Synthesis of naphthalene derivatives via nitrogen-to-carbon transmutation of isoquinolines

Abstract

Heteroarene skeletal editing is gaining popularity in synthetic chemistry. Transmuting single atoms generates molecules that have distinctly varied properties, thereby fostering potent molecular exchanges that can be extensively used to synthesize functional molecules. Herein, we present a convenient protocol for nitrogen-carbon single-atom transmutations in isoquinolines, which is inspired by the Wittig reaction and enables easy access to substituted naphthalene derivatives. The reaction uses an inexpensive and commercially available phosphonium ylide as the carbon source to furnish a wide range of substituted naphthalenes. The key to the success of this transformation is the formation of a triene intermediate through ring opening, which undergoes 6π-electrocyclization and elimination processes to afford the naphthalene product. Furthermore, this strategy enables the facile synthesis of ¹³C-labeled naphthalenes using ¹³CH₃PPh₃I as a commercial ¹³C source and facilitates modifying the directing group for C─H functionalization.

The nitrogen atom in isoquinoline is directly exchanged for a carbon atom by an inexpensive Wittig reagent.

INTRODUCTION

Naphthalene is an important moiety; naphthalene derivatives exhibit a wide range of antagonistic and therapeutic activities (Fig. 1A) (1). Naphthalene is also an interesting building block for the synthesis of aromatic derivatives with optical and electronic properties and other organic materials (2). Naphthalene synthesis is an active research area owing to its numerous applications in the pharmaceutical, medicinal, and agrochemical industries (3–6). Traditionally, substituted naphthalenes are synthesized using electrophilic aromatic substitution chemistry. However, the regioselectivity of such reactions can be difficult to control and depends on already-present functional groups (7, 8). The development of useful regioselective synthesis methodologies for polysubstituted naphthalene derivatives has attracted much attention in the organic synthesis and pharmaceutical industries.

Fig. 1. Nitrogen-to-carbon transmutation of isoquinolines.

(A) Structures of biologically active naphthalene derivatives. (B) Isoquinolines used as directing groups in direct C─H functionalization chemistry. (C) Current N-to-C transmutation strategies. (D) Mechanism of the Wittig reaction. (E) Using a phosphonium ylide as a carbon source through a strategy involving intramolecular rearrangement and 6π-electrocyclization (this study). Me, methyl; Ph, phenyl.

Isoquinoline is a readily accessible and versatile intermediate in organic synthesis (9–12). In the past decade, considerable effort has been devoted to the development of site-selective C─H functionalizations to provide powerful tools for modifying the isoquinoline moieties in drugs and materials (13–17). In contrast to naphthalene, the nitrogen-containing heterocyclic feature of isoquinoline offers multiple differentiable positions and a broad diversity of substitution patterns as a consequence. Isoquinolines have been used as directing groups for direct C─H functionalization processes (Fig. 1B) (18–21). An isoquinoline is a valuable starting material for the synthesis of a naphthalene if the nitrogen atom in the isoquinoline can be successfully substituted by a carbon atom, which concurrently expands the applications of isoquinoline C─H functionalization reactions.

Recently, formal single-atom insertion or deletion reactions that modify aromatic skeletons have emerged as powerful skeletal-editing tools (22–24). These reactions can directly modify the skeletal structure of an aromatic ring in a drug molecule to explore potential interests. However, directly modifying an aromatic ring by swapping one atom without affecting ring size and aromaticity remains challenging, although it has been recognized as a highly desirable transformation (25, 26). Given the prevalence of nitrogen-containing heterocyclic rings in biologically active molecules, directly modifying valuable core structures through atom transmutations involving the nitrogen atoms is of particular importance, and it is especially interesting for evaluating structure-activity relationships in medicinal-chemistry settings. Despite recently reported methods for C-to-N transmutations involving molecular scaffolds (27–31), reverse N-to-C transmutations remain limited despite their applicability to drug-discovery applications (Fig. 1C). Morofuji, Kano, and coworkers reported the conversion of pyridines into benzene scaffolds as Zincke intermediates using a multistep protocol (32, 33). Recently, Studer and Boswell and their colleagues independently disclosed pyridine-to-benzene conversions through cycloaddition-retrocycloaddition strategies (34, 35). Very recently, Sorensen, Greaney, Song, and Glorius and their colleagues independently reported swapping N for C in a pyridine using a Zincke ring-opening and ring-closing sequence (36–39). However, owing to the unique structure of isoquinoline, these two methods are not applicable to the N-to-C transformation of isoquinoline. Directly converting a nitrogen-containing heterocycle into its corresponding all-carbon aromatic-ring system in a manner that does not alter the molecular periphery remains chemically challenging.

