Cu(i) catalysis for selective condensation/bicycloaromatization of two different arylalkynes: direct and general construction of functionalized C–N axial biaryl compounds

Selective condensation/bicycloaromatization of two different arylalkynes is firstly developed under ligand-free copper(i)-catalysis, which allows the direct synthesis of C–N axial biaryl compounds in high yields with excellent selectivity and functional group tolerance. Due to the critical effects of Cu(i) catalyst and HFIP, many easily occurring undesired reactions are suppressed, and the coupled five–six aromatic rings are constructed via the selective formation of two C(sp2)–N(sp2) bonds and four C(sp2)–C(sp2) bonds. The achievement of moderate enantioselectivity verifies its potential for the simplest asymmetric synthesis of atropoisomeric biaryls. Western blotting demonstrated that the newly developed compounds are promising targets in biology and pharmaceuticals. This unique reaction can construct structurally diverse C–N axial biaryl compounds that have never been reported by other methods, and might be extended to various applications in materials, chemistry, biology, and pharmaceuticals.


Introduction
Biaryl compounds that contain an axis and two aryl rings are common chemicals. These structural scaffolds, such as binaphthyls, biindoles, and naphthylindoles, show diverse physical, chemical, and biological properties, 1-8 thus, they are prevalent in natural products, 2 pharmaceuticals, 3 agrochemicals, 4 functional materials, 5 and dyestuffs, 6 as well as ligands 7 and synthetic reagents. 8 In fact, a search in SCI Finder indicates that there are more than 10 000 references with regard to this kind of compound.
Correspondingly, numerous methods for their preparation have been developed in recent decades. 9 However, these methods are generally based on the cross-coupling reaction between two polyaryls (Scheme 1a), such as two naphthalenes, 10 two indoles 11 and a naphthalene with an indole, 12 and the intramolecular cyclization of polyaryl alkynes (Scheme 1b), 13 in which an axis and an aromatic ring are constructed, respectively. Recently, a more step economical synthesis has been developed by the cross cyclization of prefunctionalized polyaryls with aryl alkynes which constructs both an axis and an aromatic ring (Scheme 1c). 14 Despite notable advances, challenges and tediousness in the preparation of the Scheme 1 Strategies to construct biaryl compounds.
prefunctionalized precursors, along with limited substrate scope, have seriously restricted the applications of these methods.
For example, 1,1 0 -naphthylindoles, containing C-N bondfused ve-six rings, are one type of biaryl compounds. It seems that 1,1 0 -naphthylindoles can be easily synthesized by transition-metal-catalyzed C-N coupling between indoles and naphthalenes. Due to limited substrate availability, there are fewer than 100 entries for 1,1 0 -naphthylindoles available by a search in SCI Finder. Among the limited reported compounds, 35 examples of this type of compound exhibit a broad spectrum of bioactivities, 15 such as anti-inammatory, antidiabetic, receptor antagonist, and antigout activities (Fig. 1), demonstrating that this kind of compound is a promising target in chemistry, biology, and pharmaceuticals.
Therefore, facile, efficient and general synthesis of biaryl compounds directly from easily available substrates are highly desirable, which could not only enrich such valuable compounds, but also provide a window for improving the properties and discovering new functions for this kind of compound, and thus may promote the development of other elds, such as chemistry, biology, pharmaceuticals, and materials science.
Undoubtedly, considering substrate availability and step economy, condensation/bicycloaromatization between two internal alkynes is the most direct and synthetically powerful approach towards biaryl compounds (Scheme 1d). However, this strategy is extremely challenging because it requires not only selective formation of a series of C(sp 2 )-X(sp 2 ) (X ¼ C and N) bonds to construct an axis and two aromatic rings, but also suppression of many types of easily occurring undesired reactions, such as addition, dimerization and intramolecular cyclization of the two different aryl alkynes. 16 Indeed, it has not been achieved.
Herein, with the continuous interest in the selective transformation of alkynes, 17 we develop a condensation/ bicycloaromatization reaction of easily available o-amino and ocarbonyl arylalkynes under ligand-free copper(I) catalysis, which successfully suppresses many easily occurring undesired reactions and produces diverse functionalized 1,1 0 -naphthylindoles in high to excellent yields with excellent selectivity (Scheme 2). In this tandem reaction, the C-N axis is formed by the condensation of the amino and carbonyl groups from two different arylalkynes. Two different aromatic rings, i.e., naphthalene and indole, are exclusively constructed via tandem 6endo-dig carbocyclization and 5-endo-dig nitrocyclization, respectively, without observation of any other C-N or C-C axial biaryl compounds. In addition to naphthylindoles, this method is also applicable to construct axial heterobiaryl scaffolds such as benzothiophenelindoles, quinolinelindoles, benzofuranindoles, benzothiophenyl-pyrrolopyridine, naphthyl-pyrrolopyridine, and benzofuranyl-pyrrolopyridine, which are not or hardly synthesized by other methods. Moreover, a very broad range of functional groups are readily incorporated into the C-N axial biaryl compounds due to the availability of substrates and very mild reaction conditions. Western blotting has demonstrated that functionalized 1,1 0 -naphthylindoles have strong anti-inammatory bioactivity. Thus, this unique condensation/ bicycloaromatization represents an ideal strategy for the synthesis of functionalized C-N axial biaryl compounds.
(2) Low reactivity of the condensation  between arylamino and arylcarbonyl groups. The relatively weak nucleophilicity of the N-atom of arylamine disfavors its attack at the carbonyl group; thus, other undesired reactions preferentially occur. (3) Competitive reactions of o-amino arylalkynes with o-carbonyl arylalkynes. Other types of condensation and condensation/cyclizations between the two different alkynes may take place, which reduce the cross selectivity. For example, three types of annulation reactions oen occur for the two alkynes, i.e., 5-endo-dig, 24 5-exo-dig, 23 and 6-endo-dig cyclization. 23,26 Therefore, selectively achieving 1,1 0 -naphthylindoles by this strategy is extremely challenging.
We realized this strategy enlightened by our previous work of Cu(I)-catalyzed 6-endo-dig cyclization of terminal alkynes, 2bromoaryl ketones, and amides for the synthesis of 1-naphthylamines, 17a in which the formation of the enamine intermediate favors the carbocyclization of the C-C triple bond. We envisioned that an enamine intermediate might be produced by condensation between two arylalkynes substituted with an amino and carbonyl under suitable conditions, which may trigger 6-endo-dig carbocyclization of carbonyl alkyne and 5endo-dig nitrocyclization of amino alkyne, and the above undesired reactions could be suppressed. Thus, C-N axial biaryl compounds would be selectively produced.
The Cu(I) catalyst and HFIP are the most pertinent to the successful realization of the transformation. The reaction parameters are shown in Table 1 and summarized as follows: (1) the choice of Cu(I) catalyst was critical to this condensation/ bicycloaromatization reaction (Table 1, entries 1-4). Cu(I) salt could activate the ketone by the interaction of Cu(I) with its Oatom, which improves its electrophilicity and facilitates the addition of amino to carbonyl groups, which favors the condensation between two different alkynes. Although Cu(I) could also catalysze the cyclization reactions of the two alkynes, respectively (see ESI, Scheme S3 †), the condensation reaction preferentially occurred in the reaction system (the intermediate 1e 0 was formed in 10 min, vide infra). Among Cu(I) catalysts, CuI exhibited the best efficiency (Table 1, entry 4). In comparison, Cu(II) and Pd(II) catalysts have much stronger metal-p interactions with C-C triple bond than Cu(I) catalysts, which strongly activate the C-C triple bond and lead to the preferential occurrence of many undesired reactions of the substrates ( Table  1, entries 5-8), such as hydration of 1a, intramolecular cyclization of 1a and 1b, and di/polymerizations of 1a and 1b. For example, when CuCl 2 and PdCl 2 were used as the catalysts, more than 10 products were detected, and the desired product was only observed in 8% and 6% yields, respectively (Table 1, entries 5 and 8; for details, see the ESI, Table S1, † entries 5 and 9). Overall, the Cu(I) salt could catalyze the condensation between two amino alkynes and ketone alkynes, 6-endo-dig carbocyclization and 5-endo-dig nitrocyclization, avoiding the occurrence of the undesired reactions of the substrates. (2) Fluorinated alcohols are essential for the reaction as solvents (Table 1, entries 4, 9 and 10). Remarkably, the yield of the desired product was increased to 86% when HFIP was employed as the solvent (  a Reaction conditions: unless otherwise stated, all reactions were performed with 1a (0.05 mmol), 1b (1.0 equiv., 0.05 mmol), catalyst (10 mol%), and additive (2.0 equiv., 0.1 mmol) in the solvent (0.5 mL) at 100 C for 24 h under N 2 atmosphere. b GC yield based on 1a using dodecane as an internal standard. c 1a (0.05 mmol), 1b (0.06 mmol). d HFIP (1.0 mL). e HFIP (1.5 mL). TFE (2,2,2-triuoroethanol), PFTB (peruoro-tert-butyl alcohol), IPA (iso-propyl alcohol), TBA (tert-butyl alcohol). 120 C, slightly reduced the reaction selectivity, and an 82% yield of 1c was observed (Table 1, entry 20). (4) The addition of Na 2 CO 3 (84%), K 3 PO 4 (87%), t-BuOK (88%) and Cs 2 CO 3 (91%) slightly improved the yield of the desired product. In comparison, the reaction failed with acidic additives, such as TsOH and acetic acid, and other reactions, such as hydration, dimerization, and intramolecular cyclization of the substrates, predominantly occurred (Table 1, entries 25 and 26; see the ESI,  Table S5 †). Finally, by improving the ratio of 1a and 1b to 1 : 1.2, a 94% yield of 1c was obtained with excellent selectivity (Table 1, entry 27). It was noted that any of the other products of the cross-coupling and cross-coupling/bicycloaromatization between 1a and 1b were not detected under the optimal reaction conditions, which also strongly facilitated the production isolation.
Next, the scope of o-acetylphenyl alkynes was investigated (Scheme 3B). All three kinds of o-acetylphenyl alkynes (o-acetylphenyl aryl alkyne, o-acetylphenyl alkyl alkyne, and o-acetylphenyl heteroaryl alkyne) exhibited high reactivity in the reaction, and a variety of N-(3-aryl, -alkyl, and -heteroaryl) naphthyl indoles (29c-43c) were readily furnished in high to excellent yields (72-93%). Similar compatibility of functional groups to that of o-aminophenyl alkynes (1c-28c) was observed, as well as electronic and steric effects. The successful incorporation of reactive hydroxyl (44c, 59% yield) and alkenyl (45c, 75% yield) groups into the desired products further veried the outstanding functional group tolerance of the reaction.
In addition, substrates with multiple substituents on the oaminophenyl and o-acetylphenyl rings also worked well (Scheme 3C), delivering the desired naphthylindoles in 78-91% yields (46c-59c). This feature makes the reaction practical for the synthesis of polysubstituted naphthylindoles, which cannot be achieved by other methods. Notably, o-aminophenyl alkynes also reacted smoothly with an o-propionylphenyl alkyne, producing naphthylindoles in 72-88% yields (60c-63c) with incorporation methyl group into 2-position of naphthyl ring.
major product but takes dynamic equilibrium with the substrates. The reaction of 1b with aniline (1a 0 ) instead of 1a gave 1-naphthylamine 1e in 90% yield (Scheme 5, eqn (2)). This result demonstrated that 6-endo-dig carbocyclization of 1b with arylamines via the formation of the enamine intermediate could proceed smoothly. The above results also suggested that imine, formed from the condensation of amino and carbonyl groups, could tautomerize to the enamine form, which triggers nucleophilic attack to the adjacent C-C triple bond to form 1naphthylamine. Compound 1e 0 was subjected to intramolecular cyclization to produce 1,1 0 -naphthylindole (2c) in almost quantitative yield (99%, Scheme 5, eqn (3)). However, the treatment of 1b with 2-phenylindole (2f) did not give the desired product (Scheme 5, eqn (4)). The above results indicated that 6endo-dig carbocyclization occurred prior to 5-endo-dig nitrocyclization. Indeed, intermediate 1e 0 could be isolated at the initial stage (10 min) of the reaction of 2a with 1b, which further conrmed the suggestion (ESI, Scheme S5, † eqn (F)). Next, deuterium labeling experiments were performed (Scheme 5, eqn (5), for details, see the ESI, Schemes S7-S10 †). The reaction was carried out by replacing HFIP with (CF 3 ) 2 -CDOD (HFIP-d 2 ) under standard conditions, and the corresponding deuterated product (2c-d, 24 h) was obtained in 92% yield with deuterium incorporation at the C-2 (85% deuteration) and C-4 (91% deuteration) positions on the naphthalene ring and the C-3 position (83% deuteration, 90% deuteration in situ) on the indole ring (Scheme S7, † eqn (K), Fig. S41 †). The different D contents suggested that they were derived from different pathways. It was also conrmed that the H/D exchange did not occur between the solvent (HFIP-d 2 ) and the C-H bond of the naphthalene ring in intermediate E and product c (ESI, Scheme S10, † eqn (S)-(V)). This result suggested that the D atom of C-2 and C-4 positions of naphthalene ring in product 2c-d did not directly derived from solvent by H/D exchange between the solvent with product 2c and intermediate 1e 0 .
As shown in eqn (5) of Scheme 5, the D content of C2 and C4 on the naphthalene ring of intermediate 1e 0 -d (isolated at 10 min, ESI, Scheme S7, † eqn (K)) are 80% and 92%, respectively. The D content of C4 on the naphthalene ring of intermediate 1e 0 -d (92%) and product 2c-d (91%) were close to that of the reaction system (deuterium content of the total reactive H and D that could be exchanged, 89.5-93.0%), which suggested that the D atom of C-4 position on the naphthalene ring was derived from the reaction system via protonation of vinyl copper intermediate D (vide infra).
The D content of C-2 position on the naphthalene ring of intermediate 1e 0 -d (80%, 10 min) and product 2c-d (85%) were lower than those of C4 (92% and 91%) and the reaction system, but higher than that of substrate 1b-d (44% D, 10 min, Scheme S7, † eqn (K)). These results suggested that the D atom of C-2 was derived from two pathways: (1) substrate 1b-d by H/D exchange of substrate 1b with the reaction system; (2) direct H/D exchange during condensation and 6-endo-dig carbocyclization (A to D, vide infra). The in situ deuterium content (90%) of the C-3 position on the indole ring of 2c-d was also close to that of the reaction system (89.5-93.0%), which was much higher than D content of substrate 2a-d (about 30%, see ESI, Scheme S7, † eqn (K)). Direct cyclization of intermediate 1e 0 using HFIP-d as solvent (ESI, Scheme S10, † eqn (S)) showed that the D content of the C-3 position on the indole ring of the product was about 90% from 10 min to 24 h, which was also close to the reaction system (89.5-93.0%). Whereas, the D content of the N-H of the 1e 0 -d were ca. 31% D and 44% D at 10 min and 1 h, respectively. These results suggested D atom of C3 mainly derived from the protonation of vinyl copper intermediate F with the reaction system.
On the basis of the above results and literature reports, 17a,29 a tentative mechanism for the selective condensation/ bicycloaromatization reaction is proposed (Scheme 6). First, the interaction of Cu(I) with the O-atom of the carbonyl of oacetylphenyl alkyne b forms intermediate A, and then the nucleophilic attack of the N-atom of o-aminophenyl alkyne a on its carbonyl group leads to the formation of a-amino alcohol B, followed by dehydration to produce imine intermediate C 0 , then the Cu(I) promotes tautomerization of imine C 0 to enamine C. 29 Intramolecular 6-endo-dig carbocyclization of C gives intermediate D, which simultaneously proceeds with deprotonation and protonation to produce (N-naphthyl)aminophenyl alkyne E. Intermediate E undergoes 5-endo-dig nitrocyclization with subsequent protonation to produce desired product c.

Bioactivity properties
Finally, to evaluate the bioactivity of this new type of naphthylindole, we initially tested the potential cytotoxicity of compounds 15c (A1), 44c, and 69c toward BV-2 cells by the MTT assay. Compounds 15c, 44c, and 69c at the indicated concentrations (0, 3, 10, 30, and 100 mM) showed no signicant cytotoxicity on BV-2 cells (Fig. S46 †). Then, we detected the effect of compound 15c on the expression of phosphorylated NF-kB (p-P65) and IL-6 in BV-2 microglial cells by western blotting. BV-2 microglial cells were pretreated with 10 mM A1 (15c) for 2 h before being stimulated with 100 ng mL À1 LPS for another 6 h. A1 (15c) attenuated LPS-induced NF-kB activation in BV-2 microglial cells (LPS 100 ng mL À1 ) compared with the control group, **P < 0.01, vs. the LPS group, *P < 0.05 (Fig. 2a). Moreover, we found that A1 (15c) also attenuated LPS-induced IL-6 production in BV-2 microglial cells in a concentrationdependent manner (Fig. 2b). In view of the above results, we speculate that compound 15c has anti-inammatory bioactivity. Indeed, compound 15c has comparable anti-inammatory effect to indomethacin (an indole derivative, non-steroidal anti-inammatory drug) on the expression of 1L-1b and TNFa in BV-2 microglial cells (see ESI, Fig. S47 †). These results demonstrate that this method has great potential application in drug and biology.

Conclusions
We have developed a condensation/bicycloaromatization reaction of two different arylalkynes under ligand-free copper catalysis for simple, efficient, and step-economic synthesis of Scheme 7 Substrate scope of axially chiral compounds. Reaction conditions: (1) a (0.2 mmol), 4b (1.2 equiv.), CuI (5 mol%), Cs 2 CO 3 (2.0 equiv.) HFIP, 80 C, 50 min; (2) PdCl 2 (5 mol%), L (6 mol%) HFIP, 80 C, 15 h. C-N axial biaryl compounds. Due to the critical effects of Cu(I) catalyst and HFIP, many easily occurring undesired reactions are suppressed, and the coupled ve-six aromatic rings are constructed via the selective formation of two C(sp 2 )-N(sp 2 ) bonds and four C(sp 2 )-C(sp 2 ) bonds. The formation of the enamine intermediate between the two different alkynes might be the key step for the reaction, which triggers selective tandem 6-endo-dig carbocyclization and 5-endo-dig nitrocyclization. Many types of biaryl compounds, such as naphthylindole, benzothio-phenelindole, quinolinelindole, benzofuranindole, benzothiophenyl-pyrrolopyridine, naphthyl-pyrrolopyridine, and benzofuranyl-pyrrolopyridine, are obtained in high yields (up to 95%), which are not or rarely synthesized by other methods. The compatibility of a very broad range of functional groups, such as alkyl, halogens (F, Cl, Br, and I), OMe, OH, vinyl, ester, NO 2 , CF 3 , CN, N,O-heterocycle, aryl, heteroaryl (pyridine, thiophene, furan, ferrocene, and carbazole), and even reactive functional groups [i.e., acetyl and alkynyl], demonstrates that structurally and functionally diverse C-N axial biaryl compounds can be easily prepared from simple starting materials. The realization of moderate enantioselectivity of chiral naphthylindoles veries its potential for the most direct asymmetric synthesis of atropoisomeric biaryls. Western blotting experiments demonstrated that these newly developed 1,1 0naphthylindole derivatives, such as 15c, have strong anti-inammatory bioactivity and are promising targets in biology and pharmaceuticals. In summary, this unique reaction can prepare structurally diverse C-N axial biaryl compounds that have not been reported by other methods, represents an ideal strategy for the synthesis of functionalized C-N axial biaryl compounds, and may be highly useful in materials, chemistry, biology, and pharmaceuticals. Further work is in progress to exploit the related bioactivity properties of C-N axial biaryls and improve the enantioselectivity of this transformation.

Author contributions
Y. Z. and J. D. conceived the work and coordinated the project. Q. S. conducted the experiments. J. D. initiated the project. Q. S. advanced and completed the project. Y. L., M. Y., and Y. C. carried out the western blotting experiment. W. Y. and Q. S. carried out asymmetric synthesis. H. T., L. S., S. X., and L. L. assisted the project. Y. Z. wrote the manuscript with Q. S., J. D., and S. Y.

Conflicts of interest
The authors declare that there are no conicts to declare.