Theoretical evaluation of the carbene-based site-selectivity in gold(III)-catalyzed annulations of alkynes with anthranils

Abstract

The gold(III)-catalyzed annulations of alkynes with anthranils were evaluated using DFT calculations. A unified rationale for the Br-migration on α-imino gold(III)-carbene was proposed, from which an unprecedented “N-donation/abstraction substitution” mechanism was established using the substituted anthranils, while direct C–H nucleophilic attack was involved with the unsubstituted anthranils. The controlling factors guiding the site-selectivity were uncovered. These computational studies provide insight for developing new α-imino gold(III)-carbene mediated reactions.

---

Gold(III)-catalyzed annulations of alkynes¹ with isoxazole derivatives have attracted significant interest since they provide atom- and step-economical access to diverse aza-heterocycles found in numerous natural products, organic functional materials and pharmaceuticals.² Considerable effort has been devoted to developing gold carbene-promoted reactions³ for the selective synthesis of azaheterocyclic scaffolds. In this context, three fascinating annulations have recently been developed by the group of Hashmi to construct diverse azacycles through the tuning of nucleophilic sites for gold(III) carbenoid intermediates (Scheme 1)^4,5 Using KAuBr₄ as a catalyst, the reaction of substituted-anthranil 2M with N-benzyl ynamide 1a is proposed to undergo C^a-nucleophilic attack to generate the quinoline-fused polyazaheterocycle P_a (Scheme 1(1)),⁴ while with N-furanylmethyl ynamide 1b undergoes exclusive C^b-attack, producing 2-aminopyrrole P_b (Scheme 1(2)).⁴ Interestingly, the annulation of unsubstituted-anthranil 2N with ynamide 1t under AuBr₃ catalysis, results in the formation of the unprotected 7-acylindole P_t, where C^t attack is favoured (Scheme 1(3)).⁵

Scheme 1 Gold(III)-catalyzed annulations of N-aryl ynamides with anthranils reported by Hashmi's group.^4,5

The divergence of the nucleophilic site-selectivity at the gold carbene³ is very interesting but means that access to desired azaheterocycles in a facile and general pattern is challenging. Hence, we decided that it was important to provide an in-depth understanding of the reaction mechanisms and, especially, the inherent origins of the site-selectivity, in order to help generalize these reactions and make them predictable. Mechanistically, according to the general proposal,⁶ the reactions in Scheme 1 are assumed to proceed via a direct nucleophilic attack of the C^a, C^b or C^t atom at the carbene C atom. By employing an exhaustive DFT evaluation (see the Computational details in the ESI†), we explored whether the direct attack pathway or other alternative mechanisms could be used to rationalize the observed annulation products. Furthermore, the inherent controlling factors affecting the site-selectivity of these reactions were thoroughly investigated. We anticipate that these results will be informative for the future design of related catalytic reactions.

For the 1a system, the DFT-computed free energy profiles are given in Fig. 1. The reaction is initiated by coordination of KAuBr₄ with the C⋯C bond of 1a to give Au⋯π coordinated complex IM1. Then the 2M N atom, acting as a nucleophile, attacks the alkyne internal C atom of IM1 due to the greater electron-deficiency of the internal C over the terminal C, as indicated by the calculated NBO charges (0.385e for the former and −0.443e for the latter). Through TS2-3, the necessary Au(III) carbenoid IM3 can then be formed by cleaving the N–O bond^7,8 with a barrier of 10.9 kcal mol⁻¹. From IM3, C^a nucleophilic attack at carbene C will exclusively produce product P_a. Therefore, we firstly evaluated the direct attack mechanism.⁹ Unexpectedly, the desired C^a-attack denoted as TS3-9 is 3.7 kcal mol⁻¹ less stable than the C^t-attack viaTS3-4. Therefore, we explored a new mechanism, initiated by one Br(Au) migration. As shown in Fig. 1(black line), one Br(Au) transfers to the carbene C atom to give IM5. The calculations indicate that, through TS3-5, Br-migration occurs with a barrier of only 0.2 kcal mol⁻¹ and is exergonic by 10.8 kcal mol⁻¹. Subsequently, an unprecedented “SN2-type N-donation/abstraction substitution” achieves a formal C^b-attack, in which the sp²–N atom substitutes with the migrated-Br viaTS5-7 (N-donation substitution) and is then replaced by a C^b atom viaTS7-8 (N-abstraction substitution), affording the five-membered cyclized intermediate IM8. After ring expansion through the C(Au) migration to the C^a atom, the target C^a-attack complex IM9 is obtained with an energy demand of 14.6 kcal mol⁻¹ relative to IM8. From the formal C^a-attack pathway (black line), one can clearly see that three energetically comparable TSs leading to P_a, TS3-5, TS5-7, and TS7-8, are lower in energy than TS3-4 (red line) and TS3-9 (blue line). Two crucial factors might be responsible for such energy differences. One is the extremely flat Br-migration due to much smaller structural deformation with slight change of the Br–Au–C(carbene) angle (see Fig. S1 in the ESI†). And the other may be related to the high stability of IM5 after Br-migration, which remarkably reduces the potential energy surface of subsequent processes, making the formation of P_a easier. From IM3, other possible reaction scenarios are disfavoured energetically and are given in Fig. S2 of the ESI.†

