Synthesis of 3-Substituted Pyrrolidines via Palladium-Catalyzed Hydroarylation.

Metal-catalyzed reactions have revolutionized synthetic chemistry, allowing access to unprecedented molecular architectures with powerful properties and activities. Nonetheless, some transformations remain sparse in number, or out of reach, even with the diverse modern catalytic chemical arsenal, including bimolecular alkene hydroarylation reactions. We report here a broad-scope, palladium-catalyzed pyrroline hydroarylation process that gives 3-aryl pyrrolidines, a class of small molecules with potency in a diverse range of biological scenarios. Thus, whereas N-acyl pyrrolines usually undergo palladium-catalyzed arylation to give alkene products, the corresponding reactions of N-alkyl pyrrolines deliver products of hydroarylation, pyrrolidines. The process has broad substrate scope and can be used to directly deliver drug-like molecules in a single step from readily available precursors.

Introduction

Small molecules with saturated and unsaturated heterocyclic cores are ubiquitous in biochemistry, and much attention has been paid to the manufacture of such structures in academia and industry. Nitrogen-containing saturated rings are particularly privileged structures in biology, and there has therefore been intense interest in the design and use of these heterocycles as drug-like molecules: ca. 60% of the currently US Food and Drug Administration-approved small molecule drugs contain such a motif (Vitaku, et al., 2014). Within the N-heterocycle-containing drug library, the pyrrolidine motif is a very frequently seen structure, and these five-membered rings are widely used in drug discovery, with even relatively simple pyrrolidines often possessing great potency (Figure 1).

Figure 1

Saturated Five-Membered Nitrogen Heterocycles

(A) 3-Aryl pyrrolidines are privileged structures, exhibiting powerful effects in a diverse range of biological scenarios, such as leishmaniasis, histone deacetylation, neurotransmission, and gene transcription.

(B) 1-Propyl-3-aryl pyrrolidines are potent and selective ligands for serotonin and dopamine receptors.

Methods to deliver functionalized pyrrolidines directly by catalytic processes are relatively scarce, with the majority of reported methods involving ring-construction, rather than peripheral modification of intact pyrrolidines. In addition, certain classes of substituted pyrrolidines lend themselves rather more favorably to ring modification than others: thus, although there have been elegant efforts for both non-catalytic (Kerrick and Beak, 1991, Beak et al., 1994, Gelardi et al., 2013, Jain et al., 2017) and catalytic (Chatani et al., 2000, Shaw et al., 2016) conversion of unsubstituted pyrrolidines to 2-substituted derivatives, there are a few methods that efficiently deliver 3-substituted pyrrolidines, a structural class with diverse biological activity (Hutton et al., 2014, Wong et al., 2012, Wang et al., 2018, Holechek et al., 2018, Grandel et al, nd WO, 2007118899A1; Drescher et al., n.d. WO, 2006040182A1; Sonesson et al., 1997). For the latter compounds, there is usually a requirement for a directing group (Affron and Bull, 2016, Feng et al., 2015, Mondal et al., 2016) or N-protection, which limits direct access to structurally simple bioactive N–H or N-alkyl pyrrolidines (such as bioactive N-propyl compounds, Figure 1B).

The Mizoroki-Heck (MH) reaction (Braese and de Meijere, 2014, Oestreich, 2009) was one of the earliest reported method (Mizoroki et al., 1971, Heck and Nolley, 1972) to execute direct, substoichiometric catalytic modification of simple alkenes; for cycloalkenes, the MH reaction proceeds by overall functionalization of an sp3—rather than an sp2—CH bond, due to the stereoelectronic control for β-hydride elimination in the key palladium(II) intermediates 1 (Figure 2A). In addition to the parent cycloalkene systems, the reaction can be applied to 2,3-dihydropyrans (Mata et al., 2007, Wu and Zhou, 2014) (2, X=O) and the analogous N-carbamoyl pyrrolines (Sonesson et al., 1996, Carpes and Correia, 2002, Montes de Oca and Correia, 2003, Garcia et al., 2005, Peixoto da Silva et al., 2007, Finelli et al., 2015) (2, X=N–C[O]R) (Figure 2), but the reactions can be unpredictable (cf. Figures 2A and 2B). MH reactions of N–H or N-alkyl azacycloalkenes can be further complicated by competing oxidation processes, and there are few reports of effective MH reactions for this class of substrate; the fact that many biologically active piperidines and pyrrolidines have this substitution pattern is a drastic limitation to the method. In addition, higher N-alkyl analogs (which can possess enhanced activity (Ekenstam et al., 1957; Figure 1B) are difficult to access directly using existing MH methodology, often requiring deacylation-alkylation strategies. We recently reported (Sweeney et al., 2018) conditions to effect the MH reaction of 1-propyl tetrahydropyridine in an improved, gram-scale, protecting-group-free route to the drug molecule preclamol (3-PPP, 4) (Figure 2C). The observed regiochemistry was ascribed to the intermediacy of a chelated palladium complex 5.

