Multibond Forming Tandem Reactions of Anilines via Stable Aryl Diazonium Salts: One-Pot Synthesis of 3,4-Dihydroquinolin-2-ones

: A fast and e ﬀ ective one-pot tandem process that generates Heck coupled products from readily available anilines via stable aryl diazonium tosylate salts was developed. The mild and simple procedure involves rapid formation of aryl diazonium salts using a polymer-supported nitrite reagent and p -tosic acid, followed by a base-free Heck − Matsuda coupling with acrylates and styrenes. Using 2-nitroanilines as substrates, the one-pot tandem process was extended for the direct synthesis of 3,4-dihydroquinolin-2-ones. In this case, following diazotization and Heck − Matsuda coupling to give methyl cinnamates, addition of hydrogen and reutilization of the palladium catalyst for reduction of the nitro group and hydrogenation of the alkene resulted in e ﬃ cient formation of 3,4-dihydroquinolin-2-ones. The synthetic utility of this one-pot, four-stage process was demonstrated with the ﬁ ve-pot synthesis of a quinolinone-based sodium ion channel modulator.


■ INTRODUCTION
The Mizoroki−Heck coupling of aryl halides with olefins in the presence of a palladium catalyst and a base has become a general method for carbon−carbon bond formation. 1 This powerful transformation allows the stereoselective synthesis of highly substituted alkenes, key synthetic building blocks for organic chemistry. 2 Despite the many applications of the Mizoroki−Heck reaction, the transformation can be limited by the requirement of a base, elevated temperatures, and olefin migration. A less-utilized variant of this cross-coupling is the Heck−Matsuda reaction that uses aryl diazonium salts as more reactive aryl halide surrogates (Scheme 1a). 3 The increased electrophilic nature of aryl diazonium salts allows the Heck− Matsuda reaction to proceed at lower temperatures and often without the need for ligands or a base. 4 Under these milder conditions, alkene migration is minimized, generating cleaner products.
Despite these advantages, the safety hazards associated with aryl diazonium salts has restricted the widespread use of the Heck−Matsuda reaction. This has partly been countered by replacing aryl diazonium chlorides and acetates that are unstable above 0°C, with more stable tetrafluoroborate crystalline salts. 5 However, these compounds still require storage at low temperatures and in the dark. To minimize the safety issues of handling and storing aryl diazonium salts, onepot methods in which the aryl diazonium salts are formed in situ and directly subjected to a Heck−Matsuda reaction have been developed. 6 Examples include a procedure reported by Andrus and coworkers that used an imidazolium carbene as a palladium ligand for the base-free coupling of a focused series of substrates in 17−62% yield (Scheme 1b). 7 In recent years, the Felpin group made a major contribution in understanding and advancing the applications of the Heck−Matsuda reaction. 8 This included the development of a one-pot tandem synthesis of Heck adducts from anilines using catalytic amounts of diazonium salts via a double catalytic cycle (Scheme 1c). 8c This process was found to be general for the efficient coupling of a wide range of substituted anilines with methyl acrylate, giving the Heck adducts in excellent yields over a reaction time of 48−65 h.
In 2008, the groups of Filimonov and Chi reported the synthesis and characterization of aryl diazonium tosylate salts. 9 These were prepared from anilines using a polymer-supported nitrite reagent under mild conditions and were found to have particularly high thermal and aging stability. Despite these properties, aryl diazonium tosylate salts are still reactive and have been utilized in standard substitution and cross-coupling reactions. 9,10 We recently demonstrated that aryl diazonium tosylate salts could be generated in situ and subjected to an iodination reaction for the one-pot tandem synthesis of aryl iodides from anilines. 11 Because of the relatively stable nature of aryl diazonium tosylate salts and this initial demonstration of their application in one-pot tandem processes, we were interested in further exploiting these compounds for additional multistep transformations. Herein, we now report a rapid, mild, and nonhazardous one-pot tandem diazotization and Heck reaction of anilines for the general preparation of cinnamates and styrenes (Scheme 1d). The highly chemoselective nature of this tandem process is demonstrated with halogenated anilines that following diazotization and Heck reaction can undergo further functionalization via additional cross-coupling reactions. We also describe an extension of the tandem process with 2-nitroanilines in which the Heck adducts can be directly converted to 3,4-dihydroquinolin-2-ones by the addition of hydrogen and reutilization of the palladium catalyst.

