Gold-catalyzed [4+3]- and [4+2]-annulations of 3-en-1-ynamides with isoxazoles via novel 6π-electrocyclizations of 3-azahepta trienyl cations.

New gold-catalyzed [4+3]-annulations of 3-en-1-ynamides with isoxazoles afford 4H-azepines efficiently; this process involves 6π electrocyclizations of gold-stabilized 3-azaheptatrienyl cations. In the presence of Zn(OTf)2, the resulting 4H-azepines undergo skeletal rearrangement to furnish substituted pyridine derivatives. We subsequently develop new catalytic [4+2]-annulations between the same 3-en-1-ynamides and isoxazoles to deliver substituted pyridine products using Au(i)/Zn(ii) catalysts. This work reports the first success of the 6π electrocyclizations of heptatrienyl cations that are unprecedented in literature reports.

Introduction

Electrocyclizations of acyclic conjugated π-motifs are powerful tools to access five-, six- and seven-membered carbocycles;1 prominent examples include Nazarov cyclizations of pentadienyl cations2 and 6π electrocyclizations of trienes,3 which have found widespread applications in organic synthesis.

In the context of seven-carbon π-motifs, heptatrienyl anions I undergo facile 8π electrocyclizations via rapid interconversions among various anion configurations (Scheme 1).4 In contrast, heptatrienyl cations III5 exclusively undergo Nazarov reactions because of the difficulties of forming all σ-cis configured cations V that have a high energy state.5 1-Aza- and 1-oxaheptatrienyl cations6 were also reported to follow Nazarov cyclizations. The realization of a 6π electrocyclization of conjugated seven-membered cations is formidable but challenging. This work reveals the first success of such seven-membered cyclizations of gold-stabilized 3-azaheptatrienyl cations V′ to form azacyclic products 3–4via a new C–C bond formation.

Scheme 1

Electrocyclizations of conjugated π-motifs.

The advent of gold catalysis has inspired new annulations between alkynes and poor nucleophiles.7 N–O containing nucleophiles serve as useful building blocks to construct valuable azacyclic frameworks.7 Ye and Hashmi reported interesting [3+2]-annulations of isoxazoles or benzisoxazoles with electron-rich ynamides, yielding substituted pyrrole derivatives through aza-Nazarov cyclizations of the key intermediate [eqn (1)].7,8 These [3+2]-annulations were extensively expanded to other N–O heterocycles including benzisoxazoles, 1,2,4-oxadiazoles, 1,4,2-dioxazoles and 4,5-dihydro-1,2,4-oxadiazoles, yielding additional five-membered azacycles as depicted in [eqn (1)].9 Here, we report two distinct [4+3]- and [4+2]-annulations between 3-en-1-ynamides and isoxazoles using varied catalysts. An Au(i) catalyst alone delivers 4H-azepines 3–4 through 6π electrocyclizations of intermediates V′ [eqn (2)] whereas a combined action of Au(i)/Zn(ii) on the same reactants furnishes highly functionalized pyridines 5 [eqn (3)]. With our convenient synthesis, the synthetic utility of new 4H-azepines 3–4 is also reported.10

Results and discussion

We examined the reactions of 3-methyl-3-en-1-ynamide 1a with 3,5-dimethylisoxazole 2a using various gold catalysts. Heating this mixture (1a/2a = 1 : 2 ratio) in hot DCE with 5 mol% LAuCl/AgNTf2 [L = p(t-Bu)2(o-biphenyl) and IPr] afforded a [4+3]-annulation product, 4H-azepine 3a, in 64% and 75% yields respectively (Table 1, entries 1–2). Under these conditions, a low loading (1.2 equiv.) of 3,5-dimethylisoxazole 2a gave 3a in a decreased yield, ca. 62% (entry 3). With a 10 mol% catalyst, IPrAuCl/AgNTf2 gave a clean reaction, yielding desired 3a up to 91% (entry 4). We tested other phosphine ligands such as PPh3 and P(OPh)3, yielding desired 3a in satisfactory yields (78–81%, entries 5–6). Other counter anions such as OTf– and SbF6– were also effective in producing 3a in 85–88% yields (entries 7–8). AgNTf2 alone was not active at all (entry 9).

Table 1

[4+3]-Annulations over various gold catalysts

Entry	Catalyst [mol%]	x	Time [h]	Yield b [%]
				1a	3a	1a–H′/1a–H′′
1 c	LAuCl/AgNTf2 [5]	2	3	20	64	—
2 d	IPrAuCl/AgNTf2 [5]	2	7	12	75	7 [2.5 : 1]
3	IPrAuCl/AgNTf2 [5]	1.2	7	23	62	5 [1 : 1]
4	IPrAuCl/AgNTf 2 [10]	2	3	—	91	Trace
5	PPh3AuCl/AgNTf2 [10]	2	3.5	—	81	5 [1.25 : 1]
6	[PhO]3PAuCl/AgNTf2 [10]	2	3.5	—	78	13 [1.1 : 1]
7	IPrAuCl/AgSbF6 [10]	2	2.5	—	85	6 [1.4 : 1]
8	IPrAuCl/AgOTf [10]	2	2	—	88	Trace
9	AgNTf2 [10]	2	15	33	—	11

[1a] = 0.15 M.

