Introduction

Due to its value to form a large array of medicinal heterocyclic scaffolds, Ugi reaction is an old but hot topic, especially for its challenge on asymmetric catalytic version [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28].

In 2010, Ji group constructed imino-pyrrolidine-thione scaffold directly using the Ugi–Smiles reaction of isocyanides, heterocyclic thiols and gem-dicyano olefins. However, this process needed two steps and could only afford low yield [29]. In 2017, our group utilized porous CeO2 nanorod as catalyst to synthesize the 4-thiazolyl imino-pyrrolidine-thiones [30] with yield up to 98%, in which low-reactive aromatic aldehyde bearing electron-donating group, unactivated benzo[d]oxazole-2-thiol and benzo[d]thiazole-2-thiol being weakly acidic were expanded as substrates firstly and successfully. However, due to the lower reactivity, benzo[d]imidazole-2-thiol and its derivatives could not participate in the Ugi reaction even in the presence of the porous CeO2 nanorod with high catalytic function.

Benzimidazole is the bioisostere of natural nucleotides and plays an important role in the biological field. It is an efficient way to introduce benzimidazole framework into molecules to obtain compounds with great bioactivity. Its low reactivity can be explained from its structure. The order of electron-withdrawing induction effect of N, O, S on imidazole, oxazole and thiazole is O (3.5) > N (3.0) > S (2.6), while the order of their electron-donating conjugation effect is N > O > S. Therefore, their electrophilic ability is N > O > S. That is, the electron cloud density is imidazole > oxazole > thiazole. Benzoxazole and benzothiazole can react under the catalysis of nano-CeO2 but benzimidazole cannot, indicating that the thiol group participated in the reaction in the form of negative ions and then the smaller electron cloud density of benzoxazole and benzothiazole can stabilize the anion more effectively. It means we have to develop new catalyst with higher catalysis for introducing benzimidazole into new compounds with biological potential via Ugi–Smiles reaction.

Tröger’s base (TB, Fig. 1), which was found in 1887 [31], has attracted great attention in molecular recognition [32, 33], bioorganic chemistry [34, 35], supramolecular chemistry [36, 37] and so on. Their distinctive non-twisted V-shape nitrogen-containing eight-member fused skeleton [38, 39] enables TB and its derivatives to capture appropriate molecules efficiently and display great potential as catalyst.

Fig. 1
figure 1

Structure of Tröger’s base

Full size image

To take full advantage of TB framework, herein we used a Schiff base derived from TB-NH2 and (R)-BINOL-(CHO)2 synthesized in our laboratory as efficient catalyst for the four component Ugi–Smiles reaction of cyclohexyl isocyanide, malononitrile, aromatic aldehydes and low-reactive benzo[d]imidazole-2-thiol to afford thioimidazolidinone derivatives (6). It was also used as the efficient ligand to promote the CuI-catalyzed Ullmann reaction of the product of Ugi–Smiles reaction to give benzo [4, 5] imidazo[1,2-a]pyrrolo[3,4-c]quinoline compounds (7).

Results and discussion

General Procedures and Apparatus were shown in the Supporting Information.

Six Schiff base catalysts (1a–1f, Fig. 2) were synthesized in our laboratory, and their \(\left[ \alpha \right]_{\text{D}}^{20}\) and pH values were tested for the catalyst screen [40,41,42].

Fig. 2
figure 2

The Schiff base derived from Tröger’s base and (R)-BINOL

Full size image

The catalytic activity of 1 (10 mol%) to the Ugi–Smiles reaction was evaluated firstly. As expected, the reaction of 2-mercapto-benzothiazole (2a, 0.2 mmol, 1.0 eq.), benzaldehyde (3a, 0.2 mmol, 1.0 eq.), cyclohexyl isocyanide (4, 0.2 mmol, 1.0 eq.) and malononitrile (5, 0.2 mmol, 1.0 eq.) could be catalyzed by 1 successfully at room temperature for 16 h (Table 1, Entry 6–11) to give 6a in CH3CN with 38–54% yield.

Table 1 Optimization of the reaction conditions
figure a
Full size table

The reaction conditions were then optimized based on the yield of 6a. As shown in Table 1, it was found that only 11% yield was obtained without catalyst (Entry 1). The yield was improved but still lower than 30% (21–29%, Entry 2–5) with TB derivatives or (R)-BINOL as catalyst. When the catalyst 1 was added, the yield increased obviously (Entries 6–11), among which 1a gave the best result (54%) due to its most reactive sites, reasonable space structure and suitable steric hindrance. The catalysts bearing two TB frameworks (Entries 6–8) were more effective than that bearing single one (Entries 9–11), which indicated that the combination of TB and (R)-BINOL frameworks improved the catalytic activity literally.

