Development of Novel and Efficient Processes for the Synthesis of 5-Amino and 5-Iminoimidazo[1,2-a]imidazoles via Three-Component Reaction Catalyzed by Zirconium(IV) Chloride

General and efficient approaches for the synthesis of new 5-amino and 5-iminoimidazo[1,2-a]imidazoles were developed through a three-component reaction of 1-unsubstituted 2-aminoimidazoles with various aldehydes and isocyanides mediated by zirconium(IV) chloride. The protocols were established considering the reactivity of the starting substrate, which varies depending on the presence of a substituent on the 2-aminoimidazole moiety. A library of new N-fused ring systems with wide structural diversification, novel synthetic, and potential pharmacological interest was obtained in moderate to good yields.


INTRODUCTION
The development of innovative synthetic approaches that allow rapid access to a wide variety of new heterocyclic derivatives is of crucial interest. The use of multicomponent reactions (MCRs) offers significant advantages in organic synthesis, such as the combination of chemical transformations of three or more different starting materials in a one-pot procedure without isolating the intermediates (Dömling and Ugi, 2000;Dömling, 2006;Abdelraheem et al., 2017;Bariwal et al., 2018;Murlykina et al., 2018). In 1998, the groups of Groebke, Blackburn, and Bienayme simultaneously developed a new subclass of MCRs to produce a series of azineand azole-fused aminoimidazoles, using diverse 2-aminoazines or 2-aminoazoles, aldehydes and isocyanides in the presence of Lewis or Brønsted acid catalysts (Bienayme and Bouzid, 1998;Blackburn, 1998;Groebke et al., 1998).
This reaction recently attracted much attention in organic and medicinal chemistry because of its simplicity, efficiency and the ability to generate diverse compound libraries (Devi et al., 2015;Kaur et al., 2016;Shaaban and Abdel-Wahab, 2016;Shaabani and Hooshmand, 2016).
The imidazo-imidazole scaffolds hold a special place in this category, since they are found in compounds showing a wide range of pharmaceutical activities. In particular, they have been described as antifungal (Lila et al., 2003), antithrombotic (Imaeda et al., 2008;Fujimoto et al., 2009), anxiolytic and anti-depressive agents (Han et al., 2005;Tellew and Luo, 2008;Zuev et al., 2010). In addition, they have been reported as androgen receptor agonists and antagonists that are useful in the treatment of a variety of disorders (Zhang et al., 2006). Several synthetic methods based on multistep synthesis have been employed by our group and others to prepare these medicinally important N-fused imidazoles (Langer et al., 2001;Poje and Poje, 2003;Adib et al., 2008;Saima et al., 2012;Chen et al., 2013;Grosse et al., 2015;Castanedo et al., 2016;Loubidi et al., 2016;Kheder and Farghaly, 2017).
On the other hand, compounds containing an imidazo[1,2a]imidazole moiety have been understudied (Compernolle and Toppet, 1986;Kolar and Tisler, 1995;Mas et al., 2002). Only one paper was found in the literature for the preparation of 5-aminoimidazo[1,2-a]imidazole compounds by MCRs starting from 1,5-disubstituted 2aminoimidazoles (Pereshivko et al., 2013). However, the reported protocol gave poor to moderate yields and showed some limitations with 1-unsubstituted 2-aminoimidazole substrates. Hence, there is a need to develop a new, more efficient and general method for the preparation of these derivatives.
In this context, and in continuation of our ongoing search for innovative small molecules, we report herein novel and straightforward approaches for the synthesis of new series of 5-amino and 5-iminoimidazo[1,2-a]imidazoles starting from 1-unsubstituted 2-aminoimidazoles and using zirconium(IV) chloride as catalyst. To the best of our knowledge, this is the first report using 1-unsubstituted 2-aminoimidazoles in MCRs.

