One-Pot Synthesis of Polynuclear Indole Derivatives by Friedel–Crafts Alkylation of γ-Hydroxybutyrolactams

The Friedel–Crafts reaction of novel 3,5-diarylsubstituted 5-hydroxy-1,5-dihydro-2H-pyrrol-2-ones was used for low cost, one-pot preparation of polycyclic indole derivatives structurally similar to Ergot alkaloids.


Introduction
A vast pool of indole derivatives, both naturally occurring [1] and synthetic [2], have long been a valuable source of potential leads for novel therapeutics, as the majority of indoles are biologically active exhibiting, among others, antitumor [3], anti-Alzheimer's [4], antiviral [5], antituberculosis [6], antimalarial [7] and antibacterial [8] properties. In this context, a subclass of polynuclear indoles, such as cryptolepine [9], isocryptolepine 1 [10,11] and neocryptolepine [12] alkaloids or paullone 2 derivatives [13], stand somewhat alone, because their preparation usually involves time-consuming, multi-step synthetic procedures ( Figure 1). Nevertheless, as such multicyclic skeletons are frequently part of bioactive natural substances, there is significant scientific interest toward them. Thus, Ergot alkaloids, for instance, have been the subject of various studies in recent years [14][15][16][17], including a number of total syntheses of Lysergic acid 3 [18][19][20][21]. Herein we would like to present a convenient route to a principally new type of polycyclic indole moiety 4, which resembles the skeletons of the Ergot alkaloids and may possess some potent bioactivities as well.

Introduction
A vast pool of indole derivatives, both naturally occurring [1] and synthetic [2], have long been a valuable source of potential leads for novel therapeutics, as the majority of indoles are biologically active exhibiting, among others, antitumor [3], anti-Alzheimer's [4], antiviral [5], antituberculosis [6], antimalarial [7] and antibacterial [8] properties. In this context, a subclass of polynuclear indoles, such as cryptolepine [9], isocryptolepine 1 [10,11] and neocryptolepine [12] alkaloids or paullone 2 derivatives [13], stand somewhat alone, because their preparation usually involves time-consuming, multi-step synthetic procedures ( Figure 1). Nevertheless, as such multicyclic skeletons are frequently part of bioactive natural substances, there is significant scientific interest toward them. Thus, Ergot alkaloids, for instance, have been the subject of various studies in recent years [14][15][16][17], including a number of total syntheses of Lysergic acid 3 [18][19][20][21]. Herein we would like to present a convenient route to a principally new type of polycyclic indole moiety 4, which resembles the skeletons of the Ergot alkaloids and may possess some potent bioactivities as well.

Results and Discussion
Our previous research has shown that (indol-3-yl)acetamides [22] and 2-(indol-3yl)acetohydroxamic acids [23] demonstrate significant submicromolar in vitro anti-cancer activity against different types of tumor cell lines, while the latter compounds also showed the rare ability to reverse the differentiation of glioma cells, somewhat similar to the mechanism of astrocytes. Moreover, upon the evaluation of a mouse model, some of Molecules 2023, 28, 3162 2 of 12 those indole hydroxamic acids were able to reduce melanoma xenotransplant growth [24]. Encouraged by such results, we turned our attention to 3,5-diarylsubstituted 5-hydroxy-1,5dihydro-2H-pyrrol-2-one 5, that was earlier found [25] to undergo a Friedel-Crafts reaction with anilines and phenols. We speculated that the interaction of 5 with indoles 6 would lead to previously unknown 4-(indol-3-yl)butyramide 7, which are, basically, cyclic analogs of the above indole acetamides/hydroxamic acids (Scheme 1).