We sought to develop a complementary and mechanistically distinct N-to-C transmutation method that is applicable to a wide range of isoquinolines and provides access to substituted naphthalenes. When considering how to achieve N-to-C transmutation, we drew inspiration from the Wittig reaction, in which a phosphonium ylide reacts with a carbonyl compound and introduces a carbon atom into a molecule while deleting an oxygen atom (Fig. 1D) (40, 41). Herein, we introduce a method for the N-to-C transmutations of isoquinolines that affords all-carbon aromatic rings and uses methyltriphenylphosphonium bromide as the carbon source (Fig. 1E). We envisage that phosphonium ylide reacts with isoquinoline under an appropriate set of conditions to afford intermediate I that then undergoes an intramolecular rearrangement to give intermediate II. Subsequent 6π-electrocyclization and ring closure yields azaphosphetane III, which undergoes elimination to yield the naphthalene product with the release of iminophosphorane.

RESULTS

Reaction optimization

We began our investigation by examining the N-to-C transmutation of isoquinoline 1a in the presence of methyltriphenylphosphonium bromide and ^tBuONa. Unfortunately, the desired naphthalene 1 was not formed, which is possibly ascribable to the low electrophilicity of the isoquinoline moiety. We reasoned that the isoquinoline moiety should be successfully activated through the formation of the corresponding isoquinolinium salt. Therefore, we tested our hypothesis by preparing and reacting a series of isoquinolinium salts (Fig. 2A). We found that only alkyl isoquinolinium salts (1as and 1bs) effectively provided the desired naphthalene product 1, with the methyl isoquinolinium salt (1as) facilitating the desired transformation most efficiently, whereas other isoquinolinium salts did not afford the desired product. Encouraged by this preliminary result, we investigated various bases (Fig. 2B, entries 1 to 4) and were glad to find that isoquinolinium salt 1as was converted into naphthalene 1 in the highest yield when ^tBuONa (2 equiv) was used (Fig. 2B, entry 2); a lower yield was obtained when ^tBuONa (1 equiv) was used. The use of other bases did not improve the reaction outcome. Further screening revealed that polar aprotic MTBE was the best choice of solvent (Fig. 2B, entries 6 to 8).

Fig. 2. Reaction optimization.

(A) Preliminary experiments. Reactions were conducted using different isoquinolinium salts (0.2 mmol) in the presence of ^tBuONa (2.0 equiv) and CH₃PPh₃Br (1.0 equiv) in MTBE at 120°C. (B) Control experiments. Isolated yields are given. t-Bu, tertiary butyl; Bn, benzyl.

Substrate scope

Substrate scope was examined based on the optimal reaction conditions determined above (Fig. 3). With isoquinolines bearing either alkyl or aryl substituents at C-1 (2 to 7), the N-to-C transmutation occurred smoothly; however, C-1 aryl-substituted isoquinolines afforded the products in relatively low yields. A wide range of isoquinolines with varying substitution patterns underwent facile atom transmutation to afford the corresponding substituted naphthalenes. Various functional groups, including bromo (11, 18, 25, and 31), chloro (12), fluoro (13), ether (14 and 16), thioether (15), and alkene (16 and 17), were tolerated in this transformation. Heterocyclic moieties, such as thiophene (19, 20, and 28), furan (21 and 29), phenoxathiine (22), benzothiophene (23), and dibenzothiophene (30), were also compatible, which suggests that N-to-C transmutation selectively occurs at the isoquinoline ring.

Fig. 3. Substrate scope of N-to-C transmutation.

Standard reaction conditions: isoquinoline (0.2 mmol) and CH₃I (0.8 mmol) in MeCN at 90°C for 12 hours, then CH₃PPh₃Br (0.2 mmol) and ^tBuONa (0.4 mmol) in MTBE at 120°C overnight. Isolated yields after chromatography from isoquinolines are shown. i-Pr, isopropyl.