Catalyst 1a showed good catalytic efficiency (Table 1, Entry 6, 54%) in acetonitrile. Unfortunately, it is difficult to separate the pure product in the media. And the reaction did not perform well in pure water (Table 1, Entry 12) because the starting materials were poorly water-soluble. Interestingly, when a 3:1 (v/v) mixture of CH3CN/H2O was used as the reaction medium, the product 6a was formed in a higher yield (Table 1, Entry 18, 81%). Water accelerated the rate of the reaction through hydrophobic aggregation, which effectively gathered the organic reagents at the surface of the heterogeneous catalyst and enabled them interact with each other at a high concentration. And the polar solvent could also stabilize the iminium ion of the Ugi reaction. Besides, it is fortunate to find that the highest yield was obtained after 12 h (Table 1, Entries 19–22), which indicated that the longer reaction time afforded more side reactions.

Increasing the loading of 1a from 5 to 10 mol% led to a significant increase in the yield of 6a from 38 to 86% (Table 1, Entry 25 vs Entry 21), which highlighted the importance of the catalyst loading to the success of the reaction. However, no further increase was observed in the yield when the catalyst loading was increased to 20 mol% (Table 1, Entry 26, 83%). The optimum temperature for the reaction was determined to be 25 °C. Increasing the temperature will lead to an increase of by-products and reduction in the yield of 6a (Table 1, Entries 21, 23 and 24).

Above all, substrates and 10 mol% of 1a was stirred in a solvent mixture of CH3CN and H2O (VCH3CN:VH2O = 3:1) at room temperature for 12 h was the optimum condition.

With the optimized conditions in hand, we proceeded to explore the substrate scope of this transformation with a variety of aromatic aldehydes and substituted benzo[d]imidazole-2-thiol (Table 2). The results of Table 2 showed that the reaction afforded high yields (80–96%) of the desired products with wide substrate scope. It was clear that unfunctionalized weakly acidic 1H-benzo[d]imidazole-2-thiols were used as substrates successfully. There is no regular relationship between the yield and the steric hindrance and electronic effect of substituents on aromatic aldehydes and benzo[d]imidazole-2-thiol.

Table 2 Ugi–Smiles formation of different product
figure c
Full size table

The ee% value of all products was determined (HPLC, OD-H column, MeCN/Hexane = 2:8), and it was regrettable that no chiral product was found. It may be because the cavity of the catalyst is too large and the TB fragment is not chiral, and the chiral center of (R)-BINOL is far away from the substrate, which leads to the weak chiral control ability of catalyst. Therefore, to get a chiral organocatalyst from TB derivatives, it is necessary to design the chiral TB fragment into a backbone.

In order to understand the mechanism, 1H NMR analysis (400 MHz) was applied to monitor the reaction process. The catalyst 1a (0.0173 g, 10 mol%, 0.02 mmol), 2-mercaptobenzimidazole (0.0300 g, 1.0 eq., 0.2 mmol), malononitrile (0.0132 g, 1.0 eq., 0.2 mmol), benzaldehyde (0.0212 g, 1.0 eq., 0.2 mmol), cyclohexyl isocyanide (0.0218 g, 1.0 eq., 0.2 mmol) and acetonitrile (3 mL) were added in a 10-mL round bottom flask and stirred the mixture at room temperature. 0.20 mL of the mixture was taken out, and the solvent was removed under vacuum, and then, the residue was dissolved in CDCl3 for NMR analysis. The reaction process was monitored by repeating the operation at 10-minute intervals.

As the signals in the 0–8 ppm range merged together, the research focused on the changes of signals in the 8–14 ppm range (Fig. 3).

Fig. 3
figure 3

(a) The 1H NMR of 1a. (b) The changes of the signals as the reaction proceeded

Full size image

As the reaction proceeded, the signals of Ha (Scheme 1, δ = 12.8 ppm) and Hb (Scheme 1, δ = 8.7 ppm) gradually decreased and disappeared, while two new signals (δ = 9.0–9.1 ppm and 10.7–10.8 ppm) appeared at 205 min. Furthermore, the new signal at δ 10.7–10.8 ppm tended to shift slightly to the lower field as time went on.

Scheme 1
scheme 1

The tautomerization of 1a

Full size image

From Fig. 3, it can be speculated that 1a underwent deprotonation at the phenolic hydroxyl (Ha) to give oxygen anion. The shift of the signal of Hb indicated that the C=N bond was the catalytic active site which caused the reduction of the density of the surrounding electron cloud. The signal of Hc (δ=8.1 ppm) showed a lowfield shift to 10.7–10.8 ppm, which may be due to the influence of C=N bond and the quaternization of the bridgehead N atoms (Scheme 1, i).

In short, the result indicated that phenolic hydroxyl, C=N bond and the bridgehead N atoms participated in the catalytic process. By connecting the TB skeleton with BINOL, the new catalyst possessed more catalytic active sites.

