Alkyne Hydroacylation: Switching Regioselectivity by Tandem Ruthenium Catalysis

Abstract

By using tandem ruthenium-catalysis, internal alkynes can be coupled with aldehydes for the synthesis of β,γ-unsaturated ketones. The catalyst promotes alkyne transformations with high regioselectivity, with examples that include the differentiation of a methyl versus ethyl substituent on the alkyne. Mechanistic studies suggest that the regioselectivity results from a selective allene formation that is governed by allylic strain.

In contrast to an enzyme, which is typically specific, a single transition-metal complex can catalyze diverse transformations. This generality can be exploited in tandem catalysis, a strategy where one catalyst promotes two or more transformations in a cascade.¹ Ru-complexes are especially versatile in developing efficient cascade reactions, including examples that feature olefin metathesis,² reductive C–C bond coupling,³ and amide synthesis.⁴ It occurred to us that tandem catalysis⁵ could be used to overcome challenges in hydroacylation, the addition of an aldehyde C–H bond across an unsaturated π-bond.⁶ Towards expanding ketone hydroacylation, we reported a cascade involving alcohol oxidation, enantioselective ketone reduction, and lactol oxidation to generate lactones, a process in which all three transformations were promoted by Noyori’s Ru-catalyst.^5a Herein, we apply tandem Ru-catalysis to achieve the hydroacylation of alkynes with unprecedented regiocontrol to afford β,γ-unsaturated ketones.

The coupling of an aldehyde to an alkyne is a promising route to ketones, which has been studied using both transition metal catalysts (e.g., Ni,⁷ Rh,^8–9 Ru,¹⁰ Ir,¹¹ Au,¹² Cu¹³) and organocatalysts (e.g., N-heterocyclic carbenes) (Figure 1a). In alkyne hydroacylation, α,β-unsaturated ketones are afforded as the major constitutional isomer for both intramolecular and intermolecular variants. Most of these methods require terminal or symmetric alkynes to achieve high regioselectivity. Others require aldehydes bearing directing groups to promote C–H bond functionalization in preference to decarbonylation.

Figure 1.

Intermolecular Hydroacylation of Alkynes

In light of these challenges, we proposed using a metal-hydride catalyst to perform a cascade involving (1) alkyne-allene isomerization and (2) allene-aldehyde coupling (Figure 1b). This protocol would tolerate a wider scope of coupling partners and provide access to an alternative ketone isomer. We judged this cascade to be feasible on the basis of literature precedence for metal-hydride catalysis of the independent transformations. Yamamoto demonstrated alkynes as convenient surrogates for allenes in Pd-catalyzed hydroamination.¹⁵ Ishii¹⁶ and Breit¹⁷ achieved alkyne isomerizations by Ir and Rh metal-hydride pathways, respectively. Krische proposed that Ru(0) isomerized alkynes to allenes, albeit by a C–H propargyl activation mechanism.¹⁸ While Willis has reported allene hydroacylations with chelating aldehydes,¹⁹ we envisioned the resulting allene would undergo metal-hydride addition to generate a metal-allyl species that could couple to non-chelating aldehydes (Figure 1b). This mechanistic step has been shown in the coupling of allenes and aldehydes to generate homoallylic alcohols.²⁰ In our proposal, subsequent β-hydride elimination would generate the desired β,γ-unsaturated ketone and regenerate the metal-hydride catalyst. Thus, we set out to identify a single metal catalyst that could promote both transformations in sequence to afford a novel alkyne hydroacylation.

To test our hypothesis, we chose the coupling of benzaldehyde 1a and 2-pentyne 2a as the model. The unsymmetrical alkyne 2a is a challenging substrate for regioselective transformations due to the similar steric bulk of methyl and ethyl substituents (A-values of 1.74 versus 1.79 kcal/mol, respectively).²¹ An initial survey of commercially available metal hydrides complexes (Rh, Ru, Ir) revealed that RuHCl(CO)(PPh₃)₃ was a promising catalyst precursor (Table 1). Regioselectivity for β,γ-unsaturated ketone 3aa over 4aa and 5aa depended on the structure of the ligand chosen (entries 1–7). Only traditional hydroacylation products (α,β-unsaturated ketones) were obtained (4aa and 5aa) using non ferrocene-based ligands, BINAP and Tangphos (entries 1–2). Switching to ferrocene-based ligands gave the desired β,γ-unsaturated product 3aa (entries 3–6).²² Josiphos L1 provided the highest reactivity and regioselectivity (entry 5). Removing the backbone methyl group in Josiphos L1 resulted in a decrease in regioselectivity (entry 7 vs 5). Increasing the catalyst loading to 4 mol% improved the yield (entry 8). We used racemic L1 in these initial studies. The stereocenter of Josiphos L1 has little impact on enantioselectivity²³ but appears essential for both regioselectivity and reactivity.²⁴

Table 1.

