Radical aryl migration enables diversity-oriented synthesis of structurally diverse medium/macro- or bridged-rings

Medium-sized and medium-bridged rings are attractive structural motifs in natural products and therapeutic agents. Due to the unfavourable entropic and/or enthalpic factors with these ring systems, their efficient construction remains a formidable challenge. To address this problem, we herein disclose a radical-based approach for diversity-oriented synthesis of various benzannulated carbon- and heteroatom-containing 8–11(14)-membered ketone libraries. This strategy involves 1,4- or 1,5-aryl migration triggered by radical azidation, trifluoromethylation, phosphonylation, sulfonylation, or perfluoroalkylation of unactivated alkenes followed by intramolecular ring expansion. Demonstration of this method as a highly flexible tool for the construction of 37 synthetically challenging medium-sized and macrocyclic ring scaffolds including bridged rings with diverse functionalities and skeletons is highlighted. Some of these products showed potent inhibitory activity against the cancer cell or derivative of human embryonic kidney line in preliminary biological studies. The mechanism of this novel strategy is investigated by control experiments and DFT calculations.

The authors do not detail the importance of scaffold diversity within the introduction, which if included would highlight the significance of the library generated (i.e. the molecules are diverse through their functional group, appendage, stereochemical and skeletal features).
In this case, although the scaffolds appear diverse, the authors may wish to visually summarise the 'diversity' of the library through chemoinformatic analysis i.e. PMI/PCA plots. The scaffolds which have been generated are highly interesting and valuable for screening collections and the diverse nature of the molecules could be further highlighted.
The biological assessment which is included does not appear to be comprehensive. In addition, this information should appear under a separate sub-heading not added at the end of a non-related paragraph. The authors do not justify why the 'randomly selected compounds' 3Ia, 13 and 15 were tested and why the particular cell lines were chosen.

Figures -
The compound numbering is confusing to the reader and would benefit from some revision. - The 'table of contents' should not have R1 labelled as this is never modified. Additionally, the positioning of R1, R2 and R3 is different in the table of contents to that in Table 1 and 2, Figures 3 and 5. These should be consistent. - Overlapping bonds within Figure 2 need to be corrected.
- Figure 5 is confusing and should be separated into two figures and the text also split accordingly. It is also separated from its figure legend.

Typographical errors
There are numerous typographical errors within the manuscript which need to be addressed. Some examples include: - The introduction paragraph features numerous errors and is confusing to the reader.
-Page 4 paragraph 1 should read 'With this strategy, We report herein the first practical strategy for selective...' and this whole first sentence should be reworded as it is too long. -Methods -Phosphonylation reactions, the experimental should read 'The product was extracted with EtOAc (3 x 10 mL)' and 'organic layers' Reference 35 and 39 quote 'Nature' whereas 2, 3 and 18 this is given as 'Nat.' these should be consistent.

Supplementary information: -
Starting materials 1a-1x appear to be novel, yet no spectra for these molecules is included in the supplementary information. If the molecules are not novel, they should include a reference.
-Typographical error in title of Table S2a 'of reaction' - Table S2c features bold oxygen and iodine atoms -General information: should read 'q, quartet'

Reviewer #2 (Remarks to the Author):
This manuscript describes a synthetic methodology for the generation of libraries of compounds with medium to large-ish rings. Some evidence for the involvement of radical species is included, as well as a DFT study.
I will focus my comments on the computational work as that is where my main expertise lies. As reported, this aspect of the study is not particularly interesting or original, with much more extensive computational mechanistic studies reported in the literature, e.g. the work on Cucatalysed Ullmann coupling by the groups of Houk/Buchwald (10.1021/ja100739h) and later followed up by Fu and co-workers (DOI: 10.1021/ja104264v). Indeed, I appreciate that this was a relatively minor addition to back up the mechanistic postulate, but I was left wondering whether metal involvement had at all been considered, and if not, why not.
In terms of the computational methodology, I was somewhat troubled that the main body of the paper indicated that the M06-L functional was used, while the ESI repeatedly referred to B3LYP-D3, including when the optimised geometries are presented. The reason I started looking at this more closely was that selectivity is explained by a 1 kcal/mol energy difference between potentially competing pathways and I was left wondering how robust this result would be to changes in computational approach. Now, experimentally, the bottom-line obviously is that this works, but the computational work presented did not leave me entirely satisfied that all possibilities had been explored, nor that the balance might not have tipped the other way with some exploration of conformational freedom or a different computational approach. I would not necessarily expect calculations to match up with experimental ee, but a bigger energy difference would make conclusions drawn a little more comfortable. At the very least, I would like to see a more extensive discussion of mechanistic possibilities from the point of view of computation, including perhaps an estimate of computational uncertainty/noise, and stronger links with the experimental results if a metal involvement can indeed be ruled out from that point of view.
As minor comments, I would have liked to see vibrational frequencies for the TS's included in the ESI, some mention of chemical space maps (e.g. by the groups of Beratan and Reymond) could strengthen the case for the synthetic methodology, and a greater focus on the key conclusions one can derive from substrate scope and optimisations, rather than a description of individual results, would certainly enhance this work for me.

