Rhodium(I)-catalysed “ene-type” cycloisomerization of N-[2-(2-alkyn-1-yl)phenyl]carbodiimides leading to 3-(cis-alken-1-yl)-2-aminoquinolines

Abstract

Cycloisomerization of N-[2-(2-alkyn-1-yl)phenyl]carbodiimides 1 catalysed by Rh(dppp)₂Cl under heating conditions afforded 3-(cis-alken-1-yl)-2-(substituted amino)quinolines in up to 87% yield with high cis-selectivity. Scope, limitations, a proposed mechanism for these reactions, and a one-pot synthesis of their trans-isomers from 1 with iodine-promoted olefin isomerization are described.

---

Transition metal-catalysed ene-type cycloisomerization of α,ω-enynes is a facile and highly atom-economical method to access carbo- and heterocycles containing 1,4-diene moieties.^1,2 This ene-type reaction is also applicable to allene-ynes for the construction of a cross-conjugated triene system (eqn (1)),^3–7 which participates in various Diels–Alder (DA) reactions,^5a–c,6 including diene-transmissive DA reactions.^5b,c,8 Malacria and co-workers reported the first example of cycloisomerization of allene-ynes by using a stoichiometric amount of [CpCo(CO)₂].⁴ Brummond and co-workers succeeded in a catalytic version of this reaction for the first time by using cationic rhodium(I) and iridium(I) catalysts.^5d Around the same time, Shibata and co-workers also succeeded with a rhodium-catalysed ene-type cycloisomerization of allene-ynes and proposed their reaction mechanism.⁶ Since then, several research groups have developed their own reactions leading to various carbo- and heterocycles with a cross-conjugated triene system.⁷ However, as far as we are aware, the use of heterocumulene-ynes instead of allene-ynes has not yet been explored, except in our own research.

As part of our continuing programme on the synthesis of nitrogen heterocycles using functionalized heterocumulenes,^9–12 we have been investigating rhodium-catalysed cycloaddition reactions of carbodiimide-ynes.¹³ For example, we reported rhodium(I)-catalysed Pauson–Khand (PK)-type reactions of N-alkynyl-, N-(o-alkynylphenyl)-, and N-[2-(2-alkyn-1-yl)phenyl]carbodiimides or N-(3-propyn-1-yl)carbodiimides leading to pyrrole-, indole-, or quinoline-fused pyrrolin-2-ones, respectively (Schemes 1a and b).^10,11 We also achieved full intramolecular [2 + 2 + 2]-cycloaddition of N,N′-bis-[2-(2-alkyn-1-yl)phenyl]carbodiimides that deliver structurally unique penta- to heptacyclic L-shaped compounds (Scheme 1c).¹² Schemes 1b and c suggest that the possible intermediates, nitrogen-containing rhodacycles, are attractive synthetic intermediates to access variously functionalized quinoline ring systems. To exploit their diverse reactivity, we investigated Rh(I)-catalysed “ene-type” cycloisomerization of 1. Herein, we report a unique type of cycloisomerization to access 3-(cis-alken-1-yl)-2-aminoquinolines 2 with high control of olefin geometry (Scheme 1d).

Scheme 1 Our previous contributions and present work on heterocycle synthesis using carbodiimide-ynes via rhodium-catalysed cycloaddition and cycloisomerization reactions.

To examine the feasibility of the hetero-ene-type cycloisomerization reactions, we selected N-isopropyl-N′-(2-(octyn-2-yl)phenyl)carbodiimide (1a) as the test substrate, with 1a being heated in the presence of several rhodium catalysts (Table 1). A phosphine-free rhodium catalyst, [Rh(cod)Cl]₂, was ineffective, and 1a was recovered with 95% yield after heating in xylene at 130 °C for 2 h (entry 1). Next, we examined Wilkinson's catalyst (Rh(PPh₃)₃Cl), which exhibited high catalytic activities in the [2 + 2 + 2]-cycloaddition of carbodiimide-diynes¹² (Scheme 1c) and in the cycloisomerization reaction of allene-ynes;⁶ however, the reaction of 1a provided only complex mixtures (entry 2). Then, we explored a bidentate phosphine ligand, 1,3-bis(diphenylphosphino)propane (dppp), because it showed high catalytic activity in the previously reported PK reactions of carbodiimide-ynes (Schemes 1a and b).^10,11 When [Rh(cod)Cl]₂ and dppp were mixed, we were delighted to obtain the cycloisomerization product, 3-(cis-1-hexenyl)-2-(isopropylamino)quinoline (cis-2a), albeit with a low yield of 25% (entry 3). While switching the rhodium complex from [Rh(cod)Cl]₂ to [Rh(cod)OH]₂ slightly improved the yield (entry 4 vs. 3), the use of Rh(dppp)₂Cl significantly improved the yield of 2a (entry 5, 64% yield). The reaction in di-n-butyl ether (entry 6), at a lower temperature of 110 °C (entry 7), or with 6 mol% of the catalyst (entry 8) resulted in lower yields of 2a than under conditions in entry 5. Interestingly, cis-2a was formed exclusively over the trans-isomer in entries 3–8. The structure of the phosphine ligand is critical for this reaction; the use of a one-carbon longer or shorter bidentate ligand, such as 1,2-bis(diphenylphosphino)ethane (dppe) and 1,4-bis(diphenylphosphino)butane (dppb), failed to form 2a (entries 9 and 10). The use of a cationic rhodium complex, Rh(dppp)₂BF₄, completely suppressed the formation of 2a (entry 11).