Ligand Effects on Hydroacylation of 2-Pentyne^a

Entry	Ligand	Yield (%)b
		3aa	4aa	5aa
1	BINAP	0	1	3
2	Tangphos	0	5	18
3	dppf	11	43	40
4	dcypf	30	0	25
5	L1	54	0	2
6	L2	32	3	58
7	L3	41	0	8
8c	L1	81	1	2

^a1a (0.20 mmol), 2a (0.60 mmol), RuHCl(CO)(PPh₃)₃ (2 mol%), Ligand (2 mol%), toluene (0.5 mL), 80 °C, 18 h.

^bDetermined by ¹H NMR or GC-FID with 1,3,5-trimethoxybenzene (0.05 mmol) as the internal standard.

^cRuHCl(CO)(PPh₃)₃ (4 mol%), L1 (4 mol%), 15 h.

With this protocol, we explored the hydroacylation of 2-pentyne 2a with various aldehydes 1 (Table 2). Moderate to good yields (66–85%) and high regioselectivities (>20:1) were obtained with aldehydes 1 possessing varying electronic properties (entries 1–6). The coupling of 2a tolerated aldehydes 1g–h with increased steric bulk (72–78% yield, entries 7–8). Heteroaromatic β,γ-unsaturated ketones were obtained through the hydroacylation of 2-pentyne 2a with aldehydes bearing furan (1i–j), thiophene (1k), and indole (1l) rings (entries 9–12).

Table 2.

Hydroacylation with Various Aldehydes^a

Entry	R in 1	Yield (%)b
1	Ph (1a)	81 (3aa)
2	4-MeOC6H4 (1b)	69 (3ba)
3	4-MeC6H4 (1c)	85 (3ca)
4	4-FC6H4 (1d)	79 (3da)
5	4-ClC6H4 (1e)	74 (3ea)
6	4-BrC6H4 (1f)	66 (3fa)
7	3-MeOC6H4 (1g)	78 (3ga)
8	2-MeOC6H4 (1h)	72 (3ha)
9	2-Furyl (1i)	71 (3ia)
10	3-Furyl (1j)	76 (3ja)
11	2-Thienyl (1k)	87 (3ka)
12	3-N-Ts-indolyl (1l)	83 (3la)

^a1 (0.20 mmol), 2a (0.60 mmol), RuHCl(CO)(PPh₃)₃ (4 mol%), L1 (4 mol%), toluene (0.5 mL), 80 °C, 15–18 h.

^bIsolated yields of 3, >20:1 regioselectivity based on ¹H NMR or GC-FID analyses of the reaction mixtures.

We next examined the scope of alkynes (Table 3) and found that secondary (Cy) and tertiary (^tBu) alkyl substituted alkynes were suitable coupling partners (entries 1–2). High isolated yields (83–95%) and regioselectivities (>20:1) were observed in the hydroacylation of 1-phenyl-1-propyne with various aldehydes (entries 3–9). Alkynes bearing aryl groups of varying electronic and steric properties could be coupled with aldehyde 1b with 61–83% yields (entries 10–15). Switching the methyl group to an ethyl group by using in 1-phenyl-1-butyne resulted in no desired β,γ-unsaturated ketone product (entry 16). This lack of reactivity may be ascribed to increased allylic strain disfavoring β-hydride elimination to form the allene (Scheme 1).

Table 3.

Hydroacylation with Various Alkynes^a

Entry	3	R or R1 in 3	Yield (%)b
1		Cy	68 (3bb)
2		tBu	93 (3bc)
3		Ph	95 (3ad)
4		4-MeOC6H4	94 (3bd)
5		3-MeOC6H4	90 (3gd)
6		2-Furyl	87 (3id)
7		3-Furyl	94 (3jd)
8		2-Thienyl	88 (3kd)
9		3-N-Ts-indolyl	83 (3ld)
10		4-ClC6H4	77 (3be)
11		4-BrC6H4	61 (3bf)
12		4-CF3C6H4	82 (3bg)
13		4-MeOC6H4	82 (3bh)
14		3-MeC6H4	83 (3bi)
15		2-MeC6H4	67 (3bj)
16		Ph	0 (3ak)

^a1 (0.20 mmol), 2a (0.24–0.40 mmol), RuHCl(CO)(PPh₃)₃ (4 mol%), L1 (4 mol%), toluene (0.5–1.0 mL), 80–90 °C, 15–18 h.