Fig. 1 Calculated Gibbs energy profile in DCE for the formation of quinoline-fused polyazaheterocycle P_a established in the present work. The relative free energies are given in kcal mol⁻¹.

Following IM9, the C^a-attached H atom transfers to the sp²–N atom resulting in the formation of P_a. The calculated results indicate that the (C^a)H firstly transfers to one adjacent (S)O atom of the Ts fragment and then is trapped by the sp²–N atom with the assistance of KBr. The H-shift step viaTS9-10 surmounts an energy demand of 6.5 kcal mol⁻¹ and affords the extremely exergonic intermediate IM10 owning to the formation of the aromatic benzene ring.¹⁰IM10 then experiences a KBr Br-assisted H-abstraction/donation, smoothly giving rise to the target product P_a after dehydration. Along the reaction coordinate, the C^b-substitution by the N atom, corresponding to the IM7 → IM8 transformation, is considered as the rate-determining step of the whole reaction and has a barrier of 23.1 kcal mol⁻¹.

In the case of the 1b annulation, the proposed mechanism of the 1a-system, Au(III)-coordination → anthranil N nucleophilic addition → N–O cleavage → Br-migration → N-donation/abstraction substitution, is applied to rationalize the C^b-attack to the C(Au) atom for formation of key intermediate IM18 (1b Cycle, Fig. 2). Then, the reaction is predisposed to chemoselectively cleave the C^b–O bond of the furan moiety viaTS18-24 to provide IM26. The alternative C(Au)-migration (TS18-21, Fig. S3, ESI†) was also examined and found to be highly unfavourable (−34.7 kcal mol⁻¹ for TS18-21 and −44.9 kcal mol⁻¹ for TS18-24). With the C^b–O cleavage, the Br-assisted stepwise H-shift finally produces P_b and regenerates the KAuBr₄ active catalyst. Finally, the 1t reaction mechanism was considered (1t Cycle, Fig. 2). After Au(III)-coordination, N nucleophilic addition, N–O cleavage and Br-migration,¹¹ C^t-substitution with Br(Au) is selectively favoured, from which the C^t-attached H-shift and subsequent dehydration, results in P_t formation. The calculated results for other competitive pathways involved in 1b- and 1t-systems are illustrated in Fig. S4 and S5 in the ESI.†

Fig. 2 Plausible catalytic cycles for the Au(III)-catalyzed selective annulations of N-aryl ynamides with anthranils based on the current calculations.

We then turned our attention to uncover how the nucleophilic site-selectivity at C^a, C^b and C^t atoms are controlled in these reactions. The relative free energies for the transition states (TSs) of key Br-migration, N-donation substitution and N-abstraction substitution leading to C^a/C^b selectivity, as well as that of C^t-attack resulting in C^t-selectivity are summarized in Fig. 3(i). Note that, in 1a-/1b- and 1t-systems, the N-donation substitution TSs (TS5-7/TS16-17 and TS3′-5′ in Fig. 1 and Fig. S4, S5, ESI†) exhibit very similar characteristics for the reaction regions with selected (sp²)N⋯C(Au) distances of 1.934 Å, C(Au)⋯Br distance of ∼3.000 Å and (sp²)N–C(N)–C(Au) angle of 89.6°. Intriguingly, compared with the N-donation substitution TSs, the competitive C^t-attack TSs exhibit almost opposite stability changes (0.2 vs. 3.5 in the 1a-system, −2.9 vs. 7.4 in the 1b-system and 4.9 vs. −2.7 kcal mol⁻¹ in the 1t-system, respectively).

Fig. 3 (i) The relative free energies (in kcal mol⁻¹) for the crucial site-selective transition states, as well as (ii) optimized geometries with selected structural parameters for C^t substitution transition states in 1a-/1b- and 1t-systems. The bond distance and bond angle are given in Å and °, respectively.

