An Effective Synthesis of Previously Unknown 7-Aryl Substituted Paullones

A straightforward three-step procedure affording a wide range of novel 7-aryl substituted paullone derivatives was developed. This scaffold is structurally similar to 2-(1H-indol-3-yl)acetamides—promising antitumor agents—hence, could be useful for the development of a new class of anticancer drugs.


Introduction
Paullones are an important class of biologically active compounds that were found to be, for instance, efficient inhibitors of various kinases [1][2][3][4][5], including diazepam-binding sites [6] and agonists of GABA-A receptors [6]. Other types of activities shown by this kind of molecules are antitumor [3,7,8], anti-inflammatory [9], anti-parasitic [10], and antiviral [11] ones. Thus, recently, several members of the paullone family ( Figure 1) were subjects of studies evaluating their applicability to treat some common illnesses and conditions, such as Alzheimer's disease [2], osteoporosis [12], and leishmaniasis [13]. Although a large number of paullone derivatives are known, including 7-alkylidene [14] and 7-trifluoromethyl [15] ones, compounds possessing an aryl substituent at C-7 (7, Scheme 1) are somewhat underrepresented and, therefore, understudied. At the same time, in our opinion, such molecules should be of great interest, since they are structurally similar to indolyl-3-arylacetylhydroxamic acids 3, which demonstrated very promising in vitro activity against cancer cells resistant to apoptosis, as well as against multi-drug resistant cell lines [16][17][18]. However, according to the initial in vivo tests, those hydroxamic acids 3, while being readily accessible from indoles 1 and nitroalkanes 2 in polyphosphoric Although a large number of paullone derivatives are known, including 7-alkylidene [14] and 7-trifluoromethyl [15] ones, compounds possessing an aryl substituent at C-7 (7, Scheme 1) are somewhat underrepresented and, therefore, understudied. At the same time, in our opinion, such molecules should be of great interest, since they are structurally similar to indolyl-3-arylacetylhydroxamic acids 3, which demonstrated very promising in vitro activity against cancer cells resistant to apoptosis, as well as against multi-drug resistant cell lines [16][17][18]. However, according to the initial in vivo tests, those hydroxamic acids 3, while being readily accessible from indoles 1 and nitroalkanes 2 in polyphosphoric acid (PPA) (Scheme 1a) [16,19], have shown inefficiency in the remediation of cancer tumor in mice, supposedly, because of a highly unfavorable pharmacokinetic profile.
As it was found, the concentration of 3 in plasma goes below the activity thresho level within one hour, most likely due to the facile glucuronidation of hydroxamic ac functionality [17]. This result prompted us to search for the structural analogs of 3 th would be more resilient towards such metabolic degradation. Our first attempt involvi protection of the hydroxamic acid moiety with the methoxy group proved to be unsu cessful because the obtained derivatives 4 demonstrated only marginal antitumor activ (Scheme 1b) [18]. Another approach relied upon the replacement of hydroxamic acid w primary amide function to obtain compounds 5. To do so, two alternative synthetic pa ways, both based on reactions of indoles 1 with nitrostyrenes 2, were developed. Accor ing to the first one, the reaction takes place in a mixture of PPA (both as a solvent a reagent) and PCl3 (as a reductant) (Scheme 1c) [20]. The second method is a stepwise quence where the intermediate spirane 6 (Scheme 1d) [21,22] formed in H3 PO3 (a reacti media); then, it was reduced to the corresponding amide 5 with sodium borohydri (Scheme 1e) [23]. To our satisfaction, either approach works well and the obtained 2-(1 indol-3-yl)acetamides 5 demonstrated in vitro submicromolar antitumor activity sign cantly exceeding the performance of hydroxamic acid analogs 3. Supposedly, the stru turally related cyclic amides, i.e., paullones 7 bearing aryl substituents R 3 at C-7, wou also possess the desired activity as well as required metabolic stability. Herein, we d close the method for preparation of such compounds.