Results and Discussion
Our previous research has shown that (indol-3-yl)acetamides [22] and 2-(indol-3yl)acetohydroxamic acids [23] demonstrate significant submicromolar in vitro anti-cancer activity against different types of tumor cell lines, while the latter compounds also showed the rare ability to reverse the differentiation of glioma cells, somewhat similar to the mechanism of astrocytes. Moreover, upon the evaluation of a mouse model, some of those indole hydroxamic acids were able to reduce melanoma xenotransplant growth [24]. Encouraged by such results, we turned our attention to 3,5-diarylsubstituted 5-hydroxy-1,5dihydro-2H-pyrrol-2-one 5, that was earlier found [25] to undergo a Friedel-Crafts reaction with anilines and phenols. We speculated that the interaction of 5 with indoles 6 would lead to previously unknown 4-(indol-3-yl)butyramide 7, which are, basically, cyclic analogs of the above indole acetamides/hydroxamic acids (Scheme 1).
Keeping that in mind, we initially attempted to reproduce the original conditions of our work [25] using microwave heating for the xylene, only to obtain the desired product at a disappointing 11% yield (Table 1, entry 1). Next, we tried different acidic additives with or without solvents as activators (entries 2-7), but such a treatment produced no effect. However, implementing the conditions of p-Toluenesulfonic acid (TsOH) in ethanol (EtOH) in a microwave reactor showed the first sign of positive feedback, as the yield increased up to 17% (entry 8). Switching to dimethyl sulfoxide (DMSO) and conventional heating in an oil bath further improved the product outcome (37%, entry 9). After varying the amounts TsOH (entries 10-12), we have finally achieved the decent result using just one equivalent of TsOH (55%, entry 13). It should be noted that substituted γ-hydroxy-γ-butyrolactams [26], which the 5 belongs to, act as a convenient source of corresponding N-acyliminium ions upon treatment with acids [27]. The latter undergoes the electrophilic substitution with electron-rich arenes, including indole derivatives [28][29][30].
Keeping that in mind, we initially attempted to reproduce the original conditions of our work [25] using microwave heating for the xylene, only to obtain the desired product at a disappointing 11% yield (Table 1, entry 1). Next, we tried different acidic additives with or without solvents as activators (entries 2-7), but such a treatment produced no effect. However, implementing the conditions of p-Toluenesulfonic acid (TsOH) in ethanol (EtOH) in a microwave reactor showed the first sign of positive feedback, as the yield increased up to 17% (entry 8). Switching to dimethyl sulfoxide (DMSO) and conventional heating in an oil bath further improved the product outcome (37%, entry 9). After varying the amounts TsOH (entries 10-12), we have finally achieved the decent result using just one equivalent of TsOH (55%, entry 13).
This finding allowed us to proceed with the synthesis of a small, focused library of indole butyramide 7, while simultaneously evaluating the scope and limitations of this procedure (Scheme 2). As can be seen in the Scheme 2, the reaction appears to be quite tolerant to different substituents in both indoles 6 and 5-hydroxy-1,5-dihydro-2H-pyrrol-2one 5, providing reliable 37-55% yields across a range of tested reactants.
Once we have learned how to run the Friedel-Crafts reaction between pyrrol-2-one 5 and indole 6, we were ready to start assembling the corresponding polycyclic indoles 4. The key step here was to introduce pyrrol-2-one fragment 5 to the C-4 position of indole 6, in order to then perform the desired intramolecular cyclization. At this point, we speculated that the Knoevenagel condensation of indole-4-carbaldehyde 8 with 2,4diaryl-4-oxobutyronitrile 9a-i should result in 4-((1H-indol-4-yl)methyl)-5-hydroxy-3,5diaryl-1,5-dihydro-2H-pyrrol-2-one 10a-i, which upon further heating will transform to 7,9a-diaryl-2,6,9,9a-tetrahydro-8H-indolo[7,6,5-cd]indol-8-one 4a-i (Scheme 3). To our satisfaction, the desired polycyclic indole 4 were indeed obtained this way in a single step, at generally good yields and, importantly, without the need for isolating the intermediate 10. At the same time, we were able to isolate and characterize indolyl hydroxypyrrolone 10a (R 1 = R 2 = Ph), in agreement with the proposed reaction pathway. Finally, while the reaction turned out to be rather insensitive to the structural features of the starting cyanoketone 9 (Scheme 3), our attempts to use other substituents, R 1 and R 2 , rather than the aromatic, were unfruitful. Thus, a cyanoketone 9 prepared from benzylideneacetone (R 1 = Me, R 2 = Ph) failed to produce both corresponding γ-hydroxybutyrolactam 5 and intermediate 10 under standard conditions. This finding allowed us to proceed with the synthesis of a small, focused library of indole butyramide 7, while simultaneously evaluating the scope and limitations of this procedure (Scheme 2). As can be seen in the scheme 2, the reaction appears to be quite tolerant to different substituents in both indoles 6 and 5-hydroxy-1,5-dihydro-2H-pyrrol-2-one 5, providing reliable 37-55% yields across a range of tested reactants.  Table 1. Screening of reaction conditions for the reaction of 2-phenylindole 6a with 5-hydroxy-3,5diphenyl-1,5-dihydro-2H-pyrrol-2-one 5a.
with or without solvents as activators (entries 2-7), but such a treatment produced no effect. However, implementing the conditions of p-Toluenesulfonic acid (TsOH) in ethanol (EtOH) in a microwave reactor showed the first sign of positive feedback, as the yield increased up to 17% (entry 8). Switching to dimethyl sulfoxide (DMSO) and conventional heating in an oil bath further improved the product outcome (37%, entry 9). After varying the amounts TsOH (entries 10-12), we have finally achieved the decent result using just one equivalent of TsOH (55%, entry 13). Once we have learned how to run the Friedel-Crafts reaction between pyrrol-2-one 5 and indole 6, we were ready to start assembling the corresponding polycyclic indoles 4.
The key step here was to introduce pyrrol-2-one fragment 5 to the C-4 position of indole 6, in order to then perform the desired intramolecular cyclization. At this point, we speculated that the Knoevenagel condensation of indole-4-carbaldehyde 8 with 2,4-diaryl-4oxobutyronitrile 9a-i should result in 4-((1H-indol-4-yl)methyl)-5-hydroxy-3,5-diaryl-1,5dihydro-2H-pyrrol-2-one 10a-i, which upon further heating will transform to 7,9a-diaryl-2,6,9,9a-tetrahydro-8H-indolo[7,6,5-cd]indol-8-one 4a-i (Scheme 3). To our satisfaction, the desired polycyclic indole 4 were indeed obtained this way in a single step, at generally good yields and, importantly, without the need for isolating the intermediate 10. At the same time, we were able to isolate and characterize indolyl hydroxypyrrolone 10a (R 1 = R 2 = Ph), in agreement with the proposed reaction pathway. Finally, while the reaction turned out to be rather insensitive to the structural features of the starting cyanoketone 9 (Scheme 3), our attempts to use other substituents, R 1 and R 2 , rather than the aromatic, were unfruitful. Thus, a cyanoketone 9 prepared from benzylideneacetone (R 1 = Me, R 2 = Ph) failed to produce both corresponding γ-hydroxybutyrolactam 5 and intermediate 10 under standard conditions. The plausible mechanism of this transformation (Scheme 4) should include the Knoevenagel condensation of indole-4-carbaldehyde 8 with 2,4-diaryl-4-oxobutyronitrile 9 to produce the expected adduct 11, which further undergoes proton transfer, leading to acrylonitrile 12. Nucleophilic attack of methoxide anion on the nitrile group, followed by a subsequent nucleophilic attack of the nitrile nitrogen atom on the carbonyl group, results in the formation of intermediate 10, which gives the reactive iminium cation 13 upon heating. The latter attacks the C-3 position of indole to form the desired polynuclear structure 4.
Molecules 2023, 28, 3162 5 of 13 The plausible mechanism of this transformation (Scheme 4) should include the Knoevenagel condensation of indole-4-carbaldehyde 8 with 2,4-diaryl-4-oxobutyronitrile 9 to produce the expected adduct 11, which further undergoes proton transfer, leading to acrylonitrile 12. Nucleophilic attack of methoxide anion on the nitrile group, followed by a subsequent nucleophilic attack of the nitrile nitrogen atom on the carbonyl group, results in the formation of intermediate 10, which gives the reactive iminium cation 13 upon heating. The latter attacks the C-3 position of indole to form the desired polynuclear structure 4.

