Due to their importance in, for instance, natural products,² bioactive molecules,³ pharmaceuticals,⁴ nutrients (vitamins),⁵ agrochemicals,⁶ dyes,⁷ liquid crystals,⁸ and functional polymers,⁹ heteroaryl units have been recognized as essential structural motifs in various realms. The significance has motivated organic chemists to develop new methods and strategies for more efficiently constructing bonds on heteroaryl scaffolds. One of the most frequently utilized strategies for this purpose is transition metal catalysis, where diverse types of bonds can now be constructed on heteroaryl rings.¹⁰ Another option for functionalizing heteroaryl rings is the nucleophilic aromatic substitution (S_NAr) reaction; however, this reaction has often been unsatisfactory. Despite its long history of use, the negative impression seems to be ascribed to critical limitations. Aromatic compounds are intrinsically electron-rich due to their (4n + 2)π electrons but must react with electron-rich nucleophiles in the S_NAr process (Scheme 1a). This demand has narrowed the scope of substrates. Thus, anionic nucleophiles with highly electropositive metals and/or electron-poor heteroaryl electrophiles with one or more electron-withdrawing groups have been utilized (Scheme 1a).¹¹ This substrate combination appears most frequently in the conventional S_NAr reaction via an addition-elimination sequence where a Meisenheimer intermediate is involved. This mechanism can be viewed as the cause of the negative image of the conventional S_NAr reaction. However, the appearance of the concerted S_NAr reaction has triggered a major breakthrough.¹² The key feature thereof is that heteroaryl electrophiles without EWGs can serve as substrates, while metal nucleophiles are still needed in most cases (Scheme 1b).^12,13 The expanded scope of the heteroaryl electrophile is due presumably to an alternate mechanism involving a single transition state that does not require the disruption of aromaticity by way of the Meisenheimer intermediate, thereby lowering the activation energy of the process.

Our research group has been engaged in developing new Lewis-acid-catalyzed reactions, of which indium Lewis acids serve as the genesis of our study.^14,15 In 2000, we reported for the first time that indium salts are suited for activating the C≡C bond of alkynes;¹⁶ the inspiration of our indium chemistry stems from the unique carbophilic nature of allylindium reagents, which can survive under aqueous conditions without undergoing hydrolysis and thus cleavage of the C-In bond, and can successfully add to carbonyl compounds.¹⁷ Since our initial report, we have been continuing the use of indium salts as Π-Lewis acid catalysts for the activation of C≡C and C=C bonds,¹⁸ and the resulting indium-activated carbon electrophiles have been mainly utilized for the S_EAr (_E = electrophilic) reaction using (hetero)aryl nucleophiles.^14f,19 Even a C-C bond, albeit requiring the assistance of a directly connecting heteroaryl ring, can be cleaved by indium salts.^{19b,e,f,g,h,20} A series of these studies are based on our research project: "Activation of Hydrocarbon Functional Groups Classified into C≡C, C=C, C-C, and C-H²¹ mainly by Indium Lewis Acids".²²

The C-C bond cleavage, described above, is observed during the indium-catalyzed three-component alkylation of pyrroles or indoles with alkynes or carbonyl compounds and nucleophiles (Nu) (Scheme 2). We considered at the time that the coordination of the heteroaryl ring to the indium salt (InX₃ = In) would be crucial to trigger the C-C bond cleavage. Furthermore, it was anticipated that the coordination should occur on the π-face rather than the heteroatom of the heteroaryl ring, due to the carbophilicity of In. We therefore envisioned that utilizing this coordination mode could enable the direct activation of the heteroaryl ring itself. Some findings obtained by investigations performed based on the working hypothesis are discussed and summarized in the ensuing sections.

The first achievement is the S_NAr-based heteroaryl-heteroaryl bond-forming reaction²³ presented in the original article.²⁴ The topics that have not been discussed in the original article and that are crucial for this Discussion Addendum are addressed here. Interestingly, only the electron-donating OMe group serves as a leaving group (Scheme 3; Ac = acetyl), in marked contrast to the typical S_NAr reaction where EWGs like Cl and NO₂ act as leaving groups. With the more electron-rich 2,5-dimethoxythiophene (2b), the reaction occurs even at room temperature (rt).

Compounds 2 are electrophiles that react with electron-rich 1a. However, 2 is clearly more favorable with higher Π-electron density. The behavior of 2 might at first seem unusual but provides a useful insight into a reaction mechanism. The result of Scheme 4, giving 3ba-d and 3'ba-d from deuterated 1,2-dimethylindole (1b-d), is also crucial for mechanistic considerations.

