Synthesis of Indoloazepinone Scaffolds Using Sequential Photochemical and Photocatalytic Reactions

Abstract

Indoloazepinone scaffolds show promise as anticancer compounds; however, current methods for their synthesis rely on azepinone ring formation from prefunctionalized indoles. Herein, we report an alternative strategy for the rapid synthesis of indoloazepinones from dichloromaleimides and anilines using sequential photoinduced reactions, including a photochemical [5 + 2] cycloaddition and a photoredox-catalyzed dechlorinative indole formation. Construction of the indole core late in the synthesis allowed straightforward diversification of the benzenoid ring with a variety of functional groups.

Indoloazepinones 1 are a subclass of fused azepinones that form the core structure of a range of bioactive molecules (Scheme a). For example, indole-analogues of the pyrroloazepinone natural product hymenialdisine were found to be potent protein kinase and cytokine inhibitors, therefore showing potential for the treatment of Alzheimer’s disease and cancer. − Additionally, indoloazepinones have recently shown promise as crop protectants, due to their antiviral, fungicidal, and insecticidal activities.

1. (a) Structures of Bioactive Fused Indoloazepinones; (b) Previous Synthesis; (c) Proposed Synthesis.

Previous syntheses of indoloazepinone derivatives generally involve construction of the azepinone around an indole core. This utilizes amide coupling of indole-2-carboxylic acid with an ester derivative of β-alanine followed by hydrolysis and cyclodehydration to form an azepinedione ring (Scheme b). While this strategy allows facile synthesis of indoloazepinone analogues through diversification of the azepinone ring, the requirement for prefunctionalized indoles means that modifying the benzenoid ring of the indole is much more challenging. We considered that an alternative strategy involving late-stage construction of the indole core would represent a more versatile approach for the synthesis of diversely functionalized indoloazepinones. Specifically, we reasoned that the indole ring of indoloazepinedione 2 could be readily installed onto dichloro-azepinedione 3 in a two-step process involving amination with aniline 4 and photocatalytic dechlorinative cyclization onto the aromatic ring. , Importantly, 3 is readily accessed by an intramolecular [5 + 2] photocycloaddition of dichloro-maleimide 5, thus providing a concise synthesis of 2 via sequential photoinduced reactions of simple and readily accessible substrates. Herein, we report the successful realization of this strategy, wherein late-stage assembly of the indole core provides access to indoloazepinediones with readily modifiable substitution around the benzenoid ring.

Construction of the dichloro-azepinedione 3 was achieved in two steps from dichloro-maleimide 6 (Scheme ). Alkylation of 6 through a Mitsunobu reaction with 4-penten-1-ol provided alkene-tethered maleimide 5, which was transformed to 3 using the [5 + 2] photocycloaddition methodology developed by the Booker-Milburn group. Subsequent regioselective nucleophilic substitution of the more electrophilic β-keto-chloride of 3 with aniline provided the aminated chloro-azepinedione 7a in 85% yield.

2. Synthesis of the Indoloazepinedione Precursor.

With key intermediate 7a in hand, we moved to the development of the photocatalytic cyclization to construct the indole ring (Table ). Inspired by previous reports of photocatalytic oxidative indole formation from N-aryl enamines, , we initially investigated cyclization of 7a in DMSO-d ₆ using Ir­(ppy)₃ as the photocatalyst. We were pleased to observe a 50% yield of 2a after 24 h of irradiation with blue LEDs, along with 36% of unreacted starting material 7a (entry 1). Addition of sodium acetate to neutralize the HCl byproduct resulted in the full consumption of 7a and increased the yield of 2a to 79% (entry 2). Further improvements were made upon reducing both the reaction time (entry 3) and catalyst loading (entry 4), providing 2a in 89% yield. In addition, the concentration could be increased from 0.05 to 0.1 M without impacting the result (entry 5); however, further increasing the concentration to 0.2 M led to a minor decrease in the yield of 2a (entry 6). Finally, control reactions showed that both the photocatalyst and light were essential for the reaction (entries 7–8).

1. Optimization of the Indole Formation .

Entry	X mol %	Time (h)	[7a] (M)	% 2a	% 7a
1	2.5	24	0.05	50	36
2	2.5	24	0.05	79	0
3	2.5	1	0.05	86	0
4	1.0	1	0.05	89	0
5	1.0	1	0.1	89 (85)	0
6	1.0	1	0.2	80	0
7	1.0	1	0.1	0	89
8	0	1	0.1	0	95

^aYields were determined by ¹H NMR analysis using 1,3,5-trimethoxybenzene as an internal standard.

^bReaction performed without NaOAc.

^cYield of isolated product.

^dReaction carried out in the dark.

We then proceeded to investigate the scope of this indoloazepinedione synthesis by variation of the aniline used for the construction of aminated chloroazepinediones 7 (Scheme ). It is important to note that we unexpectedly observed a small drop in yield of 2a upon changing the solvent from DMSO-d ₆ to anhydrous, nondeuterated DMSO. Since this trend was also observed for several other substrates, we continued the investigations into the scope of the reaction using DMSO-d ₆ as the solvent. A variety of 5-substituted indole derivatives were successfully synthesized (2a–2k). The reaction was found to be relatively insensitive to the electronic effects of the substituent, with electron-donating (2b–2c, 2g) and electron-withdrawing groups (2h–2j) tolerated. Products substituted with synthetically useful halides (2d–2e), boronic esters (2f), ketones (2h), carboxylate esters (2i), and nitriles (2j–2k), were formed in moderate to good yields (54–81%). While the para-methoxy substrate 7c reacted successfully to form 2c, the unprotected phenol derivative 7l failed to cyclize, instead providing only a low yield of 37% of hydrodechlorination product 7l′.

