Azide trapping of metallocarbenes: generation of reactive C-acylimines and domino trapping with nucleophiles

Abstract

Azide-tethered diazocarbonyl compounds undergo copper-catalyzed conversion to transient C-acylimines. These reactive intermediates can be trapped with a variety of carbon nucleophiles, giving rise to complex 3-indolinone frameworks, including those with adjacent tetra-substituted carbon centers, in a single transformation.

---

Introduction

The generation and rearrangements of ammonium ylides offer an effective route to complex nitrogen-containing targets.¹ Particularly in the case of nitrogen heterocycles, an attractive approach involves intramolecular addition of a transient metallocarbene to a pendent tertiary amine. This process directly affords a cyclic ammonium ylide 1 which can undergo a subsequent rearrangement process (Scheme 1). Most typically this involves migration of a nitrogen substituent to the neighboring ylide carbon to give 2.² Less nucleophilic amide nitrogens can also participate, though complications from competing formation of carbonyl ylides can arise.³

Organic azides can intercept a variety of electrophilic species in synthetically useful fashion,⁴ such as aziridination of electron-deficient alkenes⁵ or Schmidt rearrangement.⁶ We speculated that replacement of basic amine with azide in the reaction sequence above might permit cyclization to betaine 3, which was expected to undergo loss of dinitrogen to form cyclic C-acylimine 4.^7–11 If this reaction were carried out in the presence of a suitable nucleophile, 4 could be trapped to provide adduct 5 through a domino process. With two molecules of dinitrogen as the only by-products, this transformation would also have significant potential as a green process. Herein, we report the results of our preliminary exploration of this reaction, demonstrating effective generation and trapping of the acylimine with β-dicarbonyl compounds, silyl ketene acetals, an electron-rich 1,3-diene, and N-methylindole.

Scheme 1 Metallocarbene trapping by nitrogen nucleophiles.

Results and discussion

To test the viability of this hypothetical process, we required a substrate which held the diazoketone and azide moieties in close proximity, and with a molecular formula whose ratio of (carbon + oxygen) atoms to nitrogen atoms permitted safe handling.¹² Anthranilic acid-derived substrate 6a could easily be prepared in two steps from readily available starting materials, and was amenable to handling and purification on scales of <1 g (Scheme 2).¹³ Treatment of this compound with 10 mol% Cu(acac)₂ in toluene at room temperature with no added nucleophile led to rapid consumption of 6a and the formation of multiple highly coloured products. Purification of this mixture yielded small quantities of a bright red solid, which was identified as alkylideneindolone 7a, apparently formed by nucleophilic trapping of the intermediate 4a by the acetylacetonate ligand from the catalyst, followed by autoxidation of the adduct. Importantly, 3H-indole-3-one 4a was not isolated in this or any subsequent experiment.¹⁴ The ready formation of 7a in the presence of catalytic amounts of acetylacetonate nucleophile indicates the high reactivity of this intermediate.¹⁵ With this in mind, we carried out the same reaction, this time in the presence of 1 equiv. of silyl ketene acetal 8a. In this case, known alkylideneindolone 7b ¹⁶ was obtained in good yield, along with traces of 7a. Notably, no evidence was seen for competing reaction of 8a with the intermediate metallocarbene, indicating that intramolecular capture by azide to generate 4a is kinetically favored.

Scheme 2 Preparation and preliminary reactions of 6a.

Several other catalysts were screened, and of these copper(II) bis(hexafluoroacetylacetonate) (Cu(hfacac)₂) was found to give optimal results, affording product in good yield with no competing trapping by the ligand (Table 1, entry 1). A variety of traps (2 equiv.) were then examined under these conditions. To avoid potential catalyst deactivation, we focused on carbon nucleophiles as opposed to heteroatom nucleophiles that might bind tightly to the copper complex. Efficient trapping was seen with tetra-substituted silyl ketene acetal 8b, affording indolone 7c in excellent yield (entry 2). A dehydrogenated product analogous to 7a,b could not be formed in this case due to the exocyclic quaternary centre; however, slow oxidation to a different product was observed (Scheme 3). Upon prolonged exposure to air, 7c produced 2-hydroxyindolone 10. Further oxidation was observed when 10 was allowed to stand in CDCl₃, presumably the result of elimination due to traces of acid.¹⁷