Figure 2

Mizoroki-Heck Arylation of Unsaturated Heterocycles Proceeds by Overall sp3-Functionalization

(A) Carba- and heterocycloalkene MH arylation favors allylic functionalization, controlled by β-hydride accessibility.

(B) Mizoroki-Heck-Matsuda arylations of N-acyl pyrrolines and tetrahydropyridines often give mixtures of products.

(C) MH arylations of N-alkyl tetrahydropyridines are controlled by chelation.

In theory, interception of 1 and 5 by a hydride source could lead to saturated products, rather than alkenes; in practice, such hydroarylation reactions (Beletskaya and Cheprakov, 2000, Namyslo et al., 2010) are narrowly confined to conjugate-like additions (Figure 3A) (Cacchi and Arcadi, 1983), constrained alkenes (Larock and Johnson, 1989, Bai et al., 1996), and intramolecular processes (Figure 3B) (Larock and Babu, 1987, Burns et al., 1988, Kong et al., 2017, Diethelm and Carreira, 2015). Although a rhodium-catalyzed process has been reported (Figure 3C) (So et al., 2013), to date, only one intermolecular palladium-catalyzed hydroarylation reaction to give pyrrolidines has been described (Figure 3D) (Gurak and Engle, 2018); we report here a novel, broad-scope, palladium-catalyzed hydroarylation process (Figure 3E), directly furnishing 3-substituted pyrrolidines efficiently.

Figure 3

Reductive Mizoroki-Heck Processes Are Rare, and Confined to a Small Number of Reaction Types

(A) Conjugate-type additions.

(B) Intramolecular.

(C) Rh-catalyzed pyrroline hydroarylation.

(D) N-acylpyrroline reductive MH.

(E) This work: reductive Mizoroki-Heck reaction—palladium-catalyzed hydroarylation of pyrrolines.

Results and Discussion

During the course of the optimization of the route to preclamol shown in Figure 2C, it became clear that redox side reactions were significant competitors to the desired arylation process; in addition to the monoarylated product, hydroarylated product 6 was also obtained in ca. 3% yield (Figure 4A). We supposed that the hydride necessary to deliver 6 originated in the substrate (present in excess), leading to dihydropyridiniums 7, reactive species notorious for their propensity for side reaction (such as dimerization [Baldwin et al., 1998]), and we deduced that this would explain the relatively low yields of MH products compared with the parent cycloalkenes. Based on this analysis, we proposed that reaction of the lower homolog, pyrroline, should proceed more efficiently, since the analogous oxidized by-product (pyrrole 8, Figure 4B) would be stable and therefore would neither initiate nor participate in further side reactions. If this were the case, we assumed that the latter reaction would cleanly deliver hydroarylated product (pyrrolidine 9) rather than the traditional olefinic product.

Figure 4

Hijacking Redox Side Reactions to Deliver an Efficient Alkene Hydroarylation Reaction

(A) Tetrahydropyridine MH reaction: redox processes are deleterious.

(B) Pyrroline MH reaction: harnessing pyrrole stability to favor hydroarylation.

We were, therefore, gratified to observe that reaction of pyrroline 10 with iodide 11 under the MH conditions previously identified in the tetrahydropyridine series gave pyrrolidine 9a as the only coupled product (Figure 5 yields calculated using 19F nmr); N-propylpyrrole was also obtained, in approximately equal yield, confirming the hydride source to be the excess substrate and validating the original hypothesis.

