Ludmila A.
Oparina
,
Kseniya V.
Belyaeva
,
Nikita A.
Kolyvanov
,
Igor A.
Ushakov
,
Denis N.
Tomilin
,
Lyubov N.
Sobenina
,
Anton V.
Kuzmin
and
Boris A.
Trofimov
*
A. E. Favorsky Irkutsk Institute of Chemistry, Siberian Branch, Russian Academy of Sciences, 1 Favorsky Str., Irkutsk 664033, Russia. E-mail: boris_trofimov@irioch.irk.ru
First published on 17th November 2023
The merging of fundamentally and biologically important heterocyclic structures, such as pyrroles, indoles, and imidazoles, in molecules of dihydropyrrolo[1′,2′:3,4]imidazo[1,2-a]indoles functionalized by E-acylethenyl groups was achieved in water on the platform of commercial 3H-indoles using readily available acylpyrrolylacetylenes as additional building blocks. The yields of the ensembles reached 88% (100 °C, 8 h), which is noticeably better than that produced using conventional solvents (MeCN, MeNO2, DMSO, and CF3CH2OH), wherein the reaction proceeds at a rate of approximately 10 times longer (96 h) with lower yields of the target products (not higher than 80%). The reaction presumably proceeds in a micellar-like microreactor self-assembled in a two-phase aqueous medium that secures favorable mutual orientation of the merging molecules, providing a facile [2+3] concerted cycloaddition to finally form fused polyheterocyclic systems. The performed DFT calculations are in agreement with such mechanistic considerations, particularly those underlying the crucial role of water in the studied cascade process.
![]() | ||
Fig. 1 Representative bioactive molecules bearing imidazo[1,2-a]indole and pyrrolo[1,2-c]imidazole units. |
The merging of both structural motifs into one molecule might lead to a new group of polyheterocyclic compounds with practically important biological properties. As mentioned above, this goal could be reached on the platform of 3H-indoles provided that they would comply with the cyclization with acylpyrrolylacetylenes. As demonstrated in this communication, this expectation came true even, to our delight, in a highly eco-friendly version with water as the reaction medium.
Entry | Solvent | Conditions | Conversion (2a)b % | Yield (3a)b % |
---|---|---|---|---|
a 3H-indole 1a (0.25 mmol), benzoylpyrrolylacetylene 2a (0.25 mmol), solvent 1 mL. b Yields determined by 1H NMR using durene as an internal standard. c The starting pyrrolylacetylene 2a was recovered nearly quantitatively. d ND = not detected. e 10 mol% t-BuOK was added. f 10 mol% DABCO was added. g 10 mol% PdCl2(MeCN)2 was added. h Under MW activation. | ||||
1 | MeOH | 20–25 °C, 96 h | 0c | NDd |
2 | MeOH | 60 °C, 48 h | 20 | 18 |
3 | MeCN | 60 °C, 48 h | 45 | 42 |
4 | MeCN | 60 °C, 72 h | 59 | 52 |
5 | MeCN | 80 °C, 48 h | 79 | 73 |
6 | MeCN | 80 °C, 96 h | 94 | 85 |
7 | MeNO2 | 80 °C, 96 h | 91 | 78 |
8 | DMSO | 80 °C, 96 h | 86 | 66 |
9 | CF3CH2OH | 80 °C, 96 h | 73 | 60 |
10e | MeCN | 80 °C, 40 h | 95 | 7 |
11f | MeCN | 80 °C, 40 h | 70 | 26 |
12g | MeCN | 80 °C, 40 h | 71 | 51 |
13h | MeCN | 100 °C, 5 h | 10 | 9 |
At room temperature (MeOH), the reaction led only to the recovery of the starting materials, the target product 3a was not detected (entry 1). After 48 h at 60 °C in MeOH, the conversion of initial indole 1a and acetylene 2a was approximately 20%, and the yield of compound 3a was 18% (entry 2). In MeCN, under similar conditions, the reaction proceeded more efficiently: the product yield was 42% and the conversion of the starting pyrrolylacetylene 2a was 45% (entry 3). At a longer reaction time (72 h) in the same solvent, the yield increased up to 52% (entry 4). A further temperature increase to 80 °C (96 h) led to a nearly quantitative conversion of pyrrolylacetylene 2a and an 85% (1H NMR) yield of 3a (entry 6). In an attempt to attain better yields, we tried some other solvents, including, DMSO, MeNO2, and CF3CH2OH, but higher yields were not reached (60–78%, entries 7–9). All attempts to optimize the reaction conditions using a catalyst or activator were not successful. The addition of 10 mol% of tert-BuOK had a detrimental effect on the product yield, furnishing just 7% of 3a in MeCN at 80 °C for 40 h (entry 10). The same reaction in the presence of 10 mol% DABCO or PdCl2(MeCN)2 gave no improvement over the catalyst-free reaction (cf. entries 11, 12, and 5), while MW irradiation was ineffective (entry 13). Thus, MeCN, 80 °C, 96 h were found to be the best conditions for the assembly of acylethenylpyrroloimidazoindoles at this stage of the study.