Table 1 Cyclization and trapping of diazo azide 6a^a

Entry	Substrate	Nu	Product	Yield^b (%)
a Standard procedure: a solution of 6a in PhMe (0.04 M) was added dropwise over 1 h by syringe pump to a solution of Cu(hfacac)2 (10 mol %) and the trap (2 equiv.) in PhMe (0.04 M) at rt.b All yields given are for isolated product after chromatographic purification.c Small quantities of 7a were obtained, but yield was not determined.
1	6a			96
2	6a			93
3	6a			78
4	6a		7a	—^c
5	6a			83
6	6a			53
7	6a			75

---

Scheme 3 Oxidation of 7c.

Deliberate inclusion of Na(acac) 8c afforded adduct 7a in good yield; acetyl acetone itself also provided this product, but at a disappointingly slow rate (entries 3 and 4). Alternatively, diethyl bromomalonate 8e afforded 7d in good yield without the need for added base (entry 5); in this case, the exocyclic alkylidene group is presumably formed via elimination rather than autoxidation. N-Methylindole 8f gave known indolylindolone 7e ¹⁸ (entry 6), and use of the Danishefsky diene 8g led to tricyclic 4-pyridone 7f in good yield (entry 7), presumably via stepwise Mannich/Michael process with 4a,¹⁹ followed by autoxidation. Trapping was also attempted using triethylsilane and phenyl boronic acid pinacol ester; however, these reactions did not yield any discernable adducts.

A second substrate 6b bearing an additional carboxy stabilizing group was prepared via diazotransfer reaction with the keto ester 9, which was readily accessible from the corresponding benzoic acid via the Ti-crossed-Claisen protocol reported by Tanabe and coworkers²⁰ (Scheme 4). This compound could also be subjected to the domino azide coupling/nucleophilic trapping process, though higher temperatures were required to consume the doubly stabilized diazo starting material (Table 2). Thus, treatment with Cu(hfacac)₂ in toluene at reflux in the presence of silyl ketene acetals 8a,b furnished adducts 7g,h (entries 1 and 2). Diminished yields of 7h can be attributed to the steric demand encountered during the formation of two contiguous quaternary centres. Given the lack of hydrogens at C-2 of the indolinone ring, no autoxidation of 7g,h was observed.

Other traps were also effective (entries 3 and 4), including N-methylindole 8f and the trimethylsilyl enol ether of acetophenone (8h). The corresponding allyl ester 6c ²¹ also furnished adducts, albeit in slightly diminished yields (entries 5 and 6). Interference with the intermediate metallocarbene by the pendent allyl group may contribute to yield erosion, though we were unable to detect any cyclopropane-containing impurities. The effects of ring substitution were also evaluated with substrates 6d,e ²¹ (entries 7–10). An electron-withdrawing chloro substituent was well tolerated, affording adducts 7m–o, as was a methyl group adjacent to the azide (7p). Notably, compatibility with halo substituents suggests that further elaboration of the indolinone products via cross-coupling processes should be possible.

Scheme 4 Preparation of distabilized substrate 6b.

Table 2 Cyclization and trapping of doubly stabilized diazo azides^a

Entry	Substrate	Nu	Product	Yield^b (%)
a Standard procedure: a solution of 6 in PhMe (0.04 M) was added dropwise over 1 h by syringe pump to a solution of Cu(hfacac)2 (10 mol %) and the trap (2 equiv.) in PhMe (0.04 M) at reflux.b All yields given are for isolated product after chromatographic purification.c The OTBS silyl ketene acetal was used in place of the OTMS version.
1	6b	8a		89
2	6b	8b		48
3	6b	8f		72
4	6b			71
5	6c	8f		63
6	6c	8h		52
7	6d	8a^c		73
8	6d	8f		76
9	6d	8h		76
10	6e	8h		68

---

An interesting observation was made when 6a was treated with Cu(hfacac)₂ in the presence of acid chloride 12a (Scheme 5). In this case, tetracyclic indoloquinazoline 13 was formed, albeit in variable yields. Compound 13 is the alkaloid natural product tryptanthrin,²² whose derivatives possesses a number of promising biological activities.²³ This one-step synthesis is presumed to occur through sequential addition of the azido group of 12a to imine 4a (or its copper complex),²⁴ followed by N-acylation and elimination of dinitrogen. Consistent yields could be obtained if aniline 12b was used in place of 12a. In this case, an oxidation step must occur following assembly of the tetracyclic scaffold. It is notable that all trapping examples in Tables 1 and 2 involved carbon π-nucleophiles, while 12a,b trapped via nitrogen. The scope of heteronucleophilic traps merits further study.

