure 10 was given as Figure S1, and corresponding results for Figure 11 and 12 were summarized in Figure S2 and S3. As shown below, the overall free energy profiles in Figure S2 and S3 is lower than that in Figure 11 and 12, although the free energy profile in Figure S1 is higher than that in Figure 10. More importantly, these data indicate that the tendency shown in B3LYP-D3(BJ) calculated data is very similar to that obtained by M06, and the conclusions would remain the same. Consequently, computational results obtained by the method B3LYP-D3(BJ) is consistent with the data in our manuscript. Figure S2 and S3 are added in the Supporting Information for clarity.

Question 3:
The authors declare only M06 DFT approach in the main text of the paper, though the supplementary information contains a lot of additional data, such as another method for the structures, the applied basis sets and solvation model. I would recommend also describing these important details in the main text shortly. Answer: Thanks for the comments. We have revised the manuscript and added the description of basis sets and solvation model in page 11. The revised text is also attached below: Computational study. Density functional theory (DFT) calculations were carried out using the Gaussian 09 program (Gaussian Inc., Wallingford, CT) to determine the diastereoselectivity for the hydroboration step to stereoselectively form 26 from 24 (see Figure 10). The method B3LYP with 6-31G(d) basis set was used for geometry optimizations in gas phase. M06 functional with larger basis set 6-311+G(d) was employed for solvation single point calculations in THF based on gas-phase stationary points using the continuum solvation model SMD (see SI for details). Question 4: On the pictures 10, 11 and 12 the title of the y-axis should be ∆G(M06). It is probably only a minor problem with the text format, but in the file, which I was able to download, the symbol "∆" is corrupted. Answer: Our response: Thanks for this kind reminder. We have revised the title of the y-axis in Figure 10, 11 and 12.

Reviewer 4：
Question 1: Compound 10: The completeness of the data is poor, despite the crystallographic symmetry being monoclinic. No reasoning for this is given either in the CIF or the supplementary information. This should be addressed either by collecting more data to achieve at least 98% completeness to 0.8Å or by stating a valid reason for the lack of data in either the supplementary information or the CIF. Information regarding the crystal _exptl_crystal_description and _exptl_crystal_colour both missing from the CIF. This information should be included. Answer: The poor quality of the X-ray crystallographic data for compound 10 was the result of the crystal that was slightly moved unexpectedly during the data collection of this compound, and this result is not awarded until far afterwards. Due to the material of compound 10 was not available, and it is difficult to resynthesize it, we could not cultivate the new crystals for its X-ray crystallographic studies. We therefore used the original crystal data for further conversion after XRD analysis. During the time of our refine the X-ray data of compound 10, we deleted the last groups of images, as a result, the R factor is greatly improved, and seems to be meaningful. We hope that the current X-ray crystallographic data with the 95.9% completeness is acceptable to support the structure of compound 10.
In addition, we also added a sentence in our revised text to describe this point. "and its structure was tentatively confirmed through X-ray crystallography; however, the data were of insufficient quality to allow a definitive determination of the structure." (see the Track Change Version at page 5, second paragraph, line 2) Question 2: Compound 16: Information regarding the crystal _exptl_crystal_description and _exptl_crystal_colour both missing from the CIF. This information should be included. Ratio Observed / Unique Reflections (too) Low (25%). This implies that the data is weak or non-existant and that only 25% of the collected data are statistically 2 sigma above the background noise of the experiment. This is unacceptable. This is not addressed in either the supplementary information or the CIF. Perhaps recollection for a longer exposure or collection of the data at a synchrotron would yield better data.
Structure contains some questionable bond distances e.g. Small Average Phenyl C-C Distances and a C21 -C22 Sp3 carbon bond of 1.615(8) Å. These should be addressed wither through restraints or through refinement against a better data set.
Large thermal parameters on main molecule, again indicative of a poor data set or lack of suitable restraints.
Information regarding cell indexation missing. Check. Temperature of cell measurement and data collection reported as 293K. Check. Information regarding the crystal also missing e.g. description and colour.

Answer:
The XRD data for compound 16 have been re-collected as suggested. The CCDC number has been provided in the SI.