Reviewer #3 (Remarks to the Author):
The authors reported here several interesting transformations via novel radical aryl migrations to construct medium-ring motifs. Besides, excellent chirality transfers were observed starting from enantioenriched tertiary alcohols. Potential applications (biological studies) and mechanism elucidation were conducted in this research. The work is well organized and covers a relatively wide range of examples.
I have a few minor comments that may be taken into consideration before publication: 1. For azidation reaction, the author did provide optimization table in SI, however, more information about side products (3Aa' and 3Aa') is missing. It would help the readers to understand the reaction better if the ratio could be provided in some entries, especially cases with low yields. And the authors did not mention why they changed to new conditions in the case of 3Ma.
2. For trifluromethylation reaction, in some instances, the yields were rather poor (30-40%). Is it due to competitive 1,2-oxytrifluromethylation? It would help to understand mechanism, providing 1,2oxytrifluromethylation path was considered in DFT calculation.  Our response: We very much appreciate these comments of the referee and sincerely thank the referee for recommending publication of this work in Nat. Commun., which is a great acknowledgement of the significance of our work.

Comment 2:
The article is currently written in a 'report' style and needs to be formatted correctly to fit the Nature Communications guidelines/style. Multiple paragraphs are excessively long and would benefit from being separated to aid the flow of the article. In addition, the article is extremely long (22 pages and over 6500 words including references) which exceeds the 12-page/ 5000-word limit. In this case it should be shortened considerably e.g. through summarising the substrate scope and experimental results.
Our response: We thank the referee for bringing this issue to our attention. Normally, the manuscript for an article is 12 pages and 5000 word for main text. According to your suggestions, we have shortened the manuscript appropriately and the special editing did not affect any of the conclusions of the published paper. Since we are writing the article in a 'report' style, we will see whether it will meet the criteria when it is formatted into its template. If not, we will consider shortening the manuscript again to fit the guidelines of Nature Communications. Our response: We appreciate the referee for pointing out this issue. We have separated 'results' and 'discussion' into two parts in the new manuscript.
Comment 4: Additionally, the tense changes throughout the article and this is confusing within some paragraphs.
Our response: We thank the referee for informing us these problems and apologize for these mistakes. Following the referee's valuable suggestion, we have reviewed the manuscript carefully and revised the tense properly. Meanwhile, we highlighted the correction in the revised manuscript. Our response: We sincerely thank the referee for this valuable suggestion and have conducted principal component analysis (PCA) to evaluate the diversity of our compound library compared with natural products, drug-like compounds and drugs (see Figs. S4-10 and Tables S6 and S7 in the revised Supplementary Information for details). The results reveal comparablely diverse but distinct chemical spaces occupied by our synthetic molecules and similar benzannulated medium-ring natural products. Compared with the naturally occurring couterpart, our compound library displays less overlaps with the chemical spaces determined by drug-like compounds and drugs. Both of these two features indicate a wide and less-explored chemical space spanned by our compound library, thus demonstrating its great potential for future drug discovery. We have added the correponding results into the revised manuscript and Supplementary Information.    Figures 3 and 5. These should be consistent.

Supplementary
Our response: We sincerely appreciate the referee for bringing the 'compound numbering' issue to our attention. We have reorganized the compound numbering through the whole manuscript and supplementary information. For example, we changed the labeling of azido-substituted products to '3A-3M' from '3Aa-3Ma' and CF 3 -substituted products to '4A-4X' from '4Ab-4Xb'.
We also thank the referee for pointing out the mistake concerning substituents on aryl ring and olefin moiety and apologize for it. Accordingly, we have revised it in our revised manuscript to make sure that all are in consistent. For example, R 1 is connected to olefin moiety and R 3 is removed in the update version such as the following structure 1. Figure 2 need to be corrected.