3. Substrate Scope .

^a Yields are of isolated products. The regioisomeric ratios (r.r.) were determined by ¹H NMR analysis.

^b Reaction performed using 1 mmol of 7a.

Having investigated the scope of the reaction for para-substituted anilines, we then turned our attention to substrates with meta-substitution. A mixture of 4- and 6-substituted indole derivatives (2m–2q) were obtained in good overall yields (63–84%) as a mixture of 4- and 6-substituted indole regioisomers. While the regioisomers of alkyl-substituted products 2m–2o were found to be inseparable, it was possible to separate the regioisomers of the methoxy (2p/2p′) and nitrile (2q/2q′) derivatives. Generally, a small preference for the 4-substituted regioisomer was observed, except for the sterically demanding tert-butyl product 2o, which displayed a slight preference for the formation of 6-substituted regioisomer 2o′. These results indicate that sterics have a greater (see 2m vs 2o), albeit still relatively minor, impact on the regioselectivity compared to electronics (see 2p vs 2q). Finally, 7-methoxy-substituted indole 2r was formed in a 68% yield.

Next, we investigated the mechanism of the indole forming reaction from chloro-azepinediones 7. Previous reports of indole synthesis via photoinduced cyclization of N-aryl enamines have been proposed to proceed through both single electron transfer (SET), , energy transfer, and direct photoexcitation pathways. While the essential role of the photocatalyst rules out a direct photoexcitation pathway, activation of 7 by excited state Ir­(ppy)₃ could occur by both single electron reduction or energy transfer. The feasibility of these two pathways was supported by comparison of the reduction potentials and triplet energies of 7a and Ir­(ppy)₃, which show the excited state photocatalyst is a strong enough reductant (E _1/2 [Ir^IV/*Ir^III] = −1.7 V and E _p [7a/7a ^•– ] = −1.1 V vs SCE in MeCN) and has a high enough triplet energy (E _T = 243 kJ/mol for Ir­(ppy)₃ and 236 kJ/mol for 7a) for SET and energy transfer, respectively. Evidence for the SET mechanism was provided by the successful formation of 2a in 73% yield when the photocatalyst was changed from Ir­(ppy)₃ to eosin Y (EY), which is a strong enough excited state reductant (E _1/2 [EY ^+•/*EY] = −1.1 V vs SCE in MeCN/H₂O), but has a significantly lower triplet energy (E _T = 182 kJ/mol) than 7a (Scheme a). This suggests that the reaction likely proceeds through a SET pathway; , however, we cannot rule out the possibility of an energy transfer pathway also operating when Ir­(ppy)₃ is used as the photocatalyst. , Our proposed mechanism for the SET pathway involves single electron reduction of chloroazepinedione 7a by photoexcited Ir­(ppy)₃ to form radical anion 8 and the reduced photocatalyst (Ir^IV) (Scheme b). Elimination of a chloride anion from 8 generates vinylic radical 9, which cyclizes onto the phenyl ring to give cyclohexadienyl radical 10. Finally, the oxidation of 10 to cation 11 by Ir^IV and deprotonation produces indole 2a.

4. (a) Comparison of catalyst efficiencies. (b) Proposed mechanism.

These mechanistic studies also revealed that eosin Y was an effective photocatalyst for indole formation and, therefore, showed potential as a cheaper alternative to Ir­(ppy)₃. To further demonstrate this, several other indoloazepinediones (2b–2d and 2g) were synthesized using eosin Y as the photocatalyst (Scheme a). Although slower reaction rates were observed with eosin Y, products 2b–2d were formed in only slightly reduced yields compared to those obtained using Ir­(ppy)₃. Interestingly, the yield of acetamide 2g increased from 47% to 66% when using eosin Y; however, the origin of this increase in efficiency is currently unclear.

In conclusion, we have developed a method for the synthesis of indoloazepinone scaffolds using sequential photoinduced reactions, including a photochemical intramolecular [5 + 2] cycloaddition of a dichloro-maleimide and photoredox-catalyzed indole formation from aminated chloro-azepinediones. Formation of the indole relies on single-electron-reduction-induced dechlorinative vinylic radical formation and cyclization onto a tethered aniline. Notably, construction of the indole at a late-stage in the synthesis allows facile diversification of the benzenoid ring, with products substituted in the 4-, 5-, 6- and 7-positions readily accessible.

The data underlying this study are available in the published article and its Supporting Information.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.orglett.5c02977.

• Experimental procedures, characterization data, ¹H and ¹³C NMR spectra for all novel compounds, crystal data for 7e, and DFT calculations (PDF)

The authors declare no competing financial interest.

Safety Statement: Caution! Ultraviolet light is damaging to biological tissues. Caution is required when working with the lamp to prevent exposure and protective eyewear must be used at all times.