Rh(III)-Catalyzed Regioselective Synthesis of Pyridines from Alkenes and α,β-Unsaturated Oxime Esters

Abstract

α,β-Unsaturated O-pivaloyl oximes are coupled to alkenes by Rh(III) catalysis to afford substituted pyridines. The reaction with activated alkenes is exceptionally regioselective and high yielding. Mechanistic studies suggest that heterocycle formation proceeds via reversible C-H activation, alkene insertion and a C-N bond formation/N-O bond cleavage process.

Substituted pyridines are the most abundant heteroaromatic structures in medicinal chemistry.¹ Consequently, extensive research has focused on developing methods to access this motif.² Limitations associated with more established carbonyl condensation³ and [2+2+2] cycloaddition⁴ strategies have inspired new approaches to the pyridine core, and several elegant reports demonstrate the recent progress in this field.⁵

Previous work (Cheng,^6a,b Chiba,^6c Rovis^6d):

(1)

Bergman and Ellman^6e:

(2)

This work:

(3)

Rhodium-catalyzed coupling of α,β-unsaturated oximes and alkynes represents a particularly useful contribution to pyridine synthesis (eq 1, 2).^6,7 In a recent report, Cheng demonstrated the Rh(I)-catalyzed synthesis of polysubstituted pyridines from α,β-unsaturated oximes and internal, symmetrical alkynes.^6a Several terminal alkynes, historically avoided due to their propensity to side reactions,⁸ were reported by Bergman and Ellman to be competent substrates with a Rh(I)/phosphite system (eq 2).^6c The reactions with internal, unsymmetrical and terminal alkynes, however, suffer from low regioselectivities. Investigations by our group found that a bulky tert-butyl cyclopentadienyl ligand allowed for enhanced selectivity with internal, unsymmetrical alkynes.^6d

Despite these advances, inconsistent regioselectivities and a restricted terminal alkyne scope still limit this methodology. We envisioned that use of an alkene in the place of the alkyne, with an external oxidant in addition to the N-O bond internal oxidant,^9,10 could potentially address these problems and provide a complementary method for pyridine synthesis (eq 3).

We chose to examine the reaction of oxime derivatives of 1-acetyl-1-cyclohexene (1a) with ethyl acrylate in the presence of [RhCp*Cl₂]₂ (Cp* = pentamethylcyclopentadienyl) in 2,2,2-trifluoroethanol (TFE) (Table 1). With silver(I) acetate as an oxidant, the parent oxime affords none of the desired pyridine 3aa, and instead undergoes intramolecular cyclization to isoxazole 4 (entry 1). Isoxazole formation is similarly observed with the O-acetyl oxime ester of 1a (entry 2). The O-pivaloyl derivative (Piv = C(O)tBu), on the other hand, furnishes the desired pyridine 3aa in 30% yield (entry 3). Importantly, the 6-substituted pyridine is formed as a single regioisomer. An extensive screen of reaction conditions (see Supporting Information) revealed a 0.3 M solution of dichloroethane (DCE): acetic acid (AcOH) (2:1) to be optimal, providing pyridine 3aa in 65% yield (entry 5).

Table 1.

Reaction Optimization^a

entry	R	solvent	temp. (°C)	yield 3aa (%)b	yield 4 (%)b
1	H	TFE	74	0	35
2	Ac	TFE	74	0	35
3	Piv	TFE	74	30c	0
4	Piv	DCE/AcOH	85	45	0
5d	Piv	DCE/AcOH	85	65	0

^a1.2 equiv 2a, 0.15 M solution.

^bDetermined by ¹H NMR.

^cFormed as a ~4:1 mixture of Et and CH₂CF₃ esters.

^d0.3 M solution.

With these optimized conditions in hand, we examined the reaction of various O-pivaloyl ketoximes 1 and ethyl acrylate (Chart 1).¹¹ The oxime ester substrates are accessed easily from the corresponding ketones, hydroxylamine hydrochloride and pivaloyl chloride. Reactivity dramatically increases when the oxime ester (1) lacks substitution at the β-position. α-Alkyl and α-aryl substrates afford the 2,3,6-trisubstituted pyridines in excellent yields. Vinyl oxime ester 1h delivers the disubstituted pyridine 3ha, but in lower yield. Notably, the primary alkyl chloride 1e is tolerated in the presence of the Ag(I) oxidant, giving 3ea in good yield. In all cases, 6-substituted pyridines are formed with complete selectivity.

Chart 1. Oxime Ester Scope^a.

^aConditions: 1 (0.21 mmol), 2a (0.25 mmol), [RhCp*Cl₂]₂ (0.005 mmol) and AgOAc (0.44 mmol) in 0.7 mL DCE/AcOH (2:1) for 14h. ^b0.12 mmol scale.

We next investigated the alkene component of the reaction (Chart 2). Electron-deficient alkenes afford the desired products in excellent yields as single regioisomers. The reaction of 1c and styrene is also completely regioselective; however, the 6-phenylpyridine product 3cf undergoes a second C-H activation event,¹² resulting in a mixture of 3cf and 2-styrenyl-3cf (eq 4). On the other hand, product alkenylation is not observed with oxime ester 1g. Presumably, the slightly larger ortho-substituent in 3gf discourages alkenylation.¹³ The reaction of oxime 1g with various substituted styrenes provides the 2-aryl pyridines (3) as single regioisomers.

Chart 2. Activated Alkene Scope^a.