Comment 8: Overlapping bonds within
Our response: We thank the referee for informing us this mistake and have revised it in the revised manuscript.

S9
Comment 9: Figure 5 is confusing and should be separated into two figures and the text also split accordingly. It is also separated from its figure legend.
Our response: We thank the referee for this valuable suggestion and have separated this figure to Figure 5 and Figure 6 and the corresponding paragraph into two paragraphs.
Comment 10: There are numerous typographical errors within the manuscript which need to be addressed. Some examples include:-The introduction paragraph features numerous errors and is confusing to the reader.
Our response: We apologize for our carelessness and give our appreciation to the referee for such kind guidance to us. Accordingly, we have carefully reviewed the whole manuscript and polished our writing again.
Comment 11: -Page 4 paragraph 1 should read 'With this strategy, We report herein the first practical strategy for selective...' and this whole first sentence should be reworded as it is too long.
Our response: We thank the referee for this good suggestion and have reorganized this sentence as following: "With this strategy, we report herein the first practical strategy for selective and diversity-oriented synthesis of benzannulated 8 to 11(14)-membered cyclic ketones along with concurrent installation of various functional groups from readily available starting materials. This strategy was realized through concerted remote 1,4-or 1,5-aryl migration/ring expansion sequence triggered by radical azidation, trifluoromethylation, phosphonylation, sulfonylation, or perfluoroalkylation of unactivated alkenes." Comment 12: -Page 7 -'alkyenyl alcohols 1 featuring a six-membered ring' and '1M containing an internal alkene' Our response: We thank the referee for suggestion of a proper way to exemplify these sentences and have corrected them in the revised manuscript.

Comment 13: -Page 10 -'switching the benzylic alcohol (1O) to an aliphatic substrate (1V).'
Our response: Thanks to the referee, we have revised this description as "the reaction was not significantly affected by switching the benzylic alcohol (1O) to an aliphatic substrate (1V)." Our response: Thanks to the referee, we have changed the tense in that sentence as "…we expected that the stereochemical information of the tertiary alcohol would be completely transferred to the remote new-formed carbon chiral center in a highly stereoselective way…" Comment 15: -Page 15 -'Consequently, the resultant functionalized...' S11

Referee 2 Comment 1: This manuscript describes a synthetic methodology for the generation of libraries of compounds with medium to large-ish rings. Some evidence for the involvement of radical species is included, as well as a DFT study.
Our response: We appreciate the referee's summary of the findings in this work. Our response: Many thanks for this suggesting consideration of mechanism involving a metal complex. We apologize for not addressing this issue in the original manuscript and this suggestion has been followed by performing the additional DFT calculations on the reaction of the alkyl radical 1Q 1 with Cu II species. In these calculations, the broken symmetry solution of M11 method was used to locate the open-shell-singlet-diradical transition state (TS-SD). The theoretical study reported by Li and coworkers indicated that "when one-electron Cu I -reductant was available, Togni's reagent will be easily reduced via single-electron transfer (SET) to produce CF 3 • free radical and Cu II species, and Then the CF 3 free radical facilely attacks the