■ RESULTS AND DISCUSSION
In several studies of the Heck−Matsuda reaction, nitrosubstituted anilines have been found to be problem substrates. 5,6d,12 The high redox potential and a favorable homolytic dediazonization pathway of nitro-substituted aryl diazonium salts have been used to explain the facile decomposition of these compounds and the complex mixtures from attempted Heck−Matsuda reactions. Felpin and coworkers demonstrated that their one-pot diazotization and Heck−Matsuda reaction process could overcome these issues, allowing the efficient generation of coupled products (Scheme 1c). 8c Because of the issues associated with nitro analogues and to investigate whether our one-pot tandem process could also overcome these problems, our study began by investigating the one-pot synthesis of aryl diazonium tosylate salts and Heck− Matsuda coupling reaction using 4-nitroaniline (1a) as a substrate. Initially, a one-pot tandem process in which all the reagents were added together was investigated. This included the polymer-supported nitrite reagent, which prevents the release of nitrogen oxides and is easily prepared by the ion exchange of tetraalkylammonium functionalized resins, such as Amberlyst A-26 with aqueous solutions of sodium nitrite. 9,11 Initially, the use of tetrafluoroboric acid as a proton source was compared with p-tosic acid ( Table 1, entries 1 and 2). A higher conversion and cleaner reaction were observed for p-tosic acid, and so this was used for further optimization. An increase in temperature to 60°C resulted in a higher conversion and a much shorter reaction time (entry 4). Increasing both the palladium acetate catalyst loading to 15 mol % and the reaction time to 1.5 h allowed full conversion (entry 5). Finally, lowering the catalyst loading to 5 mol % under the optimized conditions still gave full conversion and an isolated yield of 68% for cinnamate 2a. Throughout this optimization study and in accordance with the Felpin one-pot process, 8c no issues of dediazonization and decomposition of the aryl diazonium tosylate salt intermediate were observed. It should also be noted that during the early stages of optimization of this process, a wide range of palladium catalysts (e.g., PdCl 2 , Pd(PPh 3 ) 4 , Pd(tfa) 2 , Pd(hfacac) 2 , Pd(acac) 2 , Pd 2 dba 3 , PdCl 2 (dppf) etc.) and associated ligands (e.g., (n-Bu) 3 P, (2furyl) 3 P, S-Phos, Dave-Phos, X-Phos, etc.) were trialed; however, the standard base-free conditions using palladium acetate gave the best results.
Following the optimization of a rapid and mild one-pot tandem process for diazotization and Heck coupling, the scope of this transformation was explored using various anilines and methyl acrylate (Scheme 2). Irrespective of the electronic nature or the substitution pattern of the aniline, all examples investigated were fully converted under the standard conditions to the corresponding Heck adduct in yields of 53−83%. Interestingly, the one-pot diazotization and Heck− Matsuda coupling of halogenated anilines was completely chemoselective, giving compounds 2d−2f, 2i, and 2l as the sole products. These results, particularly for 2e, 2f, and 2i, exemplify the superior electrophilic nature of diazonium salts compared to halides. The one-pot process was also investigated for the synthesis of E-stilbenes. Reaction of 4nitroaniline (1a) with styrene or 4-fluorostyrene gave Estilbenes 2n and 2o in 74 and 46% yields, respectively.
As the one-pot process with halogenated anilines gave only Heck−Matsuda adducts and none of the Mizoroki−Heck products, we wanted to exploit this chemoselectivity for the two-pot synthesis of orthogonally functionalized aryl com-Scheme 1. Synthesis of Heck Adducts from Anilines The Journal of Organic Chemistry Article pounds. 13 The one-pot diazotization and Heck−Matsuda coupling of 4-bromoaniline (1e) was scaled up, allowing the multigram synthesis of methyl (E)-3-(4-bromophenyl)acrylate (2e) (Scheme 3). Having used the amino group to implement the first cross-coupling process, 2e was then subjected to a Mizoroki−Heck reaction with styrene. 14 This gave E-stilbene 3 in 72% yield. In a similar fashion, arylation or allylation by Suzuki−Miyaura reaction of 2e completed the efficient two-pot synthesis of acrylates 4a−4d. This study shows that a combination of the one-pot diazotization and Heck−Matsuda process with a further cross-coupling reaction allows the highly controlled and selective introduction of multiple unsaturated moieties onto a central arene core.
A key objective of this project was to expand the scope of the one-pot diazotization and Heck−Matsuda process to include 2-nitroanilines as substrates. It was proposed that the resulting 2-nitrophenyl acrylates could be converted to 3,4dihydroquinolin-2-ones by extending the one-pot process to include reduction and cyclization steps. Before attempting the extended one-pot process, a series of substituted 2-nitroanilines was initially examined as substrates for the one-pot twostep tandem process (Scheme 4). Using the previously developed standard conditions, a wide range of acrylates 6a− 6n bearing various functional groups and substitution patterns were prepared in 56−87% yields. Again, complete chemoselectivity was observed with brominated (5i) and iodinated (5j) anilines.
As 2-nitroanilines were found to be excellent substrates for the one-pot diazotization and Heck−Matsuda reaction, the use of these in an extended one-pot multibond forming process for the preparation of 3,4-dihydroquinolin-2-ones was next investigated. 3,4-Dihydroquinolin-2-ones are important heterocyclic scaffolds for synthesis and are found in a wide range of pharmaceutically active agents. 15 Because of this importance, numerous methods have been developed for their synthesis, including one-pot processes. For example, Jiao and coworkers used a combination of radical and ionic processes for the reaction of 2-iodoanilines with ethyl acrylate for the preparation of 3,4-dihydroquinolin-2-ones in 17