Proposed reaction mechanisms that take the above observations into account are shown in Scheme 5 by the reaction of HetAr-D 1-d with 2a. First up is the Π-face coordination of 2a to In to give complex 4a, in which In serves as a transient EWG to make 2a electrophilic enough and thus to induce the nucleophilic attack of 1-d via path a and/or b, giving allylindium-type intermediates 5-d and/or 5'-d, respectively. Subsequent D⁺ transfer to their α and/or γ sites to give 6-d and 6'-d²⁵ followed by the aromatizing elimination of MeOH(D) yields 3-d and 3'-d. This reaction mechanism nicely explains the results of Schemes 3 and 4. Thus, the role of the MeO group is to enhance the π-electron density of the thiophene ring and facilitate the complexation of 2a with electrophilic In. The 23% loss of the D atom observed should be attributed to the final step that can release both MeOH and MeOD. Moreover, the formation of 3-d and 3'-d due to the proposed deuteration of the C-In bond supports the probability of π-face coordination mode 4a in which the carbon atoms of 2a directly interact with In.

Worthy of note is that neither heteroaryl-metal nucleophiles nor EWGs-substituted heteroaryl electrophiles are necessary for this strategy. Moreover, the S_NAr reaction between two electron-rich heteroarenes is unique.²⁶ To the best of our knowledge, no catalytic S_NAr reaction involving heteroarene-metal π-complexs²⁷ has been presented other than reports based on our strategy (vide infra).²⁶ Next, we envisioned that electron-rich compounds other than 1 could be suitable for the indium-catalyzed S_NAr reaction.

The electron-rich compound that we next focused on is an amine, thereby allowing the synthesis of a broad range of heteroarylamines.²⁸ Representative results obtained when using MeO-(benzo)thiophenes 2 are presented in Table 1. In(NTf₂)₃ is more effective than In(OTf)₃ for these reactions. As nucleophiles 7, primary and secondary alkyl/aryl amines with cyclic/acyclic structures can be used. With 3-bromo-4-methoxythiophene (2d), the MeO-selective amination uniquely occurs, thus leaving the Br group intact in product 8gd. If low-boiling amines are desired as nucleophiles, their salts, 7m and 7n, are good choices (8me and 8ne). This reaction features high functional group compatibility: besides functional groups listed in Table 1, C(sp²)-I, -CF₃, -CN, -OH, C(sp³)-OH, pyridyl, thiazolyl, benzyl, and C=C are all tolerated.

Heteroaryl electrophiles 2 besides MeO-(benzo)thiophenes are also capable of participating in the reaction (Table 2).

Furthermore, alcohols and thiols can be used instead of amines in this strategy,²⁹ and Scheme 6 displays representative examples.

We expected that combining two of our indium-catalyzed reactions, one of which is the S_NAr amination²⁸ and the other is the addition of heteroarenes to a C≡C bond,¹⁹ could provide expedient access to heteroaryl[b]quinolines (HA[b]Qs).³⁰ A working hypothesis is illustrated in Scheme 7. The initial step would be the S_NAr amination of 2 by 13a via Π-coordination 4 to afford 14. The C≡C bond of 14 would then be activated as in 15 to induce the intramolecular S_EAr addition of the heteroaryl ring, thereby providing 16. Aromatization of 16 would result in the generation of HA[b]Qs 17.

To verify the working hypothesis, the annulation of o-ethynylaniline (13a) with 3-methoxybenzothiophene (2e) shown in Scheme 8 was tested.

