C1-Substituted N-Alkyl Tetrahydroisoquinoline Derivatives through V-Catalyzed Oxidative Coupling

In recent years, oxidative coupling has emerged as a powerful tool for the construction of organic molecules.1 In essence, oxidative coupling produces a new CX bond from CH and XH fragments in the presence of a catalyst and a sacrificial oxidant [Equation (1)]. A major advantage of this strategy over traditional metal-catalyzed cross-coupling reactions is that neither coupling partner requires pre-functionalization, which helps to reduce waste and to streamline synthesis.2(1)

• (1)

Tetrahydroisoquinoline (THIQ) derivatives make up an extensive class of organic molecules, including numerous natural products and pharmaceutically active compounds.3 As a result of their biological importance, new synthetic routes to obtain substituted THIQs remain an active area of research. Oxidative coupling reactions on the THIQ core are well established within the literature; however, these studies have been largely restricted to N-aryl derivatives,4 which limits further transformations and, therefore, their synthetic utility. Extension of this methodology to include N-alkyl THIQs represents an important advance in this area and would provide access to pharmaceutically interesting compounds. A few examples of oxidative coupling with N-alkyl5 and N-acyl5e, 6 THIQs have been reported; however, each method is limited to a single class of nucleophile. With this in mind, we sought to develop a simple, practical method for the coupling of various N-alkyl THIQs with a wide variety of nucleophiles.

It is generally accepted that the key step in the oxidative coupling of tertiary amines is the formation of a reactive iminium ion often derived from the corresponding radical cation (Scheme 1 a).4b, 7 As this strategy appears to be less effective for N-alkyl derivatives, an alternative approach is required. Hwang and Uang have shown that a series of commercially available amine N-oxides could be converted to the corresponding iminium ion by treatment with a simple vanadium catalyst.8 We reasoned that this strategy could be extended to tertiary tetrahydroisoquinoline N-oxides formed in situ from the parent amine (Scheme 1 b).9

Scheme 1. Approaches to reactive iminium ions. a) Approach used in literature (SET=single electron transfer). b) Alternative approach used in this study.

To begin our study, N-benzyl THIQ 1 was chosen as an appropriate test substrate, as the benzyl group can be potentially removed and thus allows further functionalization of the amine. Treatment of 1 with 1 equiv of meta-chloroperoxybenzoic acid (m-CPBA) and 5 mol % vanadyl acetylacetonate, [VO(acac)₂], in nitromethane (acting as both nucleophile and reaction solvent) provided the coupling-product 2 in 77 % isolated yield (Scheme 2).

Scheme 2. Vanadium-catalyzed CC bond formation.

Encouraged by these results, we tried to optimize the reaction conditions with regard to catalyst loading, oxidant, and reaction temperature (Table 1). In the absence of the vanadium catalyst, no product formation was observed, and reducing the catalyst loading to 1 mol % resulted in a sharp decrease in the isolated yield (entries 1 and 2). Increasing the catalyst loading from 5 to 10 mol % improved the yield from 77 to an excellent 92 % (entries 3 and 4). Higher catalytic loadings had no further positive effect on the yield (entry 5). Accordingly, 10 mol % loading was used in further experiments. Then, we examined the use of alternative oxidants with an aim to improve the atom economy of our transformation. However, the use of oxygen as the terminal oxidant gave no desired product (entry 8). tBuOOH and urea hydrogen peroxide (UHP) were also not efficient, providing 2 in low yields after a reaction time of 16 h (entries 6 and 7). Although the use of tBuOOH and UHP is desirable from an environmental standpoint, m-CPBA was adopted as the oxidant of choice in light of the reduced reaction times and superior yields. The impact of an elevated temperature was also examined, but no positive effect was observed (entry 9).

Table 1. Screening of reaction conditions.^[a]

Entry	Amount of catalyst [mol %]	Oxidant	Time [h]	Yield^[b] [%]
[a] All reactions were performed at room temperature, except for entry 9, which was performed at 80°C. [b] Isolated yield. [c] Yield determined by using NMR spectroscopy.
1	–	m-CPBA	4	0
2	1	m-CPBA	4	27
3	5	m-CPBA	4	77
4	10	m-CPBA	4	92
5	20	m-CPBA	4	84
6	10	tBuOOH	16	34^[c]
7	10	UHP	16	28^[c]
8	10	O2	16	0
9	10	m-CPBA	4	79

Having established an optimized set of conditions, we examined the substrate scope of the reaction. Fortunately, the reaction conditions developed above were compatible with a wide range of nucleophilic partners. Furthermore, N-methyl THIQ derivatives were also effective substrates under these conditions (Table 2). The fact that N-methyl derivatives are efficient electrophiles is important as this motif is prevalent in many biologically active alkaloids.3

Table 2. Substrate scope under standard conditions.^[a]