Scheme 5 Synthesis of tryptanthrin.

Conclusions

Domino azide-metallocarbene coupling/nucleophilic addition has been achieved, forming substituted indolone systems by sequential formation of adjacent C–N and C–C bonds. With unsubstituted diazoketone precursors, rapid autoxidation occurs after nucleophilic trapping. A variety of nucleophiles can be used, including active methylenes, silyl ketene acetals, Danishefsky's diene, or N-methylindole. Use of 2-azido- or 2-aminobenzoyl chloride allows for 1-step construction of the natural product tryptanthrin. Doubly stabilized diazoketones also undergo efficient cyclization and nucleophilic capture, generating a variety of ester-substituted indolinones. Variation of ring substitution had no observable effect on the efficiency of the process. Further studies of this process will be reported in due course.

Acknowledgements

We thank NSERC for support of this work, and the University of Alberta for a Queen Elizabeth II Graduate Scholarship (TMB).

Notes and references

1. J. A. Vanecko, H. Wan and F. G. West, Tetrahedron, 2006, 62, 1043–1062.
2. F. G. West and J. S. Clark, in Nitrogen, Oxygen and Sulfur Ylide Chemistry, ed. S. J. Clark, Oxford University Press, Oxford, 2002, pp. 115–134.
3. A. Padwa, J. P. Snyder, E. A. Curtis, S. M. Sheehan, K. J. Worsencroft and C. O. Kappe, J. Am. Chem. Soc., 2000, 122, 8155–8167.
4. (a) A. Rostami, Y. Wang and F. G. West, Org. Lett., 2007, 9, 703–706; (b) Review: S. Bräse, C. Gil, K. Kneper and V. Zimmermann, Angew. Chem., Int. Ed., 2005, 44, 5188–5240.
5. K. B. Hong, M. G. Donahue and J. N. Johnston, J. Am. Chem. Soc., 2008, 130, 2323–2328.
6. S. Grecian and J. Aube, in Organic Azides: Syntheses and Applications, ed. S. Bräse and K. Banert, Wiley, Chichester, 2010, pp. 191–237.
7. For two important examples of intramolecular addition of azides to metallocarbenes, see: (a) A. G. H. Wee and J. Slobodian, J. Org. Chem., 1996, 61, 2897–2900; (b) A. Blond, R. Moumné, G. Bégis, M. Pasco, T. Lecourt and L. Micouin, Tetrahedron Lett., 2011, 52, 3201–3203.
8. For a recent report of the addition of organic azides to gold carbenoids, see: H. Liu, X. Li, Z. Chen and W.-X. Hu, J. Org. Chem., 2012, 77, 5184–5190.
9. Doyle and coworkers have recently described the intermolecular reaction of rhodum metallocarbenes with aryl azides to form imines that can be trapped in subsequent cycloadditions processes: M. D. Mandler, P. M. Truong, P. Y. Zavalij and M. P. Doyle, Org. Lett., 2014, 16, 740–743.
10. The isoelectronic coupling of a metallocarbene with a diazo compound has been reported in both intra- and intermolecular forms: (a) M. P. Doyle, W. Hu, I. M. Phillips and A. G. H. Wee, Org. Lett., 2000, 2, 1777–1779; (b) J. H. Hansen, B. T. Parr, P. Pelphrey, Q. Jin, J. Autschbach and H. M. L. Davies, Angew. Chem., Int. Ed., 2011, 50, 2544–2548.
11. Notably, Driver has recently shown that o-styryl azides undergo rhodium catalyzed indole formation via apparent nitrene intermediates: C. Jones, Q. Nguyen and T. G. Driver, Angew. Chem., Int. Ed., 2014, 53, 785–788.