Thus, it is predicted that the site-selectivity is closely related to the geometric constrains of the C^t-attack TS. The distinctive degree of the trans-influence of the Br atom is mainly responsible for the bifurcated selectivity at C^a/C^b and C^t atoms. It can be seen in Fig. 3 (ii) that a weaker trans-influence is exhibited in the 1t-system with a Br–C(carbene)–C^t bond angle of 151.6° (166.6° in the 1a-system and 166.4° in the 1b-system), which, in turn, makes the interaction of the C^t with the C(carbene) atom easier. As a result, smaller energy demand is required for the 1t-system. In contrast, because of the greater Br trans-influence, the resultant larger energy demand forces the 1a- and 1b-systems to follow the N-donation substitution and thus C^a/C^b selectivity is favoured.

After favourable N-donation substitution followed by N-abstraction substitution in the 1a-/1b-systems, C^a- and C^b-selectivity was further evaluated. Two selective-determining TSs for the ring-expansion with C^a-selectivity and ring-opening with C^b-selectivity, TS8-9vs.TS8-9′ in the 1a-system and TS18-21vs.TS18-24 in the 1b-system, were compared and given in Fig. 4. For the 1b-system, the main driving force for the favoured ring-opening might originate from the greater electrostatic O⋯C² interaction.¹² As shown by the calculated NBO charges, the charges on O and C² atoms in TS18-24 are −0.45 and 0.39 e, whereas those of C(Au) and C^a in TS18-21 are −0.42 and −0.18 e, respectively. Clearly, the stronger electrostatic attraction involved in TS18-24 is more advantageous to cleave the C^b–O bond and thereby generates a relatively lower energy consumption. Nevertheless, significantly different from five-membered ring-opening in the 1b-system, TS8-9′ in the 1a-system is characteristic of a six-membered ring-opening, in which the C^b–C⁴ bond rupture will generate a substantially unstable terminal C cation, thereby resulting in a greater energy requirement. Consequently, the reaction is forced to follow a C^a-selective ring-expansion viaTS8-9.

Fig. 4 The schematic structures with NBO charges (in e) for competitive C^a and C^b selective-determining transition states in 1a- and 1b-systems.

In summary, the mechanisms and origins of multiple nucleophilic site-selectivity of Au(III)-catalyzed annulations of substituted-anthranil 2M with N-benzyl ynamide 1a and N-furanylmethyl ynamide 1b, as well as that of unsubstituted-anthranil 2N and N-benzyl ynamide 1t, have been elucidated using DFT calculations. A unified and rational mechanism is presented in Fig. 2. After Au(III) coordination → 2M/2N N nucleophilic addition → N–O cleavage to afford gold carbene intermediate C, further Br(Au)-migration affords a common bromide intermediate D. Then, the three competitive reaction sites, C^a, C^b and C^t, selectively attack the Au-attached C atom, resulting in the distinct products. For 1a-/1b-annulations, a common key issue concerning a formal C^b-attack is established and involves an unprecedented “SN2-type N-donation/abstraction substitution” process. Subsequently, the reaction undertakes either C^a-migration → H-shift → Br-assisted H-shift → dehydration to form the C^a-attack product P_a in the 1a-system or C–O cleavage → Br-assisted H-shift to furnish the C^b-attack product P_b in the 1b-system. For P_t formation in the 1t system, the bromide intermediate, undergoes selective C^t-substitution with Br(Au), C^t-attached H-shift and dehydration.

For these annulation reactions, the present study indicates that the strength or weakness of the Br(Au) trans-influence dominates the nucleophilic selectivity for the C^a/C^b or C^t atoms. Furthermore, the C^a site-selectivity in the 1a-system can be attributed to the rather unstable terminal C cation in the ring-opening TS leading to C^b-attack, while the strong C⋯O electrostatic attraction in the crucial ring-expansion TS is responsible for the preferred C^b-selectivity in the 1b-system. In summary, we anticipate that the controlling factors involved in the nucleophilic site-selectivity observed for these reactions can be applied to other transition metal carbene-promoted reactions.

This work was jointly supported by the National Natural Science Foundation of China (21873055), the Natural Science Foundation of Shandong Province (ZR2019MB016 and ZR201709240033), Key Laboratory of Emergency and Trauma (Hainan Medical University), Ministry of Education (KLET-201903). TDJ wishes to thank the Royal Society for a Wolfson Research Merit Award and the Open Research Fund of the School of Chemistry and Chemical Engineering, Henan Normal University for support (2020ZD01).

Conflicts of interest

There are no conflicts to declare.