Entry	NuH	Product			Yield^[b] [%]
[a] Reaction conditions: amine (0.44 mmol), nucleophile (0.88 mmol), [VO(acac)2] (0.044 mmol), CH2Cl2 (2.0 mL), RT, 16 h. [b] Isolated yield. [c] Nucleophile (2.0 mL) used as solvent. [d] 1:2.7 mixture of diastereoisomers as determined by using ^1H NMR spectroscopy. [e] 3 equiv of 1-phenyl-1-trimethylsilyloxyethylene used as nucleophile. [f] Reaction conditions: amine (0.44 mmol), alkyne (0.44 mmol), [VO(acac)2] (0.044 mmol), CuCl2⋅2 H2O (0.044 mmol), MeOH (2.0 mL), 40 °C, 16 h.
1			2	R^1=H, R^2=Bn	92^[c]
2			3	R^1=H, R^2=Me	60^[c]
3			4	R^1=Me, R^2=Bn	67^[c,d]
4			5	R^1=H, R^2=Bn	72^[c]
5			6	R^1=H, R^2=Me	70^[c]
6			7	R^1=OMe, R^2=Me	53^[c]
7			8	R^1=H, R^2=Bn, R^3=Ph	30^[c]
8			8		50^[e]
9			9	R^1=OMe, R^2=Me, R^3=p-C6H4Cl	40^[c]
10			10	R^2=Bn	60
11			11	R^2=Me	60
12			12		73
13			13	R^1=H, R^2=Bn, R^3=Ph	67^[f]
14			14	R^1=H, R^2=Me, R^3=Ph	50^[f]
15			15	R^1=OMe, R^2=Me, R^3=p-C6H4Cl	35^[f]

During our exploratory studies using different nucleophiles, extended reaction times often resulted in higher isolated yields. For this reason, a reaction time of 16 h was adopted. Additionally, if the nucleophile could not be used as the reaction solvent, 2 equiv of the nucleophile and dichloromethane were used. Nitromethane reacted with both N-methyl and N-benzyl THIQ to give coupling products 3 and 2 in 60 % and 92 % yield, respectively (Table 2, entries 1 and 2). Nitroethane was also an effective nucleophile, providing 4 as a 1:2.7 mixture of diastereomers in 67 % yield (entry 3). Acetone served as an effective nucleophilic solvent, providing products 5–7 in 53–72 % yield (entries 4–6). In contrast, acetophenones were less effective, forming 8 and 9 in 30 and 40 % yield, respectively (entries 7 and 9). The yield of 8 could be improved to a more acceptable 50 % by using the corresponding preformed silyl enol ether (entry 8). Given that malonates are usually highly reactive, this class of nucleophile was surprisingly ineffective under the standard reaction conditions. The reasons behind this lack of reactivity are currently unclear.

Conversely, heteroaromatics such as indole and 2-ethyl pyrrole reacted, providing 10–12 in 60–73 % yield (entries 10–12). Simple addition of copper(II) chloride allowed for aryl acetylene derivatives to be coupled in moderate to good yields (entries 13–15). The yields of 13, 14, and 15 were higher, when methanol was used as a solvent, probably because of the formation of hemiaminal ethers acting as a stable reservoir for the intermediate iminium ions.7

Unfortunately, our attempts to characterize the by-products formed in this transformation were unsuccessful and only an ill-defined mixture of products was obtained. For the products derived from N-benzyl THIQ, the regioselectivity of the transformation was confirmed by using MS, 2 D NMR spectroscopy, or X-ray crystallography (see the Supporting Information).

We also investigated the use of an internal nucleophile. To this end, 16 was synthesized from tetrahydroisoquinoline and methyl vinyl ketone. Exposure of 16 to the standard conditions provided tricyclic compound 17 in 30 % yield (Scheme 3). Analysis of the reaction mixture by using thin layer chromatography (TLC) revealed the presence of methyl vinyl ketone, suggesting that 16 was not sufficiently stable under the reaction conditions. The synthesis of 17, albeit in low yield, is interesting, as it forms the core structure of a series of biologically active compounds and highlights a potential application of this transformation.10

Scheme 3. Intramolecular reaction.

A proposed catalytic cycle for the reaction is shown in Scheme 4. Oxidation of N-alkyltetrahydroisoquinoline by m-CPBA gives the corresponding N-oxide 18, which reacts with [VO(acac)₂]. Abstraction of a proton, possibly through a six-membered transition state (19), gives the corresponding iminium ion 20. This is subsequently trapped by an appropriate nucleophile, which provides the observed products and releases the catalyst for further reaction.

Scheme 4. Proposed catalytic cycle.

To provide support for this mechanistic proposal, we investigated each step of the catalytic cycle separately. Treatment of 1 with m-CPBA gave the corresponding N-oxide 18 (R=Bn). When submitted to [VO(acac)₂] in acetone, 18 was converted to product 5 in 77 % isolated yield. This result is in good agreement with the results obtained from our one-pot procedure, which gave 72 % yield (Table 2, entry 4).

Finally, the newly developed method was applied to the synthesis of a pharmaceutically active compound. We identified methopholine1121, an opioid analgesic, as a potential target, which could be accessed through the hydrogenation of coupling product 15. Treatment of 15 with Pd(OH)₂ and hydrogen provided 21 in near quantitative yield (Scheme 5).

Scheme 5. Synthesis of methopholine.

In summary, we have developed a simple method for the oxidative coupling of N-alkyl tetrahydroisoquinolines by using a broad range of nucleophiles. Nitroalkanes, ketones, heteroaromatics, and alkynes could be utilized to give the cross-coupled product in moderate to good yields. The synthetic utility of the transformation has been exemplified by the synthesis of 17 and the analgesic methopholine. We envision that the broad nucleophile scope may allow for the synthesis of other pharmaceutically active compounds in the future.

Experimental Section

General procedures for the oxidative coupling and the ¹H and ¹³C NMR spectra of all starting materials and products are available in the Supporting Information. CCDC 836580http://www.ccdc.cam.ac.uk/cgi-bin/catreq.cgi contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from the Cambridge Crystallographic Data Centre through www.ccdc.cam.ac.uk/data_request/cif.

Acknowledgements

We are grateful to Prof. Ben List, the Max-Planck-Institut, and the Alexander von Humboldt-foundation (Scholarship to K.M.J.) for financial support and to Daniela Bartels for 2 D NMR analysis.