The treatment of 13a and 2e with 5 mol% of In(NTf₂)₃ under the heating conditions delivered the desired benzothieno[3,2-b]quinoline 17ae, albeit in a low yield of 11%. Switching the catalyst to In(ONf)₃ (Nf = SO₂C₄F₉) gave not only 17ae but also a small amount of o-acetylaniline (18a). The carbonyl group in 18a was assumed to be formed by indium-catalyzed hydration of the C≡C bond with H₂O present in the reaction mixture. Hence, it was proposed that 17ae could be formed via the S_NAr amination of 2e with 18a followed by intramolecular nucleophilic addition of the benzothienyl ring to the carbonyl group and dehydration. Based on this proposal, the reaction of 13a with 2e was carried out with added H₂O, and as anticipated, the yields of both 17ae and 18a were raised. Prolonging the reaction time from 24 h to 36 h further improved the yield of 17ae to 61% with the complete consumption of 18a. Due to these results, we conducted the direct annulation of 18a with 2e and obtained 17ae in 92% yield by using catalyst InBr₃, as also shown in Scheme 8. These results show that InX₃ activates the benzothienyl ring of 2e and the carbonyl group of 18a. The ability for both the Π- and σ-electron-welcoming characteristics of InX₃ presents diverse opportunities for reactions.^18b We have utilized this reactivity^{19c,d,i,j,20,31} and further demonstrate indium's utility as a two-way activator. Representative results mainly focusing on the scope of 18 are thus collected in Table 3. For example, 2-propyl (18b), CF₃ (18c), and a series of aryl (18d-h) groups are available as R¹. The carbonyl group between two aryl rings (18h), the OH group (18i), and the acetal moiety (18j) remained untouched. Various thieno[2,3-b]quinolines 17ka-ga can be also obtained from 2a instead of 2e.³²

This method followed by a two-step transformation enables to synthesize cryptolepine derivatives, which represent a significant structural motif with anti-malarial and -cancer activities.³³ Thus, for instance, the indium-catalyzed annulation of 18c with 2j can be used to construct 17cj, which, when followed by the methylation and treatment with aq. Na₂CO₃, delivers 20 (Scheme 9).

The chemistry of the indium-heteroarene π-complex can be further applied to a distinct type of reaction: indium-catalyzed formal N-arylation and N-alkylation of pyrroles.³⁴ This transformation involves a unique nitrogen-nitrogen exchange strategy, or in other words, a pyrrole-ring opening-closing strategy. A working hypothesis that was developed before embarking on this study is depicted in Scheme 10.

Upon the treatment of pyrrole (21a) and amine 7 with an indium catalyst (In), we expected the sequence of reaction steps illustrated in Scheme 10. Thus, coordination of 21a to In would make 21a electrophilic and promote reaction with 7, giving the enamine intermediate 24 via 23. Isomerization of 24 to imine 25 and coordination of its nitrogen atom to In would generate 26, which could participate in ring opening and closing to produce 28 that incorporates the nitrogen atom of 7. This sequence can be regarded as a variation on the Paal-Knorr pyrrole synthesis.³⁵ The bond formation when preparing N-aryl- and N-alkylpyrroles from pyrroles is usually made directly on the nitrogen atom. Accordingly, this indium-catalyzed process is totally distinct from the general approach and thus unique.³⁶

The N-arylation and N-alkylation of pyrroles are carried out by two methods: method A with solvent 1,4-dioxane and method B with no solvent. Representative results are summarized in Table 4.

The synthesis of 28ao-go indicates the scope of pyrroles 21 that can be formed in the reaction, and the other products found in Table 4 demonstrate the scope of amines 7. When 1,2-phenylenediamine (7q) is used, only one amino group reacted with 2-methylpyrrole (21b) to yield 28bq. With 5-amino-2-methylindole (7t), the N-arylation chemoselectively occurred on the pyrrole ring, and the indolyl N-H thus remained unmodified, producing 28ct in a high yield. No racemization was observed in the reaction of (S)-1-phenylethylamine (7w), suggesting that no pyrrolyl-N-C bond-forming step is involved in this reaction.

Although the results of mechanistic studies are not provided herein, it was demonstrated that the mechanistic proposal of Scheme 10 is plausible.³⁴

This Discussion Addendum started with a brief history of the indium π-Lewis acid that is crucial in promoting our original chemistry and influencing a subsequent series of studies utilizing the indium-heteroarene π-complex (Figure 1). Since the first discovery of the heteroaryl-heteroaryl bond-forming reaction in which the π-complex between In and the MeO-substituted heteroarene participates, we have developed a number of new reactions: the nitrogen-, oxygen-, and sulfur-heteroaryl bond-forming reactions as well as the annulation reaction through the nitrogen-heteroaryl bond formation followed by the intramolecular carbon-heteroaryl bond formation. These reactions are unique because of occurring catalytically on electron-rich heteroaryl rings and should thus be classified as a distinct type of S_NAr reaction from the conventional and concerted ones.³⁷ Moreover, the π-complex has been demonstrated to be applicable to the formal N-arylation and N-alkylation of pyrroles.

We are continuing to dedicate our efforts to the chemistry of the indium-heteroarene π-complex, with the anticipation of presenting our new findings in upcoming articles.