The Journal of Organic Chemistry
Article while the Lautens groups developed a one-pot Rh/Pd catalyzed conjugate addition and amidation process of 2chlorophenyl boronic acids with acrylamide to generate a series of these compounds in 53−96% yields. 17 Felpin and coworkers also reported a one-pot synthesis of 3,4-dihydroquinolin-2ones from 2-nitroaryl tetrafluoroborate diazonium salts using a Heck−Matsuda reaction with acrylates followed by the addition of charcoal and a reduction and cyclization sequence. 8b In our study, we wanted to demonstrate that 3,4-dihydroquinolin-2-ones could be prepared directly from 2nitroanilines using the palladium catalyst to effect the Heck− Matsuda coupling and the hydrogenation/reduction steps without the use of any additives. Using 2-nitroaniline (5a), the one-pot tandem diazotization and Heck−Matsuda sequence was performed as before (Scheme 5). On complete conversion to methyl acrylate 6a, the reaction mixture was placed under an atmosphere of hydrogen. After 24 h, this gave 3,4dihydroquinolin-2-one 7a in 73% overall yield. Having shown that a one-pot diazotization/Heck−Matsuda/reduction/cyclization sequence could lead directly to 3,4-dihydroquinolin-2-ones, the scope and limitations of this process were investigated. Alkyl substituted 2-nitroanilines and a substrate bearing a fluorine substituent were readily converted to the corresponding 3,4-dihydroquinolin-2-ones 7a−7d and 7g, in yields of 64−79%, under these conditions. Electron-rich methoxy substituted 2-nitroanilines 5e and 5f could also be converted to 3,4-dihydroquinolin-2-ones 7e and 7f, however, the hydrogenation/reduction step at atmospheric pressure was slow, leading to mixtures of the target 3,4-dihydroquinolin-2ones and methyl 2-nitrophenylpropionate intermediates after 24 h. 18 To effect cleaner, more efficient syntheses of these compounds, the one-pot process was repeated by conducting the reduction stage under pressure at 2.5 bar. This gave 3,4dihydroquinolin-2-ones 7e and 7f as the sole products in 74 and 57% yields, respectively. A limitation of this one-pot process is that 2-nitroanilines bearing labile carbon−halogen bonds are subject to dehalogenation at the reduction stage. For example, the attempted use of 5-chloro-2-nitroaniline (5h) as a substrate for this process gave 5-chloro-3,4-dihydroquinolin-2one in 28% yield but as a 1:1 inseparable mixture with the deschlorinated product 7a. Our interest in preparing halogenated 3,4-dihydroquinolin-2-ones was for the potential structural diversification of these through cross-coupling reactions after the one-pot process. To overcome this limitation, crosscoupling reactions were conducted prior to the one-pot process for the efficient two-pot synthesis of 7-aryl-3,4dihydroquinolin-2-ones. Suzuki−Miyaura reaction of 4-iodo-2-nitroaniline (5j) with various aryl boronic acids (see Supporting Information for full details) was then followed by the one-pot process, which gave 7h−7j in 57−73% yield. While the standard diazotization and Heck−Matsuda steps for the 4-aryl-2-nitroanilines could be conducted under standard conditions, again the use of 2.5 bar of pressure for the reduction stage allowed the most efficient synthesis of these analogues.
The synthetic potential of the 3,4-dihydroquinolin-2-ones prepared from the extended one-pot process was then demonstrated with the synthesis of a pharmaceutically active target. N-Acetic acid derived 3,4-dihydroquinolin-2-ones bearing 6-or 7-aryl groups are late stage sodium channel blockers and have the potential to be used in the treatment of cardiovascular diseases and diabetes. 19 In this study, a sodium channel modulator, 3,4-dihydroquinolin-2-one 11, was prepared via a five-pot synthesis (Scheme 6). As shown above, the extended one-pot process was used to convert 5-methoxy-2nitroaniline (5e) to the corresponding 3,4-dihydroquinolin-2-