Table 1 Screening of reaction conditions

Entry	Catalyst (mol%)	Solvent	Temp (°C)	Time (h)	Yield (%)
a cod: 1,5-cyclooctadiene.b No reaction.c Not detected.d dppp: 1,3-bis(diphenylphosphino)propane.e Based on NMR.f dppe: 1,2-bis-(diphenylphosphino)ethane.g dppb: 1,4-bis(diphenylphosphino)butane.
1	[Rh(cod)Cl]2 (5)^a	Xylene	130	2.0	N.R.^b
2	Rh(PPh3)3Cl (10)	Xylene	130	2.0	N.D.^c
3	[Rh(cod)Cl]2 (5) + dppp (22)^d	Xylene	130	1.0	25^e
4	[Rh(cod)OH]2 (5) + dppp (20)	Xylene	130	1.0	33
5	Rh(dppp)2Cl (12)	Xylene	130	1.0	64
6	Rh(dppp)2Cl (12)	nBu2O	130	1.0	55^e
7	Rh(dppp)2Cl (12)	Toluene	110	1.5	55
8	Rh(dppp)2Cl (6)	Xylene	130	1.0	40^e
9	Rh(dppe)2Cl (10)^f	Xylene	130	1.0	N.D.
10	Rh(dppb)2Cl (10)^g	Xylene	130	1.0	N.D.
11	Rh(dppp)2BF4 (10)	Xylene	130	1.0	N.D.

---

With the optimized reaction conditions (Table 1, entry 5) in hand, we explored the scope of the cycloisomerization reaction of carbodiimide-ynes 1 (Table 2). Prominent features observed are as follows. (a) The cis-alkenyl quinolines were formed nearly exclusively in all entries. (b) With regard to the substituent at the alkyne terminus (CH₂R¹), the benzyl group (R¹ = Ph) is the most suitable for this reaction (entries 16–20) compared with n-pentyl (R¹ = nBu, entries 1–5), methyl (R¹ = H, entries 6–10), and isobutyl (R¹ = iPr, entries 11–15) groups. (c) With regard to the substituent of the carbodiimide terminus (R²), a relatively larger substituent such as isopropyl (Table 1, entry 5; Table 2, entries 8, 13, and 18) or cyclohexyl (entries 3, 9, 14, and 19) resulted in a better yield; however, the t-butyl group seems to be too bulky for the reaction (entry 4).

Table 2 Scope of substituents of 1

Entry	1	R^1	R^2	Time (h)	Yield (%)
a In toluene at 110 °C.
1	1b	nBu	nPr	1	25
2	1c	nBu	Bn	0.5	30
3	1d	nBu	cHex	1	48
4	1e	nBu	tBu	1	N.R.
5	1f	nBu	Ph	1	34
6	1g	H	nPr	1	25
7	1h	H	Bn	1.5	9
8^a	1i	H	iPr	1	48
9^a	1j	H	cHex	2	54
10^a	1k	H	Ph	1	18
11	1l	iPr	nPr	1.5	16
12	1m	iPr	Bn	0.5	55
13	1n	iPr	iPr	0.5	55
14	1o	iPr	cHex	1	43
15	1p	iPr	Ph	0.5	67
16	1q	Ph	nPr	1	67
17	1r	Ph	Bn	1	49
18	1s	Ph	iPr	0.8	77
19	1t	Ph	cHex	1	73
20	1u	Ph	Ph	0.1	87