C=C bond, leading to trifluoromethyl alkyl radical intermediate. These two processes of CF 3 free radical generation and C-CF 3 bond formation are thermodynamically favorable. In addition, trifluoromethyl alkyl radical intermediate binding with the Cu II complex took place on the singlet-diradical state potential energy surface (PES),
which has lower reaction barrier than that via the triplet/closed-shell-singlet PES (ACS catalysis, 2015(ACS catalysis, , 5, 2458(ACS catalysis, -2468." Based on these mechanistic insights, reaction of the alkyl radical intermediate 1Q 1 with Cu II species was performed exclusively on the singlet-diradical state. As shown in Figure S3 in the revised Supplementary Information, transformation of 1Q 1 to INT requires the activation free energy of 20.9 kcal/mol via transition state TS-SD with the forming Cu-C bond distance being 2.97 Å. In the transition state TS-SD, NBO analysis demonstrates that the spin density was dominantly located on the Cu atom (0.62) and C(-0.92) atom with the calculated spin expectation values (<S 2 >) of 0.99, indicating that a typical singlet-diradical character of TS-SD. Comparison of the reaction barriers between the reaction of the alkyl radical 1Q 1 with Cu II species (20.9 kcal/mol, Figure S3) and radical aryl migration process followed by ring expansion without the Cu II species (14.8 kcal/mol, Figure 4d of our response to comment 4) reveals that the former is energetically unfavorable and thus the mechanism involving metal species of titled reactions could be reasonably ruled out. The above discussion and Figure S3 have been included in the revised manuscript and Supplementary Information. On the other hand, please kindly note that such reactions with some substrates (for the construction of products 4N, 4W, 4X) could be realized in the presence of organic base as the catalyst. Therefore, in some cases, the mechanism involving metal species should be ruled out.
On the other hand, pleased kindly note that Houk/Buchwald (10.1021/ja100739h) and Fu and co-workers (DOI: 10.1021/ja104264v) works have already been cited in the revised manuscript as Ref. 49 and 50. Figure S3. The calculated relative free energies (ΔG sol ) in 1,4-dioxane with SMD model at the M11/6-31+G**/SDD/Aug-cc-PVTZ level are given in kcal/mol. The selected bond lengths are in Å.

Supplementary
Comment 3: In terms of the computational methodology, I was somewhat troubled that the main body of the paper indicated that the M06-L functional was used, while the ESI repeatedly referred to B3LYP-D3, including when the optimised geometries are presented.
Our response: We apologize for the writing errors in the original manuscript. The B3LYP-D3 functional has been replaced by the functional M11 in revised manuscript S13 and Supplementary Information as to that the latter was employed in the updated calculations (see also the response to comment 4 of Referee 2).
Comment 4: "The reason I started looking at this more closely was that selectivity is explained by a 1 kcal/mol energy difference between potentially competing pathways and I was left wondering how robust this result would be to changes in computational approach. Now, experimentally, the bottom-line obviously is that this works, but the computational work presented did not leave me entirely satisfied that all possibilities had been explored, nor that the balance might not have tipped the other way with some exploration of conformational freedom or a different computational approach. I would not necessarily expect calculations to match up with experimental ee, but a bigger energy difference would make conclusions drawn a little more comfortable.
Our response: Many thanks for bringing our attention to the activation free energy difference between two key transition states responsible for the experimental selectivity and for suggesting comparison of density functional-dependent ee values. This concern has been addressed by performing the additional DFT calculations using several other functionals, including M11, M06L, B3LYP, BP86 and B3P86 functionals together with B3LYP-D3 methods previously used in the original manuscript. The results (see below) have been included in the following Table. Among these functionals, M11 functional gave ee values ranging from 83.6% to 90.4% on basis of the calculated energy differences between transition states TS 1Q1 and TS 1Q1' (1.43~1.77 kcal/mol), which were comparable to that found in the experimental one (ee: 96%). In addition, the B3LYP-D3 functional yields an ee value of 87% in the solvent (∆∆G sol-cpcm =1.58 kcal/mol). However, the other functions such as M06L, conventional B3LYP, BP86 and B3P86 provide poor and even incorrect prediction for the experimental results. As such, the predicted ee