The Journal of Organic Chemistry
Article one 7e on a gram scale in 57% yield. Quinolin-2-one 7e was alkylated in quantitative yield using t-butyl chloroacetate and sodium hydride. A highly regioselective iron(III) triflimidecatalyzed halogenation with N-bromosuccinimide (NBS) was used to functionalize the 6-position of the aryl ring. 20 The combination of catalytic amounts of iron(III) chloride and the ionic liquid [BMIM]NTf 2 forms iron(III) triflimide in situ, which acts as a powerful Lewis acid for the activation of NBS. This reaction gave 6-bromo-3,4-dihydroquinolin-2-one 9 as the sole product in 81% yield. Suzuki−Miyaura coupling of 9 with 4-chlorophenyl boronic acid under standard conditions gave 6aryl-3,4-dihydroquinolin-2-one 10 in 75% yield. Finally, TFA mediated deprotection of the t-butyl ester completed the synthesis of sodium ion channel modulator 11 in 32% overall yield from 2-nitroaniline 5e.

■ CONCLUSIONS
In summary, a rapid one-pot tandem process involving diazotization and base-free Heck−Matsuda coupling of anilines with acrylates and styrenes was developed. The use of a particularly mild procedure to form stable aryl diazonium tosylate salts using a polymer-supported nitrite reagent and ptosic acid as the proton source allowed the synthesis of a wide range of Heck adducts. A particular feature of this process was the chemoselective coupling of brominated and iodinated anilines, which could then be further functionalized using additional cross-coupling reactions, leading to orthogonally substituted arenes. Using 2-nitroanilines as substrates, the onepot process was extended for the direct synthesis of 3,4dihydroquinolin-2-ones. Following the diazotization and Heck−Matsuda steps, reutilization of the palladium catalyst without the requirement of additional ligands or additives allowed hydrogenation of the alkene, reduction of the nitro group, and in situ cyclization to give a series of 3,4dihydroquinolin-2-ones. The potential of these one-pot multibond forming processes for application in total synthesis and medicinal chemistry was demonstrated with the five-pot preparation of a sodium ion channel modulator. Investigation of further applications of stable aryl diazonium salts in one-pot multistep processes is currently underway.

■ EXPERIMENTAL SECTION
All reagents and starting materials were obtained from commercial sources and used as received unless otherwise stated. Dry solvents were purified using a solvent purification system. Brine refers to a saturated solution of sodium chloride. All reactions were performed in oven-dried glassware under an atmosphere of argon unless otherwise stated. Flash column chromatography was carried out using silica gel (40−63 μm) and neutral aluminum oxide (50−200 μm). Aluminumbacked plates precoated with silica gel 60 (UV 254 ) were used for thin layer chromatography and were visualized under ultraviolet light and by staining with KMnO 4 or ninhydrin. 1 H NMR spectra were recorded on a NMR spectrometer at 400 MHz and data are reported as follows: chemical shift in ppm relative to tetramethylsilane or the solvent as the internal standard (CDCl 3 , δ 7.26 ppm), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet or overlap of nonequivalent resonances, integration). 13 C NMR spectra were recorded on a NMR spectrometer at 101 MHz and data are reported as follows: chemical shift in ppm relative to tetramethylsilane or the solvent as internal standard (CDCl 3 , δ 77.0 ppm), multiplicity with respect to hydrogen (deduced from DEPT experiments, C, CH, CH 2 , or CH 3 ). IR spectra were recorded on a FTIR spectrometer; wavenumbers are indicated in cm −1 . Mass spectra were recorded using electron impact or electrospray techniques. HRMS spectra were recorded using a dual-focusing magnetic analyzer mass spectrometer. Melting points are uncorrected.
General Procedure for the Preparation of the Polymer-Supported Nitrite. To a stirred solution of sodium nitrite (0.55 g, 8.00 mmol) in water (20 mL) was added Amberlyst A26 hydroxide form resin (1.00 g, 4.00 mmol). The resulting mixture was stirred at room temperature for 0.5 h, and then polymer-supported resin was filtered and washed with water until the pH of filtrate became neutral. The content of prepared polymer-supported nitrite was 3.5 mmol of NO 2 per g of resin. 9 Methyl (E)-3-(4′-Nitrophenyl)acrylate (2a). 8c To a stirred solution of 4-nitroaniline (1a) (0.028 g, 0.20 mmol), polymer-supported nitrite (0.17 g, containing 0.60 mmol of NO 2 ), p-toluenesulfonic acid monohydrate (0.10 g, 0.60 mmol), and palladium(II) acetate (0.0020 g, 0.010 mmol) in methanol (2 mL) was added methyl acrylate (0.090 mL, 1.0 mmol). The reaction mixture was heated to 60°C and stirred for 1.5 h. The mixture was cooled to room temperature and filtered, and the resin was washed with methanol (2 mL (24), 102 (26).

■ ACKNOWLEDGMENTS
Financial support from the University of Glasgow (studentship to R.J.F.) and EPSRC (EP/K503903/1) is gratefully acknowledged.