Notes and references

1. (a) D. J. Gorin and F. D. Toste, Nature, 2007, 446, 395; (b) R. Dorel and A. M. Echavarren, Chem. Rev., 2015, 115, 9028; (c) A. Furstner, Chem. Soc. Rev., 2009, 38, 3208; (d) L. Fensterbank and M. Malacria, Acc. Chem. Res., 2014, 47, 953; (e) L.-W. Ye, X.-Q. Zhu, R. L. Sahani, Y. Xu, P.-C. Qian and R.-S. Liu, Chem. Rev., 2020 DOI:10.1021/acs.chemrev.0c00348; (f) R. L. Sahani, L.-W. Ye and R.-S. Liu, Adv. Organomet. Chem., 2020, 73, 195; (g) E. Aguilar and J. Santamaría, Org. Chem. Front., 2019, 6, 1513.
2. (a) S. Hahn, S. Koser, M. Hodecker, P. Seete, F. Rominger, O. Miljanic, A. Dreuw and U. H. F. Bunz, Chem. – Eur. J., 2018, 24, 6968; (b) L. Ji, A. Friedrich, I. Krummenacher, A. Eichhorn, H. Braunschweig, M. Moos, S. Hahn, F. Geyer, O. Tverskoy, J. Han, C. Lambert, A. Dreuw, T. Marder and U. H. F. Bunz, J. Am. Chem. Soc., 2017, 139, 15968; (c) Z. Zeng, H. Jin, X. Song, Q. Wang, M. Rudolph, F. Rominger and A. S. K. Hashmi, Chem. Commun., 2017, 53, 4304.
3. (a) M.-H. Tsai, C. Wang, A. S. K. Raj and R.-S. Liu, Chem. Commun., 2018, 54, 10866; (b) R. L. Sahani and R. Liu, Angew. Chem., Int. Ed., 2017, 56, 1026; (c) R. L. Sahani and R.-S. Liu, Angew. Chem., Int. Ed., 2017, 56, 12736; (d) G. Ung, M. Soleilhavoup and G. Bertrand, Angew. Chem., Int. Ed., 2013, 52, 758.
4. Z. Zeng, H. Jin, M. Rudolph, F. Rominger and A. S. K. Hashmi, Angew. Chem., Int. Ed., 2018, 57, 16549.
5. H. Jin, L. Huang, J. Xie, M. Rudolph, F. Rominger and A. S. K. Hashmi, Angew. Chem., Int. Ed., 2016, 55, 794.
6. (a) A.-H. Zhou, Q. He, C. Shu, Y.-F. Yu, S. Liu, T. Zhao, W. Zhang, X. Lu and L.-W. Ye, Chem. Sci., 2015, 6, 1265; (b) L. Li, T.-D. Tan, Y.-Q. Zhang, X. Liu and L.-W. Ye, Org. Biomol. Chem., 2017, 15, 8483; (c) H. Li, J. Liu, A. A. Ogunlana and X. Bao, Org. Chem. Front., 2017, 4, 1130; (d) S. S. Giri and R.-S. Liu, Chem. Sci., 2018, 9, 2991.
7. A. A. Ogunlana and X. Bao, Chem. Commun., 2019, 55, 11127.
8. (a) K. Wang, Y. Liu, Q. Wu, L. Liu, Y. Li, T. D. James, G. Chen and S. Bi, Mol. Catal., 2020, 480, 110647; (b) Y. Liu, X. Yang, L. Liu, H. Wang and S. Bi, Dalton Trans., 2015, 44, 5354; (c) Y. Liu, K. Wang, B. Ling, G. Chen, Y. Li, L. Liu and S. Bi, Catal. Sci. Technol., 2020, 10, 4219.
9. (a) T. Li, W. Rong, T. Zhang and J. Li, ChemCatChem, 2020, 12, 5276; (b) C. Zhu, L. Kou and X. Bao, Chin. J. Chem., 2020, 38, 57; (c) A. K. Sahoo, R. Vanjari, S. Dutta, B. Prabagar and V. Gandon, Chem. – Asian J., 2019, 14, 4828.
10. (a) M. J. Dewar, Angew. Chem., Int. Ed. Engl., 1971, 10, 761; (b) M. Randić, Chem. Rev., 2003, 103, 3449; (c) Z. W. Davis-Gilbert, X. Wen, J. D. Goodpaster and I. A. Tonks, J. Am. Chem. Soc., 2018, 140, 7267.
11. It is confirmed by the IRC calculations that, in the 1t-system, the N–O bond breaks with simultaneous Br(Au)-migration, implying that the energy surface of the Br-migration is very flat. This fact can be indirectly corroborated by the corresponding step in the 1a- and 1b-systems, where the activation barriers are only 0.2 (Fig. 1) and 0.7 kcal mol⁻¹ (Fig. S4, ESI†), respectively.
12. In addition, to examine the contribution of the O⋯C² electrostatic attraction to O–C^b cleavage, the furan O atom is replaced by a C atom (Fig. S6, ESI†). It was found that the ring-opening transition state is 2.8 kcal mol⁻¹ higher in energy than the ring-expansion transition state, indicating that a strong electrostatic attraction between the O and C² is a factor in the C^b-selectivity.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0cc07440c

---