Scheme 4. Proposed mechanistic rationale for the formation of polycyclic indoles 4.
One may notice that the given one-pot cascade transformation does not require any acid in order to form reactive N-acyliminium cation 13. Meanwhile, the authors of somewhat conceptually similar work [28] generated such ionic species 14 separately by heating the corresponding 5-hydroxy-1H-pyrrol-2(5H)-one with a 6M HCl solution, at 50 °C for 2 h (Scheme 5). Assumingly, as in our case, the imine-type reactive intermediates 13 can be formed under thermal conditions as well, which makes the step of acid addition unnecessary. Additionally, the intramolecular characteristic of the following then Friedel-Crafts cyclization should greatly favor quenching of the reactive N-acyliminium ion, therefore forcing the reaction to reach completion.

Conclusions
An efficient protocol for the synthesis of potentially bioactive indole butyramide 7 was developed. Intramolecular variation of this reaction leads to the formation of polynuclear indole derivatives 4, which are structurally similar to the ergot alkaloids and One may notice that the given one-pot cascade transformation does not require any acid in order to form reactive N-acyliminium cation 13. Meanwhile, the authors of somewhat conceptually similar work [28] generated such ionic species 14 separately by heating the corresponding 5-hydroxy-1H-pyrrol-2(5H)-one with a 6M HCl solution, at 50 • C for 2 h (Scheme 5). Assumingly, as in our case, the imine-type reactive intermediates 13 can be formed under thermal conditions as well, which makes the step of acid addition unnecessary. Additionally, the intramolecular characteristic of the following then Friedel-Crafts cyclization should greatly favor quenching of the reactive N-acyliminium ion, therefore forcing the reaction to reach completion. The plausible mechanism of this transformation (Scheme 4) should include the Knoevenagel condensation of indole-4-carbaldehyde 8 with 2,4-diaryl-4-oxobutyronitrile 9 to produce the expected adduct 11, which further undergoes proton transfer, leading to acrylonitrile 12. Nucleophilic attack of methoxide anion on the nitrile group, followed by a subsequent nucleophilic attack of the nitrile nitrogen atom on the carbonyl group, results in the formation of intermediate 10, which gives the reactive iminium cation 13 upon heating. The latter attacks the C-3 position of indole to form the desired polynuclear structure 4.

Scheme 4. Proposed mechanistic rationale for the formation of polycyclic indoles 4.
One may notice that the given one-pot cascade transformation does not require any acid in order to form reactive N-acyliminium cation 13. Meanwhile, the authors of somewhat conceptually similar work [28] generated such ionic species 14 separately by heating the corresponding 5-hydroxy-1H-pyrrol-2(5H)-one with a 6M HCl solution, at 50 °C for 2 h (Scheme 5). Assumingly, as in our case, the imine-type reactive intermediates 13 can be formed under thermal conditions as well, which makes the step of acid addition unnecessary. Additionally, the intramolecular characteristic of the following then Friedel-Crafts cyclization should greatly favor quenching of the reactive N-acyliminium ion, therefore forcing the reaction to reach completion.

Conclusions
An efficient protocol for the synthesis of potentially bioactive indole butyramide 7 was developed. Intramolecular variation of this reaction leads to the formation of polynuclear indole derivatives 4, which are structurally similar to the ergot alkaloids and

Conclusions
An efficient protocol for the synthesis of potentially bioactive indole butyramide 7 was developed. Intramolecular variation of this reaction leads to the formation of polynuclear indole derivatives 4, which are structurally similar to the ergot alkaloids and therefore represent a new class of potential pharmacophores. Investigation of biological activity of these compounds is currently under way.

General Information
NMR spectra, 1 H and 13 C, were measured in solutions of CDCl 3 or DMSO-d 6 the Bruker AVANCE-III HD instrument (at 400.40 or 100.61 MHz, respectively). Residual solvent signals were used as internal standards in DMSO-d 6 (2.50 ppm for 1 H, and 40.45 ppm for 13 C nuclei) or in CDCl 3 (7.26 ppm for 1 H, and 77.16 ppm for 13 C nuclei). HRMS spectra were measured on the Bruker maXis impact (electrospray ionization in MeCN solutions, employing HCO 2 Na-HCO 2 H for calibration). IR spectra were measured on FT-IR spectrometer Shimadzu IRAffinity-1S, equipped with an ATR sampling module. Spectral data are provided in the Supplementary Materials (Figures S1-S50). Reaction progress, purity of isolated compounds, and R f values were monitored with TLC on Silufol UV-254 plates. Column chromatography was performed using silica gel (32-63 µm, 60 Å pore size). Melting points were measured with the Stuart SMP30 apparatus. All reagents and solvents were purchased from commercial venders, and were used as received.