Results and Discussion
Presently, there are many studies dedicated to the synthesis of various members the paullone family. Most of them employ a rather standard approach of assembling t indole subunit via Fischer reaction [5,7,[24][25][26][27][28][29] of arylhydrazines with a ketone 10 whi in turn, are accessed by the intramolecular condensation of anthranilic acids 8 followi the decarboxylation of the resulting -keto-acids 9 (Scheme 2). Several examples utilizi the lactonization of 2-(о-aminophenyl)indole-3-acetic esters 11 have also been report [30][31][32][33][34]. As it was found, the concentration of 3 in plasma goes below the activity threshold level within one hour, most likely due to the facile glucuronidation of hydroxamic acid functionality [17]. This result prompted us to search for the structural analogs of 3 that would be more resilient towards such metabolic degradation. Our first attempt involving protection of the hydroxamic acid moiety with the methoxy group proved to be unsuccessful because the obtained derivatives 4 demonstrated only marginal antitumor activity (Scheme 1b) [18]. Another approach relied upon the replacement of hydroxamic acid with primary amide function to obtain compounds 5. To do so, two alternative synthetic pathways, both based on reactions of indoles 1 with nitrostyrenes 2, were developed. According to the first one, the reaction takes place in a mixture of PPA (both as a solvent and reagent) and PCl 3 (as a reductant) (Scheme 1c) [20]. The second method is a stepwise sequence where the intermediate spirane 6 (Scheme 1d) [21,22] formed in H 3 PO 3 (a reaction media); then, it was reduced to the corresponding amide 5 with sodium borohydride (Scheme 1e) [23]. To our satisfaction, either approach works well and the obtained 2-(1H-indol-3-yl)acetamides 5 demonstrated in vitro submicromolar antitumor activity significantly exceeding the performance of hydroxamic acid analogs 3. Supposedly, the structurally related cyclic amides, i.e., paullones 7 bearing aryl substituents R 3 at C-7, would also possess the desired activity as well as required metabolic stability. Herein, we disclose the method for preparation of such compounds.