Based on the above results and literature report [28], we suggested a possible mechanism for the reaction (Scheme 2). Firstly, the catalyst 1a attacked the methylene on malononitrile and captured a proton. After undergoing the Knoevenagel condensation, olefin I was formed. Then cyclohexyl isonitrile attacked I to form intermediate II. Catalyst 1a captured a proton from mercapto on 1H-benzo[d]imidazole-2-thiol; then, the sulfur anion attacked the alkynyl on intermediate II. Subsequently, Smiles rearrangement (formed intermediate III) followed by nucleophilic addition of the amino group onto the cyano group (formed intermediate IV) afforded the final product 6. Both the alkali and active site of catalyst promoted the reaction.

Scheme 2
scheme 2

The proposed probable reaction process

Full size image

The Ullmann reaction [44] is a coupling between aryl halides and copper. The classical Ullmann reaction is limited to electron-deficient aryl halides and requires harsh reaction conditions. Modern variants of the Ullmann reaction employing palladium and nickel have widened the substrate scope and rendered reaction conditions milder. Yields are generally still moderate and catalysts are highly toxic, expensive and depended on unstable and highly toxic organic phosphine ligands, however [45]. Stoichiometric metal experiments showed that, compared to Pd and Ni, low-price Cu is much easier to generate Ar–N, Ar–O, Ar–F and Ar–CF3 by reducing the reaction barrier, but the oxidation addition reaction of Cu is much slower than the first two metals. This process is also a quick step for the Cu catalytic Ullmann reaction. In order to reduce the barrier of Cu oxidation addition process, it is a good method to use suitable ligands.

As Schiff base played an important role in the C–C, C–N or C–O coupling reaction as excellent ligand, catalysts 1a was used as ligands to catalyze Ullmann reaction, and imidazole-containing five-membered fused ring compound 7 was obtained in high yields (Scheme 3) from 6.

Scheme 3
scheme 3

Synthesis of 7 from 6 via Ullmann reaction

Full size image

Firstly, the optimum conditions were explored using 6e as substrate. It is obvious that the reaction could not go smoothly without ligand (Table 3, Entry 1). When 1a was added as ligand, to our delight, the yield was significantly increased (Table 3, Entry 2–7). There may be two factors promoting the reaction: (1) 1 coordinated to CuI by C=N bond and hydroxyl; (2) the cavity of 1a made it beneficial to capture substrate. Then, the inorganic base (Table 3, Entry 2, 8–10) and the amount of catalyst (Table 3, Entry 14–15) was screened. In summary, the optimized reaction was using 1a (10 mol%)-CuI (10 mol%) as co-catalyst, KOH (2.0 eq.) as base in toluene and refluxing for 12 h.

Table 3 The influence of different conditions for synthesis of 7a
Full size table

7 was then synthesized under the optimized reaction condition (Table 4). The ee% value of 7 was also determined (HPLC, OD-H column, MeCN/Hexane = 2:8), and it was found that the product had also no chirality. The possible reasons were the same as 6.

Table 4 Synthesis of substituted products 7
Full size table

Finally, the antitumor activities on human three positive breast cancer cells (MCF-7), human three negative breast cancer cells (MDA-MB-231), human alveolar epithelial cell (A549) and cytotoxicity on human bronchial epithelial cell (HBE) of all products in vitro were evaluated by MTT method. Compounds having inhibition effect (IC50 < 50 μg/mL) are listed in Table 5.

From Fig. 4 and Table 1 in SI (pp 12–13), A2–NH2, 6r, 6y showed effect of great inhibition on MCF-7 and MDA-MB-231 and of very low toxicity on HBE. Besides, 6c and 6 l showed high antitumor activity and specificity on MCF-7. The effect of IC50 of 6c, 6j and 6l on MCF-7 reached nanogram levels (0.27, 0.15 and 0.06 μg/mL, respectively) is particularly noteworthy, which were 49.7-, 9.1- and 65.5-fold of the effect of IC50 on normal cell HBE, respectively. Meanwhile, the effect of IC50 of 6v on MDA-MB-231 also reached nanogram levels (0.55 μg/mL), which was 2.8-fold of the effect of IC50 on HBE. These compounds showed the profound value for further study in new drug development.

Fig. 4
figure 4

The inhibition of products on tumor and normal cells (μg/mL)

Full size image

Due to the poor water solubility, it was difficult for 7 to penetrate into cells and no biological activity was observed.

Conclusion

In summary, the Ugi–Smiles and Ullmann reactions were promoted by a Schiff base (1a) derived from a Tröger’s base-NH2 and the (R)-BINOL-(CHO)2 efficiently by contributing three catalytic sites (phenolic hydroxyl, C=N bond and the bridgehead N atoms) and alkaline. Low-reactive unfunctionalized 1H-benzo[d]imidazole-2-thiols were used as substrates successfully. Correspondingly, thioimidazolidinone derivatives (6) and benzo [4, 5] imidazo[1,2-a]pyrrolo[3,4-c]quinoline compounds (7) were synthesized efficiently under mild condition. Some products showed great inhibition effect on cancer cells (IC50 on MCF-7 and MDA-MB-231 reached nanogram levels) and low toxicity to normal cell, showing their great potential in new drug development.