RESULTS AND DISCUSSION
In order to find a MCR protocol that can afford an efficient formation of 5-aminoimidazo[1,2-a]imidazoles, we initially performed a model reaction using ethyl 2-aminoimidazole-4carboxylate 1 with benzaldehyde and t-octyl isocyanide under different conditions (Table 1). Two possible regioisomers can be formed in this case 4a or 4a ′ according on which side of the 2-aminoimidazole 1 that reacts.
In our last study, we showed that the Lewis acid zirconium(IV) chloride delivered an efficient catalytic effect for the MCR (Driowya et al., 2018). This catalyst was therefore chosen for the present optimization study.
First, the reaction was carried out in methanol at room temperature in presence of 5 mol% of ZrCl 4 , but unfortunately, no product was observed (entry 1). Poor yields of the expected product 4a or 4a ′ were obtained when the reaction was performed under heating in ethanol with either 5 or 10 mol% of ZrCl 4 (entries 2 and 3). The reaction time was significantly reduced to 10 min under MW irradiation at 140 • C, and the yield was relatively improved to 38% (entry 4). The use of n-BuOH as solvent instead of EtOH under MW irradiation resulted in an improvement of the yield to 62% (entry 5), whereas the use of PEG-400 gave a moderate yield (entry 6). Interestingly, the reaction in PEG-400 under classical heating at 75 • C during 4 h provided a very good yield (entry 7). Moreover, the optimal amount of catalyst (10 mol%) was confirmed, since the use of 5 mol% resulted in a lower yield (entry 8). Finally, replacing ZrCl 4 by other catalysts such as p-TsOH or ZnCl 2 was associated with a significant decrease in the yield of the product 5-aminoimidazo[1,2-a]imidazole 4a or 4a ′ (entries 9 and 10).
Hence, the optimized reaction conditions were found to be ZrCl 4 (10 mol%) as catalyst and PEG-400 as solvent with heating at 75 • C during 4 h. This system was used before for the preparation of imidazo[1,2-a]pyridines (Guchhait and Madaan, 2009).
In order to disclose the structure of the formed regioisomer of this reaction, we carried out a single-crystal X-ray analysis of the product (Figure 1). The results reveal the formation of the regioisomer 4a. The regioselectivity of this reaction can be explained by the presence of intermolecular hydrogen bonds on the ethyl 2-aminoimidazole-4-carboxylate 1, orienting the synthesis toward the formation of only the regioisomer 4a. Moreover, the structure of compound 4a is stabilized by intramolecular N-H. . . O and intermolecular N-H...N interactions as observed in the crystalline structure (see the crystallographic section on the Supplementary Material).
With these reaction conditions in hand, we next explored the scope and limitation of our methodology with diverse 2-aminoimidazoles and isocyanides in the presence of a range of aldehydes as shown in Table 2. A large chemical library of 5-aminoimidazo[1,2-a]imidazole derivatives 4a-f and 5a-i was designed in generally good yields. The reactions proceeded well with both ethyl 2-aminoimidazole-4-carboxylate  1 mmol), t-octyl isocyanide (1.1 mmol), solvent (2 mL: entries 1-5; 1 mL: entries 6-10). b Yield of isolated product. c Starting materials were recovered.
However, a poor yield was obtained when using propionaldehyde (entry 12).
The impact of isocyanide on our reaction was also investigated; the tert-octyl isocyanide and tert-butyl isocyanide gave similar good results, whereas the cyclohexyl isocyanide showed slightly lower yields.
The MCR involving the unsubstituted 2-aminoimidazole 3 using our conditions did not yield any product. This can be explained by the poor reactivity of the starting substrate due to the absence of an electron-withdrawing group.
In order to find another strategy allowing us to synthetize 5aminoimidazo[1,2-a]imidazoles starting from the unsubstituted 2-aminoimidazole 3, we first carried out the condensation of the latter with p-anisaldehyde as a first step model reaction under different conditions ( Table 3). The free amine 3 was prepared from the commercially available 2-aminoimidazole sulfate (see Supplementary Material). Initially, conventional or MW heating of the reaction in different solvents with ZrCl 4 (10 mol%) as catalyst provided the desired imine 6a in poor yields (entries 1-4). The use of ZnCl 2 as catalyst furnished a very low yield (entry 5), while p-TsOH and InCl 3 gave slightly higher yields (entries 6 and 7). Very interestingly, using a reduced catalytic amount of InCl 3 (2 mol% instead of 10 mol%) under conventional or MW heating in ethanol produced a real improvement in terms of yield and reaction time (entries 9 and 11). This may explain the non-reactivity observed in the MCR  ( Table 2, entry 16), because of the instability of the formed imine under such acidic conditions. However, the prolonged reaction time noted with ZrCl 4 (2 mol%) or when no catalyst was used, revealed the influence of InCl 3 on this condensation reaction (entries 8 and 10, respectively). After developing these optimized conditions for the first reaction step, we next focused on finding the best conditions for the second step, which is based on the [4+1] cycloaddition reaction of the formed imine with an isocyanide.