Results and Discussion
Presently, there are many studies dedicated to the synthesis of various members of the paullone family. Most of them employ a rather standard approach of assembling the indole subunit via Fischer reaction [5,7,[24][25][26][27][28][29] of arylhydrazines with a ketone 10 which, in turn, are accessed by the intramolecular condensation of anthranilic acids 8 following the decarboxylation of the resulting β-keto-acids 9 (Scheme 2). Several examples utilizing the lactonization of 2-(o-aminophenyl)indole-3-acetic esters 11 have also been reported [30][31][32][33][34]. We speculated that paullone derivatives 7 could be easily accessed as well via synthetic methodologies previously developed in our laboratories, such as the acetamidation of electron-rich arenes with nitroalkenes [35]. It was shown, however, that the reaction of the corresponding 2-arylindoles with nitrostyrenes 2 leads to the formation of 2-quinolones 12 instead of desired paullones 7 [19] (Scheme 3a). An attempt to introduce an ortho-amine function into the aryl substituent did not afford the paullones either, resulting in indoloquinolines 13 related to the isocryptolepine alkaloid as the only isolable products (Scheme 3b) [36]. This prompted us to explore alternative approaches based on the classical Fischer indolization reaction. To explore this venue, however, one would need an efficient route to suitable ketolactam precursors 16. Originally, we assumed that this could be achieved through the direct cyclization of readily available cyanoketones 14 [37] to keto-lactams 16; however, none of the attempts to carry out this transformation were successful affording only black tar-like residue. Therefore, a two-step protocol, involving initial acid-mediated hydrolysis Scheme 2. Common approaches to synthesis of the paullone derivatives.
We speculated that paullone derivatives 7 could be easily accessed as well via synthetic methodologies previously developed in our laboratories, such as the acetamidation of electron-rich arenes with nitroalkenes [35]. It was shown, however, that the reaction of the corresponding 2-arylindoles with nitrostyrenes 2 leads to the formation of 2-quinolones 12 instead of desired paullones 7 [19] (Scheme 3a). An attempt to introduce an orthoamine function into the aryl substituent did not afford the paullones either, resulting in indoloquinolines 13 related to the isocryptolepine alkaloid as the only isolable products (Scheme 3b) [36]. This prompted us to explore alternative approaches based on the classical Fischer indolization reaction. We speculated that paullone derivatives 7 could be easily accessed as well via synthetic methodologies previously developed in our laboratories, such as the acetamidation of electron-rich arenes with nitroalkenes [35]. It was shown, however, that the reaction of the corresponding 2-arylindoles with nitrostyrenes 2 leads to the formation of 2-quinolones 12 instead of desired paullones 7 [19] (Scheme 3a). An attempt to introduce an ortho-amine function into the aryl substituent did not afford the paullones either, resulting in indoloquinolines 13 related to the isocryptolepine alkaloid as the only isolable products (Scheme 3b) [36]. This prompted us to explore alternative approaches based on the classical Fischer indolization reaction. To explore this venue, however, one would need an efficient route to suitable ketolactam precursors 16. Originally, we assumed that this could be achieved through the direct cyclization of readily available cyanoketones 14 [37] to keto-lactams 16; however, none of the attempts to carry out this transformation were successful affording only black tar-like residue. Therefore, a two-step protocol, involving initial acid-mediated hydrolysis To explore this venue, however, one would need an efficient route to suitable ketolactam precursors 16. Originally, we assumed that this could be achieved through the direct cyclization of readily available cyanoketones 14 [37] to keto-lactams 16; however, none of the attempts to carry out this transformation were successful affording only black tar-like residue. Therefore, a two-step protocol, involving initial acid-mediated hydrolysis of the nitrile function, was implemented (Scheme 4) to give the expected acids 15 with yields ranging from 51 to 78%. Having access to a variety of keto-lactams 16, we were ready to run the Fisher indolization only to find out, disappointingly, that this reaction does not proceed under the commonly used conditions. Therefore, screening studies of the reaction between keto-lactam 16a and phenylhydrazine 17 under various conditions were undertaken (Table 1). Thus, it was found that reaction in PPA containing 80% P2O5 produces only trace amounts of targeted paullone 7aa (entries 1,2) and PPA with an increased concentration of P2O5 gives even worse results (entries 3,4). Similarly, an unsatisfactory outcome was observed with methanesulfonic acid (MsOH) (entry 5). Marginal improvement was achieved when a reaction in the presence of 80% PPA was carried out in organic solvent (EtOAc or EtOH), as the yields were increased to almost 30% (entries 6,7). At this point, it seemed evident that the root of the problem lies in the formation of the key intermediatecorresponding hydrazone-which is problematic in PPA for some reasons. This issue was addressed by carrying out the indolization reaction in two steps (entry 8). First, 3-phenyl-3,4-dihydro-1H-benzo[b]azepine-2,5-dion (16a) and phenylhydrazine (17) were allowed to react in ethanol at room temperature for 30 min in the presence of acetic acid (entry 8.1). Then PPA containing 80% P2 O5 was added to the reaction mixture, which was then heated for another 30 min at 70 °С (entry 8.2). Under these conditions, the desired product 7aa was obtained in 79% yield. Therefore, with the optimized conditions in hand, we proceeded with the synthesis of a series of paullones 7 (Scheme 6). As could be seen, the reaction is rather tolerant to the nature of the aryl substituent both in hydrazine and ketolactam 16 as the yields stays in the range between 67 and 86%. Thus, it was found that reaction in PPA containing 80% P 2 O 5 produces only trace amounts of targeted paullone 7aa (entries 1,2) and PPA with an increased concentration of P 2 O 5 gives even worse results (entries 3,4). Similarly, an unsatisfactory outcome was observed with methanesulfonic acid (MsOH) (entry 5). Marginal improvement was achieved when a reaction in the presence of 80% PPA was carried out in organic solvent (EtOAc or EtOH), as the yields were increased to almost 30% (entries 6,7). At this point, it seemed evident that the root of the problem lies in the formation of the key intermediate-corresponding hydrazone-which is problematic in PPA for some reasons. This issue was addressed by carrying out the indolization reaction in two steps (entry 8). First, 3-phenyl-3,4-dihydro-1Hbenzo[b]azepine-2,5-dion (16a) and phenylhydrazine (17) were allowed to react in ethanol at room temperature for 30 min in the presence of acetic acid (entry 8.1). Then PPA containing 80% P 2 O 5 was added to the reaction mixture, which was then heated for another 30 min at 70 • C (entry 8.2). Under these conditions, the desired product 7aa was obtained in 79% yield. Therefore, with the optimized conditions in hand, we proceeded with the synthesis of a series of paullones 7 (Scheme 6). As could be seen, the reaction is rather tolerant to the nature of the aryl substituent both in hydrazine and keto-lactam 16 as the yields stays in the range between 67 and 86%. Molecules 2023, 28, x FOR PEER REVIEW 6 of 16 Scheme 6. Library of 7-aryl paullones (7) synthesized by the procedure developed herein.