Product yields are reported after separation from a silica column.

L = p(t-Bu)2(o-biphenyl).

IPr = 1,3-bis(diisopropylphenyl)-imidazol-2-ylidene. Ms = methanesulfonyl, DCE = 1,2-dichloroethane, and Tf = trifluoromethanesulfonyl.

Suitable substituents of 3-en-1-ynamides 1 are crucial to achieve 6π cyclizations of 3-azaheptatrienyl cations V′ [eqn (2)]. We tested the reactions on 3-en-1-ynes 1b–1m bearing a C(3)-substituent to circumvent aza-Nazarov cyclizations as reported in Ye's work.7 Herein, only entries 9 and 10 showed the presence of 3-azanorcaradienes 3′. We examined these [4+3]-annulations on 3-methyl-3-en-1-ynamides 1b–1e bearing various sulfonamides NTsR4 (R4 = Me, cyclopropyl, benzyl and N(n-C4H9)(–SO2Bu)), affording the desired 4H-azepines 3b–3e in high yields (84–90%, Table 2, entries 1–4). Nevertheless, this new annulation becomes less efficient for 3-en-1-ynamide 1f bearing an oxazolidin-2-one to yield product 3f in 64% yield (entry 5).

Table 2

[4+3]-Annulations with various 3-en-1-ynamides

[1] = 0.15 M.

Product yields are reported after separation from a silica column. EWG = electron withdrawing group.

We altered the C(3)-substituents as in substrates 1g–1i; their resulting products 3g–3h (R1 = isopropyl and cyclopropyl) were obtained in 74–79%, and 3i (R1 = Ph) with only 58% yield (entries 6–8). Notably, when a long n-butyl group was present as in species 1j and 1k, their corresponding reactions afforded compounds 3j/3j′ = 5/1 and 3k/3k′ = 11.1 : 1, respectively, in 55% and 68% yields (entries 9–10). For E-configured trisubstituted 3-en-1-yne 1l (R1 = Me, R2 = Ph and R3 = H), 4H-azepine 3l and pyrrole 6l were obtained in equal proportions (entry 11). When a cyclohexenyl group was present for alkene as in species 1m, pyrrole product 6m was dominant over azepine 3m (entry 12). Accordingly, preferable 3-en-1-ynes comprise a small R2 or R3 substituent whereas R1 must be substituted. Herein, the structures of 4H-azepines 3b and 3l, and pyrrole species 6m were confirmed with X-ray diffraction.11

Isoxazoles of a wide scope are compatible with these [4+3]-annulations, as depicted in Table 3. The reaction of unsubstituted isoxazole 2b with model 3-en-1-ynamide 1b afforded the desired 4H-azepine 4a in 84% yield, together with pyrrole 7a′ in only 8% yield (entry 1). Mono-substituted 3-methyl or 5-methyl isoxazoles 2c and 2d are also suitable for these annulations to afford compounds 4b and 4c in 75% and 87% yields, respectively (entries 2–3). We prepared additional 3,5-disubstituted isoxazoles 2e–2i with R1 = alkyl and phenyl, and R2 = alkyl; their annulations proceed smoothly to produce desired 4d–4h in 69–85% yields (entries 4–8). For di-substituted isoxazoles 2j and 2k bearing R2 = Ph, 4H-azepines 4i and 4j were obtained in 61% and 71% yields respectively, together with their rearrangement products 5i and 5j in 15–30% yields (entries 9–10). Compounds 4a and 5i were characterized by X-ray diffraction.11

Table 3

[4+3]-Annulations with various isoxazoles

Entry	(R1, R2)	2	Time [h]	Yield [%]	4
(1)	H, H	2b	4	84	4a (X-ray)
				8	7a′
(2)	H, Me	2d	3	75	4b
(3)	Me, H	2c	3	87	4c
(4)	Et, Et	2e	6	85	4d
(5)	n-Bu, n-Bu	2f	7	81	4e
(6)	Me, n-Bu	2g	3	82	4f
(7)	n-Bu, c-Pr	2h	2	77	4g
(8)	Ph, n-Bu	2i	4	69	4h
(9)	Ph, Ph	2j	6.5	61	4i
				30	5i (X-ray)
(10)	Me, Ph	2k	4	71	4j
				15	5j

[1b] = 0.15 M.

Product yields are reported after separation from a silica column.

Our convenient synthesis of 4H-azepines provides new synthetic utilities; several new functionalizations are depicted in Scheme 2. NaBH4-reduction of species 3b delivered an alcohol derivative 7a in 84% yield. Selective hydrogenation of the same species afforded 2-aza-1,3-dien-5-one 7b in 71% yield. A final treatment of 4H-azepine 3b with NBS in acetone afforded compound 7c, of which the molecular structure was determined by 1H NOE spectra.