^bIsolated yields of 3, >20:1 regioselectivity based on ¹H NMR or GC-FID analyses of the reaction mixtures.

Scheme 1.

Tandem Ru-Catalysis: Mechanism for Hydroacylation of Alkynes

To support the proposed allene intermediate, we prepared phenylallene 6 independently and then subjected this allene to benzaldehyde 1a under otherwise standard conditions (eq. 1). This control experiment provides the same product 3ad as that obtained in the hydroacylation 1-phenyl-1-propyne. While allene 6 was completely consumed, multiple unidentified side products were observed. We propose that tandem catalysis overcomes possible decomposition pathways by forming allenes in low concentration. A similar phenomenon was observed by Krische in his recent study on the synthesis of homoallylic alcohols.¹⁸

(1)

On the basis of literature and our own observations, we propose the mechanism featuring tandem Ru-catalysis (Scheme 1). The reversible insertion of 2-pentyne 2a into ruthenium hydride (Ru–H) can yield Ru-vinyl intermediate A or B. The interception of intermediate A with aldehyde 1 generates isomer 5, whereas interception with B generates isomer 4. Indeed, we observed formation of both 5aa and 4aa (~3:1 ratio) when using BINAP or TangPhos (Table 1, entries 1–2). While formation of both A and B are feasible, we propose that β-hydride elimination from B to generate the 1,3-disubstituted allene D will be disfavored due to higher allylic strain. In contrast, the β-hydride elimination of metal-vinyl complex A will form 1-substituted allene C in situ and regenerate a Ru–H species.²⁶ The reinsertion of Ru–H into the terminal allene C will occur regioselectively to yield Ru-allyl E. Trapping of aldehyde 1 through a six-membered intermediate F yields ruthenium alkoxide G.²⁰ A second β-hydride elimination from G delivers the observed ketone 3 and regenerates the Ru–H.

When the hydroacylation was performed with deutero-alkyne 2d-d₃, the deuterium label was scrambled into the α-, β- and γ-positions of ketone 3ad-d₃ (eq. 2). The incorporation of deuterium into both α- and β- positions of ketone 3ad-d₃ suggests that the Ru–H species generated from both β-H elimination steps (A to C and G to 3) are indistinguishable (Scheme 1). The observed incorporation of hydrogen at the γ-position of product 3ad-d₃ indicates the reversibility of β-hydride elimination in allene formation (Scheme 1). This hypothesis is supported by deuterium scrambling in the isotope-labeling study with aldehyde 1a-d₁ (eq. 3).

(2)

(3)

This tandem strategy can be used to access other useful motifs. For example, we investigated the coupling of aldehyde 1b with 1-(trimethylsilyl)propyne 7 for the formation of silyl dienol ether 8 (Scheme 2). A 1,3-silyl migration²⁷ of intermediate 9 generates a nucleophile that can be used in asymmetric vinylogous aldol reactions.²⁸

Scheme 2.

Silyl Dienol Ether Prepared by Tandem Catalysis

In addition, through a procedure involving tandem Ru-catalysis followed by base-catalyzed isomerization, we can generate α,β-unsaturated ketones 4 with 70–87% yields (Scheme 3). This one-pot protocol allows us to differentiate between the methyl and ethyl groups on 2-pentyne 2a with high regiocontrol and generate α,β-unsaturated ketones that would be difficult to access using traditional hydroacylation methods.

Scheme 3.

Regioselective Hydroacylation of 2-Pentyne.

By identifying the Ru-L1 catalyst, we have been able to expand the power of aldehyde-alkyne cross couplings through a cascade mechanism that enables unique regioselectivity. Due to the formation of a 1-substituted allene intermediate, our protocol differentiates a methyl from an ethyl substituent on 2-pentyne with high regioselectivity. Future studies will focus on stereoselective variants. More importantly, insights from these studies will contribute to the emerging use of alkynes as allene surrogates and guide future developments in tandem catalysis.