General Information
NMR spectra ( 1 H, 13 C and 19 F) were recorded on a Bruker AVANCE-III HD spectrometer (400, 101, and 376 MHz, respectively) equipped with a BBO probe in CDCl3 or DMSO-d6 using residual solvent signals as the internal standard. Spectral data are provided in the Supplementary Materials (S1-S75). High-resolution mass spectra were registered in MeCN solutions on a Bruker maXis impact (electrospray ionization, using HCO2 Na-HCO2 H for calibration). IR spectra were measured on a FT-IR spectrometer Shimadzu IRAffinity-1 S equipped with an ATR sampling module. Reaction progress, the purity of isolated compounds, and Rf values were assessed by TLC on Silufol UV-254 plates. Column chromatography was performed on silica gel (32-63 μm, 60 Å pore size). Melting points were measured on Stuart SMP30 apparatus. Cyanoketones 14a-d,f,h,j,k [37] and 14e [38] were synthesized according to the previously reported procedures and were identical to those described. All the reagents and solvents were purchased from commercial vendors and used as received.

General Information
NMR spectra ( 1 H, 13 C and 19 F) were recorded on a Bruker AVANCE-III HD spectrometer (400, 101, and 376 MHz, respectively) equipped with a BBO probe in CDCl 3 or DMSO-d 6 using residual solvent signals as the internal standard. Spectral data are provided in the Supplementary Materials ( Figures S1-S75). High-resolution mass spectra were registered in MeCN solutions on a Bruker maXis impact (electrospray ionization, using HCO 2 Na-HCO 2 H for calibration). IR spectra were measured on a FT-IR spectrometer Shimadzu IRAffinity-1 S equipped with an ATR sampling module. Reaction progress, the purity of isolated compounds, and R f values were assessed by TLC on Silufol UV-254 plates. Column chromatography was performed on silica gel (32-63 µm, 60 Å pore size). Melting points were measured on Stuart SMP30 apparatus. Cyanoketones 14a-d,f,h,j,k [37] and 14e [38] were synthesized according to the previously reported procedures and were identical to those described. All the reagents and solvents were purchased from commercial vendors and used as received.

General Procedure for Preparation of 4-(2-Aminophenyl)-2-aryl-4-oxo-butanenitriles 14g,i
These compounds were prepared in analogy to the method described in [37]. A 25 mL round-bottom flask was charged with 3-aryl-2 -aminochalcone (2.00 mmol), acetic acid (120 mg, 0.114 mL, 2.00 mmol), and DMSO (6 mL). The mixture was vigorously stirred, and a solution of KCN (260 mg, 4.00 mmol) in water (0.5 mL) was added dropwise. Then, the reaction vessel was equipped with a reflux condenser, and the mixture was stirred at 50 • C for 1 h, while the reaction progress was monitored by TLC. Upon complete conversion, the mixture was diluted with water (30 mL) and extracted with dichloromethane (4 × 15 mL). Combined organic extracts were washed with water (4 × 15 mL), concentrated in vacuum, and purified by preparative column chromatography on silica gel eluting with 1:4 EtOAc/hexane.

General Procedure for Preparation of 7-Arylpaullones (7aa-ak,ba,ca)
A 5 mL round-bottom flask equipped with magnetic stirring bar was charged with the corresponding keto-lactam 16a-k (0.50 mmol), phenylhydrazine (0.50 mmol), ethanol (0.25 mL), and AcOH (30 mg, 0.029 mL, 0.50 mmol). The reaction mixture was left stirring at room temperature for 30 min. After this, 0.5 g of 80% PPA was added and stirring continued at 70 • C for another 30 min. The reaction mixture was then poured into water (20 mL) and carefully neutralized with concentrated aqueous ammonia solution. The formed precipitate was filtered off, open-air dried, and purified by column chromatography using mixture EtOAc/Hex, followed by recrystallization from EtOAc.