Differentiation between enamines and tautomerizable imines in the oxidation reaction with TEMPO

Enamine and imine represent two of the most common reaction intermediates in syntheses, and the imine intermediates containing α-hydrogen often exhibit the similar reactivity to enamines due to their rapid tautomerization to enamine tautomers. Herein, we report that the minor structural difference between the enamine and the enamine tautomer derived from imine tautomerization results in the different chemo- and regioselectivity in the reaction of cyclohexanones, amines and TEMPO: the reaction of primary amines furnishes the formal oxygen 1,2-migration product, α-amino-enones, while the reaction of secondary amines under similar conditions generates exclusively arylamines via consecutive dehydrogenation on the cyclohexyl rings. The 18O-labeling experiment for α-amino-enone formation revealed that TEMPO served as oxygen transfer reagent. Experimental and computational studies of reaction mechanisms revealed that the difference in chemo- and regioselectivity could be ascribed to the flexible imine-enamine tautomerization of the imine intermediate containing an α-hydrogen.


1.
The authors are not especially rigorous in defining the reaction stoichiometry. For at least one example from each of the two reaction classes, a rigorous quantification of all reaction (by)products, including the fate of TEMPO species, should be provided. The mechanisms in Figure 2 would seem to implicate a need for more equivalents of TEMPO than are used in the optimized reaction conditions. I may have missed it, but an assessment of the reaction outcome as a function of TEMPO equivalents should be included for each of the two reaction classes..

2.
The details provided for DFT calculations are woefully inadequate in the Supporting Information. Whereas experimental details are well documented, a compilation of xyz coordinates for all computed structures should be provided (this content should be standard requirement for all reputable journals). In addition, the authors do little more than provide a computed pathway for their favored mechanism, without any discussion of alternate pathways. This type of presentation does not inspire confidence...

a.
Computed barriers are not compared to the experimental barriers.
b. Free energies for HAT from the N-H of the enamine tautomer are shown in Fig. 2d); how do these compare to the energies for HAT from the C-H of the imine tautomer. c.
How does the energy and energy barrier of HAT from the beta C-H of the primary aminederived enamine compare to those of the secondary amine-derived enamine. d.
TS2A in Fig 2c show a cluster of three water molecules and it show the H-transfer to the oxygen, rather than the nitrogen. None of these features is justified and compared to relevant alternative pathways.
In short, I don't consider the computational presentation suitable for publication. Although much of the requested information may be presented in the SI to avoid distracting in the main article, it needs to be included. The switch in selectivity is intriguing and compelling, and these empirical observations provide the main justification for publication.

Response:
We thank the reviewer for his(her) positive comment on our findings.
Reviewer #1 Comment 2. I find the DFT-based mechanistic analysis to be only marginally compelling, mainly providing specificity for the authors intuitive view of the mechanism (i.e., not really providing a rigorous mechanistic assessment of reaction pathway), Response: More comprehensive investigation has been carried out per the reviewer's advice.
(1) Kinetics instead of thermodynamics of hydrogen abstraction from all possible sites is now provided to directly evaluate the relative reactivity on different sites. Reviewer #1 Comment 3. and the synthetic impact of this work is marginal (i.e., I don't expect other groups will use this chemistry in synthesis). In the end, I support publication Nature Communications, but not with strong conviction and will defer to the editor to assess the merits in light of other reviews.
The other reaction presented by us is the reaction of cyclohexanones with secondary amine and TEMPO to produce aryl amines via consecutive dehydrogenation of enamine intermediate. Aryl amines are ubiquitous building blocks for various organic molecules, electronic materials and pharmaceutical agents. Although aryl amines are generally synthesized using metal-catalyzed cross-coupling of aryl halides with amines, such as Buchwald-Hartwig coupling and Ullmann-type aminations, our method opens up a new avenue to aryl amine synthesis using readily available cyclohexanones as starting materials , and thus provides a concise alternative to the existing methods for syntheses of aryl amines. Importantly, our method would have an impact on retrosynthetic disconnections of the targeted products since starting materials in our method are different from the ones in Buchwald-Hartwig coupling and Ullmann-type aminations.
The broad generality of these two reactions and the importance of their products in organic synthesis should therefore leads to a non-negligible impact on the synthetic chemistry. We have thoroughly discussed the synthetic merits of our two reactions in the revised manuscript.
We sincerely hope the reviewer concurs after reading this clarification.
Reviewer #1 Comment 4. The authors are not especially rigorous in defining the reaction stoichiometry. For at least one example from each of the two reaction classes, a rigorous quantification of all reaction (by)products, including the fate of TEMPO species, should be provided. The mechanisms in Figure 2 would seem to implicate a need for more equivalents of TEMPO than are used in the optimized reaction conditions. I may have missed it, but an assessment of the reaction outcome as a function of TEMPO equivalents should be included for each of the two reaction classes.
Response: During screening the reaction parameters for these two reaction classes, we established the reaction stoichiometry including yields of main product and byproducts, and the fate of TEMPO. In response to this suggestion, we have added these data to the results of optimization studies on two model reactions (please see the parts A and B of "Reaction condition Screening " in Supplementary materials) .
By GC-MS analysis, we observed that TEMPO was converted into the corresponding amine, 2,2,6,6-tetramethylpiperidine (TEMPH), in both classes of reactions. The calculated mechanism for α-amino enone formation reaction (Figure 2d) indicated that generation of one equivalent of α-amino enone consumed two equivalents of TEMPO, in which TEMPO was converted to one equivalent of the corresponding hydroxylamine (TEMPOH) and one equivalent of the corresponding amine (TEMPH). Thus, the TEMPO equivalents implicated in the mechanism appear to be more than 1.5 equivalents of TEMPO used in the optimized conditions. In the calculated mechanism for aryl amine formation (Figure 2c), 4 equivalents of TEMPO were required to abstract H-atom and generate TEMPOH, which appear to be more than 2.5 equivalents used in the optimized conditions. However, in fact, TEMPOH generated from TEMPO is unstable and spontaneously disproportionates to a 2:1 mixture of TEMPO and TEMPH (Org. Biomol. Chem. 2003, 1, 3232). TEMPO from the disproportionation of TEMPOH would re-participate in reaction, which accounts for the equivalents of TEMPO used in optimized conditions, and the observed conversion of TEMPO to TEMPH in both classes of reactions. To avoid the confusion, we have added the discussion about the re-generation of TEMPO through the disproportionation of TEMPOH in the mechanism part.
We have determined the reaction outcomes with varying TEMPO equivalents, of which results have been added to the revised Supplementary Materials.