Additional trisubstituted systems containing F, Cl, Br, I, MeO, and MeS groups were also tolerated to afford trisubstituted products 32 to 37. Such densely substituted analogs are appealing to medicinal and materials chemists, although their syntheses are often challenging. We also note that Wittig reaction–sensitive functionalities are not maintained in this reaction (e.g., ketone 38a was converted into alkene 38). However, ketones are compatible with the reaction conditions following protection (39). Quinolines 40a and 41a were not converted into the desired products. Last, isoquinoline 42a failed to be converted into the desired products. A large amount (80%) of the starting material is recovered without any desired N-to-C transmutation product. This result indicated that C1 substitution of isoquinolines was crucial for the success of this atom transmutation process.

Synthetic utility

¹³C-labeled molecules are widely used in mechanistic and nuclear magnetic resonance (NMR) studies. Therefore, developing efficient methods that facilitate such approaches is of interest. The recent advent of atom transmutation provides opportunities for installing isotopic labels (42–45). Commercially available ¹³C-labeled methyltriphenylphosphonium iodine can be used to synthesize multisubstituted naphthalene rings labeled with the stable carbon-13 isotope at specific positions (Fig. 4A). Subjecting an isoquinoline to the optimized reaction conditions with ¹³C-labeled methyltriphenylphosphonium iodine afforded the corresponding labeled naphthalene derivatives 43 to 47.

Fig. 4. Application potential of N-to-C transmutation.

(A) Access to ¹³C-labelled naphthalenes. (B) Selectively benzylic alkylation. (C) Modify directing group after C–H activations. (D) Scale-up reaction.

In recent years, considerable effort has been directed toward the development of methods for benzylic alkylation in structurally complex molecules. However, distinguishing multiple benzylic positions in a complex molecule is difficult. The innovative approach developed herein provides a naphthalene-synthesis solution that uses the corresponding isoquinoline as the precursor. Isoquinoline 48b, which contains two distinct benzylic positions, was selected as an example. Because of the particular acidity of the protons of isoquinoline-2-alkyl groups, alkylation selectively occurred at the benzylic position adjacent to the isoquinoline moiety, and 48a was then smoothly converted into naphthalene 48 in 41% yield under standard conditions (Fig. 4B).

Transition metal–catalyzed C─H functionalization has been shown to be a potent technique for directly transforming arenes into diverse valuable products, particularly when isoquinolines are used as efficient directing groups. However, these directing groups, which are crucial for these selective transformations, are often difficult to remove or convert into other functional groups, which restrict the broader applications of these methodologies as a consequence. The method developed herein provides opportunities for the abovementioned transformations; that is, converting an isoquinoline directing group into a naphthalene through atom swapping. As noted above, combining this methodology with existing selective C─H functionalization protocols enabled the development of an approach for the selective synthesis of substituted naphthalenes (Fig. 4C). For example, 1,2-diarylethanes bearing two aryl groups are fundamental structural motifs in a number of bioactive compounds and drugs. Isoquinoline 1a can be converted into 1,2-diarylalkane 49 by C─H/C─H coupling involving the heteroarene (46) in combination with N–C transmutation. In addition, 2-silyl-1-arylalkane 50 (47) and 2-alkyl-1-arylalkane 51 (48) were synthesized from isoquinoline 1a using a similar strategy. To demonstrate the utility of the current method, we reacted 1as on the gram scale, which provided 1 in 50% yield (Fig. 4D).

Mechanistic studies

We also conducted experiments aimed at exploring the reaction mechanism. The reaction of isoquinoline 7a or 1a with CD₃-substituted methyltriphenylphosphonium iodine afforded deuterium-labeled D-7 and D-1 (Fig. 5A). These results confirm that the C─H motif in the products originates from methyltriphenylphosphonium iodine. At the same time, the alpha C─H position of 1a was also deuterated. This H/D exchange may occur before the transmutation, due to the acidity of the isoquinoline’s alpha C─H. Next, the formation of the triphenylphosphinimine was confirmed by electrospray ionization high-resolution mass spectrometry (ESI-HRMS) and NMR (Fig. 5B). Last, when isopropyl, benzyl, ethyl, or carbethoxymethyl phosphonium ylide was used, the reaction was observed to completely shut down, which may be attributed to the obstruction of the 6π-electrocyclization after the formation of the substituted intermediate II (Fig. 5C). Collectively, these observations support the proposed mechanism shown in Figure 1E.

Fig. 5. Mechanistic studies.