Next, we evaluated the scope and limitations of this cycloaddition (MeCN, 80 °C, 96 h) using a variety of 3H-indoles 1 and acylpyrrolylacetylenes 2 (Scheme 1). The results showed that the yields of pyrroloimidazoindoles 3 were in the range of 9–80% (mostly 40–76%), and were very sensitive to the nature of the substituents in 3H-indole 1 and pyrrolylacetylene 2. In all the cases, the E-configuration of the acylethenyl substituents was observed. The reaction tolerated 3H-indoles bearing different substituents at the C5 position. Indole 1b containing the electron-donating Me-group was more reactive than those with electron-withdrawing groups to give pyrroloimidazoindole 3b in good yield (58%). The moderately electron-withdrawing substituents (F, Cl) in 3H-indoles diminished the yields of 3c (36%) and 3d (35%). The presence of a strong acceptor at position 5 of indole (cyano group in 1e and nitro group in 1f) did not prevent, but significantly reduced, the effectivity of the cycloaddition reaction; the yields of adducts 3e and 3f were from low to modest (9–15%). Apparently, the introduction of the electron-accepting substituents partially withdraws the electron density from the nitrogen atom and makes it less basic/nucleophilic.
A limitation of the method was observed in the case when the Me-group at the C-2 position of 3H-indole was replaced by the phenyl group, possibly due to its electron-accepting nature and steric screening of the position 2 by its ortho-hydrogen. Therefore, the phenyl substituent in 1g impeded the reaction, and no conversion of this substrate was observed and the expected cycloaddition product was detected in only trace amounts.
Benzo[e]indole 1h more easily reacted with pyrrolylacetylene 2a, forming cycloadduct 3h in 72% yield. The effect of the extra benzene ring fused with the indole molecule likely consists of the conjugation extension over the whole system that improves the electron communication, thereby enhancing the basicity/nucleophilicity of the indole nitrogen.
We, then estimated the reaction scope relative to the acetylene substrates. The experiments showed that pyrrolylacetylenes 2b,c containing 2-furoyl and 2-thenoyl groups instead of the benzoyl group proved to be good substrates for this reaction; the yields of the corresponding cycloadducts 3i–l were 70–80%.
The substituents in the pyrrole ring of the starting acetylenes 2, both alkyl and aryl, slowed down the reaction and decreased the yields of the product 3m–o to 40–43%. This can be due to the steric screening of the reacting pyrrole nitrogen by bulky substituents.
Taking into account the growing interest in green approaches in modern organic chemistry and understanding that much more attractive in many regards would be the realization of the reaction under study in water (cheaper, safest, non-toxic, and the most environmentally friendly solvent), we performed the process in the aqueous medium. To our surprise, the results exceeded all our expectations. When we heated a heterogeneous mixture of 3H-indole 1a and pyrrolylacetylene 2a in H2O at 100 °C, soon after the reaction mixture looked pseudo-homogeneous, the insoluble stuff started to precipitate out and after 8 h, almost quantitative conversion of the initial reactants was reached, and the expected cycloadduct 3a was isolated (chromatography, silica gel) in 83% yield.
This approach was further examined and a wide range of acylpyrrolylacetylenes 2 was found to successfully undergo the cycloaddition with 3H-indoles in water under the above conditions (Scheme 2).
Generally, various 3H-indoles and acylpyrrolylacetylenes participate in the reaction to exclusively form E-adducts in higher yields (20–88%) and much faster (8 h instead of 96 h) compared to those observed in MeCN (Table 1 and Scheme 1). Although the same regularities in the substituent effect occurred, in all the cases, the water-based approach provided better yields. For example, 5-cyano- 1e and 5-nitro- 1f 3H-indoles gave products 3e and 3f in 34% and 20% yields (Scheme 2), while in acetonitrile the yields were 15% and 9% (Scheme 1), respectively. Spirocyclic 3H-indoles 1i and 1j with the 3,3-attachment of the cyclohexane ring afforded polycyclic dihydropyrroloimidazoindoles of the original spirocyclic structure 3p–s, in reasonable to good yields (56–84%), which are first representatives of a novel family of fused polyheterocyclic systems. However, similar to the reaction in MeCN, with 2-phenyl-3H-indole 1g in water, only trace amounts of pyrroloimidazoindole 3g were detected in the reaction mixture. The same as in MeCN, retarding influence of the substitution in the pyrrole ring on the cyclization appears in the aqueous medium. So, the bulky alkyl groups (n-Bu and n-Pr) in the pyrrole ring of benzoylpyrrolylacetylene 2g expectedly retarded the reaction, and the yield of the corresponding cycloadduct 3v was only 41%. No side products were detected in the reaction mixtures. Additionally, we established that ethyl 3-(4,5,6,7-tetrahydroindol-2-yl)-2-propynoate (2h) complies well with the aqueous version of the reaction, delivering cycloadduct 3w in 46% yield under the above non-optimized (for this type of electron-deficient acetylenes) conditions. Generally, it is evident that the reaction with every reactant pair 1/2 needs to be separately optimized to achieve the best yield of the target cycloadducts for each case.