^aConditions: see Chart 1. For 1g: 1.05 equiv 2. ^b1.2 equiv 1c. ^c75 °C.

(4)

Alkyl alkenes are also competent coupling partners, but produce separable mixtures of the 6- and 5-substituted pyridines 3 and 5 (Table 2).¹⁴ Surprisingly, the reaction with vinylcyclohexane 2p is significantly more selective for 3 than that with vinylcyclopentane 2o (Cyp = cyclopentyl); intermediate selectivity is observed with 3-methylpentene 2q (entries 3–5). The reason for this anomaly is currently under investigation. Notably, the reaction of 3,3-dimethylbutene 2r affords the 5-substituted product 5cr exclusively, albeit in low yield, suggesting that there is a steric component to regiocontrol.

Table 2.

Alkyl Alkene Scope^a

entry	alkene	R	3:5b	yield (%)
1	2m	CH2OPh	2.2:1	66
2	2n	nHex	1:1.3	72
3	2o	Cyp	1.3:1	67
4	2p	Cy	7.2:1	69
5	2q	sBu	3.4:1	50c
6	2r	tBu	<1:20	17c,d

^aConditions: see Chart 1.

^bDetermined by ¹H NMR.

^c5 equiv 2.

^d0.54 mmol scale.

β-Substituted acrylates (6) also participate as the alkene component (Table 3). Both (Z)- and (E)-isomers of 6 react with 1c to afford the tetrasubstituted pyridines in similar yields. The reaction is selective for 2-picolinate products (7) even when the β-substituent of the alkene is a ketone (6c, entry 5).

Table 3.

Internal Alkene Scope^a

entry	alkene	CO2R	R	7:8b	yield (%)
1	(Z)-6a	CO2Me	CO2Me	n/a	71
2	(E)-6a	CO2Me	CO2Me	n/a	54
3c	(Z)-6b	CO2Et	Me	>20:1	47
4c	(E)-6b	CO2Et	Me	>20:1	51
5	(E)-6c	CO2Me	C(O)Me	4.7:1	41

^aConditions: see Chart 1.

^bDetermined by ¹H NMR.

^c5 mol % [RhCp*Cl₂]₂.

Contemplating the mechanism of the reaction, we considered two general pathways for C-N bond formation from rhodacycle A (Figure 1). In pathway A, β-hydride elimination generates azatriene B, and subsequent electrocyclization and elimination of PivOH furnishes 3. Such Rh-catalyzed C-H alkenylations have been demonstrated with various directing groups,^{10b,12,15,16,17} including oxime ethers.¹⁸ Moreover, 6π-electrocyclization of an azatriene intermediate¹⁹ is proposed in the Rh(I)-catalyzed reaction of oximes and alkynes.^6a,b,e To probe the possibility of pathway A, we used cis-fused 1,2-dihydronaphthalene as our alkene component (eq 5). If pathway A were operational, no product would be expected since a trans relationship between rhodium and the β-hydride in intermediate A would preclude β-hydride elimination.²⁰ In fact, tricyclic product 10 is formed in 32% yield, implicating an alternate pathway for C-N bond formation.

Figure 1.

Two possible pathways for C-N bond formation.

Pathway B in Figure 1 depicts a C-N bond forming/N-O bond cleaving event that is followed by oxidative aromatization to 3. This type of process is invoked in the Rh(III)-catalyzed synthesis of pyridines from oximes and alkynes^6c,d and analogous reactions of hydroxamic acid derivatives.¹⁰ In particular, Glorius demonstrates that the product of the reaction of N-oxybenzamides and alkenes hinges upon the O-substituent.^10b Namely, N-methoxybenzamides afford olefinated products while N-pivaloxybenzamides form tetrahydroisoquinolones, presumably via pathways analogous to A and B, respectively. Accordingly, the O-methyl derivative of 1a does not afford the pyridine product under our conditions.

(5)

Isotope experiments provided further insight into the mechanism of pyridine formation.²¹ Particularly, when the reaction is performed with AcOD to approximately 50% conversion, deuterium incorporation is observed at the β-position of 1c and at the 2-methyl and 5-H of 3ca. The observed deuteration at the β-position of 1c implies that C-H activation is reversible. Near complete deuterium incorporation at the 2-methyl of 3ca signals a deprotonation step; since there is no deuteration at the same position of 1c, the deprotonation most likely occurs after the first irreversible step of the reaction.²²

With these observations in mind, we propose the reaction mechanism in Scheme 2. After generation of acetate catalyst I, reversible C-H activation forms rhodacycle II. Dissociation of an acetate ligand leads to cationic complex III that may be in equilibrium with chelated complex III′. Alkene coordination and presumed irreversible migratory insertion provides rhodacycle V, and deprotonation at the α-position gives neutral complex VI. It is possible that chelation of the O-pivalate substituent prevents β-hydride elimination at this stage due to coordinative saturation.²³ Instead, a C-N bond forming/N-O bond cleaving event provides complex VII after tautomerization.²⁴ Final β-hydride elimination furnishes the pyridine product. Reductive elimination of complex VIII followed by oxidation of the Rh(I) species regenerates the active catalyst.²⁵

In conclusion, we have developed a novel pyridine synthesis from α,β-unsaturated O-pivaloyl oximes and alkenes. The reaction is general, efficient and highly regioselective with activated alkenes. Mechanistic studies support a pathway that proceeds by reversible C-H activation, alkene insertion and a C-N bond formation/N-O bond cleavage process.

Scheme 1.

Proposed Reaction Mechanism