(A) Deuterium labelling studies. (B) The detection of triphenylphosphinimine. (C) Control experiments. (D) DFT calculations on the proposed reaction pathway.

Density functional theory (DFT) calculations were used to examine the energetics of the proposed pathways (Fig. 5D). The nucleophilic addition of the substrate (sub) with the phosphonium ylide is slightly endergonic to form IM-1. Then, deprotonation of the intermediate IM-1 was exergonic by 31.6 kcal/mol to produce IM-2. The C─N bond cleavage barrier from IM-2 through TS-1 was found to be 12.8 kcal/mol, affording intermediate IM-3. IM-3 subsequently underwent 6π-electrocyclization via TS-2, leading to the formation of IM-4 with an activation barrier of 31.1 kcal/mol. The formation of product was highly exergonic (37.5 kcal/mol) compared to IM-4 and was irreversible. Overall, the calculations indicate that 6π-electrocyclization is the rate-determining step. The activation free energy barrier (31.1 kcal/mol) is reasonable considering the reaction conditions (e.g., 120°C).

DISCUSSION

In summary, we developed a method for the synthesis of substituted naphthalenes that involves N-to-C single-atom transmutations in isoquinolines. This reaction features a one-step process and precisely swaps the N atom for a CH unit without affecting the remaining atoms. Mechanistic studies revealed an efficient single-atom transmutation pathway involving ring-opening, 6π-electrocyclization, and elimination steps, and provides an alternative protocol for advancing the development of skeletal editing chemistry. The developed method demonstrates great potential for the preparation of substituted naphthalenes in academia and industry.

MATERIALS AND METHODS

General procedures

Unless stated otherwise, all reactions were carried out using predried glassware under an inert atmosphere using standard Schlenk techniques, or in a MIKROUNA Super (1220/750) glovebox. Solvents were treated before use according to the standard methods. Dry and degassed solvents were obtained from Energy Chemical or by distillation over the appropriate drying agent and were stored under a protective gas atmosphere. Thin-layer chromatography was performed using Thin-Layer Chromatography (BM000001) from Energy and visualized by ultraviolet irradiation and/or phosphomolybdic acid. Column chromatography was carried out on silica gel (300 to 400 mesh) using a forced flow of eluent at 0.3 to 0.5 bar pressure. Flash column chromatography was carried out using silica gel (200 to 300 mesh) at increased pressure. ¹H-NMR and ¹³C-NMR spectra were recorded on a WNMR-I spectrometer (400 MHz ¹H, 100 MHz ¹³C, 376 MHz ¹⁹F). The spectra were recorded in CDCl₃ as the solvent at room temperature. HRMS were recorded using Bruker Daltonics Micro Tof-Q II mass spectrometer (ESI), SHIMADZU GCMS-QP2010 Plus (EI). Melting points were measured on an SGWX-4A meting point apparatus and are uncorrected.

General condition for nitrogen-carbon transmutation

In an N₂-filled glovebox, an oven-dried 10-ml sealable tube equipped with a Teflon-coated magnetic stir bar was charged successively with isoquinoline substrates (0.2 mmol, 1.0 equiv), CH₃I (0.8 mmol, 4.0 equiv), and MeCN (0.5 ml). The tube was then sealed with a Teflon screw cap, moved out of the glovebox, and placed on a hotplate preheated to 90°C with vigorous stirring for 12 hours. After the reaction was cooled to room temperature, Et₂O (6 ml) was added to the mixture, which resulted in rapid crystallization to give a yellow solid. Filter off the precipitate and wash the precipitate with Et₂O (3 × 5 ml). Then, in an N₂-filled glovebox, an oven-dried 10-ml sealable tube equipped with a Teflon-coated magnetic stir bar was charged successively with isoquinoline salts, CH₃PPh₃Br (0.2 mmol, 1.0 equiv), ^tBuONa (0.4 mmol, 2.0 equiv), and MTBE (2 ml). The tube was then sealed with a Teflon screw cap, moved out of the glovebox, and placed on a hotplate preheated to 120°C with vigorous stirring for 12 hours. After the reaction was cooled to room temperature, the solvent was removed under vacuum, and the resulting pure product was purified by column chromatography over silica gel.

Supplementary Materials

This PDF file includes:

Supplementary Text

Tables S1 and S2

Figs. S1 to S3

References