---

To clarify the effects of substituents adjacent to the alkynyl carbon, the cycloisomerization reactions of 3 and 5 were examined. The reaction of 3 bearing a methyl group at the benzylic position, proceeded to form 4-methylquinoline derivative 4, albeit with a lower yield than that of 4-unsubstituted quinoline 2i (eqn (2) vs. Table 2, entry 8). Meanwhile, the cycloisomerization reactions of 5, in which a methyl group was introduced to the other adjacent carbon to the C≡C bond, were completely suppressed, and the formation of 6 was not detected (eqn (3)). Accordingly, it is suggested that the presence of a trans-substituent with regard to the quinolyl group in the product is unfavourable for this reaction, and causes selective formation of cis-2 (vide infra).

We further evaluated the reaction of naphthyl analogues 7 as substrates. The reactions of 7 proceeded under the above optimized conditions to produce 2-(substituted amino)-4-(cis-hexen-1-yl)benzo[g]quinolines 8 with acceptable yields (54–59%) and high cis-selectivity (Scheme 2).

Scheme 2 Rhodium-catalysed cycloisomerization of 7.

To probe a preliminary mechanistic consideration, we performed N-[2-(1,1-dideuterio-2-alkyn-1-yl)phenyl]-N′-isopropylcarbodiimide (1a-d₂) under the above optimized reaction conditions (eqn (4)). The reaction produced 2-isopropylamino-3-(cis-1-deuterio-1-hexenyl)-4-deuterioquinoline (cis-2a-d₂) with high D-content, indicating that one of the benzylic deuterium atoms in 1a-d₂ migrates to the 1-alkenyl position of cis-2a-d₂ via formal internal [1,3] migration during the cyclization.

Based on the above observations, a possible mechanistic pathway for the formation of 2 is shown in Fig. 1a. (i) Co-ordination of the rhodium complex to the alkyne bond and the external C=N bond to form A, from which (ii) oxidative cyclization takes place to construct rhodacycle B. (iii) Formal [1,3]-migration of the benzylic proton produces intermediate C, concurrent with gaining pyridine aromatic stabilization energy, from which (iv) β-hydride migration to the rhodium leads stereoselectively to cis-alkenyl quinoline intermediate cis-D. (v) Finally, reductive elimination of the rhodium species from cis-D results in formation of cis-2 and regeneration of the catalyst. Notably, in the present ene-type reaction, the alkynyl function (C≡CCH₂R¹) serves as an “ene” unit and the external heterocumulenic C=N bond serves as an “enophile” unit. This is quite a contrast to the ene-type reaction of allene-ynes, in which the alkynyl group function (C≡C) and the allenyl function (C=CCH₂R) serve as an “enophile” and an “ene” unit, respectively (eqn (1)).

Fig. 1 (a) Assumed reaction mechanism. (b) Assumed TSs for β-hydride migration to form cis- and trans-alkenyl groups.

As a possible rationale for the observed cis-selectivity, transition states (TSs) of β-hydride migration in C to cis- and trans-alkenyl quinoline complex D are depicted in Fig. 1b. In the E-TS, the substituent on the alkyne terminus (R¹) may constitute a barrier to the phenyl group on the dppp, hampering the syn-hydride migration. Meanwhile, in the Z-TS, R¹ faces the opposite side of the ligand and hence the hydride would more smoothly migrate with syn-elimination leading to cis-D.

Finally, we found that molecular iodine¹⁴ effectively promotes cis- to trans-isomerization of 2. For example, stirring cis-2s and iodine (3 equiv.) in dichloromethane at room temperature for 24 h afforded trans-2s with 94% yield. In addition, trans-2 was obtained using a one-pot reaction from 1 (Scheme 3), and the yields compared favourably with those of the ene-type reaction (Scheme 3 vs. Table 2), implying high yield of the olefin-isomerization step. The progress of isomerization confirms that the trans-isomer is more thermodynamically stable than the corresponding cis-isomer, and hence the present cycloisomerization is the kinetically controlled reaction.

Scheme 3 One-pot synthesis of trans-alkenylquinolines 2.

In conclusion, rhodium(I)-catalysed cycloisomerization of N-(propargylphenyl)carbodiimides 1 results in formation of 3-(cis-alken-1-yl)-2-(substituted amino)quinolines 2 with high cis-selectivity. This reaction constitutes the first example of a metal-catalysed “ene-type” cycloisomerization of heterocumulene-ynes, featuring the alkynyl function–“ene” unit and one heterocumulenic C=N bond–“enophile” unit. Facile access to 2-(trans-alkenyl)-3-aminoquinoline derivatives from 1 was also developed.