12. Both azides and diazo groups are potentially explosive, and it is recommended that the (C + O)/N ratio be ≥3 to ensure against unanticipated detonation of azide-containing compounds: T. Keicher and S. Löbbecke, in Organic Azides: Syntheses and Applications, ed. S. Bräse and K. Banert, Wiley, Chichester, 2010, pp. 3–27.
13. Because the (C + O)/N ratio for 6a is below 3, we limited the scale of its preparation to no more than 1 g, and individual reactions using 6 were generally run at a scale of 50–100 mg. When stored under argon at –20 °C in a foil-wrapped container, 6a was found to be stable for at least 2 weeks. Compounds 6b–e appear to lack any special sensitivity, and are amenable to use in gram-scale reactions.
14. (a) Isolation of compound 4a from a mushroom pigment: S. Takekuma, H. Takekuma, Y. Matsubara, K. Inaba and Z. Yoshida, J. Am. Chem. Soc., 1994, 116, 8849–8850; (b) Formation of 4a from microwave-assisted cyclization of an O-phenyloxime: F. Portela-Cubillo, J. S. Scott and J. C. Walton, Chem. Commun., 2007, 4041–4043.
15. While substituted 3H-indol-3-ones are isolable, the unsubstituted example 4a was reported to undergo rapid decomposition, and is a postulated transient intermediate in indigo synthesis: (a) W. M. Abdou, M. A. I. Salem and A. A. Sediek, Bull. Chem. Soc. Jpn., 2002, 75, 2481–2485; (b) L. Li, M. Han, M. Xiao and Z. Xie, Synlett, 2011, 1727–1730; (c) O. Neunhoeffer and G. Lehmann, Chem. Ber., 1961, 94, 2965–2967; (d) J. Gosteli, Helv. Chim. Acta, 1977, 60, 1980–1983.
16. K. Okuma, N. Matsunaga, N. Nagahora, K. Shioji and Y. Yokomuri, Chem. Commun., 2011, 5822–5824.
17. (a) L. Greci, G. Tommasi, P. Bruni, P. Sgarabotto and L. Righi, Eur. J. Org. Chem., 2001, 3147–3153; (b) J. Boyer, V. Bernardes-Genisson, V. Farines, J.-P. Souchard and F. Nepveu, Free Radical Res., 2004, 38, 459–471.
18. P. Astolfi, M. Panagiotaki, C. Rizzoli and L. Greci, Org. Biomol. Chem., 2006, 4, 3282–3290.
19. (a) Y. Yuan, X. Li and K. Ding, Org. Lett., 2002, 4, 3309–3311; (b) P. J. Alaimo, R. O'Brien III, A. W. Johnson, S. R. Slauson, J. M. O'Brien, E. L. Tyson, A.-L. Marshall, C. E. Ottinger, J. G. Chacon, L. Wallace, C. Y. Paulino and S. Connell, Org. Lett., 2008, 10, 5111–5114; (c) D. Lanari, O. Piermatti, F. Pizzo and L. Vaccaro, Synthesis, 2012, 44, 2181–2184.
20. T. Misaki, R. Nagase, K. Matsumoto and Y. Tanabe, J. Am. Chem. Soc., 2005, 127, 2854–2855.
21. Substrates 6c–e were prepared in analogy to 6b. See ESI for details.†.
22. Reviews: (a) A. M. Tucker and P. Grundt, ARKIVOC, 2012, 546–569; (b) Y. Jahng, Arch. Pharmacal Res., 2013, 36, 517–535.
23. (a) L. A. Mitscher and W. Baker, Med. Res. Rev., 1998, 18, 363–374; (b) Y. Takei, T. Kunikata, M. Aga, S. Inoue, S.-I. Ushio, K. Iwaki, M. Ikeda and M. Kurimoto, Biol. Pharm. Bull., 2003, 26, 365–367; (c) M.-C. Recio, M. Cerda-Nicolas, O. Potterat, M. Hamburger and J.-L. Rios, Planta Med., 2006, 72, 539–546; (d) S. Yang, X. Li, F. Hu, Y. Li, Y. Yang, J. Yan, C. Kuang and Q. Yang, J. Med. Chem., 2013, 56, 8321–8331.
24. S. Zhu, J. Dong, S. Fu, H. Jiang and W. Zeng, Org. Lett., 2011, 13, 4914–4917.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental procedures and spectral data for substrates 6a–e and adducts 7a, c–p, 10, 11, and 13. See DOI: 10.1039/c4ra06044j

---