Figure 5

N-Alkyl Pyrroline Arylation: an Efficient Bimolecular Reductive Mizoroki-Heck Reaction

Encouraged by the first-pass reaction, we next undertook a screening study (key data given in Table 1), which indicated conditions using bromide 12 as substrate with 4 mol% loading of Pd catalyst (entry 11) as being optimal, when considering stoichiometry, reaction time, and yield.

Table 1

Optimization of Reaction Conditions

Entry	x/Eq.	y/mol %	z/mol %	Hal	Additivea	Baseb	t/hr	Yield (%)c
1	4	5	5	I	AgNO3	DABCO	17	20
2	4	5	7.5	I	Cu(OTf)2	DMpipd	17	78
3	4	5	7.5	I	Zn(OTf)2	DMpip	17	47e
4	4	5	7.5	Br	Cu(OTf)2	DMpip	17	77
5	4	5	7.5	Br	Zn(OTf)2	DMpip	17	32f
6	3	1	1.5	Br	Cu(OTf)2	DMpip	17	62
7	3	5	7.5	Br	None	DMpip	20	0g
8	3	1	1.5	Br	Cu(OTf)2	DMpip	90	71
9	3	2	3	Br	Cu(OTf)2	DMpip	26	71
10	3	3	4.5	Br	Cu(OTf)2	DMpip	26	78h
11	3	4	6	Br	Cu(OTf)2	DMpip	17	77i
12	3	5	7.5	Br	Cu(OTf)2	DMpip	17	80
13	2.5	3	4.5	Br	Cu(OTf)2	DMpip	26	75
14	2.5	5	7.5	Br	Cu(OTf)2	DMpip	26	76

DABCO = 1,4-diazabicyclo[2.2.2]octane.

1 equivalent.

5 equivalents.

Estimated from19F NMR.

N, N-Dimethylpiperazine.

Fluorobenzene obtained in 26% yield.

Fluorobenzene obtained in 61% yield.

Fluorobenzene obtained in 97% yield.

97% conversion.

100% conversion.

Traditional silver(I) additives delivered low yield of coupled product (entry 1), but the use of Zn(OTf)2 was productive (entries 3 and 5), though less efficient than Cu(OTf)2 (cf. entries 2 and 3, and entries 4 and 5), although only protodehalogenation was observed (in 97% yield) when no additive was present (entry 7); these data indicate that the additive is acting as a Lewis acid (vide infra).

Having identified an efficient and practical protocol, we next moved to examine the scope of the process and were gratified to observe that a diverse range of aryl bromides underwent the reductive MH reaction, delivering 3-substituted pyrrolidines 9a–9t generally in good yields (Figure 6). The process was also applicable to N-benzylpyrroline, giving the synthetically tractable pyrrolidines 13a and 13b.

Figure 6

Scope of Pyrroline Reductive Mizoroki-Heck Arylation

The power of this method is exemplified in the preparation of nanomolar dopamine antagonist 9k, which is accessed in one step using the protocol described above, compared with the multi-step process, which is the only previously described synthetic strategy to obtain 9k (Figure 7).

Figure 7

Palladium-Catalyzed Hydroarylation of Pyrrolines: an Improved Entry To Dopamine Receptor Antagonists

(A) This work: one-step synthesis of nanomolar compound 9k.

(B) Reported synthesis of 9k (Sonesson et al., 1997).

With regard to the precise mechanism in play, it is not unreasonable to assume that an intermediate such as cationic complex 14 (Figure 8) is involved (although isolation of complex 14 remains elusive, the importance of N-coordination to this process was confirmed by the use of N-sulfonyl pyrrolines under the reaction conditions, whereupon the only products obtained were 3-aryl 2-pyrrolines): 14 (formed by ligand exchange of halide for pyrroline, promoted by Cu(OTf)2 [Anson et al., 2006, Fang et al., 2014]) can rapidly be converted to palladium hydride 15 (generating pyrrole 8 as by-product), which reductively eliminates the hydroarylated product and returns the catalyst to the cycle.

Figure 8

Plausible Mechanism for Palladium-Catalyzed Pyrroline Hydroarylation

In summary, we have described a broad-scope palladium-catalyzed pyrroline hydroarylation to prepare pyrrolidines: the method is operationally simple and delivers potent bioactive small molecules in short order, and in good yields. The precise mechanistic features of these reactions are a focus of our research at present, and these data will be disclosed elsewhere, in due course.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.