Scheme 2

New functionalization of 4H-azepines.

The Lewis-catalyzed rearrangement of 4H-azepines 3–4 to substituted pyridines 5 [eqn (3)] is unprecedented in 4H-azepine chemistry.10 We undertook such novel [4+2]-annulations between 3-en-1-ynamides 1 and isoxazoles 2 using Au(i)/Zn(ii) in a relay series, as depicted in Table 4. In the reactions of various 3-substituted 3-en-1-ynamides 1 (R1 = methyl, n-butyl, cyclopropyl and isopropyl) with 3,5-dimethylisoxazole 2a, substituted pyridines 5a–5d were obtained in satisfactory yields (51–73%, entries 1–4). In entry 1, if the reaction was performed with combined Au(i)/Zn(ii) catalysts in a non-relay operation, compounds 5a and 3b were isolated in 35% and 28% yields respectively. For 3-en-1-ynamide 1a bearing a NMs(n-butyl), the corresponding product 5e was obtained in 63% yield (entry 5). We tested the reactions on 3,5-disubstituted isoxazoles 2e–2f & 2h bearing all alkyl substituents, producing desired 5f–5h in good yields (69–78%, entries 6–8). For such disubstituted isoxazoles bearing R4 = Ph, the reactions afforded the desired pyridine derivatives 5i and 5j in 75–80% yields (entries 9–10). The molecular structures of compounds 5a and 5i were characterized by X-ray diffraction.11

Table 4

[4+2]-Annulations between 3-en-1-ynamides and isoxazoles

Entry	(R1, R2, EWG)	1	(R3, R4)	2	Time [h]	Yield [%]	5
(1)	Me, Me, Ts	1b	Me, Me	2a	19	73 (35) c	5a (X-ray)
(2)	n-Bu, Me, Ts	1k	Me, Me	2a	33	64	5b
(3)	c-Pr, Me, Ts	1h	Me, Me	2a	20	56	5c
(4)	i-Pr, Me, Ts	1g	Me, Me	2a	15	51	5d
(5)	Me, n-Bu, Ms	1a	Me, Me	2a	28	63	5e
(6)	Me, Me, Ts	1b	n-Bu, n-Bu	2f	19	78	5f
(7)	Me, Me, Ts	1b	Et, Et	2e	16	69	5g
(8)	Me, Me, Ts	1b	nBu, c-Pr	2h	20	75	5h
(9)	Me, Me, Ts	1b	Ph, Ph	2j	24	80	5i (X-ray)
(10)	Me, Me, Ts	1b	Me, Ph	2k	30	75	5j

[1] = 0.15 M.

Product yields are reported after separation from a silica column.

The value in parentheses is reported using a mixture of IPrAuCl/AgNTf2 (10 mol%) and Zn(OTf)2 (20 mol%) in hot DCE (70 °C, 48 h); 3b was also isolated in 28% yield.

Scheme 3 rationalizes the crucial roles of substituents of 3-en-1-ynamides in the chemoselectivity that relies on two conformational structures DversusD′. The N-attack of isoxazole at gold-π-ynamide A is expected to form a gold-carbene D′, which can be visualized as a gold-stabilized cycloheptatrienyl cation. Conformation D is favorable with R = H, which prefers aza-Nazarov reactions.12 When a C(3)-substituent is present (R = alkyl and aryl), all σ-cis configured species D′ are the preferable geometry to induce novel 6π electrocyclizations. This ring closure is expected to proceed through an attack of enamide at the alkenylgold moiety that is also visualized as a gold-stabilized cation. Additional C(4)-substituents render the formation of cations D′ difficult, thus yielding pyrrole 6 as byproducts. A loss of an acidic proton from seven-membered cations E is expected to yield azepines 3–4. 4H-Azepines 3–4 bear an enone conjugated with a triene; this extensive conjugation is very stable to impede a 6π electrocyclization of their triene moieties unless a Lewis acid is present. Zn(OTf)2 likely coordinates with the carbonyl of 4H-azepine 3 to generate a 2-azapentadienyl cation F bearing a zinc enolate, further enabling an intramolecular cyclization to generate species G. A 1,2-acyl shift14 of species G delivers the observed product 5.13

Scheme 3

A plausible reaction mechanism.

Conclusions

In summary, this work describes new gold-catalyzed [4+3] annulations15 of 3-substituted 3-en-1-ynamides with isoxazoles to form 4H-azepines. A relay catalysis is also developed with Au(i)/Zn(ii) catalysts to achieve [4+2] annulations from the same reactants. The mechanisms of gold-catalyzed [4+3] annulations involve unprecedented 6π electrocyclizations of 3-azacycloheptatrienyl cations to form 4H-azepines 3–4 efficiently. Control experiments confirm that 4H-azepines 3–4 are catalyzed by Zn(OTf)2 to undergo new rearrangement reactions to form substituted pyridine derivatives.

Conflicts of interest

The authors declare no conflict of interest.

Click here for additional data file.

Click here for additional data file.