The structures of dihydropyrrolo[1′,2′:3,4]imidazo[1,2-a]indoles 3a–w were proven by NMR (1H, 13C, 15N, including 2D correlations, Fig. 2), IR spectroscopy, and mass spectrometry.
In the 2D NOESY spectra, cross-peaks between the signal of the olefin proton (δ 6.70 ppm) and the protons of the C-9 atom of indole moiety and the C-3 atom of furan ring were observed. The vicinal 3JC11a,H coupling constant 8.1 Hz corresponds to a trans position of the olefinic proton with respect to C-11a, that is, the E configuration of the adduct 3i (Fig. 2). Similar NOE effects in the NMR spectra of other pyrrolo[1′,2′:3,4]imidazo[1,2-a]indoles 3 were also observed.
To gain insights into the mechanism of pyrroloimidazoindole 3 formation, we conducted DFT calculations of [2+3] cycloaddition 3H-indole 1a with pyrrolylacetylene 2a. The level of the theory used was B2PLYP-D3BJ(PCM)/Def2-TZVPPD//wB97XD(PCM)/Def2-TZVPP. The results are depicted in Scheme 3.
DFT analysis was already applied by us24,25 to the initial step of the interaction between N-nucleophilic center and electron-deficient acetylenes. According to this approach, the interaction of 1a and 2a was initiated by the formation of H-bonded pre-reaction complex A (Scheme 3). The subsequent attack of the nitrogen lone electron pair of 1a on Cβ atom of the triple bond in acetylene 2a produces the intermediate Bvia a high-lying transition state (TSAB, ΔΔG‡ = 23.8 kcal mol−1) that makes this interaction the rate-determining step of the whole process. The rotation of the pyrrole ring around the C2–C bond in intermediate B results in intermediate C (a conformational isomer of intermediate B). The latter may undergo an intramolecular H-shift to afford a quasi-stable intermediate D through the high-lying transition state TSCD, ΔΔG‡ = 19.9 kcal mol−1. However, a more favorable transformation C → D was initiated by water molecule precoordination so that the oxygen atom was H-bonded with the NH-pyrrole hydrogen, while one of the hydrogen atoms was coordinated with the C-nucleophilic center, intermediate C′ (′ – denotes the associated H2O molecule). This slightly increases the free energy of the system due to the entropy lowering (ΔΔG = 2.9 kcal mol−1). However, the further transformation C′ → D′ (→ D) proceeded significantly easily via the concerted low-lying transition state (TSC′D′, ΔΔG‡ 1.1 kcal mol−1). After 180° reorientation of the pyrrole anionic part (D⇄E), the N-anionic center of E attacks the carbocationic center of the indole moiety to give the adduct F through the corresponding TSEF (ΔΔG‡ = 10.0 kcal mol−1). The most interesting aspect here is how the reorientation of the benzoyl substituent (Z → E isomerization) occurs. The possible candidates for this role could be intermediates E or F. In the former case, predicted activation free energy is substantially high (ΔG‡ = 21.1 kcal mol−1 relative to the energy of the non-reacting system [1a + 2a]) compared with the latter. Therefore, the most probable and thermodynamically favorable transformation is F → 3a. The activation free energy ΔG‡ of the latter is only 2.4 kcal mol−1 with respect to the energy of the nonreacting [1a + 2a] system or 16.4 kcal mol−1 compared with the free energy of the intermediate F.
We also considered the relative probability of the two alternative processes, which may occur when the intermediate E′ attacks the carbocation position 2 of the indole (Fig. S1, ESI†). The first one is the oxygen center for the water attack on the carbocationic center (intermediate F′). The second one is the attacks of this position by the pyrrole N-anionic center (intermediate G′). DFT predicted the former one to be 4.8 kcal mol−1 less favorable compared with the latter option. This is well explained in terms of the HSAB concept, the softer base, a pyrrolate anion, should attack C2 carbon atom more easily than in the case of the harder oxygen base.
The quantum-chemical calculations are in agreement with the experimentally observed decisive effect of the water medium in the synthesis of pyrroloimidazoindole 3. The mechanism of [2+3]-cycloaddition in the aqueous medium could be rationalized as a micellar-like autocatalytic process. In the two-phase water/reactant mixture, the reacting indoles 1 and acetylenes 2 are self-organized in cross-aggregates of 1 (Lewis base)/2 (Lewis acid) types, which should be thermodynamically preferable compared to the homo-aggregates (1/1, 2/2). Consequently, in the cross-aggregates, the effective concentration of the reactants is much higher than that in a solution, and the reactant molecules accept the disposition corresponding to their dipole moments and inner electron distribution to produce the pre-reaction complexes A. All the intermediates represent typical surfactants consisting of hydrophobic (indole) and hydrophilic (acylpyrrolylacetylene) counterparts, both of non-ionic (A) and ionic (B–E) nature and hence accordingly26,27 can catalyze the micelle formation.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3nj05049a |
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2024 |