Reviewer #1 Comment 5. Computed barriers are not compared to the experimental barriers.
Response: We agree that making a direct comparison between the computational and experimental barriers can evaluate the reliability of computed mechanism. From the reaction of TEMPO with cyclohexanone and primary amine, we isolated the α-aminoxylated ketone (please see eq. 3, Scheme 3), which is in agreement with the calculation results that α-site acts as the reactive site for imine molecule B, demonstrating the reliability of computed mechanism.
Reviewer #1 Comment 6. Free energies for HAT from the N-H of the enamine tautomer are shown in Fig. 2d); how do these compare to the energies for HAT from the C-H of the imine tautomer.
Response: The Fig. 2a and b have been revised to provide the barriers for hydrogen abstraction with TEMPO from all probable reactive sites, instead of the thermodynamics only. As shown in Fig. 2b, the barriers for hydrogen abstraction from all the C-H sites of the imine tautomer are unfeasible because of the high barriers.

Reviewer #1 Comment 7. How does the energy and energy barrier of HAT from the beta C-H
of the primary amine-derived enamine compare to those of the secondary amine-derived enamine.
Response: As shown in the revised Fig. 2a and b, hydrogen abstraction from the beta C-H of the primary amine-derived enamine LM1B experiences a barrier of 35.9 kcal/mol, showing a lower activity as compared to hydrogen abstraction from the beta C-H of secondary aminederived enamine A (33.9 kcal/mol).

Reviewer #1 Comment 8. TS2A in Fig 2c show a cluster of three water molecules and it show
the H-transfer to the oxygen, rather than the nitrogen. None of these features is justified and compared to relevant alternative pathways.
Response: Actually, both oxygen and nitrogen sites of TEMPO can act as the reactive site for H B elimination to generate of arylamine species. As shown in Figure S8a (the revised SI), the pathway for H B transfer to nitrogen site has been located, showing a slightly lower feasibility as compared to H B transfer to oxygen site.
Reviewer #1 Comment 9. In short, I don't consider the computational presentation suitable for publication. Although much of the requested information may be presented in the SI to avoid distracting in the main article, it needs to be included.  (Org. Lett. 2013, 15, 1646-1649. Within the newly formed adduct, hydrogen bond enables easy proton-transfer and therefore accelerates addition of primary amine to the C-N double bond of the imine intermediate to produce the targeted imine and release 4-methylanthranilic acid. In contrast, due to steric hindrance, relatively bulky secondary amine is difficult to react with the imine intermediate from the condensation of ketone and 4-methylanthranilic acid. As a result, 4-methylanthranilic acid impedes the condensation of ketone with secondary amine.

Response
We've changed the main text accordingly to include this justification.
Reviewer #1 Comment 11. In line 108, the authors use the successful reactivity with unsubstituted cyclohexanone to conclude that steric hindrance is not a factor in driving the formation of α-amino-enones. However, this statement is not prefaced by any hypothesis or precedent indicating that sterics should play a major role. In this context, 2-substituted cyclohexanones are not demonstrated in any of the synthetic example, nor is there any comment regard the effect of substituents in this position.
Response: In our previous work (J. Am. Chem. Soc. 2016, 138, 5623.), we established that Cu-catalyzed β-amination of acyclic ketone with a variety of amines via formation of α,βunsaturated ketone, which are sensitive to steric hindrance. In our model reaction, 4-tert-butyl cyclohexanone was used a substrate to lead to discovery of α-amino-enone formation reaction. To emphasize that the α-amino-enone formation reaction stems from the reactivity of tautomerizable imine rather than steric factor, we gave an example of unsubstituted cyclohexanone in the investigation on substrate scope.
Indeed, 2-substetuted cyclohexanones did not participate in this reaction likely due to steric hindrance. In the revised manuscript, we have described this limitation.
Reviewer #1 Comment 12. Specific products discussed in the α-amino-enone scope (starting on line 103) should be clarified with their compound numbers in the appropriate products.

Response:
We have used compound numbers to describe the products in question for clarity.

Reviewer #1 Comment 13.
There is no optimization table in the text for the dehydrogenative aromatization reaction, and there is no indication in the text that there is an optimization table located in the SI (line 132).
Response: Nature Communications limits the number of Figures and Schemes to no more than six. We therefore put the optimization table for dehydrogenative aromatization in the SI.
For clarity, we have indicated this in the main text.
Reviewer #1 Comment 14. Schemes 1 and 2 do not contain any detailed footnote information on how the yields were obtained (GC yields or isolated NMR yields) or indicate different conditions. For example, the reaction to make product 5u in Scheme 2 used 6 eq. of TEMPO, but this is not clarified in the graphic.