Question 3:
Compound 21: Much experimental information missing from the CIF (device type, cell measurement information, crystal description/colour, absorption correction information). This information must be included. Relatively large residual peak in the Fourier difference map. This could be indicative of accounted for disorder or twinning, an error in the absorption correction or bad reflections in the refinement.
Temperature of cell measurement and data collection reported as 293K. Check. Low completeness of data collection, despite the monoclinic crystallographic symmetry. More reflections should be collected to achieve 97% completeness to 0.83Å when using a CuKα source. Answer: The XRD data for compound 21 have been re-collected as suggested. The CCDC number has been provided in the SI.
Question 4: Compound 26: The authors don't appear to have even attempted to ensure that this CIF is competed to publication standards. This CIF is purely a direct output CIF from Shelx97 without inclusion/combination with other CIFs regarding the experiment.
Basic crystallographic information is missing from the CIF _symmetry_cell_setting. All experimental information missing from the CIF (device type/method, cell measurement information, crystal description/colour/dimensions, absorption correction information, data collection/reduction software). These MUST be included.
Absolute parameter shift to su on final refinement much greater than 0.2 (9.308). This indicates the refinement did not converge. Additional refinement steps must be performed to achieve convergence. If convergence is not achieved then this indicates an error within the model or the data.
Large residual peak in the Fourier difference map. This could be indicative of accounted for disorder or twinning, an error in the absorption correction or bad reflections in the refinement.
Structure contains solvent accessible voids. These are not accounted for in the supplementary information or CIF. Assignment of additional peaks in the difference map or use of SQUEEZE could be employed to account for this void space.
Ratio Observed / Unique Reflections (too) Low (46%). This implies that the data is weak or non-existant and that only 46% of the collected data are statistically 2 sigma above the background noise of the experiment. This is unacceptable. This is not addressed in either the supplementary information or the CIF. Perhaps recollection for a longer exposure or collection of the data at a synchrotron would yield better data.
R1 value is high. This is not accounted for in the supplementary information or CIF. Collection of better data or use of a correct model should be used in the refinement.
Low completeness of data collection, despite the monoclinic crystallographic symmetry. More reflections should be collected to achieve 98% completeness to 0.8Å when using a MoKα source.
Temperature of cell measurement and data collection reported as 293K. Check. Answer: The XRD data for compound 26 have been re-collected as suggested. The CCDC number has been provided in the SI.
In data for compound 26, a B alert was reported by http://checkcif.iucr.org/, this was due to manually removed crystalized solvent. The solvent molecule could not be assigned clearly, though the crystal was cultivated from ether-methanol-water. This explanation has also been included in the SI.
Question 5: Compound 29: Information regarding the crystal _exptl_crystal_description and exptl_crystal_colour both missing from the CIF. This information should be included。 Flack parameter >0.5. This could be indicative of miss assigned chirality and the structure being inverted.
Large variations in thermal parameters for non-solvent atoms. Should be addressed with appropriate restraints.

Answer:
The crystal _exptl_crystal_description and _exptl_crystal_colour have been added.
As the enantiomer could be assigned by reference to an unchanging chiral center from compound 26 in the synthetic procedure, this XRD data was collected using Mo K\a as diffraction source, which led to big flack parameter. Mo K\a was used instead of Cu K\a for the crystal was not big enough, and the time for collecting data was limited. Method for determining absolute configuration has also been included in the CIF.
The data has been examined again, and the large variations in thermal parameters came from gem-dimethyl lactone of B ring. This could be attributed to flexible conformation of 5-membered-ring, and the terminal methyl magnify this variations.
Question 6: Compound 31: O1 refined to 0.285(14) occupancy. Is this partial occupancy justified? Combined with short O163 -O1 D -A distance, is the atom correctly assigned?
Flack parameter meaningless (su>>value). This should be revised or addressed in the supplementary information or CIF. Method for determining absolute configuration has not been stated in the CIF.
Information regarding the crystal _exptl_crystal_description, _exptl_crystal_colour and dimensions missing from the CIF. This information should be included.
Large thermal parameter discrepancy and C41 large ADP max/min ratio. These should be refined using suitable restraints to generate a realistic model.
Centre of gravity of residues not within the unit cell. Residues should reside mainly within the unit cell. Answer: The No. 1 atom has been carefully considered, and according to the surrounding environments and the number of atoms, this atom could only be assigned to be insufficiently occupied oxygen from water, and the relative hydrogen was not added due to insufficient occupancy. Additionally, the short O163-O1 distance could also be attributed to a hydrogen bond. This explanation has also been included in the SI.
Because of the same reason as compound 29, this data was also collected using Mo K\a. _chemical_absolute_configuration has been included in the CIF.
The crystal _exptl_crystal_description and _exptl_crystal_colour have been added. The data was refined again using suitable restraints as suggested. It did give a better C41 ADP max/min ratio, which might be acceptable. Besides, the center of gravity of residues has also been adjusted within the unit cell.
Question 7: Compound 32: Flack parameter meaningless (su>>value). This should be revised or addressed in the supplimenary information or CIF.
Ratio of maximum/minimum residual density could be an indication of missed disorder/twinning or the inclusion of bad reflections in the refinement.
Information regarding the crystal _exptl_crystal_description and _exptl_crystal_colour both missing from the CIF. This information should be included. Answer: Due to the same reason as compounds 29 and 31, this data was also collected using Mo K\a. _chemical_absolute_configuration has been included in the CIF.
The data has been examined carefully again and we have to admit that there is no disorder or twinning. Instead, the unsatisfying ratio of maximum/minimum residual density was due to the poor, but also the best we can get, crystal quantity. Different from the final natural product, this intermediate is not that stable at room temperature.
The crystal _exptl_crystal_description and _exptl_crystal_colour have been added.