value of the titled reactions involving radical species significantly depends on the functionals employed in the calculations. Collectively, M11 functional was the most suitable one for mirroring the experimental enantioselectivity both in gas phase and in solvent. Consequently, we recalculated the potential energy surfaces of two reaction pathways responsible for the stereoselective reactions and replaced the original Figure 4d calculated with B3LYP-D3 functionals by the update Figure 4d shown below. Accordingly, the sections about DFT calculations have been carefully organized in the revised manuscript (see Figure 4d and the paragraph marked in yellow on page 14-15 of the revised manuscript) :   Fig. 4d, DFT calculation for mechanistic investigation. The calculated potential energy surfaces for aryl migration and ring expansion processes at M11/BS level of theory in 1,4-dioxane with the full optimization using smd model (BS refers to the used basis sets. For C, H, O, F atoms, the 6-31+G** basis set were used and for Br atom, the Aug-cc-PVTZ basis set was used).
Comment 5: At the very least, I would like to see a more extensive discussion of mechanistic possibilities from the point of view of computation, including perhaps an estimate of computational uncertainty/noise, and stronger links with the experimental results if a metal involvement can indeed be ruled out from that point of view.
Our response: According to the calculated results in the response for comment 2, some more mechanistic possibilities have been discussed in the revised manuscript, as described in the following text. S15 "To probe the origin of the observed stereoselectivity, we further investigated the reaction mechanism computationally using M11 method. The generally assumed alkyl sp 3 -carbon-centered radical species 1Q 1 was chosen as the starting point to locate two reaction pathways responsible for the stereoselective reactions. The calculated results revealed that this reaction occurs stepwise, involving the formation of the bicyclic rings 2Q 1 /2Q 1' and the ring expansion to result in the formation of intermediates 3Q 1 /3Q 1' (Fig. 4d). Some notable points from these calculations are as follows: (i) The addition of sp 3 -carbon-centered radical species to the aryl group (from 1Q 1 to 2Q 1 /2Q 1' ) is exothermic.
In addition, the reversed process (2Q 1 /2Q 1' to 1Q 1 ) has higher reaction barriers than the subsequent ring expansion (2Q 1 /2Q 1' to 3Q 1 /3Q 1' ). As such, reaction barriers of the aryl migration step could be used to mirror the observed stereoselectivity. (ii) The energy of transition state TS 1Q1 , leading to the major (S)-product 3Q 1 , is lower than that of transition state TS 1Q1' forming the minor (R)-product 3Q 1' by 1.4 kcal/mol when (R)-1Q* is employed as the reactant. This difference in energy appears to result from steric repulsion between the trifluoroethyl group and the six-membered ring and agrees well with experiment findings (for 4Q* from (R)-1Q in a highly stereoselective way as shown in Fig. 4b). (iii) As going from 1Q 1 to 3Q 1 , the sequential cascade reaction is exothermic by 2.0 kcal/mol along with the formation of the more stable neutral ketyl radical species. To understand the factors that contribute to the more stability of radical species 3Q 1 than that of 1Q 1 , we performed the NBO analysis of spin density distribution of the above two species (Fig. 4d). The spin density of 1Q 1 is exclusively localized on the carbon atom of substrate, while that of 3Q 1 is partially distributed on the carbon atom with significant radical character on the O atom and the phenyl group. This reveals that the presence of a OH group as a π-donating moiety can significantly stabilize the resultant radical species 3Q 1 , providing an important driving force that makes the formation of ketyl radical species 3Q 1 thermodynamically favorable and increases the power of the overall process substantially to generate medium-sized ring systems. DFT calculations on an alternative mechanism involving a Cu species have also been performed, showing that the reaction of the alkyl radical 1Q with Cu II species is highly energetically unfavorable, as revealed by a 6.1 kcal higher reaction barrier of the forming Cu intermediate than the radical aryl migration process of alkyl radical intermediate 1Q 1 (Fig. S3 in Supplementary Information). Therefore, all these experimental and calculated results, together with the observed highly efficient radical chirality transfer during the ring expansion, are in support of our initial proposal as shown in Fig. 2c, in which the selective addition of a variety of in situ generated radicals to unactivated alkenes could trigger ring expansion via remote 1,4-or 1,5-aryl migration processes."