Notes and references

1. For reviews on metal catalysed cycloisomerizations, see: (a) C. Aubert, O. Buisine and M. Malacria, Chem. Rev., 2002, 102, 813–834; (b) C. Obradors and A. M. Echavarren, Acc. Chem. Res., 2013, 47, 902–912; (c) D.-H. Zhang, Z. Zhang and M. Shi, Chem. Commun., 2012, 48, 10271–10279.
2. For a review on rhodium(I)-catalysed cycloisomerization, see: (a) M. Fujiwara and I. Ojima, in Modern Rhodium-Catalyzed Organic Reactions, ed. P. A. Evans, Wiley-VCH, Weinheim, 2005, p. 129; (b) For a pioneering work on rhodium(I)-catalyzed ene-type reaction of ene-ynes, see: P. Cao, B. Wang and X. Zhang, J. Am. Chem. Soc., 2000, 122, 6490–6491.
3. For a review: C. Aubert, L. Fensterbank, P. Garcia, M. Malacria and A. Simonneau, Chem. Rev., 2011, 111, 1954–1993.
4. D. Llerena, C. Aubert and M. Malacria, Tetrahedron Lett., 1996, 37, 7027–7030.
5. (a) K. M. Brummond, T. O. Painter, D. A. Probst and B. Mitasev, Org. Lett., 2007, 9, 347–349; (b) K. M. Brummond and L. You, Tetrahedron, 2005, 61, 6180–6185; (c) B. Mitasev, B. Yan and K. M. Brummond, Heterocycles, 2006, 70, 367–388; (d) K. M. Brummond, H. Chen, P. Sill and L. You, J. Am. Chem. Soc., 2002, 124, 15186–15187; (e) S. Mao, D. Probst, S. Werner, J. Chen, X. Xie and K. M. Brummond, J. Comb. Chem., 2008, 10, 235–246.
6. T. Shibata, Y. Takesue, S. Kadowaki and K. Takagi, Synlett, 2003, 268–270.
7. (a) C. Mukai, F. Inagaki, T. Yoshida, K. Yoshitani, Y. Hara and S. Kitagaki, J. Org. Chem., 2005, 70, 7159–7171; (b) C. Mukai, F. Inagaki, T. Yoshida and S. Kitagaki, Tetrahedron Lett., 2004, 45, 4117–4121; (c) X. Jiang and S. Ma, J. Am. Chem. Soc., 2007, 129, 11600–11607.
8. For a review on diene-transmissive DA reaction, see: H. Hopf and M. S. Sherburn, Angew. Chem., Int. Ed., 2012, 51, 2298–2338.
9. T. Saito, Y. Sonoki, T. Otani and N. Kutsumura, Org. Biomol. Chem., 2014, 12, 8398–8407.
10. (a) T. Saito, K. Sugizaki, T. Otani and T. Suyama, Org. Lett., 2007, 9, 1239–1241; (b) For cobalt-catalyzed PK reactions, see: C. Mukai, T. Yoshida, M. Sorimachi and A. Odani, Org. Lett., 2006, 8, 83–86.
11. T. Saito, N. Furukawa and T. Otani, Org. Biomol. Chem., 2010, 8, 1126–1132.
12. (a) T. Otani, T. Saito, R. Sakamoto, H. Osada, A. Hirahara, N. Furukawa, N. Kutsumura, T. Matsuo and K. Tamao, Chem. Commun., 2013, 49, 6206–6208; (b) K. Tateno, R. Ogawa, R. Sakamoto, M. Tsuchiya, T. Otani and T. Saito, Org. Lett., 2014, 16, 3212–3215.
13. For rhodium-catalyzed asymmetric [2 + 2 + 2] cycloaddition of alkenyl carbodiimides with alkynes, see: (a) R. T. Yu and T. Rovis, J. Am. Chem. Soc., 2008, 130, 3262–3263; (b) For rhodium-catalyzed [2 + 2 + 2] cycloaddition of carbodiimides with diynes, see: M. Ishii, F. Mori and K. Tanaka, Chem.–Eur. J., 2014, 20, 2169–2174.
14. R. G. Dickinson and H. Lotzkar, J. Am. Chem. Soc., 1937, 59, 472–475.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra11846d

---