The resulting imine 6a was isolated and reacted with tert-octyl isocyanide under several catalytic conditions ( Table 4). No reaction occurred when using the same conditions as for the condensation step (entry 1). Increasing the catalytic amount of InCl 3 to 10 mol% produced the desired product in poor yields with either conventional or MW heating methods (entries 2 and 3). It is interesting to note that the 5-aminoimidazo[1,2-a]imidazole product formed was unstable and underwent a dehydrogenation reaction in situ to generate the corresponding stable oxidized compound 5-iminoimidazo[1,2-a]imidazole 7a.
We already observed this type of oxidation in our previous work on the synthesis of imidazo[1,2-b]pyrazoles by MCRs (Driowya et al., 2018).
The use of other catalysts such as p-TsOH, ZnCl 2, and ZrCl 4 under microwave irradiation did not produce any significant improvement in the reaction yield (entries 4-6). However, using ZrCl 4 as catalyst and replacing EtOH by n-BuOH as solvent for the reaction showed a slight increase in the yield to 40% (entry 8), which was the optimum result obtained for this reaction. The same conditions used under conventional heating resulted in a significant decrease in the yield. The low yield and the difficulty of this reaction can be explained by the instability of the imine in the acid medium.
With these optimized conditions in hand, we succeeded in achieving the one-pot two-step procedure without isolating the imine by removing EtOH at the end of the first reaction step. The isolated product 7a was obtained with a global yield of 32%. This protocol was next extended to the synthesis of series of 5-iminoimidazo[1,2-a]imidazoles 7a-i starting from the unsubstituted 2-aminoimidazole and exploring a wide range of Frontiers in Chemistry | www.frontiersin.org  aldehydes and isocyanides ( Table 5). As mentioned previously, the 5-aminoimidazo[1,2-a]imidazole products formed were unstable and led directly to the corresponding imine forms 7ai. Despite the low yields obtained, it was nevertheless possible  to produce the targeted compounds, which proved unsuccessful with the methods developed previously or with those cited in the literature. The chemical space of our synthetized compounds was then enlarged, by removing the tert-octyl groups of 5-aminoimidazo[1,2-a]imidazoles 4a-d and 5a-g using TFA as cleavage agent in DCM and giving access to the primary amine compounds 8a-d and 9a-g, respectively, with yields ranging from 26 to 79% (Table 6). In a similar way, the primary imine imidazo[1,2-a]imidazoles 10a, 10b, 10d, 10e, and 10g were prepared in good yields from their corresponding Schiff base derivatives 7a, 7b, 7d, 7e, and 7g by deprotection of tert-octyl groups using the same conditions (Table 7). Unfortunately, the reaction was unsuccessful when the substituent R 1 was 2,4,6-trimethoxyphenyl (entry 3) or ethyl (entry 6).

CONCLUSION
In summary, we have designed highly efficient protocols of multicomponent isocyanide-based reactions catalyzed by zirconium(IV) chloride which offer the synthesis of a library of new functionalized 5-amino and 5-iminoimidazo[1,2a]imidazoles in moderate to good yields. The optimized processes were successively applied to a large number of substituted (or unsubstituted) 2-aminoimidazoles, aldehydes and isocyanides. In addition, the use of inexpensive zirconium(IV) chloride as catalyst delivered an efficient catalytic effect for the reactions with a greater purity of isolated products compared to other catalysts.

DATA AVAILABILITY
All datasets generated for this study are included in the manuscript and/or the Supplementary Files.

AUTHOR CONTRIBUTIONS
MD designed and performed the experiments, then was responsible for writing the manuscript. RG realised the X-ray analysis for the compound 4a. PB and GG directed the project and revised the manuscript. Dömling, A., and Ugi, I. (2000).