Comment 6: As minor comments, I would have liked to see vibrational frequencies for the TS's included in the ESI.
Our response: We apologize for not including the imaginary vibrational frequencies of the calculated transition states, which have been added in the revised Supplementary Information. Comment 7: some mention of chemical space maps (e.g. by the groups of Beratan and Reymond) could strengthen the case for the synthetic methodology, and a greater focus on the key conclusions one can derive from substrate scope and optimisations, rather than a description of individual results, would certainly enhance this work for me.

S16
Our response: We sincerely thank the referee for this valuable suggestion and have conducted principal component analysis (PCA) to evaluate the diversity of our compound library compared with natural products, drug-like compounds and drugs (see Figs. S4-10 and Tables S6 and S7 in the revised Supplementary Information for  details). The results reveal comparablely diverse but distinct chemical spaces occupied by our synthetic molecules and similar benzannulated medium-ring natural products. Compared with the naturally occurring couterpart, our compound library displays less overlaps with the chemical spaces determined by drug-like compounds and drugs. Both of these two features indicate a wide and less-explored chemical space spanned by our compound library, thus demonstrating its great potential for future drug discovery. We have added the correponding results into the revised manuscript and Supplementary Information. See also the response to Comment 5 of Referee 1.

Referee 3
Comment 1: The authors reported here several interesting transformations via novel radical aryl migrations to construct medium-ring motifs. Besides, excellent chirality transfers were observed starting from enantioenriched tertiary alcohols. Potential applications (biological studies) and mechanism elucidation were conducted in this research. The work is well organized and covers a relatively wide range of examples. Overall, I think these are very remarkable results and I recommend to publish in Nature Communications subject to the above issues being addressed. Our response: We very much appreciate the comments of the referee and sincerely thank the referee for recommending publication of this work in Nature Communications.
Comment 1: For azidation reaction, the author did provide optimization table in SI, however, more information about side products (3Aa' and 3Aa') is missing. It would help the readers to understand the reaction better if the ratio could be provided in some entries, especially cases with low yields. And the authors did not mention why they changed to new conditions in the case of 3Ma. Our response: We thank the referee for bringing this issue to our attention. In designing this project, we assumed that both 3A' and 3A'' might be generated. During screening the reaction conditions, we almost observed no formation of 3A' and 3A''. The low yield of 3A is due to the decomposition of 1A. However, when we optimized the reaction conditions for synthesis of 4N, we observed a 65% yield of oxytrifluoromethylation product 4N' in the presence of Cu(I) catalyst. When we changed to base as the catalyst, the desired ring expansion product 4N can be detected as a single isomer. These results have been discussed in the revised manuscript.
With regard to preparation of 3M, the reaction is quite messy and the desired product was observed in quite low yield in the presence of CuCN for the 1,2-disubstituted alkene substrate 1M. Our screening of additive revealed that 1,10-phenanthroline (1.5 equiv) could enhance the yield of 3M, so we utilized this reaction condition. These results have been added in the revised manuscript.
It should be noted that in the revised manuscript, we have reorganized the compound numbering according to the comment of Referee 1. For example, we changed the labeling of azido-substituted products to '3A-3M' from '3Aa-3Ma' and CF 3 -substituted products to '4A-4X' from '4Ab-4Xb'.
Comment 2: For trifluoromethylation reaction, in some instances, the yields were rather poor (30-40%). Is it due to competitive 1,2-oxytrifluromethylation? It would help to understand mechanism, providing 1,2-oxytrifluromethylation path was considered in DFT calculation. Our response: We greatly thank the referee for providing such a question. For trifluoromethylation reaction, we have screened a variety of parameters to inhibit the 1,2-oxytrifluoromethylaiton reaction. For those whose yield is slightly low, almost no 1,2-oxytrifluoromethylation product was formed and the substrate was decomposed. However, in our expansion of substrate scope, we found that the reaction efficiency depends largely on the structure of substrates. Based on these results, at the end of that paragraph we gave a possible conclusion that tuning the opening-ring size can have a profound influence on the control of the reaction outcome, probably owing to the presence of different favorable conformations and transition states during the reaction (page 9 in the revised manuscript). For the DFT calculations, we have performed the additional DFT calculations on the process of the reaction of the alkyl radical 1Q with Cu II species, which could give 1,2-oxytrifluromethylation product from this pathway. As shown in Figure S3 in Supplementary Information, transformation of 1Q 1 to INT requires the activation free energy of 20.9 kcal/mol via transition state TS-SD with the forming Cu-C bond distance being 2.97 Å. Comparison of the reaction barriers between the reaction of the alkyl radical 1Q with Cu II species (20.9 kcal/mol, Figure S3) and radical aryl migration process followed by ring expansion without the Cu II species (14.8 kcal/mol, Figure 4d in the revised manuscript) reveals that the latter is energetically favorable and the mechanism involving metal species of titled reactions could be reasonably ruled out with a higher reaction barrier.
See also the response to Comments of Referee 2. Our response: We thank the referee for bringing this issue to our mind. We apologize for this mistake and have revised this part in the revised manuscript as follows: (R)-4A: 98% ee, HPLC analysis [Daicel Chiralpak AD-H, isopropanol/hexane = 10/90, 1.0 mL/min, λ= 214 nm, t R (minior) = 6.6 min, t R (major) = 9.2 min].

Comment 4: Did authors ever try substrate with unactivated alkene disubstituted at the internal position?
Our response: We thank the referee for this valuable question. We have synthesized internal disubstituted alkene 1X. Unfortunately, the trifluoromethylation reaction provided the abnormal product 4X in only moderate yield. The probable mechanism was shown below: