Nickel(II)-Catalyzed Formal [3+2] Cycloadditions between Indoles and Donor–Acceptor Cyclopropanes

This article describes the development of a nickel-catalyzed regio- and diastereoselective formal [3+2] cycloaddition between N-substituted indoles and donor–acceptor cyclopropanes to synthesize cyclopenta[b]indoles. Optimized reaction conditions provide the desired nitrogen-containing cycloadducts in up to 93% yield and dr 8.6:1 with complete regioselectivity. The substrate scope showed high tolerance to various substituted indoles and cyclopropanes, resulting in the synthesis of six new cyclopenta[b]indoles and the isolation of five derivatives previously reported in the literature. In addition, a mechanistic proposal for the reaction was studied through online reaction monitoring by ESI-MS, allowing for the identification of the reactive intermediates in the Ni(II) catalyzed process. X-ray crystallography confirmed the structure and relative endo stereochemistry of the products. This method enables the fast and efficient construction of fused indolines from readily accessible starting materials.

In this regard, the Kerr group pioneered the synthesis of this skeleton through cycloaddition of donor-acceptor cyclopropanes and 3-methylindoles as a two synthon catalyzed by Yb(OTf)3 (Scheme 1A) [45][46][47].Subsequently, in 2013, Xie an reported the use of Cu(II) catalysts (Scheme 1A), in combination with BOX-typ ligands, for an enantioselective version of this cyclization [48].Recently, during th opment of this project, Doyle's group reported the use of 3-alkylindoles as an int strategy for the construction of carbocyclic systems fused to indoles via a Ni(II) ca stereoretentive [3+2] cycloaddition, followed by a one-pot Sc(III)-catalyzed decar tive ring expansion to afford the corresponding cyclopenta[b]indoles and dihydrobazoles (Scheme 1B) [49].In the initial stage catalyzed by Ni(II), the authors desc clopenta[b]indole formation via [3+2] cycloaddition with good retention of the chi the starting product, but with a low diastereoselectivity, the inconvenience of w elegantly overcome in a subsequent rearrangement/decarboxylation stage cataly Sc(OTf)3 [49].On the other hand, other research groups have explored the use of 3doles as more reactive species in [3+2] cycloadditions, using vinylcyclopropanes in Pd(0)-catalyzed processes to obtain cyclopenta[b]indole systems (Scheme 1C) Following a similar strategy, but employing phenylaziridines or oxiranyl-dicarbo instead of VCPs, other authors have described the efficient synthesis of pyrrolo[2 dole and furo [3,4-b]indole scaffolds [54][55][56].These cyclopropanes can be activat variety of Lewis acids, including those based on non-precious metals that are hig sirable for their abundance and cost-effectiveness [57][58][59][60].
In this context, and despite the precedents for the use of non-precious metals t cyclopenta [b]indole skeletons from skatole derivatives, this field is still underdev Therefore, starting from the general concept of opening metal-catalyzed cycloprop rivatives, the main objective of this work is the application of non-precious-meta catalysts in the diastereoselective synthesis of cyclopenta[b]indole-type skeletons gle synthetic stage by opening donor-acceptor cyclopropanes and guiding thei quent formal [3+2] cycloaddition with C3-substituted indoles (Scheme 1D).In this context, intermolecular [3+2] cycloadditions have emerged as powerful for the construction of five-membered rings in a single synthetic step.Significant contributions to this field have been made through catalytic reactions of vinyldiazo compounds [29,35], azomethine ylides [36], and trimethylenemethane (TMM) [37] and methodologies employing donor-acceptor cyclopropanes [20].This latter strategy is particularly appealing because it would provide a straightforward and atom-economical strategy for gaining access to cyclopenta [b]indole cores [38][39][40][41][42][43][44].
In this context, and despite the precedents for the use of non-precious metals to access cyclopenta[b]indole skeletons from skatole derivatives, this field is still underdeveloped.Therefore, starting from the general concept of opening metal-catalyzed cyclopropane derivatives, the main objective of this work is the application of non-precious-metal-based catalysts in the diastereoselective synthesis of cyclopenta[b]indole-type skeletons in a single synthetic stage by opening donor-acceptor cyclopropanes and guiding their subsequent formal [3+2] cycloaddition with C3-substituted indoles (Scheme 1D).

Results and Discussion
We started our study by employing different Lewis acids based on non-precious metals, using N-benzylskatole (1a) and cyclopropane 2a as model substrates.As an initial control experiment, we attempted to reproduce the reported reaction conditions when employing Cu(OTf)2 as a catalyst [48], affording the desired cycloadduct 3aa at a 79% yield and a diastereoismeric ratio (dr) of 5:1, consistent with the literature reports (Table 1, entry 1) [48].Next, the reactivity of different metal salts was evaluated.The use of Co(ClO4)2•6H2O led to 42% 3aa with a modest increase in diastereoselectivity (dr = 6:1, Table 1, entry 2).On the other hand, Zn(ClO4)2•6H2O showed a similar enhancement of diastereoselectivity (dr = 6.4:1) but with a significant decrease in yield (28%, Table 1, entry 3).Interestingly, the use of Ni(ClO4)2•6H2O as the catalyst led to 3aa in an excellent 95% yield and with a 6:1 ratio of isomers (Table 1, entry 4).This diastereomeric ratio is considerably higher than the 1:1 ratio described by the Doyle group under similar reaction conditions [49].Furthermore, the use of Ni(II) salts, such as Ni(ClO4)2•6H2O, in cycloaddition reactions of donor-acceptor cyclopropanes has been previously described in the literature [56,[61][62][63], which supports the observed results.Other metal salts of Mn 2+ , Fe 2+ , Fe 3+ , and Al 3+ were also evaluated (see Table S1 in the Supplementary Materials), but they were inefficient in obtaining cyclopenta[b]indole 3aa.Based on these results, we continued our study by evaluating different Ni 2+ salts.When using a catalytic load of 10 mol%, Ni(ClO4)2•6H2O showed a significantly slower conversion, although it led to a 48% yield (Table 1, entry 5).On the other hand, when Ni(OAc)2•4H2O was employed, a negligible conversion and reaction yield were observed after 24 h (Table 1, entry 6).Other salts such as NiSO4•6H2O, Ni(NO3)2•6H2O, or NiCl2 were evaluated, providing the desired product in modest yields and incomplete conversions (Table 1, entries 7 to 9).Notably, the use of Ni(OTf)2 (Table 1, entry 10) resulted in a significant increase in the yield of the product, reaching 78%, with a diastereomeric ratio of 5:1.At this point, we evaluated the influence of different solvents on the reaction.The use of DCM led to a decrease in yield (47%) with a dr = 7.7:1 (Table 1, entry 11), while with CHCl3, it provided only a 25% yield with an important decrease in diastereoselectivity (Table 1, entry 12).Given the reports of Tang and Doyle [48,49], we also evaluated the use of toluene as a solvent, observing a poor 29% yield and no diastereoselectivity (Table 1, entry 13).Acetonitrile and ethyl acetate were also evaluated, although the starting material was fully recovered in both cases (Table 1, entries 14 and 15).To consult the other solvents evaluated, see Table S2 in the Supplementary Materials.

Results and Discussion
We started our study by employing different Lewis acids based on non-precious metals, using N-benzylskatole (1a) and cyclopropane 2a as model substrates.As an initial control experiment, we attempted to reproduce the reported reaction conditions when employing Cu(OTf) 2 as a catalyst [48], affording the desired cycloadduct 3aa at a 79% yield and a diastereoismeric ratio (dr) of 5:1, consistent with the literature reports (Table 1, entry 1) [48].Next, the reactivity of different metal salts was evaluated.The use of Co(ClO 4 ) 2 •6H 2 O led to 42% 3aa with a modest increase in diastereoselectivity (dr = 6:1, Table 1, entry 2).On the other hand, Zn(ClO 4 ) 2 •6H 2 O showed a similar enhancement of diastereoselectivity (dr = 6.4:1) but with a significant decrease in yield (28%, Table 1, entry 3).Interestingly, the use of Ni(ClO 4 ) 2 •6H 2 O as the catalyst led to 3aa in an excellent 95% yield and with a 6:1 ratio of isomers (Table 1, entry 4).This diastereomeric ratio is considerably higher than the 1:1 ratio described by the Doyle group under similar reaction conditions [49].Furthermore, the use of Ni(II) salts, such as Ni(ClO 4 ) 2 •6H 2 O, in cycloaddition reactions of donor-acceptor cyclopropanes has been previously described in the literature [56,[61][62][63], which supports the observed results.Other metal salts of Mn 2+ , Fe 2+ , Fe 3+ , and Al 3+ were also evaluated (see Table S1 in the Supplementary Materials), but they were inefficient in obtaining cyclopenta[b]indole 3aa.Based on these results, we continued our study by evaluating different Ni 2+ salts.When using a catalytic load of 10 mol%, Ni(ClO 4 ) 2 •6H 2 O showed a significantly slower conversion, although it led to a 48% yield (Table 1, entry 5).On the other hand, when Ni(OAc) 2 •4H 2 O was employed, a negligible conversion and reaction yield were observed after 24 h (Table 1, entry 6).Other salts such as NiSO 4 •6H 2 O, Ni(NO 3 ) 2 •6H 2 O, or NiCl 2 were evaluated, providing the desired product in modest yields and incomplete conversions (Table 1, entries 7 to 9).Notably, the use of Ni(OTf) 2 (Table 1, entry 10) resulted in a significant increase in the yield of the product, reaching 78%, with a diastereomeric ratio of 5:1.At this point, we evaluated the influence of different solvents on the reaction.The use of DCM led to a decrease in yield (47%) with a dr = 7.7:1 (Table 1, entry 11), while with CHCl 3 , it provided only a 25% yield with an important decrease in diastereoselectivity (Table 1, entry 12).Given the reports of Tang and Doyle [48,49], we also evaluated the use of toluene as a solvent, observing a poor 29% yield and no diastereoselectivity (Table 1, entry 13).Acetonitrile and ethyl acetate were also evaluated, although the starting material was fully recovered in both cases (Table 1, entries 14 and 15).To consult the other solvents evaluated, see Table S2 in the Supplementary Materials. 1 Reaction conditions: Indole 1a (1.0 eq.) and cyclopropane 2a (1.2 eq.), was treated with [M] (X mol%) in the indicated solvent (0.1 M) at rt and time (h). 2 Based on recovery of starting material. 3 Isolated yield. 4 Calculated from the 1 H NMR. 5 24 h rt, and then 24 h at 50 °C.
Having identified Ni(OTf)2 as the most suitable promoter of the [3+2] cycloaddition, we initially evaluated different substituted indoles.Unfortunately, we observed the formation of the desired products with low yields and diastereoselectivities (Table S3 in the Supplementary Materials).For this reason, we decided to assess the influence of different ligands commonly employed in this type of reaction, including bidentate nitrogenous ligands, bisphosphines, and N,N'-dioxides (Table 2) [56,64,65].We started by evaluating nitrogen-and phosphorous-based bidentate ligands such as Bipy, Phen, dppe, d F ppe, L2-PrPr2, and L3-PrPr2 (Table 2, entries 1 to 6), but unfortunately, none of these catalysts improved the results observed with Ni(OTf)2 (see Table 1, entry 10).On this basis, we decided to reevaluate Ni(ClO4)2•6H2O, considering that it proved to be a competent catalyst (Table 1, entry 4) and taking advantage of the labile nature of the perchlorate ion to facilitate the formation of cationic metal-ligand complexes [66].With the use of bidentate ligands such as Phen or dppe, the formation of compound 3aa was observed with low yields (Table 2, entries 7 and 8).In contrast, bisphosphines such as dfppe and dppBz afforded the product in modest 43% and 53% yields, respectively (Table 2, entries 9 and 10), while with dppf, no reactivity was observed (Table 2, entry 11).Gratifyingly, the evaluation of rac-BINAP [67,68] resulted in a significant acceleration of the reaction, providing the desired 3aa product in a 73% yield with a diastereoisomeric ratio of 6:1 after 5 h at room temperature (Table 2, entry 12).For more details on ligand optimization, see Table S4 in the Supplementary Materials.The effect of temperature was also evaluated, with us observing that at both 50 and 80 °C, there was a decrease in the isolated yield of 3aa, possibly due to the degradation of the starting materials, along with a decrease in the diastereoisomeric ratio ( 1 Reaction conditions: Indole 1a (1.0 eq.) and cyclopropane 2a (1.2 eq.), was treated with [M] (X mol%) in the indicated solvent (0.1 M) at rt and time (h). 2 Based on recovery of starting material. 3Isolated yield. 4 Calculated from the 1 H NMR. 5 24 h rt, and then 24 h at 50 • C.
Having identified Ni(OTf) 2 as the most suitable promoter of the [3+2] cycloaddition, we initially evaluated different substituted indoles.Unfortunately, we observed the formation of the desired products with low yields and diastereoselectivities (Table S3 in the Supplementary Materials).For this reason, we decided to assess the influence of different ligands commonly employed in this type of reaction, including bidentate nitrogenous ligands, bisphosphines, and N,N'-dioxides (Table 2) [56,64,65].We started by evaluating nitrogen-and phosphorous-based bidentate ligands such as Bipy, Phen, dppe, d F ppe, L 2 -PrPr 2 , and L 3 -PrPr 2 (Table 2, entries 1 to 6), but unfortunately, none of these catalysts improved the results observed with Ni(OTf) 2 (see Table 1, entry 10).On this basis, we decided to reevaluate Ni(ClO 4 ) 2 •6H 2 O, considering that it proved to be a competent catalyst (Table 1, entry 4) and taking advantage of the labile nature of the perchlorate ion to facilitate the formation of cationic metal-ligand complexes [66].With the use of bidentate ligands such as Phen or dppe, the formation of compound 3aa was observed with low yields (Table 2, entries 7 and 8).In contrast, bisphosphines such as dfppe and dppBz afforded the product in modest 43% and 53% yields, respectively (Table 2, entries 9 and 10), while with dppf, no reactivity was observed (Table 2, entry 11).Gratifyingly, the evaluation of rac-BINAP [67,68] resulted in a significant acceleration of the reaction, providing the desired 3aa product in a 73% yield with a diastereoisomeric ratio of 6:1 after 5 h at room temperature (Table 2, entry 12).For more details on ligand optimization, see Table S4 in the Supplementary Materials.The effect of temperature was also evaluated, with us observing that at both 50 and 80 • C, there was a decrease in the isolated yield of 3aa, possibly due to the degradation of the starting materials, along with a decrease in the diastereoisomeric ratio (Table 2, entries 13 and 14). 1 Reaction conditions: Indole 1 (1.0 eq.) and cyclopropane 2a (1.2 eq.), was treated with [Ni] (10 mol%) and ligand (10 mol%) in DCE (0.1 M) at rt and time (h). 2 Based on recovery of starting material. 3Isolated yield. 4 Calculated from the 1 H NMR. 5 Carried out at 85 °C. 6Carried out at 50 °C.
The results listed in Table 2 suggest that the rigidity and bite angle of bidentate phosphines might be crucial factors in the system's reactivity.Ligands such as dppe (flexible ligand) and dppBz (rigid ligand) [69], with approximately 90° angles, exhibited a reduced yield and a significant increase in reaction time compared to the BINAP version, which features a greater bite angle (approximately 93°) than the previously mentioned ligands [70].Ligands with larger angles, such as dppf (99°) [71], also fail to provide ideal conditions for this [3+2] cycloaddition.We primarily evaluated a possible enantioselective version of the cycloaddition using chiral bisphosphines such as (R)-BINAP, (R)-DM-BINAP, and (S)-DTBM-DEGPHOS.Among these, only the BINAP derivatives effectively catalyzed the reaction, yielding up to 81%, but not exceeding 40% ee in any case (see details in Table S5 in the Supplementary Materials).Further studies to develop a highly enantioselective version are under development.
Then, with optimized conditions in hand, we investigated the scope of the reaction by varying the substitution patterns of both substrates (Scheme 2).Regarding cyclopropanes, diverse functional groups on the aryl ring, such as OMe, H, Cl, Br, F, and NO2, were well-tolerated, affording the corresponding products with yields ranging from 51% to 90%, although with lower diastereomer selectivity (3aa-3ag).In particular, electron-donor groups led to shorter reaction times (2 to 3 h), while strongly attracting substituents like the nitro group significantly extended the reaction time (5 days).However, the use of a cyclopropane bearing a 2,4-dichloro-substituted arene did not lead to the corresponding cycloadduct 3af, with us instead recovering the starting material.This result may be explained by steric hindrance having occurred between cyclopropane and the catalyst, rather than there being an electronic effect.Importantly, we also evaluated a   1 Reaction conditions: Indole 1 (1.0 eq.) and cyclopropane 2a (1.2 eq.), was treated with [Ni] (10 mol%) and ligand (10 mol%) in DCE (0.1 M) at rt and time (h). 2 Based on recovery of starting material. 3Isolated yield. 4 Calculated from the 1 H NMR. 5 Carried out at 85 °C. 6Carried out at 50 °C.

Entry
The results listed in Table 2 suggest that the rigidity and bite angle of bidentate phosphines might be crucial factors in the system's reactivity.Ligands such as dppe (flexible ligand) and dppBz (rigid ligand) [69], with approximately 90° angles, exhibited a reduced yield and a significant increase in reaction time compared to the BINAP version, which features a greater bite angle (approximately 93°) than the previously mentioned ligands [70].Ligands with larger angles, such as dppf (99°) [71], also fail to provide ideal conditions for this [3+2] cycloaddition.We primarily evaluated a possible enantioselective version of the cycloaddition using chiral bisphosphines such as (R)-BINAP, (R)-DM-BINAP, and (S)-DTBM-DEGPHOS.Among these, only the BINAP derivatives effectively catalyzed the reaction, yielding up to 81%, but not exceeding 40% ee in any case (see details in Table S5 in the Supplementary Materials).Further studies to develop a highly enantioselective version are under development.
Then, with optimized conditions in hand, we investigated the scope of the reaction by varying the substitution patterns of both substrates (Scheme 2).Regarding cyclopropanes, diverse functional groups on the aryl ring, such as OMe, H, Cl, Br, F, and NO2, were well-tolerated, affording the corresponding products with yields ranging from 51% to 90%, although with lower diastereomer selectivity (3aa-3ag).In particular, electron-donor groups led to shorter reaction times (2 to 3 h), while strongly attracting substituents like the nitro group significantly extended the reaction time (5 days).However, the use of a cyclopropane bearing a 2,4-dichloro-substituted arene did not lead to the corresponding cycloadduct 3af, with us instead recovering the starting material.This result may be explained by steric hindrance having occurred between cyclopropane and the catalyst, rather than there being an electronic effect.Importantly, we also evaluated a The results listed in Table 2 suggest that the rigidity and bite angle of bidentate phosphines might be crucial factors in the system's reactivity.Ligands such as dppe (flexible ligand) and dppBz (rigid ligand) [69], with approximately 90 • angles, exhibited a reduced yield and a significant increase in reaction time compared to the BINAP version, which features a greater bite angle (approximately 93 • ) than the previously mentioned ligands [70].Ligands with larger angles, such as dppf (99 • ) [71], also fail to provide ideal conditions for this [3+2] cycloaddition.We primarily evaluated a possible enantioselective version of the cycloaddition using chiral bisphosphines such as (R)-BINAP, (R)-DM-BINAP, and (S)-DTBM-DEGPHOS.Among these, only the BINAP derivatives effectively catalyzed the reaction, yielding up to 81%, but not exceeding 40% ee in any case (see details in Table S5 in the Supplementary Materials).Further studies to develop a highly enantioselective version are under development.
Then, with optimized conditions in hand, we investigated the scope of the reaction by varying the substitution patterns of both substrates (Scheme 2).Regarding cyclopropanes, diverse functional groups on the aryl ring, such as OMe, H, Cl, Br, F, and NO 2 , were well-tolerated, affording the corresponding products with yields ranging from 51% to 90%, although with lower diastereomer selectivity (3aa-3ag).In particular, electron-donor groups led to shorter reaction times (2 to 3 h), while strongly attracting substituents like the nitro group significantly extended the reaction time (5 days).However, the use of a cyclopropane bearing a 2,4-dichloro-substituted arene did not lead to the corresponding cycloadduct 3af, with us instead recovering the starting material.This result may be explained by steric hindrance having occurred between cyclopropane and the catalyst, rather than there being an electronic effect.Importantly, we also evaluated a vinylcyclopropane (VCP), a less stabilized 1,3-dipole system, widely used with palladium catalysis.After 5 days of reaction, only traces of the product 3ah were observed by mass spectrometry.Inspired by these results, we decided to evaluate a styrylcyclopropane derivative, a more resonance-stabilized VCP.Gratifyingly, this substrate provided cyclopenta[b]indole 3ai at a 93% yield with a dr of 1.7:1.Regarding the effect of the indolic ring, substitutions such as OMe, Cl, or Br in positions 5 and 6 were well-tolerated, affording the desired products with good yields and diastereoisomeric ratios of between 8.6:1 to 4:1, albeit with longer reaction times (3ca-3fa).In contrast, introducing a nitro group in C5, disubstituted C2,C3, or an electron-withdrawing group in the C3 position of the indole did not lead to the desired tricyclic compounds (3ga-3ja).Similarly, it was observed that the N-protection of the indole ring with attractor groups such as Ts, Boc, or Bz did not lead to the formation of the desired skeleton (3ka-3ma).These results are consistent with a decreased nucleophilicity in the indole ring due to the electronic effects of the abovementioned substituents.On the other hand, incorporating a methyl group at the C7 position of the ring did not provide the desired product (3ba), resulting in a complex mixture of products in all cases.This result, along with the case of double substitution in the phenyl ring of cyclopropane (3af), indicated that this cycloaddition is also sensitive to steric factors in both reaction components.Finally, we evaluated sktole (deprotected N) in the reaction, but the formation of 3na was not observed and the starting material was recovered.This observation indicated that this Ni(II)-based catalytic system differs from the reaction described by Doyle's group [49].The structure of all reaction products was completely characterized by spectroscopic and spectrometric techniques.Additionally, the structure and relative endo stereochemistry could be confirmed by X-ray analysis of the 3ac crystal (CCDC: 2302398) [72].
Molecules 2024, 29, x FOR PEER REVIEW 6 of 20 vinylcyclopropane (VCP), a less stabilized 1,3-dipole system, widely used with palladium catalysis.After 5 days of reaction, only traces of the product 3ah were observed by mass spectrometry.Inspired by these results, we decided to evaluate a styrylcyclopropane derivative, a more resonance-stabilized VCP.Gratifyingly, this substrate provided cyclopenta[b]indole 3ai at a 93% yield with a dr of 1.7:1.Regarding the effect of the indolic ring, substitutions such as OMe, Cl, or Br in positions 5 and 6 were well-tolerated, affording the desired products with good yields and diastereoisomeric ratios of between 8.6:1 to 4:1, albeit with longer reaction times (3ca-3fa).In contrast, introducing a nitro group in C5, disubstituted C2,C3, or an electron-withdrawing group in the C3 position of the indole did not lead to the desired tricyclic compounds (3ga-3ja).Similarly, it was observed that the N-protection of the indole ring with attractor groups such as Ts, Boc, or Bz did not lead to the formation of the desired skeleton (3ka-3ma).These results are consistent with a decreased nucleophilicity in the indole ring due to the electronic effects of the abovementioned substituents.On the other hand, incorporating a methyl group at the C7 position of the ring did not provide the desired product (3ba), resulting in a complex mixture of products in all cases.This result, along with the case of double substitution in the phenyl ring of cyclopropane (3af), indicated that this cycloaddition is also sensitive to steric factors in both reaction components.Finally, we evaluated sktole (deprotected N) in the reaction, but the formation of 3na was not observed and the starting material was recovered.This observation indicated that this Ni(II)-based catalytic system differs from the reaction described by Doyle's group [49].The structure of all reaction products was completely characterized by spectroscopic and spectrometric techniques.Additionally, the structure and relative endo stereochemistry could be confirmed by X-ray analysis of the 3ac crystal (CCDC: 2302398) [72].Based on the experimental results and previous reports of [3+2] cycloadditions between donor-acceptor cyclopropanes and indoles [48,49,55], a plausible stepwise Based on the experimental results and previous reports of [3+2] cycloadditions between donor-acceptor cyclopropanes and indoles [48,49,55], a plausible stepwise mechanism is proposed (Scheme 3).Based on the crystal structure of the cationic BINAP/Ni 2+ complex (A) [73,74] and the labile nature of the perchlorate ion for the formation of metal-ligand complexes [66], we propose that an exchange of the labile ligands of the complex occurs when it interacts with the 1,1-dicarbonyl system of the donor-acceptor cyclopropane 2. This exchange leads to the formation of intermediate B, which, in turn, stabilizes the formation of the polarized species C. In view of the steric demand, the nucleophilic attack of the C3 position of indole 1 on the sp 3 stabilized carbon of cyclopropane, which bears a cationic character (δ + ), leads to the formation of the intermediate D [75,76].Subsequently, an intramolecular cyclization occurs through the attack of the malonic anion on the iminium ion generating the indole.Finally, the decoordination of the obtained cyclopenta[b]indole 3aa regenerates the catalytic species A.
Molecules 2024, 29, x FOR PEER REVIEW 7 of 20 mechanism is proposed (Scheme 3).Based on the crystal structure of the cationic BINAP/Ni 2+ complex (A) [73,74] and the labile nature of the perchlorate ion for the formation of metal-ligand complexes [66], we propose that an exchange of the labile ligands of the complex occurs when it interacts with the 1,1-dicarbonyl system of the donor-acceptor cyclopropane 2. This exchange leads to the formation of intermediate B, which, in turn, stabilizes the formation of the polarized species C. In view of the steric demand, the nucleophilic attack of the C3 position of indole 1 on the sp 3 stabilized carbon of cyclopropane, which bears a cationic character (δ + ), leads to the formation of the intermediate D [75,76].Subsequently, an intramolecular cyclization occurs through the attack of the malonic anion on the iminium ion generating the indole.Finally, the decoordination of the obtained cyclopenta[b]indole 3aa regenerates the catalytic species A.
To gain further insight into the proposed mechanism, we performed online reaction monitoring by ESI-MS, a technique of great relevance in the advancement of organic catalysis [77][78][79], since molecules and transient intermediates [80] of high polarity and moiety complexity can be easily studied by mass spectrometry.ESI-MS has proven to be a remarkable "ion-fishing" technique, as it gently transfers preformed ions in solution directly to the gas phase.Thus, Figure 2 shows the screening of the reaction of 1a and 2a catalyzed with Ni(ClO4)2⋅6H2O (10 mol%) and rac-BINAP (10 mol%) in DCE by online reaction monitoring by ESI-MS.The goal of this study was to intercept the cationic species resulting from the Ni-catalyzed reaction using ESI-MS in the positive-ion mode (Scheme 3) [81][82][83].The main advantage of using ESI is its capacity to transfer ions to the gas phase without inducing unwanted side reactions, and the composition of ESI-transferred ions often reflects that in solution [77][78][79].The ESI-MS spectra collected for such a reaction are particularly clean and mechanistically enlightening.Shortly after four to five minutes of reaction of Ni(ClO4)2⋅6H2O and rac-BINAP, two cationic species directly related to the proposed catalytic cycle (Scheme 3) were detected as major ions (Figure 2a): [A + ClO4 − ] + Scheme 3. Proposed mechanism for formal [3+2] cycloaddition.
To gain further insight into the proposed mechanism, we performed online reaction monitoring by ESI-MS, a technique of great relevance in the advancement of organic catalysis [77][78][79], since molecules and transient intermediates [80] of high polarity and moiety complexity can be easily studied by mass spectrometry.ESI-MS has proven to be a remarkable "ion-fishing" technique, as it gently transfers preformed ions in solution directly to the gas phase.Thus, Figure 2 shows the screening of the reaction of 1a and 2a catalyzed with Ni(ClO 4 ) 2 •6H 2 O (10 mol%) and rac-BINAP (10 mol%) in DCE by online reaction monitoring by ESI-MS.The goal of this study was to intercept the cationic species resulting from the Ni-catalyzed reaction using ESI-MS in the positive-ion mode (Scheme 3) [81][82][83].The main advantage of using ESI is its capacity to transfer ions to the gas phase without inducing unwanted side reactions, and the composition of ESI-transferred ions often reflects that in solution [77][78][79].The ESI-MS spectra collected for such a reaction are particularly clean and mechanistically enlightening.Shortly after four to five minutes of reaction of Ni(ClO 4 ) 2 •6H 2 O and rac-BINAP, two cationic species directly related to the proposed catalytic cycle (Scheme 3) were detected as major ions (Figure 2a): [A + ClO 4 − ] + of m/z 779, and [A + Cl − ] + of m/z 715.After 10 min of reaction, B was observed as m/z 1059 (Figure 2b).The isotopic patterns of Ni species were in accordance with the calculated compounds.In several studies, it is proposed that cycloadditions with catalyst-activated donoracceptor cyclopropanes can occur via the SN2 or SN1 pathway [49,[84][85][86][87].However, according to the ESI-MS experiments and synthetic results in this work, the SN1 pathway could not be discarded.Therefore, C and D are undoubtedly present in the reaction mixture in the course of the reaction.This observation, made according to ESI-MS/MS experiments, suggests that intermediate D is the only species that produces the final product 3aa.In several studies, it is proposed that cycloadditions with catalyst-activated donoracceptor cyclopropanes can occur via the S N 2 or S N 1 pathway [49,[84][85][86][87].However, according to the ESI-MS experiments and synthetic results in this work, the S N 1 pathway could not be discarded.Therefore, C and D are undoubtedly present in the reaction mixture in the course of the reaction.This observation, made according to ESI-MS/MS experiments, suggests that intermediate D is the only species that produces the final product 3aa.

General Section
All the reactions were conducted in dry solvents under a N 2 atmosphere unless otherwise stated.All reagents used were obtained from commercial suppliers and used without further purification.The abbreviation "rt" refers to reactions carried out at approximately 25 • C. Reaction mixtures were stirred using Teflon-coated magnetic stirring bars.Reaction temperatures were maintained using Thermowatch-controlled silicone oil baths.The reactions were monitored by thin-layer chromatography (TLC), which was performed on silica gel Merck 60 F 254 , and the components were visualized by observation under UV light (254 and 365 nm) and/or by treating the plates with p-anisaldehyde, oleum, phosphomolybdic acid, or cerium nitrate solutions, followed by heating.Flash chromatography was carried out on silica gel (63-200 µm) unless otherwise stated.Drying was performed with anhydrous Na 2 SO 4 .Concentration refers to the removal of volatile solvents via distillation using a Büchi rotary evaporator R-300 followed by residual solvent removal under a high vacuum.Melting points were determined using a Stuart SMP3 apparatus.Infrared spectra were measured using a Perkin-Elmer FT-IR Spectrometer Spectrum Two (Llantrisant, UK) with KBr pellets.NMR spectra were recorded in CDCl 3 , at 300, 400, 500, or 700 MHz (Bruker Advance III, Oxford, UK).Chemical shifts were reported in parts per million (δ) using the residual solvent signals (CDCl 3 : δ H 7.26, δ C 77.16) as the internal standards for the 1 H and 13 C NMR spectra and coupling constants (J) in Hz.Carbon types and structure assignments were determined from DEPT-NMR and two-dimensional experiments (HSQC and HMBC, COSY and NOESY).Mass spectra (ESI-MS) were acquired using an Agilent 1200 ESI/APCI Q-TOF in tandem with an Agilent Mass Q-TOF 6520 (Santa Clara, CA, USA).For the crystal structure determination, data were collected by applying the omega and phi scan method on a Bruker D8 VENTURE PHOTON III-14 diffractometer (Karlsurehe, Germany) using an Incoatec multilayer mirror monochromated with Mo-Kα radiation (λ = 0.71073Å) from a microfocus sealed tube source at 100 K, with a detector resolution of 7.3910 pixels mm −1 .For details of the synthesis and characterization of indoles 1a-m and donor-acceptor cyclopropanes 2a-i, see the Supplementary Materials.

Conclusions
In conclusion, the nickel-catalyzed regio-and diastereoselective formal [3+2] cycloaddition of N-benzylskatoles and donor-acceptor cyclopropanes was developed.The reaction tolerated different monosubstitutions in a series of indoles and cyclopropanes, providing yields of up to 93%, with dr 8.6:1 and complete regioselectivity.Of the synthesized compounds, we were able to determine six new cyclopenta[b]indoles and obtain five previously reported derivatives.Based on the results obtained and the bibliographic precedents, a reaction mechanism was proposed and studied via online reaction monitoring by ESI-MS.All the proposed intermediates were detected experimentally, allowing us to establish the active species of the catalytic cycle and the important role of rac-BINAP in the proposed mechanism.This method enables the rapid construction of the medicinally relevant cyclopenta[b]indole scaffold using an inexpensive nickel catalyst.
the proposed mechanism depicted in Scheme 3, the search began for the species C and D of m/z 1165 that have the same m/z.The formation of intermediates C and D of m/z 1165 was observed after 19 min of reaction time, as shown in Figure 2d.The overall appearance of the spectrum and relative intensities of the ions changed little from 19 to 120 min of reaction in solution, as revealed by continuous ESI-MS monitoring.Intermediates C and D in Figure 2e are isobars, but they have a different bond behavior, as consistent with Scheme 2. Furthermore, the ESI-MS/MS spectrum of the ion of m/z 1165 did not change with time, showing that these intermediates participate in the catalytic process.However, spectra of samples taken after 20 min to 2 h of reaction (Figure 2e) showed that the ion of m/z 1165 dissociates in the gas phase mainly into two pathways.A fragment ion of m/z 265 (2a) was formed from intermediate C, whereas a dissociated ion of m/z 485 could be distinguished from intermediate D, which afforded the observed final product 3aa, as depicted in Figure 2e,f.Furthermore, this study demonstrates the important role of BINAP in the catalytic cycle, since its participation in forming all the key intermediates of the proposed mechanism was detected by ESI-MS. of m/z 779, and [A + Cl − ] + of m/z 715.After 10 min of reaction, B was observed as m/z 1059 (Figure 2b).The isotopic patterns of Ni species were in accordance with the calculated compounds.Then, following the proposed mechanism depicted in Scheme 3, the search began for the species C and D of m/z 1165 that have the same m/z.The formation of intermediates C and D of m/z 1165 was observed after 19 min of reaction time, as shown in Figure 2d.The overall appearance of the spectrum and relative intensities of the ions changed little from 19 to 120 min of reaction in solution, as revealed by continuous ESI-MS monitoring.Intermediates C and D in Figure 2e are isobars, but they have a different bond behavior, as consistent with Scheme 2. Furthermore, the ESI-MS/MS spectrum of the ion of m/z 1165 did not change with time, showing that these intermediates participate in the catalytic process.However, spectra of samples taken after 20 min to 2 h of reaction (Figure 2e) showed that the ion of m/z 1165 dissociates in the gas phase mainly into two pathways.A fragment ion of m/z 265 (2a) was formed from intermediate C, whereas a dissociated ion of m/z 485 could be distinguished from intermediate D, which afforded the observed final product 3aa, as depicted in Figure 2e,f.Furthermore, this study demonstrates the important role of BINAP in the catalytic cycle, since its participation in forming all the key intermediates of the proposed mechanism was detected by ESI-MS.

Table 1 .
Screening of the metal salts 1 .

Table 2 .
Screening of the ligand influence 1 .